【作者】 王颖；

【导师】 彭省临； 董军；

【作者基本信息】 中南大学 ， 国土资源信息工程， 2010， 博士

【摘要】 为危机矿山寻找可接替资源已成为当前成矿预测研究领域的热点问题。一方面,危机矿山具有矿山勘探开发程度高、地质勘探原始资料积累充分、评价与找矿向深边部三维空间发展的特点。另一方面,成矿预测学中的许多重大基础问题的研究常常涉及到非线性问题,如矿床与成矿事件及各种成矿控制因素之间不显示简单的线性关系。因此,建立适应于危机矿山深边部及外围找矿的矿产资源评价新理论新方法和找矿创新体系成为当务之急。多属性决策问题广泛存在于社会、经济、管理、工程等领域。由于现实世界的大部分事物或现象都具有或多或少的模糊性,因此,处理模糊型非线性现象的多属性模糊优选决策模型是解决综合评价问题的重要方法之一Meta分析方法是当今比较流行的一种对同一主题下多个独立实验结果进行综合的统计方法,具有强大的系统研究功能。Meta分析思想的精髓是“对具备特定条件的、同主题的诸多研究结果进行综合后统计”。Meta分析经历了三个发展阶段,时至今日,已经广泛的应用于教育学、医学、心理学以及生态学等方面,但在地质学研究领域尚未见运用,本文首次将该方法应用于各种找矿信息的重要性比较。本文结合“十一五”国家科技支撑计划项目——“铜陵地区危机铜矿山大比例尺定位预测技术示范研究”,以铜陵铜矿区和天马山—金口岭—铜官山地区作为研究区,开展了基于模糊多属性决策理论与Meta分析模型的隐伏矿定位预测研究。取得了如下主要成果：(1)在探查技术有效性研究中引入Meta分析。项目组对铜山铜矿区前山南测区开展了高频大地电磁测深法EH-4、可控源音频大地电磁测深CSAMT、时间域瞬变电磁法TEM三种方法的联合测量工作。获得了6个内容翔实的二维测深剖面,本文对这6个剖面进行了10×10米,20×20米,40×40米的网格化工作,最终到了18个网格化剖面。然后对这18个剖面的数据进行了量化。最后利用Meta分析对所有数据进行了统计分析,结果表明,在前山南测区,EH-4和CSAMT两种方法的找矿预测有效性相当,且均比TEM法的有效性高得多。(2)提出了基于模糊数学的Meta分析模型。利用Meta分析进行探查技术有效性研究的过程中,发现Meta分析得出的是统计学意义上的结果,回答的是诸如“甲比乙有效性好还是差”的定性问题,因而,结论是定性的。而模糊数学是运用多层次模糊综合评判的方法,系统的综合由统计学方法得出的若干单项研究结果,进而可得出模糊性的定量结论。本文将模糊数学引入到了Meta分析中,构建了基于模糊数学的Meta分析模型。并利用该模型进行了CSAMT法与TEM法对预测线有效度的研究。最终结果认为,在铜山铜矿区前山南测区,CSAMT法比TEM法的找矿预测有效度高17.7%。(3)首次使用Meta分析对成矿指标的重要性进行了定量分析和比较。对模糊层次分析法而言,构建一个符合实际情况的模糊互补判断矩阵是该分析法的难点和关键所在。这是因为当人们在对诸多评判指标重要性进行两两比较时,受到太多人为因素的影响。因此,本文首次使用Meta分析的方法对各指标的重要性进行了定量分析和比较,达到了预期目的。利用Meta分析,能够客观地获得各成矿信息重要性的排序。(4)在多元成矿信息综合集成研究中引入了多属性模糊优选决策模型。基于各个矿区成矿机制与成矿演化模式具有复杂性、独特性及非线性的特点,将多属性模糊优选决策模型引入多元信息综合集成研究领域。并且为了更好的确定评判指标的权重,对层次分析法进行了改进,引入了模糊层次分析法,构建了模糊层次分析法的数学模型。利用该模型,对天马山-金口岭-铜官山地区进行了多元成矿信息综合预测与评价,并结合地质分析,圈定了12个成矿有利度较高的靶区对照其它相关资料及野外实地验证,结果表明：利用该模型所圈定的靶区成矿条件有利,找矿潜力大。(5)研究并设计了基于Meta分析的隐伏矿定位预测算法。有效地求解出实际线(面)与预测线(面)的吻合度,是Meta分析方法能否被成功引入到成矿预测研究领域的关键,也是构建数字矿床模型的基础。本文设计了求解吻合度的算法,并开发了相应软件。运用上述软件,并结合RevMan,可较好的完成基于Meta分析的成矿信息权重比较及探查技术有效性比较等研究工作。更多还原

