Application Research Of Genetic Algorithm And Artificial Neural Networks In The Prediction Of Mine Water Gushing-out

Mine Water Gushing-out as a mine safety production of natural disasters, It often led to a serious loss of coal mine production safety, Restricting the development of coal industry and economic efficiency .So, exploring the mechanism of water inrush, predicting for practical theories and methods of production of coal mine safety has always been an important element.Mine water gushing-out is an extremely complex nonlinear dynamic system, we can not completely accurate mathematical language to describe, using traditional methods and techniques is difficult to reveal the internal pattern currently. BP neural network can solve nonlinear problems in the simulation and structural issues and shows the superiority of the larger, you can directly use of the network input and output variables for network training to achieve the projected goal. So, a new optimal algorithm of BP Neural Network based on genetic algorithm was proposed to form a mine water gushing-out prediction mode, and researched this idea. The main contents are as follows:(1) We studied the current methods of prediction, and compared the advantages or drawbacks among those methods. Then we chose the method of neural network prediction for the main study which based on mine water gushing-out mechanism and influencing factors.(2) This paper focuses on genetic algorithm and BP neural network combined forecasting model. First, we introduced to the BP neural network and genetic algorithm theory, and analyzed the Advantages and disadvantages of the two algorithms. We proposed that using genetic algorithm improved BP neural network program, through the neural network structure, connection weights and learning the rules of optimization. Then, Established a genetic algorithm and BP neural network combined prediction model. In this paper, the BP neural network model and genetic algorithm combined with the BP algorithm for network models were used to predict mine water gushing-out, and used the MATLAB as the experimental platform. With the help of the measured data as training samples for learning and prediction, we analyzed different simulation results of graphics and data. The results showed that the combination of two algorithms for the prediction of Mine Water Gushing-out were fault-tolerant, faster convergence and more accurate than the traditional method. Finally, through those experiment proved that combination of two algorithms is reasonable, and it is feasible and effective to apply this algorithm to predict mine water gushing-out.(3) At last, this paper designed mine water gushing-out prediction system based on prediction model of this paper, and completed the system overall architecture. By B/S framework, we completed user-based water gushing-out prediction system and prediction function.