Main Generator Online Monitoring and Diagnostics

Main Generator On-Line Monitoring
and Diagnostics
SED WARNING:
N
Please read the Export Control
A L
LICE
and License Agreement on the

R I
back cover before removing the

M AT E Wrapping Material. Technical Report
EPRI Powering Progress
R E P O R T S U M M A R Y

and Diagnostics
The monitoring of large electric turbine generators has become a
topic of interest and investigation in the electric utility industry due to
pressure to extend plant life and reduce operating costs. This is
particularly true in the nuclear power industry because the costs and
consequences of losing a unit are more significant than for fossil or
hydro plants. This document provides information about common
generator monitoring, diagnostics, sensors, and instrumentation.
BACKGROUND Improving the amount or type of monitoring is one possible

way to extend generator life. Improved monitoring will allow more foresight in
machine operation through a better knowledge of the machine’s condition and
INTEREST CATEGORIES will avoid major failures before they occur due to early warning of degradation
or problems.
Nuclear plant operations
and maintenance
OBJECTIVES
Plant electrical systems and
equipment • To identify various approaches to large turbine generator monitoring
Engineering and technical • To provide a description of common generator and auxiliary system sensors
support and instrumentation
Maintenance practices • To provide guidance in evaluating system monitoring needs
KEYWORDS APPROACH The project team reviewed pertinent technical literature regarding
the subject of turbine generator monitoring and diagnostics. The information
Monitoring from the literature was condensed and the experiences of generator engineers
Diagnostics at Ontario Hydro were used to add insight. The resulting document was re-
Predictive maintenance viewed by a technical advisory group composed of a select number of experi-
Generator enced generator engineers.
Turbine generator
Maintenance RESULTS A description of the various monitoring philosophies is given,
starting from simple monitoring with alarms, up to elaborate “expert systems”
with on-line trending and diagnostic capabilities. A description of common
generator and auxiliary system sensors and instrumentation is provided. Infor-
mation is furnished on the types of monitoring devices, what they measure or
sense, and the possible interpretations of the readings.
In addition to the common sensors, there are numerous specialized monitoring
devices or systems that provide information about specific generator problems
that can occur. Descriptions are provided on each of the known systems.
EPRI TR-107137s Electric Power Research Institute December 1996

EPRI PERSPECTIVE In today’s economic and technological climate,
utility engineers are tasked with much more than monitoring a few
simple generator parameters. To make decisions regarding generator
monitoring, equipment engineers need an in-depth understanding of
monitoring and diagnostic systems, and there are many different types
of monitoring systems and specialized monitoring devices available.
This EPRI-NMAC effort was designed to provide utility engineers with
information necessary to increase their knowledge and awareness of
existing monitoring systems, as well as to identify potential needs or
benefits of monitoring upgrades.
PROJECT
WO 3814-25
EPRI Project Manager: James Sharkey
Nuclear Power Group
Contractor: Ontario Hydro
For further information on EPRI research programs, call

EPRI Technical Information Specialists, 415/855-2411.
and Diagnostics
TR-107137
Final Report
December 1996
Prepared by
Ontario Hydro
Principal Investigator
G.S. Klempner
Prepared for
Nuclear Maintenance Applications Center
1300 W.T. Harris Boulevard
Charlotte, North Carolina 28262
Operated by
Electric Power Research Institute
3412 Hillview Avenue
Palo Alto, California 94304
EPRI Project Manager
James Sharkey
Nuclear Power Group
DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
THIS REPORT WAS PREPARED BY THE ORGANIZATION(S) NAMED BELOW AS AN ACCOUNT OF
WORK SPONSORED OR COSPONSORED BY THE ELECTRIC POWER RESEARCH INSTITUTE, INC.
(EPRI). NEITHER EPRI, ANY MEMBER OF EPRI, ANY COSPONSOR, THE ORGANIZATION(S) BELOW,
NOR ANY PERSON ACTING ON BEHALF OF ANY OF THEM:
A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH

RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM
DISCLOSED IN THIS REPORT, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY
OWNED RIGHTS, INCLUDING ANY PARTY’S INTELLECTUAL PROPERTY, OR (III) THAT THIS REPORT
IS SUITABLE TO ANY PARTICULAR USER’S CIRCUMSTANCE; OR
(B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING
ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN
ADVISED OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE
OF THIS REPORT OR ANY INFORMATION, APPARATUS, METHOD, PROCESS, OR SIMILAR ITEM
DISCLOSED IN THIS REPORT.
ORGANIZATION(S) THAT PREPARED THIS REPORT
Ontario Hydro
ORDERING INFORMATION
Cost: $10,000
Requests for copies of this report should be directed to the Nuclear Maintenance
Applications Center (NMAC), 1300 W.T. Harris Boulevard, Charlotte, NC 28262,
800/356-7448. There is no charge for reports requested by NMAC member utilities.
Electric Power Research Institute and EPRI are registered service marks of Electric Power
Research Institute, Inc. Copyright © 1996 Electric Power Research Institute, Inc. All rights
reserved.
EPRI Licensed Material
Main Generator On-Line Monitoring and Diagnostics
CONTENTS
1.0 INTRODUCTION ....................................................................................... 1
2.0 GENERATOR MONITORING PHILOSOPHIES ....................................... 3
2.1 Basic Monitoring with Static Alarm Limits ................................... 4
2.2 Dynamic Monitoring with Load Varying Alarm Limits ................. 5
2.3 Artificial Intelligence and Diagnostic Systems ............................ 8
3.0 SENSORS AND INSTRUMENTATION ................................................... 13
3.1 Generator Electrical Parameters ................................................. 13

3.1.1 Generator Output Power .................................................... 13
3.1.2 Generator Reactive Power ................................................ 14
3.1.3 Stator Three Phase Currents ............................................. 14
3.1.4 Generator Terminal Voltage ............................................... 15
3.1.5 Field Current ...................................................................... 15
3.1.6 Field Voltage ...................................................................... 15
3.1.7 Frequency .......................................................................... 15
3.1.8 Volts per Hertz ................................................................... 16
3.1.9 Negative Sequence ........................................................... 16
3.1.10 Generator Main Breaker Status ......................................... 18
3.1.11 Generator Field Breaker Status ......................................... 18
3.2 Stator Core and Frame ................................................................. 18

3.2.1 Core Temperatures ............................................................ 18
3.2.2 Core Clamping Plate Temperature .................................... 19
NMAC Tech Note iii

3.2.3 Core End Flux Screen Temperature .................................. 19

3.2.4 Core Vibration .................................................................... 19
3.2.5 Frame Vibration ................................................................. 20
3.2.6 Liquid Level in Generator Casing ...................................... 20
3.3 Stator Winding .............................................................................. 21

3.3.1 SCW Inlet Temperature ..................................................... 21
3.3.2 Bulk SCW Outlet Temperature ........................................... 21
3.3.3 Conductor Bar Hose Outlet Temperature .......................... 22
3.3.4 Conductor Bar Slot Temperature ....................................... 22
3.3.5 Stator Winding Differential Temperature ............................ 23
3.3.6 SCW Inlet Pressure ........................................................... 24
3.3.7 Stator Winding Differential Pressure .................................. 24
3.3.8 Hydrogen Gas in SCW ...................................................... 24
3.3.9 Stator Winding Ground Alarm ............................................ 25
3.4 Rotor .............................................................................................. 25

3.4.1 Rotor Winding Temperature ............................................... 25
3.4.2 Rotor Winding Ground Alarm ............................................. 27
3.4.3 Shaft Speed ....................................................................... 27
3.4.4 Rotor Vibration ................................................................... 28
3.4.5 Bearing Metal Temperature ............................................... 29
3.4.6 Bearing Inlet Oil Temperature ............................................ 29
3.4.7 Bearing Outlet Oil Temperature ......................................... 29
3.4.8 Hydrogen Seal Metal Temperature .................................... 30
3.4.9 Hydrogen Seal Inlet Oil Temperature ................................. 30
3.4.10 Hydrogen Seal Outlet Oil Temperature .............................. 30
3.4.11 Hydrogen Seal H2/Seal Oil Differential Pressure ............... 30
3.4.12 Rotor Fan Differential Pressure ......................................... 31
3.5 Excitation System ......................................................................... 31

3.5.1 AC Power into the Exciter .................................................. 31
3.5.2 DC Power out of the Exciter .............................................. 31
3.5.3 Main Exciter, Cooling Air Inlet Temperature ....................... 32
3.5.4 Main Exciter, Cooling Air Outlet Temperature .................... 32
3.5.5 Collectors and Brushgear, Cooling Air Inlet Temperature .. 32
3.5.6 Collectors and Brushgear, Cooling Air Outlet
Temperature ....................................................................... 32
3.5.7 Collectors and Brushgear, Cooling Air Filter
Differential Pressure .......................................................... 32
3.6 Hydrogen Cooling System ........................................................... 33

3.6.1 Bulk Hydrogen Supply Pressure ........................................ 33
3.6.2 Generator Casing Hydrogen Gas Pressure ....................... 33
3.6.3 Cold Hydrogen Gas Temperature ...................................... 34
3.6.4 Hot Hydrogen Gas Temperature ........................................ 34
3.6.5 Hydrogen Dew-Point Temperature .................................... 34
3.6.6 Hydrogen Purity ................................................................. 34
3.6.7 Hydrogen Make-Up or Leakage Rate ................................ 35
3.6.8 Raw Service Water Inlet Pressure to Hydrogen Coolers ... 35
iv NMAC Tech Note

3.6.9 Raw Service Water Inlet Temperature to Hydrogen

Coolers .............................................................................. 35
3.6.10 Raw Service Water Outlet Temperature from Hydrogen
Coolers .............................................................................. 36
3.7 Lube Oil System............................................................................ 36

3.7.1 Lube Oil Cooler Inlet Temperature (Bearing Outlets) ......... 36
3.7.2 Lube Oil Cooler Outlet Temperature (Bearing Inlets) ......... 36
3.7.3 Lube Oil Pump, Oil Outlet Pressure ................................... 36
3.7.4 Lube Oil System, Oil Outlet Pressure (Bearing Inlets) ....... 37
3.7.5 Lube Oil Filter, Differential Oil Pressure ............................. 37
3.7.6 Lube Oil Tank Level ........................................................... 37
3.7.7 Lube Oil Flow ..................................................................... 37
3.7.8 Raw Service Water Inlet Pressure to Lube Oil Coolers ..... 37
3.7.9 Raw Service Water Inlet Temperature to Lube Oil
Coolers .............................................................................. 38
3.7.10 Raw Service Water Outlet Temperature from Lube
Oil Coolers ......................................................................... 38
3.8 Seal Oil System ............................................................................. 38

3.8.1 Seal Oil Cooler Inlet Temperature (H2 Seal Outlets) .......... 38
3.8.2 Seal Oil Cooler Outlet Temperature (H2 Seal Inlets) .......... 38
3.8.3 Seal Oil Pump, Oil Outlet Pressure ................................... 39
3.8.4 Seal Oil System, Oil Outlet Pressure (H2 Seal Inlets) ........ 39
3.8.5 Seal Oil Filter, Differential Oil Pressure ............................. 40
3.8.6 Seal Oil Tank Level ............................................................ 40
3.8.7 Seal Oil Flow ...................................................................... 40
3.8.8 Seal Oil Vacuum Tank Pressure ........................................ 40
3.8.9 Raw Service Water Inlet Pressure to Seal Oil Coolers ...... 41
3.8.10 Raw Service Water Inlet Temperature to Seal Oil
Coolers .............................................................................. 41
3.8.11 Raw Service Water Outlet Temperature from Seal
Oil Coolers ......................................................................... 41
3.9 Stator Cooling Water System ...................................................... 41

3.9.1 SCW Cooler, SCW Inlet Temperature
(Stator Winding Outlet) ...................................................... 42
3.9.2 SCW Cooler, SCW Outlet Temperature
(Stator Winding Inlet) ......................................................... 42
3.9.3 SCW Pump, SCW Outlet Pressure .................................... 42
3.9.4 SCW System, SCW Outlet Pressure
(Stator Winding Inlet) ......................................................... 43
3.9.5 SCW Filter (Strainer), SCW Differential Pressure ............. 43
3.9.6 SCW Flow .......................................................................... 44
3.9.7 SCW Conductivity .............................................................. 44
3.9.8 SCW Make-Up Tank Level ................................................. 45
3.9.9 SCW System H2 Detraining Tank Level ............................. 45
3.9.10 Raw Service Water Inlet Pressure to SCW Coolers .......... 45
3.9.11 Raw Service Water Inlet Temperature to SCW Coolers .... 45
3.9.12 Raw Service Water Outlet Temperature from SCW
Coolers .............................................................................. 46
NMAC Tech Note v

3.10 Sensors and Instrumentation ...................................................... 46

3.10.1 Generator Electrical Parameters ....................................... 46
3.10.2 Stator Core and Frame ...................................................... 46
3.10.3 Stator Winding ................................................................... 47
3.10.4 Rotor .................................................................................. 47
3.10.5 Excitation System .............................................................. 48
3.10.6 H2 Cooling System ............................................................. 48
3.10.7 Lube Oil System ................................................................ 49
3.10.8 Seal Oil System ................................................................. 49
3.10.9 Stator Cooling Water System ............................................ 50
4.0 SPECIALIZED MONITORING DEVICES ............................................... 51
4.1 Stator Winding Partial Discharge Monitoring ............................ 51

4.1.1 RF Monitoring .................................................................... 52
4.1.2 Capacitive Coupling ........................................................... 53
4.1.3 Stator Slot Coupler ............................................................ 54
4.2 Rotor Winding Shorted Turns Detection .................................... 57
4.3 Stator Endwinding Vibration Monitoring .................................... 60
4.4 Shaft Voltage and Current Monitoring ........................................ 60
4.5 Generator Condition Monitor/Tagging Compounds .................. 61
4.6 Torsional Vibration Monitoring .................................................... 63
4.7 Stator Cooling Water Chemistry Monitoring .............................. 63

4.7.1 SCW Oxygen Content ....................................................... 63
4.7.2 SCW H2 Content ................................................................ 64
4.7.3 Copper/Iron Content .......................................................... 64
4.7.4 SCW pH Value ................................................................... 64
5.0 DECIDING ON A MONITORING SYSTEM ............................................. 65
5.1 Need ............................................................................................... 65
5.2 Economics ..................................................................................... 65
5.3 Risk ................................................................................................ 67
5.4 System Maintenance and Obsolescence ................................... 67
6.0 REFERENCES ........................................................................................ 69
vi NMAC Tech Note

LIST OF FIGURES
Figure 1 Dynamic Temperature Monitoring .................................................... 7
Figure 2 Stator Winding Hierarchy ............................................................... 12
Figure 3 Large Generator Negative Sequence Capability ........................... 17
Figure 4 Determination of Rotor Winding Hot Spot Multiplier ...................... 26
Figure 5 RF Monitor ..................................................................................... 52
Figure 6 SSC Position in Slot, Bore View .................................................... 55
Figure 7 Cross-Section of Stator Slot Containing an SSC ........................... 56
Figure 8 Rotor Winding Shorted Turn .......................................................... 57
Figure 9 Rotor Winding Shorted Turns Detector, Typical Flux Probe

Signal On-Load .............................................................................. 59
Figure 10 Rotor Winding Shorted Turns Detector, Typical Signal

with Shorted Turn Indicated ........................................................... 59
NMAC Tech Note vii

1
INTRODUCTION
The monitoring of large electrical generators has become a high priority in

the electric utility industry due to the pressure to extend unit life and reduce
operating costs. This is particularly true for the nuclear portion of the indus-
try because the costs and consequences of losing a unit are more significant
than for fossil, hydraulic, or other types of plants. Improving the amount or
type of generator monitoring is a possible way to help extend generator life.
Better monitoring can help avoid major failures before they happen because
of earlier warning, and allows additional flexibility in operation by operators
being more familiar with the performance of the machine.
There are numerous types of generator monitoring systems and approaches
to generator monitoring. This document is an attempt to identify the different
approaches to generator monitoring and to provide utilities with options for
the level of monitoring they need for a particular station and its role in their
power system.
In addition, a description of the various monitoring devices presently avail-
able is provided, along with how they are most often used and how they
could be used in the various types of monitoring systems.
Further, some guidance is provided as to how to evaluate the need for
generator monitoring. This means making decisions, such as:
• Which type of system should be used?
• How extensively should each monitoring device or sensor should be used?
• Is the monitoring cost effective?
NMAC Tech Note 1

2
GENERATOR MONITORING
PHILOSOPHIES
Generator on-line monitoring and diagnostics cover a wide range of ap-

proaches. These start from minimal monitoring with few sensors and simple
alarms up to elaborate expert systems with extensive diagnostic capability.
The level of system sophistication for the most basic monitoring involves the
operator simply keeping track of generator operation by periodically check-
ing specified operating parameters on various gauges and indicators and
relying on those sensors that are connected to an alarm system to give warn-
ing when a predetermined high limit is reached.
As hardware and software capability in data loggers and computers has
progressed, the ability to provide better monitoring has increased dramati-
cally in past years. Computers can now handle many more sensors at one
time and can scan them for information far more often. Faster CPUs, in-
creased memory capability, and more sophisticated data acquisition hard-
ware are part of the reason for the progress in monitoring.
In addition to the hardware, software capability and flexibility have also
grown exponentially. Improved software has provided the industry with the
ability to take simple sensor inputs and derive very detailed information
about machine performance from them. Sensor inputs can now be combined
to provide artificial indications of problems that would otherwise not be
available. Computer modeling of various generator components predicts
how they will react during load changes and operating events. The use of
NMAC Tech Note 3

such techniques allows closer tracking of sensors and the ability to diagnose
problems at a much earlier stage in their development.
The information gathered from sensor readings by the monitoring systems
can be stored as an archived history about the performance of a machine.
This can be used for long term trending and maintenance management of the
equipment.
In addition, the use of graphical user interfaces has allowed much more
meaningful presentation of the data collected so that operators can interpret
the information faster and more accurately. Readings are now presented in
both numeric and graphical form. Short term trends are used to compare
various operating parameters while operators attempt to diagnose problems
based on such things as temperature rises with load increase and so on.
Because of the significant advances in computers, generator monitoring has
become very sophisticated. Along with this sophistication comes a high price
tag to install a complex “expert system.” However, in some cases, a high level
of sophistication is not always necessary. A utility must assess its needs based
on the equipment under consideration. A large 1000 megavolt amperes (MVA)
nuclear unit might warrant the installation of an expert system, while a 100
MVA co-generation (co-gen) unit might need a more basic level of monitoring.
When deciding on the correct approach for generator monitoring, it is always a
good idea to first understand what the needs are. Then, to decide, the monitor-
ing options that can be employed must be understood and the cost to provide
them known. The following is a brief description of some of the types of moni-
toring approaches that can be used on large generators in operation today.
Basic monitoring implies that the generator itself has very few installed
2.1 Basic Monitoring
with Static sensors and that only the most necessary and basic operating parameters are
Alarm Limits selected for permanent monitoring. Alarms are usually set at predetermined
high limits.
Nearly all generators have their main electrical parameters connected to a
unit computer so that the operators are aware of the load point of the genera-
tor and where they are operating in relation to the limits of the generator. The
main electrical parameters would include such things as:
• Megawatts (MW)
• Volt-amps reactance (MVars)
• Stator phase currents
• Terminal voltage
• Frequency
• Field current
• Field voltage
4 NMAC Tech Note

