Proceedings of the 37th Hawaii International Conference on System Sciences - 2004

Component Monitoring and Dynamic Loading Visualization from Real Time Power Flow Model Data
A. P. Sakis Meliopoulos School of Electrical and Computer Engineering Georgia Institute of Technology Atlanta, GA 30332-0250 George J. Cokkinides Department of Electrical Engineering University of South Carolina Columbia, SC 29803 Thomas J. Overbye Department of Electrical and Computer Engineering University of Illinois at UrbanaChampaign Urbana, IL 61801

Abstract: The technology of intelligent electronic devices in power systems has exploded and with it the available real time data. The data are typically used to extract a real time model of the system via traditional state estimation methods. The traditional approach estimates only the system voltages and uses a small part of the available information. This paper presents a new approach for better utilization of the available information. Specifically, we propose the use of existing data for estimating detailed operational models of major power equipment. For example, the detailed models can be in the form of electro-thermal models which then will allow the monitoring of device temperatures, dynamic loading, etc. In general, the real time model can be as simple or as complex depending on the type and quality of available data. For example, from typical SCADA data, the real time operational model of a generator can be extracted. This model can provide the operating margins, etc. Addition of other data, such as ambient temperature, etc. can also provide an electro-thermal model in real time. The paper demonstrates the approach with the example of a power transformer. The methodology is applied to extract an electro-thermal model of the transformer. The real time electro-thermal model provides: (a) transformer temperatures including hot spot, (b) transformer loss of life and (c) transformer dynamic loading. The approach has been simulated in a multitasking environment using an electric power system model with a time-function electric load. The operation of the system is simulated by solving the power flow at user selected time intervals. As the time progresses the electric load changes and a power flow solution determines by computation the operation of the system. At each time step, the real time model of the selected power devices are extracted via statistical estimation methods. The real time model provides vital information for these devices. Their operation conditions can be visualized and animated as desired. The paper presents these applications.

Keywords: power flow, real-time model, state estimation, visualization, animation.

1. Introduction
The technology of intelligent electronic equipment (IED) has advanced to the point that is changing the legacy SCADA systems of electric power systems. The trend has resulted in an explosion of information that is available in real time. This information is available locally (at the substation) or most times it is communicated to the control center. While the amount of available information has exploded, the efficient utilization of this information by operators has been lagging. It is clear that new methods of manipulating the available information and presenting this information to the operator in a meaningful way are needed. Extensive efforts are under way towards this goal. One particular approach is to present this information to the operator with animated visualization techniques. In this approach, the visualized information can be raw information collected by legacy SCADA systems and/or supplemented by IEDs or it can be filtered data via the traditional state estimator. This paper is an attempt to push this approach a little bit further. Specifically, it is possible to utilize available real time data for extracting more detailed operational models of power system components. For example, this data can be used to extract a real time electro-thermal model of a transformer, of a generator, etc. The advantage of this approach is that the real time electro-thermal model can be used to determine a multiplicity of useful information, for example, transformer hot spot temperature, loss of life of the transformer, dynamic loading limits, etc. This information can be presented to operators in an animated visualization. Since todays electric power systems are operated ever closer to limits, this approach can be quite useful to operators.

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

Proceedings of the 37th Hawaii International Conference on System Sciences - 2004

The paper presents the basic approach with an example of a power transformer. Specifically, we describe the extraction of the electro-thermal model of a power transformer from real time data and its utilization to determine (a) hot spot temperature, (b) loss of life and (c) transformer dynamic loading. The exact same approach can be used to extract an elector-thermal model of a synchronous generator. Similar operating information can be then defined for the generator. The electro-thermal model can be visualized and animated so that the operator can have a revealing picture of the operating condition and limitations in one simple glance. The approach is demonstrated with the aid of a power system simulator. The power system simulator uses a time function electric load model. As time progresses, the electric load varies and drives the operation of the system. At each instant of time the overall system operating conditions are determined via a power flow solution for the specific load level at that instant. The SCADA system is simulated by extracting data from the power flow solution and providing inputs to the estimators that extract the real time transformer and generator models. The estimated electrothermal models are presented via animation and visualization techniques. The overall approach is demonstrated on an example electric power system, the IEEE RTS system. The system has been modified to include a time function electric load model. Snapshot displays of the animations are given in the paper. Live demonstration of the animations will be given during the presentation of the paper. The paper presents the transformer electro-thermal model, the estimation of the real time electro-thermal model from SCADA data and additional real time data and specific results from the real time electro-thermal model, such as hot spot temperature, loss of life and transformer dynamic loading. All of this information is visualized and animated. The paper describes the simulation environment for the demonstration of these ideas and provides snapshots of the visualizations. During the presentation of the paper, a live demonstration of the visualization and animation techniques will be provided.

