一种基于改进LLE方法的传感器故障诊断方法研究

针对非线性系统传感器故障诊断难以解决的问题,提出了一种新的基于局部嵌入映射(LLE)的方法,解决了非线性数据的特征映射问题。首先,改进了基于分形维估计的内在维数的估计,通过线性拟合解决了线性区域的自动确定。然后,将故障状态与空间分布结合起来,通过确定数据点在空间超球内的分布完成故障的检测,在这个过程中将超球的确定与LLE算法中基于核函数的样本外数据扩展结合起来,大大减少了计算量,提高了算法的实时性,从而为复杂非线性传感器的故障诊断提供了一种新的有效的方法。

投资时间序列的混沌分维研究

以某市固定资产投资同比增长数据为根据 ,利用混沌动力学中的相空间重构技术、幅值谱分析、庞加莱截面等方法进行分析 ,并计算其关联分维 ,得出其饱和嵌入维数为 5、关联分维约为 1 7,初步确定某市固定资产投资增长过程具有混沌特性 ,有 5个因素影响着某市固定资产投资的同比增长幅度 ,其中 2个因素起决定作用·该法可用于复杂的经济系统的分析研究·

动态系统结构与分形公司

传统的静态企业结构不能适应快速变化的环境，企业需要可以快速重构的动态结构，以内部变化来适应外部变化。系统的构成单元应该具有决策智能、自组织、协商合作和自治能力。本文研究了企业动态结构的需求与特点，同时介绍了分形公司的动态结构思想。

分形公司概念及理论

在世纪之交，市场环境动荡变化，全球各国纷纷研究和探索“新的制造模式”，提出了种种新思想，如敏捷制造、仿生制造和分形公司等．本文对分形公司的基本概念、结构和设计方法进行了介绍和探讨。

海拉尔盆地贝尔凹陷基岩潜山特征及储层预测

The lithology of the buried hill of Triassic Budate group in Beier depression is epimetamorphic clastic rock and volcanic clastic rock stratum. Recently the favorable hydrocarbon show was discovered in buried hill of base rock, and large-duty industrial oil stream was obtained in some wells in Beier depression. Based on the information of seismos and wells, the tectonic framework, tectonic deformation times and faulted system of the Beier depression are comprehensively studied, then configuration, evolutional history, genetic type and distributed regularity of buried hill are defined. According to observing description and analysis of core sample, well logging and interpretive result of FMI, the lithological component, diagenetic type and diagenetic sequence of buried hill reservoir are confirmed, then reservoir space system of buried hill is distinguished, and vegetal feature, genetic mechanism and distributed regularity of buried hill fissure are researched, at the same time the quantitative relationship is build up between core fissures and fissures interpreted by FMI. After that fundamental supervisory action of fault is defined to the vegetal degree of fissure, and the fissure beneficial places are forecasted using fractal theory and approach. At last the beneficial areas of Budate group reservoir are forecasted by reservoir appraisal parameters optimization such as multivariate gradually regression analysis et. al. and reservoir comprehensive appraisal method such as weighing analyze and clustering procedure et. al. which can provide foundation for the next exploratory disposition. Such production and knowledge are obtained in this text as those: 1. Four structural layers and two faulting systems are developed, and four structural layers are carved up by three bed succession boundary surfaces which creates three tectonic distortional times homology. Three types of buried hill are divided, they are ancient physiognomy buried hill, epigenetic buried hill, and contemporaneous buried hill. 2. Reservoir space of Budate buried hill is mainly secondary pore space and fissure, which distributes near the unconformity and/or inside buried hill in sections. The buried hill reservoir experienced multi-type and multi-stage diagenetic reconstruction, which led to the original porosity disappeared, and multi secondary porosity was created by dissolution, superficial clastation and cataclasis et. al. in diagenetic stage, which including middle crystal pore, inter crystal pore, moldic pore, inter particle emposieu, corrosion pore space and fissure et. al. which improved distinctly reservoir capability of buried hill. 3. The inner reservoir of buried hill in Beier depression is not stratigraphic bedded construction, but is fissure developing place formed by inner fault and broken lithogenetic belt. The fissures in inner reservoir of buried hill are developed unequally with many fissure types, which mainly are high angle fissure and dictyonal fissures and its developing degree and distribution is chiefly controlled by faulting. 4. The results of reservoir comprehensive evaluate and reservoir predicting indicates that advantageous areas of reservoir of buried hill chiefly distributes in Sudeerte, Beixi and Huoduomoer, which comprehensive evaluate mainly Ⅱand Ⅲ type reservoir. The clues and results of this text have directive significance for understanding the hydrocarbon reservoir condition of buried hill in Beier depression, for studying hydrocarbon accumulated mechanism and distributed regularity, and for guiding oil and gas exploration. The results of this text also can enrich and improve nonmarine hydrocarbon accumulated theory.

