| Avoiding the use of mathematic models,the Fuzzy Logical Control (FLC) system meets the control demands of the tunnel ventilation process well which is a nonlinear distribution system. But because of its lacking of learning and adaptive ability,many problems has emerged when using FLC:the membership functions of the fuzzy variables cannot be changed,the fuzzy logical rules cannot be modified automatically when environmental variables such as traffic model,average exhaust,etc,are changed. For this reason,the paper uses Fuzzy Neural Network Control (FNNC) system to improve the control process.From the view of engineering,the characteristic that the best conclusion of one rule does not influence that of others is confirmed by simulation program,which is called the best rule sorting independence. On the basis of this,logical rules are modified one by one and the optimized rules are gained. It takes much less time to regulate rules by using the above method than the traditional one. In addition,membership functions are adjusted by using fuzzy neural network based on Back Propagation (BP) algorithm,so they match for the rules better and a more reasonable fuzzy combination is got. The course continues until the electrical power that jet fans consume is essentially steady,that is to say,the updated fuzzy combination will not lead to the change of the energy consumption any longer. Then the final optimized fuzzy combination based on any specific traffic model is gained. Finally,the feasibility of the control model is tested by using different traffic models and by decreasing the average exhaust,etc. the simulation result reveals the pollutant concentration fluctuates near the control goal as closely as possible.The FNNC system can modify the logical rules and regulate the parameters of membership functions automatically. Compared with the FLC system,the control process becomes more automatic,the time and the energy consumption becomes less.
|
PDF Full Text Request