Research On The Optimization Algorithm Of Residual Current Protection For Low-voltage Distribution Network

Residual current device(RCDs),a type of protective equipment in low-voltage systems,are widely used to prevent and avoid leakage accident of power grid and protect the safety of life and property.At present,the operation of an RCD is based on sensing the root mean square value of residual current in an electrical circuit.The circuit will be interrupted on the action of a relay when the residual current exceeds a predetermined level(30 mA for human safety),known as the tripping current.Although such devices offer a large degree of protection,they are prone to nuisance tripping or maloperation in the actual operation of the grid due to the lack of the ability to diagnose electric shock type and identify touch current,which reduces the reliability and the rate of proper commissioning for RCDs.To resolve the above problems,the research on the optimization algorithm of residual current protection for low-voltage distribution network is developed under the special support of State Grid Corporation’s Scientific and Technological Project:"Research on the methods of rapid extraction and identification of transient information of the electric shock".In this dissertation,the original data and waveforms of the electric shock signals of the organisms in different typical scenes are obtained through the establishment of an electric shock experiment platform.On this basis,the time-frequency characteristics of the residual transient current are analysed,and then,the multi-source information fusion feature of the residual transient current component are extracted,Finally,the detection of the time of electric shock,the diagnosis of the type of electric shock and the identification of the touch currrent are determined accurately.The above resesarch lay a solid theoretical and technical foundation for the replacement of the residual current protection principle in China.The main contents of this work are summarized as follows:(1)The time-frequency characteristics of the residual current in the event of electric shock are analyzed by using the pseudo-Wigner-Ville distribution.The characteristics of the spectrum of the residual transient current with time are obtained which reveals the spectral distribution pattern of the residual current in the electric shock.(2)The multi-dimensional features of the residual current in the time-frequency domain are extracted by using the statistics,the singular value decomposition and the local mean decomposition,and the spatial distribution of these multi-dimensional feature information is explored.Furthermore,the multi-source information fusion feature index of the residual transient current is constructed using the principal component analysis,and the residual current fault information is maximized by using the minimum feature.(3)By using the characteristics of the skewness of the noise and sinusoidal signal,the increments of absolute value of skewness for the residual current are utilized as a sign of the presence of an electric fault.Therefore,a novel method for detecting the electric shock fault time based on the fundamentals of skewness is proposed,and this method provides a new way to realize high precision electric shock fault time detection.(4)The particle swarm optimization(PSO)algorithm is used to optimize the penalty factor and kernel parameter of support vector machine(SVM)and the multi-source information fusion features are used as feature values for fault classification,an identification model of electric shock fault type based on parameter-optimized support vector machine is established.Fault diagnosis examples show that three kinds of fault types such as the direct electric shock for plants,the direct or indirect electric shock for animals can be identified effectively by multiple features support vector machines model,and both accuracy and generalization are better than single characteristic support vector machines method.(5)A novel model for identifying and calculating the touch current component is proposed,which is based on least squares support vector machine.The technical problem in the detection and extraction of the touch current component from the residual current is resolved using this model.The experimental results show that the model has the advantages of fast training speed and low detection error,which can overcome the defects of neural network such as long training time and easy to fall into the local minimum.