Small Signal Stability Analysis of Wind Turbine Penetration in Sulselrabar Interconnection System

This paper present and explain the analysis of power flow and small-signal stability of the electrical system in the islands of South Sulawesi, East and West (Sulselrabar) located in Indonesia. The research focused on the Eigen values analysis that used to investigate small-signal stability performance of the main electricity grid when the fault conditions with and without control equipment as well as the inclusion of wind turbines large scale on the system. Power Systems Analysis Toolbox (PSAT), which integrated in MATLAB is used to develop a network topology in Sulselrabar. This study allowed us to determine the optimal location of the placement of Power System Stabilizer (PSS) to dampen the oscillation. The simulation result shows the placement of PSS on two generators can add system stability and help the system to resume operation at the balance point. Time domain simulation methods are used to see the frequency response of the rotor speed.


Introduction
The growth of electricity consumers lead to increase problems in the stability of the power system and become an important issue in the planning of electric power systems [1][2][3][4]. During the system under normal conditions, the balance between the needs of active and reactive power supply can be assured. The stability of the system will has a significant effect when all the isolated electricity networks interconnected as planned in the development of electricity networks [1][2][3][4][5].
Because of its large and complex system [6], the low frequency oscillation phenomenon needs more attention to the potential oscillation power problems that have a direct impact on the stability of the system [5][6][7][8][9]. The stability of the power system components which is related to the system in providing balance and improving the state during disturbances [7], one of the effects of these disorders is the electromechanical oscillation.
The electromechanical oscillation consists of two mode which are the local oscillation mode and inter-area oscillation mode [8]. This study focuses on the phenomenon of small-signal in the electrical network systems in South Sulawesi, East and West (Sulselrabar), one of the islands in eastern Indonesia. With small signal stability analysis, the performance of the power system in a minor disturbances state can be evaluated [10]. Small Signal Stability can be defined the system's ability to stay in synchronization when subjected to small perturbations [11]. The small disturbances phenomenon has frequency range between 0 to 2 Hz [5]. Whereas the local oscillation mode has International Conference on Computation in Science and Engineering IOP Publishing IOP Conf. Series: Journal of Physics: Conf. Series 1090 (2018) 012034 doi : 10.1088/1742-6596/1090/1/012034 typical frequency range of 1-2 Hz and the inter-area oscillation mode has a range of frequencies below 1 Hz [8].
Power System Analysis Toolbox (PSAT) software is used to analysis study of power system [12]. This toolbox capable of perform power flow studies [13], the stability of the small-signal and time domain simulation to see the response and rotor angular velocity. To provide damping during the placement of Power System Stabilizer (PSS) isolation is connected to the Automatic Voltage Regulator (AVR) become priority [7]. PSS placement is based on eigenvalue damping which obtained when simulation is done.

Research Methodology
These case studies are used as a test case to the electric grid interconnection system Sulselrabar. Furthermore, the electrical system topology Sulselrabar modeled and modified using a Power Systems Analysis Toolbox (PSAT), which has been integrated in Matlab Simulink®. This toolbox is open source software used to analyze and study the power system [14]. The toolbox can run the power flow, continuation power flow, optimal power flow, small-signal stability analysis and time domain simulation. All operations can be assessed by means of Graphical User Interfaces (GUI) and simulink which provides a variety of tools for network design.

Power System Modeling
All dynamic models of the system are available and have been documented in the manual PSAT. Include synchronous generator model, turbine governor and Automatic Voltage Regulator (AVR) model. Simple version of the synchronous generator contained in PSAT model system, where the dynamic parameter of each of the synchronous generator as the input to d-axis and q-axis. Dynamic parameters are shown in Table 1 All mechanical damping and disregarded of the effect on the overall constellation generator. AVR models used is the model 3 and used for all generators but with different parameters. At the PSAT, Turbine Governor represent in two types of models, namely: models 1 and 3. Where the thermal power plant in the state and represented the first model (1) and the hydro plant model expressed and represented the third model (3). However, in this study, Model 3 [15] is used. In this model, some properties of the system regarded as inelastic penstock where the inertia of the water to be considered, and also the ideal turbine. For simplicity of analysis [18][19][20][21], static load models are also used [1]. As for wind turbine used doubly fed induction generator (DFIG).