All of these have operating limits that, if exceeded, can cause damage to one
or more of the generator components.
In addition to the electrical parameters, there are other operating values that
must be monitored so that the generator operating limits are adhered to.
Some of these include:
• Shaft speed
• Hydrogen gas pressure and temperature
• Lube and seal oil temperatures
• Stator cooling water temperature
• Pressure and conductivity
• Bearing vibration
• Raw service water temperature
These are some of the more critical parameters, and all of them tell the
operator something about the condition of the generator or one of its compo-
nents. In addition, all of them have specific operating limits that, when they
are exceeded, have certain consequences.
Not all parameters have the same level of priority. For example, exceeding
shaft speed limits by 10% while operating at steady load will have far greater
consequences in terms of machine damage than exceeding a component
temperature limit by 10% or a bearing vibration by 10%.
Another point is that not all parameters have to be monitored to guarantee
proper operation of the generator. For example, all of the stator cooling water
(SCW) outlet temperatures being normal during on-load condition tells an
operator that there must be water flow in the system. Therefore, it is not
absolutely necessary to monitor the SCW flow itself. Monitoring the stator
cooling water outlet temperature is sufficient to safely operate the generator
and detect when problems occur.
There is also a wide range in the number of sensors installed by the various
vendors in their generators over the years. There are some machines with no
core or stator winding thermocouples (TCs) or resistance temperature detectors
(RTDs) installed and some that have as many as two dozen core TCs and an
RTD in every stator winding slot and on every stator winding outlet hose. The
variation is extensive and is generally dependent on the vendor, but the utility
can request additional monitoring sensors to be installed during manufacture
if they are willing to pay more to get them. Most vendors will oblige.
Whatever sensors are installed and connected to the monitoring system, it

2.2 Dynamic
Monitoring with always a good idea to get the most out of what you do monitor. Dynamic
Load Varying monitoring of sensors is an effective way to do this. Dynamic monitoring is a
Alarm Limits term that simply means alarm limits that change in relation to the generator
NMAC Tech Note 5

load point, rather than waiting for an absolute high alarm limit to be reached.
Relying on static high limits can sometimes mean that a problem
has progressed too far for corrective action by the time the operator is noti-
fied by the alarm limit being exceeded.
The premise behind dynamic monitoring is to mathematically predict what a
particular sensor or group of sensors should be reading at any operating point
and compare it with the actual sensor reading. This way, the difference be-
tween the two can be closely monitored, and if the deviation is more than a
previously determined limit, it can be brought to the attention of the operator.
The obvious advantage is that much earlier warning can be obtained to look
for long-term problem trending, as well as more immediate failure modes.
To do this, a mathematical model of the generator parameter being moni-
tored must be available, and then it must be customized to the machine
being monitored. This mathematical model then becomes an artificial sensor
or an indicator of a problem. The following is a brief description of how this
type of sensor is constructed.
One of the best examples of an artificial indicator built from sensor readings
into a mathematical model is that of temperature measurement for a water-
cooled stator winding hose outlet.
To build this indicator in its simplest form, you must look at what affects the
stator winding temperature during all modes of operation, but specifically
when the generator is connected to the system and loaded. In the case of the
stator winding hose outlet sensor, you are not concerned with the stator
winding temperature when the generator is off-line because no current flows
in the winding then. Fault current could flow when the generator is on open
circuit and a failure of the ground wall insulation occurs, but this is a case
where the generator ground fault relay protection comes into play, and the
stator winding temperature monitoring is a secondary issue. The main
concern is the temperature of the stator winding when the machine is on-line
and stator current is flowing in the winding.
To begin, the stator winding hose outlet temperature is at least that of the
water inlet temperature. Therefore, the first component of a stator bar hose
outlet temperature, Tout, is the temperature of the cooling water in, Tin. (For
direct hydrogen-cooled stator windings, the coolant inlet temperature is
derived from the cold hydrogen gas temperature.)
Stator bar temperature increases as electrical current flows in the copper of
the winding. The relationship of temperature to electrical current is well
known as T∝I2. Therefore, if the generator is at full load where the stator
current is theoretically at its maximum (Iref), then the temperature of the
stator bar hose outlets will be some temperature above the cooling water
inlet temperature. The difference between the cooling water inlet and outlet
temperatures is the temperature rise, dTref, at this reference load, due to the
heat input from the stator bar I2R losses.
6 NMAC Tech Note

The temperature difference between Tout and Tin obviously changes as the
generator loading (operating stator current, Is) is increased and decreased.
Applying the relationship T∝I2, you can use Is and Iref in the form (Is / Iref )2 to
account for generator load changes. Therefore, the basic formula to calculate
stator winding hose outlet temperatures can be written as:
Tout = Tin + dTref * (Is / Iref )2
In this relationship, the portion of the function (Is / Iref)2 is equal to one, as it
should be, when fingerprinting of the stator winding temperatures is done at
the reference load. As Is becomes lower, at lower loads, the temperature
calculated for Tout decreases proportionally [1].
Using the above formula, the difference between the measured reading and
the calculated value can be closely monitored. An alarm value (for example,
5˚C) can then be added to the calculated value to produce the dynamic alarm
limit as follows.
Talarm = Tout + 5˚C
If the deviation is more than the calculated alarm limit, Talarm, as shown in
Figure 1, it is then brought to the attention of the operator.
deg C
100
Calculated alarm value,

T alarm
80
Temperature
60
Actual temperature
40 5 deg C measured by sensor
20
0
Time
Figure 1
Dynamic Temperature Monitoring
NMAC Tech Note 7

It should be noted that the algorithms above are in their simplest form. Other
factors must be included for complete accuracy and these are generally
known by the OEM. For example, the stator bar expected temperature
calculation can also be enhanced to include a factor to allow for variable
coolant flow in water cooled stator windings. Likewise, for direct hydrogen
cooled stator windings, this model can be enhanced to incorporate changes
in hydrogen gas pressure. When implementing these types of models, the
utility should consult the OEM before implementation.
Now consider the direct hydrogen cooled stator winding. If the gas pressure
factor is not included, the model allows diagnosis of overheating due to
drops in casing hydrogen pressure. If the factor is included, then the model
will always account for casing pressure changes and never be able to predict
temperature rises on pressure variation because the model is now dynamic
on pressure as well as coolant inlet temperature and stator current. There-
fore, to determine if one of the model variables is affecting the outlet tem-
perature, you need to either depend on separate indicators to advise the
operator that a limit has been exceeded or use combinations of the basic
model with each of the variables sequentially removed.
The distinct advantage of using this type of indicator in conjunction with
direct sensor readings is the capability to predict expected values over the
entire load and power factor range of the generator and compare them to the
actual readings. This allows a much improved and closer degree of monitor-
ing on specific generator components rather than simply relying on a maxi-
mum limit before an alarm is incurred. Using this dynamic monitoring
method, you can look for deviations of only a few degrees above normal (for
temperature relationships) at any load and be provided with much faster
warning of impending problems in the generator, long before measured
parameters get anywhere near their absolute limits.
Monitoring systems with diagnostic capability that use artificial intelligence

2.3 Artificial
Intelligence and software are referred to as expert systems. The purpose of an expert system for
Diagnostic Systems monitoring and running diagnostics on large steam turbine-driven genera-
tors is to collect, analyze, and interpret sensor information from the genera-
tor and auxiliary systems and to provide early diagnosis of developing
problems in the generator and its associated systems. The expert system
should provide an easily understandable description of the suspected prob-
lem and recommendations to correct it or to bring the unit to a safe operating
condition in a timely and appropriate manner.
The main advantage of an expert system is its ability to look at all available
sensor data in real time, correlate it as an “expert” would, and continuously
update the diagnosis, based on changing sensor readings. This allows opera-
tors to react quickly at the onset of the majority of generator problems that
can be experienced during operation and avert major failures. In addition, an
expert system gives station maintenance engineers a tool to closely monitor
8 NMAC Tech Note

and log the performance of the generator and make better maintenance
decisions from the data collected during operation.
There are a variety of types of expert systems in use today on large rotating
machines. Their deterministic capabilities rely on such methods as rule-based
systems, pattern recognition, neural networks, Bayesian belief networks, and
so on. Regardless of the type of expert system implemented, all of them have
a number of common elements:
• Knowledge base containing the equipment facts, component relation-
ships, mathematical models, etc.
• Data acquisition hardware
• Software for sensor inputs
• Artificial intelligence (AI) software more commonly referred to as an
inference engine to perform the reasoning function between the knowl-
edge base and sensor inputs
• Graphical user interface to allow the operator to interact with the system
• Installation software to allow changes and updates to the system by the
experts
• Simulator for off line testing and training in some cases
Within the various elements of expert systems, some interesting and novel
techniques have been developed in recent years to provide accurate analysis
of impending problems. These include mathematical modeling techniques
used for logical and probabilistic determination of large generator problems.
Additionally, methods have been developed for combining sensor inputs to
create mathematical indicators of problems, and techniques have been
developed for dynamic tracking of problem indicators over the full power
factor and load range of the generator.
An expert system generally consists of a computer for monitoring and
processing data, external data acquisition hardware for collection of the
generator and auxiliary systems sensor inputs, and the software that forms
the basis of the expert system.
The computer is usually a fairly powerful type of machine such as a worksta-
tion. This is mainly because real time expert systems have large memory
requirements and need fast CPU speeds. However, it is now possible to put
even the larger expert monitoring systems on a PC.
The data acquisition system is used to collect raw sensor data from the
generator and auxiliary systems. It can consist of a stand-alone data logger or
the equivalent, or the computer data acquisition system for the existing unit.
The number of sensors monitored varies, depending on the particular gen-
erator and how extensively it is instrumented. The expected readings of the
monitored sensors are determined during installation and configuration of
the expert system by “heat run” tests on the generator at various loads and
NMAC Tech Note 9

power factors to “fingerprint” the expected machine behavior. The finger-

print data is used to produce scaling factors for specific formulas developed
to track sensor inputs over the entire load and power factor range (that is,
dynamic tracking) and to set maximum sensor limits according to insulation
class, machine rating, and other machine-specific parameters. Such formulas
include stator winding temperature as a function of coolant inlet temperature
and stator phase current as previously described, and stator core tempera-
tures as a function of MW, MVars, cold hydrogen gas temperature and
pressure, and so on.
Examples of sensors are:
• Direct temperature readings from TCs and RTDs
• Pressure readings
• Voltage measurements
• Current measurements
• Equipment status (for example, breaker open/closed, pump on/off, tank
level high/low, and so on)
These are the instruments that are usually hardwired directly to the data
acquisition system. The readings are used in their raw form both in terms of
the measurement value and units.
Within the software is found the AI software, the knowledge base, installa-
tion software, simulation software, and third-party software for such things
as the graphical user interface.
Within the knowledge base are the general and specific generator information
on problems and indicators of problems, which the AI software must process.
The knowledge base is generally a refined database that attempts to incorpo-
rate industry-wide common information about generator operation and
troubleshooting. In addition, the specific information on the generator being
monitored and its auxiliary systems are structurally mapped into the database.
The knowledge base consists of the possible generator and auxiliary system
problems. Attached to the problem network or table are the indicators, which
consist of as many sensors and problem indicators as available from the
installed generator instrumentation. The problem indicators are, in effect, the
sensor inputs or combinations of sensor inputs that convey the information
that some operating parameter or limit has been exceeded and that a real
problem is occurring. This deviation from the normal is determined by the AI
software as it processes the differences between actual sensor readings and
expected readings.
The expert system software uses the sensor inputs to look for deviations in
readings that indicate generator problems, and the software reports the
relevant information to the display monitor for operator interaction. Dy-
namic monitoring lends itself to the expert system application extremely
well. The example given in the section on dynamic monitoring has shown
10 NMAC Tech Note

how the individual hose outlet temperature from one stator bar is calculated
so that it can be compared to the actual sensor reading. However, each stator
bar is also associated with a particular stator winding parallel and phase.
This is a very important to consider because some stator winding problems
are not simply related to only one stator bar, but to a particular parallel,
phase or the whole winding. An expert system can handle this complexity
extremely well as shown below.
On a two-pole, two parallel-path stator winding, for example, one of the
phase parallels could be affected if a parallel connector on the collector end
ring bus were open circuited. This would cause all the current on the associ-
ated phase to divert from the open parallel to the other parallel. In effect,
there would be no current flow in one parallel, and the other parallel would
be attempting to carry full phase current and subsequently overheat. You
would, therefore, expect to see all the bars associated with the overheated
parallel register a high temperature alarm and the other parallel go down in
temperature because no stator current is flowing.
Suppose that for the same winding configuration one phase loses all cooling
water flow due to plugging of the coolant path, sometimes made possible by
the configuration of the water delivery system to the winding. You would then
expect to see all the bars associated with this phase register a high temperature
alarm because the phase is still carrying current but is not being cooled.
Finally, consider the whole winding in the case where the cooling water flow
is greatly reduced but is still flowing. The temperature monitoring indicators
for the stator bars will see normal stator current and inlet water temperature,
but reduced flow. Therefore, all the stator bar hose outlet temperatures
should be in alarm because all will be reading higher than the calculated
expected value for each bar.
The point of the last three examples is that you do not want to have every
stator bar hose outlet temperature in alarm when the root problem is not
related to the bars themselves. Therefore, a further method is required to
establish that the problem is not with the stator bars but rather is rooted in
the connectors or the cooling water delivery system, and so on.
To do this, you can use the stator winding diagram to form a sensor network
and map out which bars belong to each of the three phases and subsequently
which of these are in each parallel as shown in Figure 2 [1]. Using the stator
bar temperature models, the expert system can then reason that the problem
is related to, for example, the red phase only because the winding mapping
tells it that only the bars in the red phase are overheating. Therefore, the
expert system would report simply that the red phase is overheating, rather
than all the stator bars in that phase. The graphing abilities of the computer
can then be used to track the temperature of the affected phase compared to
the other phases and to load changes.
NMAC Tech Note 11

Whole Stator
Winding
Red Phase White Phase Blue Phase
Red Red White White Blue Blue

Parallel #1 Parallel #2 Parallel #1 Parallel #2 Parallel #1 Parallel #2
Top Top Top Top

Bar #1 Bar #2 Bar #3 Bar #n
Bottom Bottom Bottom Bottom

Bar #24 Bar #25 Bar #26 Bar #m
Figure 2
Stator Winding Hierarchy
The degree of overheating can be further determined by using the individual

bar temperature models to give the average temperature of the bars in the
affected phase or the hottest bar in the affected phase, and so on. Handling
the temperature reporting is discretionary and is simply a matter of choice in
this case. Hence, there is additional flexibility built in with this type of
approach.
12 NMAC Tech Note

3
SENSORS AND INSTRUMENTATION
Regardless of the type of monitoring system employed on a generator, all

systems require sensor inputs to provide the information about the various
machine components and operating parameters. This section lists the basic
sensors that are installed on generators in service today. It provides a brief
description of the type of sensor or instrument, the units in which the output
is given, what each detects, and how they are used.
In addition, some sensor information can be combined with other sensor
indications to form an indicator of a problem or existing generator condition.
This section also describes how certain relevant sensors can be used to better
advantage in this manner.
When discussing generators in general, the machine is usually broken down
into components and sub-systems that have common elements. The same is
true when discussing generator monitoring.
3.1.1 Generator Output Power

3.1 Generator
Electrical Generator output power is the real megawatt (MW) power output from the
Parameters generator. It is principally determined by the steam input to the turbine, but
it is a function of the stator terminal voltage (Vt) and stator current (Ia), and
the generator power factor (pf).
MW = MVA * pf
where: MVA = √3 * Vt * Ia
NMAC Tech Note 13

These parameters are measured and monitored to keep track of the load point
of the machine and allow the operator to control the operation of the generator.
MW overload on the generator is the main concern in terms of monitoring.
MW overload means that the stator current limit has probably been exceeded
and will affect the condition of the stator winding. The stator terminal voltage
might also have been exceeded during overload, depending on the main
transformer tap settings. Excessive terminal voltage will affect core heating.
Transient MW events from the system or internally in the machine also show
up as transients in the stator current or terminal voltage.
3.1.2 Generator Reactive Power

Generator reactive power is the volt-amps reactance (MVar) output from the
generator. It is principally determined by the field current input to the rotor,
but it is also a function of the stator terminal voltage and current. MVars are
also measured and monitored to keep track of the load point of the machine
and allow the operator to control the operation of the generator.
MVars = √(MVA2 - MW2)
Excessive maximum and minimum MVar loading on the generator is the
main monitoring concern.
Exceeding the maximum MVar loading means that the field current limit on
the rotor has probably been exceeded in the lagging power factor range, and
will affect the condition of the rotor winding. The stator terminal voltage
might also have been exceeded during excessive MVar loading, depending
on the main transformer tap settings. Excessive terminal voltage will affect
core heating.
Going below the minimum MVar loading means that the field current on the
rotor has been reduced to a very low level so that the generator is operating
in the leading power factor range. When the MVars are reduced beyond the
design limits, the possible problems that can occur are overheating the stator
core end, going below the minimum terminal voltage limit, and losing
stability from slipped poles.
Transient MVar events from the system or internally in the machine also
show up as transients in the stator current or terminal voltage.
3.1.3 Stator Three Phase Currents

The three phase currents flowing in the stator winding are alternating
currents (ac), and produce losses in the copper that directly affect the tem-
perature of the winding. In addition, vibration and bar bounce forces are
induced in the stator windings in proportion to the electrical current flowing.
Temperature and vibration affect the electrical and mechanical integrity of
both the inter-strand and groundwall insulation, and the stator bar surface
coatings. The mechanical integrity of the copper strands is also affected by
temperature and bar bouncing.
14 NMAC Tech Note

The stator currents are monitored and used to provide an indication of stator
current overload and phase current unbalance, and they can be used to
calculate the negative sequence currents flowing in the rotor.
3.1.4 Generator Terminal Voltage

Terminal voltage is a function of magnetic flux, rotor speed, and the stator
winding configuration. Excessively high voltage on the stator winding can
break down the groundwall insulation and deteriorate the stator bar surface
coatings due to the electrical breakdown phenomenon.
The stator terminal voltage is monitored for abnormal terminal voltage
(either too high or too low), and for the degree of phase voltage unbalance.
Monitoring the generator terminal voltage is critical while synchronizing the
generator with the system. The terminal voltage of the generator must be
matched in magnitude, phase, and frequency to that of the system voltage
before closing the main generator breakers. This ensures smooth closure of the
breakers and connection to the system with no mal-synchronization occurring.
3.1.5 Field Current

