Transformer Internal Fault Detection Based On Symbolic Dynamic

Transformer as the core electrical equipment of the electric power system,and it's safety and reliablity directly or indirectly affect the running of the whole electric power system.With the development of economy,the huge demand for all kinds of energy promotes the expansion of the electric power industry,people pay more and more attention on the protection of transformer equipment.Therefore,It is of great significance to work in the transformer state overhaul technically and economically.State overhaul work is focused on real-time evaluation of transformer internal condition to make accurate runtime,to determine the degree of transformer internal fault,according to the result of condition assessment and the corresponding planning strategy.It not only reduce the maintenance cost,at the same time,it improves the reliability of the system.In this paper,the transformer internal fault detection based on symbolic dynamics,mainly for the target device current parameters,instead of other sensors,it is a kind of economical and reliable method for on-line monitoring.This paper can be divided into two parts,first of all,create the mathematical simulation model of transformer and establish transformer internal fault physical model according to it.Then use symbolic dynamics algorithm to verify the data from test,simulation and real experiments,This article first categorized transformer type and structure,established transformer fault tree analysis model on the basis of it,select easy winding fault as the analysis of the failure of the reference object.Then focus on analysis of transformer internal fault.It demonstrates the nonlinear nature of transformer working environment based on the mathematical simulation model and proves the validity of the symbolic dynamics for testing,and also realize the transformer internal fault simulation according to the improved extension mathematical model.Then,symbolic dynamics of transformer internal fault detection framework is proposed.It introduces common methods and put forward improvement innovative maximum entropy classification for the division of the phase space reconstruction,and then establish the markov state transfer probability matrix to extract the signal characteristic information,contrast pattern characterized by the different signal.This paper uses two kinds of partition construction algorithm for fault signal detection,and analysis the results according to the of experimental.Experiment mainly includes three parts,firstly uses of sine signal screening test set for two kinds of algorithm parameters,preliminary proves the effectiveness of the two kinds of symbolic dynamics algorithm to detect the signal.Then to obtain current signal data through the simulation of transformer internal fault experiment,the results show that two kinds of symbolic dynamics algorithm can quickly and efficiently detect the fault signal.And it provides state overhaul plan strategy for the ransformer on the basis of the better algorithm.Then compare the algorithm to another different one from other scholars systematically.Finally it uses real data collected to verify the symbol dynamics algorithm again,the result accord with the actual situation.This paper proves that the symbolic dynamics algorithm can detect internal fault for transformer effectively step by step.