搜索

x

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于优化核极限学习机的风电功率时间序列预测

李军 李大超

引用本文:
Citation:

基于优化核极限学习机的风电功率时间序列预测

李军, 李大超

Wind power time series prediction using optimized kernel extreme learning machine method

Li Jun, Li Da-Chao
PDF
导出引用
  • 针对时间序列预测, 在单隐层前馈神经网络的基础上, 基于进化计算的优化策略, 提出了一种优化的核极限学习机(optimized kernel extreme learning machine, O-KELM)方法. 与极限学习机(extreme learning machine, ELM)方法相比, 核极限学习机(kernel extreme learning machine, KELM)方法无须设定网络隐含层节点的数目, 以核函数表示未知的隐含层非线性特征映射, 通过正则化最小二乘算法计算网络的输出权值, 它能以极快的学习速度获得良好的推广性. 在KELM的基础上, 分别将遗传算法、模拟退火、微分演化三种进化算法用于模型的结构输入选择、正则化系数以及核参数的优化选取, 以进一步提高网络的性能. 将O-KELM方法应用于标准Mackey-Glass混沌时间序列预测及某地区的风电功率时间序列预测实例中, 在同等条件下, 还与优化的极限学习机(optimized extreme learning machine, O-ELM)方法进行比较. 实验结果表明, 所提出的O-KELM方法在预测精度上优于O-ELM方法, 表明了其有效性.
    Since wind has an intrinsically complex and stochastic nature, accurate wind power prediction is necessary for the safety and economics of wind energy utilization. Aiming at the prediction of very short-term wind power time series, a new optimized kernel extreme learning machine (O-KELM) method with evolutionary computation strategy is proposed on the basis of single-hidden layer feedforward neural networks. In comparison to the extreme learning machine (ELM) method, the number of the hidden layer nodes need not be given, and the unknown nonlinear feature mapping of the hidden layer is represented with a kernel function. In addition, the output weights of the networks can also be analytically determined by using regularization least square algorithm, hence the kernel extreme learning machine (KELM) method provides better generalization performance at a much faster learning speed. In the O-KELM, the structure and the parameters of the KELM are optimized by using three different optimization algorithms, i.e., genetic algorithm (GA), differential evolution (DE), and simulated annerling (SA), meanwhile, the output weights are obtained by a least squares algorithm just the same as by the ELM, but using Tikhonovs regularization in order to further improve the performance of the O-KELM. The utilized optimization algorithms of the O-KELM are respectively used to select the set of input variables, regularization coefficient as well as hyperparameter of kernel function. The proposed method is first applied to the direct six-step prediction for Mackey-Glass chaotic time series, under the same condition as the existing optimized ELM method. From the analysis of the simulation results it can be verified that the prediction accuracy of the proposed O-KELM method is increased by about one order of magnitude over that of the optimized ELM method. Furthermore, the DE-KELM algorithm can achieve the lowest root mean square error (RMSE). The O-KELM method is then applied to real-world wind power prediction instance, i.e., the Western Dataset from NERL. The 10-minute ahead single-step prediction as well as 20-minute ahead, 30-minute ahead, 40-minute ahead multi-step prediction for wind power time series are respectively implemented to evaluate the O-KELM method. Experimental results of each of the short-term wind power time series predictions at different time horizons confirm that the proposed O-KELM method tends to have better prediction accuracy than the optimized ELM method. Moreover, the GA-KELM algorithm outperforms other two O-KELM algorithms at future 10-minute, 20-minute, 40-minute ahead prediction in terms of the RMSE value. The DE-KELM algorithm outperforms other algorithms at future 30-minute ahead prediction in terms of the normalized mean square error (NMSE) and the RMSE value. The results from these applications demonstrate the effectiveness and feasibility of the proposed O-KLEM method. Therefore, the O-KELM method has a potential future in the field of wind power prediction.
      通信作者: 李军, lijun691201@mail.lzjtu.cn
    • 基金项目: 国家自然科学基金(批准号: 51467008)资助的课题.
      Corresponding author: Li Jun, lijun691201@mail.lzjtu.cn
    • Funds: Project supported by the National Natural Science Foundation of China (Grant No. 51467008).
    [1]

    Wang X C, Guo P, Huang X B 2011 Energy Procedia 12 770

    [2]

