From the SelectedWorks of Lin Gao

June 2010

Analysis of Power System Frequency Responses With Hydro Turbines Incorporating Load Shedding

Contact Author

Start Your Own SelectedWorks Available at: http://works.bepress.com/gao/4

Notify Me of New Work

Analysis of power system frequency responses with hydro turbines incorporating load shedding
Lin Gao Junrong Xia
Institute of Turbomachinery Xian Jiaotong University Xian, China
AbstractAs shown in many recent works, the loadfrequency behavior can be represented by system frequency response models especially for an islanding process. In this paper, a frequency response model incorporating an under-frequency load-shedding scheme is presented for islanding power system simulations. The model was applied for a practical power system with both hydro and fossil-fired power generators. The governing system models are involved in the model, which have great impact on the system dynamics. Parameter identification method is used to find and verify the model parameters. Similar responses were observed between the simulations and the accident records. The simulation results showed disadvantages of the primary frequency control systems of hydro power generators at the beginning of the load disturbances. An closed fossil fired power plant is proposed to serve against the load disturbances and improved stability was observed according to the system simulation results. Index Termsfrequency response model, hydro power sytem, load shedding.

Yiping Dai

II.

SYSTEM MODELLING

The dynamic modeling of a power system is often based on some crucial parameters that have an important effect on the frequency during isolated operation of the system after large disturbances. In most recent researches these parameters are the governor droop, , the reheater time constant, TR , the water starting time constant, TW , the inertia constant, M ,and the damp constant, D . The parameters of PID controller should also be considered. The popular DEH (digital electro-hydraulic) control system cause the system dynamics different with the original ones. A const governor droop is not accurate for the governing system model due to the uses of PID controllers. Besides, more parameters are needed for a system model incorporating with a load-shedding system because tiny frequency deviation may cause great load fluctuations according to the load-shedding scheme. A. Steam turbine unit and its governing system model

I.

INTRODUCTION

The power system may break into a number of islanded subsystems during a load disturbance or an accident, for instance, the over loading or the outage of the transmission line. This may cause a non-correctable emergency and the load shedding may be the only alternative left against the frequency decrease[1]. Significant amount of recent work has been carried out on load shedding based on under frequency considerations [2-10]. Most of them tried to improve the load shedding procedures and/or to reduce the amount of the shedding load belonging to the customers. The generation unit characteristics are of great importance for a load-shedding procedure design or optimizations. But simplistic models are adopted for most of their works. The frequency regulation systems are often simplified as a proportional coefficient, which have great impact on the system dynamics, especially for modern digital-electrical control systems. Dead-band, PID controller and other nonlinear components did dramatically change the original dynamics of the prime movers [11-13], which are ignored for most studies. A more precise model is needed for modern islanded system dynamic analysis with loadshedding. Hydro turbine units, steam turbine unit, and their governing systems are modeled in this paper for islanded system dynamic analysis in cooperating the under frequency load-shedding systems. Some essential nonlinear components are adopted and the models are used for a practical power system analysis.

Figure 1.

Sketch of reheat steam turbine unit

KP

Pr fr
f
1/

VMax
KI s

PMax

1 sTS + 1 1 sTL + 1
K LP

PMin

sK D

1 sTR + 1
K IP

1 sTH + 1
K HP

PG

Figure 2. Steam turbine unit and its governing system model

978-1-4244-5046-6/10/$26.00 2010 IEEE

ICIEA 2010

Reheated steam turbine units are used widely for modern power system. A classic reheat steam turbine consists of three volumes, which has great effect on the system dynamic responses, as shown in Fig. 1. Then, its primary frequency control system can be illustrated as Fig.2. with the consideration of PID controller and some necessary nonlinear elements which are necessary to the analysis on dynamic frequency responses to load disturbances. B. Hydro turbine and its governing system model Hydro turbines spread in some special areas where the local power systems may have different characteristics from an general subsystem, especially for a high proportion of hydro power subsystem. The turbine and its governing system model are available in reference [11] and [12]. A typical system model can be illustrated as Fig. 3.
PMax

simplified as a first order component with pure delay. If the frequency decreases below the given threshold, the corresponding timer will be active. If the frequency is higher than the threshold value, the corresponding timer will be reset. The load-shedding trigger time was set for each timer for prudent action.
f

