Building HVAC Cascaded Model for Zone Temperature Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Model predictive control systems for building HVAC face challenges in capturing temperature dynamics, particularly due to the presence of both linear and nonlinear components, and the lack of available training data for heat disturbances, which complicates the prediction and control of building zone temperatures.
Innovation Solution
A cascaded model comprising a disturbance model and a physics model is used, where the disturbance model predicts heat disturbances as a function of exogenous parameters and the physics model predicts temperature based on heat disturbances and HVAC inputs, allowing for the generation of control signals without requiring direct measurement of heat disturbances, and enabling training without explicit heat disturbance data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single predictive model is used to capture both linear and nonlinear temperature dynamics, then the model structure becomes simpler, but the model accuracy deteriorates due to inability to properly represent both linear and nonlinear components
Solution Approach 1:
The predictive model is segmented into two separate models: a linear model for capturing linear temperature dynamics and a nonlinear model for capturing nonlinear temperature dynamics. This segmentation allows each model to specialize in its respective domain, improving overall prediction accuracy while maintaining manageable complexity through modular architecture.
2Reliability
If models requiring heat disturbance data as training data are used, then the model can capture heat disturbance effects, but the training process becomes difficult due to lack of available heat disturbance measurement data
Solution Approach 1:
An intermediary model (the linear model) is introduced that relates measurable quantities (temperature, HVAC inputs) to the unmeasurable heat disturbance. This intermediary allows the system to infer heat disturbance effects indirectly through observable temperature dynamics, enabling training without direct heat disturbance measurements.
3Measurement precision
If a nonlinear model is used to capture all temperature dynamics, then the model can represent complex heat transfer phenomena, but the model becomes difficult to train due to lack of training data and computational complexity
Solution Approach 1:
The model is segmented into linear and nonlinear components, where the linear model handles straightforward conduction and the nonlinear model handles convection and radiation. This division reduces training difficulty for each individual model while collectively capturing all temperature dynamics phenomena.
Solution Approach 2:
The model transitions from a single complex nonlinear model to a combination of linear and nonlinear models with different parameter sets. This parameter change strategy allows each model to be trained on appropriately scaled and transformed data, reducing training difficulty while maintaining comprehensive dynamic representation.
Data Source
AI summary
A controller for HVAC equipment stores a cascaded model that includes a disturbance model configured to predict a heat disturbance affecting the building zone as a function of one or more exogenous parameters and a physics model configured to predict a temperature of the building zone as a function of the heat disturbance and an amount of heating or cooling provided to the building zone by HVAC equipment. The processing circuit is configured to execute a combined training procedure to determine parameters of the disturbance model and parameters of the physics model, generate control signals for the HVAC equipment using the disturbance model to predict the heat disturbance and applying the heat disturbance as an input to the physics model, and operate the HVAC equipment to provide the heating or cooling to the building zone in accordance with the control signals.