Based Scada System, The Auxiliary Gas Scheduling Neural Network Model

This paper introduces SCADA system and basic conceptions of the nerve network in a systematic way; explains the fundament, the mathematical expression, the algorithm and measure of BP algorithm in detail; presents the design principle, the system structure, the function structure and the communication disposition of the transport and distribution of city gas-SCADA system in a summary fashion. There are some problems in the forecast of the gas transport and distribution, such as experience, uncertainty, irrationality, etc. However, the artificial nerve network simulating human brain movement has many advantages, such as the extremely strong non-line shape approach, the parallel processing, the self-training, self-studying, self-organizing and the fault-tolerant ability. Then it can overcome some defects of the traditional forecasting method and fast accurately get the forecasting result to apply nerve network theory to the gas-dispatching forecast. The main researching contents is to take advantage of BP algorithm and set up the gas forecasting models of the daily and monthly demand, according to the historical data gathered by SCADA system. Moreover, the paper includes how to set up the gas forecasting models of the yearly demand, utilizing BP algorithm and curve combined method. These models have the ability to deal with previous experience intelligently and analyze the running result and error. So these models can effectively forecast gas-demanding quantity in new gas dispatching and also provide reference and plan compare for dispatchers. In additional, it sets up the dynamically simulated model of pipe-network movements to utilize BP algorithm, which can keep watch on the pipe-network movements and realize dynamic imitation. Above-mentioned models find a new way besides the traditional forecasting theory and method, shake off the dependence of personal experience and set up a kind of effective job pattern.For guaranteeing the gas pipe-network running safe, the paper explores the dynamically localization to leak gas in city gas pipe-network by the model classify and the distinguishing function of BP nerve network.