The current flowing in the rotor winding is direct current (dc), and produces
I2R (current2 · resistance) losses that directly affect the temperature of the
winding. The temperature of the rotor winding cannot be measured directly
but rather by resistance using the measured field current and the measured
field voltage to calculate the resistance of the rotor winding. This will be
discussed in detail in Section 3.4.1.
Excessive field current causes field winding overheating and high terminal
voltage.
If the field current is too low, it can cause the terminal voltage to decrease
below its minimum allowable value, or it can also cause core end heating. In
addition, if the field current is reduced too far, magnetic coupling between
the stator and rotor fields is weakened so that loss of synchronism occurs
from slipped poles. This is more commonly referred to as instability.
3.1.6 Field Voltage

Increasing the field voltage increases the field current, proportional to the
rotor winding resistance. Field voltage is monitored, but is not usually used
for alarms or trips. It is used to calculate rotor winding resistance and,
subsequently, the rotor winding average and hot spot temperature. Auto-
matic voltage regulator problems can cause the field voltage to become too
high and in turn cause the excitation to increase beyond design limits.
3.1.7 Frequency
Frequency, measured in cycles per second or hertz (Hz), refers to the electri-
cal frequency of the generator. It is monitored for abnormal deviation from
the system frequency, which is 60 Hz in North America.
NMAC Tech Note 15

Frequency is more a consideration for turbine blade, hydrogen (H2) seal, and
bearing operation. There are many subsynchronous vibration modes associ-
ated with low frequency operation that could fail these components and
must be avoided. In terms of generator operation, overfrequency and
underfrequency are the main concerns.
Overfrequency is most often the result of an instantaneous load reduction
when the generator is synchronized to the system or from excessive excita-
tion when the generator is in the open circuit condition at rated speed and
terminal voltage.
During on-load operation, fast load reductions cause the current in the stator
winding to decrease rapidly and the terminal voltage to increase rapidly due
to a high level of field excitation still applied. In this case automatic action is
taken to decrease the steam input to the turbine in order to match the load
requirements and to quickly reduce field current in order to keep the termi-
nal voltage within limits (for example, the AVR).
Underfrequency is generally caused by a system event rather than by the
generator itself. The effect on the generator however, is almost always an
attempt by the system to extract excessive current from the stator and to drag
the rotor speed down. This also has the effect of depressing the stator termi-
nal voltage. To offset this, the excitation system for the generator will nor-
mally go into field forcing to try and maintain rated terminal voltage. There-
fore, it can be seen that there is a possibility of sustaining overheating in both
the stator and rotor windings during this type of event. Protection against
overloading in these components is usually provided.
3.1.8 Volts per Hertz

Volts per hertz (V/Hz) is the ratio of terminal voltage to generator electrical
frequency. It is actually used as a generator supplementary start protection
during the open circuit condition rather than for on-load operation. It is
primarily in place to protect the generator from overfluxing during open
circuit operation.
During the open circuit condition, core fluxing and, hence, heating are at a
minimum and the risk of overheating the core is low. However, if the termi-
nal voltage should rise beyond the design capability of the stator core, there
is a danger of overfluxing the core.
3.1.9 Negative Sequence

Negative sequence refers to negative sequence currents in the generator
rotor, induced by unbalance in the three stator phase currents. The unbal-
anced condition is usually due to a machine or system problem.
There are two components of negative sequence to consider. The first is the
continuous I2 component, which refers to the amount of phase unbalance the
generator can tolerate for an infinite operating period. The second is the
16 NMAC Tech Note

transient component called I22t, which refers to the degree of short term phase
unbalance that the generator can withstand.
For large steam turbine generators, a typical continuous I2 value of 8 would
be normal. This means that the generator could carry a continuous phase
unbalance in the stator winding of 8% or .08 per unit (p.u.) of the rated stator
current without damaging any of the generator components, specifically the
rotor. A typical transient value for I2 (the I22t component) would be 10. This
means that the generator could withstand 100% or 1 p.u. phase unbalance for
10 seconds (that is, 1 p.u. phase current times 10 seconds = an I22t value of 10)
[2]. The overall relationship for negative sequence capability of a generator is
expressed exponentially as shown in Figure 3.
I 2
P.U.
1
.9
.8
.7 2
I 2 t = 10
.6
.5
.4
.3
.2
.1 continuous I = 8
2
.08
0
10 40 100 120
Time (seconds)
Figure 3
Large Generator Negative Sequence Capability
There is always a small natural degree of unbalance in these three phase

currents, but they are not harmful below the continuous I2 value. When the
degree of unbalance becomes significant, it appears as 120 Hz currents
flowing in the surface of the rotor body and wedges, which can overheat the
rotor forging and wedges. The symptoms can arise as high vibration due to
thermal unbalance in the rotor, and eventual component failure can happen
should overheating occur. Relay protection is generally provided to detect
the level of negative sequence currents and initiate a generator trip.
NMAC Tech Note 17

3.1.10 Generator Main Breaker Status

Generator main breaker status is a simple indicator of the position of the
generator breaker. It indicates if the generator is connected to the system.
3.1.11 Generator Field Breaker Status

Generator field breaker status is an indicator of the position of the dc field
breaker. It indicates that the excitation system is connected to the rotor for
field current application.
3.2.1 Core Temperatures

3.2 Stator Core and
Frame Stator core temperatures are monitored by TCs or RTDs embedded between
the stator core laminates at strategic locations. In the radial direction, these
locations are most often in the tooth center, in the core yoke a few inches
below the slot bottom, and in the core yoke centered between the slot bottom
and the back of the core. In the axial direction they are positioned at both
core ends, generally from the first core packet up to about one and a half feet
inward, and in the core center. This provides coverage to fully monitor the
various core heating modes including global overheating, core end overheat-
ing, and localized overheating from local core faults.
Global overheating of the stator core can occur from overfluxing by excessive
field current application either on-line or off-line. An indication that this
condition is occurring is also evident from the stator terminal voltage being
above the allowable limit. It can also occur during overload, in which case
the terminal voltage and field current will almost certainly be above limits.
Another cause of global core overheating can occur when the hydrogen gas
pressure drops too low. In this case, the core is not getting sufficient cooling.
Hydrogen gas purity also affects core temperature, but a drop in purity
would have to be severe for overheating to occur.
The other cause of global core overheating is high hydrogen gas temperature.
This is most likely to occur due to a problem with the hydrogen coolers, such
as low flow or loss of raw cooling water to the H2 coolers. There have also
been cases where the raw water supply is too warm in the summer months
in southern areas, and even at full raw water flow, the cold H2 gas could not
be maintained below allowable limits. Again, core heating will be higher but
overheating is unlikely. The condition should be corrected though because
the stator core interlaminar insulation life will be adversely affected in the
long run by extended operation at higher than rated temperatures.
Core end overheating is a condition related to leading power factor opera-
tion. In this overheating mode, the core center is not affected. During leading
power factor operation, the interaction of the magnetic fields in the core ends
is such that there is a higher degree of axial flux impingement on the core
that tends to saturate the iron at the core ends. As the power factor or field
current is reduced, saturation increases, and subsequently the core end
18 NMAC Tech Note

temperatures increase also. The level of temperature rise in the core ends is
dependent on a variety of generator design features, and depending on the
design variations, some machines see this effect more than others.
Local core overheating is a condition that is related to a small area of the core,
usually due to a localized defect or foreign body present in the stator. It can
also be associated with a breakdown of the interlaminar insulation or short-
ing of the core laminates.
Interlaminar insulation breakdown can occur from voltage spikes due to
system- or machine-related events, fretting from loose iron, poor space block
welds, and so on. Shorts across laminates usually occur from either foreign
objects in the bore, or damage to the core from the rotor skid plate or from
hammer strikes during re-wedging activities.
3.2.2 Core Clamping Plate Temperature

Core compression or clamping plate overheating is also related primarily to
leading power factor operation and is affected by the interaction of the
magnetic fields at the ends of the generator, the same as the core. TCs or
RTDs are attached to the core compression plates to monitor for excessive
temperature rise.
In addition to the end iron effects from leading power factor operation, the
clamping plates are also susceptible to overheating by abnormalities in the H2
gas (that is, low gas pressure or high gas temperature) because they are also
cooled by hydrogen.
3.2.3 Core End Flux Screen Temperature

Core end flux screens are employed to reduce the saturation effect on the
stator core end iron and core compression plates. Again, this occurs primarily
during leading power factor operation. The net effect of the flux screens is to
control the stator end temperatures below the Class B insulation level of
130˚C. The flux screens, therefore, generate considerable losses and heat from
the circulating currents induced in the screens.
Temperature monitoring with TCs or RTDs is provided to ensure that the
high temperature limit is not exceeded.
In addition to the end iron effects from leading power factor operation, the
clamping plates are also susceptible to overheating by abnormalities in the H2
gas (that is, low gas pressure or high gas temperature) because they are also
cooled by hydrogen. However, there are some designs in existence that use a
water circuit in the flux screens and use stator cooling water for primary heat
removal. In such cases, the H2 gas has a lesser effect.
3.2.4 Core Vibration

Vibration in the stator core is naturally produced by the unbalanced magnetic
pull in the air gap, originating from the unequal magnetic field distribution
NMAC Tech Note 19

of the rotor. The pole or direct axis carries the main flux while the winding or
quadrature axis carries only the leakage and stray fluxes. Therefore, a large
difference in magnetic force is inherent between the two axes. A large mag-
netic force is generated in the pole axes, and a weak magnetic force is present
in the winding axes. Because each pole has a north and a south (or a + and -)
associated with it, an unbalanced magnetic pull is generated at twice the line
frequency or 120 Hz.
The core must be maintained in a tight condition or fretting will occur
between the laminates. Minor fretting tends to deteriorate the interlaminar
insulation, but if the core becomes too loose, the laminates and the space
blocks can become fatigued, resulting in pieces of loose core material break-
ing off and causing damage.
Monitoring of core vibrations can be done with accelerometers mounted on
the core back in strategic locations to determine the magnitude and phase of
both radial and tangential vibration modes.
3.2.5 Frame Vibration

Frame vibration is also excited by the unbalanced magnetic pull and by any
vibration produced in the core. There are also known cases of vibration
resonances occurring on the frame as a result of the frame’s having a reso-
nant frequency near the line frequency or twice the line frequency. The
occurrence of resonances is generally due to design error. Resonant frequen-
cies can be corrected either by adding mass to the frame to bring the natural
frequency down or by stiffening the frame to drive the natural frequency
higher, the object being to move the frame’s natural frequency away from the
exciting frequencies by at least 10%.
Frame vibration can cause severe frame damage if it initiates cracks in the
frame welds or in the frame members themselves. Residual damage from the
high vibrations associated with frame vibration are likely to be transmitted to
other components of the generator if the situation becomes severe.
Monitoring of frame vibration can also be done with accelerometers
mounted on the keybars, frame ribs, or casing structure in strategic locations
to determine the magnitude and phase of both radial and tangential vibra-
tion modes.
3.2.6 Liquid Level in Generator Casing

The majority of large generators are equipped with liquid detectors. As fluids
collect in the generator casing, they are drained to a tank that is equipped
with a level detector connected to an alarm. The liquid detector does not
differentiate between the fluids that collect in it, only that the allowable
liquid level in the stator casing has been exceeded.
When the liquid detector alarm is triggered, a sample must be taken of the
liquid to determine the source. The possibilities are seal oil, SCW, and H2
cooler service water.
20 NMAC Tech Note

3.3.1 SCW Inlet Temperature

3.3 Stator Winding
Pure demineralized water is used as the cooling medium for stator windings
in large generators. It is commonly referred to as the stator cooling water
(SCW). Water cooling of stator windings is usually done on generators of 300
MW size and larger. Below that, cooling is done indirectly by hydrogen gas
or air, and also the stator core acts as a heat sink to dissipate the losses
generated in the stator winding.
Generally, the SCW inlet temperature is maintained below 50˚C, but usually
operates in the 35–40˚C range. High inlet water temperature causes the water
outlet temperature from the stator winding to be correspondingly high as
well. This increases the possibility of the outlet water temperature exceeding
the boiling point. Strict attention must be paid to the inlet SCW temperature
because the outlet SCW water temperature is directly affected.
3.3.2 Bulk SCW Outlet Temperature

The outlet temperature for the SCW is normally maintained below 75˚C. The
upper limit is approximately 90˚C. Overheating of the stator bars and possi-
bly boiling of the SCW are likely to occur if the limit is exceeded.
Monitoring the bulk SCW outlet temperature provides indication of a range of
problems that might be occurring in the stator winding, but often this tempera-
ture is indicative of a global stator winding heating problem rather than a
localized winding problem. If other stator winding components are water
cooled, such as the stator terminals and phase connectors, they can be affected
by a heating problem. In many generator designs, the terminals and phase
connectors are cooled by hydrogen and the stator bars are cooled by water.
The main problems indicated by a high bulk SCW outlet temperature are
stator current overload, low flow or a reduction of the cooling water flow to
the stator winding, or high inlet cooling water temperature. These conditions
are also indicated by high temperature readings on all slot RTDs and/or
SCW outlet temperature sensors.
Stator current overload means that the stator current is above the maximum
allowable limit and that the I2R losses in the copper of the winding have
exceeded the cooling capacity of the SCW system.
Low flow or reduced cooling water in the winding can occur from:
Plugging (of the stator bars). Plugging can occur from corrosion of the
hollow copper conductor bar strands, debris, or dust plugs left in the stator
water cooling path after a generator overhaul.
Leaks. Low cooling water flow in the generator from a leak would have to be
a substantial leak and would most likely involve a problem with the SCW
system, external to the generator. Leaks can originate at piping joints, from
cooler tube failure, and so on.
NMAC Tech Note 21

Reduced flow from the SCW system. Reduced flow from the SCW system
may mean a pump problem or a clogged SCW filter/strainer.
A high temperature at the cooling water inlet to the generator also implies an
SCW system problem. A likely cause is the SCW coolers not getting enough
raw service water for heat removal from the SCW itself.
3.3.3 Conductor Bar Hose Outlet Temperature

The SCW outlet temperature of the individual conductor bars are generally
maintained below 75˚C. Operating above 90˚C will overheat a stator bar and
possibly allow boiling to occur in the individual stator bar if the upper limit is
exceeded. Boiling can cause stator bar strand rupture and an eventual ground
fault. Monitoring of individual stator bar hose outlet temperatures provides an
indication of a problem in the specific bar being monitored and a high tem-
perature at a stator bar hose outlet is indicative of a localized bar problem.
High stator bar hose outlet temperatures can be caused by plugged strands
from debris, corrosion products, or hydrogen gas locking, or by collapsed
strands from magnetic termites that collapse the hollow strands by magnetic
attraction and vibration effects. In addition, broken strands, broken electrical
connectors, and bar bouncing can cause stator bar temperature problems.
Plugging in a conductor bar is one of the most common problems that can be
sensed by hose outlet temperature monitoring. Plugging refers to a flow block-
age in the stator conductor bars by one of the above mechanisms, with the main
result likely to be overheated bars due to insufficient cooling water flow.
If the plugging is partial, SCW flow continues but not in sufficient quantity
to cool the affected stator bar. Therefore, the higher temperature of the SCW
outlet is picked up by the sensor.
If there is full blockage of the coolant flow in a conductor bar for one of the
above reasons, the condition might also be indicated by a high slot RTD
reading (discussed in next section) but a lower SCW outlet temperature. This
happens when the RTD is still reading the high bar temperature but the SCW
outlet sensor is detecting the temperature of the cooling gas since there is no
cooling water outlet flow. In addition, a coolant flow blockage also can be
indicated by a core monitor alarm if the stator winding insulation is burning.
A severe condition can occur that causes boiling of the stator cooling water.
For such an occurrence, some manufacturers provide an automatic runback
scheme to bring the machine to minimum load to avoid overheating. All
OEMs recommend reducing the load during such a condition.
3.3.4 Conductor Bar Slot Temperature

Conductor bar slot temperature monitoring refers to an embedded TC or
RTD between the top and bottom bars in the slot. The same conditions are
generally monitored with slot sensors as those described for hose outlet
sensors with the following additional information.
22 NMAC Tech Note

A well instrumented generator stator winding employs an RTD or TC em-

bedded between the top and bottom conductor bars in each slot, and a TC
installed on the cooling water outlet hose of each individual stator bar.
However, this is not always the case. In some machines, there might be only
one or the other, or only 25% of the slots might contain sensors.
With only slot temperature monitoring, there is a possibility that a high
temperature reading is caused by either the top or bottom bar in the slot.
With only hose outlet temperature monitoring of the stator bar coolant,
overheating might go undetected where there is total flow blockage in a bar.
The hose outlet sensor temperature actually reverts to the temperature of the
hydrogen cooling gas circulating over the monitoring location because there
is no flow of cooling water and, therefore, the true temperature of the stator
bar is unknown.
The optimum monitoring situation is where both slot and hose outlet tempera-
ture monitoring are employed. This allows identification of the affected bar or
bars and provides a positive means to check the actual temperature of each bar.
In cases where only one strand in a stator bar is plugged, there might be only
a minimal effect on the discharge temperature of the stator bar and its perfor-
mance. As more strands become plugged, the partial blockage of the coolant
flow in one bar increases and results in a much more identifiable overheating
situation.
3.3.5 Stator Winding Differential Temperature

Stator winding differential temperature refers to a condition of temperature
unbalance between individual stator bars or phases. The condition can be
caused by phase unbalance due to a system problem, localized bar-to-bar
temperature differences due to a design problem, plugging, other flow
restrictions, and high resistance or broken electrical joints. An attempt is
normally made to keep the temperature differential to less than 10˚C. The
condition is also indicated by high readings at the slot RTD or SCW outlet
temperature sensors.
The differential pressure across the stator winding from the SCW inlet to the
outlet of the generator can be monitored to ensure that the design pressure
drop across the whole stator winding is at the correct level. Correct differen-
tial pressure is an indication that the SCW flow rate is above the minimum
recommended level for adequate cooling of the stator winding as a whole.
This type of monitoring is not indicative of problems in individual stator
conductor bars.
When the pressure differential is higher than normal or higher than recom-
mended by the manufacturer, a partial blockage somewhere in the stator
winding or SCW system may be indicated. To determine the location of the
blockage, other testing and monitoring is required.
NMAC Tech Note 23

If the pressure differential is much higher than expected, a large general

obstruction to the flow of the SCW might be present. If the pressure drop is
very low across the generator, this can indicate a different problem, such as a
large leak of SCW from the external system piping before the generator inlet.
3.3.6 SCW Inlet Pressure

