基于LSTM网络的住宅负荷短期预测
|
摘要
在智能电网中,若未考虑海量的住宅负荷和气象数据的相关性,就会导致输入信息冗余,为此提出一种长短期记忆(long short-term memory,LSTM)网络的住宅负荷短期预测方法。该方法首先利用LSTM网络对负荷数据、日期类型、气象数据进行动态建模,然后采用主成分分析对气象等数据进行特征选择以过滤掉数据间的冗余信息,最后使用自适应矩估计(adaptive moment estimation, ADAM)算法优化后的LSTM网络参数提高模型的泛化能力。采用美国马萨诸塞州某小区公寓实测数据进行短期负荷预测,结果验证了所提方法的有效性和实用性。
Input information redundancy is caused if without considering correlation of massive residence load and weather data in the smart grid. Therefore, this paper proposes a kind of short-term residence load forecast method based on long short-term memory(LSTM) network. This model firstly makes use of the LSTM network for dynamical modelling for load data, date types and weather data, and then it adopts the principal component analysis method to select data and filter out redundant information. Finally it uses the adaptive moment estimation(ADAM) algorithm to optimize parameters of the LSTM network so as to improve generalization ability of the model. Short-term load forecast was carried out by using measured data of a residential apartment in Massachusetts, USA. The results verify effectiveness and practicability of the proposed method.
Input information redundancy is caused if without considering correlation of massive residence load and weather data in the smart grid. Therefore, this paper proposes a kind of short-term residence load forecast method based on long short-term memory(LSTM) network. This model firstly makes use of the LSTM network for dynamical modelling for load data, date types and weather data, and then it adopts the principal component analysis method to select data and filter out redundant information. Finally it uses the adaptive moment estimation(ADAM) algorithm to optimize parameters of the LSTM network so as to improve generalization ability of the model. Short-term load forecast was carried out by using measured data of a residential apartment in Massachusetts, USA. The results verify effectiveness and practicability of the proposed method.
引文
[1] 杨威,曾智健,陈皓勇,等.广东电力市场需求侧响应交易机制研究[J].广东电力,2017,30(5):25-34.YANG Wei,ZENG Zhijian,CHEN Haoyong,et al.Research on demand response trading mechanism in Guangdong electricity market[J].Guangdong Electric Power,2017,30(5):25-34.
[2] 丁晓,孙虹,郑海雁,等.基于配用电大数据的短期负荷预测[J].电力工程技术,2018,37(3):21-27.DING Xiao,SUN Hong,ZHEN GHaiyan,et al.Distribution and consumption big data based short-term load forecasting[J].Electric Power Engineering Technology,2018,37(3):21-27.
[3] 吴亚雄,谢敏.基于BP神经网络灰色回归组合模型的年最大负荷预测[J].南方能源建设,2017,4(2):46-50,57.WU Yaxiong,XIE Min.Annual peak load forecasting based on combination model of back propagation neural network and grey regression[J].Southern Energy Construction,2017,4(2):46-50,57.
[4] 梁智,孙国强,卫志农,等.基于变量选择与高斯过程回归的短期负荷预测[J].电力建设,2017,38(2):122-128.LIANG Zhi,SUN Guoqiang,WEI Zhinong,et al.Short term load forecasting based on variable selection and Gaussian process regression[J].Electric Power Construction,2017,38(2):122-128.
[5] 方八零,李龙,赵家铸,等.动态相似与静态相似相结合的短期负荷预测方法[J].电力系统保护与控制,2018,46(15):29-35.FANG Baling,LI Long,ZHAO Jiazhu,et al.Short-term load forecasting based on the combination of dynamic similarity and static similarity[J].Power System Protection and Control,2018,46(15):29-35.
[6] 刘建华,李锦程,杨龙月,等.基于EMD-SLSTM的家庭短期负荷预测[J].电力系统保护与控制,2019,47(6):40-47.LIU Jianhua,LI Jincheng,YANG Longyue,et al.Short-term household load forecasting based on EMD-SLSTM[J].Power System Protection and Control,2019,47(6):40-47.
[7] HOCHREITER S S.Long short-term memory[J].Neural Computation,1997,9(8):1735-1742.
