The Application Of Fuzzy Neural Network On Underground Gas-Pipeline Risk Analysis

The underground gas pipeline is the significant parts of the " life blood project " in city , the underground gas pipeline not only the surface is vast ,the line is long, but also that apting to affect by the environment , different man-made damage and natural calamitys. With the underground gas pipeline quantity increasing and the operating time longth, the underground gas pipeline is living with lots of questions such as pipeline design , element manufacture , the problem which fixed and administration reveals out one by one,but the underground gas pipeline is builded under the sbterranean , the work environment is considerably complex , gives maintenance and protection to bring extremely great hardship once more , resuling in the gas piping mishap more frequently in town. The gas piping once lose the effectiveness will bring the calamity ,affect people's existence and wealth. Hence in order to make sure the underground gas pipeline secure, unearth pipelines-latent capacity, decreases gas pipelines failure to brought ceases gas and the economy decrease, it is very significant to underway the fault tree analysis and risk analysis on underground gas pipeline.On the base of gathering home and abroad datum widely and living up the underground gas pipeline researches, and combining our country gas pipeline actual work situation, This research has been carried on the quality risk analysis research. This research according to collect the underground gas pipeling data in field,use compensation fuzzy neural network to establish mathematical model and calculate and anilyle and obtaine the gas pipeline failure risk.For the sake of more valid analysis and modification the questions that ariseed in assessing, we adopts establishing the mathematical model earlier , afterwards assumes the analog computation in the computer, finally by means of the experiment to certificate , that may thrift more labours power , resource and financial, and avert the test blindly , the test step clearly , the aim is clear. This task merely analyses the underground gas pipeline death stage and carries on risk analysis research. Complete 9 respects research work as the following below.1. Gathering the literature datum , evaluation criterion and operation that is administerd and keeping the minutes and examine in gas pipeline corruption datum , history mishap and manipulate miss, and correlation criterion in searching home and abroad , subject study orientation and make sure the technique course, and comprehend the Guiyang gas pipeline operation situation that was exploted at present, and covers up building technology that etcing present situation and guard arrangement, the corporation administers and so on.2. Carefully analysing underground gas pipeline failure cause completely. The underground gas pipeline failure effect is divided into five types( Corruption effect element and third side effect element, design effect element and manipulate effect element and man as a result of the dependability) 130 base events with covering up . In fault tree analysis hostiry,man as a result of the dependability is the first to join the concluate and analysis.3.This research first choose compensation fuzzy neural network model pattern as calculation and analysis mathematical model, and determinate fault tree compensates fuzzy neural network input element weight. Use progress step analytic approach and CR unanimousle evidence to bring to fix the base event element weight , offering the compensation fuzzy neural network calculation flaut tree fuzzy importance degree that supplyd the foundation of training data structure.4. Structure and calculation fuzzy fault tree . use compensation fuzzy neural network calculate fault tree top event fuzzy effectiveness prc bability and calculation fault tree fuzzy important degree , finds out the main reason that lose effectiveness with underground gas pipeline, and carries on the...