| In order to evaluate the rock cuttability quickly,accurately and effectively,the rock indentation test by a conical pick,drilling test and TBM excavation test were analyzed,and the characteristic parameters of rock breaking by cutting tool were taken as the evaluating index of rock cuttability,which are utilized to evaluate the rock cuttability of mechanized rock breaking.Among them,the rock cuttability analysis based on the rock indentation test mainly studies the influence of different rock properties and uniaxial confining stress on the rock cuttability,where the pick indentation test of granite,marble,red sandstone and phosphate rock were performed under different uniaxial confining stress.The peak indentation force,peak indentation depth,cutting work and specific energy were taken as the characteristic parameters of rock breaking performance of the cutter,and the regression analysis(two-dimensional regression analysis and three-dimensional regression analysis)and intelligent algorithm(generalized neural network GRNN and support vector machine GA-SVM combined with genetic algorithm)were used to investigate the rock cuttability under different uniaxial compressive strength,tensile strength,brittleness index and uniaxial confining pressure stress.Finally,the established prediction models are applied to the mechanized rock breaking prediction of Kaiyang Phosphate Mine in Guizhou Province.In addition,the relationship between the parameters while drilling(thrust,rotational speed,torque,and rate of penetration)and rock cuttability was analyzed through drilling tests.Firstly,the grouting concrete drilling test is carried out at Datengxia hydropower station and the corresponding drilling parameters are obtained.Then,the specific energy was used to evaluate rock cuttability,and the regression model between drilling parameters and specific energy were established through regression analysis,and the specific energy model proposed by Teale was modified accordingly.Meanwhile,the prediction models of specific energy were established by using random forest RF algorithm and GA-SVM.Finally,the uniaxial compressive strength model under different drilling bits was established by combining the specific energy SE and drillability index I_d,and the drilling test results of Datengxia Hydropower station were substituted into the prediction model to verify its accuracy.The excavating parameters(thrust,rotation speed,torque and penetration rate)of EPB TBM,single shield TBM and double shield TBM in different tunnel excavation were collected from published papers to analysis rock cuttabilities.Then,the specific energy is also taken as the evaluation parameter of rock cuttability,and the relationship between excavation parameters and specific energy were analyzed by two-dimensional regression.Meanwhile,the corresponding regression models were established.Since the two-dimensional regression model unable effectively reflect the relationship between the above parameters,the three-dimensional regression models of the penetration rate and specific energy were established.The three-regression model shows that there is a good statistical relationship between torque,penetration rate and specific energy.Therefore,the GA-SVM,RF and GRNN were used to establish specific energy prediction models with torque and penetration rate as input variables and specific energy as output variables.Then,the prediction performance of the above three models was compared,and the results show that GRNN prediction model has the best performance,RF prediction model takes the second place,GA-SVM prediction model is the worst,and can’t meet the engineering requirements.Through the above research,rock cuttability can be comprehensively evaluated before and during mechanical rock breaking,which can provide parameter design guidance and data source for mechanical rock breaking to a certain extent. |