The SCW inlet pressure to the generator is generally kept at a design level to
ensure cooling water flow to the stator winding, but it should be 5 psi below
the hydrogen gas pressure to minimize the possibility of water leakage into
the generator.
3.3.7 Stator Winding Differential Pressure

There is a normal differential pressure that exists across the stator winding
based on design factors. A higher than normal differential cooling water
pressure across the stator winding can indicate that there is a cooling water
flow problem. This can be due to plugging or hydrogen gas locking.
3.3.8 Hydrogen Gas in SCW

Hydrogen leaking into the SCW can occur at fittings, at joints because of
cracks or separations, from a bad or porous braze, from corrosion, due to
cracked strands, and from holes due to magnetic termites. It is aggravated by
vibration and causes high conductivity of the SCW as hydrogen contamina-
tion increases. Hydrogen gas locking, which can also lead to bar overheating,
can occur if the leak is large enough.
Hydrogen gas locking is a condition that can occur if the leakage rate of
hydrogen into an SCW leak is large enough. Hydrogen gas locking can cause
a cooling water flow restriction and eventually overheat the affected stator
bar. Another form of gas locking can occur from gas bubbles if the SCW
starts to boil in the bar.
Westinghouse and Parsons employ a counter system to determine the amount
of hydrogen lost from the SCW system during on-line operation of the genera-
tor. The counter is a device used to collect a measured volume of hydrogen gas
vented from the SCW storage or make-up tank and then it releases the col-
lected gas to the atmosphere. The number of releases times the volume of the
collection chamber indicates the hydrogen into SCW leak rate.
SLMS is General Electric’s Stator Leak Monitoring System, which is an
automated collection and measurement system for the vented gas in the
SCW system. It has a secondary purpose of helping to keep the SCW dis-
solved oxygen content at the correct level. This is a preferable method to vent
line bagging and is similar to the vent line counter.
24 NMAC Tech Note

3.3.9 Stator Winding Ground Alarm

Stator winding grounds occur when there is a failure of the stator winding
ground wall insulation. In addition to the stator winding conductor bars, the
phase connectors and terminals are also protected by this alarm. In fact this is
more a protection scheme than a monitoring parameter and is generally used
to trip the generator.
However, a ground alarm is also provided that is initiated by instantaneous
overvoltages due to either machine transients or system spikes. The alarm is an
indication that an event has occurred, but does not necessarily mean that a
winding insulation failure has occurred. Some basic electrical testing is usually
carried out to prove that the ground insulation is still viable and the unit is
returned to service if the tests show no failure has occurred. In this sense, the
stator winding ground alarm might be considered an odd form of monitoring.
3.4.1 Rotor Winding Temperature

3.4 Rotor
The rotor winding temperature is generally measured as a function of the
winding resistance. This produces the average temperature of the winding
but does not indicate the temperature of the hottest part of the winding. The
hot spot in the rotor winding is dependent on a knowledge of the rotor
winding cooling circuit for the location of the hot spot. The OEM should be
able to provide a profile of the temperature distribution across the axial
length of the rotor winding. The difference between the hot spot and the
average winding temperatures produces a multiplier factor for calculation of
the hot spot. The multiplier factor is determined as shown in Figure 4.
NMAC Tech Note 25

Temp.
Rise
(deg C)
70
Rotor Winding Maximum Temperature Rise
60
50
Rotor Winding Average Temperature Rise
40
30
20
Rotor Winding Axial Temperature Profile
10 Above Cold H2 Gas Temperature
0
Endwinding Slot Portion Endwinding
Hot Spot Multiplier (M) 60

–– = 1.5
40
Figure 4
Determination of Rotor Winding Hot Spot Multiplier
With the hot spot multiplier known, the rotor winding hot spot temperature
can be calculated for any load and cold operating gas temperature as follows:
Ths = ( Tave - Tcold H2 ) x M + Tcold H2
where: Tave = Rave / Rref ( K + Tref ) - K (IEEE 115) [3]
Rave = Vf / If , calculated average winding resistance from
measured field voltage, Vf, and field current, If
Rref = winding resistance measured in factory by OEM at
known reference temperature, Tref
K = 234.5 (copper constant)
Tcold H2 = measured value of the cold hydrogen gas temperature
M = Ths / Tave , hot spot multiplier from OEM winding
temperature profile data
The rotor winding temperature varies with field current, hydrogen gas
temperature, and pressure.
Rotor winding high temperatures usually occur as a result of overloading or
undercooling. They are caused mainly by excessive field current in the on-
line mode, low hydrogen gas pressure, high hydrogen cold gas temperature,
26 NMAC Tech Note

or ventilation problems in the rotor that block the hydrogen cooling gas from
flowing to part of the winding. Excessive rotor winding temperature causes
deterioration of the rotor winding inter-turn and ground wall insulation.
Advanced effects of the rotor winding exceeding its temperature limit can be
overheating or burning of the insulation or a rotor thermal unbalance.
Reducing field current to remove the energy input to the rotor winding allows
the winding temperature to come down to a tolerable level and stabilize.
Failure of the inter-turn insulation causes rotor winding shorted turns.
Failure of the ground wall or slot liner insulation causes rotor grounds. The
retaining ring is also insulated from the rotor winding and is affected by high
winding temperatures. Failure of the retaining ring insulation also constitutes
a rotor winding ground.
3.4.2 Rotor Winding Ground Alarm

Rotor winding grounds are actually a leakage of field current to ground. As
previously stated, grounds can occur through or over the slot liner and the
retaining ring insulation. In addition, grounds can occur to the rotor forging
through the insulation systems of:
• The sliprings
• The radial studs
• The upshaft leads or bore copper conductors
• The rotor winding main leads
Grounds can also occur external to the rotor itself, out in the excitation
system.
Rotor winding ground protection is applied to the generators to provide
warning of a ground. The rotor winding ground alarm is almost always con-
nected as an alarm and not an actual generator trip. It is left up to the operator
or suitable authority to decide whether the unit should be taken off-line.
While it is common for utilities to operate with a single rotor ground for
short periods of time until a convenient outage, this practice has a high
degree of inherent risk. Should a second ground occur anywhere on the rotor,
very high currents are circulated through the two ground points, creating
overheating in the affected rotor components. The overheating effect is
extremely serious and damaging to the machine. In all probability, a double
rotor ground is likely to cause a catastrophic rotor failure.
3.4.3 Shaft Speed

Shaft speed refers to the shaft rotational speed. In North America, this is 3600
RPM for two-pole generators and 1800 RPM for four-pole generators. For both
two- and four-pole machines, this translates to 60 Hz operation of the generator.
NMAC Tech Note 27

The speed is generally measured by a probe mounted next to the rotor,

looking at a toothed wheel or key phasor on the rotor shaft. The speed signal
is used for monitoring, but the frequency of the generator is usually taken
from the electrical output of the generator in Hz.
Shaft speed monitoring is particularly useful when looking at the vibration
profile of the rotor during run-up and run-down when the generator goes
through its first and second critical speeds. Shifts in critical speeds during
run-up and run-down can be an indicator of a change in shaft stiffness due to
some problem within the rotor, for example, a cracked shaft.
3.4.4 Rotor Vibration

Rotor vibration refers to vibration monitoring on the shaft and the bearings.
Shaft vibration is the movement of the shaft itself in relation to the bearing
mounts or generator footing that the vibration probe is normally mounted on.
Bearing vibration is the movement of the bearings relative to the generator
footing.
Vibration is measured in units of displacement (that is, thousandths of an
inch or micro meters, peak to peak on the displacement signal obtained). All
generators operate under strict bearing and shaft vibration guidelines and
limits set down by the manufacturer, but they are often modified by the
utility from operating experience.
High rotor vibrations can damage the rotor or one of its components, the
bearings, H2 seals, the slipring, and so on. High rotor vibration is caused by a
number of mechanical or thermal unbalance problems.
Rotor mechanical unbalance is caused by conditions such as the following:
• Loss of a balance weight
• Oil whirl
• Bearing loading
• Coupling alignment
• Misaligned retaining rings
• H2 seal or oil wiper rubs
• Foundation resonances
• General rotor structural component problem
The level of severity is most often determined by the magnitude of vibration
present and can require an outage to correct the source of the vibration or to
apply balancing weights to offset the unbalance. For mechanical vibration
problems, vibration levels generally remain constant regardless of field
current changes but vary with the shaft rotational speed.
28 NMAC Tech Note

Rotor thermal unbalance is a condition where there is uneven heating in the

rotor shaft due to a number of possible influences with the resulting effect
being high rotor vibration. Some of the possible causes are rotor shorted
turns, negative sequence heating, or blocked rotor winding ventilation ducts.
Vibrations due to thermal unbalance are usually load dependent, that is,
vibrations increase with increases in field current.
3.4.5 Bearing Metal Temperature

Rotor bearing metal temperature is a function of the lube oil inlet tempera-
ture at the bearings. In addition, bearing temperature is affected by:
• Vibration
• Alignment
• Oil condition
• Bearing preloads
• Electrical arcing between the rotor shaft and the bearings
• Low lube oil flow to the bearings
High bearing metal temperature also causes high temperatures at the bearing
oil outlet for the affected bearing. Excessive bearing temperatures might
result in an overheated bearing and a subsequent rub or wipe.
The bearing metal temperature of the upper and lower bearing halves or of
the individual pads in a tilting pad bearing is monitored by thermocouples
embedded in the babbit material of the bearing.
3.4.6 Bearing Inlet Oil Temperature

High inlet oil temperature causes the bearing temperatures to rise and can
affect their operation.
High inlet oil temperature is almost always a result of poor operation of the
bearing lube oil system and affects all generator and turbine bearings. If the
bearing oil inlet temperature exceeds its limits, investigation of the root cause
should begin with the lube oil system.
3.4.7 Bearing Outlet Oil Temperature

Bearing oil outlet temperatures are indicated most commonly by a TC or RTD
in the oil outlet stream from each individual bearing. Overheating of the
individual bearings is indicated by this parameter and should correlate with
the bearing metal temperature readings.
Bearing oil outlet temperatures can also indicate a global problem with the
lube oil system if all bearing oil outlet temperatures are high. In this case, it is
possible that the bearing inlet oil from the lube oil system has a temperature
problem from the lube oil system.
NMAC Tech Note 29

3.4.8 Hydrogen Seal Metal Temperature

Hydrogen seal metal temperatures are a function of the seal oil inlet tempera-
ture. In addition, H2 seal temperatures are affected by mechanical problems
such as seal rubs from misalignment, electrical mechanisms such as arcing,
and oil-related causes such as low inlet oil flow to the seals. High H2 seal
metal temperatures also cause high H2 seal oil outlet temperatures of the
affected seal.
Overheating of the seal metal can cause a failure of the seal and subsequent
serious damage not only to the seals but also to other generator components.
A hydrogen seal failure is extremely dangerous because hydrogen is likely to
escape from the generator casing at the location of the failure and ignite, not
just due to the high temperature from the failure point, but hydrogen leaking
from a pressure vessel often will self-ignite.
H2 seal metal temperatures are monitored by thermocouples embedded in
the seal ring material.
3.4.9 Hydrogen Seal Inlet Oil Temperature

High inlet oil temperature causes the H2 seal temperatures to rise and can
affect their operation.
High inlet oil temperature is most almost always the result of poor operation
of the seal oil supply system and it affects both generator H2 seals. If the H2
seal oil inlet temperature exceed its limits, investigation of the root cause
should begin with the seal oil supply system.
3.4.10 Hydrogen Seal Outlet Oil Temperature

H2 seal oil outlet temperatures are indicated most commonly by a TC or RTD
in the oil outlet stream from each individual seal. Overheating of the indi-
vidual H2 seals is indicated by this parameter and should correlate with the
H2 seal metal temperature readings.
H2 seal oil outlet temperatures can also indicate a global problem with the
seal oil supply system if all H2 seal oil outlet temperatures are high. In this
case, it is possible that the inlet oil has a temperature problem from the seal
oil supply system.
3.4.11 Hydrogen Seal H2/Seal Oil Differential Pressure

A high degree of hydrogen loss from the generator can cause the H2/seal oil
differential pressure to increase. Similarly, seal oil inlet pressure to the
generator being too high can also cause the differential pressure to increase.
One of the problems of high H2/seal oil differential is excess seal oil leaking
into the generator, which can overload the drainage system and also cause oil
to mist and coat the generator internal components. The problem with this is
that oil has a tendency to promote loosening of components such as the
stator wedges and so on.
30 NMAC Tech Note

The differential pressure between the oil inlet to the H2 seals and the casing H2
gas pressure is continuously monitored. It is always maintained in the range of
5–15 psi, with the seal oil pressure being the higher of the two. The higher seal
oil pressure keeps the hydrogen from escaping the generator casing.
Loss of the H2 seal at either end of the generator causes hydrogen leakage at
the seal, creating a dangerous condition for both personnel and the internal
generator components.
When hydrogen under pressure is allowed to leak uncontrolled, it has the
ability to self-ignite. When it does, the flame is invisible to the eye. The only
clues to the presence of a hydrogen fire are heat given off and the formation
of water droplets because the escaping hydrogen combines with the oxygen
in air to make water.
The other problem that leaking hydrogen causes is the reduction in cooling of
the generator components due to the reduction in H2 pressure.
3.4.12 Rotor Fan Differential Pressure

Rotor fan differential pressure is an indication of the operation of the H2
cooling circuit inside the generator. The fan differential pressure is the pres-
sure required to overcome the pressure drop across the entire hydrogen gas
cooling circuit inside the generator, so that there is a sufficient flow of hydro-
gen to cool the generator components.
A significant increase in the fan differential pressure is an indication that
there might be a flow restriction somewhere in the cooling circuit. Low fan
differential can indicate a problem with the fan itself.
3.5.1 AC Power into the Exciter

3.5 Excitation
System The ac power into the exciter is the power consumed by the excitation sys-
tem. In static excitation systems, it is the power delivered by the ac supply
transformer to the excitation cubicles; in rotating excitation systems, it is the
input to the main rotating exciter. For any particular load, the ac power
should be at a certain level. A check can be made of the consumption to see if
it is in the correct range for the load output.
3.5.2 DC Power out of the Exciter

The dc power out of the exciter is the power delivered to the rotor winding
after rectification. The difference between the ac power in and dc power out
should be only the normal losses. A large differential between the two is an
indicator of a possible excitation system fault.
Whether static or rotating, the main exciter can be considered the power
stage of the excitation system. When the power stage of the exciter has failed,
it can result in a partial or complete loss of excitation to the generator. For a
static excitation system, this would include the thyristor bank or transformer.
NMAC Tech Note 31

For a rotating exciter, this would include problems with the diode wheel or
rotating alternator.
3.5.3 Main Exciter, Cooling Air Inlet Temperature

The cooling air temperature into the main excitation system, whether a rotating
exciter or a static exciter, is generally the ambient temperature of the power-
house. Cooling air is usually taken from the powerhouse, filtered, directed over
the excitation equipment for cooling, and discharged back into the powerhouse.
The cooling air inlet temperature varies with the seasons as the powerhouse
general atmosphere temperature changes. Therefore, temperature problems
relating to the cooling air supply for the exciters are more likely to occur in
the summer.
3.5.4 Main Exciter, Cooling Air Outlet Temperature

The cooling air temperature out of the main exciter (for example, static or
rotating) is an indicator of the overheating of the main excitation compo-
nents. The cooling air outlet temperature also varies with the cooling air inlet
temperature; therefore, the temperature rise of the outlet cooling air above
the inlet cooling air of the main exciter should be monitored.
3.5.5 Collectors and Brushgear, Cooling Air Inlet Temperature

The collector and brushgear inlet air is normally the air from the powerhouse
turbine hall at its ambient value. It is filtered, directed over the collectors and
brushgear for cooling, and discharged back into the turbine hall.
3.5.6 Collectors and Brushgear, Cooling Air Outlet Temperature

The cooling air outlet temperature from the collectors and brushgear is an
indicator of overheating of these components. Overheating is generally
caused by current overloading or improper operation of the brushgear,
indicating a problem with brushgear operation.
The outlet air temperature also varies with the cooling air inlet temperature;
therefore, the temperature rise of the outlet cooling air above the inlet cooling
air should be monitored.
3.5.7 Collectors and Brushgear, Cooling Air Filter Differential Pressure

The collectors and brushgear air filter being clogged or dirty causes reduced
cooling air flow to the collectors and brushgear and can also cause above-
normal temperatures and poor operation of the collectors and brushgear.
32 NMAC Tech Note

Hydrogen is provided in large generators for improved cooling of the inter-

3.6 Hydrogen
Cooling System nal generator components. Hydrogen is the primary coolant for the stator
core and frame, the rotor, and the stator winding in generators with directly
cooled stator windings. Hydrogen coolers and a cooling system are provided
to remove the heat absorbed by the hydrogen gas.
3.6.1 Bulk Hydrogen Supply Pressure

Hydrogen supply pressure is monitored to indicate when there is a problem
with the external hydrogen system pressure controls or simply that the
system external hydrogen supply is low and requires replenishing.
The effect of low external hydrogen supply pressure is no available hydrogen
make-up if the generator casing should lose gas pressure for some reason.
3.6.2 Generator Casing Hydrogen Gas Pressure

Hydrogen cooling effectiveness of the generator internal components is propor-
tional to the hydrogen gas pressure. Therefore, the design H2 pressure in the
generator must be maintained at all times the generator is in operation or the
temperature limits of the various components cooled by hydrogen are affected.
Low H2 gas pressure results in temperature rise of all the internal generator
components (stator winding, rotor, core, and so on) and eventually causes
overheating of the same components if the condition becomes severe enough.
The dielectric strength of the hydrogen gas is also proportional to H2 pres-
sure. Therefore, reducing the H2 pressure below its rated value also decreases
the insulating capability and the cooling capacity of the hydrogen gas.
Low generator casing gas pressure can occur if there is a leak of hydrogen
gas from the generator. This can occur at casing joints such as the end doors
or by a leak into the stator cooling water, H2 cooler service water, and seal oil.
A small amount of hydrogen will continually permeate through any Teflon™
hoses in the stator water cooling circuit and provide a slow source of hydro-
gen loss into the SCW.
There is always some continual loss of H2 gas that is made up by the external
hydrogen system, but if the external supply is low, the casing pressure cannot
be increased as needed.
It is also important to maintain the correct hydrogen pressure so that it is
always lower than the seal oil pressure. This ensures that hydrogen does not
escape from the H2 seals.
The casing hydrogen pressure is also maintained above the stator cooling
water pressure and the raw service water pressure of the H2 coolers. These
systems are designed to handle hydrogen ingress into the water and discharge
it safely to the atmosphere. It is important to keep the H2 pressure higher than
the water pressure in these systems so that water does not leak into the genera-
tor and create a possible failure mechanism or a high H2 dew-point condition.
NMAC Tech Note 33

