基于随机森林理论的配电变压器重过载预测
|
摘要
针对使用传统分类器预测配变重过载会因为重过载样本率较低而带来的总正确率很高,重过载预测正确率却很低这一问题,将重抽样与随机森林理论引入分类模型中,构建重抽样-随机森林分类器对配变重过载进行预测。首先将观测中重过载样本和正常样本按照一定的比例进行抽样形成新的子样本,重复上述过程获得大量的新子样本。接着根据随机森林理论构建一系列的分类器,并用新子样本对分类器进行训练,得到分类模型。最终的预测结果由所有分类器预测结果的众数所决定。对山东省某市的配变进行重过载预测,并将上述方法与传统分类器进行比较。结果表明,新方法在预测配变日重过载类型、重过载开始与结束时间、重过载严重程度方面有较高的准确率。
In view of problem that prediction of heavily overloaded distribution transformers with traditional classifier brings higher overall correct rate and lower heavy overload correct rate due to low sampling rate of heavy overload, a resampling and random forest regression method is introduced to classification model to predict heavily overloaded distribution transformers with resampling and random forest regression classification model. Firstly, heavy overload samples and normal samples from observation were randomly sampled to form a new sub sample according to a certain ratio, thus large number of new sub samples were obtained by repeating above process. Then, according to random forest theory, a series of classifiers are constructed, and the classification model is trained with sub samples. Final prediction is decided by all classifiers' prediction results. In comparison of performances between above method and traditional classifiers in prediction of heavily overloaded distribution transformers in Shandong Province, results show that the new method has higher accuracy in predicting heavy overload types, starting and ending times and heavy overload severity.
In view of problem that prediction of heavily overloaded distribution transformers with traditional classifier brings higher overall correct rate and lower heavy overload correct rate due to low sampling rate of heavy overload, a resampling and random forest regression method is introduced to classification model to predict heavily overloaded distribution transformers with resampling and random forest regression classification model. Firstly, heavy overload samples and normal samples from observation were randomly sampled to form a new sub sample according to a certain ratio, thus large number of new sub samples were obtained by repeating above process. Then, according to random forest theory, a series of classifiers are constructed, and the classification model is trained with sub samples. Final prediction is decided by all classifiers' prediction results. In comparison of performances between above method and traditional classifiers in prediction of heavily overloaded distribution transformers in Shandong Province, results show that the new method has higher accuracy in predicting heavy overload types, starting and ending times and heavy overload severity.
引文
[1]郝晓晓.住宅小区用电负荷及配变容量的测算[J].华北电力技术,2011(8):16-18.Hao Xiaoxiao.Measurements and calculation of capacity of distribution transformer and power load in residential block[J].North China Electric Power,2011(8):16-18(in Chinese).
[2]吴潇雨,和敬涵,张沛,等.基于灰色投影改进随机森林算法的电力系统短期负荷预测[J].电力系统自动化,2015,39(12):50-55.Wu Xiaoyu,He Jinghan,Zhang Pei,et al.Power system short-term load forecasting based on improved random forest with grey relation projection[J].Automation of Electric Power Systems,2015,39(12):50-55(in Chinese).
[3]赵莉,候兴哲,胡君,等.基于改进k-means算法的海量智能用电数据分析[J].电网技术,2014,38(10):2715-2720.Zhao Li,Hou Xingzhe,Hu Jun,et al.Improved K-means algorithm based analysis on massive data of intelligent power utilization[J].Power System Technology,2014,38(10):2715-2720(in Chinese).
[4]焦润海,苏辰隽,林碧英,等.基于气象信息因素修正的灰色短期负荷预测模型[J].电网技术,2013,37(3):720-725.Jiao Runhai,Su Chenjun,Lin Biying,et al.Short-term load forecasting by grey model with weather factor-based correction[J].Power System Technology,2013,37(3):720-725(in Chinese).
[5]周佃民,管晓宏.基于神经网络的电力系统短期负荷预测研究[J].电网技术,2002,26(2):10-13.Zhou Dianmin,Guan Xiaohong.A short-term load forecasting system based on bp artificial neural network[J].Power System Technology,2002,26(2):10-13(in Chinese).
[6]王大鹏,汪秉文.基于变权缓冲灰色模型的中长期负荷预测[J].电网技术,2013,37(1):167-171.Wang Dapeng,Wang Bingwen.Medium-and long-term load forecasting based on variable weights buffer grey model[J].Power System Technology,2013,37(1):167-171(in chinese).
