基于樽海鞘群优化网络模型的计量装置状态识别方法

State identification method of metering device based on salps optimal network model

为提升用电信息采集系统客户服务部门的数据分析能力,提升客户服务精益化管理水平,针对计量装置状态远程识别准确率低的问题,提出了一种基于樽海鞘群优化网络模型的计量装置状态识别方法。该模型首先利用小波分解对样本集曲线类特征数据进行分解,并获取状态影响因子,然后利用LSTM长短期记忆网络进行计量装置状态分类,并计算损失函数后进行反馈调参,最后采用樽海鞘群算法优化LSTM网络的调参过程,待损失函数低于阈值后,固定参数输出模型。实验表明,樽海鞘群算法寻优调参可降低模型参数调参时间,并提高了分类算法的精准度。

In order to improve the data analysis ability of the customer service department of the electricity consumption information acquisition system and improve the lean management level of customer service,a measurement device state based on the optimization network model of the salps group is proposed to solve the problem of low accuracy of remote identification of the state of the metering device.The model first uses wavelet decomposition to decompose the curve-like characteristic data of the sample set,and obtains the state influence factor,then uses the LSTM long short-term memory network to classify the state of the metering device,and calculates the loss function for feedback adjustment.The algorithm optimizes the parameter adjustment process of the LSTM network.After the loss function is lower than the threshold,the fixed parameters are output to the model.Experiments show that the optimization and tuning of the salps swarm algorithm can reduce the time for parameter tuning of the model parameters and improve the accuracy of the classification algorithm.