3.6.3 Cold Hydrogen Gas Temperature

The cold H2 gas temperature is the ambient operating temperature of the
generator and is controlled by H2 cooler operation. It should be maintained at
its operating setpoint, which for most generators is between 30–40˚C. The
maximum allowable cold gas temperature by IEEE/ANSI standards is 46˚C [3].
The operation of the hydrogen coolers is usually either a once-through
system with temperature control throttled by varying the raw water flow to
the coolers or by use of raw water recirculation with temperature control
carried out by a temperature control valve that varies the warm raw water
recirculation level.
3.6.4 Hot Hydrogen Gas Temperature

The hot H2 gas temperature is a measure of the heat absorbed by the cold
hydrogen gas from cooling the generator components. The hot gas tempera-
ture should always be maintained with design limits set by the manufacturer.
In addition to a problem with one or more of the generator components
causing the H2 gas to become excessively warm, problems with the H2
cooling system can also cause high hydrogen gas temperature. This can occur
from reduced efficiency of the H2 coolers by either high raw water inlet
temperature or low raw water flow to the coolers.
3.6.5 Hydrogen Dew-Point Temperature

Hydrogen dew-point temperature is an indicator of the moisture content of
the H2 gas in the generator casing. Moisture is undesirable for the stator and
rotor insulation systems, because it can initiate insulation failure by electrical
tracking, and for various steel components in the generator, due to its rusting
and corrosive effects. It is recommended to maintain the dew point at less
than 0˚C [3].
High moisture content in the hydrogen gas is often due to hydrogen dryer
malfunction, contaminated hydrogen supply, poor seal oil quality, or a
malfunction in the seal oil system on the vacuum treatment plant. Stator
cooling water or H2 cooler raw water leakage into the generator can also
cause a high dew point but this is a less common cause.
Moisture is most likely to collect in the generator during shutdown periods when
the internal components begin to cool and the dew-point temperature rises.
3.6.6 Hydrogen Purity

Hydrogen purity is typically maintained above 95% to ensure high cooling
efficiency of the H2 gas. If the purity drops too low, an explosive mixture can
occur below 70%.
34 NMAC Tech Note

3.6.7 Hydrogen Make-Up or Leakage Rate

The hydrogen make-up or leakage rate is monitored to indicate a loss of H2
somewhere in the generator. Hydrogen can leak into the stator cooling water
and result in H2 gas locking in the stator winding, or it can leak into the H2
coolers. It can also be entrained in the seal oil at an excessive rate or leak
from generator casing at joints or connections. Hydrogen leakage can also
occur due to a problematic hydrogen seal, which has dangerous implications
due to the heat from the seal.
The allowable daily leakage is always specified by the manufacturer. How-
ever, most generators should be capable of maintaining a total daily leakage
rate under 1000 cubic feet per day, although some manufacturers would
argue that this is excessive.
3.6.8 Raw Service Water Inlet Pressure to Hydrogen Coolers

The service water inlet pressure to the H2 coolers is monitored to ensure a
continuous flow of raw cooling water to the H2 coolers and ensure proper
heat removal from the H2 gas.
Low raw water inlet pressure to the H2 coolers is an indication of low flow to
the coolers, which will affect the cooling of the H2 gas. Low service water flow
to the H2 coolers can be caused by either a failure of the supply pump or a large
leak before the generator. Low pressure or flow to the H2 coolers results in
reduced cooler efficiency and causes the cold H2 gas temperature to rise.
3.6.9 Raw Service Water Inlet Temperature to Hydrogen Coolers

Raw water from the local lake or river is circulated through the hydrogen
cooler tubes on one side to remove the heat from the hydrogen flowing over
the cooler tubes on the other side of the heat exchanger.
The service water inlet temperature to the H2 coolers is monitored because it
is important in maintaining the correct cold H2 gas temperature. It is also
monitored for comparison to the outlet water temperature and subsequent
determination of the differential temperature. This allows an estimation of
the cooler efficiency to be made.
High service water inlet temperature to the H2 coolers usually occurs in the
summer months. If the temperature of the inlet service water is too high, it
can cause an H2 cooler reduced efficiency problem. This results in the tem-
perature of the cold H2 gas rising above its normal setpoint.
Most generators use a once-through raw water system, but H2 coolers with a
warm service water recirculation system are not uncommon. These have
somewhat more control of the inlet service water temperature to the coolers
but are more useful in times when the service water is too cold.
NMAC Tech Note 35

3.6.10 Raw Service Water Outlet Temperature from Hydrogen Coolers

The raw service water outlet temperature is monitored to assist in detecting
problems with the H2 coolers. The service water outlet temperature should be
proportional to the inlet service water temperature and the generator loading.
Low service water flow to or in the H2 coolers can cause high outlet water
temperature. The low flow condition can be due to silting or fouling of the
cooler tubes. This results in reduced H2 cooler efficiency and causes the
temperature of the cold H2 gas and subsequently the generator internal
components to rise.
3.7.1 Lube Oil Cooler Inlet Temperature (Bearing Outlets)

3.7 Lube Oil System
Heat exchangers are provided for heat removal from the lube oil. Raw water
from the local lake or river is circulated on one side of the cooler to remove
the heat from the lube oil circulating on the other side of the heat exchanger.
The inlet water temperature is monitored for comparison to the outlet water
temperature and subsequent determination of the differential temperature.
This allows an estimation of the cooler efficiency to be made.
3.7.2 Lube Oil Cooler Outlet Temperature (Bearing Inlets)

The temperature of the rotor bearings varies with the inlet temperature of the
lube oil. It is important to keep the inlet lube oil temperature within the
recommended limits for the proper operation of the bearings and to reduce
the possibility of bearing overheating.
High lube oil inlet temperature to the rotor bearings occurs when the lube oil
outlet temperature from the lube oil system is too high. This can be caused by
reduced lube oil cooler efficiency, due to low service water flow to the lube oil
coolers or due to high service water inlet temperature to the lube oil coolers.
3.7.3 Lube Oil Pump, Oil Outlet Pressure

The lube oil pump outlet pressure is monitored to provide a warning of the
loss of lube oil flow due to a supply pump problem.
Low lube oil pump outlet pressure implies a problem with or failure of the
lube oil pump and creates a no- or low-flow condition of lube oil to the rotor
bearings. The result is a rise in the rotor bearing temperatures if the oil flow
is not re-established immediately.
Failure of the source of lube oil supply is extremely serious. The unit is
usually shut down as soon as possible to avoid bearing damage and possible
damage to the rotor shaft.
There are often redundant lube oil pumps for backup and, in some cases, a
dc emergency pump to allow safe unit shutdown if the shaft-driven or ac
pumps are unavailable for some reason.
36 NMAC Tech Note

3.7.4 Lube Oil System, Oil Outlet Pressure (Bearing Inlets)

The lube oil system outlet pressure (inlet pressure to bearings) is monitored
to ensure that there is lube oil flow to the bearings.
Fouled or plugged lube oil coolers cause a flow blockage and reduced lube
oil flow to the rotor bearings. Fouled or plugged lube oil coolers on the lube
oil side are most likely the result of debris in the lube oil or lube oil that is in
poor condition, containing sludge that has built up in the lube oil coolers.
A large leak in the lube oil system before the bearings also creates a low
pressure (low flow) problem.
3.7.5 Lube Oil Filter, Differential Oil Pressure

Lube oil filter differential pressure is used to indicate the operating condition
of the filters and the overall lube oil system. Low differential pressure on the
filters is an indication of reduced lube oil flow, while high differential pres-
sure is an indication of high lube oil flow or a plugged filter.
Fouled or plugged lube oil filters indicate that the filters have collected an
excessive amount of sludge or debris from the lube oil and the condition of
the lube oil is in question, or that the lube oil filter has not been cleaned or
maintained for some time. A plugged lube oil filter causes a lube oil flow
blockage resulting in reduced lube oil flow to the rotor bearings.
3.7.6 Lube Oil Tank Level

The lube oil tank level is monitored to ensure that there is always a supply of
lube oil for the lube oil pumps to draw on.
Low lube oil level in the tank implies a significant loss of lube oil in the system.
A lube oil system leak is one possible reason for a low lube oil tank level.
3.7.7 Lube Oil Flow

Lube oil flow to the bearings is monitored to ensure proper operation of the
bearings. Lube oil flow can be indicated by a flow indicator or by low lube oil
pressure at the rotor bearings.
Low lube oil flow to the rotor bearings can be caused by a flow blockage
somewhere in the lube oil system, a failure of the lube oil supply pump, or a
large lube oil leak from the lube oil system before the inlet to the bearings.
Low lube oil flow results in the bearing temperatures increasing above their
normal operating range. The design lube oil flow should be maintained to
avoid damage to the bearings.
3.7.8 Raw Service Water Inlet Pressure to Lube Oil Coolers

The service water inlet pressure to the lube oil coolers is monitored to ensure
continuous flow of raw cooling water to the coolers and ensure proper heat
removal from the lube oil.
NMAC Tech Note 37

Low raw water inlet pressure to the lube oil coolers is an indication of low
flow to the coolers, which affects the cooling of the lube oil. Low service
water flow to the lube oil coolers can be caused by either a failure of the
service water supply pump or a large leak before the coolers. Low pressure
or flow to the lube oil coolers results in reduced cooler efficiency and causes
the lube oil temperature to rise.
3.7.9 Raw Service Water Inlet Temperature to Lube Oil Coolers

Raw service water is circulated through the lube oil coolers on one side to
remove the heat from the lube oil flowing through the other side of the heat
exchanger.
The service water inlet temperature to the lube oil coolers is monitored to
maintain the correct lube oil inlet temperature to the rotor bearings. It is also
monitored for comparison to the service water outlet temperature of the
coolers and subsequent determination of the differential temperature. This
allows an estimation of the cooler efficiency to be made.
High service water inlet temperature to the lube oil coolers usually occurs in
the summer months. If the temperature of the inlet service water is too high,
it can cause a reduced efficiency problem in the cooler. This results in the
temperature of the lube oil rising above its normal operating level.
3.7.10 Raw Service Water Outlet Temperature from Lube Oil Coolers
problems with the lube oil coolers. The service water outlet temperature
should be proportional to the inlet service water temperature.
Low service water flow to or in the lube oil coolers causes high outlet water
temperature from reduced efficiency of the coolers. This results in the tem-
perature of the lube oil rising above its normal operating level. The low flow
condition can originate from silting or fouling of the cooler tubes.
3.8.1 Seal Oil Cooler Inlet Temperature (H2 Seal Outlets)

3.8 Seal Oil System
Heat exchangers are provided for heat removal from the seal oil. Raw water
the heat from the seal oil circulating on the other side of the heat exchanger.
3.8.2 Seal Oil Cooler Outlet Temperature (H2 Seal Inlets)

The temperature of the hydrogen seals varies with the inlet temperature of
the seal oil. It is important to keep the inlet seal oil temperature within the
recommended limits for the proper operation of the seals and to reduce the
possibility of H2 seal overheating.
38 NMAC Tech Note

High seal oil inlet temperature to the rotor H2 seals occurs when the seal oil
outlet temperature from the seal oil system is too high. This can be caused by
reduced seal oil cooler efficiency, due to low service water flow to the seal oil
coolers or high service water inlet temperature to the seal oil coolers.
3.8.3 Seal Oil Pump, Oil Outlet Pressure

The seal oil pump outlet pressure is monitored to provide a warning of the
loss of seal oil flow due to a supply pump problem and to ensure that the
pressure is maintained within limits for the H2/seal oil differential pressure.
Seal oil pressure that is too high can cause excess seal oil leakage into the
generator, which will cause oil contamination problems with many of the
internal generator components.
Low seal oil pump outlet pressure implies a problem with or failure of the
seal oil pump and creates a no- or low-flow condition of seal oil to the rotor
H2 seals. The result is a rise in the H2 seal temperatures if the oil flow is not
re-established immediately.
Failure of the source of seal oil supply is extremely serious because seal oil is
required to maintain the hydrogen pressure in the generator casing. In the
event of a serious failure of the H2 seals, the unit is usually shut down as
soon as possible and purged with CO2 to avoid a hydrogen fire, seal damage,
and possible damage to the rotor shaft.
There are often redundant seal oil pumps for backup and, in some cases, a dc
emergency pump to allow safe unit shutdown if the shaft-driven or ac
pumps are unavailable for some reason.
3.8.4 Seal Oil System, Oil Outlet Pressure (H2 Seal Inlets)
The seal oil system outlet pressure (inlet pressure to the H2 seals) is moni-
tored to ensure that there is oil flow to the H2 seals. The system outlet pres-
sure is also monitored to ensure that the pressure is maintained within limits
for the H2/seal oil differential pressure and proper operation of the H2 seals.
Seal oil pressure that is too high can cause excess seal oil leakage into the
generator, which will cause oil contamination problems with many of the
internal generator components.
Reduced seal oil flow to the H2 seals is an indicator of fouled or plugged seal
oil coolers. Fouled or plugged seal oil coolers on the oil side are most likely
the result of debris in the seal oil or seal oil that is in poor condition, contain-
ing sludge that has built up in the seal oil coolers.
A large leak in the seal oil system before the bearings also creates a low
Low H2 seal inlet pressure is dangerous if it goes below the hydrogen gas
pressure in the generator casing. This allows hydrogen leakage from the
generator, creating a safety hazard.
NMAC Tech Note 39

3.8.5 Seal Oil Filter, Differential Oil Pressure

Seal oil filter differential pressure is used to indicate the operating condition
of the filters and the overall seal oil system. Low differential pressure on the
filters is an indication of reduced seal oil flow, while high differential pres-
sure is an indication of high seal oil flow or a plugged filter.
Fouled or plugged seal oil filters indicate that the filter has collected an
excessive amount of sludge or debris from the seal oil and that the condition
of the seal oil is in question, or the seal oil filter has not been cleaned or
maintained for some time. A plugged seal oil filter causes a seal oil flow
blockage, resulting in reduced seal oil flow to the H2 seals.
3.8.6 Seal Oil Tank Level

The seal oil tank level is monitored to ensure that there is always a supply of
seal oil for the seal oil pump to draw on.
Low seal oil level in the tank implies a significant loss of seal oil in the
system. A seal oil system leak or excessive oil leakage into the generator is a
possible reason for a low seal oil tank level.
3.8.7 Seal Oil Flow

Seal oil flow to the H2 seals is monitored to ensure proper operation of the
seals. Seal oil flow can be indicated by a flow indicator or by low seal oil
pressure at the H2 seal oil inlets.
Low oil flow to the H2 seals can be caused by a flow blockage somewhere in the
seal oil system, a failure of the seal oil supply pump, or a large seal oil leak from
the seal oil system before the inlet to the H2 seals. Low seal oil flow results in the
H2 seal temperatures increasing above their normal operating range. The design
seal oil flow should be maintained to avoid damage to the seals.
3.8.8 Seal Oil Vacuum Tank Pressure

The seal oil vacuum tank pressure is maintained near vacuum to assist in
removal of entrapped hydrogen and moisture from the seal oil.
Loss of the vacuum causes improper operation of the seal oil vacuum system
and can result in oil foaming in the seal oil vacuum tank. This allows mois-
ture to carry over into the generator casing and adversely affects the dew-
point temperature of the hydrogen gas.
Failure of the seal oil vacuum system requires the hydrogen dryers (if pro-
vided) to work harder to keep the dew point of the H2 gas at design levels.
Seal oil vacuum treatment malfunction can occur from failure of the seal oil
vacuum pump or a leak in the vacuum tank.
40 NMAC Tech Note

3.8.9 Raw Service Water Inlet Pressure to Seal Oil Coolers

The service water inlet pressure to the seal oil coolers is monitored to ensure
removal from the seal oil.
Low raw water inlet pressure to the seal oil coolers is an indication of low
flow to the coolers, which affects the cooling of the seal oil. Low service water
flow to the seal oil coolers can be caused by either a failure of the service
water supply pump or a large leak before the coolers. Low pressure/flow to
the seal oil coolers results in reduced cooler efficiency and causes the seal oil
temperature to rise.
3.8.10 Raw Service Water Inlet Temperature to Seal Oil Coolers

Raw service water is circulated through the seal oil coolers on one side to
remove the heat from the seal oil flowing through the other side of the heat
exchanger.
The service water inlet temperature to the seal oil coolers is monitored to
maintain the correct seal oil inlet temperature to the rotor H2 seals. It is also
High service water inlet temperature to the seal oil coolers usually occurs in
the summer months. If the temperature of the inlet service water is too high,
it can cause a reduced efficiency problem in the cooler. This results in the
temperature of the seal oil rising above its normal operating level.
3.8.11 Raw Service Water Outlet Temperature from Seal Oil Coolers
problems with the seal oil coolers. The service water outlet temperature
should be proportional to the inlet service water temperature.
Low service water flow to or in the seal oil coolers causes high outlet water
perature of the seal oil rising above its normal operating level. The low flow
All water-cooled stator windings require an external system to deliver

3.9 Stator Cooling
Water System cooling water to the stator winding (and other water-cooled components).
The SCW system is regarded as a closed system, and every major generator
manufacturer has a specific philosophy regarding its design and the quality
of the water used for cooling generator components. For stator windings, the
water must be conditioned by the SCW system to operate in conductor bars
that are energized at high voltage and carrying high current. Therefore, the
water must be very pure with minimal conductivity. To accomplish this, the
NMAC Tech Note 41

following are the general components that are found in all SCW systems
from the different manufacturers.
In all SCW systems, the various parameters involved are monitored at
different points in the system. The general parameters monitored are:
• SCW inlet to generator temperature
• SCW outlet from generator temperature
• SCW inlet pressure to generator
• SCW outlet pressure from generator
• Differential pressure across filters and strainers
• System SCW flow
• SCW conductivity
• Storage and make-up tank levels
In addition, the raw service water for the SCW coolers is also monitored for
the inlet and outlet temperature of the coolers.
3.9.1 SCW Cooler, SCW Inlet Temperature (Stator Winding Outlet)

Heat exchangers are provided for heat removal from the SCW. Raw water
the heat from the demineralized SCW circulating on the other side of the heat
exchanger.
3.9.2 SCW Cooler, SCW Outlet Temperature (Stator Winding Inlet)

Generally, the SCW inlet temperature is maintained below 50˚C, but usually
operates in the 35–40˚C range. It is desirable to keep the stator core, winding,
and hydrogen gas temperatures in the same range to minimize thermal
differentials.
High SCW outlet temperature from the SCW system is often caused by
reduced SCW cooler efficiency from either low service water flow to the
SCW coolers or high service water inlet temperature to the SCW coolers.
High SCW outlet temperature from the SCW system causes the SCW inlet
temperature to the stator winding to be too high and subsequently the
temperature of the stator winding rises above its normal operating range.
3.9.3 SCW Pump, SCW Outlet Pressure

