| 许逢明.黑龙江多宝山矿集区三维地质建模与深部找矿预测[J].地质与勘探,2022,58(3):629-652 |
| 黑龙江多宝山矿集区三维地质建模与深部找矿预测 |
| Three-dimensional geological modeling and deep prospecting prediction of the Duobaoshan ore concentration area in Heilongjiang Province |
| 投稿时间：2021-08-10 修订日期：2022-01-03 |
| DOI：10.12134/j.dzykt.2022.03.014 |
| 中文关键词: 三维地质建模 找矿预测 靶区圈定 钻探验证 多宝山矿集区 黑龙江省 |
| 英文关键词: three-dimensional (3D) geological modeling, prospecting prediction, target delineation, drilling verification, Duobaoshan ore concentration area, Heilongjiang Province |
| 基金项目:国家重点研发计划深地资源勘查开采专项项目（No.2017YFC0601305）和中国地质调查局地质调查项目（DD20201162）联合资助 |
|
| 摘要点击次数: 1788 |
| 全文下载次数: 621 |
| 中文摘要: |
| 随着三维地质信息技术的发展,三维地质建模和深部成矿预测研究对实现深部找矿突破的作用越来越重要。本次基于已知矿区的中段、勘探线剖面图及钻孔数据等资料，依托Micromine平台，从矿床尺度（比例尺1:1000）建立了黑龙江多宝山铜矿、铜山铜矿的地质体和蚀变矿化体两类三维地质实体模型。以地表地质图、图切地质剖面、钻孔数据及岩石物性数据为约束，基于重磁电三维物性反演，综合解译了93条重磁电剖面。在综合剖面地质解译成果的基础上，利用平行剖面显式建模技术，从矿集区尺度（比例尺1:2.5万）建立了多宝山矿集区84km2范围内3000m以浅三维地质体实体模型。在三维地质体模型的基础上，以多宝山、铜山铜矿体三维地质模型为先验数据模型，根据区域找矿模型，利用“立方体预测模型”法，提取与成矿有关二值变量（地层、岩体、构造）、连续变量（重、磁、电）等7个勘探变量，利用GeoCube3.0逻辑斯蒂回归深层次机器学习集成模块，开展了研究区的勘探变量集成，经由ROC模型评估后验概率，在可靠后验概率的基础上，利用C-V分形方法计算阈值以圈定深部找矿有利靶区，综合圈定深部找矿靶区8处，预估铜潜在资源量634万吨。按照靶区综合评级排序，对最优靶区A-8开展钻探验证，经证实深部发现厚大铜钼矿体，实现了多宝山矿集区深部找矿的重大突破。 |
| 英文摘要: |
| With the development of 3D geological information technology, the role of 3D geological modeling and deep metallogenic prognosis and research increases in importance to achieve deep prospecting breakthroughs. Grounded upon the data of the middle section of the known mining area, the exploration profile and drilling data, two types of 3D geological solid models for the geological bodies and altered mineralization bodies of the Duobaoshan copper mine and the Tongshan copper mine were established from the scale of the deposit (at a scale of 1:1000). Constrained by surface geological maps, map-cut geological sections, borehole data and petrophysical data, 93 gravity, magnetic, electroseismic profiles were comprehensively interpreted . On the basis of comprehensive geological interpretation on the profiles and the parallel profile explicit modeling technology, a shallow 3D physical model of 3,000m geological body was constructed from the scale of the mining area (at a scale of 1:25,000). Upon the 3D geological body models and with the 3D geological model of the Duobaoshan and Tongshan copper ore bodies used as the prior data model, seven exploration variables including the metallogeny-related binary variables (stratum, rock mass, structure) and continuous variables (gravity, magnetism, electricity) were extracted according to the deep prospecting prediction model and with the method of "cube prediction model". The GeoCube3.0 deep machine learning integration modules (logistic regression methods) were used for integration of exploration variables in the study area, and the posterior probability was also evaluated through ROC model. On the basis of the reliable posterior probability, the concentration-volume fractal method was used to calculate the threshold to delineate favorable target areas for deep prospecting, and six deep prospecting target areas were comprehensively delineated to estimate the volume of copper resources at 6.34 million tons. According to the comprehensive ranking of the target area, drilling verification was carried out on the optimal target area A-8. It was confirmed that thick copper-molybdenum ore bodies were discovered in the deep part, and a major breakthrough in deep prospecting in the Duobaoshan ore concentration area was realized. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