【Abstract】 In order to ensure the sustainable development of the social economy of our country, it becomes presently an outstanding problem to explore replaceable resources for the crisis mines. On one hand, the crisis mines have the following characteristics:high intensity in mine exploration and development, sufficient accumulation in the firsthand information of geological exploration, development towards the deep and marginal three-dimensional space for evaluation and ore prospecting. On the other hand, many important and basic problems in mineralization and predictive research often involve non-linear problems. For instance, the simple linear relation can not be revealed between the deposits and control factors of mineralization. Therefore, it became an urgent affair to establish the new method and theory for evaluating mineral resources in new prospecting innovation systems that are suitable for the crisis mine prospecting in the deep, marginal and peripheral locations.Multi-attribute decision-making problems exist in many fields, such as society, economy, management, projects, etc. Because most phenomena in the real world have fuzziness more or less, the fuzzy optimization model of multi-attribute decision-making characteristics for dealing with the nonlinear phenomenon of fuzzy type turns to be one of the most important methods in the comprehensive evaluation.Meta-analysis is a statistical method that systematically combines, analyzes and comprehensively evaluates several relevant trials. The soul of the Meta-analysis is that similar studies with specific conditions and the same themes are statistical after synthesis. The Meta-analysis has experienced three developmental stages. It has been widely used in Education, Psychology, Ecology and Medicine, but is new in geology today. In this paper, the Meta analysis is first used to investigate the importance of the various prospecting information.The thesis is supported by the following project:The national science and technology supporting project of" eleven five "-"The study of technology demonstration on the location prediction of large-scale crisis copper mines in Tongling areas". Taking Tongling copper mine and Tianmashan-Jingkouling-Tongguanshan regions as the research area, the study of concealed deposit location prediction is carried out by means of the Fuzzy optimize model of multi-attribute decision-making characteristics and the Meta-analysis model. The main conclusions are listed as follows:(1) The Meta-analysis is applied to the comparison of availability of several explorations. The project teams conduct the work by using the methods of the EH-4, CSAMT, TEM in Qianshannan survey area of Tongshan copper mine. In order to obtain six two-dimensional depth profiles that have abundant contents, the thesis carries out the gridding work of 10×10m,20x20m,40x40m, and gets the result of 18 gridding profiles for making the quantitative analysis of the data. In addition, we carried out an statistical analysis to all the data by using the method of the Meta-analysis. The results show that the EH-4 and CSAMT methods are equivalent on the basis of effectiveness, and the TEM is less effective than above two.(2) The Meta-analysis model is further developed on the basis of fuzzy mathematics. In the process of investigating the availability of several explorations by using the Meta-analysis, we find that the Meta-analysis can produce the results (such as" values") of statistical significance, so that it answers the question of "is A better or worse than B in terms of effectiveness". In this situation, the conclusion is qualitative. In spite of that, the fuzzy mathematics uses the method of judge synthetically and it is multi-level and fuzzy. It systematically integrates the results of several individual events obtained from the statistics method, meaning that quantitative conclusions can be drawn from the use of fuzzy mathematics. Thus, the thesis applies the fuzzy mathematics to the Meta-analysis in chapter 6, constructs the Meta-analysis model based on fuzzy mathematics, and carries out the availability research of the CSAMT and TEM methods by using the model. The results show that the CSAMT is 17.7% higher in availability than the TEM when they are used in the Qianshannan survey area. (3) Quantitative analysis and comparison have been conducted, for the first time, to evaluate the metallogenic index using the Meta-analysis method. The difficulty and the sticking point are to construct a fuzzy complementary judge matrix that is suitable for the actual conditions. This is because that the matrix is influenced too much by human factors when people compare the judging indexes with each other. Therefore, this thesis uses the Meta-analysis method to carry out the quantitative analysis and to investigate the importance of different index. As a result, the anticipated goal has been achieved. This indicates that the weightiness sequencing can be obtained by using the Meta-analysis in mineralization information.(4) This study applies the Fuzzy optimize model of multi-attribute decision-making characteristics to the multivariate metallogenic information meta-synthesis. The metallogenic mechanism and metallogenic evolution model have the characteristics of complexity, uniqueness and non-linearity. Furthermore, the study introduces the FAHP (fuzzy analytic hierarchy process) to make use of the judging indexes quota, and to construct the mathematical model of the FAHP. The study also applies the multivariate metallogenic information synthetic prediction and evaluation to the Tianmashan-Jingkouling-Tongguanshan regions, and integrates the geological analysis, delineates 12 target areas which have highly advantage degree in metallogenic aspects. Compared with the other relevant materials and the field verification, the results show that using this kind of model to delineate target areas has some remarkable advantages in prospecting.(5) The concealed mineral location prediction software is developed on the basis of the Meta-analysis. Whether it can be used to identify the degree of coincidence between the practical lines (surfaces) and the predicted lines (surfaces) is the key point that we introduce the Meta-analysis method into metallogenic prognosis study area. This progress is the foundation of constructing a digital mineral deposit model. We design algorithms to identify the degree of coincidence, and develop the corresponding software. As a result, we recommend a kind of Meta-analysis software, RevMan, in detail so that it can be wildly used. The software described above has been used to quantitatively conduct the study of concealed mineral location prediction on the basis of the Meta-analysis.更多还原