基于模型迁移和SVM的煤矿电网单相接地电容电流预测方法

Prediction method of single-phase grounded capacitance current in coal mine power grid based on model migration and SVM

针对智慧化矿山背景下煤矿电网单相接地电容电流预测精度低和智能预测实现困难的问题,提出了一种单相接地电容电流智能预测方法。采用模型迁移思想扩充矿用橡套软电缆的电容电流样本数据,以解决智能预测中电容电流数据不完备的问题;基于Sobol’敏感性方法分析了影响单相接地电容电流的关键因素及其各因素的交互关系,确定智能预测模型输入特征量;在此基础上,通过基于稀疏技术的支持向量机建立煤矿电网单相接地电容电流智能预测模型,并引入鲸鱼算法优化预测模型超参数,克服了电容电流数据样本容量小的不足。采用厂家测试数据及多家煤矿电网现场实测数据进行试验,结果表明:电缆绝缘层外径和绝缘层内径是影响单相接地电容电流的关键因素;所提方法预测煤矿电网单相接地电容电流的平均误差为2.26%,相较于现有预测方法,误差分别下降了34.19%,24.91%和7.40%。该方法实现了煤矿电网单相接地电容电流的准确预测,并为其智能化预测提供了新思路。

In response to the challenges of low prediction accuracy and difficulties in implementing intelligent prediction for the capacitive current of the single-phase grounded electrical network in the context of intelligent mining,this paper proposes an intelligent prediction method for the capacitive current of single-phase grounded electrical networks.For the problem of incomplete capacitance current data in intelligent prediction,the model transfer is used to expand the sample data of rubber sheathing flexible cable used in mining.Based on Sobol’sensitivity method,the key factors affecting the single-phase grounding capacitance current and their interaction are analyzed,and the input characteristic of the intelligent prediction model is determined.And the support vector machine based on sparse technology is used to establish the intelligent prediction model of single-phase grounding capacitor current in coal mine power grid,and whale algorithm is introduced to optimize the prediction model’s hyperparameters,which overcomes the shortage of small sample size of capacitor current data.The proposed method is experimentally tested using test data from manufacturers and field measurements from multiple coal mine electrical networks.The results indicate that the outer diameter and inner diameter of the cable insulation layer are the key factors affecting the single-phase grounding capacitance current.The average error of the proposed method is 2.26%,which is 34.19%,24.91%and 7.40%lower than that of the existing method.The method presented in this paper can effectively improve the prediction accuracy of single-phase grounded capacitance current in coal mine power grid,and provide a new idea for its intelligent prediction.