[7]戴远航,陈磊,闵勇,等.电网每日运行评价指标体系研究[J].电网技术,2015,39(6):1611-1616.Dai Yuanhang,Chen Lei,Min Yong,et al.Study on the daily operation assessment system of power grid[J].Power System Technology,2015,39(6):1611-1616(in Chinese).
[8]朱利鹏,陆超,孙元章,等.基于数据挖掘的区域暂态电压稳定评估[J].电网技术,2015,39(4):1026-1032.Zhu Lipeng,Lu Chao,Sun Yuanzhang,et al.Data mining based regional transient voltage stability assessment[J].Power System Technology,2015,39(4):1026-1032(in Chinese).
[9]朱建全,李颖,谭伟,等.基于特性融合的电力负荷建模[J].电网技术,2015,39(5):1358-1364.Zhu Jianquan,Li Ying,Tan Wei,et al.Characteristic fusion based on electric load modeling[J].Power System Technology,2015,39(5):1358-1364(in Chinese).
[10]罗军,何光宇,张思远,等.基于负荷点相似的地区短期负荷预测新方法[J].电网技术,2007,31(6):67-71.Luo Jun,He Guangyu,Zhang Siyuan,et al.A new method to forecast regional short-term load based on similar load point[J].Power System Technology,2007,31(6):67-71(in Chinese).
[11]贺建章,王海波,季知祥,等.面向智能电网的配电变压器重过载影响因素分析[J].电网技术,2017,41(1):279-284.He Jianzhang,Wang Haibo,Ji Zhixiang,et al.Analysis of factors affecting distribution transformer overload in smart grid[J].Power System Technology,2017,41(1):279-284(in Chinese).
[12]吴越强,吴文传,李飞,等.基于鲁棒Holt-Winter模型的超短期配变负荷预测方法[J].电网技术,2014,38(10):2810-2815.Wu Yueqiang,Wu Wenchuan,Li Fei,et al.Ultra-short term load forecasting using robust holt-winter in distribution network[J].Power System Technology,2014,38(10):2810-2815(in Chinese).
[13]Li M,Zhou Q.Distribution transformer mid-term heavy load and overload pre-warning based on logistic regression[C]//Power Tech,2015 IEEE.Eindhoven,Netherlands:IEEE,2015:1-5.
[14]李楠,段隆振,陈萌.决策树C4.5算法在数据挖掘中的分析及其应用[J].计算机与现代化,2008(12):160-163.Li Nan,Duan Longzhen,Chen Meng.Research and application of data mining algorithm based on decision tree[J].Computer&Modernization,2008(12):160-163(in Chinese).
[15]方匡南,吴见彬,朱建平,等.随机森林方法研究综述[J].统计与信息论坛,2011,26(3):32-38.Fang Kuangnan,Wu Jianbin,Zhu Jianping,et al.A review of technologies on random forests[J].Statistics&Information Forum,2011,26(3):32-38(in Chinese).
[16]吴磊,房斌,刁丽萍,等.融合过抽样和欠抽样的不平衡数据重抽样方法[J].计算机工程与应用,2013(21):172-176.Wu Lei,Fang Bin,Diao Liping,et al.Imbalanced data resampling based on oversampling and under-sampling[J].Computer Engineering&Applications,2013(21):172-176(in Chinese).
[17]徐建忠,李晖,钟云南,等.一种城网配变重过载短期预警模型的建模方法:CN104699890A[P].2015-01-07.
[18]孙雅明,宋轶博.基于Petri Nets和遗传算法结合的配变变压器故障的恢复研究[J].电力系统保护与控制,2004,32(15):14-18.Sun Yaming,Song Yibo.Study of the transformer service restoration based on the combination of Petri Net and GA in distribution substation[J].Relay,2004,32(15):14-18(in Chinese).
[2]吴潇雨,和敬涵,张沛,等.基于灰色投影改进随机森林算法的电力系统短期负荷预测[J].电力系统自动化,2015,39(12):50-55.Wu Xiaoyu,He Jinghan,Zhang Pei,et al.Power system short-term load forecasting based on improved random forest with grey relation projection[J].Automation of Electric Power Systems,2015,39(12):50-55(in Chinese).
[3]赵莉,候兴哲,胡君,等.基于改进k-means算法的海量智能用电数据分析[J].电网技术,2014,38(10):2715-2720.Zhao Li,Hou Xingzhe,Hu Jun,et al.Improved K-means algorithm based analysis on massive data of intelligent power utilization[J].Power System Technology,2014,38(10):2715-2720(in Chinese).
