A Study Of Exception Detection In Power Distribution Network With Large Random Matrices

Exception detection in power distribution network has been an important part in its handling process.The development of technologies related to WAMS(wide area measurement system)has improved the quality of data collected in power grids.Consequently the detection and location process could be advanced when using large random matrices theories to analyze those globally collected data.Firstly his dissertation provided the mathematical model of PMUs(phase measurement unit)and its solution,and practiced it in the simulation of power distribution network.Secondly three approaches are introduced to detect exception supplied by large random theories,which are spectral distribution judgement,maximum-minimum eigenvalue judgement and trace judgement.Large random matrices are built of data collected from PMUs,showing their different features under the judgements above.Three approaches are also practiced with different short circuit and harmonic wave exceptions,and compared in fields of sensitivity and robustness.Finally the dissertation proposed the method of exception location with PCA(principal component analysis),which is also tested in different short circuit and harmonic wave cases.The result of simulation shows that three approaches of detection and method of location can be both acute and prompt,which provides a new way in exception handling of power distribution network.