基于ISABO-SVM的冲击地压危险等级预测
Prediction of rock burst risk level based on ISABO-SVM
李忠勤 1刘赵龙1
作者信息
- 1. 黑龙江科技大学 电气与控制工程学院,哈尔滨 150022
- 折叠
摘要
为有效提高冲击地压危险等级预测的准确率,建立了基于改进的减法平均算法(ISA-BO)优化支持向量机(SVM)的冲击地压危险等级预测模型.通过提取声发射信号的上升时间和绝对能量等5 个时域特征参数,结合煤岩体抗压强度,构建了冲击地压预测的特征参数数据库;根据特征参数数据库计算每个样本的综合危险指数,将冲击地压危险性划分为四个等级.以特征参数数据库和危险等级为模型输入和输出,通过ISABO优化SVM预测冲击地压危险等级过程中的核参数和惩罚因子.结果表明,与传统SVM模型以及减法平均算法(SABO)优化后的SVM模型相比,建立的ISABO-SVM模型预测准确率提高至98%.
Abstract
This paper intends to improve the prediction accuracy of rock burst risk level by establis-hing an improved subtraction average algorithm to optimize the support vector machine prediction model.By extracting five time-domain characteristic parameters including rising time and absolute energy of a-coustic emission signals,and combined with the compressive strength of coal and rock mass,the study is performed by constructing the characteristic parameter database of rock burst prediction;calculating the comprehensive risk index of each sample according to the characteristic parameter database,and dividing the risk of rock burst into four grades;taking the characteristic parameter database and the risk level as the model input and output;and predicting the kernel parameters and penalty factors in the process of the risk level of rock burst by optimizing SVM with ISABO.The results show that compared with the tradi-tional SVM model and the SVM model optimized by the subtraction average algorithm,the prediction ac-curacy of the established ISABO-SVM model is improved to 98%.
关键词
冲击地压/声发射/抗压强度/减法平均算法/支持向量机Key words
rock burst/acoustic emission/compressive strength/subtraction-average-based optimi-zer/support vector machine引用本文复制引用
基金项目
黑龙江省省属高等学校基本科研业务费科研项目(2020-KYYWF-06867)
黑龙江省揭榜挂帅科技攻关项目(2021ZXJ02A02)
出版年
2024
国家科技期刊平台
NSTL
万方数据