支持向量机建模的二维滑坡模型控制系统研究

滑坡模型作为研究滑坡地质灾害的重要实验平台是一个装载一定量沙石土壤的箱体.二维滑坡模型是在一定模拟降雨条件下通过抬升装载有砂石土料的箱体实现模拟山体滑坡行为.基于对抬升过程中箱体运动规律的分析,设计了一个智能液压控制系统,通过控制阀门开启时间实现了箱体抬升角度的精确控制.采用了最小二乘支持向量机滚动窗算法实现滑坡模型实验箱抬升过程进行动态建模和预测控制,实现了模型平台抬升过程中的倾角检测与控制.本控制系统避免了低速大负载条件下液压系统的脉动和爬行对滑坡模型的影响.实际运行表明本系统满足滑坡模型的平台倾角精度要求.

Research on Two-Dimensional Landslide Model Control System Based on Support Vector Machine Modeling

Landslide model,an important experiment platform for studying geologic hazard,is a cabinet loaded by a certain volume of the gravel earth material.Two-dimensional model simulates landslide behavior by uplifting the experiment cabinet at certain angle under simulated rainfall.An intelligent hydraulic control system which by controlling the valve opening time to achieve a precise control of the angle of the box uplift is presented based on motion analysis of the cabinet during uplift.Thanks to predictive control algorithm of least squares support vector machine (LS-SVM) model,random real-time change of equivalent load applied to the hydraulic cylinder system is effectively solved.The control system effectively eliminates effect of crawling and pulse lifting on two-dimensional landslide model in the condition of low speed and high load of hydraulic system.The actual operation shows that this system meets the accuracy requirements of the platform inclination landslide model.