基于因子分析的PSO-ELM岩溶发育预测模型研究

Study on prediction model of karst development based on factor analysis and PSO-ELM

岩溶发育是由多个影响因素共同作用的结果,具有成因复杂、隐蔽性强等特点,对地面设施和人员安全构成了潜在威胁。对岩溶发育进行评估及预测,可以在岩溶致灾前采取预防措施,减少岩溶灾害带来的损失。以武汉市某岩溶区工程为例,对岩溶区溶洞的赋存规律进行分析,确定地下稳定水位埋深、覆盖层厚度、基岩层数等9个影响因素,利用因子分析提取5个公因子,提出了一种在因子分析的基础上利用粒子群算法(PSO)优化极限学习机(ELM)的岩溶发育预测模型,利用现场的60组实测数据作为样本进行学习预测,以其中50组作为训练集,其余10组作为测试集,对预测模型的精度进行验证,结果表明:PSO-ELM预测模型的预测值与实际值吻合较好。将PSO-ELM预测模型与ELM预测模型的预测结果进行对比分析,发现PSO-ELM预测模型的精度更高。

Karst is the result from the joint action of many factors,which is characterized by complex formation and good concealment,and poses a threat to the safety of ground facilities and personnel.The assessment and prediction of karst development are expected to take precautions to minimize any losses caused by karst ahead of the disasters.Taking a project in a karst area of Wuhan city as a reference,the occurrence of karst caves in the karst region was studied,while nine factors such as depth of stable underground water level,thickness of overburden and number of bedrock layers were determined,five common factors were extracted for factor analysis.As a result,a prediction model of karst development based on factor analysis and optimized ELM by particle swarm optimization algorithm(PSO)was established,by which 60 groups of data measured from the project site were utilized as samples for learning prediction,among which 50 groups were used for training,and the rest 10 groups were used for testing to verify the accuracy of the prediction model.The results show that the predicted results of the PSO-ELM model was in good agreement with the facts.From the comparative analysis,we can see that the prediction results of PSO-ELM model are of higher accuracy than PSO-ELM model.