Font Size: a A A

Research On XX Enterprise Building Energy Consumption Forecast Based On Machine Learning Method

Posted on:2021-01-30Degree:MasterType:Thesis
Country:ChinaCandidate:H Y LiuFull Text:PDF
GTID:2392330602982094Subject:Engineering
Abstract/Summary:PDF Full Text Request
Energy is a very important factor that affects the development of an enterprise.Analyzing and predicting various types of energy consumption within the enterprise is beneficial to the enterprise's energy management,reducing daily energy consumption,achieving energy conservation,emission reduction,and sustainable development.This article takes the XX Enterprise Technology Center as the research object,collects relevant data,analyzes the energy consumption situation,and analyzes the energy consumption situation for four automobile test laboratories,science and technology buildings,and strength,vehicle,engine performance,and engine emissions.Part of the energy consumption prediction model is established to help enterprises manage energy consumption and reduce waste.First of all,through analyzing the energy consumption data of lighting and office equipment in refrigeration equipment rooms,science and technology buildings,the main influencing factors of each part of energy consumption are determined.Then use multiple regression,SVM and LS-SVM to model and predict the daily energy consumption of the refrigeration equipment room;use multiple regression and least square regression to model and predict the daily energy consumption of office equipment;use multiple regression and CART decision tree pairs Modeling and forecasting the daily energy consumption of lighting equipment.At the same time,by evaluating the prediction accuracy of various models,the most realistic model is selected.Finally,the intensity,vehicle,engine performance and energy consumption data of the emission laboratory are analyzed,and the least square polynomial fitting algorithm is used to fit the hourly energy consumption of the four automobile laboratories.Combine polynomials to establish energy consumption prediction models for each laboratory.Through model evaluation,it is found that the fitting accuracy of each model is good,which can effectively reflect the energy consumption of the enterprise,and has certain theoretical and application value for guiding the implementation of energy use and energy-saving measures of the enterprise.
Keywords/Search Tags:Energy Consumption Prediction, Multiple Regression, Support Vector Machine, CART Decision Tree, Polynomial Fitting
PDF Full Text Request
Related items