The Application Research Of Artificial Neural Networks In Open Pit Optimization Of Blasting Parameters

Artificial neural network of artificial intelligence as an important branch has been widelyused in the mining of the research of the problem, to artificial neural network (Ann) in open pitmine blasting of research is one of the most important one of the branch.The existing neuralnetwork in open pit mine blasting in most studies focus on the study of blasting,this paper ismainly based on artificial neural network research,make its more conducive to the mineblasting.This paper takes kinds of blasting parameters as the research object,in order to achieve theaccurate study of open-pit mine blasting block size for the purpose of.Presents a method ofblasting parameters based on BP neural network.This paper use MATLAB7.0software toolboxthrough to normalize the data,and pass on the network function,training function,performancefunction,the number of neuron is trained contrast,choose good training effect to create thesystem optimization of blasting parameters based on BP neural network,and based on the BPneural network optimization,to achieve the precise optimization of blasting fragmentation.The main research contents:(1) Collecting, sorting and blasting parameters optimization of the characteristic data,onthe analysis of the blasting parameters optimization of both at home and abroad based on thepresent research.(2) Based on the study of the original blasting effect prediction method,by using artificialneural network to the horse home tower open-pit coal mine rock blasting a relevant analysis oftest data,find the piece of degree distribution and the inherent law of blasting parameters,makefull use of the artificial neural network has highly parallel ability,good fault tolerance,highlynonlinear,strong self ability to adapt and self-learning ability,MATLAB language orientedblasting parameters is established the BP neural network prediction model,prediction ofblasting parameters in reverse. (3) Establish standard neural network performance is not good, not good reverseprojections for blasting parameters. Therefore, we are in the standard BP neural network basedon the optimization. Commonly used optimization method is additional momentum methodand adaptive rate method. We choose to increase the number of neurons method and particleswarm optimization (pso) algorithm with BP algorithm of standard joint optimization methodto optimize the BP neural network.(4) The optimized network model for the original blasting fragmentation prediction modelis improved,and the ideal results by the optimized network model are deduced correspondingblasting parameters,overcomes the original model through adjusting the blasting parameters isneeded to predict the low efficiency of ideal blasting effect.By using the improved back BP neural network prediction model of reasonable field datacollected from training samples,upon examination,the model precision accords with arequirement completely. For blasting parameters of artificial neural network BP neural networkprediction model for research,to solve the problem of the unreasonable blasting parameters ofpractical significance and practical guiding significance to the production of the mine later.