CT

Monitored Transformer
PT

CT

CT

PT PT

CT

CT

PT PT PT Temp Tap

CT

Data Collection

To Control Center

Figure 1: Transformer Monitoring System Configuration This data is utilized to determine a real time electrothermal model of the transformer. Subsequently the real time model is utilized to compute the loss of life of the transformer and the available dynamic loading margin. The three computational procedures will be described next.

3. Real-Time Electro-Thermal Model

The electro-thermal model estimates the internal transformer temperature distribution from the measured quantities, i.e. the transformer terminal voltages and currents, top oil temperature, tank temperature, load tap changer position, and ambient temperature. Figure 2 illustrates the equivalent circuit of the transformer electro-thermal model. It consists of 12 temperature nodes corresponding to six coil temperatures, three magnetic core temperatures, oil, tank, and ambient temperature. A capacitor is connected at the coil, core, oil, and tank nodes, representing the thermal capacity of the corresponding components. A resistive network links the 12 nodes. This network represents the thermal conductances among the various components. A higher order model can be used for increased precision.
Thermal Conductance Matrix

2. Estimation of Transformer Electrothermal Model

Transformer terminal voltages and currents, tap position and possibly tank temperature, ambient temperature, top of oil temperature may be monitored via a variety of devices, for example, SCADA, relays, metering devices, etc. Assume that this data is available at the control center via the available communication channels. The collected data is illustrated in Figure 1. This data is streaming into the control center at regular time intervals determined by the SCADA scanning rate.

T1a Q1a Q1b C1

T1b Qmb C2

Tmb Qmc C8

Tmc

Toil

Ttank Tamb

C9

C10

C11

Figure 2. Thermal Model Equivalent Circuit

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

Proceedings of the 37th Hawaii International Conference on System Sciences - 2004

The dynamics of the above network are described by the following differential equation:

B LOL = exp( ( A + ~ ))h T

where:

(2)

d C T = GT + Q dt
where: C is the thermal capacitance matrix, G is the thermal conductance matrix, T is the temperature vector, and Q is the heat source vector.

(1)

~ T is the average hot spot temperature over the

time interval (t-h,t) in degrees Kelvin A, B are constants dependent on transformer

design Above equation is integrated over time to provide the loss of life over this time. The loss of life result can be better appreciated when normalized. The normalization is provided with the following equation.
T B A+ ~ T

The thermal capacitances are computed from the volume and the specific heat constant of the corresponding components. The thermal conductances are estimated from known temperature gradients between coils, oil, tank, and the environment. The transformer losses are computed from the measured voltage and current waveforms at the transformer terminals. Computed losses include winding ohmic losses as well as magnetic core losses. These losses are treated as heat sources in the electro-thermal model (Q1a, Q1b,,Qmc). For a three-phase transformer (delta-wye or autotransformer) a total of nine heat sources are computed: six winding ohmic losses and three magnetic core losses (one for each magnetic core leg). These heat sources are shown in Figure 2. Coil losses are computed by multiplying the coil resistance by the squared value of the RMS winding currents. For a three-phase transformer (delta-wye or autotransformer) a total of six winding heat sources are computed. Core loss is a function of the core magnetic flux. The model estimates the flux from measurements and then the core loss is computed. The end result of the estimation module is the state of the electro-thermal model in real time. This model is used to compute the loss of life and the dynamic loading of the transformer.

e LOL pu =
0

dt
(3)

where T is the elapsed time. It is known from the standard that if the transformer operates under constant hot spot temperature equal to the permissible value, the normalized LOL will be exactly 1.0.

