Building Load Control Using Classifier-Based Digital Model Automation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The manual creation of digital building models for controlling heating, air conditioning, and ventilation systems is time-consuming and costly, limiting the efficiency and cost-effectiveness of load control in buildings.
Innovation Solution
An automated method using a classifier trained with data from similar buildings to determine building model parameters, allowing for efficient creation and control of digital building models, which includes determining relevant building data such as geometry, thermodynamic, and system technology data, and using machine learning techniques like artificial neural networks to optimize model parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If manual methods are used to create digital building models, then model accuracy can be ensured, but the process becomes time-consuming and expensive
Solution Approach 1:
The patent uses building data from existing calibrated buildings as templates to create digital models for new buildings. Instead of manually creating each model from scratch, the system copies and adapts parameters from similar buildings, significantly reducing creation time while maintaining accuracy through the classifier-based adaptation process
Solution Approach 2:
The system performs preliminary calibration of building models in advance by training a classifier on data from multiple calibrated buildings. This pre-trained classifier can then quickly determine parameters for new buildings without requiring time-consuming manual calibration each time, resolving the contradiction between accuracy and time consumption
2Manufacturing precision
If manual methods are used to create digital building models, then model quality can be maintained, but costs increase
Solution Approach 1:
The patent reuses building data and parameters from existing calibrated buildings to create models for new buildings. This copying approach eliminates the need for expensive manual creation processes while maintaining model quality through the classifier-based parameter determination that adapts copied data to match specific building characteristics
Solution Approach 2:
The system creates a universal classifier that can determine parameters for multiple different building types by learning from diverse training data. This single classifier serves multiple functions across different buildings, reducing the need for expensive specialized manual work for each individual building while maintaining quality standards
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method (10) for controlling loads in a building, in particular heating systems, air conditioning systems and ventilation systems, wherein the method comprises the following steps: - determining building data for a building to be controlled (12); - determining building model parameters for a digital building model (14); wherein the determination of the building model parameters includes the use of a classifier previously trained with training data and the classifier is designed to determine the building model parameters depending on the determined building data; - controlling the loads in the building depending on the determined building model (16).