The SCW pump outlet pressure is monitored to indicate that there is SCW flow
to the stator windings. There are usually two ac motor-driven pumps (one in use
42 NMAC Tech Note

and the other as a backup) used to deliver the cooling water to the windings. In
some instances, a dc motor-driven pump is used for emergency shutdown.
An SCW supply pump failure implies loss of SCW supply to the generator.
The loss of SCW pump outlet pressure indicates an SCW flow problem and
results in a rise in the temperature past the normal operating range for the
stator winding and other components cooled by the SCW system. The gen-
erator can be operated at some minimum specified load without stator
cooling water, but this is generally less than 25% of the rated full load (con-
sult the generator manual for the OEM-set limits). If the unit is operated
above this minimum load, the stator winding is likely to overheat.
3.9.4 SCW System, SCW Outlet Pressure (Stator Winding Inlet)

The SCW system outlet pressure (inlet pressure to the stator winding) is
monitored to ensure that there is SCW flow to the winding. Loss of SCW to
the stator winding causes loss of coolant to the stator winding and a subse-
quent rise in the stator winding temperature.
Fouled or plugged SCW coolers cause a flow blockage and reduced SCW
flow to the winding. Fouled or plugged SCW coolers on the SCW side are
most likely the result of debris in the SCW or corrosion of the SCW coolers
due to poor SCW chemistry. Backflushing is often done to remove whatever
debris or other problem is causing the flow blockage.
A large leak in the SCW system before the generator also creates a low
3.9.5 SCW Filter (Strainer), SCW Differential Pressure

Filters, strainers, or a combination of both are employed for the removal of
debris from the SCW. Strainers are ordinarily sized to remove debris in the
20–50 micron range and larger; filters are sized for debris in the range of 3
microns and larger. Filters and strainers can be either mechanical or organic
type. Strainers are, as a rule, mechanical devices and can be cleaned by
backflushing procedures that are outlined in the manufacturer’s manual.
Filters are usually of the organic type with replaceable cartridges that should
be replaced with new cartridges if they are found dirty or fouled.
Fouled or plugged SCW filters or strainers are often caused by the collection
of debris removed from the SCW to a point where the amount of debris is
excessive. This is an indication that the filters or strainers might require
replacement or backflushing.
It is very important to ensure that the SCW filters and strainers are functioning
as required. Any solid particles in the SCW that are not collected by the filters
will make their way to the inlets of the stator conductor bars. The hollow
strands of the bars represent the narrowest portion of the SCW flow path and
also have numerous turns and angles in the internal flow path due to the
endwinding formation of the stator bars and the Roebel twisting employed.
The hollow strands become a likely location for debris to build up.
NMAC Tech Note 43

The SCW filter/strainer differential pressure is used to indicate the operating

condition of this equipment. Low differential pressure on the filters can
indicate a reduced SCW flow, while high differential pressure can indicate a
high SCW flow or a plugged filter.
3.9.6 SCW Flow

SCW flow to the stator winding can be measured directly by a flow measure-
ment device or indicated indirectly by the SCW inlet pressure to the generator.
Continuous cooling water flow is essential to carry generated heat away.
Flow velocities are design-specific and are based on such things as heat-
carrying capacity of the water, cross-sectional flow area in each bar, corrosion
effects on the copper, and so on. It is, therefore, important to have some
indication that there is SCW flow, especially during high load operation.
Low SCW flow can be due to a flow blockage, a large SCW leak in the
external SCW system, or a failure of the SCW supply pumps. This results in
SCW starvation of the stator winding and causes the temperature of the
winding to rise above the normal operating levels. The degree of low flow is
generally dependent on the root cause and urgency for corrective action,
based on the temperature of the stator winding components.
In addition to debris and copper corrosion build-up, a flow blockage in the
stator winding water path inside the generator can occur from hydrogen into
SCW leaks. When this occurs, there is a possibility of H2 gas locking, which
causes an SCW flow problem.
As an aside, there is almost invariably a certain amount of SCW that leaks
from any opening or fault present in the stator bar, even though the hydro-
gen is maintained at a higher pressure than the SCW. This is due to the
capillary action of water. The leaking SCW is detrimental to the groundwall
insulation and can cause a bar failure if left uncorrected.
3.9.7 SCW Conductivity

A deionizing subsystem is required to maintain low conductivity in the SCW,
normally in the order of 0.1 microsiemens/centimeter (µS/cm). High con-
ductivity of the demineralized cooling water can cause electrical flashover to
ground by tracking, particularly at the Teflon hoses where an internal track-
ing path to ground exists.
High conductivity in the SCW can be the result of a failure of the SCW
deionizer, but it can also be caused by internal erosion of the stator conductor
bar strands from cavitation or by chemical attack on the copper itself.
Failure of the stator cooling water system deionizer allows the conductivity
of the stator cooling water flowing through the stator winding to rise. The
deionizer usually polishes a small percentage of the SCW on a continual
basis (about 5% of the total SCW flow). If left uncorrected, the conductivity
of the SCW continues to rise slowly, risking a possible phase to ground
44 NMAC Tech Note

leakage current through the increasingly conducting SCW. A ground fault is

possible in the worst case.
3.9.8 SCW Make-Up Tank Level

In the event SCW is lost, or when filling the SCW system after shutdown and
draining, the SCW system requires replenishing. Therefore, a storage tank to
hold sufficient make-up water is required. Some systems are open to the
atmosphere, while others maintain a hydrogen blanket on top of the water to
keep the level of oxygen at a minimum. The storage tank arrangement is
manufacturer-specific, depending on the desired water chemistry.
3.9.9 SCW System H2 Detraining Tank Level

Because no SCW system is leak perfect, there is some ingress of hydrogen (for
example, permeation of hydrogen through Teflon hoses) and natural collec-
tion of other gases, such as oxygen, in the SCW system. A means for venting
these gases is required. Usually, the excess gases are vented to the atmo-
sphere. In some systems, the venting process is monitored or counted, and in
other systems, it is not. This is manufacturer-specific.
3.9.10 Raw Service Water Inlet Pressure to SCW Coolers

The service water inlet pressure to the SCW coolers is monitored to ensure
removal from the SCW.
Low raw water inlet pressure to the SCW coolers is an indication of low flow
to the coolers, which affects the cooling of the SCW. Low service water flow
to the SCW coolers can be caused by either a failure of the service water
supply pump or a large leak before the coolers. Low pressure or flow to the
SCW coolers results in reduced cooler efficiency and causes the SCW tem-
perature to rise.
3.9.11 Raw Service Water Inlet Temperature to SCW Coolers

Raw service water is circulated through the SCW coolers on one side to
remove the heat from the SCW flowing through the other side of the heat
exchanger.
The service water inlet temperature to the SCW coolers is monitored to
maintain the correct SCW inlet temperature to the stator winding. It is also
High service water inlet temperature to the SCW coolers usually occurs in the
summer months. If the temperature of the inlet service water is too high, it
can cause a reduced efficiency problem in the cooler. This results in the
temperature of the SCW rising above its normal operating level.
NMAC Tech Note 45

3.9.12 Raw Service Water Outlet Temperature from SCW Coolers

problems with the SCW coolers. The service water outlet temperature should
be proportional to the inlet service water temperature.
Low service water flow to or in the SCW coolers causes high outlet water
perature of the SCW rising above its normal operating level. The low flow
This section shows the types of sensors used in on-line monitoring.

3.10 Sensors and
Instrumentation
3.10.1 Generator Electrical Parameters
Sensor Type Units

Generator Output Power MW Transducer MW
Generator Reactive Power MVar Transducer MX
Stator Three Phase Current Current Transformer Amps ac
Stator Terminal Voltage Potential Transformer Volts ac
Field Current DC CT/Shunt Amps dc
Field Voltage Voltage Transducer Volts dc
Frequency Transducer Hertz
Volts per Hertz Mathematical Model p.u.
Negative Sequence Mathematical Model %
Generator Main Breaker Status Relay Contact Open/Closed
Generator Field Breaker Status Relay Contact Open/Closed
3.10.2 Stator Core and Frame
Sensor Type Units

Core Temperature TC/RTD Deg. C/F
Core End Clamping Plate Temperature TC/RTD Deg. C/F
Core End Flux Screen Temperature TC/RTD Deg. C/F
Core Vibration Accelerometer/Fiber Optics mils/µm, p-p
Frame Vibration Accelerometer/Fiber Optics mils/µm, p-p
Liquid Level in Generator Casing Tank with Level Indicator High/ Normal
46 NMAC Tech Note

3.10.3 Stator Winding
Sensor Type Units

SCW Inlet Temperature TC/RTD Deg. C/F
Bulk SCW Outlet Temperature TC/RTD Deg. C/F
Conductor Bar Hose Outlet Temperature TC/RTD Deg. C/F
Conductor Bar Slot Temperature TC/RTD Deg. C/F
Stator Winding Differential Temperature TC/RTD Deg. C/F
SCW Inlet Pressure Pressure Transducer Psi/Kpa
Stator Winding Differential Pressure Pressure Transducer Psi/Kpa
Hydrogen Gas in SCW Special Cu. Ft. per Day
Stator Winding Ground Alarm Relay Contacts Open/Closed
3.10.4 Rotor
Sensor Type Units

Rotor Winding Temperature Mathematical Model Deg. C/F
Rotor Winding Ground Alarm Relay Contacts Open/Closed
Shaft Speed Speed Probe/Transducer RPM
Rotor Vibration Vib. Probe/Transducer mils/µm, p-p
Bearing Metal Temperature TC/RTD Deg. C/F
Bearing Inlet Oil Temperature TC/RTD Deg. C/F
Bearing Outlet Oil Temperature TC/RTD Deg. C/F
Hydrogen Seal Metal Temperature TC/RTD Deg. C/F
Hydrogen Seal Inlet Oil Temperature TC/RTD Deg. C/F
Hydrogen Seal Outlet Oil Temperature TC/RTD Deg. C/F
Hydrogen Seal, H2/Seal Oil Differential Pressure Pressure Transducer Psi/Kpa
Rotor Fan Differential Pressure Pressure Transducer Psi/Kpa
NMAC Tech Note 47

3.10.5 Excitation System
Sensor Type Units

AC Power into Exciter Transducer MW
DC Power out of Exciter Transducer MW
Main Exciter, Cooling Air Inlet Temperature TC/RTD Deg. C/F
Main Exciter, Cooling Air Outlet Temperature TC/RTD Deg. C/F
Collectors and Brushgear, Cooling Air TC/RTD Deg. C/F
Inlet Temp.
Collectors and Brushgear, Cooling Air TC/RTD Deg. C/F
Outlet Temp.
Collectors and Brushgear, Cooling Air Pressure Transducer Psi/Kpa
Filter Diff. Press.
3.10.6 H2 Cooling System
Sensor Type Units

Bulk Hydrogen Supply Pressure Pressure Transducer Psi/Kpa
Generator Casing Hydrogen Gas Pressure Pressure Transducer Psi/Kpa
Cold Hydrogen Gas Temperature TC/RTD Deg. C/F
Hot Hydrogen Gas Temperature TC/RTD Deg. C/F
Hydrogen Dewpoint Temperature Dewpoint Meter Deg. C/F
Hydrogen Purity Purity Meter/ Katherometer %
Hydrogen Make-Up or Leakage Rate Special Cu. Ft. per Day
Raw Service Water Inlet Pressure to H2 Cooler Pressure Transducer Psi/Kpa
Raw Service Water Inlet Temp. to H2 Cooler TC/RTD Deg. C/F
Raw Service Water Outlet Temp. from H2 Coolers TC/RTD Deg. C/F
48 NMAC Tech Note

3.10.7 Lube Oil System
Sensor Type Units

Lube Oil Cooler Inlet Temperature TC/RTD Deg. C/F
(Bearing Outlets)
Lube Oil Cooler Outlet Temperature TC/RTD Deg. C/F
(Bearing Inlets)
Lube Oil Pump, Oil Outlet Pressure Pressure Transducer Psi/Kpa
Lube Oil System, Oil Outlet Pressure Pressure Transducer Psi/Kpa
(Bearing Inlets)
Lube Oil Filter, Differential Oil Pressure Pressure Transducer Psi/Kpa
Lube Oil Tank Level Tank Level Indicator Low/ Normal/
High
Lube Oil Flow Flow Transmitter GPM/LPS
Raw Service Water Inlet Pressure to Lube Pressure Transducer Psi/Kpa
Oil Coolers
Raw Service Water Inlet Temp. to Lube TC/RTD Deg. C/F
Oil Coolers
Raw Service Water Outlet Temp from Lube TC/RTD Deg. C/F
Oil Coolers
3.10.8 Seal Oil System
Sensor Type Units

Seal Oil Cooler Inlet Temperature TC/RTD Deg. C/F
(H2 Seal Outlets)
Seal Oil Cooler Outlet Temperature TC/RTD Deg. C/F
(H2 Seal Inlets)
Seal Oil Pump, Oil Outlet Pressure Pressure Transducer Psi/Kpa
Seal Oil System, Oil Outlet Pressure Pressure Transducer Psi/Kpa
(H2 Seal Inlets)
Seal Oil Filter, Differential Oil Pressure Pressure Transducer Psi/Kpa
Seal Oil Tank Level Tank Level Indicator Low/ Normal/
High
Seal Oil Flow Flow Transmitter GPM/LPS
Seal Oil Vacuum Tank Pressure Pressure Transducer Psi/Kpa
Raw Service Water Inlet Pressure to Seal Pressure Transducer Psi/Kpa
Oil Coolers
Raw Service Water Inlet Temp. to Seal TC/RTD Deg. C/F
Oil Coolers
Raw Service Water Outlet Temp. from Seal TC/RTD Deg. C/F
Oil Coolers
NMAC Tech Note 49

3.10.9 Stator Cooling Water System
Sensor Type Units

SCW Cooler, SCW Inlet Temperature TC/RTD Deg. C/F
(SW Outlet)
SCW Cooler, SCW Outlet Temperature TC/RTD Deg. C/F
(SW Inlet)
SCW Pump, SCW Outlet Pressure Pressure Transducer Psi/Kpa
SCW System, SCW Outlet Pressure (SW Inlet) Pressure Transducer Psi/Kpa
SCW Filter (Strainer), SCW Differential Pressure Pressure Transducer Psi/Kpa
SCW Flow Flow Transmitter GPM/LPS
SCW Conductivity Conductivity Cell µ siemens (S)
SCW Make-Up Tank Level Tank Level Indicator Low/Normal/High
SCW System H2 Detraining Tank Level Tank Level Indicator Low/Normal/High
SCW Cooler, LPSW Inlet Pressure Pressure Transducer Psi/Kpa
SCW Cooler, LPSW Inlet Temperature TC/RTD Deg. C/F
SCW Cooler, LPSW Outlet Temperature TC/RTD Deg. C/F
50 NMAC Tech Note

4
SPECIALIZED MONITORING DEVICES
The previous section dealt with the most common sensors and monitoring
instrumentation that are installed on a generator. This section deals with the
more specialized monitoring devices that are not normally connected to the
unit main computer or generator monitoring system. These are usually
installed as stand-alone devices that provide information to the operator
apart from the main monitoring system; however, that is not to say that some
have not been connected to the main monitoring system in some form on
some machines in service.
Partial discharge (PD) in the large generator application is largely associated

4.1 Stator Winding
Partial Discharge with the high voltage stator conductor bars. PD occurs in the stator winding
Monitoring most commonly in the form of:
• Slot discharges from bar bouncing
• Slot discharges from loss of the high resistive coating for the slot
• Internal discharges from insulation voids or delaminations
• Surface discharge due to contamination or moisture in the winding
All of the above mechanisms, which promote PD, create areas of voltage
stress that allow electrical charge to build up and discharge. The effect is to
cause possible damage to the voltage grading systems on the bars, the
interstrand insulation, or the groundwall insulation.
NMAC Tech Note 51

A failure of the stator winding insulation is very costly to fix, both in terms of
capital cost to repair or replace a stator bar and the outage time required to
complete the work. Therefore, much effort has been invested over the years
in developing techniques to identify the occurrence of PD in the stator
winding. There are methods for off-line detection of PD, but they do not
cover the operating effects that also promote PD activity, such as those due to
thermal and vibration effects.
To provide the best PD detection, an on-line method of monitoring is re-
quired. There are now three basic approaches to on-line PD monitoring,
which all have attained recognized status as viable methods. The following is
a description of each.
4.1.1 RF Monitoring
Radio frequency (RF) monitoring is a technique used to detect electrical
sparking and arcing or stator winding PD inside the generator. This type of
monitoring operates on the premise that arcing in the stator winding causes
RF currents to flow in the neutral of the winding. The types of problems or
failure mechanisms that cause these RF currents to flow are:
• Conductor bar strand cracking
• Electrical joint failure
• Partial discharges from insulation problems
To monitor these currents, a high frequency current transformer (CT) is
placed around the neutral grounding lead before the neutral grounding
transformer, as shown in Figure 5.
Main
Generator Transformer
System
RF-CT
RF
Monitor
Neutral Grounding
Transformer
Figure 5
RF Monitor
52 NMAC Tech Note

The output of the CT is fed to a RF monitor for signal processing and analy-
sis. The signals from the CT are filtered to examine those that are in the
correct frequency range for radio frequency arcing. The monitor generally
has a setpoint or an alarm limit that can be adjusted to a predetermined level
at which the RF activity is known to be excessive. When the level of RF
activity increases to where the setpoint is reached, the operator is notified of
an RF problem in the generator.
It is difficult, however, to distinguish between the sources of RF arcing, and it
is not always possible to identify the root cause. In most cases, one can only
say that arcing at a certain level is occurring. In addition, filtering does not
completely eliminate noise; therefore, this also becomes a problem in signal
discrimination. This is especially true for large turbo-generators where the
noise generated from the sliprings/brushgear and the shaft grounding
brushes is considerable.
4.1.2 Capacitive Coupling

Capacitive coupling has been in use since the early 1950s and was developed
as an alternative to RF monitoring [4, 5, 6]. In contrast to RF monitoring, where
detection of PD is at the generator neutral, capacitive coupling is done at the
line ends of the generator winding, that is, at the output of each of the phases.
This is an improvement because PD can be detected on a per phase basis.
To measure the PD activity on each phase, a tuned capacitor or “capacitive
coupler” is connected to each of the generator terminals. In turn, the couplers
are then connected to an resistive-capacitive (RC) bandpass filter for which the
output is a signal containing the high frequency PD pulses. The PD pulses are
then displayed on a high-speed oscilloscope where they can be categorized
(that is, negative and positive PD pulses) and quantified (that is, the magni-
tudes of the negative and positive PD pulses) by a technician with sufficient
skill. Although capacitive coupling has proven to be instrumental in on-line
detection of PD, it does requires a great deal of experience in distinguishing the
PD pulses from noise in the generator and from the system.
One of the disadvantages of this test is that the capacitive couplers are
normally temporary on steam turbine generators and require live line instal-
lation, a very dangerous procedure with risk to the technician.
After the couplers are installed, three tests are usually done to further iden-
tify the type of PD activity occurring. The first is done off-line but with the
generator on open circuit at rated terminal voltage, the second is done with
the generator on-line at low load, and the third, also on-line, is done at high
load. Measurements are taken on each phase for each of the three conditions.
The loading on the generator is varied to distinguish between the types of PD
activity present. Because temperature and vibration have a significant effect
on the level of PD activity, changing the loading conditions causes variations
in terminal voltage and stator current. Hence, there are variations in winding
temperature and stator bar bounce forces or bar vibration.
NMAC Tech Note 53

