﻿ 基于时间序列的电力负荷数据分析 Analysis of Electric Load Data Based on Time Series

Vol.05 No.02(2016), Article ID:17547,11 pages
10.12677/AAM.2016.52028

Analysis of Electric Load Data Based on Time Series

Shuo Yuan, Liding Chen, Guopeng Sun, Jinguan Lin*

Department of Mathematics, Southeast University, Nanjing Jiangsu

Received: Apr. 23rd, 2016; accepted: May 10th, 2016; published: May 13th, 2016

ABSTRACT

Time series analysis method is one of the important tools in the field of power load forecasting. It mainly describes the law of the historical data dynamic change over time to predict the future value by establishing a relevant model. In this paper, Winter’s exponential smoothing method and seasonal ARIMA model are applied to model estimating on the power load data, and the authors use the Mean Absolute Percentage Error (MAPE) to measure the accuracy. The results prove that both of them have high fitting and forecasting precision.

Keywords:Electric Load, ARIMA, Time Series, Forecasting, Exponential Smoothing

1. 引言

2. 模型介绍

2.1. 温特线性与季节指数平滑法

2.2. ARIMA模型

2.2.1. ARMA模型

2.2.2. 求和ARIMA模型

2.2.3. 季节乘积ARIMA模型

(1)

(2)

3. 电力负荷数据案例分析

3.1. 基于温特线性与季节性指数平滑法的探究性分析

(1) 模型初值的确定

Figure 1. Original values of electric power load from May 1, 2013 to May 11, 2013

(2) 最优平滑系数的确定

(3) 预测与讨论

3.2. 基于ARMA模型的探究性分析

Figure 2. Seasonal difference plot of the previous ten days

Table 1. The ACF of seasonal difference

Table 2. Unit root test

3.3. 基于乘积季节ARIMA模型的探究性分析

Table 3. Parameter estimation and test of ARMA(1,1)

Table 4. Parameter estimation of noconstant model

Table 5. Autocorrelation check of residuals

Table 7. ACF and PACF of seasonal difference

Table 8. Comparison of seasonal ARIMA model

Table 9. Parameter estimation

Table 10. Parameter estimation of noconstant model

4. 结论

Figure 3. Comparison of predicted values and original values on May 11th.

Figure 4. Comparison of predicted values and original values on May 11th.

Analysis of Electric Load Data Based on Time Series[J]. 应用数学进展, 2016, 05(02): 214-224. http://dx.doi.org/10.12677/AAM.2016.52028

1. 1. 牛东晓. 电力负荷预测技术及其应用[M]. 北京: 中国电力出版社, 1998: 40-54, 59-88, 95-134, 185-101.

2. 2. 邓祖新. 数据分析方法和SAS系统[M]. 上海: 上海财经大学出版社, 2006: 390-443.

3. 3. 吴仁群. 经济预测与决策[M]. 北京: 中国人民大学出版社, 2015: 34-36, 92-140.

4. 4. 梁慧玲, 赵嘉阳, 林玉蕊. 温特线性与季节指数平滑法在气温中的应用[J]. 科技和产业, 2015(9): 145-148.

5. 5. Hanzák, T. (2012) Holt-Winters Method with General Seasonality. Kybernetika Praha, 1, 1-15.

6. 6. 王燕. 应用时间序列分析[M]. 北京: 中国人民大学出版社, 2012: 41-110.

7. 7. 孙国红, 沈跃, 徐应明, 等. 基于Box-Jenkins方法的黄河水质时间序列分析与预测[J]. 农业环境科学学报, 2011, 30(9): 1888-1895.

8. 8. Brockwell, P.J. and Davis, R.A. 时间序列: 理论与方法[M]. 田铮, 译. 北京：高等教育出版社, 2001: 200-202, 214-254.

9. 9. Kumar, S.V. and Vanajakshi, L. (2015) Short-Term Traffic Flow Prediction Using Seasonal ARIMA Model with Limited Input Data. European Transport Research Review, 7, 1-9. http://dx.doi.org/10.1007/s12544-015-0170-8

10. 10. Box, G.E.P. 时间序列分析: 预测与控制[M]. 王成璋, 等, 译. 北京: 机械工业出版社, 2011: 116-125.

11. 11. 邓盛川, 于德亮, 齐维贵. 基于乘积季节ARIMA模型的供热负荷预报[J]. 沈阳工业大学学报, 2011(3): 321-325.

12. 12. Eni, D. and Adeyeye, F.J. (2015) Seasonal ARIMA Modeling and Forecasting of Rainfall in Warri Town, Nigeria. Journal of Geoscience & Environment Protection, 3, 91-98. http://dx.doi.org/10.4236/gep.2015.36015

13. 13. 叶舟, 陈康民. 温斯特线性与季节性指数平滑法在电力负荷预测中的应用及改进[J]. 电力建设, 2000, 21(8): 13-16.