| With the development of human society, geological mining condition became more and more complicated with gradually increasing production system and complex and varied digging environment. However, requirement for mine are growing including safety, production, automation as well as monitoring and controlling. The traditional mine construction model cannot satisfy the demand above. It’s urged to create another management model and technological means to establish green, intelligent and sustainable smart mine, which includes innovation in reserve estimation method is incapable of coping the mine environment in complicated mass data condition, difficult to support smart mine construction, badly in need of researching and developing a new reserve estimation method with high performance and high precision. Ground on this point, focusing on complex mine’s geological condition and mass data environment, the author carried out a series of exploratory research, trying to develop a kind of new kriging reserve estimation method with intelligent features, then verifying the feasibility by experiment.The research started with conceptual model of traditional kriging reserve estimation, comprehensively analyzing the key steps and implementation plan involved in the layers of data acquisition, model construction, variogram calculation, and result output. On the foundation above cutting-edge technology such as genetic algorithm, dynamic programming and parallel GPU computing and so on were introduced to establish conceptual model of smart kriging reserve estimation. Through analyzing the core features of smart mine and three key steps in smart mine construction, namely instrumentation, interconnection, and intelligence, the conceptual model of smart kriging reserve estimation was integrated in smart mine’s three-layer framework to form implementation plan of smart kriging reserve estimation oriented to smart mine. On this basis, the author emphasized discussing the smart process needed in intelligent layer. First, for intelligence issue of variogram calculation, the key problem in kriging calculation flow, the author adopted optimizing search to conduct parallel computing to whole samples, using local output to accelerate calculation process of experimental variogram, then employing genetic algorithm’s auto-fitting step to improve the precision of calculation result; Second, expounding the techniques necessary to the whole smart reserve estimation process one by one, such as, 3D modeling, dynamic calculation of ore body reserve, spatial kriging fast interpolation, dynamic calculation of best boundary grade, as well as space cutting of 3D model of the calculation results, and so on. Last, based on the data from Shimensi tungsten copper mine, systematic application experiment was conducted to verify the all constructed scheme. The main research contents include:(1) Analyzing intelligent kriging reserve estimation system’s function and its necessity to smart mine construction, researching data processing procedure and requirement of data transmission; determining the functional and logical structure of the system.(2) According to kriging’s 3D reserve estimation procedure combined with framework features of smart mine system engineering researching the smart kriging reserve estimation’s framework and its algorithm oriented to smart mine.(3) Studying the smart application issue involved in smart kriging reserve estimation and analyzing its intelligent algorithm, establishing the initial scheme applied to data management and application of reserve estimation.(4) Researching variogram optimizing algorithm during the process of smart kriging reserve estimation, and analyzing the limitation of traditional experimental variogram, adopting parallel computing scheme based on GPU for key calculating steps and emphasizing on parallel scheme’s disadvantages in experimental variogram calculation designing new solution strategies and framework.(5) Researching the theoretical fitting optimizing algorithm of experimental variogram of smart kriging reserve estimation method. On the foundation of analyzing the characteristics and performance of fitting process, studying the issue of result optimization by using smart algorithm to conducting matching namely, employing linear programing of the weight coefficient grounded on lag distance reciprocal to establish objective function, then using genetic algorithm to achieving solution by auto-fitting.(6) Based on the data of drill hole and geological section and etc. of Shimensi tungsten copper mine,3D geologic model of the mine lot was established to discuss and study the function of 3D geologic modeling technique in smart reserve estimation.(7) Designing and developing smart kriging reserve estimation system. Using real data of Shimensi tungsten copper mine to do experiment of reserve estimation, verify the feasibility of theories, methods and techniques in the thesis.To sum up, the author proposed smart kriging reserve estimation implementation plan and software system oriented to smart mine through theoretical research, technological development and practical application. From dynamic, automatic and intelligent perspective of reserve estimation, it provides technical support to smart mine construction as well as a new high efficient strategy to reserve estimation. Then the feasibility was verified by application examples.In conclusion, the innovation points of the research are as follows:(1) A smart kriging reserve estimation scheme was put forward by organically integrating several automatic and intelligent methods. The scheme can support the smart calculation of ore reserve of different minerals in different environment, which provided a new way to smart mine’s construction and application from a certain perspective.(2) Aiming at complicated source and large scale volume of mineral grade data, a intelligent variogram calculation solution was proposed to overcome the disadvantages of traditional experimental variogram that has slow calculating speed and high dependence to experts’experience in theoretical matching. This scheme includes conducting optimizing search to experimental variogram, fast output by parallel computing and auto-matching to output results by genetic algorithm etc.(3) The block mode’s space cutting mapping algorithm in mineral 3D reserve estimation results overcame the defect of traditional kriging reserve estimation which is with single result and fully used results information in ore body’s 3D block models, guiding the dynamic reserve estimation and mining in later period.The findings in the thesis promotes the automation of real reserve estimation as well as smart mine construction in certain degree, which has been applied in the project, " 3-D reserve estimation subsystem with multi-method" by far. |
PDF Full Text Request