Research On Fault Location And Layout Optimization Of Distribution Network Based On Compressed Sensing Technology

Distribution network undertakes the task of power supply directly in power system.For the reason of its wide coverage,complex topology,various types of loads and changing operating environment,these features lead to the results of more fault-prone,trouble locating and the reliability of power supply is also affected seriously.Therefore,realizing the accurate and fast fault location in distribution network can not only shorten the interruption duration and reduce the blackout area,but also improve the reliability and stability of the power system significantly.The paper presents a fault location method based on compressed sensing for distribution network,and the measuring points are optimized based on this method.The main contents of research are as follows:Firstly,the relationship between the node voltage sags and the node impedance matrix is established according to the Kirchhoff's voltage law.The analytical formula of voltage sags between fault node and non-fault nodes is deduced based on symmetrical component method.Then the full cycle Fourier algorithm is used to calculate the voltage sag value before and after the fault,In order to determine the duration of voltage sag,discrete wavelet transform(DWT)is used to calibrate the fault onset time,the fault type is recognized by the amplitude difference of three phase voltage before and after fault.Secondly,the relationship among the node fault voltage sag and the node impedance matrix and the node injection current is used as the constraint condition of the mathematical model,This paper constructs a mathematical model of node injection current reconstruction based on the principle of compressed sensing,which is based on the principle of compressed sensing.So the problem of fault location in distribution network is transformed into a mathematical model with minimum l1 norm.Then the method of solving the mathematical model is analyzed,and the dual alternating direction algorithm(DADM)is used to reconstruct the node fault injection current vector.As a consequence,the effectiveness and accuracy of the fault location method for distribution network based on compressed sensing technology is verified by simulation and experiment on a simple distribution network topology with 23 nodes based on PSCAD.Eventually,the adjacent nodes of the fault nodes are chosen as the optimization algorithm based on the observation area(MRA)during the observation point optimization method.Establish a considerable number of measuring points to achieve the global minimum as the objective function,the dual constraints of the high probability reconstruction constraint and observable region of voltage sag constraint are used as the constraint conditions of the optimization objective function.Finally,the objective function is planned for several times,at last,the configuration scheme of fault voltage sag of distribution network with multi-branches and multi-nodes is determined.The simulation and experiment test topology with 134 nodes from actual power grid are carried out.The effectiveness and robustness of the proposed method are verified by numerical examples.