基于随机介质理论的多尺度储层非均质性定量描述

Application of long-term exploration for oil and gas shows that the reservoir technology of prediction is one of the most valuable methods. Quantitative analysis of reservoir complexity is also a key technology of reservoir prediction. The current reservoir technologies of prediction are based on the linear assumption of various physical relationships. Therefore, these technologies cannot handle complex reservoirs with thin sands, high heterogeneities in lithological composition and strong varieties in petrophysical properties. Based on the above-mentioned complex reservoir, this paper conducts a series of researches. Both the comprehending and the quantitative analysis of reservoir heterogeneities have been implemented using statistical and non-linear theories of geophysics. At the beginning, the research of random media theories about reservoir heterogeneities was researched in this thesis. One-dimensional (1-D) and two-dimensional (2-D) random medium models were constructed. The autocorrelation lengths of random medium described the mean scale of heterogeneous anomaly in horizontal and deep directions, respectively. The characteristic of random medium models were analyzed. We also studied the corresponding relationship between the reservoir heterogeneities and autocorrelation lengths. Because heterogeneity of reservoir has fractal nature, we described heterogeneity of reservoir by fractal theory based on analyzing of the one-dimensional (1-D) and two-dimensional (2-D) random medium models. We simulated two-dimensional (2-D) random fluctuation medium in different parameters. From the simulated results, we can know that the main features of the two-dimensional (2-D) random medium mode. With autocorrelation lengths becoming larger, scales of heterogeneous geologic bodies in models became bigger. In addition, with the autocorrelation lengths becoming very larger, the layer characteristic of the models is very obvious. It would be difficult to identify sandstone such as gritstone, clay, dense sandstone and gas sandstone and so on in the reservoir with traditional impedance inversion. According to the obvious difference between different lithologic and petrophysical impedance, we studied multi-scale reservoir heterogeneities and developed new technologies. The distribution features of reservoir lithological and petrophysical heterogeneities along vertical and transverse directions were described quantitatively using multi-scale power spectrum and heterogeneity spectrum methods in this paper. Power spectrum (P spectrum) describes the manner of the vertical distribution of reservoir lithologic and petrophysical parameters and the large-scale and small-scale heterogeneities along vertical direction. Heterogeneity spectrum (H spectrum) describes the structure of the reservoir lithologic and petrophysical parameters mainly, that is to say, proportional composition of each lithological and petrophysical heterogeneities are calculated in this formation. The method is more reasonable to describe the degree of transverse multi-scale heterogeneities in reservoir lithological and petrophysical parameters. Using information of sonic logs in Sulige oil field, two spectral methods have been applied to the oil field, and good analytic results have been obtained. In order to contrast the former researches, the last part is the multi-scale character analysis of reservoir based on the transmission character of wave using the wavelet transform. We discussed the method applied to demarcate sequence stratigraphy and also analyzed the reservoir interlayer heterogeneity.

