SVM model for estimating the parameters of the probability-integral method of predicting mining subsidence

A new mathematical model to estimate the parameters of the probability-integral method for mining subsidence prediction is proposed. Based on least squares support vector machine (LS-SVM) theory, it is capable of improving the precision and reliability of mining subsidence prediction. Many of the geological and mining factors involved are related in a nonlinear way. The new model is based on statistical theory (SLT) and empirical risk minimization (ERM) principles. Typical data collected from observation stations were used for the learning and training samples. The calculated results from the LS-SVM model were compared with the prediction results of a back propagation neural network (BPNN) model. The results show that the parameters were more precisely predicted by the LS-SVM model than by the BPNN model. The LS-SVM model was faster in computation and had better generalized performance. It provides a highly effective method for calculating the predicting parameters of the probability-integral method.     

冬瓜山矿体稳定性和岩石力学特性分析

通过对冬瓜山矿体岩石力学特性的分析,探讨了矿体特征及围岩性质、原岩应力、岩石物理力学性质、岩体的整体稳定性,以及矿岩构造条件、采场结构参数和采场回采顺序对巷道稳定性的影响.结果表明:冬瓜山矿体表现出软岩的性质,属于高应力软岩;冬瓜山矿体地应力场的分布规律可以作为矿山设计及岩体稳定性评价的依据;矿岩构造条件对矿床开采过程巷道的稳定性不会造成较大的影响;选择合适的采场结构参数和采场回采顺序,可减小对巷道稳定性的影响.     

林歹铝矿崩落采矿法借模拟试验优选采场结构与参数

本文简述了林歹矿床的地质开采技术条件,矿块崩落采矿法的选择、结构、工艺特点、和提出模拟试验的理由。论述了模拟试验的目的、任务、目标函数,主要因素和水平范围的确定。以严密地数理统计分析,建立了主要采场结构因素与回贫指标关系的数学模型。最后根据铝氧厂对铝土矿矿石质量要求,优选出采场最佳结构参数,并预报了回贫指标。为工业试验设计、施工、生产提供了理论依据,并对类似矿山也具有借鉴作用。     

An improved method to assess the required strength of cemented backfill in underground stopes with an open face

Backfill is increasingly used in underground mines to reduce the surface impact from the wastes produced by the mining operations. But the main objectives of backfilling are to improve ground stability and reduce ore dilution. To this end, the backfill in a stope must possess a minimum strength to remain self-standing during mining of an adjacent stope. This required strength is often estimated using a solution proposed by Mitchell and co-workers, which was based on a limit equilibrium analysis of a wedge exposed by the open face. In this paper, three dimensional numerical simulations have been performed to assess the behavior of the wedge model. A new limit equilibrium solution is proposed, based on the backfill displacements obtained from the simulations. Comparisons are made between the proposed solution and experimental and numerical modeling results. Compared with the previous solution, a better agreement is obtained between the new solution and experimental results for the required cohesion and factor of safety. For large scale (field) conditions, the results also show that the required strength obtained from the proposed solution corresponds quite well to the simulated backfill response.     

Strategic sill pillar design for reduced hanging wall overbreak in longhole mining

Steeply dipping, vein and tabular orebodies are traditionally extracted with longitudinal retreat mining methods such as Eureka and Avoca in a bottom-up sequence with delayed backfill. To increase productivity, sill pillars in the orebody are used to separate mining zones thus allowing production to take place simultaneously in two or more zones. While such mining methods are productive, they may be accompanied with high volumes of hanging wall overbreak causing significant unplanned ore dilution. In this work, it is shown through a mine design case study of a narrow vein deposit that a sill pillar could also play a significant role in limiting hanging wall overbreak. To demonstrate the role of sill pillar, a novel numerical modelling scheme is proposed to account for progressive stope wall overbreak. A numerical modelling approach of element death and rebirth is developed to allow for the detected stope overbreak to be immediately removed and replaced with backfill material before upper-level stope extraction. It is further shown that the average overbreak volume could be reduced by as much as 33% when the sill pillar is strategically placed in the lower half of a mine plan.     

Analysis of factors and countermeasures of mining subsidence in Kaiyang Phosphorus Mine

1 INTRODUCTION After useful mineral body has been mined and dragged out, mining goaf is formed and country rocks would probably lose original equivalence. When span of the goaf is long enough, even if the roof is hard and intact, the rock around the goaf would move towards the mining openings, and this movement would pass on to the adjoining rock mass until it reaches the surface and ground displacement can be observed. In other words, underground mining may cause ground cracks and surface…     

A novel artificial intelligent model for predicting air overpressure using brain inspired emotional neural network

Blasting is the live wire of mining and its operations, with air overpressure(AOp) recognised as an end product of blasting. AOp is known to be one of the most important environmental hazards of mining.Further research in this area of mining is required to help improve on safety of the working environment.Review of previous studies has shown that many empirical and artificial intelligence(AI) methods have been proposed as a forecasting model. As an alternative to the previous methods, this study proposes a new class of advanced artificial neural network known as brain inspired emotional neural network(BIENN) to predict AOp. The proposed BI-ENN approach is compared with two classical AOp predictors(generalised predictor and McKenzie formula) and three established AI methods of backpropagation neural network(BPNN), group method of data handling(GMDH), and support vector machine(SVM). From the analysis of the results, BI-ENN is the best by achieving the least RMSE, MAPE, NRMSE and highest R, VAF and PI values of 1.0941, 0.8339%, 0.1243%, 0.8249, 68.0512% and 1.2367 respectively and thus can be used for monitoring and controlling AOp.     

