Available online at www.sciencedirect.com
ScienceDirect Energy Procedia 100 (2016) 261 – 265
3rd International Conference on Power and Energy Systems Engineering, CPESE 2016, 8-12 September 2016, Kitakyushu, Japan
Study on an Improved Model Predictive Control Strategy with Power Self-coordination for VSC-MTDC Jiang Binkaia, Wang Zhixina,*, Zhou Jianlongb, Shi Lic a
Department of Electrical Engineering, Shanghai Jiao Tong University, 800 Dongchuan Road, Shanghai 200240, China b Jiaxing Renewable Power Electrical Co., Ltd, 1888 West Zhongshan Road, Jiaxing Zhejiang 314031, China c Shanghai Najie Electric Co., Ltd, 2000 Guanghua Road, Shanghai 201111, China
Abstract Multi-terminal HVDC transmission based on the development of HVDC is an important way of power transmission of new energy in the future. But its control strategy is complex, especially the coordinated control between stations. A three-terminal parallel VSC-MTDC simulation model in the MATLAB/Simulink platform is built in the article. An improved model predictive control strategy with power self-coordination is put forward. A single converter station takes the model predictive control with power consumption as the cost function. And the coordinated control between converter stations is fulfilled by DC power control which is controlled by the cost function in model predictive control. Thus, coordinated control between converter stations and the control of converter are of organic combination through the cost function. When disturb occurs in any converter station except the main converter or one converter station is out of work, the main converter station provides the power compensation to ensure the system keep working, Without the need for additional coordinated controller design, greatly reducing the economic consumption, improving economic benefit. And each converter station is relatively independent, the system stability is better. In general, the simulation results prove the effectiveness of this method, which provides a new way for VSC-MTDC control. 2016The TheAuthors. Authors. Published Elsevier © 2016 Published by by Elsevier Ltd.Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of CPESE 2016. Peer-review under responsibility of the organizing committee of CPESE 2016 Keywords: VSC-MTDC; power self-coordination; model predictive control; MATLAB/Simulink
1. Introduction With the depletion of traditional energy, new energy has been gotten more and more attention. Of all the new energy sources, wind energy has its unique advantages: the large capacity of development, clean and low cost. These
* Corresponding author. Tel.: +8618916218890 E-mail address:
[email protected]
1876-6102 © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Peer-review under responsibility of the organizing committee of CPESE 2016 doi:10.1016/j.egypro.2016.10.175
262
Jiang Binkai et al. / Energy Procedia 100 (2016) 261 – 265
advantages make it favored in the grid system study. At present, the long distance transmissions of large offshore wind farms usually adopt the VSC-HVDC system [1,2]. Based on two-terminal HVDC transmissions, multi-terminal HVDC technology [3,4], that is MTDC, has gradually improved and applied. Compared with the two-terminal HVDC system, multi-terminal DC transmission system has to consider the coordination control between each station besides the control of converter. Thus, effective control strategy is essentially important for the stable operation of VSC-MTDC transmission system. For converter control, lots of control strategies has been put up and applied such as direct power control (DPC), PID neural network (PIDNN) control and double closed-loop control [5,6,7]. Considering the complexity of control strategy for VSC-MTDC, An improved model predictive control with power self-coordination is put forward in the article. In the method, a single converter station takes the model predictive control with power consumption as the cost function. And the coordinated control between converter stations is fulfilled by DC power control which is controlled by the cost function in model predictive control. A 3terminal VSC-MTDC system and the improved model predictive controller are built in MATLAB/Simulink, and the validity and rationality of the designed controller is verified. 2. Model predictive control 2.1. The mathematical model of VSC Due to the structural similarity of the rectifier and the inverter, a three-phase voltage grid-connected inverter is taken for an example. Assuming the three phase voltages of grid are of balance, the current equations can be expressed in α, β coordinates [8]. diD °° L dt ® ° L diE °¯ dt
Discrete the equation (1):
U D eD RiD
(1)
U E eE RiE
iD k 1 iD k UD k eD k RiD k °L Ts ° ® ° L iE k 1 iE k U k e k Ri k E E E ° Ts ¯
(2)
From the equation (2): °iD k 1 ° ® °i k 1 °¯ E
And then from the equation (3):
Ts § RTs · ¨1 ¸ iD k UD k eD k L ¹ L © Ts § RTs · ¨1 ¸ iE k U E k eE k L ¹ L ©
°id k 1 iD k 1 cos T iE k 1 sin T ® °¯iq k 1 iD k 1 sin T iE k 1 cos T
(3)
(4)
Actually, this is the predictive current. As the grid voltage maintains stable, it can be assumed it is constant. So the active power and the reactive power can be got:
° P k 1 ed id k 1 eqiq k 1 ® °¯Q k 1 eqid k 1 ed iq k 1
Actually, this is the predictive power.
