| 基于支持向量机的城市燃气日负荷预测方法研究 |
张超1,2,刘奕1( ),张辉1,黄弘1 |
| 2. 中国标准化研究院, 北京 100191 |
| Study on urban short-term gas load forecasting based on support vector machine model |
| Chao ZHANG1,2,Yi LIU1(),Hui ZHANG1,Hong HUANG1 |
| 1. Institute for Public Safety Research, Department of Engineering Physics, Tsinghua University, Beijing 100084, China 2. China National Institute of Standardization, Beijing 100191, China |
摘要:
| |||
| 文章导读 | |||
| 摘要天然气是绿色、高效的能源,在工业生产和居民生活中应用日益广泛。天然气日负荷预测有助于科学合理地进行供应调配,预测结果对于实际工作具有重要参考价值。该文采用支持向量机方法对华北某城市的燃气实际日负荷数据进行了分析,建立了城市燃气日负荷预测模型。讨论了影响燃气日负荷变化的若干主要因素及其对燃气负荷预测建模的影响,分析了数据规则化方法对预测模型准确性的影响。该文建立的模型,对于全年负荷的预测误差小于5%; 对于供暖期负荷的预测误差约为2%,结果较好。该文对建模影响因素和预测准确性的讨论,对类似问题有一定借鉴意义。 | |||
| 关键词 ：燃气负荷预测,支持向量机,数据规则化方法 | |||
| Abstract：Natural gas is green and efficient energy which is widely used in industrial production and daily life. Daily gas load forecasting is helpful for scientifically and rationally supplying. Therefore, the forecasting results are beneficial to practical work. A forecasting model for daily gas loads was developed based on support vector machine theory. The gas load data of a North-China city were taken as a sample to verify the forecasting accuracy, with the main factors that affect the daily gas load as well as their effects on model accuracies being discussed. Several data normalization methods were used with the forecasting accuracy based on normalization methods analyzed. The developed model performs well with the error less than 5% for the through-year data, and less than 2% for the heating period data. The discussion about forecasting accuracies in this paper may be helpful for similar problems. | |||
| Key words：gas load forecastingsupport vector machinedata normalization method | |||
| 收稿日期: 2010-10-05 出版日期: 2015-04-16 | |||
| |||
| 基金资助:国家自然科学基金资助项目(70833003, 90924001);北京市科技专项(Z09050600910902) | |||
| 引用本文: |
| 张超, 刘奕, 张辉, 黄弘. 基于支持向量机的城市燃气日负荷预测方法研究[J]. 清华大学学报（自然科学版）, 2014, 54(3): 320-325. Chao ZHANG, Yi LIU, Hui ZHANG, Hong HUANG. Study on urban short-term gas load forecasting based on support vector machine model. Journal of Tsinghua University(Science and Technology), 2014, 54(3): 320-325. |
| 链接本文: |
| http://jst.tsinghuajournals.com/CN/或 http://jst.tsinghuajournals.com/CN/Y2014/V54/I3/320 |
图表:
| |||||||||||||||||||||||||||||||||||||||||||||||||
| 燃气日负荷与影响因素之间的相关性分析 | |||||||||||||||||||||||||||||||||||||||||||||||||
 | |||||||||||||||||||||||||||||||||||||||||||||||||
| 燃气日负荷与日平均温度之间的关系 | |||||||||||||||||||||||||||||||||||||||||||||||||
 | |||||||||||||||||||||||||||||||||||||||||||||||||
| 基于支持向量机的燃气日负荷预测模型结构图 | |||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||
| 支持向量机模型的输入向量 | |||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||
| 影响因子定性信息的量化规则 | |||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||
| 将全年数据作为训练样本的预测结果 | |||||||||||||||||||||||||||||||||||||||||||||||||
 | |||||||||||||||||||||||||||||||||||||||||||||||||
| 将全年数据作为训练样本,对未来10 d燃气负荷进行预测的结果 | |||||||||||||||||||||||||||||||||||||||||||||||||
 | |||||||||||||||||||||||||||||||||||||||||||||||||
| 将全年数据作为训练样本,对未来10 d、 20 d、 30 d负荷进行预测的结果 | |||||||||||||||||||||||||||||||||||||||||||||||||
