| Rock mass blasting is one of the most effective technologies for rock mass excavation in tunnel engineering. The numerical modeling of tunnel blasting serves as a link between the theoretical developments and the engineering practice. It simulates the dynamic blasting of the rock mass by employing mathematical physical models approximating the real situation, with a view to revealing the actual transient blasting process and the subsequent effect, and obtaining a better understanding of the rock mass blasting and failure mechanism. At the same time, the credibility of the numerical simulation can be testified by in-situ large-scale experiments. A process consisting of numerical simulation, physical experiment and theoretical investigation has been involved in various mechanical and engineering fields for a systematic and integrated analysis.The dynamic Finite element analysis commercial package ANSYS/LS-DYNA is employed to assist the analysis carried out in this project. The numerical modal for the blasting of the jointed rock mass is efficiently constructed using the ANSYS specified parametric language. The Lagrange algorithm is adopted in the analysis. The unbounded rock mass is modeled by non-reflecting artificial boundary conditions. Relevant materials in this study include rock, joints, air and explosive, with corresponding material types in the numerical model being concrete, rate-independent bilinear kinematic hardening model, high explosive and null-material modals, respectively. Six selected sections of the Gaoling tunnel are examined in this study to explore the failure characteristics and the overbreak/underbreak mechanisms of the jointed rock mass blasting. The DOE technique is adopted to investigate the influence of the geological properties and the blasting parameters on the jointed rock mass blasting. Satisfactory blasting phenomenon is achieved by adjusting the designing blasting parameters according to optimization method.The main conclusions are summarized as follows:1.The influence of joints to the blasting phenomenon is intuitively observed by examining the Six typical sections, i.e., RK343+665, RK343+667, RK343+669, RK343+698, RK343+719 and RK343+735 of the Gaoling tunnel, that is, the closed joints enforce negligible influence on the blasting, whereas the influence from the filled joints mainly depends upon the spacial arrangement of the blastholes:(1)With blastholes being positioned inside the joints, the joint strength decreases dramatically after blasting, secondary stretching faults are resulted from the stress wave reflection between the rocks at the two sides of the joint, leading to severe damages, even to the surrounding rock out of the designing scope.(2)When blastholes are within the vicinity of the joints, the blasting energy can be partially absorbed by the joints, with a correspondingly partial reflection, which may result in massive overbreak.(3)When a couple of or more than two sets of joints are closely located or intersect, a back-and-forth reflection of the stress wave will be created, yielding stretching failure of the rocks in between the joints. This, however, may lead to underbreak if the rocks, on the other side of the joints, are protected without any blasthole arranged.( 4) It is therefore suggested for rock mass blasting with filled joints that Large-diameter blastholes should be designed with little or no explosive. No explosive should be placed inside the joints.2. The characteristics of the joints, i.e., the orientation, width, and the spacing, have different influence on the jointed rock mass blasting, varying from negligible effect to significant overbreak/underbreak, which may unnecessarily increase the blasting cost. Based on the ISIGHT platform, Optimal Latin Hypercube Method is employed in the smooth blasting design of the jointed rock mass. 43 sets of simulations are carried out with the input variables being joint- blasthole center angle, joint width, joint spacing, blasthole spacing, the Intensive factor of peripheral holes and number of explosive, and the output variables being the amount of overbreak or underbreak.(1)Mutil-factor least squared method is utilized to obtain the following conclusions: the weighting ratio of each factor influencing the overbreak/underbreak of the smooth blasting of the joined rock mass is localized within 10%. The first five factors are: the linear affect of the joint- blasthole center angle to the overbreak which takes up to 9.8%; the linear primary effect of the explosive and the interact effect of the joint width JW1 and the joint spacing JD both take up to 7.3%;the second primary effect of Jw2 takes up 6.8% and the interact affect the joint width Jw2 and the joint spacing Jd takes up to 6%.(2)RBF neutral network is adopted in the multi-dimensional spacial approximation of the data constructive modal. a non-linear interact phenomenon is observed with relative complexion.3.Experimental optimization design provides valuable information for acquiring more appropriate blasting parameters. An optimal software platform ISIGHT is used to perform the optimized numerical modeling of the joint rock mass blasting. The optimal blasthole spacing and the Intensive factor of peripheral holes are obtained by inspecting two objective values: the volume of the overbreak/underbreak and the effectiveness of the explosive energy. More variables and objective values can be taken into consideration in this analysis with more practical significance.The main innovation spots in this paper:1.The complex jointed rock mass blasting model is constructed in this study, with a satisfactory simulation of the overbreak and the underbreak of the tunnel blasting. A suggestion that no explosive should be placed inside the filled jointed is provided. Discussions are presented in terms of the energy point of view with empty blasthole.2.The interaction effect between the geological information and blasting parameters are taken into account. The least squared multi-factor formulations are presented. Radial based neutral network is used for the construction of spacial date approximation, which ensures the visualization of the multi-factor interaction effect in multi-dimensions.
|
PDF Full Text Request