Leak detection based on estimation of distribution algorithm and wavelet transform

Because of the problems that low location positioning accuracy and more number of iterations of traditional artificial intelligence algorithms that applied to detect the gas leakage. This paper proposes a distribution estimation algorithm based on wavelet transform and the method is used to diagnosis of pipeline leakage. Through the wavelet analysis to the signal, singular value information about upstream and downstream signals can be got and be used as the initial population in the method. And probability model of the individual´s distribution in the space is established and sample the model. Gradually, the optimal probability vector within the solution space can be improved. The simulation results show that distribution estimation algorithm was applied to gas pipeline leakage detection positioning has a better position precision and good fault tolerance.