基于显式分形插值的高保真地震数据重建方法研究

Based on the fractal theories, contractive mapping principles as well as the fixed point theory, by means of affine transform, this dissertation develops a novel Explicit Fractal Interpolation Function（EFIF）which can be used to reconstruct the seismic data with high fidelity and precision. Spatial trace interpolation is one of the important issues in seismic data processing. Under the ideal circumstances, seismic data should be sampled with a uniform spatial coverage. However, practical constraints such as the complex surface conditions indicate that the sampling density may be sparse or for other reasons some traces may be lost. The wide spacing between receivers can result in sparse sampling along traverse lines, thus result in a spatial aliasing of short-wavelength features. Hence, the method of interpolation is of very importance. It not only needs to make the amplitude information obvious but the phase information, especially that of the point that the phase changes acutely. Many people put forward several interpolation methods, yet this dissertation focuses attention on a special class of fractal interpolation function, referred to as explicit fractal interpolation function to improve the accuracy of the interpolation reconstruction and to make the local information obvious. The traditional fractal interpolation method mainly based on the randomly Fractional Brown Motion (FBM) model, furthermore, the vertical scaling factor which plays a critical role in the implementation of fractal interpolation is assigned the same value during the whole interpolating process, so it can not make the local information obvious. In addition, the maximal defect of the traditional fractal interpolation method is that it cannot obtain the function values on each interpolating nodes, thereby it cannot analyze the node error quantitatively and cannot evaluate the feasibility of this method. Detailed discussions about the applications of fractal interpolation in seismology have not been given by the pioneers, let alone the interpolating processing of the single trace seismogram. On the basis of the previous work and fractal theory this dissertation discusses the fractal interpolation thoroughly and the stability of this special kind of interpolating function is discussed, at the same time the explicit presentation of the vertical scaling factor which controls the precision of the interpolation has been proposed. This novel method develops the traditional fractal interpolation method and converts the fractal interpolation with random algorithms into the interpolation with determined algorithms. The data structure of binary tree method has been applied during the process of interpolation, and it avoids the process of iteration that is inevitable in traditional fractal interpolation and improves the computation efficiency. To illustrate the validity of the novel method, this dissertation develops several theoretical models and synthesizes the common shot gathers and seismograms and reconstructs the traces that were erased from the initial section using the explicit fractal interpolation method. In order to compare the differences between the theoretical traces that were erased in the initial section and the resulting traces after reconstruction on waveform and amplitudes quantitatively, each missing traces are reconstructed and the residuals are analyzed. The numerical experiments demonstrate that the novel fractal interpolation method is not only applicable to reconstruct the seismograms with small offset but to the seismograms with large offset. The seismograms reconstructed by explicit fractal interpolation method resemble the original ones well. The waveform of the missing traces could be estimated very well and also the amplitudes of the interpolated traces are a good approximation of the original ones. The high precision and computational efficiency of the explicit fractal interpolation make it a useful tool to reconstruct the seismic data; it can not only make the local information obvious but preserve the overall characteristics of the object investigated. To illustrate the influence of the explicit fractal interpolation method to the accuracy of the imaging of the structure in the earth’s interior, this dissertation applies the method mentioned above to the reverse-time migration. The imaging sections obtained by using the fractal interpolated reflected data resemble the original ones very well. The numerical experiments demonstrate that even with the sparse sampling we can still obtain the high accurate imaging of the earth’s interior’s structure by means of the explicit fractal interpolation method. So we can obtain the imaging results of the earth’s interior with fine quality by using relatively small number of seismic stations. With the fractal interpolation method we will improve the efficiency and the accuracy of the reverse-time migration under economic conditions. To verify the application effect to real data of the method presented in this paper, we tested the method by using the real data provided by the Broadband Seismic Array Laboratory, IGGCAS. The results demonstrate that the accuracy of explicit fractal interpolation is still very high even with the real data with large epicenter and large offset. The amplitudes and the phase of the reconstructed station data resemble the original ones that were erased in the initial section very well. Altogether, the novel fractal interpolation function provides a new and useful tool to reconstruct the seismic data with high precision and efficiency, and presents an alternative to image the deep structure of the earth accurately.