Modeling of the Sulselrabar Power Grid
Indonesia is an archipelagic country that has a complex electrical system between the island and other islands. Sulawesi islands comprising South Sulawesi, East and West (Sulselrabar) used for this research study. Interconnection system Sulselrabar strut by several generating units where the supply of Gas Power Generation and Sengkang Steam Power, Bakaru and Poso Hydro, Jeneponto and Tello Steam is supply to the largest generation system interconnect Sulselrabar [16]. Sulselrabar interconnection system consists of 15 generators, 44 bus 47 lines and 34 load spread, whereas the voltage varies from 30, 70, 150 to 275 kV [17][18][19][20].

Case Study
The case studies in this research is to analyze the small signal before and after the inclusion of wind turbine large scale on Jeneponto bus at Sulselrabar interconnection system as shown in Figure 1 Dynamic parameters and excitation of each generator is shown in Table 1 and Table 2 to see the response of the system using power flow studies, eigenvalue and time domain simulation. The first simulation model of the system under normal conditions, whereas system must demonstrate its stability. It is important before doing the next phase simulation. After condition is stable by eigenvalue analysis, the second stage simulation is by giving disturbance to the system in the form of termination of the channel on bus of Parepare and bus of Barru. The first intervention given time is 1 second and the second intervening time is 200 second.  The third simulation is the modeling system with the addition of PSS control the event of termination of channel. The simulation further analyzes the system when the inclusion of wind turbines on a jeneponto bus with and without PSS control equipment.
PSS placement will be based on the critical eigenvalue or close to zero. There are 44 buses, 47 lines, 6 transformers, 15 generators and 34 loads will become the object of this research.
This study will use conventional thermal generators units, where for Turbine Gevernor (TG) used Model 1, while Hydroelectric unit used Model 3 TG and used to the whole system. As for wind turbine used model doubly fed induction generator (DFIG) with Nominal wind speed 15.00 (m/s).

Results and Discussions
In this paper, comparison between the system between with and without using the control damper and damper control is done. The simulation results were observed the deviation rotor angle (δ) or rotor angular velocity (ω) of each generating unit. Then, small signal stability analysis is used with the system. Figure 2 shows the PSAT Sulselrabar eigenvalue models under normal conditions, dynamics condition without PSS and dynamics condition using the PSS.  Figure 2a shows the normal conditions on the electricity network of Sulselrabar interconnection system. In this condition has not been given any input in the form of interference or load shedding. Figure 2b shows the dynamics system after being given the disruption of the termination of the transmission line between the buses 4 and 11 (bus Barru and Bus Parepare). As for Figure 2c shows the shift of eigenvalue after the addition of control given in the form of PSS. The placement of PSS is randomly based on critical or near-critical value of eigenvalue. From the results of the experiment placement of the maximum PSS is on the two generators in reducing oscillations of the system.