(5)
263
Jiang Binkai et al. / Energy Procedia 100 (2016) 261 – 265
2.2. Model predictive control theory Model predictive control technique is received extensive attention in recent years because of its relatively simple principle and it can be easy to achieve in digital. For the converter control, it can be understood as finding the right gate- driving signal to control the variable x(t) of the system. Where the driving signal can make the x(t) as close as possible to the desired reference variable x*(t). Model predictive control can only be applied to the situation in which the status of variable is finite. Compare all the possibilities, and find out the most suitable driving signal which will be chosen as the next control signal. By the analysis of transient behavior of the system variable x(t),the predictive function of the variable, which can predict the future behavior of the variable, can be gained. In general, a predictive function can obtain many results, and how to select the optimal result relates to the concept of cost function. A typical cost function is: gi=[x*(tk)-xi(tk+1)]2.In addition, the absolute value of the difference between the reference value and the predicted value is also usually used as cost function [9,10,11]. The model predictive control contains two steps. Step 1: find the predictive function of the variable. In the article, the current from the VSC is chosen as the variable of the function. According to the equation (4), the predictive function can be defined the same as equation (4). But considering the control of VSC, we take the equation (5) as the predictive function. Step 2: choose the suitable cost function. Since the equation (5) is taken as the predictive function, the cost function must have some relation with the power. And two cost functions are given for different converter stations. Given to the principle of power balance, for the main converter station VSC1, the cost function is defined as:
g | (Q* Q1 Pdc P2 P3 P1 |
(6)
and for the other two stationsˈ VSC2 and VSC3, the cost function is defined as:
g | (Q* Q | | ( P* P) |
(7)
Figure 1 shows the block diagram of model predictive control. And Figure 2 shows the flow diagram of model predictive control.
Fig. 1 Block diagram of model predictive control
Fig. 2 Flow diagram of model predictive control
264
Jiang Binkai et al. / Energy Procedia 100 (2016) 261 – 265
2.3. Realization of the power Self- coordination Taking the model predictive control as the local controller, for the DC voltage and DC power, a PI controller is designed to adjust; for the power balance between converter stations, taking the equation (6) and (7) as cost function, only the switch status which meet the requirement of power balance are chosen. Thus, the power between stations is automatically coordinated, and the need of additional coordination controller is eliminated. The main converter station (VSC1) works in DC voltage control mode, and the VSC2 and VSC3 are both in the power control mode. 3. Simulation and analysis A 3-terminal VSC-MTDC system is built in MATLAB/Simulink to verify the control strategy. The simulation parameters are listed in Table 1. The responses of the system are discussed in the following situations :(1) normal working state; (2) the power of VSC2 changes; Table 1 The simulation parameters Converter
Line
Transmission
DC
DC
Impedance
sampling
station
voltage
power
capacitor
voltage
at grid side
frequency
VSC1
10kV
9MW 2000uF
20kV
R=0.2Ω
VSC2
10kV
5MW
L=12mH
VSC3
10kV
4MW
C=0
20kHz
Figure 3 shows the response of the system when in normal working condition. The power and DC voltage are both very stable. Besides, the DC voltage is at the same value as the set value, maintaining at 20kV, which means that the model predictive control strategy is effective. 107
DC Voltage (V)
2
P1
P2
1 0.5 0 -0.5
10 4
2.1
1.5
P3 0.1
0
0.2
0.3
Time (s)
0.4
2.05 2 1.95
0.5
1.9
0
0.1
0.2
0.3
0.4
0.5
Time (s)
(a)
(b) Fig.3 System response when working normally
7
2.5
DC Voltage (V)
Active Power (W)
10
2
P1
1.5 1
0.5 0 -0.5
P3 0.1
0
P2 0.2
0.3
Time (s)
0.4
0.5
10
2.1
4
2.05
2 1.95 1.9
0.1
0