奇异性分析方法及其在储层预测中的应用研究

In research field of oil geophysical prospecting, reservoir prediction is refers to forecasting physical properties of petroleum reservoir by using data of seismic and well logging, it is a research which can guide oil field development. Singularities of seismic and logging data are caused by the heterogeneity of reservoir physical property. It's one of important methods that using singularity characteristics of seismic and logging data to study the reservoir physical property in recently. Among them, realization of reservoir quantitative prediction by analyzing singularity of the data and enhancing transition description of data is difficulty in method research. Based on wavelet transform and the fractal theory, the paper studied the singularity judgment criterion for seismic and logging data, not only analyzed quantitative relation between singularity data and reservoir physical property, but also applied it in practical reservoir prediction. The main achievements are: 1. A new method which provides singular points and their strength information estimation at only one single scale is proposed by Herrmann (1999). Based on that, the dissertation proposed modified algorithm which realized singularity polarity detection. 2. The dissertation introduced onset function to generalize the traditional geologic boundaries variations model which used singularity characteristics to represent the abruptness of the lithologic velocity transition. We show that singularity analysis reveals generic singularity information conducted from velocity or acoustic impedance to seismogram based on the convolution seismic-model theory. Theory and applications indicated that singularity information calculated from seismic data was a natural attribute for delineating stratigraphy boundaries due to its excellent ability in detecting detailed geologic features. We demonstrated that singularity analysis was a powerful tool to delineate stratigraphy boundaries and inverse acoustic impedance and velocity. 3. The geologic significances of logging data singularity information were also presented. According to our analysis, the positions of singularities indicate the sequence stratigraphic boundary, and there is subtle relationship between the singularity strength and sedimentary environment, meanwhile the singularity polarity used to recognize stratigraphic base-level cycle. Based on all those above, a new method which provided sedimentary cycle analysis based on the singularity information of logging data in multiple scales was proposed in this dissertation. This method provided a quantitative tool for judging interface of stratum sequence and achieved good results in the actual application.

基于分形和统计理论的地层非均质研究

Basin-scale heterogeneity contains information about the traces of the past sedimentary cycle and tectonic process, and has been a major concern to geophysicists because of its importance in resource exploration and development. In this paper, the sonic data of 30 wells of Sulige field are used to inverse the power-law spectra slope and correlation length which are measures of the heterogeneity of the velocity of the log using fractal and statistic correlation methods. By taking the heterogeneity parameters of different wells interpolated, we get power law spectra slope and correlation length contours reflecting the stratum heterogeneity. Then using correlation and gradient, we inverse the transverse heterogeneity of Sulige field. Reservior-scale heterogeneity influnce the distribution of remaining oil and hydrocarbon accumulation. Using wavelet modulus maximum method to divide the sedimentary cycle using Gr data, therefore we can calculate the heterogeneity parameter in each layer of each log. Then we get the heterogeneity distribution of each layer of Sulige field. Finally, we analyze the relation between the signal sigularity and the strata heterogeneity, and get two different sigularity profiles in different areas.

