Adaptive modeling method and system for MPC-based building energy control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing building energy management systems using model predictive control (MPC) face inefficiencies due to fixed thermal models that do not adapt to changes in occupancy and other thermal properties, leading to inaccuracies and potential energy waste or occupant discomfort.
Innovation Solution
An adaptive modeling method that uses probabilistic inference to identify and update the thermal model based on temperature measurements and sensor data, adjusting for changes in occupancy, insulation, and other factors to improve accuracy and reduce errors in temperature predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a fixed thermal model is used in MPC for building energy management, then the control system is simple and computationally efficient, but the model accuracy deteriorates over time as occupancy and thermal properties change
Solution Approach 1:
The patent implements dynamic model adaptation by continuously updating thermal model parameters (occupancy, insulation properties, heat capacity) based on sensor measurements and probabilistic inference. The model transitions from a static fixed structure to a dynamic system that automatically adjusts to changing building conditions, resolving the contradiction between model accuracy and complexity through automated adaptation mechanisms.
Solution Approach 2:
The system employs feedback loops where temperature sensors continuously monitor actual building conditions, compare them against model predictions, and use the discrepancies to trigger probabilistic inference processes. This feedback mechanism automatically updates model parameters when deviations exceed thresholds, maintaining high accuracy without requiring complex manual intervention.
2Reliability
If manual model adjustment is performed by expert engineers, then the model can be tuned for specific conditions, but the system requires significant engineering investment and becomes obsolete when experts leave
Solution Approach 1:
The patent implements self-service automation through probabilistic inference algorithms that automatically detect model inaccuracies and adjust parameters without human intervention. The system monitors temperature deviations, infers the likely causes (occupancy changes, insulation degradation, equipment failures), and updates the thermal model autonomously, eliminating dependency on expert engineers and ensuring continuous operation.
Solution Approach 2:
The system replaces the mechanical process of manual expert tuning with an automated computational inference mechanism. Instead of engineers physically adjusting model parameters based on experience, the system uses sensor data and probabilistic algorithms to automatically update the thermal model, substituting human expertise with an automated intelligent system.
3Measurement precision
If sophisticated thermal models are used to represent building physics, then the model accuracy is high, but the computational intensity becomes too great for real-time MPC simulations
Solution Approach 1:
The patent dynamically adjusts model parameter complexity based on operational conditions. The thermal model uses detailed physical parameters (heat capacity, insulation resistance, thermal conductance) only when and where needed, rather than maintaining full complexity continuously. This selective parameter activation maintains high prediction accuracy for critical zones while reducing overall computational energy consumption.
Solution Approach 2:
The building is divided into multiple thermal zones, each with its own simplified thermal model. Rather than creating one comprehensive complex model of the entire building, the system segments the building into manageable zones that can be simulated independently and in parallel, reducing total computational energy requirements while maintaining overall prediction accuracy.
4Productivity
If the thermal model does not adapt to changes in occupancy and building properties, then the system is stable and easy to operate, but energy efficiency deteriorates due to inaccurate predictions
Solution Approach 1:
The system uses feedback from temperature sensors and occupancy detectors to continuously monitor building conditions. When deviations from predicted behavior exceed predefined thresholds, the probabilistic inference system automatically triggers model parameter updates, enabling the system to adapt to changing conditions and maintain high energy efficiency without complex manual intervention.
Solution Approach 2:
The thermal model performs self-service adaptation by automatically detecting when its predictions diverge from actual measurements and autonomously updating its parameters through probabilistic inference. This self-adjusting capability enables the system to maintain optimal energy efficiency while avoiding the complexity of external manual tuning processes.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach enhances the accuracy of building energy management by continuously updating the thermal model, reducing errors and optimizing energy usage and costs, allowing for more efficient and autonomous control of building temperatures.
Implementation Method 1
insulation between spaces like windows and walls act as thermal resistors, slowing down the flow of thermal energy between two spaces with different temperatures
Implementation Method 2
the air and objects in a room act as thermal capacitors, storing thermal energy and slowing down how quickly power raises the temperature of the room
Implementation Method 3
The current source, PSOURCE, represents the objects and people in the room that produce thermal power. This can include the HVAC system which regulates temperature in the room, a small space heater, human bodies, and electrical equipment like computers or refrigerators
Data Source
AI summary
A method for controlling a temperature of a building using a building thermal model of the building in a model predictive control (MPC) system includes measuring the temperature and a rate of change of temperature for at least one zone that is included in the building thermal model so as to determine an error from a temperature and rate of change of temperature that is predicted by the MPC system. Possible causes of the error and a probability of each of the possible causes of error occurring are determined. An impact of each of the possible causes is evaluated so as to identify at least one of the possible causes which would reduce the error. The building thermal model is adapted based on the at least one identified possible cause and the temperature of the building is controlled using the adapted building thermal model in an MPC controller.