基于PSO-SVR算法的悬臂式掘进机工作性能预测  

Performance Prediction of Roadheader by PSO-SVR Algorithm

作　　者：傅鹤林[1] 赵一博 王立志 郭弘宇 李鲒 邓皇适 FU Helin;ZHAO Yibo;WANG Lizhi;GUO Hongyu;LI Jie;DENG Huangshi(Central South University,Changsha,Hunan 410009,China;China Railway First Survey and Design Institute Group Co.Ltd,Xi'an,Shaanxi 710043,China;CCFEB Civil Engineering Co.Ltd,Changsha,Hunan 410000,China)

机构地区：[1]中南大学,长沙410009 [2]中铁第一勘察设计院集团有限公司,西安710043 [3]中建五局土木工程有限公司,长沙410000

出　　处：《铁道工程学报》2024年第9期92-98,共7页Journal of Railway Engineering Society

基　　金：湖南省建设科技计划项目(KY202109)。

摘　　要：研究目的:悬臂式掘进机作为铣挖法中不可或缺的机械设备,其工作性能常受到围岩特征和机械设备等条件的制约。为预测悬臂式掘进机工作性能,本文以瞬时切割速率ICR作为评价指标,综合考虑围岩因素和机械设备因素,通过不同算法预测ICR,优选出基于粒子群优化的支持向量机算法(PSO-SVR),并据此建立了ICR预测系统,从而快速高效地预测铣挖隧道掘进机工作性能和施工速度。研究结论:(1)利用已开挖的铣挖法隧道数据作为训练样本,通过5种算法预测ICR,其中粒子群优化的支持向量机算法预测精度最佳;(2)通过粒子群优化算法搜寻最优惩罚系数C和核函数系数g,能有效避免陷入局部最优解问题,并显著提高模型预测精度和泛化能力;(3)十折交叉验证结果表明,PSO-SVR算法具有较好的鲁棒性,且鲁棒性高于其他4种算法;(4)依托赣州蓉江隧道工程,PSO-SVR模型可准确预测悬臂式掘进机瞬时切割速率和施工速度,预测精度显著高于经验公式和其他算法,可为铣挖机械设备选型以及铣挖隧道施工速度预测提供参考。Research purposes:The roadheader works as an indispensable mechanical equipment in the milling and excavation method.Its working performance is often constrained by the characteristics of surrounding rock and mechanical equipment conditions.To predict the working performance of roadheader,this paper takes the instantaneous cutting rate(ICR)as the evaluation index,comprehensively considers the surrounding rock factors and mechanical equipment factors,and establishes an ICR prediction system based on different algorithms.The support vector machine algorithm based on particle swarm optimization(PSO-SVR)is optimized,which supported the ICR prediction system.This system can efficiently predict the working performance and excavating speed of tunnels excavated by roadheader.Research conclusions:(1)The data of tunnels excavated by roadheader are used for the training samples of 5 different algorithms.The PSO-SVR algorithm has the best prediction accuracy.(2)The particle swarm optimization is used to search the optimal penalty coefficient C and kernel function coefficient g,which can effectively avoid falling into local optimal solution and significantly improve the prediction accuracy and generalization ability of the model.(3)The ten-fold cross validation results indicate that the PSO-SVR model has better robustness than that of the other four algorithms.(4)Based on the Ganzhou Rongjiang Tunnel,the ICR and excavating speed of the roadheader is accurately predicted by PSO-SVR model,whose prediction accuracy is significantly higher than the empirical formula and other four algorithms,providing reference for the selection of roadheader and the prediction of excavating speed of roadheader.

关 键 词：悬臂式掘进机 瞬时切割速率 工作性能预测 粒子群算法 支持向量机 

分 类 号：U455.6[建筑科学—桥梁与隧道工程]