5. Transformer Dynamic Loading

The objective of the transformer dynamic loading capability is to answer the following question: if the transformer loading is suddenly increased to P, how long it will take until the hot spot temperature of the transformer reaches the maximum permissible value? This information is useful in an emergency since it will let us know for how long we can overload the transformer without exceeding manufacturers specifications. The dynamic loading pairs (P,t) are computed via time domain simulation of the transformer operation. The initial conditions are determined by the present operating conditions. The voltages applied to the transformer are assumed constant. The load of the transformer is assumed to be P. Then the differential equations describing the electro-thermal model (1) of the transformer are numerically integrated. Specifically, using trapezoidal integration over the time interval (t-h,t) on equation (1) yields:

4. Transformer Loss of Life

The real-time model of the transformer provides an estimate of the hot spot of the transformer in real time. Specifically, the hot spot temperature is defined as the highest temperature anywhere along the coils of the transformer. Then, the loss of life of the transformer is computed per IEEE-ANSI standard 57 using the hot spot temperature. This calculation is based on the ANSI/IEEE C57.911981. Specifically, the per unit loss of life, (LOL), of the transformer in the time interval (t-h,t) is computed with the formula:

T (t ) =

2C +G h

2C ~ G T (t h ) + 2Q h

(4)

where: C is the thermal capacitance matrix, G is the thermal conductance matrix, and

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

Proceedings of the 37th Hawaii International Conference on System Sciences - 2004

~ Q is the average value of the heat sources over the time

interval h By repeatedly evaluating equation (4), the transformer temperatures are computed as a function of time. From the temperatures T(t), the hot spot of the transformer is extracted. When the permissible hot spot temperature is reached, the time t is recorded, yielding the pair (P,t). This procedure is repeated for different values of transformer loading P, yielding a number of pairs (P,t). This data is plotted providing the dynamic loading curve of the transformer. An example of the transformer dynamic loading curve is illustrated in Figure 3.

the simulation is repeated until a power level is reached where the hot spot temperature is not reached for a certain simulation time limit tMAX (typically 90 minutes). The results of the simulations are presented as a plot of transformer loading versus time to hot-spot temperature limit (as illustrated in Figure 4).

Start

t=0 T=T0 P = PMAX

Store (t,P) t=0 T= T0 P=P-P YES

Compute T(t) t=t+h Max(T(t)) >TPERM

NO

Figure 4: Transformer Dynamic Loading Display

6. Visualization and Animation

Stop YES t > tmax NO
The computed real time electro-thermal model, temperatures and dynamic loading can be visualized. The visualization may be updated as time progresses providing an animated impression of the operation of the system. The visualization of the transformer can be implemented in a variety of ways. We have elected to build a 3-D model of the transformer and color code the temperatures of the various parts of the transformer. The 3-D model can be rotated, paned, zoomed, etc. so that the user can take as close a view of the operating conditions of the transformer as he/she desires. Figures 5 and 6 illustrate two stills of the visualization representing different operating conditions of the same transformer. Note that in Figure 5 the temperatures of the various transformer coils are quite different indicating an unbalanced operation. This type of application obviously requires three phase measurements and three phase network solutions. Figure 6 illustrates near balanced operating conditions. Figure 4 provides the transformer dynamic loading for the condition of the transformer of Figure 6. It is important to note that in real time applications (or simulated demonstrations) all visualizations can be displayed simultaneously. It is also important to note that in a real time application, the transformer temperatures do not change very fast. This means that the animation of the transformer conditions will be very slow.

