摘要
光功率是衡量电力光纤通信质量的一种重要的指标。针对光功率的这一特性,提出了一种自回归移动平均-长短期记忆网络(ARMA-LSTM)光功率预测方法,该方法通过db5小波对原始光功率数据进行一层分解获得光功率数据趋势项部分和细节项部分。采用LSTM对趋势项进行建模预测,采用ARMA对细节项进行建模预测并将两种预测结果相加得到最终的预测结果。实验结果表明:该预测模型比传统预测模型的预测精度更高。
Optical power is an important index to measure the quality of power optical fiber communication. For this characteristic of optical power, an auto-regressive moving average and long-term memory network (ARMA-LSTM) optical power prediction method is proposed, this method decomposes the original optical power data layer by db5 wavelet to obtain the trend term and detail term of the optical power data. LSTM is used to model and predict trend items and ARMA is used to model and predict detailed items, and the predicted results are added together to obtain the final predicted results. The experimental results show that the prediction accuracy of the model is higher than that of the traditional prediction model.
作者
付浩
王圣达
姜山
康爱民
FU Hao;WANG Shengda;JIANG Shan;KANG Aimin(Changchun University of Science and Technology, Electronic and Information Engineering, Changchun 130022, China;Information and Communication Company, State Grid Jilin Electric Power Co., Ltd., Changchun 130021, China;Fuxin Power Supply Company, State Grid Liaoning Electric Power Co., Ltd., Fuxin Liaoning 123000, China)
出处
《光通信技术》
北大核心
2019年第1期20-23,共4页
Optical Communication Technology
基金
长春市科技局项目(17DY030)资助
作者简介
付浩(1994-),男,河南周口人,硕士研究生,主要从事智能信号检测方向的研究。参与了光纤在线检测与智能预警方法的研究和双模式城区电力光纤支撑网在线检测研究两个项目;在校期间,获得了“基于温度和电阻的光缆故障预测方法”和“一种光缆接头盒密闭性在线监测系统”两项专利。
