M-CM-GA-BP算法的地表移动变形参数预测模型
Prediction model of surface movement and deformation paramete based on M-CM-GA-BP algorith
秦忠诚 1高广慧 1李晓禾 1席天乐1
作者信息
- 1. 山东科技大学 能源与矿业工程学院,山东 青岛 266590
- 折叠
摘要
针对复杂的开采沉陷预测问题,研究22 个工作面采动地表移动变形参数变化规律,提出了一种基于M-CM-GA-BP算法求取地表移动变形参数的预测模型.通过线性加权组合预测方法和遗传算法优化BP神经网络的权值和阈值,融合多元回归模型来提高地表移动变形参数的求取精度,以地表下沉系数q为例,将该模型与其他预测模型预测性能进行对比分析,验证模型的准确性.结果表明,该模型能够有效地提高地表移动变形参数的预测精度,模型的平均相对误差为1.294、均方根误差为0.013,为地表移动变形参数预测提供了一种可行方法.
Abstract
This paper aims to address the complex mining subsidence prediction problem and propo-ses a prediction model based on M-CM-GA-BP algorithm to obtain surface movement and deformation pa-rameters by studying the variation law behind surface movement and deformation parameters at 22 working faces.The study includes optimizing the weights and thresholds of BP neural network by linear weighted combination prediction method and genetic algorithm;integrating the multiple linear regression model to improve the accuracy of surface movement and deformation parameters;analyzing the prediction perform-ance of the model compared with other prediction models,and to verify the accuracy of the model.The results show that this model can effectively improve the prediction accuracy of surface movement and de-formation parameters.The parameters with average relative error by 1.294 and the root mean square error by 0.013 provide a feasible method for the prediction of surface movement and deformation parameters.
关键词
开采沉陷/BP神经网络/地表移动变形参数/组合模型/参数预测Key words
mining subsidence/BP neural network/surface movement deformation parameters/combination model/parameter prediction引用本文复制引用
基金项目
山东省自然科学基金面上项目(ZR2021ME248)
山东省自然科学基金重点项目(ZR2020KE030)
出版年
2024
国家科技期刊平台
NSTL
万方数据