Figure 3. Dynamic Loading Algorithm Flow Chart The overall algorithm for computing the pairs (t,P) is illustrated in Figure 3. A series of simulations is performed all starting with initial temperature values (T0) taken from the present operating conditions of the transformer. Each simulation assumes a different transformer total power, P, within a user selectable range. The transformer terminal voltages and power factor are assumed constant and equal to those voltages and power factor of the present operating conditions. The terminal currents are computed from the known voltages, power factor and the assumed power P. Then, the heat source vector Q used for the simulations is computed as described in the section Electro-thermal model from the transformer terminal voltages and currents. Each simulation is carried out until the transformer hot-spot temperature reaches the maximum allowable value (TPERM). The maximum allowable hot-spot temperature is a user-specified quantity. The first simulation starts at the maximum power loading PMAX (typically about 2.5 pu). Then the power is reduced and

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

Proceedings of the 37th Hawaii International Conference on System Sciences - 2004

the transformer. As time passes, the operation of the transformer is viewed in an animated way. Figures 5 and 6 present stills of this animation. During the presentation of the paper, this simulated operation of the proposed system will be demonstrated.

8. Summary
We propose a non-intrusive approach to transformer monitoring, evaluation and dynamic loading using the existing instrumentation infrastructure. The proposed approach is supplemented with visualization and animation software that provide a self-explanatory picture of the operating conditions of the transformer. This approach is very effective and very useful for power system dispatcher. It is believed that it will be a valuable tool for the control center. The proposed approach and methodology is applicable to any major power system equipment. The real time model can be as detailed as desired. For example, the proposed methodology can be applied to estimate a detailed electro-thermal model of a generating unit. This real time model, providing information about the loadability of the generator can be very useful to system dispatchers.

Figure 5. Transformer 3-D Model View with ColorCoded Temperatures View 1

9. References
1. A. P. Sakis Meliopoulos, G. J. Cokkinides and R. I. James, Transformer Diagnostic System for Loss of Life and Coil Integrity, Proceedings of the 1998 Georgia Tech Fault and Disturbance Analysis Conference, Atlanta, Georgia, May 4-5, 1998. A. P. Sakis Meliopoulos and G. J. Cokkinides, "Effects of Harmonics on Power Transformers", Proceedings of the 2nd International Conference on Harmonics in Power Systems, Winnipeg, Manitoba, Canada, pp. 201-213, October 1986. B. C. Lesieutre and W. H. Hagman and J. L. Kirtley, Jr., An Improved Transformer Top-Oil Temperature Model for Use in an Online Monitoring and Diagnostic System, IEEE Transactions on Power Delivery, Vol. 1, pp. xx-xx, January 1997. ANSI/IEEE C57.91-1981, Guide for Loading Mineral OilImmersed Overhead and Pad-Mounted Distribution Transformers. ANSI/IEEE C57.92-1981, Guide for Loading Mineral OilImmersed Power Transformers. ANSI C57.92-1962, Guide for Loading Oil-Immersed Distribution and Power Transformers. IEC Publication 216: Guide for the Determination of Thermal Endurance Properties of Electrical Insulating Materials, IEC 1993. J. F. Lindsay, Temperature Rise of an Oil Filled Transformer with Varying Load, IEEE Transactions on

2.

Figure 6. Transformer 3-D Model View with ColorCoded Temperatures View 2

3.

7. Simulated Demonstration
In order to demonstrate the proposed approach for transformer monitoring and evaluation, we use the 24 bus IEEE RTS. This system is loaded with a time varying electric load. At pre-specified interval, the power flow problem is solved and the results are passed to the transformer monitoring and evaluation system. This system estimates the electro-thermal model of the transformer, computes the transformer temperatures, loss of life, and the dynamic loading of the transformer. The results of these analyses are fed into the visualization of

4.

5. 6. 7.

8.

0-7695-2056-1/04 $17.00 (C) 2004 IEEE

Proceedings of the 37th Hawaii International Conference on System Sciences - 2004

9.

10. 11.

12.

13.

14.

15. 16.

17.

18.

