Fault Diagnosis And Life Prediction Of Gas Equipment

With the rapid development of the natural gas industry in China,the distribution of the national natural gas industry has gradually expanded.After the 13 th Five-Year Plan,the transformation of the energy industry began,such as coal to gas and so on.As a result,the demand for natural gas will increase substantially.Therefore,how to ensure the stable transmission and distribution of gas in the future is an important research direction.Gas regulator as the most important existence in the transmission and distribution system,there can be no serious accident.However,at this stage,the operation and maintenance of gas regulator is relatively simple,usually using the planning maintenance strategy,which will not only lead to a lot of waste of manpower and material resources,but also lead to a waste of resources;At the same time,the fault diagnosis and life prediction technology of the voltage regulator is relatively backward,a set of matching life evaluation system has not been built,and the strategy of replacing the voltage regulator after reaching the service value is adopted.Therefore,based on these two situations,this paper proposes a probabilistic neural network approach to fault diagnosis and simulation of gas regulator,and establishes the optimal life evaluation model based on historical failure rate.The contents of this paper are as follows:1.Firstly,the common gas regulator and its fault state are summarized,and then the probabilistic neural network is proposed to establish the gas regulator model.The advantage of the fault diagnosis method based on this network is that the network is easy to train,and the convergence speed of the network is fast,which is suitable for real-time data processing.At the same time,it also has strong fault tolerance and expansibility.With the increase of fault data and fault mode,the diagnostic effect of the trained network will be more accurate.This can greatly reduce the manpower to judge the fault time,reduce the uncertainty caused by manual operation.Using the model distance between the test sample and the learning sample,normalized the initial probability matrix of Gao Si function,and then calculated the probability of the test sample and the learning sample.In the final simulation experiment,we use the actual fault data to train the network,and compare with the previous research results,verify the feasibility of the probabilistic neural network and draw a conclusion.2.Establish the best life evaluation model based on the statistical failure rate model,the failure rate model based on the failure rate model and the service life back and back mode after the equipment is overhauled.When the same type of equipment works for a long time in the current working environment,the failure times in recent years are counted,and the failure rate model of the equipment over time is obtained,and the effect of retrograde service after maintenance of the equipment is added during the period.After that,the operation cost,overhaul cost,fault cost and so on are estimated.According to the minimum annual cost of the estimated value,the best life evaluation model of the equipment is established.3.In the last chapter of the article,in order to find a more intelligent method of fault diagnosis and life prediction,this paper explores the laws of nature,and explores the ways of self-organization and self-repair of animal cells in nature and the limited life span of division.The model of cell division was established,the diagnosis and prediction of cell life were analyzed,and the model simulation was carried out.In the abstract level,the function of internal structure of cell corresponded to the function of components of gas regulator equipment.The research on life prediction of gas surge equipment by cell division is analyzed in this paper.It is hoped that through the research of biological system,the fault diagnosis and life prediction model suitable for equipment can be found out.