Quantitative Identification Of Rock Strength Based On Drilling Parameters Of Anchor Hole

Improving the efficiency of roadway excavation and reducing roof fall are important for realizing safe,efficient and intelligent mining,and the real-time perception of roof rock strength and structure is its key.It is an effective measure to determine the roof rock strength and structure based on drilling parameters during the construction of anchor hole.Therefore,sensitive drilling parameters are selected based on the analysis of drilling force in this paper.Different sorts of simulated rock samples are drilled to obtain the drilling parameters by the micro drilling experimental platform that is built by oneself.and the noisy parameters are denoised by wavelet threshold method.Finally,the model of MLR-SVM residual correction is established to realize the prediction of rock strength.The main contents of this study are as follows:(1)Selecting the drilling parameters based on the analysis of cutting force.Through analyzing cutting teeth force,the balance equation of force between cutting teeth and rock is established in rock breaking process,and the internal relationship between rock strength and drilling parameters is revealed.The parameters selected in this paper are drilling thrust,rotary torque,rotary speed and drilling speed.(2)Building the micro drilling experiment platform.The micro drilling experiment platform which includes rotary mechanism,propulsion mechanism,drilling control mechanism and data acquisition mechanism was constructed independently.The simulated rock samples with different strength and stratification were made,the uniaxial compressive strength of the samples was measured,and the sample drilling experiment was carried out to obtain the drilling parameters.(3)Denoising drilling parameters by wavelet threshold method.The denoising parameters(wavelet function and decomposition level)which are suitable for different drilling parameters are selected by experiment.The drilling parameters are denoised based on wavelet threshold method.The denoising effect is evaluated by SNR,RMSE and BISA,and the coupling relationship between drilling parameters is analyzed.(4)Constructing the model of MLR-SVM residual correction.This paper analyzes the principle of multiple linear regression(MLR)and support vector machine(SVM),constructs the model of MLR-SVM residual correction to predict rock strength,and compares the precision of prediction among ordinary linear regression,stepwise linear regression and the model of MLR-SVM residual correction.The results show that:(1)the strength of samples is positively correlated with drilling thrust and rotary torque,and drilling thrust is more sensitive to the strength of samples than rotary torque,while the influence of drilling speed and rotary speed on drilling thrust and rotary torque can be ignored.(2)The higher the sample strength is,the more stable the drilling process is and the smaller the fluctuation of parameters is.The drilling thrust and rotary torque increase slowly with the increase of drilling depth,but the total increment is small.(3)The results show that the curves of drilling thrust and rotary torque at the interface between two layers of rock are not completely abrupt,instead of going through a rapid rising process in a short time.It is easier to identify rock strength and interface between two layers of rock when the strength of samples is more different.(4)Taking the drilling parameters of samples for M20,M25 and M30 into ordinary linear regression model,stepwise regression model and the model of MLR-SVM residual correction,it is found that the precision of prediction of the model of MLR-SVM residual correction is 1.11%,3.53% and 8.13%respectively,and the effect of prediction is better.There are 53 figures,31 tables and 73 references in this paper.