European Association for the Development of Renewable Energies, Environment and Power Quality (EA4EPQ)

International Conference on Renewable Energies and Power Quality (ICREPQ10) Granada (Spain), 23rd to 25th March, 2010

Creation of Stability Index for Micro Grids

I. Vokony, A. Dn Dr.

Department of Electric Power Engineering, Group of Power Systems and Environment Faculty of Electrical Engineering, Budapest University of Technology and Economics 18 Egry Jzsef Street, Budapest, 1111, Hungary Phone: +36 30 640 9013 e-mail: vokony.istvan@vet.bme.hu

Abstract This paper is qualified for creation of a new stability index for micro grids (MG) in cooperation with the power system (PS), which characterizes the dynamic security of MG well. Primary objective of this paper is to examine the effect of continuously increasing integration of intelligent energy distribution networks, concerning the stability of the PS. The model used for simulating the MG and the calculation method of the new stability index are published in this paper. Simulation results of model network are described, conclusions are drawn and evaluated. Keywords
Stability index, micro grid (MG), power system (PS), power balance, time-domain simulations.

micro networks it is essential to create a suitable MG model.

2. Calculation method
Two different methods were compared. The first one is the so called TSI method. The transient stability index is used to get information about the system stability. The basic idea is that calculation of angular accelerations of the machine units at the instant fault can be used to get useful information about the system. An index can be formulated from these accelerations that will give information about the disturbance on the whole system. The calculation method is described in the following points: Assuming that a 3-fault occurs at the node of ith machine unit at t=0. Calculate the angular accelerations of the examined machine units (relative to a reference machine noted by index n), and find the maximum of these values:

1. Introduction
The development history of PS proceeded through similar phases everywhere: the initial local supply was followed by transfers of shorter distance and finally national interconnected networks came into operation. These networks were connected for the sake of mutual technical help and nowadays the networks are also used for great commercial transports, covering whole Europe. This global PS definitely has many advantages, but nowadays its disadvantages come increasingly into foreground. As a consequence of the high loading of system elements, the reliability of supply is decreasing, provided by some recently recorded cases. Network structures are needed to be established, the direction of which are more favorable and designable. They simplify the system regarding system regulation are able for island operation and in some cases they can join to the PS. The MG can be a solution for the above questions. On the other hand for the examination of these kinds of

= max k n ir
= 1..n 1 k

Calculate the average angular acceleration of the whole system:

= * rj
j =1

1 n

 Calculate the difference of each acceleration from this average:

i=
r i

(V I ji = j Vi ) y ji + V j

y 'ji 2

After calculating i for the case of the fault occurring at all examined machines, TSI can be defined as:

Following these calculations, the branch currents and the bus voltages during the fault are known. Only one more step is necessary to get the stability index, using the equal area method.

= TSI

max
i =1..n

In order to become familiar with the characteristic values, behavior and meaning of the obtained TSI values, the results with time-domain simulation of the same cases need to be compared. One of the most characteristic results of these simulations is the critical clearing time, i.e. the maximal duration of the fault before clearing, after which all machine units can return to synchronous operation in a new equilibrium. It is possible to get reliable information based on the value of Transient Stability Index. The correspondence between TSI and CCT values is strictly monotonous, so critical clearing times can be unambiguously estimated. Although this method is useful a new one has been created, which is sharper and the calculation requirements are less.

Figure 2.1.: the equal area method

Pm = * max

(P
le

max

*sin Pm )d

Pm * max Pmax *(cos le cos i ) Pm *( i le ) =

This new method is the t method. The aim of calculation method is to get the short circuit le 0 currents in the system at optional location of the short circuit. The system parameters are known. In a short circuit case on a bus there is a short circuit = Pmax (cos( 0 + ) cos i ) Pm ( i 0 ) Pm current. This current comes from the generator connected to the failure bus, and from the whole remaining system. The current deriving from the = Pmax (cos( 0 + ) cos i ) Pm ( i 0 ) 0 generator is easy to calculate. If the short circuit current can be obtained on the bus, we can calculate the remaining systems current. From these P cos( 0 + = cos i + m ( i 0 ) ) parameters it is easy to obtain the so called Pmax remaining systems impedance and complex voltage level. The first step is the short circuit current. The calculated 3-phase sc.-s are supposed to be impedance-less. The voltage levels of the buses are m 0 0 i i in the simulator software (this case the Power World simulator). This program is able to generate max an admittance matrix directly in MATLAB form, and the voltages are contained as well. The necessary equations to calculate the sc. m currents: 0 0 i i

= arccos(cos + +

P ( )) P

= arccos(cos +

P ( )) Pmax

Vh0 I zh = Z hh
V j V j0 Z hj I zh =

From this calculation the angle changing caused by the failure could be got. From this parameter the angle acceleration should be calculated, this is the .

P H S = M i = i ni = s t fn M

generators and loads.

This method is similar to the TSI method, however the way of the calculation is a bit different. The t parameter was used, to compare the angle accelerations to the time, instead of critical clearing time.

