| Rock masses are composed of rock blocks and structural planes. The deformation and stability of rock mass is mainly controlled by structural plane's characters of deformation and strength. It is hard for us to get shear strength parameters of the structural planes exactly, which are important for rock mass designing and calculating. The inaccurate parameter is becoming the bottleneck problem of the theoretical analysis and numerical modeling for rock mess. Therefore, it is significant to research on determination methods of shear strength parameters of structural planes.Considering the uncertainty of the structural planes'parameters and combined indoor shearing tests, we put forward methods and suggestions on optimizing the parameters from three ways, which are processing methods for test data, analyzing domino effect of sample preparation errors and correcting the structural planes'strength parameters, producing a new structural plane direct shear test apparatus. In this way, the influence to the test result may sharply be decreased from the whole process.Least square method is commonly used for dealing with direct shear test data of the structural plane. When using least square method, some disadvantage appears due to the obvious theory limitations. Enlightened by situ-testing, point cluster center method, random-fuzziness method, reliability analysis method are used to manage the data in order to choose the right method due to the test data condition, test sample condition, engineering parameters requirements.(1) Theoretical source of least square method is the theory of minimum squares which only consider the randomness of the test samples. It is suitable for solving the good linear, less abnormal points contained, large amount data. Even other method fails, it is still a available way for reference.(2) Point cluster center method belongs to the methods of roughly determining the shear strength of structural plane which is similar to least square method. The parameters can be determined by solving the distance of the points to the line. The prominent feature of the method is easy operating, hand-painted, intuitive processing. It is suitable for solving good linear data and coarse determination of shear strength.(3) The property of structure plane is the start point of random-fuzziness method. The best estimation is determined by iterative operation which chooses "Membership Maximum Principle" as the prerequisite. The comparison between random-fuzziness method and least square method shows the former is much more advanced, though it is limited by the difficulty of choosing the right membership function, complex calculation processes; convergence problems and so on. It is suitable to deal with the space variable quantity or generic uncertain quantity.(4) Reliability analysis method takes the distribution characteristics of the shear strength parameters into account. The shear strength parameters can only be determined by given the assurance rate that we need. But it is difficult to set the initial value, because the result may changes with it. This method which must obey the normal distribution assumption also has the convergence problems and the complex calculation processes. It is suitable for dealing with the data which is given the assurance rate and contains less abnormal points, the parameters meeting normal distribution and good correlativity or the test data is discrete.By quantitative calculation and comparison between the four methods, we can get the conclusions as follows:Firstly, when the test data is high linear, the calculated result by four methods is similar with each other; we can choose the smallest as the suggesting parameter for conservative consideration. And then, if the test data is enough but discrete, we should take the other three methods instead of to deal with the point cluster center method with the test data. The suggesting parameter is offered by comparing the different three results. Thirdly, if we meet with the test data which is not only scared but also discrete, the calculated results of least square method and point cluster center method are not satisfactory. We can mainly depend on the results calculated by random-fuzzy method and reliability analysis method.Sample preparation errors is the second main point discussed in this paper which is introduced in three aspects, the great effects on shear strength caused by sample preparation errors, two methods for correcting parameters and the confirmation by numerical simulation test. Conclusions can be summarized as follows:(1) The sample preparation errors are controlled by dip angleα, drift angleβwhich can be determined by shear direction and sample error angle. And the formula about correcting normal stress and shear stress is derived by that. The comparison about the influence on shear strength between dip angleα, drift angleβshows dip angle a plays much more important place than drift angleβdoes which can even be ignored. Then we can get the simplified formula about correcting normal stress and shear stress.(2) We find dip angleαobeys normal distribution by statistic to the sample preparation errors under the normal test condition. The dip angleαis mainly near 0±5°.(3) The structure plane's modifier formula of 'climbing direction' and 'downward direction' can be solved by putting the correcting normal stress and shear stress formula into Mohr-Coulomb formula. If given the mean angle of sample preparation errors, the result can be directly set right by this formula. Though this is a rough method, it gives easy way for making the parameter corrected.(4) We simulate the direct shear test by FLAC3D numerical simulation software. It is certified correctly that the method used for simulating is absolutely right. Besides it is simulated under different conditions when drift angleβis 0°, dip angleαchanges fromα=±1°toα=±5°. The comparison shows the cohesion and friction angle of the joints increase or decrease with the dip angleα. The results calculated by numerical simulation and theory formula are extremely similar, which verifies the correctness of the theory formula. Finally, we choose the condition as en example when the drift angleβis 5°and the dip angleαis 0°.It shows the dip angle a is critical factor which controls the impress on the shear strength of the joints.In this paper, we introduced a new apparatus for structure plane direct shear test which can be used indoor or outside. It offers easy, fast, convenience and veracious way for determining the parameters by using this apparatus. This apparatus can be divided into two parts:the box for making sample and shear testing, the shear equipment. The structure of the box is simple and easy for using. It can merge the process of making the sample and shearing in order to eliminate the influence caused by the sample preparation errors. It can extremely decrease impact to structure plane caused by the personal factor. The extra moment of couple and distortion of stress result from using the traditional apparatus. The invalid shear displacement is cleared up by limit displacement device and position restoration device. The improved direct shear system CAT can record the normal stress, the shear stress, the normal displacement and the shear displacement at the same time. It concentrates signal acquisition, signal processing and signal output. The apparatus is very suitable for fast outside direct shear test, because it is less heavy than traditional machines and it is easy for testing and taking with the regularly shape.The general steps of optimal solution method are proposed by taking the joints shear strength determination in Chenjia dam as an example. Firstly, we should do the shear test in standard, correct way. Then the test data should be modified due to the structure plane's modifier formula. Ultimately, the parameter can be determined by comparing the results calculated by different methods.
|
PDF Full Text Request