[8] GERS F A,SCHMIDHUBER J,CUMMINS F.Learning to forget:continual prediction with LSTM[J].Neural Computation,2000,12(10):2451-2471.
[9] QING Xiangyun.Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM[J].Energy,2018(10):148-157.
[10] 钟惠锋.提高电网短期负荷预测精度的研究[J].广东电力,2011,24(6):97-100.ZHONG Huifeng.Study on promoting short-term load forecasting accuracy of power grid[J].Guangdong Electric Power,2011,24(6):97-100.
[11] 嵇灵,牛东晓,吴焕苗.基于贝叶斯框架和回声状态网络的日最大负荷预测研究[J].电网技术,2012,36(11):82-86.JI Ling,NIU Dongxiao,WU Huanmiao.Daily peak load forecasting based on Bayesian framework and echo state network[J].Power System Technology,2012,36(11):82-86.
[12] 王文锦,戚佳金,王文婷,等.基于人工蜂群优化极限学习机的短期负荷预测[J].电测与仪表,2017,54(11):32-35.WANG Wenji,QI Jiajin,WANG Wenting,et al.Short-term load forecasting based on improved extreme learning machine with artificial bee colony algorithm[J].Electrical Measurement & Instrumentation,2017,54(11):32-35.
[13] 黄宇腾,韩翊,赖尚栋.深度神经网络在配电网公变短期负荷预测中的应用研究[J].浙江电力,2018,37(5):1-6.HUANG Yuteng,HAN Yi,LAI Shangdong.Application of deep neural network in short-term load prediction of public transformer of power distribution network[J].Zhejiang Electric Power,2018,37(5):1-6.
[14] 李琪,戚浩金,胡一嗔.基于主成分分析和集对分析理论的配电网项目群投资决策[J].广东电力,2017,30(4):15-20.LI Qi,QI Haojin,HU Yichen.Investment decision-making for power distribution network project group based on principal component and set pair analysis[J].Guangdong Electric Power,2017,30(4):15-20.
[15] KINGMA D P,BA J.Adam:a method for stochastic optimization[J].Computer Science,2014(4):13-21.
[16] 杨楠,南琳,张丁一,等.基于深度学习的图像描述研究[J].红外与激光工程,2018,2(2):18-25.YANG Nan,NAN Lin,ZHANG Dingyi,et al.Research on image interpretation based on deep learning[J].Infrared and Laser Engineering,2018,2(2):18-25.
[17] MA X,TAO Z,WANG Y,et al.Long short-term memory neural network for traffic speed prediction using remote microwave sensor data[J].Transportation Research Part C,2015(54):187-197.
[18] BENGIO Y,SIMARD P,FRASCONI P.Learning long-term dependencies with gradient descent is difficult[J].IEEE Trans.Neural Netw,2002,5(2):157-166.
[19] 杨子文,曾上游,杨远飞.基于二叉树型卷积神经网络信息融合的人脸验证[J].计算机应用,2017(2):155-159.YANG Ziwen,ZENG Shangyou,YANG Yuanfei.Face verification based on information fusion binary tree convolution neural network[J].Journal of Computer Applications,2017(2):155-159.
[2] 丁晓,孙虹,郑海雁,等.基于配用电大数据的短期负荷预测[J].电力工程技术,2018,37(3):21-27.DING Xiao,SUN Hong,ZHEN GHaiyan,et al.Distribution and consumption big data based short-term load forecasting[J].Electric Power Engineering Technology,2018,37(3):21-27.
[3] 吴亚雄,谢敏.基于BP神经网络灰色回归组合模型的年最大负荷预测[J].南方能源建设,2017,4(2):46-50,57.WU Yaxiong,XIE Min.Annual peak load forecasting based on combination model of back propagation neural network and grey regression[J].Southern Energy Construction,2017,4(2):46-50,57.
[4] 梁智,孙国强,卫志农,等.基于变量选择与高斯过程回归的短期负荷预测[J].电力建设,2017,38(2):122-128.LIANG Zhi,SUN Guoqiang,WEI Zhinong,et al.Short term load forecasting based on variable selection and Gaussian process regression[J].Electric Power Construction,2017,38(2):122-128.