活动构造区区域地壳稳定性研究——以金沙江水电梯级工程规划区为例

The study of regional crustal stability of active tectonic region basically includes analysis of recent activity of Earth's crust, single factor assessment, study of complexity, and comprehensive assessment of crust stability. In this thesis, some work are made as follows: · Based on abundant data from gravity field, aeromagnetic survey, magnetism, magnetotelluric deep sounding, remote sensing and geotectonic as well as earthquakes observed in recent years around this region and adjacent zones, we can get a through understanding about the structural features and activity of the earth's crust in Chuan-Dian region. The results from explosion earthquake and telluric electromagnetic sounding are consistent with the structural features of the crust manifested by the geophysical field. The data of deep geologic structures are important for us to work out a vivid three-dimensional model of the earth's crustal structure of the Jinsha River region. According to a synthesis, the author of this thesis proposes some indicators for dividing the faulted blocks. It can also be inferred that the movement of the Chuan-Dian faulted block, which is the relatively active part of southwestern China, is controlled by the boundary faults, and the intensive activity and deformation are concentrated along the boundaries of the block. · Mainly discussing respectively the mechanism and laws of active faults, earthquakes, and geological hazards activity, and their influences on the stability and security of engineering, also trying to probe into the way to assess the risk of single factor in this section. Especially with the method of fractal geometry, the thesis has discussed how to study the complexity of each factor. These geologic hazards which are distributed at the uppermost part of the crust in this region form a typical mountainous set of the active tectonic areas. The results of survey show that some slopes are liable- to -sliding with a weak layer of low shear strength. Occurrences of landslides are to a great extent related to local geological structures, in particular active faults. This is why numerous landslides have occurred simultaneously around the epicenter of a strong earthquake or the center of a strong rainfall, which are related to active faults. · The analysis of the crustal stability is based on a regional grid division, and a fuzzy comprehensive analysis method is used to determine the grade of the quality in each grid. The evaluation factors and their weights are taken from the results of the hierarchical analysis. The evaluation indexes consist of qualitative and quantitative ones. The qualitative ones can be quantified through the experts weighing system, while the quantitative ones can be obtained from statistical analysis. For quality grades, four levels are used: stable, essentially stable, sub-stable, and unstable. The results of the evaluation on Jinshajiang region demonstrate that the crustal stability become distinctly worse in the areas controlled by active deep faults. Therefore, detailed investigations on the active faulting and geologic hazards, include earthquake activity are especially necessary for those areas adjacent to the deep fault belts. On the bases of the data available and the survey results, we have made a preliminary assessment for the construction conditions and adaptability of every planned site in the middle or lower reaches of Jinsha River. Finally, the thesis prospected the vista of the study of crustal stability.

石油二次运移的物理与数值模拟实验及其应用

Facing the problems met in studies on predominant hydrocarbon migration pathways, experiments and numerical simulating were done in this thesis work to discuss the migration mechanisms. The aim is to analyze quantitatively the pathway pattern in basin scale and to estimate the hydrocarbon loss on the pathway that offer useful information for confirming the potential hydrocarbon accumulation. Based on our understandings on hydrocarbon migration and the fluid dynamic theory, a series of migration experiments were designed to observe the phenomena where kerosene is used as draining phase driven only by buoyancy force that expulses pore water. These experiments allow to study the formation of migration pathways, the distribution of non-wetting oil along these pathways, and the re-utilizing of previously existing pathways marked by residual traces etc. The types of pattern for migration pathways may be characterized by a phase diagram using two dimensionless numbers: the capillary number and the Bond number. The NMR technique is used to measure the average saturation of residual oil within the pathways. Based our experiment works and percolation concept, a numerical simulation model were proposed and realized. This model is therefore called as BP (Buoyancy Percolation) simulator, since buoyancy is taken as the main driving force in hydrocarbon migration. To make sure that BP model is applicable to simulate the process of oil secondary migration, the experimental phenomena are compared with those simulated with BP model by fractal method, and the result is positive. After then, we use BP simulator to simulate the process of migration of oil in the porous media saturated with water at different scale. And the results seem similar to those cited in literatures. In addition, our software is applied in Paris basin to predict the pathway of hydrocarbon migration happened in the Middle Jurassic reservoirs. It is found that the results obtained with our BP model are generally agree with Hindle (1997) and Bekeles'(1999), but our simulated migration pathway pattern and migration direction seem more reasonable than theirs.

