The Studies Of Physical Model Of Unified Power Flow Controller

FACTS (the Flexible AC Transmission System) technique is the outcome of combination with the power electronics, the modern control theories and the computer techniques. By fast regulating the voltage, reactance, phase angle of AC transmission system, it can achieve the flow control of active and reactive powers, therefore improve the stability of the system. And at the same time, the transmission capability of traditional system is increased, line could be fully used, and the aim of safety, reliability, economy operation of such system gives satisfaction. UPFC (the Unified Power Flow Controller) is one of the strongest equipments in function among the FACTS. Its purpose lies in providing necessary, more effective, more flexible methods for real-time control and dynamic compensation of AC Transmission System. The UPFC changes the control method of traditional equipments of FACTS which can usually regulate one or some parameters, by combining shunt and serial equipments together, and optionally affects almost all factors related to line flow in real-time, thus can flexibly control the parameters such as active power, reactive power, voltage, reactance, phase angle of power line, and resolve a series of problems in the power transmission system. Thereby, it has really fine developing foreground. And, its researches and applications have caused more and more extensive attention in the area.Based on the practical job of building a 20KVA UPFC physical model, this dissertation deeply studies the issues of principles, structures, control strategies, control methods, realization methods of UPFC with emphasis on the control methods and implementations of bottom controllers. When finished, the UPFC physical model will be an integrated equipment used for the purposes of research and teaching in laboratory.The main researches in this dissertation are:1. Investigations of UPFC principle: In the first place, the principle and regulating range in static state of UPFC are analysed in detail. Then, by compare it with other traditional FACTS equipments, its advantages are explained. Based on above, various working patterns and control strategies of UPFC are put forward.2. Main power circuit studies of UPFC physical model: In consideration of various factors like practicability and extensibility, based on a lot of analyses and studies, the main power circuits of the model use basic topologic structures of three-phase voltage sourced inverter with high power IGBT devices as their switching elements . To fully use DC bus voltage, whether for shunt current source or for series voltage source, the modulating method is selected as SVPWM (Space Vector PWM). With the SVPWM basic principle as start point, the dissertation educes a new fast algorithm which may be easy realized by software of microcontroller or DSP. This method with its anti saturated mechanism is simpler, more effectual and reliable comparing the conventional algorithm because it avoids a mass of coordinate transformations and arithmetic and triangle function calculations. Then, from the practical point of view, to the issues of capacities and generated harmonics of SVPWM based inverter, bring forward a new structure using multi SVPWM bridges. And also, its characteristics are analysed and studied in details. Further more, from above discusses, the design ideas, key issues and parameters of the main power circuit of the developed UPFC physical model are provided.3. Hardware of UPFC control system and its implementation: Considering the complexity of the task, the UPFC control system is designed according to the fundamental of multi-layer that means the whole control system is divided into a few subsystems for different purpose. Grounded on this principle, the structures for the whole system and each controller with advanced DSP chip as the core are proposed first, and then all the system hardware was designed, built and adjusted.4. Researches on signal measurement algorithm: In UPFC control system, to get good, fast and satisfactory performance, fast signal measurements are basic foundation. In order to realize the fast measurements of parameters such as RMS, active power, reactive power, phase angle and reactance, an instantaneous method, which use DFT principle in a sliding window, is put forward and studied. Then, to reduce the calculations, two algorithms that are the recursion algorithm with least calculations and the fast algorithm with avoiding the cumulation errors are proposed. Therefore, make UPFC fast control possible.5. Control of shunt current source: First, build generalized math model of current source as controlled plant in the current control system. On this foundation, three control methods, which can force the system outputs following the sinusoidal reference inputs instantaneously, are suggested, namely: basic close loop control (deadbeat control), neural network (NN) based predictive control and NN+Fuzzy logic based hybrid control. With regard to system performance and implementing feasibility, they could all be applied to the control of current source, but with different structures, algorithms and characteristics. Deadbeat control, which is a common method used in the control of power electronic device, is capable of realizing fast close loop control, however, when the control and SVPWM algorithm implemented by software of DSP or other microcontroller, the system output will delay the input more than one steps. If the reference input is sinusoidal, the delay will generate a obvious follow error in static state. Further more, this control has an inherent disadvantage that is the system performance being sensitive to the plant's parameters. Neural network (NN) based predictive control, which adds two liner NN predictors to the above controller, can effectively eliminate static error caused by calculation delay. When plant's parameters are known accurately, the static sinusoidal error may be reduced to zero. NN+Fuzzy logic based hybrid control, by introducing a fuzzy logic loop on the base of NN predictive control, can eliminate the influences of nonlinearities and uncertainties, therefore overcome the weakness of deadbeat control. Our research proves that the proposed hybrid control is best among the three methods and can be easy implemented by the designed DSP system.6. Researches on power exchange mechanisms of DC link and its bus voltage control: With the power exchange mechanism of DC bus as the start point, build reasonable math model for DC voltage control, and suggest a kind of active current modulator which ensures least harmonics generated by current source, and then, propose two methods for the DC bus voltage control, namely, basic PI control and nonlinear PI control. By theoretical derivation and comparative tests, point out that if we only consider the situation of small disturbance, these two strategies may regulate the DC voltage well, thus guarantee the power exchange between shunt current source and series voltage source of UPFC. However, if large disturbance must be taken into account, the second one without any additional computation will be better than the basic PI control and is very suitable for the developed UPFC physical model. Moreover, from two sides of static and dynamic state, derivate the formulae for theselection of capacitance, thus, provide a new method for the design of UPFC.7. Control of series voltage source: To make the output of series voltage source follow the sinusoidal reference input accurately, and ensure the system has the good robustness when the parameters of system the UPFC connected to are changed, the following researches are down in details: (1) build continuous math model with line inductance for the control of voltage source and educe its corresponding discrete model. From this, prove that because of the output filter, the system plant has weak damp. (2) To the characteristic of the weak damp, put forward the correct method by using state feedback or part of them, and therefore, make the plant has better dynamic performance and become more robust. For doing this, two methods are studied. The first one is based on the Lyapunov theory and matrix inequation, and another is based on Genetic Algorithms to enhance the system performance and robustness further. By combining above two methods, namely, using the first one to find a rough set of state feedback parameters, and then applying the second method to find the better feedback parameters which can be used when the parameters change in the larger range. The generalized plant after these corrections, therefore, has not only better dynamic performance but also satisfactory robustness. (3) Under the condition of sinusoidal reference input with certain frequency, say 50Hz, the system output usually could not follow the input perfectly because of the generalized plant's lagging or leading characteristics. To solve this problem, two strategies, namely, repetitive control and intelligent integral control, which form the closed loop further for the system output voltage, are introduced. These two methods are all capable of making system output follow the input without obvious errors, but by comparison, the latter has more advantages and suitable for the voltage control of UPFC, and the better static and dynamic performance could be obtained.