Research Of Forecast On Temperature Of Bridge Sets Massive Concrete Based On Neural Network

Nowadays, the calculator applies technical development very quickly, involve social a lot of realmses. For the flashing and construction engineering, because of needing a great deal of engineering calculation and the design demand, the calculator is applied very extensively. Currently, the infrastructure of our country is launching completely, the problem of the temperature crack of the massive concrete is very serious, domestic and international experts have the thorough research, and obtain a lot of results. Under the background, this text points out the new standpoint and method.This text according to the example of big bridge construction in Chang Jiang River of Su Tong, elaborate that the possibility that the BP neural network is appled in the massive concrete temperature estimate, use the Matlab neural network tool box to establish temperature forecast model, and use of the VC and Matlab hybrid plait distance, carry on the program design and interfaces design, provide the basis for the construction temperature on aftertime.This text main work is as follows:(1)Introduced the basic principle of the artificial neural network, argument the possibility of the artificial neural network usage of engineering practice, and realize it. (2) Detailed analyse the main cause of the massive concrete temperature change, thinks that the cement turn hot, the hot function, the structure size and temperature for boundary condition, concrete to sprinkle builds and outside temperatures of the concrete is cause the massive concrete temperature change. (3)Carry out the model of the network of BP, the calculate way of the network of BP, training method and the training parameters to realize settle etc. Points out how to use BP network tool box carries on the realization of the network' s model. (4) Resolved how to communicate VC and MATLAB, put forward a few methods, and make use of a method and carry it out, use of the VC vivid plait distance ability and the interface design abilities, put it together with the MATLAB strong science calculation ability.End, this text points out the many shortage of this research, and put forward the suggestion of further research, I hope somebody interesting on it should have the more thorough research, make the result of research near to the engineering actual and have a good engineering guide function.