f < f1

1 e1s sTs1 + 1

F1

f < f2

1 e 2 s sTs 2 + 1

F2

f < f3
1 sK ATw 1 + sK BTw

1 e 3s sTs 3 + 1

F3

fr
f

1 + sTn

1 s
s tTd 1 + sTd

1 sTS + 1
PMin

PG

f < f4
Pr

1 e 4 s sTs 4 + 1

F4

Figure 5. Diagram of a four-stage UFLS

Figure 3. Hydro turbine unit and its governing system model

C. System load model A single inertia-damping model is widely used for the power system load dynamics. Averaging behavior of the system frequency responses are assumed in this model as shown in references [13-16]. Inter-machine oscillations among different locations due to system disturbances are ignored according to such a model. The assumptions are accurate enough for small isolated system with strong connections. The shedding load is assumed to be widely distributed across the network as is typically case. Then it is suitable to submit the shedding loads directly from the system loads in the frequency response model as shown in Fig. 4
PL

D. Practical system simulation model The power system of Hetian area in West China is modeled in this paper. The system was islanded off the main network due to an accident on October 14, 2008. There were three main hydro power plants (Wuluwati, Paiziwati and Kalager), one fossil fired power plant (Lasikui, which was shut down before the accident) and 23 small hydro power plants in the islanded system as shown in Fig. 6. Before the accident, the system status is shown as Table I, where the downloaded power through the 220 kV transmission line had reached almost 41.3 % of the total loads. The accident caused the failure of the 220 kV transmission line and resulted in a great load disturbance to the islanded system. The huge unbalanced power made the system frequency decreased rapidly to 46.56 Hz and triggered all the designed load shedding procedures (four levels) in a short time.

1 sM + D

Figure 4. Load dynamic response model

where the ULFS is the under freqency load shedding system. In general, the load shedding procedure is active when the frequency f decrease below a given value f i and lasts for a short time si . A typical four-stage ULFS are shown as Fig. 5. where timers are used to generate load shedding signals. The servo systems and the delay time of other devices are

Figure 6. Power system structure of Hetian area

Item Wuluwati Kalager Paiziwati Lasikui 23 small hydro power plants Download From the main network Total Load

Output power variation (p.u.)

Power MW 12.0 7.7 4.7 0.0 8.7 23.3 56.4

Rated Power MW 60.0 11.4 10.8 36.0 / / /

Gate Position variation (p.u.)

TABLE I.

SYSTEM STATUS BEFORE THE ACCIDENT

1.0 0.5 0.0 0 1.0 10 20 Time (s) 30 40

Reverse action
0.5 0.0 0 10

Output power generation Simulation with identified parameters

20 Time (s) 30 40

The islanded system is organized by the models referred before and shown as Fig. 7. All the generation power and load demands are summed up together in the view of average parameters. The load model parameters and the load shedding procedure parameters are collected from the electric power co. as shown in Table II.

Figure 8. Hydro-turbine frequency response

0.0 -0.2 -0.4 -0.6 -0.8 -1.0 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 0.0 -0.2 -0.4 -0.6 -0.8 -1.0 0.0 -0.2 -0.4 -0.6 0 2 4 6 8 10 12 14 Time (s) 16 18 20 22 24

1 sM + D

Figure 7. Power system structure of Hetian area

TABLE II. Load Shedding Procedure I II III IV

SYSTEM STATUS BEFORE THE ACCIDENT Triger Frequency Hz 49.0 48.5 48.0 47.5 Triger Time

si
s 0.5 0.5 0.5 0.5

Shedding Power MW 18.75 3.897 4.37 2.56

crossover steam pressure p.u.

Reheater pressure p.u.

Governing stage pressure p.u.

PL

Control valve position p.u.

Test data simulated response with identified parameters

Figure 9. Steam turbine identification results

TABLE III.
Parameter Range

GOVERNING SYSTEM PARAMETER RANGES Steam turbine units

Range Parameter Range Parameter

Hydro-turbine units

E. Model parameters Model parameters are of the same importance as the model structures. System response characteristics vary significant with some different model parameters. Most turbine unit parameters are identified from effective test data. The identification method used in this paper is a GA (genetic algrithm) based method[17][18]. The simulation responses of the system are compared with another serious of test data for validation. Fig. 8 shows a hydro-turbine simulation results with the identified parameters while Fig.9 shows that of a steam turbine unit. Governing systems vary with different manufacturers but fortunately only two model structures for hydro governing systems and one for steam governing systems are enough according to the main units in the object subsystem. The parameter ranges of different governing systems in Hetian area are listed in Table III.