0.06
0.08
0.1
Time (s )
0.3
0.4
0.5
0.4
0.5
(b)
4
0.12
0.14
AC Curre nt o f VS C2 (A)
10 3 2 1 0 -1 -2 -3 0.04
0.2
Time (s)
(a)
AC Voltage of VSC2 (V)
Active Power (W)
2.5
500 0 -500 0
0.1
(c)
0.2
Time (s ) (d)
Fig.4 System response when power of VSC2 changed
0.3
Jiang Binkai et al. / Energy Procedia 100 (2016) 261 – 265
As shown in Figure 4, when the active power of VSC2 changes, decreasing from 5MW to 3MW, the power of DC system gets unbalanced. The power injection of DC networks is larger than its output power, and it makes the DC voltage rise, which changes the operating point of system. However, the system can automatically adjust the power to be balanced according to the power deviation by the power compensation of VSC1. And that makes the system be in a new operating point where the active power is balanced and the DC voltage remains the 20kV, stable. The results also show that the control strategy can improve the anti-interfere ability of system. 4. Conclusion Using the cost function in the model predictive control, we propose an improved model predictive control with power self-coordination for VSC-MTDC. The most obvious character of the control strategy is that there is no need to design the extra coordination controller. When one converter station mutates, the system can maintain stable through the PI controller and model predictive control, and the power between stations is automatically coordinated to be balanced. That can greatly improve the operation reliability of the system. The simulation results in MATLAB/Simulink show that the proposed control strategy is not complicated. Compared with the traditional control strategy, it is easier to be realized with good control effect, which provides a new way for multi-terminal HVDC control. Acknowledgements The paper was supported by Project of National Natural Science Foundation of China (51377105), Project of Science and Technology Plan of Jiaxin(2014BZ15002), and Project of Science and Technology Plan of Shanghai Minhang District (2014MH103, 2015MH103). References [1] Zhixin W, Chuanwen J, Qian A, et al. The key technology of offshore wind farm and its new development in China [J]. Renewable and Sustainable Energy Reviews, 2009, 13(1): 216-222. [2] Wu J, Wang Z X, Xu L, et al. Key technologies of VSC-HVDC and its application on offshore wind farm in China [J]. Renewable andSustainable Energy Reviews, 2014, 36(8): 247-255. [3] Zhang Wenliang, Tang Yong, Zeng Nanchao. Multi-Terminal HVDC Transmission Technologies and Its Application Prospects in China [J]. Power System Technology, 2010, 34 (9): 1-6. (in Chinese) [4] Tang Guang Fu, Luo Xiang, Wei Xiaoguang. Multi-terminal HVDC and DC-grid Technology [J]. Proceedings of the CSEE, 2013, 33 (10): 817. (in Chinese) [5] Wang Guoqiang, Wang Zhixin, Zhang Huaqiang, et al. DPC-based control strategy of VSC-HVDC converter for offshore wind farm [J]. Automation of electric power system, 2011, 31 (7): 115-119.(in Chinese) [6] Wang Guoqiang, Wang Zhixin. Application of PSO and PIDNN Controller for VSC-HVDC [J]. Proceedings of the CSEE, 2011, 31 (3): 8-13. [7] Zhang B, Yan X, Xiao X, et al. The VSC parallel structure and control technology for the centralized V2G system[C]//Industrial Electronics (ISIE), 2013 IEEE International Symposium on. IEEE, 2013: 1-6. [8] Wu Jie, Wang Zhixin, Wang Guoqiang,et al. Backstepping control for voltage source converter-high voltage direct current grid side converter [J]. Control Theory & Applications, 2013, 30(011): 1408-1413. (in Chinese) [9] Rodriguez J, Pontt J, Correa P, et al. Predictive power control of an AC/DC/AC converter[C]//Industry Applications Conference, 2005. Fortieth IAS Annual Meeting. Conference Record of the 2005. IEEE, 2005, 2: 934-939. [10] Rodríguez J, Pontt J, Silva C A, et al. Predictive current control of a voltage source inverter [J]. Industrial Electronics, IEEE Transactions on, 2007, 54(1): 495-503. [11] Yang Yong, Zhao Fangping, Ruan Yi, et al. Model Current Predictive Control for Three-Phase Grid-Connected Inverters [J]. Transactions of China Electrote, 2011, 26(6): 153-159. (in Chinese)
265