溶蚀岩体渗透水力学特性研究

Seepage control in karstic rock masses is one of the most important problems in domestic hydroelectric engineering and mining engineering as well as traffic engineering. At present permeability assessment and leakage analysis of multi-layer karstic rock masses are mainly qualitative, while seldom quantitative. Quantitative analyses of the permeability coefficient and seepage amount are conducted in this report, which will provide a theoretical basis for the study of seepage law and seepage control treatment of karstic rocks. Based on the field measurements in the horizontal grouting galleries of seepage control curtains on the left bank of the Shuibuya Hydropower Project on the Qingjiang river, a hydraulic model is established in this report, and the computation results will provide a scientific basis for optimization of grouting curtain engineering. Following issues are addressed in the report. (1) Based on the in-situ measurements of fissures and karstic cavities in grouting galleries, the characteristics of karstic rock mass is analyzed, and a stochastic structural model of karstic rock masses is set up, which will provide the basis for calculation of the permeability and leakage amount of karstic rock mass. (2) According to the distribution of the measured joints in the grouting galleries and the stochastic results obtained from the stochastic structural model of karstic rock mass between grouting galleries, a formula for computation of permeability tensor of fracturing system is set up, and a computation program is made with Visual Basic language. The computation results will be helpful for zoning of fissured rock masses and calculation of seepage amount as well as optimization of seepage control curtains. (3) Fractal theory is used to describe quantitatively the roughness of conduit walls of karstic systems and the sinuosity of karstic conduits. It is proposed that the roughness coefficient of kastic caves can be expressed by both fractal dimension Ds and Dr that represent respectively the extension sinuosity of karstic caves and the roughness of the conduit walls. The existing formula for calculating the seepage amount of karstic conduits is revised and programmed. The seepage amount of rock masses in the measured grouting galleries is estimated under the condition that no seepage control measures are taken before reservoir impoundment, and the results will be helpful for design and construction optimization of seepage curtains of the Shuibuya hydropower project. This report is one part of the subject "Karstic hydrogeology and the structural model and seepage hydraulics of karstic rock masses", a sub-program of "Study on seepage hydraulics of multi-layer karstic rock masses and its application in seepage control curtain engineering", which is financially supported by the Hubei Provincial key science and technology programme.

胜利油田低渗透砂岩油藏非均质地质建模

Guided by geological theories, the author analyzed factual informations and applied advanced technologies including logging reinterpretation, predicting of fractal-based fracture network system and stochastic modeling to the low permeable sandstone reservoirs in Shengli oilfield. A new technology suitable for precious geological research and 3D heterogeneity modeling was formed through studies of strata precious correlation, relation between tectonic evolution and fractural distribution, the control and modification of reservoirs diagenesis, logging interpretation mathematical model, reservoir heterogeneity, and so on. The main research achievements are as follows: (1) Proposed four categories of low permeable reservoirs, which were preferable, general, unusual and super low permeable reservoir, respectively; (2) Discussed ten geological features of the low permeable reservoirs in Shengli area; (3) Classified turbidite fan of Es_3 member of the Area 3 in Bonan oilfield into nine types of lithological facies, and established the facies sequences and patterns; (4) Recognized that the main diagenesis were compaction, cementation and dissolution, among which the percent compaction was up to 50%~90%; (5) Divided the pore space in ES_3 member reservoir into secondary pores with dissolved carbonate cement and residual intergranular pores strongly compacted and cemented; (6) Established logging interpretation mathematical model guided by facies- control modeling theory; (7) Predicted the fracture distribution in barriers using fractal method; (8) Constructed reservoir structural model by deterministic method and the 3D model of reservoir parameters by stochastic method; (9) Applied permeability magnitudes and directions to describe the fractures' effect on fluid flow, and presented four different fractural configurations and their influence on permeability; (10) Developed 3D modeling technology for the low permeable sandstone reservoirs. The research provided reliable geological foundation for the establishment and modification of development plans in low permeable sandstone reservoirs, improved the development effect and produced more reserves, which provided technical support for the stable and sustained development of Shengli Oilfield.

油藏数值模拟中的网格粗化和细化技术研究

The grid is a foundation of reservoir description and reservoir simulation. The scale of grid size is vital influence for the precision of reservoir simulation the gridding of reservoir parameters require reasonable interpolation method with computing quickly and accurately. The improved distant weighted interpolation method has many properties, such as logical data points selection, exact interpolation, less calculation and simply programming, and its application can improve the precision of reservoir description and reservoir simulation. The Fractal geologic statistics describes scientifically the distribution law of various geological properties in reservoir. The Fractal interpolation method is applied in grid interpolation of reservoir parameters, and the result more accorded with the geological property and configuration of reservoir, and improved the rationality and quality of interpolation calculation. Incorporating the improved distant weighted interpolation method with Fractal interpolation method during mathematical model of grid-upscaling and grid-downscaling, the softwares of GROUGH(grid-upscaling) and GFINE (grid-downscaling) were developed aiming at the questions of grid-upscaling and grid-downscaling in reservoir description and reservoir simulation. The softwares of GROUGH and GFINE initial applied in the research of fined and large-scale reservoir simulation. It obtained fined distribution of remaining oil applying grid-upscaling and grid-downscaling technique in fined reservoir simulation of Es21-2 Shengtuo oilfield, and provided strongly and scientific basis for integral and comprehensive adjustment. It's a giant tertiary oil recovery pilot area in the alkaline/surfactant/polymer flooding pilot area of west district of Gudao oilfield, and first realized fined reservoir simulation of chemical flooding using grid-upscaling and grid-downscaling technique. It has wide applied foreground and significant research value aiming at the technique of grid-upscaling and grid-downscaling in reservoir description and reservoir simulation.