Because positive PD is associated with surface discharges, one could con-

clude, for instance, that if the magnitude of the measured positive PD pulses
increased with load and if there were no increase in the magnitude of the
negative PD pulses, then there might be slot discharge activity occurring due
to loose stator wedges. Similarly, if there is equal positive and negative PD
activity, one could conclude that there is discharge in the bulk of the ground
wall insulation, and so on.
In addition to the above, efforts to verify the source of the PD pulses from the
generator, or within the isolated phase bus (IPB), or external to the system
have also been made using two capacitive couplers per phase installed on
the IPB where they can be separated by several feet. When a PD pulse is
measured on both couplers, its direction can be determined by which coupler
is the first to see the pulse. Therefore, if the coupler furthest from the genera-
tor picks up the PD pulse first, then its direction is toward the generator. And
vice versa, if the coupler closest to the generator is the first to see the pulse,
then it had to come from the generator. There has been some progress made
in the directional capability of capacitive coupling but, again, noise is a
problem and often masks the true PD being measured.
Interpretation of PD measurements is not an exact science and requires a
great deal of experience to determine the type of PD activity that a particular
machine might be experiencing.
4.1.3 Stator Slot Coupler

The stator slot coupler (SSC) is basically a tuned antenna with two ports. The
antenna is approximately 18 inches long and is embedded in an epoxy/glass
laminate with no conducting surfaces exposed. SSCs are installed under the
stator wedges at the line ends of the stator winding, so that the highest
voltage bars are monitored for the best PD detection. Because the SSC is also
installed lengthwise in the slot at the core end, its two-port characteristic
gives it inherent directional capability [4, 5, 6].
54 NMAC Tech Note

SSC under
Stator Wedges
Stator
Core
Stator Bar
under SSC and
Wedges
Endwinding
Stator
Wedges
SSC Output
Cables
Figure 6
SSC Position in Slot, Bore View
NMAC Tech Note 55

Wedge
Wedge Slider
Depth
Packing
SSC
Top Pad
Stator
Core
Top
Bar
Side
Packing Slot
Separator
Bottom
Bar
Bottom
Pad
Figure 7
Cross-Section of Stator Slot Containing an SSC
The problem of noise is virtually eliminated with the SSC. Although the SSC has
a very wide frequency response characteristic that allows it to see almost any
signal present in the slot where it is installed, it also has the characteristic of
showing the true pulse shape of these signals. This gives it a distinct advantage
over other methods, which cannot capture the actual nature of the PD pulses.
Since PD pulses occur in the 1–5 nanosecond range and are very distinguishable
with the SSC, the level of PD activity can be more closely defined.
In addition, dedicated monitoring devices have been devised to measure the
PD activity detected in the SSC. The capability for PD detection using the
SSC and its associated monitoring interface is enhanced to include measure-
ment in terms of:
• The positive and negative characteristic of the pulses
• The number of pulses
• The magnitude of the pulses
• The phase relation of the pulses
• The direction of the pulses (that is, from the slot or from the endwinding
or actually under the SSC itself at the end of the slot)
56 NMAC Tech Note

The other advantage of an SSC is that, once it is installed, measurements can

be taken at any time without the need for exposing live portions of the
generator bus work to make connections.
Shorted turns in rotor windings are associated with turn-to-turn shorts on the
4.2 Rotor Winding
Shorted Turns copper winding as opposed to turn-to-ground faults (See Figure 8).
Detection
Rotor Rotor
Tooth Wedge
Slot Liner
Creepage
Block
Copper
Winding Shorted Turn
Location
Interturn
Insulation
Axial
Radial Vent
Vent
Sub
Slot
Figure 8
Rotor Winding Shorted Turn
Rotor winding shorted turns, or inter-turn shorts, can occur from:

• An electrical breakdown of the inter-turn insulation
• Mechanical damage to the inter-turn insulation allowing adjacent turn-
to-turn contact
• Contamination in the slot that allows leakage currents between turns
When shorted turns occur, the total ampere-turns produced by the rotor is
reduced because the effective number of turns has been reduced by the
number of turns shorted. The result is an increase in required field current
input to the rotor to maintain the same load point and an increase in rotor
winding temperature.
At the location of the short, there is a high probability of localized heating of
the copper winding and arcing damage to the insulation between the turns.
NMAC Tech Note 57

This type of damage can propagate and worsen the fault so that more turns
are affected or the groundwall insulation becomes damaged and a rotor
winding ground occurs.
One of the most noticeable effects of shorted turns is increased rotor vibra-
tion due to thermal effects. When a short on one pole of a rotor occurs, a
condition of unequal heating in the rotor winding exists between poles. The
unequal heating causes bowing of the rotor and, hence, vibration. The extent
and location of the shorted turns and the heating produced governs the
magnitude of the vibrations produced.
Off-line methods for detecting shorted turns include winding impedance
measurements while the rotor speed is varied from zero to rated speed and
RSO (recurrent surge oscillation) tests based on the principle of time domain
reflectometry. In addition, a short of significant magnitude can be identified
by producing a open circuit (OC) saturation curve and comparing it to the
design OC saturation curve. If the field current required to produce rated
terminal voltage has increased from the original design curve, then a short is
likely present. The number of shorted turns can be identified by the ratio of
the new field current value over the design field current value.
All of the above methods of identifying shorted turns are prone to error and
indicate only that a short exists. They do little to help locate which slot the
short is in and require special conditions for collecting the data or for testing.
To better identify shorted turns and to employ a method that works on-line,
the shorted turns detector (STD) or search coil method has been perfected.
Each OEM has its own version of the STD, but all work essentially in the
same manner.
The STD is actually a search coil mounted on the stator core but located
strategically in the air gap. The search coil looks at the variation in the
magnetic field produced in the air gap by the rotor as it spins. The energized
rotor winding and the slotted effect of the winding arc causes a sinusoidal
signal to be produced in the winding face of the rotor. The pole face, on the
other hand, has no winding, and the signal is flatter because the variation in
the magnetic field is minimal (see Figure 9).
58 NMAC Tech Note

Slot Centers
Quadrature
Pole #2 Axis
Pole #1 Direct Pole #1

Axis
Quadrature
Axis
Direct Direct
Axis Axis
Figure 9
Rotor Winding Shorted Turns Detector, Typical Flux Probe Signal On-Load
The magnitude of the sinusoidal peaks in the winding face is dependent on
the ampere-turns produced by the winding in the various slots. If there is a
short in a slot, the peak of the signal for the affected slot is reduced (see
Figure 10). The reduction is dependent on the magnitude of the short. There-
fore, as well as knowing which slot the short is in, an estimate of the number
of shorted turns can be made fairly accurately.
Slot with Shorted Turn
Quadrature
Pole #2 Axis
Pole #1 Direct Pole #1

Axis
Quadrature
Axis
Direct Direct
Axis Axis
Figure 10
Rotor Winding Shorted Turns Detector, Typical Signal
with Shorted Turn Indicated
NMAC Tech Note 59

Problems due to saturation effects at full load can occur in analyzing the
data, and most OEMs now have a dedicated monitor connected to the STD to
automate the analysis process. This allows the STD and monitor to act as a
standalone sensor to alarm when a short turn is detected and notify the
operator for investigation.
Endwinding vibration can damage the mechanical integrity of the stator

4.3 Stator
Endwinding conductor bars and the other stator winding components that make up the
Vibration Monitoring electrical and cooling water delivery portions of the total stator winding.
Endwinding vibration is usually a symptom of loose endwinding support
structures. Checking and retightening these support structures are required
to avoid damage to the stator winding and associated hardware.
Vibration monitoring can be employed to indicate increasing vibration levels
and plan outages to make repairs.
Damage from vibration can initiate leak problems from cracked strands,
broken braze joints, and loosened fittings.
There are many different endwinding support system designs in operation.
Some are designed with the intent that they do not need retightening, but
loosening can still occur naturally over time due to vibration from the forces
induced by the high ac electromagnetic fields inside the machine. Other
designs are provided with mechanisms for periodic retightening and allow
for natural loosening. Again, the degree of natural loosening and the time it
takes depends on the design of the endwinding support system and the
method and quality of the factory installation.
Premature loosening can be enhanced by excessive seal oil ingress, which
also creates a greasing effect. The black grease is seen at ties and blocking
interfaces where fretting has occurred.
System faults again affect the integrity of the endwinding support system
due to the high forces induced during such events.
On stators that are oil free, looseness can be indicated by numerous paint
cracks at tie- and blocking-to-winding interfaces. In addition, there might be
some white or light gray powder from corona discharge at paint cracks,
which also can indicate some looseness.
During operation, voltages build on the generator rotor shaft. The sources of
4.4 Shaft Voltage and
Current Monitoring shaft voltage have been identified as:
• Voltage from the excitation system due to unbalanced capacitive cou-
pling
• Electrostatic voltage from the turbine due to charged water droplets
impacting the blades
60 NMAC Tech Note

• Di-symmetric voltage from unsymmetrical stator core stacking

• Homopolar voltage from shaft magnetization
If these voltages are not drained to ground, they rise and break down the
various oil films at the bearings, hydrogen seals, turning gear, thrust bearing,
and so on. The result is current discharges and electrical pitting of the critical
running surfaces of these components. Mechanical failure can then follow.
Inadequate grounding of the rotor shaft also allows these voltages to further
build up on the rotor shaft of the generator. Inadequate grounding might be
due to a problem with the shaft grounding brushes from wear (requiring
replacement brushes) or a problem with the associated shaft grounding
circuitry if a monitoring circuit is provided.
High shaft voltages can also be caused by severe local core faults of large
magnitude that impress voltages back on the shaft from long shorts across
the core.
Protection against shaft voltage buildup and current discharges is provided in
the form of a shaft grounding device, usually located on the turbine end of the
generator rotor shaft. The grounding device consists of a carbon brush or copper
braid, with one end riding on the rotor shaft and the other connected to ground.
Shaft voltage and current monitoring schemes are also provided in many cases
to detect the actual shaft voltage level and current flow through the shaft
grounding brushes. This has the advantage of providing a warning when the
shaft grounding system is no longer functioning properly and requires mainte-
nance. There are numerous monitoring schemes available and each OEM has
its own system provided with the turbine generator (TG) set when purchased.
For older machines with only grounding and no monitoring, a monitoring
system can often be retrofitted to the existing ground brushes. The OEM
should be consulted when upgrading the shaft monitoring.
The generator condition monitor (GCM), more commonly known as the core
4.5 Generator
Condition Monitor/ monitor, is a device used to detect overheated insulation inside the generator.
Tagging Compounds Although the original intent of the GCM was to provide advance warning of
stator core insulation overheating, the fact is that any overheated organic
material inside the generator (for example, insulation) creates pyrolysis
products in the form of particulates that can be detected by the GCM.
The detection of particulates occurs in an ion chamber within the GCM. The
ion chamber is designed to work in a hydrogen atmosphere under the high
pressures used in large generators. As hydrogen is passed through the ion
chamber, the detector inside the chamber produces a constant current through
the gas flow. The current is set (usually at 80%) by regulating the gas flow and
continuously monitoring it. When particulates (that is, pyrolysis products from
NMAC Tech Note 61

overheated insulation) are present, the current drops. If the reduction in

current goes below a preset limit (usually 50%), the GCM signals an alarm.
The alarm tells the operator that a GCM event has occurred in the generator,
and a check on the validity of the alarm is now required. To assist the operat-
ing personnel in this assessment, a filter is employed to check that particu-
lates are actually present in the hydrogen gas. The operator simply pushes a
button to divert the gas flow through a filter assembly before going to the ion
chamber. If the alarm is real, the filter removes the particulates and the GCM
current returns to normal. Releasing the filter button should bring the alarm
back in as long as the particulates are still present.
One of the problems that has been experienced over the years with the GCM
is false alarms due to oil mist contamination in the hydrogen gas. The prob-
lem has been somewhat alleviated since it was found that slightly heating the
ion chamber causes the oil mist to have no effect on the current signal [7].
Therefore, many GCMs are now equipped with heated ion chambers.
The downside to the heated ion chamber is that it has also been known to
somewhat desensitize real alarms from overheated insulation. This remains a
controversial issue, and there is evidence on both sides of the argument as to
whether heating the ion chamber affects the GCM sensitivity to actual py-
rolysis products.
To further enhance the use of the GCM and to aid in determining what the
source of any alarm is, an automatic sampling assembly is also incorporated
in the GCM. When an alarm occurs, a controlled volume of hydrogen gas is
passed through the sampler, which has a removable collector. The collector
maintains the ability to capture both the particulates and the gas that passes
through it. The collector is then sent to a laboratory for analysis of the col-
lected samples. In this way, the source of the overheating or false indication
can be determined.
The downside of autosampling is that it takes time for the samples to be
analyzed while the unit is still operating. A decision can be made to continue
running on the chance that the source of overheating is not fatal to the
generator (for example, the back of the core burning, oil mist, and so on), or
the unit can be shut down until the analysis results are obtained.
Tagging compounds [8] were also developed to assist with identification of the
source of internal generator overheating. Tagging compounds are essentially a
number of different synthesized chemicals that are chemically and thermally
stable. Each of the different “tags” is mixed in trace amounts with the colored
epoxy paints that are used to coat the various internal generator components.
When a GCM alarm occurs, a sample is taken and the sample is quickly
analyzed by the on-site equipment provided. Because the tagging compounds
are known, the location of the overheating can be quickly determined.
62 NMAC Tech Note

The torsional monitor is a device or system used to monitor the occurrences

4.6 Torsional
Vibration Monitoring of shaft torques, particularly those of a severe nature that will impact on the
remaining life of the turbine generator shaft. By measuring the pertinent
parameters and entering all the event information into a model of the turbine
generator shaft, the loss of life on the shaft can be calculated.
Torsional events are most often caused by severe system disturbances or
power system resonant frequencies that are inadvertently stimulated. These
cause the TG shaft to respond by oscillating sub-synchronously on top of the
shaft operating speed. The effect is to cause excessive oscillating torque in the
shaft, which, if not dampened, can eventually cause a runaway oscillation
that will fail the shaft.
When the event does not run away and there is no apparent damage, there
might, however, have been excess torsional stress on the TG shaft that has
reduced the life of the TG shaft and needs to be accounted for. The torsional
monitor is used for this purpose.
The parameters required to be measured are stator current, terminal voltage,
and shaft speed.
Corrosion of hollow copper strands is a common source of plugging in the

4.7 Stator Cooling
Water Chemistry stator winding and occurs mainly from improper SCW chemistry [9]. In both
Monitoring high and low oxygen systems, it is preferable to avoid dissolved oxygen
contents in the range of 200–300 parts per billion (ppb) in the SCW. This is the
range in which the maximum corrosion rate occurs and, hence, the range
with the greatest risk of plugging. In addition, the pH value also affects the
corrosion rate when it decreases below the recommended levels. As a rule, a
higher pH reduces the corrosion rate.
Hollow steel strands mixed with solid copper strands have been used for the
cooling water circuit in the stator bars. The stainless steel allows a better leak
seal at the ends of the stator bars, higher mechanical strength due to nature of
the materials used, and reduced effects of corrosion.
Corrosion products can also form in the external SCW system when the SCW
chemistry is not optimal because the external piping is also usually made of
copper. In some cases though, the external system piping is composed of
stainless steel, and in these instances, the corrosion mechanism is reduced.
However, SCW chemistry must still be closely watched since even stainless
steel can have problems if not maintained.
4.7.1 SCW Oxygen Content

The dissolved oxygen content of the SCW is controlled to prevent corrosion
of the hollow copper strands. Corrosion products can build up and block the
cooling water flow. Oxygen at 200–300 micrograms/liter (µg/l) produces the
highest corrosion rate. The content of oxygen in the SCW is normally main-
NMAC Tech Note 63

tained at less than 50 ppb in hydrogen-saturated and low oxygen systems

and without limit for open-vented or high oxygen systems.
High oxygen refers to air-saturated SCW with dissolved oxygen present in
the range of greater than 2000 µg/l (ppb) at standard temperature and
pressure (STP). The high oxygen system is based on the supposition that the
surface of pure copper forms a corrosion-resistant and adherent cupric oxide
layer (CuO), which becomes stable in the high oxygen environment. The
oxide layer is generally resistant to corrosion as long as the average water
velocity is less than 15 ft/sec.
Low oxygen refers to SCW with a dissolved oxygen content that is less than 50
µg/l (ppb). A low oxygen system is based on the supposition that pure copper
does not react with pure water in the absence of dissolved oxygen. The upper
limit is set by the corrosion rate that the water cleansing system can handle.
The lower limit is set to the level where copper does not deposit on any
insulating surface in the water circuit, such as a hose. This is to avoid electrical
tracking paths to ground. A low oxygen system has better heat transfer proper-
ties at the copper/water interface and a lower copper ion release rate.
4.7.2 SCW H2 Content

When no leaks are present in the system, hydrogen content is at a minimum.
High hydrogen in the SCW can cause gas locking if the leak rate is too high.
Excessive venting of hydrogen is an indication of a high leak rate. High
concentrations of hydrogen can also cause conductivity problems.
4.7.3 Copper/Iron Content

The amount of copper and iron in the SCW is normally less than 20 ppb.
High concentrations of either can cause conductivity problems.
4.7.4 SCW pH Value

The pH value of the SCW is manufacturer-specific. Basically, there are two
modes of operation: neutral and alkaline.
Neutral pH refers to a pH value of approximately 7. A neutral pH with low
oxygen content less than 50 µg/l works best, but high oxygen over 2000 µg/l
also works. Oxygen at 200–300 µg/l produces the highest corrosion rate.
Alkaline pH refers to a high pH value of around 8.5. Again, low oxygen
content works best; however, high oxygen also works. This method requires
an alkalizing subsystem to keep the pH at the proper level.
64 NMAC Tech Note

5
DECIDING ON A MONITORING SYSTEM
This section is designed to provide guidance on determining the right system

for the particular needs of a utility.
When selecting an alternative for generator monitoring, it is helpful to know

5.1 Need
the availability and capacity requirements of the unit. A high-end monitoring
system with superior capability is not necessarily required for a machine that
might operate only 10% of the time. On the other hand, the best available
monitoring might be the only reasonable solution if the unit is required to
run 90% of the time at full load and is a primary generating unit to the utility.
A financial evaluation is required to assess the need or benefit of installing an
improved generator monitoring system and for determining the level of
sophistication required. The evaluation should include dollar values and
probability factors for the different options or various alternatives for the
monitoring systems under consideration.
One of the reasons for minimal monitoring to be employed is cost. Perma-