    Zhang G Y, Wu Y G, Zhang Y, Dai X L 2014 Acta Phys. Sin. 63 138801 (in Chinese) [章国勇, 伍永刚, 张洋, 代贤良 2014 物理学报 63 138801]

    [3]

    Foley A M, Leahy P G, Marvuglia A, McKeogh E J 2012 Renew. Energy 37 1

    [4]

    Colak I, Sagiroglu S, Yesilbudak M 2012 Renew. Energy 46 241

    [5]

    Jung J, Broadwater R P 2014 Renew. Sustain. Energy Rev. 31 762

    [6]

    Yang Z L, Feng Y, Xiong D F, Yang Z, Zhang X, Zhang J 2015 Smart Grid 3 1 (in Chinese) [杨正瓴, 冯勇, 熊定方, 杨钊, 张玺, 张军 2015 智能电网 3 1]

    [7]

    Potter C W, Negnevitsky M 2006 IEEE Trans. Power Syst. 21 965

    [8]

    Kavasseri R G, Seetharaman K 2009 Renew. Energy 34 1388

    [9]

    Louka P, Galanis G, Siebert N, Kariniotakis G, Katsafados P, Kallos G, Pytharoulis I 2008 J. Wind Eng. Ind. Aerodyn. 96 2348

    [10]

    Liu R Y, Huang L 2012 Autom. Electr. Power Syst. 36 18 (in Chinese) [刘瑞叶, 黄磊 2012 电力系统自动化 36 18]

    [11]

    Seo I Y, Ha B N, Lee S W, Jang M J, Kim S O, Kim S J 2012 J. Energy Power Eng. 6 1605

    [12]

    Zeng J W, Wei Q 2012 IEEE Trans. Sustain. Energy 3 255

    [13]

    Huang G B, Zhu Q Y, Siew C K 2006 Neurocomputing 70 489

    [14]

    Chen H Y, Gao P Z, Tan S C, Fu X K 2014 Acta Phys. Sin. 63 200505 (in Chinese) [陈涵瀛, 高璞珍, 谭思超, 付学宽 2014 物理学报 63 200505]

    [15]

    Matias T, Souza F, Arajo R, Antunes C H 2014 Neurocomputing 129 428

    [16]

    Wan C, Xu Z, Pinson P, Dong Z Y, Wong K P 2014 IEEE Trans. Power Syst. 29 1033

    [17]

    Huang G B, Zhou H M, Ding X J, Zhang R 2012 IEEE Trans. Syst. Man, and Cybern. B: Cybern. 42 513

    [18]

    Wang X Y, Han M 2015 Acta Phys. Sin. 64 070504 (in Chinese) [王新迎, 韩敏 2015 物理学报 64 070504]

    [19]

    Zhang Y T, Ma C, Li Z N, Fan H B 2014 J. Shanghai Jiaotong Univ. (Sci.) 48 641 (in Chinese) [张英堂, 马超, 李志宁, 范洪波 2014 上海交通大学学报 48 641]

    [20]

    Haupt R L, Haupt S E 2004 Practical Genetic Algorithms (2nd Ed. with CD-ROM) (New York: John Wiley Sons) pp27-66

    [21]

    Mohamed M H 2011 Neurocomputing 74 3180

    [22]

    Feoktistov V 2006 Differential Evolution:In Search of Solutions (Secaucus, NJ, USA:Springer-Verlag New York, Inc.) pp1-82

    [23]

    Machey M C, Glass L 1977 Science 197 287

    [24]

    Potter C W, Lew D, McCaa J, Cheng S, Eichelberger S, Grimit E 2008 Wind Eng. 32 325

  • [1]

    Wang X C, Guo P, Huang X B 2011 Energy Procedia 12 770

    [2]

    Zhang G Y, Wu Y G, Zhang Y, Dai X L 2014 Acta Phys. Sin. 63 138801 (in Chinese) [章国勇, 伍永刚, 张洋, 代贤良 2014 物理学报 63 138801]

    [3]

    Foley A M, Leahy P G, Marvuglia A, McKeogh E J 2012 Renew. Energy 37 1

    [4]

    Colak I, Sagiroglu S, Yesilbudak M 2012 Renew. Energy 46 241

    [5]

    Jung J, Broadwater R P 2014 Renew. Sustain. Energy Rev. 31 762

    [6]

