Identification Of Low Frequency Oscillation In Power System Based On Adaptive Internal Mode Algorithm

The safe and stable operation of the power system is an important prerequisite for ensuring the stability and prosperity of the national economy.With the expansion of the scale of interconnected power grids and the increase of power equipment,the low frequency oscillations in power grid,which due to negative damping,forced oscillation,nonlinear and other factors,has become an important reason for threatening the stable operation of the power grid and affecting the quality of power supply.The parameters identification of low frequency oscillation signals has become a fundamental problem of great significance in modern power science research.In order to suppress and eliminate the impact of low frequency oscillation signals on interconnected power grids,further research is needed on the online identification method of low frequency oscillation signal parameters(frequency or amplitude,etc.)in the power system.This dissertation has carried out related researches on the parameter identification of low frequency oscillation signals in power systems.Mainly include the following:First,this dissertation introduces the research background and practical significance of the topic selection,briefly gives examples of low-frequency oscillations at home and abroad that endanger the stable operation of power systems,and then introduced the mechanism of low frequency oscillation and the relatively mature identification method for low frequency oscillation signal in power system at home and abroad.Subsequently,the low frequency oscillation model is studied in this dissertation and the identification target to be achieved are also given.Several commonly used parameter identification methods are briefly introduced and their characteristics are briefly analyzed.Secondly,improvements based on the adaptive internal model recognition algorithm are given.A robust adaptive algorithm based on a three-dimensional dynamic system is proposed to realize the recognition of unknown parameters such as the frequency and amplitude of periodic disturbance signals with unknown frequencies.The conditions of asymptotic convergence of dynamic systems are analyzed in turn through time scale changes,variable substitution,slow integral manifolds,averaging methods,Lyapunov stability theorem,and Mathieu equation theory.Then gives the simulation analysis of algorithm results,and compared with the original algorithm,which proves the superiority of the new algorithm.Subsequently,the algorithm discretization based on the fourth-order Runge-Kutta method is studied,and the realization of the discretization algorithm is completed by writing an m-file program for subsequent debugging and application in actual engineering.Thirdly,this dissertation presents the design of detection device based on the improved algorithm.The signal source part configures and communicates with the FPGA through TMS320C5517,and then uses DDS technology to generate the desired waveform by addressing the waveform lookup table in the memory.The signal processing part use the eight-channel sampling AD7606 sampling chip to collect the identified signal,use TMS320F28335 to discretely calculate and process the collected signals,and use DAC8552 for DA conversion and waveform result output.The hardware design of the above parts are given in turn,including the signal source part configuration circuit,signal source output circuit,DSP power supply circuit,AD acquisition circuit,DAC conversion and output circuit,communication module circuit,etc.,The corresponding software design and related flowchart are also given,which provides a certain engineering application foundation for the identification of low frequency oscillation signals of unknown frequency in power systems.Lastly,the conclusion of the whole dissertation and some expected research work are illustrated.