Open AccessModelling and Load Forecasting Using Multiple Linear Regression and Curve Fitting Method
Author(s) -
Vikas Gupta,
Seema Pal
Publication year - 2017
Publication title -
zenodo (cern european organization for nuclear research)
Language(s) - English
DOI - 10.5281/zenodo.545687
Subject(s) - linear regression , proper linear model , curve fitting , regression , regression analysis , computer science , econometrics , mathematics , statistics , bayesian multivariate linear regression
This article provides a way of predicting one week ahead load forecasting using the data of Madhya Pradesh Poorva Kshetra Vidyut Vitaran Company Ltd. (MPPKVVCL). The data is used from Jan 2015 to Dec 2015 and the result for 1 January 2016 to 7 January 2016 is obtained . Method used for the load forecasting is Multiple Linear Regression Analysis and Curve Fitting .Various parameter such as temperature, humidity, dayofweek, weekofmonth, HourofDay etc are used to make the analysis more comparative and accurate. Model for both mentioned methods is developed in MATLAB. Result obtained will be compared between the two used methods. A large variation is seen on the festival and week days. 1 January 2015 historical data shows large variation and hence result shows large variation of 9.22% obtained from Curve Fitting method and variation of 15.53% is obtained from Multiple Linear regression for 1 Jan 2016 (New Year Day) and for rest of the days i.e. from 2 January 2016 to 7 January 2016 the error hardly goes beyond 2.65%