摘要
为准确、快速地预测煤与瓦斯突出,提出了一种基于因子分析和遗传算法(GA)优化极限学习机(ELM)的煤与瓦斯突出危险性预测模型。构建10种影响因素的煤与瓦斯突出评价指标体系,采用因子分析法对评价指标体系进行分析提取,将提取出的5个公因子作为ELM的输入参数,为避免ELM输入权值和隐含层偏差随机性的影响,应用GA对ELM模型参数进行优化,构建GA-ELM模型,选取20组实例进行仿真预测,同时与传统单一的ELM、SVM和BP模型进行对比分析。结果表明:基于改进的GA-ELM模型能有效降低数据冗余、简化网络结构和提高判别精度,提出将其运用到煤与瓦斯突出的预测与实际结果具有很好的一致性。
To predict coal and gas outburst accurately and rapidly,a coal and gas outburst danger prediction model based on factor analysis and genetic algorithm(GA) to optimize the extreme learning machine(ELM) is proposed. In this paper,the evaluation index system of 10 influencing factors were constructed,and then were analysed and extracted by the factor analysis method. 5 extracting common factors were taken as input parameters of ELM. To avoid the influence of the randomness about the ELM input weight and implicit layer deviation,it applied GA to optimize the ELM model and built the GA-ELM mode. Then,20 case of the coal and outburst danger were simulated and forecasted by the mode. At the same time,the prediction results were also compared with the results of traditional single ELM,SVM and BP model. The results show that: the optimized GA-ELM model can effectively reduce data redundancy,simplify the network structure and improve the discrimination accuracy. It is proposed that the prediction model of coal and gas outburst has a good agreement with the actual results.
作者
韩永亮
李胜
胡海永
罗明坤
Han Yongliang;Li Sheng;Hu Haiyong;Luo Mingkun(Xi’an Research Institute,China Coal Technology&Engineering Group Corp,Xi’an 710054,P.R.China;College of Mining Engineering,Liaoning Technical University,Fuxin,Liaoning 123000,P.R.China;Information Research Institute of the Ministry of Emergency Management,Beijing 100029,P.R.China)
出处
《地下空间与工程学报》
CSCD
北大核心
2019年第6期1895-1902,共8页
Chinese Journal of Underground Space and Engineering
基金
国家自然科学基金(51004063)
辽宁省高等学校优秀人才支持计划(LJQ2011029)
