Research And Application Of Blasting Comprehensive Effect Prediction In GUANBAOSHAN Open Pit Mine

In mining,blasting is a complex project which is affected by many factors,and its effect will directly affect the production efficiency and safety of the subsequent process.In recent years,with the increasing demand of ore resources in China,the existing blasting prediction technology is difficult to meet the demand of blasting production.Therefore,starting from data mining,fuzzy comprehensive evaluation and neural network,combined with the historical data related to blasting in GUANBAOSHAN open pit mine,through the learning method of artificial neural network,this paper establishes a blasting effect prediction model,which provides a feasible research scheme for blasting prediction in the mine,so as to guide the safe,efficient and economic production of the mine.It mainly includes the following research contents:(1)Collect and measure relevant data of blasting evaluation indexes.In the field blasting of GUANBAOSHAN open pit mine,the blasting vibration,Boulder rate,flying stone distance and other blasting parameters are obtained by using the blasting vibration instrument,the Split_desktop blasting image processing software and the motion analysis software.(2)Based on the algorithm of feature importance selection of random forest,the main factors affecting the blasting effect are selected as the input parameters of neural network prediction model,and the input parameters of the model are preprocessed with missing value,abnormal value,normalization and other data.(3)Build the blasting effect evaluation set based on fuzzy comprehensive evaluation.First of all,select the blasting effect evaluation index by theoretical research and expert consultation,and determine the membership function of blasting evaluation index;Second,establish the judgment matrix of blasting rating index weight by AHP,and determine the weight of blasting evaluation index;Finally,evaluate the blasting effect by single factor using fuzzy comprehensive evaluation method,and use fuzzy transformation meter According to the maximum membership principle,the comprehensive evaluation result of blasting effect is obtained,which is used as the input label of neural network prediction model.(4)Establish the blasting effect prediction model based on RBF neural network.By acquiring the input parameters and labels of the network,the BP and RBF neural network models are established to predict the blasting effect,and the neural network models suitable for the blasting effect prediction are selected through comparative analysis of the models.(5)Develop the blasting effect prediction system of GUANBAOSHAN.Based on the study of blasting effect prediction model,the system framework is built by C#language,the system interface is designed by WinForm window,and the blasting database management system is built,and the blasting effect prediction system of GUANBAOSHAN is established by combining the blasting effect prediction model.According to the relevant information of GUANBAOSHAN blasting,the blasting effect is predicted and compared with the field blasting to verify the applicability and reliability of the system.