The Research Of Semi-physical Simulation And Detection System For An AVR Used In An Generator Excitation System

Safety and reliability are the basis for the survival of power stations and the entire power system,and they are also prerequisites for the sound development of the power industry.As the basic control device of the power system,the Automatic Voltage Regulator(AVR)adjusts the low-power magnetic field of the exciter,controls the output power of the rectifier,and adjusts the current of the generator's excitation winding to achieve the purpose of stabilizing generator's output voltage.However,the performance detection technology for such devices is still in its infancy,and the existing detection mechanism is not sufficient to detect the ‘hidden failure' that affects the dynamic performance of the card.The usual performance detection method is openloop test for the AVR,which can find its basic input and output characteristics.For this kind of dynamic adjustment performance of the card,open-loop test can not fully reflect the performance of the AVR.Therefore,this thesis takes this kind of card as the detection object,designs a customized intelligent detection system,builds a semiphysical simulation platform with AVR to creat a closed-loop testing environment and comprehensively detects the performance of AVR.Firstly,based on the full investigation of semi-physical simulation technology and AVR detection technology at home and abroad,this thesis combines with the actual situation of a power plant generator set and its excitation AVR.Then,the thesis points out the functional requirements of the semi-physical simulation and detection system,and proposes its design plan.Secondly,based on the requirements of semi-physical simulation and AVR performance detection,the thesis proposes a modular modeling method to establish a mathematical model in the system,and confirms the parameters in the mathematical model.The confirmation methods are divided into calculating parameter by mechanism method and identifying parameter by genetic algorithm.The final simulation verifies the accuracy of the model.Thirdly,for the established mathematical model,the computer needs a superior real-time simulation algorithm to solve the corresponding differential equations.In order to better fit the semi-physical simulation and detection system,this thesis absorbs the advantages of Runge-Kutta method and Adams-Bashforth method to build a kind of improved real-time simulation algorithm and analyzes the superiority of hybrid algorithm from stability,precision and real-time.Fourthly,the thesis expounds the detailed design and implementation process of the semi-physical simulation and detection system formed by the above research,and shows the results of the semi-physical simulation and detection system in an all-round way.The system also performs several performance tests on the AVR and verifies systematic reliability.In general,this thesis realizes the automatic closed-loop dynamic performance test on excitation AVR by the semi-physical simulation and detection system,which is with is developed by advanced semi-physical technology.The system's automation and intelligent design ideas provide a new direction for the development of the card closedloop performance detection device.