Tn Td

0.2-0.3 8-14 0.3-0.8 0.04-0.06 0.25-0.4

KP KI KD

300-450 10-50 60-200 0.03-0.04 0.03-0.3

KP KI KD

0.2-1.5 0.02-0.05 0.0-0.3 0.05 0.06-0.07

t p
Ef

p
Ef

Ef

III.

RESULTS AND DISCUSSION

It can be seen from Fig. 6 that all the power needed are generated by hydro power plants after the Lasikui power plant is shut down. Hydro turbine unit has its specialty to participate in primary frequency regulation (PFR). When the system frequency drops beyond the dead-band, the governing system needs to open the gate (guide vane) to increase the output power temporarily. But the hydro turbine power will not increased immediately but decrease for a certain degree then start to increase. Fig. 8 shows the reverse effect of a hydro turbine unit through both the testing and the simulation methods. When the accident occurred, dramatic imbalanced

power due to circuit break resulted in the fast dropping of the islanded system frequency. Soon the frequency drop beyond the dead-bands of the governing systems, then the system open the gate to try to increase the output power of the hydro turbines. But the hydro power decreased at the first time and exacerbated the already existing imbalanced power which causes a faster dropping of the frequency. Moreover, the hydro power system always has a large dead-band setting, which will further delay the action of the PFR. Both of the disadvantages compared with a fossil fired turbine unit make the system very weak against the accident. In the accident, all the load shedding procedures were triggered in a very short time and made the shedding load much larger than what it needs in theory. The emergency control strategy still could not stop the fast dropping frequency and a terrible frequency value 46.5 Hz is recorded for the accident. The whole system is simulated using the proposed mathematic models with the help of a general purpose Matlab program. The frequency of the whole procedure is shown in Fig. 9. It also shows that increased generations and dramatically decreased loads formed a reverse imbalanced power which caused the frequency increased rapidly over the nominal value in about 10 seconds.
54 52 50 48 46 0 Simulation of the practical situation 50% power generated by Fossil fired power plant 20

system. The accident triggered all four-level load-shedding procedures and the frequency got its minimum value of 46.5 Hz. Similar results are found according to the simulation results. Hydro-turbine PFR system often respond slowly relative to the steam turbine system due to their large dead-band and the reverse action. The simulation results show an advantage frequency response of the system against the accident if the 50% electric power is generated by fossil fired turbine units. Therefore, a closed fossil fired power plant is proposed to improve the system stability especially for the dry seasons.

Notation
D Fi f fi fr KA KB KD K HP KI K IP K LP KP M PG PL Pr Td THP TIP TLP Tn TR TS Tsi TW VMax
Damping coefficient of the connected loads The ith shedding load proportion Power system frquency The ith load-shedding frequency Reference frequency Hydro turbine dynamic coefficient A Hydro turbine dynamic coefficient B Differential controller coefficient Fraction of total power developed by the HP turbine section Integral controller coefficient Fraction of total power developed by the IP turbine section Fraction of total power developed by the LP turbine section Proportional controller coefficient Lumped rotation inertia of the interconnected generating units Generation power of the unit System load power Reference power Buffer time constant HP inlet piping lag time constant Reheater and IP inlet piping lag time constant Crossover and LP inlet piping lag time constant Acceleration time constant Reheater time constant Servomotor inertia time constant The ith servo system inertia time constant water starting time constant The opening speed limit speed droop of online units Permanent droop of the hydro turbine Transient droop of the hydro turbine The ith load-shedding system lag time constant

Frequency (Hz)

Time (s)

40

60

80

Figure 10. Accident simulations of Hetian area

According to prior discussions, a fossil fired power generation unit may be faster for system frequency regulation. Then, we assume that 50% electricity power is generated by fossil fired power plant (Lasikui). Same accident was simulated for the new system and the frequency responses are shown as the dotted line in Figure 9. The frequency has its lowest value as 47.55 Hz and only three UFLS procedures are triggered. The overshot is also much better than that of the practical situations. The Lasikui power plant may be better to be reopened or to be rebuilt. Such a fossil fired power plant is helpful for the islanded system especially for the dry seasons when a large download electricity power is necessary. IV. CONCLUSIONS