[5] 方八零,李龙,赵家铸,等.动态相似与静态相似相结合的短期负荷预测方法[J].电力系统保护与控制,2018,46(15):29-35.FANG Baling,LI Long,ZHAO Jiazhu,et al.Short-term load forecasting based on the combination of dynamic similarity and static similarity[J].Power System Protection and Control,2018,46(15):29-35.
[6] 刘建华,李锦程,杨龙月,等.基于EMD-SLSTM的家庭短期负荷预测[J].电力系统保护与控制,2019,47(6):40-47.LIU Jianhua,LI Jincheng,YANG Longyue,et al.Short-term household load forecasting based on EMD-SLSTM[J].Power System Protection and Control,2019,47(6):40-47.
[7] HOCHREITER S S.Long short-term memory[J].Neural Computation,1997,9(8):1735-1742.
[8] GERS F A,SCHMIDHUBER J,CUMMINS F.Learning to forget:continual prediction with LSTM[J].Neural Computation,2000,12(10):2451-2471.
[9] QING Xiangyun.Hourly day-ahead solar irradiance prediction using weather forecasts by LSTM[J].Energy,2018(10):148-157.
[10] 钟惠锋.提高电网短期负荷预测精度的研究[J].广东电力,2011,24(6):97-100.ZHONG Huifeng.Study on promoting short-term load forecasting accuracy of power grid[J].Guangdong Electric Power,2011,24(6):97-100.
[11] 嵇灵,牛东晓,吴焕苗.基于贝叶斯框架和回声状态网络的日最大负荷预测研究[J].电网技术,2012,36(11):82-86.JI Ling,NIU Dongxiao,WU Huanmiao.Daily peak load forecasting based on Bayesian framework and echo state network[J].Power System Technology,2012,36(11):82-86.
[12] 王文锦,戚佳金,王文婷,等.基于人工蜂群优化极限学习机的短期负荷预测[J].电测与仪表,2017,54(11):32-35.WANG Wenji,QI Jiajin,WANG Wenting,et al.Short-term load forecasting based on improved extreme learning machine with artificial bee colony algorithm[J].Electrical Measurement & Instrumentation,2017,54(11):32-35.
[13] 黄宇腾,韩翊,赖尚栋.深度神经网络在配电网公变短期负荷预测中的应用研究[J].浙江电力,2018,37(5):1-6.HUANG Yuteng,HAN Yi,LAI Shangdong.Application of deep neural network in short-term load prediction of public transformer of power distribution network[J].Zhejiang Electric Power,2018,37(5):1-6.
[14] 李琪,戚浩金,胡一嗔.基于主成分分析和集对分析理论的配电网项目群投资决策[J].广东电力,2017,30(4):15-20.LI Qi,QI Haojin,HU Yichen.Investment decision-making for power distribution network project group based on principal component and set pair analysis[J].Guangdong Electric Power,2017,30(4):15-20.
[15] KINGMA D P,BA J.Adam:a method for stochastic optimization[J].Computer Science,2014(4):13-21.
[16] 杨楠,南琳,张丁一,等.基于深度学习的图像描述研究[J].红外与激光工程,2018,2(2):18-25.YANG Nan,NAN Lin,ZHANG Dingyi,et al.Research on image interpretation based on deep learning[J].Infrared and Laser Engineering,2018,2(2):18-25.
[17] MA X,TAO Z,WANG Y,et al.Long short-term memory neural network for traffic speed prediction using remote microwave sensor data[J].Transportation Research Part C,2015(54):187-197.
[18] BENGIO Y,SIMARD P,FRASCONI P.Learning long-term dependencies with gradient descent is difficult[J].IEEE Trans.Neural Netw,2002,5(2):157-166.
[19] 杨子文,曾上游,杨远飞.基于二叉树型卷积神经网络信息融合的人脸验证[J].计算机应用,2017(2):155-159.YANG Ziwen,ZENG Shangyou,YANG Yuanfei.Face verification based on information fusion binary tree convolution neural network[J].Journal of Computer Applications,2017(2):155-159.