| The performance of the grounding grid is important to ensure the stable operation of the power station as well as the safety of operators and equipment of electric power system.The conductors of grounding grids are buried under grounding surface.After many years' running,they are eroded seriously in soil.Once the grounding grids have some faults,it will bring massive economical losses to the power devices and the people's lives property.Therefore,it is urgent to find an effective and feasible diagnosis method grid to avoid undesirable results in the instance of no excavating the grounding.This paper studied the method of fault diagnosis on grounding grids from the aspect of parameter identification and pattern recognition.Firstly,the grounding grid corrosion equations are established by using the application of electrical network theory,and fault branch's position are determined by solving the fault diagnosis equation based on genetic and simulated annealing algorithm.It has a high speed of convergence rates,and the advantage of local optimal searching.The results show that:the method has a high computing speed and diagnosis accuracy.The paper proposes a fault diagnosis method on grounding grids based on SVM parameters optimization.To improve classification performance,adaptive genetic algorithm is applied.The support vector machine has faster convergence speed and higher correct recognition rate after optimizedFurthermore,this paper presents a fault diagnosis method on grounding grids combine wavelet packet energy and support vector machine.The grid conductors are treated as ? models which take the grid conductor's resistance and reactance into account and the node voltage waveform can be gotten.Wavelet packet decomposition and reconstruction is used to dispose the waveform signals and construct input samples of the support vector machine,and the accurate diagnosis results can be obtained by support vector machine. |
PDF Full Text Request