Research On Parameter Selection And Effect Prediction Model Of Tunnel Blasting

With the continuous consumption of shallow mineral resources,mining is gradually developing to the deep,and the quantities of shaft and roadway construction are also greatly increased.However,the current roadway excavation engineering design is faced with the problems of lack of standards for the selection of blasting parameters,numerous surrounding rock classification methods,different geological evaluation,difficult selection of blasting parameters,poor blasting effect and so on.Therefore,it is possible to compare the engineering rock mass classification methods,It is of great significance for roadway excavation blasting engineering to convert them into one,establish a reasonable,convenient and efficient blasting parameter selection model based on this classification method,and design and develop relevant auxiliary systems.Using the methods of theoretical analysis,mathematical statistics and the combination of software engineering and roadway engineering,this paper makes a comprehensive study on the parameter selection and effect prediction of smooth blasting engineering.The main conclusions are as follows:The influencing factors of smooth blasting design are analyzed.It is considered that geological conditions are the absolute factors controlling roadway blasting design.Aiming at the problems of many surrounding rock classification methods and different evaluation of roadway blasting engineering geological conditions,this paper introduces the common surrounding rock classification methods at home and abroad,analyzes the equivalent relationship between the common classification methods and BQ rock mass classification,and transforms the surrounding rock classification of common rock mass classification methods into BQ rock mass classification through these equivalent relationships,It provides a certain reference for tunnel excavation blasting designers to face different surrounding rock classification methods in blasting design.The parameter selection model of roadway smooth blasting based on random forest is constructed.The prediction results of the verification set show that there is little difference between the actual value of the sample and the predicted value,and the model has better effect on the prediction of minimum resistance line and blast hole depth than blast hole spacing and charge density coefficient.It is proved that the random forest model is feasible to select the parameters of roadway smooth blasting,Moreover,the values of minimum resistance line and blast hole depth have more reference value for designers.The prediction model of smooth blasting effect based on support vector machine regression is constructed.Genetic algorithm,particle swarm optimization algorithm and simulated annealing algorithm are used to optimize and train 36 groups of roadway blasting samples,and 4 groups of roadway blasting samples are used to predict.The results show that the optimization effects of the three optimization algorithms are similar,and the optimization effect of genetic algorithm is good.The relative errors between the measured and predicted values of half eye rate and over and under excavation of four groups of smooth blasting samples are less than 7%.The model has good accuracy and meets the needs of practical engineering.The auxiliary system of smooth blasting design is designed and developed.The system can simply realize the functions of surrounding rock classification,smooth blasting parameter selection reference and smooth blasting effect prediction,and can meet the reference needs of designers in blasting design in engineering practice.