Description OfJoint Geometry Character And Study On The Model Of Joint Geometric Parameters Based On Fractal Theory

As the expansion of human enginering activity and working place, more and more complicated problems of rock mass structure are involved. Therefore, it’s an important assurance for guaranteeing the construction of projects safety to deeply study the characteristic of rock mass structure. For instance, in the process of construction of hydropower station, there are multiple stages to explore the rock mass structure from part to macrocosm. But a few information of rock mass structure is gotten in the prior stage, it’s a great challenge to reveal the geometrical characteristic of rock mass structure precisely and predict the engineering geological problems in foundation surface, underground powerhouse or high slope of dam abutment, and so on, according to fewer data. Monte carlo theory is widely used to simulate the rock mass structure network. There exist certain relationship between the prior random number and the latter one, produced in the process of simulating the rock mass structure based on monte carlo theory. These random numbers are not real ones, so that the rock mass structure network simulated is not strictly stochastic and is difficult to predict the geometrical characteristic of joints in rock mass in the process of excavation.Avoiding the simulation of structural plane network in rock mass, a different way is studied in this paper, which subtly studies the geometrical characteristic and difference of joints before and after excavation of foundation surface based on fractal theory.Fractal theory and differential evolutionary algorithm are used to improve the basic differential evolutionary algorithm.The optimized data(optimized hybrid population)are attained by crossing the prior measured data(basic population data) and the variation random data(variation primary population). It is predicted that the distribution of the geometrical characteristic of joints of rock mass on foundation surface after excavation, by analyzing the mathematical characteristic of the optimized data.It provides a typical research cases for the above-mentioned technical route by collecting geological datain early survey stages and recent participation of foundation surface excavation for the period of the catalogin Dagangshan hydropower construction. Geometric distribution of small sample jointwhich is collected in period of preliminary explorationfootrill is analysised by conventional statistical methods and fractal theory. An improved differential evolution algorithm model is builded to predict the geometric distribution and geometric parameters of large sample joint collected as the foundation is below the excavation. The model is validated is reasonable based on the joint data collected in period of the catalogin Dagangshansurface excavation. Some important conclusions are obtained as following:1) It found that dip direction and dip of the rock mass joint are more to normal distribution based, trace length of rock mass joint is more to log-normal distribution based, and space of rock mass joint is more to negative distribution and log-normal distribution based. By analyzing superiority interval and percentage of dip direction and dip angle, trace length, space of the rock mass joint in the way of stem-and-leaf diagram,there is a few of difference between superiority interval and percentage. So the methods routine statistics is proved to be defected in description the joint geometry characteristics.2) Based on the existing study of calculating the fractal dimension of joint orientation pole distribution, an improved subdivision principle of meshing Schmidt pole diagram is provided from the perspective of geometry, and an optimizing programming design to calculate the fractal dimension D is obtained in this paper. All of the above is one of the innovations of this article.3) Using the fractal dimension of joint orientation pole distribution, distribution characteristics of joint collected before and after the dam foundation excavation is analysised quantitatively. Compared with the traditional analysis method, Fractal dimension D is better to describe the orientationdistribution difference ofjoint collected before and after the dam foundation excavation, which provides a more open thinking for quantitative analysis of joint orientation distribution. At the same time the existing research conclusion is verified effectively:I) Under normal circumstances, the more the joints, the greater the fractal dimension D of joint orientation pole distribution; II) Orientation dispersivity of joint orientation pole distribution is put forward. The greater the dispersion, the bigger the fractal dimension D.The fractal dimension D can be used to describe the dispersion degree of joint pole distribution. The fractal dimension Din theory meet conditions:0≤D< 1.4) Under the enlightenment of Cantor set,distribution characteristics of trace length and spacing of joint is studied and analysised quantitatively. The theoretical values of trace length fractal dimension D and space fractal dimension Dmeet conditions:0≤D<1. Based on analysis of the scarcity, dispersion, non-uniformity of the date about distribution characteristics of trace length and space, the better way for quantitative description of distribution characteristics of trace length and space were proposed. This is one of innovations of this text.5) By divising the theoretical range of fractal dimension D of trace length and space, relationship between the fractal dimension D and qualitative description of trace length and space distribution index (scarcity, dispersion, non-uniformity) is established, which is the one of innovation in this article.6) By Discussing and analyzing the correlation between the fractal dimension D and the number of joints、the data interval and the relative concentration Ⅰ of the preponderance data interval. It found that the fractal dimension D have the highest correlation with the number of joints, second with the data interval, and have the lowest correlation with the relative concentration Ⅰ of the preponderance data interval.It is confirmed the biggest difference between the proposed method of fractal dimension D calculation of trace length and space and the fractal dimension D calculation method by transforming the frequency distribution function of trace length and space.7) Improving the mutation operator of the basic differential evolutionary algorithm based on fractal theory, while the primary population of the basic differential evolution algorithm increased to two:the basic population data and the primary population. The basic population data is the data of adit in the field measured, including the joint orientation, trace length and space data of rock mass.The primary population is the uniform random number which is generated by the random function. The two populations through mutation, crossover and selection operation resulting hybrid population data save both the distribution characteristics of the measured data, and increase the diversity of its data, which can better predict the regional surface after excavation the joint orientation, trace length and space data of rock mass, it is also one of the innovation in this paper.