Study On The Application Of Road Tunnel Longitudinal Ventilation Neural-Fuzzy Control System

The road tunnel ventilation system has strong non-linear characteristic and it is difficult to gain the precise mathematical model by using the traditional linear control theory, therefore, the modern control methods such as the fuzzy control become the trend for the road tunnel ventilation control. However, there are some difficulties with the establishment of the fuzzy membership functions and the rule base. The paper involves the research of the application of neural-fuzzy control method to the road tunnel longitudinal ventilation system: by learning the sample data, the neural network automatically produces the membership functions and the control rules, subsequently, the fuzzy control model is easily established; the distribution of pollutant and the fan operation is simulated under the varies traffic flow circumstances, the simulation results show that the longitudinal jet ventilation in medium-length and long road tunnels can be effectively controlled with the application of above method.The paper firstly combined the aerodynamic model, pollutant model, traffic model and the optimal operating fans selection model and then established the sample data gaining program to produce the optimal sample data; secondly, the multi-layer forward neural network and the fuzzy control is combined within an equivalent structure, after the learning of sample data, the neural-fuzzy control system for the road tunnel longitudinal ventilation is produced; thirdly, the optimal operating fans selection model is replaced by the fuzzy control model to make the tunnel ventilation simulation program, and then the control results under different traffic flow circumstances are evaluated in the MATLAB environment; finally, the adaptive ability of the neural-fuzzy control system is discussed.The above system realized the autogeneration of the membership functions and the control rules, greatly lessened the design period of the ventilation control system, and make it possible to on-line modify the control rules. The system has good control effects: the pollutant is controlled under the hygienic standard; the energy consumption is less than that of the traditional fuzzy control by 19 to 32 percent. The system has good adaptive ability: its control is still effective when the traffic flow and the basic vehicle exhaust changes and when some special traffic flows emerge. The design codes for the road tunnel longitudinal ventilation control system form a standardized and practical design flow and platform.The neural-fuzzy control system for the road tunnel longitudinal ventilation has great feasibility and practical value, and it has important implications for the current road tunnel ventilation control design.