    Yang Z L, Feng Y, Xiong D F, Yang Z, Zhang X, Zhang J 2015 Smart Grid 3 1 (in Chinese) [杨正瓴, 冯勇, 熊定方, 杨钊, 张玺, 张军 2015 智能电网 3 1]

    [7]

    Potter C W, Negnevitsky M 2006 IEEE Trans. Power Syst. 21 965

    [8]

    Kavasseri R G, Seetharaman K 2009 Renew. Energy 34 1388

    [9]

    Louka P, Galanis G, Siebert N, Kariniotakis G, Katsafados P, Kallos G, Pytharoulis I 2008 J. Wind Eng. Ind. Aerodyn. 96 2348

    [10]

    Liu R Y, Huang L 2012 Autom. Electr. Power Syst. 36 18 (in Chinese) [刘瑞叶, 黄磊 2012 电力系统自动化 36 18]

    [11]

    Seo I Y, Ha B N, Lee S W, Jang M J, Kim S O, Kim S J 2012 J. Energy Power Eng. 6 1605

    [12]

    Zeng J W, Wei Q 2012 IEEE Trans. Sustain. Energy 3 255

    [13]

    Huang G B, Zhu Q Y, Siew C K 2006 Neurocomputing 70 489

    [14]

    Chen H Y, Gao P Z, Tan S C, Fu X K 2014 Acta Phys. Sin. 63 200505 (in Chinese) [陈涵瀛, 高璞珍, 谭思超, 付学宽 2014 物理学报 63 200505]

    [15]

    Matias T, Souza F, Arajo R, Antunes C H 2014 Neurocomputing 129 428

    [16]

    Wan C, Xu Z, Pinson P, Dong Z Y, Wong K P 2014 IEEE Trans. Power Syst. 29 1033

    [17]

    Huang G B, Zhou H M, Ding X J, Zhang R 2012 IEEE Trans. Syst. Man, and Cybern. B: Cybern. 42 513

    [18]

    Wang X Y, Han M 2015 Acta Phys. Sin. 64 070504 (in Chinese) [王新迎, 韩敏 2015 物理学报 64 070504]

    [19]

    Zhang Y T, Ma C, Li Z N, Fan H B 2014 J. Shanghai Jiaotong Univ. (Sci.) 48 641 (in Chinese) [张英堂, 马超, 李志宁, 范洪波 2014 上海交通大学学报 48 641]

    [20]

    Haupt R L, Haupt S E 2004 Practical Genetic Algorithms (2nd Ed. with CD-ROM) (New York: John Wiley Sons) pp27-66

    [21]

    Mohamed M H 2011 Neurocomputing 74 3180

    [22]

    Feoktistov V 2006 Differential Evolution:In Search of Solutions (Secaucus, NJ, USA:Springer-Verlag New York, Inc.) pp1-82

    [23]

    Machey M C, Glass L 1977 Science 197 287

    [24]

    Potter C W, Lew D, McCaa J, Cheng S, Eichelberger S, Grimit E 2008 Wind Eng. 32 325

