Building energy consumption model evaluation method based on building energy management application scene

Abstract

The invention relates to a building energy consumption model evaluation method based on a building energy management application scene. The method comprises the following steps: firstly, establishinga building energy consumption model, establishing a multi-dimensional comprehensive evaluation model in combination with a building energy application scene, evaluating the energy consumption prediction performance of the building energy consumption model in historical energy consumption data of a certain building for many years in different periods according to the evaluation model to obtain a comprehensive evaluation result of the building energy consumption model, and determining a final deployment model according to the comprehensive evaluation result. The invention further provides a building energy consumption model evaluation system based on the method, and guidance is provided for building energy consumption management personnel to select a proper building energy consumption model.Aiming at a specific building energy application scene, comprehensive evaluation is carried out on a set evaluation task according to the method, and a building energy consumption model with the optimal performance is selected for deployment, so that the formulation and implementation of a green building energy management operation strategy are guided, and the purpose of the efficient and energy-saving operation of a green building energy system is achieved.

Description

technical field [0001] The present application relates to the technical field of building energy consumption, in particular to a building energy consumption model evaluation method based on building energy management application scenarios. Background technique [0002] At present, building energy consumption modeling and forecasting play an important role in green building management, especially building energy management. Traditional building energy consumption models include physical models and statistical models. With the large-scale application of smart meters (electricity consumption metering system) and sensors (building automation system) in buildings, the availability of building energy consumption and environmental data has greatly improved. Improvement, the application of data-driven modeling methods (machine learning methods) in the field of building energy modeling and forecasting is increasing. Therefore, the importance of performance evaluation of building ene...

AI Technical Summary

Problems solved by technology

However, due to the stability of the machine learning model, the performance of the same model on different training sets may be different. For building energy consumption, due to changes in operating rules and factors affecting energy consumption, there are annual and seasonal differences in energy consumption levels. , the model based on a single training set cannot meet the actual application requirements

[0004] In addition, due to the numerous modeling methods of machine learning models, building energy managers currently lack clear standards and comprehensive evaluation methods for the selection of models in the actual building energy consumption analysis

The existing performance evaluation methods for building energy consumption models mainly focus on the evaluation of the accuracy of the model. The selection method of the machine learning model training set in the transition season building energy consumption prediction is not appropriate, and the evaluation of the building energy consumption model is based on The prediction performance of the current year cannot reflect the impact of fluctuations in building energy consumption data in different years on the stability of the model, resulting in poor model performance, less involvement in the application scenarios of building energy consumption models, and historical energy consumption characteristics on which modeling is based. And there is a lack of system tools for comprehensive, hierarchical, and multi-angle performance evaluation of multiple models formed by modeling tasks in different application scenarios

Images