Both hydro and steam turbine units and their governing system are modeled in this paper for the dynamic analysis on an islanded system. A serious accident is simulated due to outage of the transmission line in cooperating a load-shedding

p t i

REFERENCES
[1] L. D. Arya, V. S. Pande, D. P. Kothari, A technique for load-shedding based on voltage stability consideration, Electrical power & Energy systems, vol. 27, pp. 506517, 2005. V. N. Chuvychin, N. S. Guroy, S. S. Venkata, R. E. Brown, An adaptive approach to load shedding and spinning reserve control during under frequency conditions. IEEE Transactions on Power Systems, Vol. 11, No. 10, 1996, pp. 18051810. S. M. Chandekar, S. G. Tarnekar, Revised load sheding schedule for power system incorporating the effect of transmission line performance Electric Power and Energy Systems, Vol. 24, 2002, pp. 379-386. W. P. Luan, M. R. Irving, J. S. Daniel, Genetic algorithm for supply restoration and optimal load shedding in power system distribution networks, IEE ProceedingsGeneration, Transmission and Distribution, vol. 149, No. 2, 2002, pp. 145-151. J. Jung, C. C. Liu, S. L. Tanimoto, V. Vittal, Adoption in load shedding under vulnerable operating condition. IEEE Transactions on Power systems, Vol. 17, No. 4, 2002. S. J. Huang, C. C. Huang, An automatic load shedding scheme including pumped-storage units, IEEE Transactions on Energy conversion, Vol. 15, No. 4, 2000, pp. 427-432. P. Paolo, Emergency load-shedding algorithm for large industrial plants, Control Engineering Practie, Vol. 10, 2002, pp. 175-181. T. Tomaz, V. Gregor , G Ferdinand, Revision of the underfrequency load-shedding scheme of the slovenian power system, Electric Power Systems Research, Vol. 77, 2007, pp. 494-500 L. Y, W. S. Kao, Study of Applying load shedding scheme with dynamic D-factor values of various dynamic load models to Taiwan power system, IEEE Transactions on power systems, Vol. 20, No. 4, 2005, pp. 1976-1984.

[2]

[3]

[4]

[5]

[6]

[7] [8]

[9]

[10] Q. Zhao, C. Chen, Study on a system frequency response model for a large industrial area load shedding, Electrical Power and Energy Systems, Vol. 27, 2005, pp. 233-237. [11] Working Group on Prime Mover and Energy Supply Models for System Dynamic Performance Studies, Hydraulic turbine and turbine control models for system dynamic studies, Transactions on Power Systems, Vol. 7, No. 1, 1992, pp.167-178. [12] A. I, K. Hirayama, T. Kageyanma, Practical hydraulic turbine model, IEEE Power Engineering Society General Meeting, Montreal, Canada, 2006, pp. 1-7. [13] Y. Dai T. Zhao Y. Tian L. Gao, Research on the primary frequency control characteristics of generators in power system, 2nd IEEE Conference on Industrial Electronics and Applications HaerbinChina2007 [14] Y. DaiT. ZhaoY. TianL. Gao, Research on the influence of primary frequency control distribution on power system security and stability, 2nd IEEE Conference on Industrial Electronics and ApplicationsHaerbinChina2007 [15] Y. DaiP. ZhaoS. Chang, Primary frequency control characteristic of a grid, 3rd IEEE Conference on Industrial Electronics and ApplicationsSingapore2008 [16] L. Gao, Y. Dai, "Impact of Overspeed Protection Control on Stability for Islanded Power System," International Journal of Emerging Electric Power Systems: Vol. 10, 2009. [17] L. Gao, Y. Dai, J. Xia, A new framework for power system identification based on an improved genetiv algorithm, 4rd IEEE Conference on Industrial Electronics and ApplicationsXian2009, pp. 1946-1951. [18] L. Gao, Y. Dai, J. Xia, Parameter Identification of hydro generation system with fluid transients based on improved genetic algorithm, Fifth international conference on natural computation, tianjin, China, pp. 398402.