花岗岩风化土微观结构及其工程地质特性研究

Saprolite is the residual soil resulted from completely weathered or highly weathered granite and with corestones of parent rock. It is widely distributed in Hong Kong. Slope instability usually happens in this layer of residual soil and thus it is very important to study the engineering geological properties of Saprolite. Due to the relic granitic texture, the deformation and strength characteristics of Saprolite are very different from normal residual soils. In order to investigate the effects of the special microstructure on soil deformation and strength, a series of physical, chemical and mechanical tests were conducted on Saprolite at Kowloon, Hong Kong. The tests include chemical analysis, particle size analysis, mineral composition analysis, mercury injection, consolidation test, direct shear test, triaxial shear test, optical analysis, SEM & TEM analysis, and triaxial shear tests under real-time CT monitoring.Based on the testing results, intensity and degree of weathering were classified, factors affecting and controlling the deformation and strength of Saprolite were identified, and the interaction between those factors were analyzed.The major parameters describing soil microstructure were introduced mainly based on optical thin section analysis results. These parameters are of importance and physical meaning to describe particle shape, particle size distribution (PSD), and for numerical modeling of soil microstructure. A few parameters to depict particle geometry were proposed or improved. These parameters can be used to regenerate the particle shape and its distribution. Fractal dimension of particle shape was proposed to describe irregularity of particle shapes and capacity of space filling quantitatively. And the effect of fractal dimension of particle shape on soil strength was analyzed. At the same time, structural coefficient - a combined parameter which can quantify the overall microstructure of rock or soil was introduced to study Saprolite and the results are very positive. The study emphasized on the fractal characteristics of PSD and pore structure by applying fractal theory and method. With the results from thin section analysis and mercury injection, it was shown that at least two fractal dimensions Dfl(DB) and Df2 (Dw), exist for both PSD and pore structure. The reasons and physical meanings behind multi-fractal dimensions were analyzed. The fractal dimensions were used to calculate the formation depth and weathering rate of granite at Kowloon. As practical applications, correlations and mathematical models for fractal dimensions and engineering properties of soil were established. The correlation between fractal dimensions and mechanical properties of soil shows that the internal friction angle is mainly governed by Dfl 9 corresponding to coarse grain components, while the cohesion depends on Df2 , corresponding to fine grain components. The correlations between the fractal dimension, friction angle and cohesion are positive linear.Fractal models of PSD and pore size distribution were derived theoretically. Fragmentation mechanism of grains was also analyzed from the viewpoint of fractal. A simple function was derived to define the theoretical relationship between the water characteristic curve (WCC) and fractal dimension, based on a number of classical WCC models. This relationship provides a new analytical tool and research method for hydraulic properties in porous media and solute transportation. It also endues fractal dimensions with new physical meanings and facilitates applications of fractal dimensions in water retention characteristics, ground water movement, and environmental engineering.Based on the conclusions from the fractal characteristics of Saprolite, size effect on strength was expressed by fractal dimension. This function is in complete agreement with classical Weibull model and a simple function was derived to represent the relationship between them.In this thesis, the phenomenon of multi-fractal dimensions was theoretically analyzed and verified with WCC and saprolite PSD results, it was then concluded that multi-fractal can describe the characteristics of one object more accurately, compared to single fractal dimension. The multi-fractal of saprolite reflects its structural heterogeneity and changeable stress environment during the evolution history.