5.2 Economics
nently installed sensors and monitoring systems can become expensive;
therefore, the tendency is to install monitoring that gives the most value for
the dollar. This is done by prioritizing the sensors that are the most important
to be connected to the monitoring system. The benefit-to-cost ratio for pur-
chasing, installing, and connecting a particular group of sensors or a com-
plete system needs to be evaluated.
NMAC Tech Note 65

As an example, the stator winding is generally regarded as one of the most

important components to monitor. This is because it is expensive to fix or
replace when a problem occurs and because, statistically, it is a problem
component. The stator winding can deteriorate due to electrical, thermal, or
mechanical effects. It requires maintenance, such as periodic wedge re-
tightening. Water-cooled stator windings can also incur problems due to
poor stator cooling water chemistry that also tends to create deterioration.
The point is that in a cost-benefit analysis, it is likely that a high standard of
stator winding monitoring on a large nuclear unit will be justifiable.
There are also a number of components of cost to consider when trying to
justify a monitoring system:
• Actual capital cost of the equipment
• Installation of additional sensors
• Installation of specialized monitoring devices
• Connection of the points to be monitored back to the monitoring system
These costs must be factored into an analysis of the benefits of having en-
hanced monitoring as opposed to maintaining the status quo.
One of the best ways to carry out an analysis is to determine the savings that
would be realized by the detection of a major failure with an enhanced moni-
toring system in place. Therefore, one needs to consider the cost of the failure
in terms of the capital loss of equipment and the revenue loss or replacement
energy cost versus the cost to install the new monitoring system and the
perceived benefits, that is, the cost/benefit ratio and the payback period.
The revenue loss is the dollar value of the lost energy sales if another utility
supplies the power. If the power is supplied by the same utility, it is termed
the replacement energy cost, which is the cost incurred by supplying the
needed energy from another source. For base load equipment, that is, a
nuclear unit, the replacement energy cost would be higher when the replace-
ment energy comes from an oil-fired unit because the cost to run the unit is
more expensive. As a rule, most utilities run the units with the lowest operat-
ing cost first. Therefore, it stands to reason that there is always likely to be a
replacement energy cost incurred during refurbishment work.
Cost/benefit ratio is a ranking used between the different viable alternatives
to determine which is the most economically attractive. Each alternative is
assessed as to its expected reliability and benefit in terms of dollar return,
and the capital and replacement energy cost to implement the alternative.
The dollar benefits are divided by the dollar cost to establish a ratio. The
highest ratio is generally the best alternative because the payback is greatest.
However, the payback period also becomes important if it is too long. In
some circumstances, although the cost/benefit ratio might be very good, the
payback period might be unacceptable in terms of utility policies on money
management. Payback periods in the order of three to five years or less are
66 NMAC Tech Note

desired in today’s economy. Therefore, the alternative most often pursued is

the one with low capital cost, low replacement energy cost, high benefit, high
cost/benefit ratio, and short payback period.
Risk is often associated with the phrase “risk of a failure.“ When a known or
5.3 Risk
generic problem exists in the generator, the probability of a failure occurring
is set higher than if no problem is suspected. Taking a risk (gambling) that
the equipment will not fail is usually done in times when the unit has a high
priority and it is desirable to defer repairs until a more convenient time.
The benefits of risk taking are the dollar value savings realized if a failure
does not occur in the timeframe specified. However, if the equipment fails
within the timeframe, the cost to the utility can be tremendous, especially if
non-performance penalties are involved.
Risk assessment should be done by highly specialized personnel with a great
deal of expertise on the equipment under consideration. This helps to mini-
mize the risk. The specialist is likely to make a more qualified judgment with
a higher probability of success.
However, one of the ways to offset risk is to depend on monitoring. With the
help of a sophisticated monitoring system, the condition of the equipment
can be watched more closely, and in many cases, a major failure can be
avoided. For many types of failure modes, no amount of monitoring will
prevent major damage, but for numerous types of slowly progressing failure
modes, increased monitoring can allow extended operation without the
machine experiencing a failure. This allows the unit to be taken out of service
gracefully and have less expensive maintenance performed rather than
suffering a forced outage.
The benefits of good monitoring in such cases are:
• Continued revenue from generating
• Lower cost to repair the problem
• No incurring replacement energy costs from forced outages
• Reduced outage times due to the avoidance of a failure
Generator monitoring systems consist largely of field sensors and devices

5.4 System
Maintenance and that are connected back to a central monitoring system. Maintenance of these
Obsolescence components is usually taken care of during the normal maintenance of the
generator during outages or repaired as required.
Monitoring system maintenance is dependent on the type of system that is
employed. If all monitoring is done by local gauges and/or the unit com-
puter, this also tends to be covered by the normal unit (hardware) mainte-
NMAC Tech Note 67

nance that is done. However, for stand-alone monitoring systems, the prob-
lem becomes more of a spare-parts-dependent situation. Monitoring systems
are not generally redundant on individual generators; therefore, if the system
goes down, monitoring is lost until it is restored.
If the problem is related to hardware, the problem component is removed
and replaced. The damaged component can then be taken away and re-
paired, if possible, or replaced with another purchased spare. It becomes
extremely important that spares are available even when the system becomes
obsolete. In today’s computer climate, this will occur rapidly. To deal with
this obsolescence, components should be chosen based on commonality of
features and the components’ availability. Alternatively, a good supply of
spares can be purchased at the time of system installation, based on pre-
dicted failure rates of the various parts of the monitoring system.
Another alternative is to employ a monitoring system that is highly depen-
dent on software that is portable from computer to computer, so that it will
operate on nearly any hardware that is installed. In this way, the obsoles-
cence of the hardware becomes less important. An additional benefit of
having a high degree of software dependence is that the monitoring program
or its generator models can be updated as improvements are available.
Reloading software is a fairly easy task and is less disruptive to the operation
of the unit than replacing hardware.
68 NMAC Tech Note

6
REFERENCES
1. G. S. Klempner, “Expert System Techniques for Monitoring and Diagnos-

tics of Large Steam Turbine Driven Generators,” presented at the IEEE/
PES Winter Power Meeting, New York, NY (February 1995).
2. Generators for Utility and Industrial Applications, GE Industrial Power
Systems, General Electric Company, GET-8022 October, 1992 (4M).
3. IEEE Standards Collection—Electric Machinery, 1995 Edition.
4. Handbook to Assess Rotating Machine Insulation Condition. EPRI Power
Plant Series, Volume 16, November 1988.
5. Dr. H. Sedding, G. S. Klempner, S. R. Campbell, and Dr. G. C. Stone, “A
New Sensor for Detecting Partial Discharges in Operating Turbine
Generators,” presented at the IEEE Winter Power Meeting, New York,
NY (February 1991).
6. Dr. H. Sedding, G. S. Klempner, S. R. Campbell, Dr. G. C. Stone, W.
McDermid, and R. G. Bussey, “Practical On-Line Partial Discharge Tests
for Turbine Generators and Motors,” presented at the IEEE Summer
Power Meeting (1993).
7. C. C. Carson, S. C. Barton, and R. S. Gill, “The Occurrence and Control of
Interference from Oil Mist in the Detection of Overheating in a Genera-
tor,” presented at the IEEE/ASME/ASCE Joint Power Generation Con-
ference, Los Angeles, CA (September 1977).
NMAC Tech Note 69

8. S. C. Barton, C. C. Carson, W. V. Ligon Jr., and J. L. Webb, “Implementa-

tion of Pyrolysate Analysis of Materials Employing Tagging Compounds
to Locate an Overheated Area in a Generator,” Paper GER-3238, pre-
sented at the IEEE PES Summer Power Meeting, Portland, OR (July
1981).
9. Primer on Maintaining the Integrity of Water-Cooled Generator Stator Wind-
ings. Electric Power Research Institute, Palo Alto, CA: September 1995.
Report TR-105504.
70 NMAC Tech Note

SINGLE USER LICENSE AGREEMENT
WARNING: This Document contains
information classified under U.S. Export THIS IS A LEGALLY BINDING AGREEMENT BETWEEN YOU AND THE ELECTRIC POWER
Control regulations as restricted from RESEARCH INSTITUTE, INC. (EPRI). PLEASE READ IT CAREFULLY BEFORE REMOVING THE
WRAPPING MATERIAL.
export outside the United States. You
are under an obligation to ensure that you have a BY OPENING THIS SEALED PACKAGE YOU ARE AGREEING TO THE TERMS OF THIS AGREEMENT. IF YOU DO
legal right to obtain access to this information NOT AGREE TO THE TERMS OF THIS AGREEMENT, PROMPTLY RETURN THE UNOPENED PACKAGE TO EPRI
and to ensure that you obtain an export license AND THE PURCHASE PRICE WILL BE REFUNDED.
prior to any re-export of this information. Special 1. GRANT OF LICENSE
restrictions apply to access by anyone that is not EPRI grants you the nonexclusive and nontransferable right during the term of this agreement to use this package
a United States citizen or a Permanent United only for your own benefit and the benefit of your organization.This means that the following may use this package:
States resident. For further information regard- (I) your company (at any site owned or operated by your company); (II) its subsidiaries or other related entities; and
ing your obligations, please see the information (III) a consultant to your company or related entities, if the consultant has entered into a contract agreeing not to
contained below in the section titled “Export disclose the package outside of its organization or to use the package for its own benefit or the benefit of any party
other than your company.
Control Restrictions.”
This shrink-wrap license agreement is subordinate to the terms of the Master Utility License Agreement between
most U.S. EPRI member utilities and EPRI. Any EPRI member utility that does not have a Master Utility License
Agreement may get one on request.
Export Control Restrictions
2. COPYRIGHT
Access to and use of EPRI Intellectual Property is granted
This package, including the information contained in it, is either licensed to EPRI or owned by EPRI and is protected by
with the specific understanding and requirement that United States and international copyright laws.You may not, without the prior written permission of EPRI, reproduce,
responsibility for ensuring full compliance with all applicable translate or modify this package, in any form, in whole or in part, or prepare any derivative work based on this package.
U.S. and foreign export laws and regulations is being under-
taken by you and your company. This includes an obligation 3. RESTRICTIONS
You may not rent, lease, license, disclose or give this package to any person or organization, or use the information
to ensure that any individual receiving access hereunder who contained in this package, for the benefit of any third party or for any purpose other than as specified above unless
is not a U.S. citizen or permanent U.S. resident is permitted such use is with the prior written permission of EPRI.You agree to take all reasonable steps to prevent unauthorized
access under applicable U.S. and foreign export laws and disclosure or use of this package. Except as specified above, this agreement does not grant you any right to patents,
regulations. In the event you are uncertain whether you or copyrights, trade secrets, trade names, trademarks or any other intellectual property, rights or licenses in respect of
your company may lawfully obtain access to this EPRI this package.
Intellectual Property, you acknowledge that it is your 4.TERM AND TERMINATION
obligation to consult with your company’s legal counsel to This license and this agreement are effective until terminated.You may terminate them at any time by destroying this
determine whether this access is lawful. Although EPRI may package. EPRI has the right to terminate the license and this agreement immediately if you fail to comply with any
make available on a case by case basis an informal assessment term or condition of this agreement. Upon any termination you may destroy this package, but all obligations of
of the applicable U.S. export classification for specific EPRI nondisclosure will remain in effect.
Intellectual Property, you and your company acknowledge 5. DISCLAIMER OF WARRANTIES AND LIMITATION OF LIABILITIES
that this assessment is solely for informational purposes and NEITHER EPRI,ANY MEMBER OF EPRI,ANY COSPONSOR, NOR ANY PERSON OR ORGANIZATION ACTING
not for reliance purposes. You and your company ON BEHALF OF ANY OF THEM:
acknowledge that it is still the obligation of you and your (A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH
company to make your own assessment of the applicable RESPECT TO THE USE OF ANY INFORMATION, APPARATUS, METHOD, PROCESS OR SIMILAR ITEM
U.S. export classification and ensure compliance accordingly. DISCLOSED IN THIS PACKAGE, INCLUDING MERCHANTABILITY AND FITNESS FOR A PARTICULAR
PURPOSE, OR (II) THAT SUCH USE DOES NOT INFRINGE ON OR INTERFERE WITH PRIVATELY
You and your company understand and acknowledge your OWNED RIGHTS, INCLUDING ANY PARTY’S INTELLECTUAL PROPERTY, OR (III) THAT THIS PACKAGE
obligations to make a prompt report to EPRI and the IS SUITABLE TO ANY PARTICULAR USER’S CIRCUMSTANCE; OR
appropriate authorities regarding any access to or use of (B) ASSUMES RESPONSIBILITY FOR ANY DAMAGES OR OTHER LIABILITY WHATSOEVER (INCLUDING
EPRI Intellectual Property hereunder that may be in violation ANY CONSEQUENTIAL DAMAGES, EVEN IF EPRI OR ANY EPRI REPRESENTATIVE HAS BEEN ADVISED
of applicable U.S. or foreign export laws or regulations. OF THE POSSIBILITY OF SUCH DAMAGES) RESULTING FROM YOUR SELECTION OR USE OF THIS
PACKAGE OR ANY INFORMATION,APPARATUS, METHOD, PROCESS OR SIMILAR ITEM DISCLOSED IN
THIS PACKAGE.
6. EXPORT
The laws and regulations of the United States restrict the export and re-export of any portion of this package, and
you agree not to export or re-export this package or any related technical data in any form without the appropri-
ate United States and foreign government approvals.
7. CHOICE OF LAW
This agreement will be governed by the laws of the State of California as applied to transactions taking place entire-
ly in California between California residents.
About EPRI 8. INTEGRATION
You have read and understand this agreement, and acknowledge that it is the final, complete and exclusive agreement
EPRI creates science and technology solutions for between you and EPRI concerning its subject matter, superseding any prior related understanding or agreement. No
waiver, variation or different terms of this agreement will be enforceable against EPRI unless EPRI gives its prior writ-
the global energy and energy services industry. ten consent, signed by an officer of EPRI.
U.S. electric utilities established the Electric Power
Research Institute in 1973 as a nonprofit research
consortium for the benefit of utility members, their
customers, and society. Now known simply as EPRI,
the company provides a wide range of innovative Programs: TR-107137
products and services to more than 1000 energy- Nuclear Power

related organizations in 40 countries. EPRI’s
multidisciplinary team of scientists and engineers
draws on a worldwide network of technical and
© 1996 Electric Power Research Institute (EPRI), Inc. All rights reserved. Electric Power Research
business expertise to help solve today’s toughest Institute and EPRI are registered service marks of the Electric Power Research Institute, Inc.
energy and environmental problems. EPRI. ELECTRIFY THE WORLD is a service mark of the Electric Power Research Institute, Inc.
EPRI. Electrify the World Printed on recycled paper in the United States of America
EPRI • 3412 Hillview Avenue, Palo Alto, California 94304 • PO Box 10412, Palo Alto, California 94303 • USA
800.313.3774 • 650.855.2121 • askepri@epri.com • www.epri.com

Main Generator Online Monitoring and Diagnostics

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Main Generator Online Monitoring and Diagnostics

Uploaded by

Copyright:

Available Formats

Main Generator On-Line Monitoring

and License Agreement on the

back cover before removing the

Main Generator On-Line Monitoring

BACKGROUND Improving the amount or type of monitoring is one possible

EPRI TR-107137s Electric Power Research Institute December 1996

For further information on EPRI research programs, call

A) MAKES ANY WARRANTY OR REPRESENTATION WHATSOEVER, EXPRESS OR IMPLIED, (I) WITH

ORGANIZATION(S) THAT PREPARED THIS REPORT

1.0 INTRODUCTION ....................................................................................... 1

2.0 GENERATOR MONITORING PHILOSOPHIES ....................................... 3

2.1 Basic Monitoring with Static Alarm Limits ................................... 4

2.2 Dynamic Monitoring with Load Varying Alarm Limits ................. 5

2.3 Artificial Intelligence and Diagnostic Systems ............................ 8

3.0 SENSORS AND INSTRUMENTATION ................................................... 13

3.1 Generator Electrical Parameters ................................................. 13

3.2 Stator Core and Frame ................................................................. 18

NMAC Tech Note iii

3.2.3 Core End Flux Screen Temperature .................................. 19

3.3 Stator Winding .............................................................................. 21

3.4 Rotor .............................................................................................. 25

3.5 Excitation System ......................................................................... 31

3.6 Hydrogen Cooling System ........................................................... 33

iv NMAC Tech Note

3.6.9 Raw Service Water Inlet Temperature to Hydrogen

3.7 Lube Oil System............................................................................ 36

3.8 Seal Oil System ............................................................................. 38

3.9 Stator Cooling Water System ...................................................... 41

NMAC Tech Note v

3.10 Sensors and Instrumentation ...................................................... 46

4.0 SPECIALIZED MONITORING DEVICES ............................................... 51

4.1 Stator Winding Partial Discharge Monitoring ............................ 51

4.2 Rotor Winding Shorted Turns Detection .................................... 57

4.3 Stator Endwinding Vibration Monitoring .................................... 60

4.4 Shaft Voltage and Current Monitoring ........................................ 60

4.5 Generator Condition Monitor/Tagging Compounds .................. 61

4.6 Torsional Vibration Monitoring .................................................... 63

4.7 Stator Cooling Water Chemistry Monitoring .............................. 63

5.0 DECIDING ON A MONITORING SYSTEM ............................................. 65

5.1 Need ............................................................................................... 65

5.2 Economics ..................................................................................... 65

5.3 Risk ................................................................................................ 67

5.4 System Maintenance and Obsolescence ................................... 67

6.0 REFERENCES ........................................................................................ 69

vi NMAC Tech Note

Figure 1 Dynamic Temperature Monitoring .................................................... 7

Figure 2 Stator Winding Hierarchy ............................................................... 12

Figure 3 Large Generator Negative Sequence Capability ........................... 17

Figure 4 Determination of Rotor Winding Hot Spot Multiplier ...................... 26

Figure 5 RF Monitor ..................................................................................... 52

Figure 6 SSC Position in Slot, Bore View .................................................... 55

Figure 7 Cross-Section of Stator Slot Containing an SSC ........................... 56

Figure 8 Rotor Winding Shorted Turn .......................................................... 57

Figure 9 Rotor Winding Shorted Turns Detector, Typical Flux Probe

Figure 10 Rotor Winding Shorted Turns Detector, Typical Signal

NMAC Tech Note vii

The monitoring of large electrical generators has become a high priority in

NMAC Tech Note 1

Generator on-line monitoring and diagnostics cover a wide range of ap-

NMAC Tech Note 3

4 NMAC Tech Note

Whatever sensors are installed and connected to the monitoring system, it

NMAC Tech Note 5

6 NMAC Tech Note