基于PSO-LSSVM算法的地下储气库洞室群爆破振动预测

超大断面小净距地下储气库洞室群开挖爆破工程中涉及到众多的影响因素,传统人工智能方法难以对爆破峰值振动速度准确预测。为了提高预测精度,引入粒子群算法,对传统的最小二乘支持向量机模型(LS-SVM)进行优化并建立粒子群最小二乘支持向量机爆破峰值振动速度预测模型(PSO-LSSVM)。以某地下储气库洞室群开挖爆破工程为研究对象,应用PSO-LSSVM模型,将PSO-LSSVM模型与LS-SVM模型、萨道夫斯基经验公式的预测结果进行对比,得到三种预测的结果平均绝对相对误差分别为:5.50%、8.56%、23.45%。由此可见,PSO-LSSVM模型的预测结果与实测数据拟合度更高,精确度更满足工程需求,可为多因素作用下类似工程爆破峰值振动速度预测提供借鉴。

Blasting Vibration Prediction of Underground Gas Storage Caverns based on PSO-LSSVM Algorithm

There are many influencing factors on the excavation blasting engineering of the underground gas stor-age caverns with large cross-section and small clear distance,and it is difficult for traditional artificial intelligence method to accurately predict the peak blasting vibration velocity. In order to improve the prediction accuracy,the par-ticle swarm algorithm was used to optimize the traditional least squares support vector machine model and to establish the particle swarm least squares support vector machine blasting peak vibration velocity prediction model. By compa-ring with the forecasted result,the average absolute relative error of three models including PSO LSSVM,LS-SVM and Sodev′s empirical formula were 5. 50%,8. 56%,23. 45%,respectively. Therefore,the prediction results of PSO LSS-VM model have higher fitting degree with the measured data,and the accuracy is in line with the engineering require-ments,which provides reference to predict the blasting vibration peak velocity of the similar projects under the muti-factor influence.