| |||||||||||||||||||||||||||||||||||||||||||||||||
| 燃气负荷分组方案及预测结果 | |||||||||||||||||||||||||||||||||||||||||||||||||
 | |||||||||||||||||||||||||||||||||||||||||||||||||
| 全年预测与分组预测的结果对比 | |||||||||||||||||||||||||||||||||||||||||||||||||
 | |||||||||||||||||||||||||||||||||||||||||||||||||
| 不同规则化方法对预测结果准确性的影响 | |||||||||||||||||||||||||||||||||||||||||||||||||
 | |||||||||||||||||||||||||||||||||||||||||||||||||
| 不同组别数值负荷特性对预测结果准确性的影响对比 |
参考文献:
| [1] | 肖久明. 城市燃气负荷的特点与预测模型的特征[J]. 煤气与热力, 2005, 25(1): 67-70. XIAO Jiuming. Characteristics of city gas load and features of forecast model[J]. Gas & Heat, 2005, 25(1): 67-70. (in Chinese) |
| [2] | 邹雪春. 城市燃气管网用气不均匀性分析及负荷预测 [D]. 广州: 广州大学, 2007. ZOU Xuechun. Non-uniformity analyses of natural gas network usage in China's cities and load prediction [D]. Guangzhou: Guangzhou University, 2007. (in Chinese) |
| [3] | 苗艳姝. 城市燃气负荷预测的研究 [D]. 哈尔滨: 哈尔滨工业大学, 2006. MIAO Yanshu. Research on the natural gas load prediction in China's cities [D]. Harbin: Harbin Institute of Technology, 2006. (in Chinese) |
| [4] | 焦文玲,秦裕琨,赵林波. 城市燃气负荷预测系统体系研究[J]. 天然气工业, 2005, 25(1): 155-158. JIAOWenling, QIN Yukun, ZHAO Linbo. Study on forecasting system of city gas load[J]. Natural Gas Industry, 2005, 25(1): 155-158. (in Chinese) |
| [5] | 苗艳姝,段常贵. 基于相似日的节假日燃气短期负荷预测[J]. 煤气与热力, 2006, 26(5): 10-14. MIAO Yanshu, DUAN Changgui. Short-term gas load prediction in holidays based on similar days[J]. Gas & Heat, 2006, 26(5): 10-14. (in Chinese) |
| [6] | 杨苑莺. 小城镇居民生活用气规律及短期负荷预测 [D]. 重庆: 重庆大学, 2006. YANG Yuanying. Regularity of daily natural gas consumption for residents in smallChinese cities and towns and short-term load prediction [D]. Chongqing: Chongqing University, 2006. (in Chinese) |
| [7] | 李国芳,徐永生,苏刚. OIHF Elman网络在燃气日负荷预测中的应用[J]. 天津城市建设学院学报, 2008, 14(4): 283-288. LI Guofang, XU Yongsheng, SU Gang. Application of OIHF Elman network in forecasting of daily natural gas load[J]. J Tianjin Institute of Urban Construction, 2008, 14(4): 283-288. (in Chinese) |
| [8] | 豆连旺,冯良. 基于神经网络的城市燃气短期负荷预测[J]. 燃气与热力, 2005, 25(12): 10-14. DOU Lianwang, FENG Liang. Short-term city gas load forecast based on neural network[J]. Gas & Heat, 2005, 25(12): 10-14. (in Chinese) |
| [9] | 李持佳,焦文玲,赵林波. 燃气短期负荷预测的小波分析综合模型[J]. 天然气工业, 2007, 27(8): 103-108. LI Chijia, JIAO Wenling, ZHAO Linbo. A synthesis wavelet analysis method for short-term gas load prediction[J]. Natural Gas Industry, 2007, 27(8): 103-108. (in Chinese) |
| [10] | 席德粹,焦文玲,李持佳,等. 上海市燃气负荷预测系统的开发与试验运行[J]. 城市燃气, 2004, 353: 14-16. XI Decui, JIAO Wenling, LI Chijia,et al.The development and operation of the gas load forecasting system in Shanghai[J]. Urban Gas, 2004, 353: 14-16. (in Chinese) |
| [11] | Mukherjee S, Osuna E, Girosi F. Nonlinear prediction of chaotic time series using support vector machines [C]// 7th IEEE Workshop on Neural Networks for Signal Processing. Amelia Island, USA: IEEE Press, 1997: 511-520. |
| [12] | Chen B J, Chang M W, Lin C J. Load forecasting using support vector machines: A study on EUNITE competition 2001[J]. IEEE Transactions on Power Systems, 2004, 19(4): 1821-1830. |
| [13] | Vapnik V N. TheNature of Statistical Learning Theory[M]. New York: Springer, 1995. |
相关文章:
|