3. Model description, simulations

For examination purposes a network analyzer software package developed by the Department of Electric Power Engineering, TU Budapest was used for the TSI calculations. With the help of the program steady states of networks (load-flow calculations), effects of several faults and breaker operations during transients can be analyzed. In addition to the steady-state analysis, dynamic simulations were also calculated. The program is able to handle two synchronous systems operating with different frequency. The t method was calculated by Power World Power Systems Analysis Software for the steady-state simulations. Power World Simulator is an interactive power system simulation package designed to simulate high voltage power system operation on a time frame ranging from several minutes to several days. The software contains a highly effective power flow analysis package, capable for solving systems of up to 100 000 buses effectively. Power World offers several optional add-ons to extend analyzing capabilities of the simulator. The model topology is shown in Fig. 2., which was created by using Power World SAS. The network model consists of:

The sources may be power and/or voltage controlled. In the model there are: two wind power parks connected to MO1S and MSZ20 buses, three 5 MW gas turbines (connected to the MO1G bus) and a large machine (connected to the G4 bus) as a system slack bus in case of parallel operation with large network. During island operation, the gas turbines are the system slack. In the two wind parks 25 x 2 MW wind turbines are in operation. The loads are frequency and voltage dependent. Different system states were calculated. The loads were changed in case of cooperated operation and in island operation as well. The island operation was established with a switch on MT1A-VN1A and MT1B-VN1B lines, and this way the two system frequencies were developed. The following examination consisted of several steps: the network operated as an island i.e. the MO1 bus and geographically close loads and sources form an island. The feasibility of island operation was examined. Different operation situations and faults were simulated; size of voltage change was examined. In cooperated and island operation the power was in- and decreased. From the results, it can be concluded that in some cases TSI provided relevant and valuable information on the dynamic security of the MG system; while in other cases, the t method has the same or even better results.
Cooperated operation M Ps[MW] dP-HTSW dP-Matlab -HTSW '-Matlab CCT MO1G 0,557 8,000 0,660 1,120 1,185 2,011 0,570 MO1S 1,751 38,440 1,810 1,890 1,034 1,080 MSZ20 #3phsc 0,700 7,980 8,040 8,040 11,481 11,481

t[sec] 0,492 Table 3.1.: calculation results in coop-op., the fault is at MSZ20 machine terminal

Fig. 2. The topology of the model network

high- and middle voltage lines and buses (400-120-20 and 10-kV) transformers lines, parallel-lines

5. Conclusion
Island operation MO1G M Ps[MW] dP-HTSW dP-Matlab -HTSW '-Matlab CCT 0,557 8,600 3,360 2,630 6,032 4,721 0,470 MO1S 1,751 38,000 4,670 1,290 2,668 0,737 MSZ20 #3phsc 0,700 7,980 7,980 7,980 11,395 11,395

t[sec] 0,542 Table 3.2.: calculation results in island-op., the fault is at MSZ20 machine terminal

4. Results of the simulations

Comparing the two characteristics, it is easy to recognize, that the t method provides accurate results. The t simulation calculations suit better at the expected characteristic, than the TSI method results. The accelerations in the two methods are very similar, 0.97 correlation was calculated. The critical clearing times however were more different.

The TSI-2 characteristic shows a clearly monotonous correlation between t values and critical clearing time. It is possible to get information on dynamic security of particular system states, based on the value of examined index, rather than by using other methods. A complete MATLAB program was created. With the help of this software, it is easy to get the TSI-2 characteristics. The Power World simulator can send the admittance matrix; afterwards the calculation is automatic. The stability indices were originally created to reduce number of calculations by substituting complicated, time-domain analyses. Nowadays this aim has lost its importance as a consequence of available computational capacity growth. However, stability indices can be very useful to monitor stability state of the system on-line by automatic calculations. In case of large system the number of calculations can be very high. The calculation of time-domain analysis is faster nowadays, but it takes plenty of time. In this cases could be useful the new method as well.

6. References
[1] Technical and Economic Feasibility of Micro grid-Based Power Systems Phil Barker, Doug Herman. Seventh EPRI Distributed Resources Conference and Exhibition Dallas, TX March 20-22, 2002 [2] Szab Lszl - Villamos hlzatok Pldatr. Kzirat. Budapest, 1982 [3] A. Faludi, L. Szabo: Power System Operation and Control (Hungarian) Lecture notes on vet.bme.hu/okt/foszak/ver/veri/index.htm, Budapest 2002. [4] Friedl, W. Fickert, L. Schmautzer, E. Obkircher, C.: Safety and reliability for Smart-, Micro- and islanded grids, SmartGrids for Distribution, 2008. IET-CIRED. CIRED Seminar [5] M. Ribbens-Pavella, P. G. Murthy, J. L. Horward, J. L. Carpentier: Transient Stability Index for on-line stability assessment and contingency evaluation. Intl. Journal of Power and Energy Systems, Vol. 4. 2, Apr. 1982.

Figure 4.1.: the characteristic by the TSI method

Figure 4.2.: the characteristic by the t method

Acknowledgement
The authors wish to thank for the financial support provided for this research project by the Foundation for the Education of Energy Engineers.