System Simulation before Entrance wind turbine.
For analysis purposes, the participation factor used to determine the bus and generator weakly damped and most affected by instability. From the analysis found that the generator 2, 6, 7, 12 and 13 represents the weakest bus with eigenvalue levels approaching critical value. There are 111 state and 31 pairs resulting complex to be initialized and observed.   Figure 3b shows the voltage profile of the whole bus is still in the voltage allowed range by the Energy Regulatory Commission from 0.95 to 1.05 p.u. Where in Figure 3a represents the voltage profile using with PSS and Figure 3b represent the voltage profile picture without the PSS. Table 3. Improvement of Voltage Magnitude Before and After using PSS For improvement of voltage magnitude as seen in Table 3, there is a change before and after the installation of PSS. Where there is improvement for each bus voltage. For example at a bus 10 (Suppa bus), an improvement of voltage from 0.99802 p.u becoming 0.99909 p.u or voltage changes 0.00106 p.u after the addition of PSS.   Table 4 and Table 5 shows participation factor of eigenvalue before and after the addition of PSS. From Table. 4. it can be seen that the value of damping ratio increased for the fifth generator. For the second generator damping ratio increased from -0.0059 to -0.0061. Also with the next generator the damping ratios will increase or will improve the eigenvalue. With the correct input parameters of the PSS, the real and imaginary parts of eigenvalue 59-60 improved from-0.0262 ± j4.4241 become -0.0269 ± j4.4226.
The next thing to observe is the rate speed of the rotor to return to a stability position after interference. From the simulation results using time domain on the PSAT, shows good damping occurs after the addition of PSS. Whereas the rotors speed dampen oscillations and back at the point of equilibrium faster than before using a PSS. This suggests that the placement of PSS and PSS proper tuning values can restore the stability of the system and reduce oscillation well in old or new balance point. Here is a graph of rotor speed (ω) on generators 2, 6, 7, 12 and 13 which are the weakest generators of the simulation results prior to the entry of wind turbines. The black graph is without the PSS and blue graph is using the PSS. From the simulation results above indicate the difference graph generator rotor speed on the state of good dynamics with and without PSS control equipment. Figure 4 shows the rotor speed response to damp the oscillation and return to the equilibrium point more slowly in the appeal using PSS. For example, it can be seen in generator two, in which the rotor speed when its dynamic insulated and downhill. Isolation and degradation the speed at the start of the first second until the seconds to 30 and continued to decline to below the rate of 49.3 Hz. Free time to survive in the value of 49.8 Hz and 49.9 Hz climb back up to speed eventually decreasing.  Slowly from the 7th to the 30th seconds the speed increases and the oscillation damping is good until it returns to its normal density. This shows the use of PSS control equipment in reducing the occurrence of better and faster oscillations than without the use of control equipment.  Tables 6 and Tables 7 show the participation factor of the eigenvalues before and after the addition of PSS after the inclusion of wind turbines. It can be seen that the damping value of the ratio increases for the five generators. For generator six, the damping ratio increased from -0.005488 up to -0.005748. Likewise with the generator further increase or improve the eigenvalue after the addition of PSS. With correct input parameters of PSS, the real and imaginary parts of the eigenvalues of each generator are improved. In the two generators eg. from -0.0426 ± 5.0118i it has improved to -0.0425 ± 5.0189i. Likewise with generators 6,7,12, and 13. Table 6. Linear Value Analysis of The System Without PSS Table 7. Linear Value Analysis of System with PSS For rotor speed, from the simulation results obtained that the entry of wind turbine in the interconnection system of Sulselrabar very influential on the stability of the system. From Figure 6, it can be seen that the fifth rotor density of the generator fluctuates from the first second to the 30th seconds. At 25 seconds the rotor speed penetrates 50.03 Hz at the upper boundary and 49.97 in lower limit.   Meanwhile, with the addition of PSS control, the rotor speed on the five generators is relatively more stable in reducing the oscillations that occur. Figure 7 shows where the speed of the five generators is more stable. The rotor speed does not exceed 50.02 on the upper boundary and 49.97 at the lower limit and gradually returns to its normal density.

Conclusions
This paper explains the power flow and small-signal stability. Through the PSAT toolbox integrated on Matlab, multi-machine system of interconnected systems modeled Sulselrabar. Conventional thermal generating units, whereas for Turbine Governor (TG) used Model 1, while using the unit Hydroelectric Model 3 TG. For wind turbine using model doubly fed induction generator (DFIG). Effective implementation and proper placement of the device PSS can cope with the dynamics of multi-machine network stability in the system and increase the critical eigenvalues as well, whether that the oscillation mode between areas and the local oscillation mode. Besides that the implementation of PSS can reduce and minimize overshoot oscillation frequency oscillation. Eigenvalues, voltage, frequency, and the factor observed with either participation in this study.