Study On The Application Of Fuzzy Neural Network With Compensation In The Evaluation Of Serious Degree Of Underground Gas Pipeline Failures Consequence

Beaucse underground gas pipeline and its subsidiary facilities are a complex system, it exits the dangers of explosion fire tapping of toxicity casualties and facilities destroyed when conveying gas, so it is a frequent phenomenon that the disastrous accident happen accompanied by the conveying or use of gas So, it is essential for effectively evaluating the serious degree of pipeline failures consequence and combine it with the probability of pipeline failures to improve precautionary measure; to provide reliable foundation for scientific decision-making of pipeline risk manegement; to reasonablely make use of limited resources of manpower physical-power and finacial-power etc; to adopt the most advisable measures for minishing the loss effectively; to improve economic performance protect enviroment , and so on, and eventually hit the target of reducing loss of accident The topic of the essay is "Study on the Application of Fuzzy Neural Network with Compensation in the Evaluation of Serious Degree of Underground Gas Pipeline Failures Consequence", it introduce Fuzzy Neural Network , which is new by technology into the evaluation of serious degree of underground gas pipeline failures consequence, intended to establish an evaluating system which better reflect the real performance and effectively predict the risk situation of pipeline, and then provide managing foundation for manager of pipeline The sum total of study in the essay:(1 ) Through the analytical study of existing evaluating technique about consequence of gas pipeline failures both here and abroad, analysis the existed limitations of them, and based on the properties of fuzzy neural network, bring forth the method of evaluating the serious degree of pipeline failures consequence based on fuzzy neural network (2) Analysis the dynamical properties of pipeline when conveying gas, bring forward simulation of the most possible accidents with hurt model, thereby figure out the casualties and direct economic losses due to accident.,(3 ) Collect existing historical data about failures consequence of gas pipeline, find out and classify influnce factors of the serious degree of pipeline failures consequence, and transform them to fuzzy language which can be used to training network with fuzzed technique Make every effort to use the original and reliable data to the utmost extent (4) Beause of the limitation of original data record, expand the original data with ambiguous information diffused technique, just for obtain more useful taining datas . In the procedure, the influence of influncing factors to consequence evaluation rests with factor weight By means of Hierarchial Analysis can obtain weight of determining affecting factors ( 5 ) Train network and verify its function with practical case, for the purpose of examining the reliability and accuracy of network(6) Carry on influencing experiment about tapping of toxicity, estimate the mechanism of diffusion density vary with time fllowing the leak of gas with Gaussian Mistry Rain Model, find out every damage aera of toxicity (aera of poisoned death...