Location prediction of blind ore-bodies in Shuijingtun Gold Mine, Zhangjiakou, China

Ｔｈｅｐｒｅｓｅｎｔｒｅｓｅｒｖｅｓｏｆｍａｎｙｇｏｌｄｍｉｎｅｓａｒｅｉｎｃｒｅａ ｓｉｎｇｌｙｅｘｈａｕｓｔｅｄ .Ｓｏｉｔｉｓｓｉｇｎｉｆｉｃａｎｔｔｏｐｒｅｄｉｃｔｂｌｉｎｄｏｒｅｂｏｄｉｅｓｉｎｔｈｏｓｅｍｉｎｅｓｔｏｅｘｐａｎｄｇｏｌｄｒｅｓｅｒｖｅｓａｎｄｐｒｏｌｏｎｇｔｈｅｉｒｐｒｏｄｕｃｔｉｖｅｌｉｆｅ[1～ 3] .Ｉｎｔｈｉｓｐａｐｅｒ,ａｎｅｘａｍｐｌｅｏｆｌｏｃａｔｉｏｎｐｒｅｄｉｃｔｉｏｎｏｆｂｌｉｎｄｏｒｅｂｏｄｉｅｓｉｎＳｈｕｉｊｉｎｇｔｕｎＧｏｌｄＭｉｎｅｉｓｉｎｔｒｏｄ…     

面向对象的矿山生产计划系统的应用

介绍作者研制的基于知识面向对象的矿山生产计划系统在山东某金矿的应用情况方案评价指标，除沿用矿山的现行指标之外，增设了另一指标———一年的冶炼回收金属量与采矿消耗金属量之比值结果表明：应用该系统编制的矿山１９９４年的采矿生产计划，明显优于矿山的手工编制计划     

混沌局域法与神经网络组合供水量预测

城市供水量是非线性、非平稳时间序列,组合预测模型能获得更高精度预测结果。通过深入分析混沌局域法与神经网络预测模型特点,提出了一种新的组合预测模型。首先,应用混沌局域法对城市日供水量进行初预测,然后,应用神经网络对预测结果进行修正。由于所提出的组合模型利用了混沌局域法及神经网络进行优势互补,能同时提高预测精度与计算效率。为验证所提出组合预测模型的可行性,采用某市7 a实测供水量数据,对混沌局域法、BPNN、RBF及GRNN神经网络4种单一预测模型及相应的3种组合模型预测精度进行定量分析,结果表明,组合预测模型精度都高于对应单一预测模型,混沌局域法与GRNN神经网络组合模型预测精度最高,且运算时间远低于单一神经网络模型运算时间。     

控顶切底房柱采矿法

控顶切底房柱采矿法采用浅孔切顶、切底、上向平行孔回采中层矿石,改进切顶房柱法的部分工艺,省去下盘脉外底部结构,回采作业在中层矿石“悬臂梁”的保护下进行,适合于矿石稳固但下盘围岩松散破碎的缓倾斜矿体的开采。     

经验知识监督的RC墩柱力学性能神经网络分析方法

本文通过研究钢筋混凝土（RC）墩柱力学性能试验数据、经验知识和机器学习的融合机制，提出了经验知识监督的RC墩柱力学性能神经网络（Knowledge-guided neural network, KGNN）分析方法。首先建立了包含761组RC墩柱拟静力试验样本的数据库；随后，基于经验知识分析了RC墩柱主要特征对其力学性能影响规律，构建了相应的数学表征方法；最后，将RC墩柱试验数据及经验知识融入人工神经网络架构和训练过程，建立了高精度、可解释、可通用且不依赖大量训练数据的RC墩柱力学性能KGNN分析模型。本文提出的KGNN分析方法与纯数据驱动神经网络（BPNN）的结果对比表明，尽管BPNN在测试集上表现更好，在分析墩柱承载力时均方误和相关系数分别为0.070和0.978，而KGNN模型结果为0.108和0.942，但BPNN所预测的墩柱特征对承载力的影响规律与经验知识并不吻合，亦即未能准确反映墩柱特征与其力学性能间的关系，发生了过拟合。而该KGNN方法不仅可以快速准确获得RC墩柱力学性能，且预测规律与经验知识吻合较好，具有更高的可靠性和实用性。     

广义回归神经网络在煤灰熔点预测中的应用

为了提高估算煤灰熔点的精度,采用广义回归神经网络(GRNN)对求解煤灰熔点问题进行了建模.将煤灰组分作为网络输入,煤灰软化温度作为网络输出,采用实验数据训练网络,训练完成的网络作为模型预测煤灰熔点.仿真结果表明,GRNN的预测值与实验值的最大相对误差为2.81%,而反向传播神经网络(BPNN)预测煤灰熔点的相对误差为3.62%.由于GRNN可应用于小样本问题的学习,GRNN比BPNN对煤灰熔点具有更好的预测和泛化能力.GRNN具有设计简单与收敛快的优点,并提高了实时处理与反映最新运行工况参数的预测能力.     