  • [1] 梅英, 谭冠政, 刘振焘, 武鹤. 基于大脑情感学习模型和自适应遗传算法的混沌时间序列预测. 物理学报, 2018, 67(8): 080502. doi: 10.7498/aps.67.20172104
    [2] 沈力华, 陈吉红, 曾志刚, 金健. 基于鲁棒极端学习机的混沌时间序列建模预测. 物理学报, 2018, 67(3): 030501. doi: 10.7498/aps.67.20171887
    [3] 魏德志, 陈福集, 郑小雪. 基于混沌理论和改进径向基函数神经网络的网络舆情预测方法. 物理学报, 2015, 64(11): 110503. doi: 10.7498/aps.64.110503
    [4] 李瑞国, 张宏立, 范文慧, 王雅. 基于改进教学优化算法的Hermite正交基神经网络混沌时间序列预测. 物理学报, 2015, 64(20): 200506. doi: 10.7498/aps.64.200506
    [5] 王新迎, 韩敏. 多元混沌时间序列的多核极端学习机建模预测. 物理学报, 2015, 64(7): 070504. doi: 10.7498/aps.64.070504
    [6] 田中大, 李树江, 王艳红, 高宪文. 短期风速时间序列混沌特性分析及预测. 物理学报, 2015, 64(3): 030506. doi: 10.7498/aps.64.030506
    [7] 刘擎超, 陆建, 陈淑燕. 基于能力区域的交通状态预测方法. 物理学报, 2014, 63(14): 140504. doi: 10.7498/aps.63.140504
    [8] 张文专, 龙文, 焦建军. 基于差分进化算法的混沌时间序列预测模型参数组合优化. 物理学报, 2012, 61(22): 220506. doi: 10.7498/aps.61.220506
    [9] 李军, 张友鹏. 基于高斯过程的混沌时间序列单步与多步预测. 物理学报, 2011, 60(7): 070513. doi: 10.7498/aps.60.070513
    [10] 李鹤, 杨周, 张义民, 闻邦椿. 基于径向基神经网络预测的混沌时间序列嵌入维数估计方法. 物理学报, 2011, 60(7): 070512. doi: 10.7498/aps.60.070512
    [11] 修春波, 徐勐. 基于混沌算子网络的时间序列多步预测研究. 物理学报, 2010, 59(11): 7650-7656. doi: 10.7498/aps.59.7650
    [12] 张春涛, 马千里, 彭宏. 基于信息熵优化相空间重构参数的混沌时间序列预测. 物理学报, 2010, 59(11): 7623-7629. doi: 10.7498/aps.59.7623
    [13] 王永生, 孙 瑾, 王昌金, 范洪达. 变参数混沌时间序列的神经网络预测研究. 物理学报, 2008, 57(10): 6120-6131. doi: 10.7498/aps.57.6120
    [14] 杨永锋, 任兴民, 秦卫阳, 吴亚锋, 支希哲. 基于EMD方法的混沌时间序列预测. 物理学报, 2008, 57(10): 6139-6144. doi: 10.7498/aps.57.6139
    [15] 王革丽, 杨培才, 毛宇清. 基于支持向量机方法对非平稳时间序列的预测. 物理学报, 2008, 57(2): 714-719. doi: 10.7498/aps.57.714
    [16] 张军峰, 胡寿松. 基于一种新型聚类算法的RBF神经网络混沌时间序列预测. 物理学报, 2007, 56(2): 713-719. doi: 10.7498/aps.56.713
    [17] 贺 涛, 周正欧. 基于分形自仿射的混沌时间序列预测. 物理学报, 2007, 56(2): 693-700. doi: 10.7498/aps.56.693
    [18] 吴延东, 谢洪波. 一种新的时间序列确定性辨识方法. 物理学报, 2007, 56(11): 6294-6300. doi: 10.7498/aps.56.6294
    [19] 李 军, 刘君华. 一种新型广义RBF神经网络在混沌时间序列预测中的研究. 物理学报, 2005, 54(10): 4569-4577. doi: 10.7498/aps.54.4569
    [20] 叶美盈, 汪晓东, 张浩然. 基于在线最小二乘支持向量机回归的混沌时间序列预测. 物理学报, 2005, 54(6): 2568-2573. doi: 10.7498/aps.54.2568
计量
  • 文章访问数:  6758
  • PDF下载量:  488
  • 被引次数: 0
出版历程
  • 收稿日期:  2016-01-18
  • 修回日期:  2016-04-15
  • 刊出日期:  2016-07-05

基于优化核极限学习机的风电功率时间序列预测

    基金项目: 国家自然科学基金(批准号: 51467008)资助的课题.

摘要: 针对时间序列预测, 在单隐层前馈神经网络的基础上, 基于进化计算的优化策略, 提出了一种优化的核极限学习机(optimized kernel extreme learning machine, O-KELM)方法. 与极限学习机(extreme learning machine, ELM)方法相比, 核极限学习机(kernel extreme learning machine, KELM)方法无须设定网络隐含层节点的数目, 以核函数表示未知的隐含层非线性特征映射, 通过正则化最小二乘算法计算网络的输出权值, 它能以极快的学习速度获得良好的推广性. 在KELM的基础上, 分别将遗传算法、模拟退火、微分演化三种进化算法用于模型的结构输入选择、正则化系数以及核参数的优化选取, 以进一步提高网络的性能. 将O-KELM方法应用于标准Mackey-Glass混沌时间序列预测及某地区的风电功率时间序列预测实例中, 在同等条件下, 还与优化的极限学习机(optimized extreme learning machine, O-ELM)方法进行比较. 实验结果表明, 所提出的O-KELM方法在预测精度上优于O-ELM方法, 表明了其有效性.

English Abstract

参考文献 (24)

目录

    /

    返回文章
    返回