Diagnostic operation of gas pipelines based on artificial neuron technologies

Abstract

The article deals with issues of diagnosing the functioning of gas pipeline systems based on one of the elements of modern information technology, the principles of artificial neuron networks. Artificial neuron networks (ANN) have been already applying in some areas of science and technology. Intellectual systems based on ANN allow solving a number of problems of pattern recognition, prediction, optimization, diagnostics and control. The technique for diagnosing the operation of gas pipelines proposed in the article based on the principles of ANN. In solving this problem, the principles for a simple pipeline operating on a squared friction mode are used. As a transfer function for the cases of presence and absence of gas leakage, was used the functional dependence of pressure loss. The analyzed changes in the operational parameters of the existing gas pipeline by applying artificial neuron networks are given in the article. The possibility of prompt and accurate determination of minor changes in the regime parameters are shown according to the dynamics of change at the output of the neuron network.