[4]焦润海,苏辰隽,林碧英,等.基于气象信息因素修正的灰色短期负荷预测模型[J].电网技术,2013,37(3):720-725.Jiao Runhai,Su Chenjun,Lin Biying,et al.Short-term load forecasting by grey model with weather factor-based correction[J].Power System Technology,2013,37(3):720-725(in Chinese).
[5]周佃民,管晓宏.基于神经网络的电力系统短期负荷预测研究[J].电网技术,2002,26(2):10-13.Zhou Dianmin,Guan Xiaohong.A short-term load forecasting system based on bp artificial neural network[J].Power System Technology,2002,26(2):10-13(in Chinese).
[6]王大鹏,汪秉文.基于变权缓冲灰色模型的中长期负荷预测[J].电网技术,2013,37(1):167-171.Wang Dapeng,Wang Bingwen.Medium-and long-term load forecasting based on variable weights buffer grey model[J].Power System Technology,2013,37(1):167-171(in chinese).
[7]戴远航,陈磊,闵勇,等.电网每日运行评价指标体系研究[J].电网技术,2015,39(6):1611-1616.Dai Yuanhang,Chen Lei,Min Yong,et al.Study on the daily operation assessment system of power grid[J].Power System Technology,2015,39(6):1611-1616(in Chinese).
[8]朱利鹏,陆超,孙元章,等.基于数据挖掘的区域暂态电压稳定评估[J].电网技术,2015,39(4):1026-1032.Zhu Lipeng,Lu Chao,Sun Yuanzhang,et al.Data mining based regional transient voltage stability assessment[J].Power System Technology,2015,39(4):1026-1032(in Chinese).
[9]朱建全,李颖,谭伟,等.基于特性融合的电力负荷建模[J].电网技术,2015,39(5):1358-1364.Zhu Jianquan,Li Ying,Tan Wei,et al.Characteristic fusion based on electric load modeling[J].Power System Technology,2015,39(5):1358-1364(in Chinese).
[10]罗军,何光宇,张思远,等.基于负荷点相似的地区短期负荷预测新方法[J].电网技术,2007,31(6):67-71.Luo Jun,He Guangyu,Zhang Siyuan,et al.A new method to forecast regional short-term load based on similar load point[J].Power System Technology,2007,31(6):67-71(in Chinese).
[11]贺建章,王海波,季知祥,等.面向智能电网的配电变压器重过载影响因素分析[J].电网技术,2017,41(1):279-284.He Jianzhang,Wang Haibo,Ji Zhixiang,et al.Analysis of factors affecting distribution transformer overload in smart grid[J].Power System Technology,2017,41(1):279-284(in Chinese).
[12]吴越强,吴文传,李飞,等.基于鲁棒Holt-Winter模型的超短期配变负荷预测方法[J].电网技术,2014,38(10):2810-2815.Wu Yueqiang,Wu Wenchuan,Li Fei,et al.Ultra-short term load forecasting using robust holt-winter in distribution network[J].Power System Technology,2014,38(10):2810-2815(in Chinese).
[13]Li M,Zhou Q.Distribution transformer mid-term heavy load and overload pre-warning based on logistic regression[C]//Power Tech,2015 IEEE.Eindhoven,Netherlands:IEEE,2015:1-5.
[14]李楠,段隆振,陈萌.决策树C4.5算法在数据挖掘中的分析及其应用[J].计算机与现代化,2008(12):160-163.Li Nan,Duan Longzhen,Chen Meng.Research and application of data mining algorithm based on decision tree[J].Computer&Modernization,2008(12):160-163(in Chinese).
[15]方匡南,吴见彬,朱建平,等.随机森林方法研究综述[J].统计与信息论坛,2011,26(3):32-38.Fang Kuangnan,Wu Jianbin,Zhu Jianping,et al.A review of technologies on random forests[J].Statistics&Information Forum,2011,26(3):32-38(in Chinese).
[16]吴磊,房斌,刁丽萍,等.融合过抽样和欠抽样的不平衡数据重抽样方法[J].计算机工程与应用,2013(21):172-176.Wu Lei,Fang Bin,Diao Liping,et al.Imbalanced data resampling based on oversampling and under-sampling[J].Computer Engineering&Applications,2013(21):172-176(in Chinese).
[17]徐建忠,李晖,钟云南,等.一种城网配变重过载短期预警模型的建模方法:CN104699890A[P].2015-01-07.
[18]孙雅明,宋轶博.基于Petri Nets和遗传算法结合的配变变压器故障的恢复研究[J].电力系统保护与控制,2004,32(15):14-18.Sun Yaming,Song Yibo.Study of the transformer service restoration based on the combination of Petri Net and GA in distribution substation[J].Relay,2004,32(15):14-18(in Chinese).