程国森 崔东文

摘 要：為提高煤层顶底板地层富水性分级判定精度，研究提出黑猩猩优化算法（ChOA）与极限学习机（ELM）相融合的判定方法。选取4个标准测试函数在不同维度条件下对ChOA的寻优能力进行仿真验证，仿真结果与粒子群优化（PSO）算法、人工蜂群（ABC）算法作对比;基于煤层顶底板地层富水性判定因子和判定分级构建ELM模型，利用ChOA优化ELM输入层权值和隐含层偏值，建立ChOA-ELM富水性分级判定模型，并构建ChOA-SVM、ChOA-BP作对比模型，通过龙固煤层顶底板地层富水性分级判定实例对ChOA-ELM、ChOA-SVM、ChOA-BP模型进行检验。结果表明：①ChOA在不同维度条件下寻优效果优于PSO、ABC算法，具有较好的寻优精度和全局搜索能力;②ChOA-ELM模型对实例训练样本和检验样本富水性分级判定准确率分别为97.5%、100%，高于ChOA-SVM、ChOA-BP模型，具有较好的判定精度和泛化能力;③ChOA能有效优化ELM输入层权值和隐含层偏值，将ChOA-ELM用于煤层顶底板地层富水性分级判定是可行的，模型及ELM权值、偏值优化方法可为相关判定研究提供参考。

关键词：富水性;分级判定;极限学习机;黑猩猩优化算法;仿真验证;参数优化

中图分类号：TV213.4;P641.5+4 文献标志码：A

doi：10.3969/j.issn.1000-1379.2021.07.012

引用格式：程国森，崔东文.黑猩猩优化算法-极限学习机模型在富水性分级判定中的应用[J].人民黄河，2021，43（7）：62-66，103.

Abstract： In order to improve the classification accuracy of the water-rich stratum in the roof and floor of the coal seam， this paper proposed a judgment method combining the chimpanzee optimization algorithm （ChOA） and the extreme learning machine （ELM）. It chose 4 standard test functions to simulate and verify the optimization ability of ChOA under different dimensional conditions. The simulation results were compared with the particle swarm optimization （PSO） algorithm and the artificial bee colony （ABC） algorithm. It based on the determination of the water richness of the roof and floor of the coal seam factor and decision grading to build an ELM model， used ChOA to optimize ELM input layer weights and hidden layer bias， established a ChOA-ELM water-rich grading decision model， and built ChOA-SVM and ChOA-BP as comparison models. The example of determining the water richness of the roof and floor strata was to test the ChOA-ELM， ChOA-SVM and ChOA-BP models. The results show that： a） ChOA has better optimization results than PSO and ABC algorithms under different dimensional conditions， and has better optimization accuracy and global search capabilities; b） ChOA-ELM model determines the water content of instance training samples and test samples. The accuracy rates are respectively 97.5% and 100% respectively， which are higher than that of the ChOA-SVM and ChOA-BP models， and have better judgment accuracy and generalization ability; c） ChOA can effectively optimize ELM input layer weights and hidden layer bias values. It is feasible to use ChOA-ELM to determine the water richness of coal roof and floor formations. The model， ELM weights and bias value optimization methods can provide references for related judgments research.

Key words： water abundance; grading judgment; extreme learning machine; chimp optimization algorithm; simulation; parameter optimization

1 研究背景

科学合理判定煤层顶底板地层富水性分级，对于全面掌握煤层顶底板地层富水性分布情况，科学提出风险预防措施，减轻或消除矿井突水隐患以及有效保护地下水资源均具有重要意义。煤层顶底板地层富水性分级判定方法根据数据来源一般分为物探法和多因素综合分析法，物探法存在工作量大、费用高且控制范围有限等问题，而多因素综合分析法应用较为广泛[1]。……