ﯾﮏ روش ﭘﯿﺶﺑﯿﻨﯽ ﺑﻠﻨﺪﻣﺪت ﺑﺎر اﻟﮑﺘﺮﯾﮑﯽ ﻣﺒﺘﻨﯽ ﺑﺮ اﺳﺘﺨﺮاج وﯾﮋﮔﯽ ﺑﺮاي ﮐﺎﻫﺶ اﺛﺮ دادهﻫﺎي ﺧﺎرج از ﻣﺤﺪوده

A Feature Extraction Based Long-Term Electricity load forecasting Framework to Reduce the Outliers Data Effects

ﭘﯿﺶﺑﯿﻨﯽ ﻣﯿﺎن-ﻣﺪت ﺑﺎر اﻟﮑﺘﺮﯾﮑﯽ اﻏﻠﺐ ﺑﺮاي ﺑﺮﻧﺎﻣﻪرﯾﺰي ﻋﻤﻠﯿﺎت ﻧﯿﺮوﮔﺎهﻫﺎي ﺣﺮارﺗﯽ و آﺑﯽ، زﻣﺎنﺑﻨﺪي ﺑﻬﯿﻨﻪ ﺑﺮاي ﺑﺎزرﺳﯽ و ﺗﻌﻤﯿﺮات و ﻧﮕﻬﺪاري ﻧﯿﺮوﮔﺎهﻫﺎ و ﺷﺒﮑﻪ ﺑﺮق اﺳﺘﻔﺎده ﻣﯽﺷﻮد. در اﯾﻦ ﻣﻘﺎﻟﻪ ﯾﮏ روش ﺗﺮﮐﯿﺒﯽ ﺑﺎ اﺳﺘﻔﺎده از ﺗﺒﺪﯾﻞ ﻣﻮﺟﮏ و ﻣﺎﺷﯿﻦ ﯾﺎدﮔﯿﺮي ﺷﺪﯾﺪ ﻣﻘﺎوم ﺑﻪ دادهﻫﺎي ﺧﺎرج از ﻣﺤﺪوده، ﺑﺮاي ﭘﯿﺶﺑﯿﻨﯽ ﺑﻠﻨﺪﻣﺪت ﺑﺎر اراﺋﻪ ﺷﺪه اﺳﺖ. دادهﻫﺎي ﺑﺎر و دﻣﺎي ﺳﺎﻋﺘﯽ، از ﭘﺎﯾﮕﺎه داده 2014 GEFCOM اﺳﺘﺨﺮاج ﺷﺪه و ﺑﻪ دو دﺳﺘﻪ آﻣﻮزش و آزﻣﺎﯾﺶ ﺗﻘﺴﯿﻢ ﺷﺪه اﺳﺖ. از ﺗﺒﺪﯾﻞ ﻣﻮﺟﮏ ﯾﮏ ﺳﻄﺤﯽ ﺑﺮاي ﺗﺠﺰﯾﻪ دادهﻫﺎ ﺑﻪﻣﻨﻈﻮر اﺳﺘﺨﺮاج وﯾﮋﮔﯽﻫﺎ و ﮐﺎﻫﺶ اﺑﻌﺎد ﻣﺎﺗﺮﯾﺲ دادهﻫﺎ اﺳﺘﻔﺎده ﻣﯽﺷﻮد. دو دﺳﺘﻪ ﻣﻘﺎدﯾﺮ ﻣﺆﻟﻔﻪﻫﺎي ﻓﺮﮐﺎﻧﺲ ﭘﺎﯾﯿﻦ )ﺗﻘﺮﯾﺐ( و ﻣﻘﺎدﯾﺮ ﻣﺆﻟﻔﻪﻫﺎي ﻓﺮﮐﺎﻧﺲ ﺑﺎﻻ )ﺟﺰﺋﯿﺎت( ﺣﺎﺻﻞ از ﺗﺠﺰﯾﻪ ﺟﻬﺖ آﻣﻮزش و ﭘﯿﺶﺑﯿﻨﯽ ﺑﻪ ﻣﺪل وارد ﺷﺪه و ﺧﺮوﺟﯽ ﻣﻘﺎدﯾﺮ ﭘﺎﯾﯿﻦ ﺑﺎ ﺧﺮوﺟﯽ ﻣﻘﺎدﯾﺮ ﺑﺎﻻي ﻣﺪل ﺟﻤﻊ ﻣﯽﺷﻮد ﺗﺎ ﭘﯿﺶﺑﯿﻨﯽ ﻧﻬﺎﯾﯽ را ﺗﺸﮑﯿﻞ دﻫﺪ. ﺟﻬﺖ ﺳﻨﺠﺶ و ﻣﻘﺎﯾﺴﻪ دﻗﺖ و ﮐﺎراﯾﯽ روش ﭘﯿﺸﻨﻬﺎدي، اﻋﻤﺎل ﺗﺒﺪﯾﻞ ﻣﻮﺟﮏ روي دادهﻫﺎ، ﺑﺮاي ﺳﻪ ﻣﺪل دﯾﮕﺮ ﻣﺎﺷﯿﻦ ﯾﺎدﮔﯿﺮي ﺷﺪﯾﺪ اﻧﺠﺎم ﮔﺮدﯾﺪه اﺳﺖ. ﻫﻤﭽﻨﯿﻦ دادهﻫﺎ ﺑﺪون اﻋﻤﺎل ﺗﺒﺪﯾﻞ ﻣﻮﺟﮏ ﺑﻪ ﭼﻬﺎر ﻣﺪل ﭘﯿﺶﺑﯿﻨﯽ دﯾﮕﺮ ﻧﯿﺰ وارد ﺷﺪه و ﻧﺘﺎﯾﺞ ﭘﯿﺶﺑﯿﻨﯽ ﺣﺎﺻﻞ ﺑﺎ روش ﭘﯿﺸﻨﻬﺎدي ﻣﻮرد ﻣﻘﺎﯾﺴﻪ ﻗﺮار ﮔﺮﻓﺘﻪ اﺳﺖ. ﻧﺘﺎﯾﺞ ارزﯾﺎﺑﯽ ﻓﻮق ﻧﺸﺎن ﻣﯽدﻫﺪ ﮐﻪ ﺗﺒﺪﯾﻞ ﻣﻮﺟﮏ و ﻣﺎﺷﯿﻦ ﯾﺎدﮔﯿﺮي ﺷﺪﯾﺪ ﻣﻘﺎوم ﺑﻪ دادهﻫﺎي ﺧﺎرج ﻣﺤﺪوده ﺑﺎﻋﺚ ﺑﻬﺒﻮد دﻗﺖ ﭘﯿﺶﺑﯿﻨﯽ ﻣﯽﮔﺮدد و ﻣﻘﺪار ﻣﯿﺎﻧﮕﯿﻦ درﺻﺪ ﺧﻄﺎي ﻣﻄﻠﻖ ﺑﻪ ﻋﺪد 3/0966 ﮐﺎﻫﺶ ﯾﺎﻓﺘﻪ اﺳﺖ. ﻣﻘﺪار ﺧﻄﺎي ﮐﻠﯽ ﻣﺤﺎﺳﺒﻪ ﺷﺪه روش ﭘﯿﺸﻨﻬﺎدي ﺑﻬﺘﺮﯾﻦ ﻧﺘﯿﺠﻪ در ﺑﯿﻦ ﺳﺎﯾﺮ ﻣﺪلﻫﺎي ﻣﺎﺷﯿﻦ ﯾﺎدﮔﯿﺮي ﺷﺪﯾﺪ و روشﻫﺎي ﺑﺪون ﭘﯿﺶﭘﺮدازش ﺑﻮده اﺳﺖ. ﺧﻄﺎي ﻓﻮق ﺑﺮ ﻣﺒﻨﺎي ﻣﻘﺪار ﻣﯿﺎﻧﮕﯿﻦ درﺻﺪ ﺧﻄﺎي ﻣﻄﻠﻖ ﺑﻪﺗﺮﺗﯿﺐ 0/4208 ﻧﺴﺒﺖ ﺑﻪ ﻣﺪل ﻣﺎﺷﯿﻦ ﯾﺎدﮔﯿﺮي ﺷﺪﯾﺪ اﺻﻠﯽ، 0/1194 ﻧﺴﺒﺖ ﺑﻪ ﻣﺪل ﺗﻨﻈﯿﻢﺷﺪه و 0/1353 ﻧﺴﺒﺖ ﺑﻪ ﻣﺪل ﺗﻨﻈﯿﻢﺷﺪه و وزندار، ﮐﺎﻫﺶ ﯾﺎﻓﺘﻪ اﺳﺖ.

Generally, medium-term electrical load forecasting is often used for the operation of thermal and hydropower plants, optimal time planning for maintenance of power plants and the power grids. However, long-term electrical load forecasting is used to manage on-time future demands and generation, transmission and distribution expansion planning. In this paper, a hybrid long-term load forecasting approach using wavelet transform and an outlier robust extreme learning machine is proposed. Hourly load and temperature data were extracted from the GEFCOM 2014 database and divided into two classes of training and test. The one-level wavelet transform is used to decompose data to extract properties and reduce the dimensions of the data matrix. Decomposed low-frequency component (approximations) and high-frequency component values (details) from wavelet analysis are entered into the model for training and forecasting. For comparison accuracy of the proposed method, wavelet transform is applied to the data for the other three extreme learning machines. Also data without wavelet transform entered into four other forecasting models and the load forecasting results are compared with the proposed method. The results of the above mentioned evaluation show that electrical load forecasting by using wavelet transform and outlier robust extreme learning machine improves forecasting accuracy and the MAPE reduces to 3.0966. The overall calculated error by the proposed method was the best result obtained between the three several models of extreme learning machines and without preprocessing model. The MAPE is 0.4208 less than the ELM, 0.944 less than the RELM, and 0.1353 less than the WRELM model, respectively