| With the country’s strong support for clean energy,photovoltaic power generation technology has developed rapidly,and DC（direct-current）power distribution has also become popular.In the DC power system,when the electronic components in the system have problems such as insulation aging,poor contact,device failure,etc.,it is easy to generate a DC fault arc in the power system.Because the fault arc is random and unstable,when burning instantaneously,it can rise extremely high temperature,which can cause damage to some electronic devices,or cause fire in severe cases,endangering the safety of the power grid.When a DC series fault arc occurs in the circuit,there will be no overcurrent phenomenon in the circuit;the circuit current will decrease;the protection device cannot detect and identify the fault arc,and cannot play an effective protective role,so the DC series fault arc becomes the cause of a major hidden danger of electrical fire in the DC power system.Under this circumstance,the research of the DC series fault arc has broad potential of development.In this thesis,the DC series fault arc under different loads is regarded as the research object,and the DC series arc fault generator is fabricated according to the UL1699-2018 AFCI standard and an experimental platform is built.The DC series fault arc signal is collected through experiments,and the collected signal is decomposed into multiple intrinsic mode functions（IMF）using the complete ensemble empirical mode decomposition of adaptive noise（CEEMDAN）,and the correlation coefficient between the IMF component and the original signal is compared to determine that the target signal processed is the voltage signal of the fault arc.The IMF component is denoised and its arrangement entropy is extracted to form an eigenvector.Finally,the sparrow search algorithm is used to optimize the extreme learning machine to detect and identify the DC series fault arc.Through multiple sets of experimental tests and model comparisons,it is concluded that the sparrow search algorithm optimized extreme learning machine fault arc identification model can accurately and quickly identify DC fault arcs,which is suitable for the detection and identification of DC series fault arcs in DC systems. |