Power Apparatus and Systems, Vol. Pas-103, pp. 25302536, September 1984. T. W. Dakin, Electrical Insulation Deterioration Treated as a Chemical Rate Phenomenon, AIEE Transactions, Vol. 67, pp. 113-122, 1948. EPRI EL-2443, Basic Transformer Life Characteristics, Vol. 1, June 1982. E. J. Murphy, Thermal Decomposition of Natural Cellulose in Vacuo, Journal of Polymer Science, Vol. 58, pp. 649-655, 1962. L. C. Flowers, Moisture Evolution Rates Calculated for Cellulose Paper Undergoing Thermal Decomposition, Insulation, Vol. 10, pp. 23-26, September 1964. W. G. Lawson, M. A. Simmons and P. S. Gale, Thermal Aging of Cellulose Paper Insulation, IEEE Transactions on Electrical Insulation, Vol. EI-12, No. 1, February 1977. F. W. Heinrichs, Bubble Formation in Power Transformer Windings at Overload Temperatures, IEEE Transactions on Power Apparatus and Systems, Vol. PAS-98, pp. 15761582, Sept/Oct 1979. A. P. Sakis Meliopoulos, Power System Grounding and Transients, Marcel Dekker Inc., 1988. A. P. Meliopoulos, F. Zhang, S. Zelingher, G. Stillmam, G. J. Cokkinides, L. Coffeen, R. Burnett, J. McBride, Transmission Level Instrument Transformers and Transient Event Recorders Characterization for Harmonic Measurements, IEEE Transactions on Power Delivery, Vol. 8, No. 3, pp. 1507-1517, July 1993. R. P. Klump, D. Schooley, T. J. Overbye, "An Advanced Visualization Platform for Real-Time Power System Operations," Proc. 14th Power Systems Computational Conference (PSCC), Sevilla, Spain, June 2002. R. P. Klump, W. Wu, G. Dooley, Displaying Aggregate Data, Interrelated Quantities, and Data Trends in Electric Power Systems, Proc. 36th Hawaii International Conference on System Sciences, Kona, HI, January 2003.

degrees from the Georgia Institute of Technology in 1974 and 1976, respectively. In 1971, he worked for Western Electric in Atlanta, Georgia. In 1976, he joined the Faculty of Electrical Engineering, Georgia Institute of Technology, where he is presently a professor. He is active in teaching and research in the general areas of modeling, analysis, and control of power systems. He has made significant contributions to power system grounding, harmonics, and reliability assessment of power systems. He is the author of the books, Power Systems Grounding and Transients, Marcel Dekker, June 1988, Ligthning and Overvoltage Protection, Section 27, Standard Handbook for Electrical Engineers, McGraw Hill, 1993. He holds three patents and he has published over 180 technical papers. Dr. Meliopoulos is the Chairman of the Georgia Tech Protective Relaying Conference, a Fellow of the IEEE and a member of Sigma Xi. George Cokkinides (M '85) was born in Athens, Greece, in 1955. He obtained the B.S., M.S., and Ph.D. degrees at the Georgia Institute of Technology in 1978, 1980, and 1985, respectively. From 1983 to 1985, he was a research engineer at the Georgia Tech Research Institute. Since 1985, he has been with the University of South Carolina where he is presently an Associate Professor of Electrical Engineering. His research interests include power system modeling and simulation, power electronics applications, power system harmonics, and measurement instrumentation. Dr. Cokkinides is a member of the IEEE/PES. Thomas J. Overbye (M '83, SM '95) received his B.S., M.S., and Ph.D. degrees in Electrical Engineering from the University of Wisconsin-Madison in 1983, 1988 and 1991 respectively. He was employed with Madison Gas and Electric Company from 1983 to 1991. Currently he is an Professor of Electrical and Computer Engineering at the University of Illinois at Urbana-Champaign. His research interests include power system modeling, restructuring and visualization.

10. Biographies
A. P. Sakis Meliopoulos (M '76, SM '83, F '93) was born in Katerini, Greece, in 1949. He received the M.E. and E.E. diploma from the National Technical University of Athens, Greece, in 1972; the M.S.E.E. and Ph.D.

0-7695-2056-1/04 $17.00 (C) 2004 IEEE