GBDT-based cold and heat load prediction method of LightGBM model

Abstract

The invention provides a GBDT-based cold and heat load prediction method of a LightGBM model. The method comprises the steps of 1, obtaining a plurality of building cases; 2, generating model data from the building case through a first model, wherein the model data comprises characteristic variable data and cold and heat load data; 3, preprocessing the model data, splitting a training set and a test set, and performing training modeling on the training set based on a LightGBM model to train and generate a cold and heat load prediction model; 4, testing the cold and heat load prediction model by adopting the data of the test set; 5, storing the cold and heat load prediction model passing the test. According to the GBDT-based cold and heat load prediction method for the LightGBM model, the load prediction regression model is constructed based on the LightGBM tree model of the boosting thought, and the cold and heat loads of different building structures can be accurately predicted aftertraining is carried out on a training set with a large amount of data.

Description

technical field [0001] The invention relates to the technical field of load forecasting, in particular to a cooling and heating load forecasting method based on the LightGBM model of GBDT. Background technique [0002] The current load forecasting methods mainly focus on single buildings or areas, through the collection and processing of existing data, to predict the future power load, cooling load and heating load. The time span of forecasting can also be divided into short-term forecasting: a few hours or a day or a week; medium-term forecasting: one month to one quarter; long-term forecasting: one year or more. Due to energy saving and emission reduction and planning area and building cooling and heating load, it is even used as a reference in building type selection. Using various forecasting methods to accurately predict the cooling and heating loads in a building or region is crucial, which enables those who design building electromechanical systems to select equipmen...

AI Technical Summary

Problems solved by technology

When we want to find similar laws from multiple buildings of the same type, like many office buildings and many hospital buildings, etc., and deduce them to a new office building or hospital building, the above method will seem a little weak

Images