露天磷矿数字配矿模型

露天磷矿选矿厂采用合理的配矿方案,对稳定入选磷矿的品位,增加选矿产率,降低生产成本有着重要的意义.为了提高入选磷矿石品位的均匀性和稳定性,提出了针对以入选磷矿品位为目标的从采矿场配矿到精矿仓配矿的整个生产环节的配矿方法.将采场内原矿、选矿厂原矿矿堆和选矿厂精矿仓均定义为矿堆,按照生产实际需求设定其属性.通过设定目标函数,实现使采矿场的开采原矿量最大,配矿方案总的运输距离最小,运输成本最低等目标.同时建立约束条件,使采矿场的开采原矿量要保证选矿厂正常生产,使每一矿堆的品位达到目标要求.最后,采集企业生产实际数据对模型加以验证,证明模型在指导露天磷矿山的生产过程中有重要的实际意义,配矿过程应成为矿山生产组织、管理的一个重要环节.     

遗传神经网络在白云鄂博铁矿爆堆质量预测系统中的应用

简要介绍了遗传算法优化BP神经网络的原理和方法，指出金属矿露天采场爆破后爆堆质量预测系统中遗传神经网络设计的基本思想和设计结果，并以包头钢铁(集团)公司白云鄂博铁矿为例验证了所设计的遗传神经网络的精确性和可行性。     

小铁山矿分段分条充填法采场结构参数的三维有限元分析

采用三维有限元数值分析方法,结合小铁山矿分段充填法实际,对不同有场结构参数条件下矿体应力分布、能量释放和采场体积闭合进行了计算,为合理、安全垂采矿体提供了必要的切合实际的结构参数。     

Optimization model of unascertained measurement for underground mining method selection and its application

An optimization model of underground mining method selection was established on the basis of the unascertained measurement theory. Considering the geologic conditions, technology, economy and safety production, ten main factors influencing the selection of mining method were taken into account, and the comprehensive evaluation index system of mining method selection was constructed. The unascertained evaluation indices corresponding to the selected factors for the actual situation were solved both qualitatively and quantitatively. New measurement standards were constructed. Then, the unascertained measurement function of each evaluation index was established. The index weights of the factors were calculated by entropy theory, and credible degree recognition criteria were established according to the unascertained measurement theory. The results of mining method evaluation were obtained using the credible degree criteria, thus the best underground mining method was determined. Furthermore, this model was employed for the comprehensive evaluation and selection of the chosen standard mining methods in Xinli Gold Mine in Sanshandao of China. The results show that the relative superiority degrees of mining methods can be calculated using the unascertained measurement optimization model, so the optimal method can be easily determined. Meanwhile, the proposed method can take into account large amount of uncertain information in mining method selection, which can provide an effective way for selecting the optimal underground mining method.     

改进PSO-BPNN的电力变压器故障诊断与模式识别

为了优化反向传播网络相关学习参数,提出一种粒子群优化辅助BP神经网络(BPNN)的新方法.以变压器油中气体体积分数百分比构造故障特征,将BP网络的初始权值和阈值进行实数编码,以对应PSO中的粒子,实现BP网络的离线训练与在线分析,对变压器故障模式做出判断.结果表明,该算法更合理地更新了粒子的位置和速度,最优地设置了全局极值,有效克服了粒子的早熟收敛,获得的故障诊断准确率高达91%,并大大提升了BP网络的收敛速度.该算法为此类设计提供了有效的模型参考.     

露天矿物料流规划系统的设计与实现

根据露天矿采剥物料的运移特点，排土场发展规划和矿山工程发展规律，提出了建立露天矿运输系统动态网络模型的基本思路与方法，运用网络技术和线性规划方法。对露天矿不同时期，不同阶段的矿岩进行最佳分析计算，使矿山的运输功最小，实现露天矿采剥物料流向流量的优化，同时给出了系统的设计流程的输出结果。     

基于FNN的智能学生综合素质评估模型研究

对学生综合素质评估体系相关要素进行了讨论，给出了学生综合素质评估体系闭环控制系统结构。提出了一种基于模糊神经网络的智能学生综合素质评估体系模型，该模型通过模糊神经网络的自适应与自学习能力，对系统中非量化因素指标的量化进行了有效的刻画，并采用遗传算法对网络结构与量化指标权重进行了优化。仿真结果证明了提出的模型不仅准确有效，而且收敛速度快。