基于支持向量机决策树的电梯载重建模研究

Study on the modeling of elevator load based on support vector machine decision tree

电梯能效评估需要考虑电梯载重影响,常用电梯载重仅提供超载提示,无连续载重测量的数据输出接口。图像处理方法可用于轿厢内人数统计,但仍无法获得准确的电梯载重量。为了能够估算载重量,利用支持向量机决策树(Support Vector Machine Decision Tree,SVMDT)分类的数据挖掘方法,对海量离线曳引机电压数据进行挖掘,建立电梯载重量与电梯母线电压之间的非线性模型,通过实验对模型进行验证,所得结果的准确性超过83. 3%。利用估算出的载重量计算电梯的能效评定指标,对比不同电梯的能效评定指标可以得到它们之间的能耗及能效情况。

The evaluation of the elevator energy efficiency has to consider the impact of the elevator load. The elevators have no interface of weight measurement and can only provide overload warning. The image procession can be used to get the number of passengers in the car,but it is still not available to get the exact load. In order to estimate the load,the data mining of Support Vector Machine Decision Tree( SVMDT) classification was used for the mining of massive off-line tractor voltage data. The nonlinear model between the weight and the elevator load voltage was established to perform the experiment to verify the model. The accuracy of the model was more than 83. 3 percent. Based on the estimated load,the energy efficiency evaluation index of the elevator was calculated. The energy consumption and energy efficiency of different elevators can be obtained by comparing the energy efficiency index.