Building Energy Model Coefficient Drift Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems fail to accurately measure and verify energy savings and peak demand reductions in buildings due to unaccounted changes in static factors, leading to inaccurate energy usage estimations and potential contractual violations.
Innovation Solution
A building management system with sensors and a controller that generates energy usage models using regression coefficients from different time periods, identifies changes in energy usage by comparing model coefficients, and automatically adjusts for changes in static factors to provide accurate energy savings calculations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If baseline models are used to estimate energy usage, then energy savings calculations can be performed, but changes in static factors cause inaccurate estimations leading to verification failures
Solution Approach 1:
The patent applies dynamics by transitioning from static baseline models to dynamic models that continuously adapt to changing conditions. The system uses recursive regression analysis to update model coefficients in real-time, allowing the energy model to adapt to changes in static factors such as building occupancy, weather patterns, and equipment usage, thereby maintaining measurement accuracy and verification reliability
Solution Approach 2:
The patent implements feedback mechanisms by continuously monitoring energy consumption data and comparing actual usage against predicted baseline values. When deviations are detected that suggest changes in static factors, the system triggers model recalibration using recursive regression, creating a closed-loop feedback system that self-corrects for changing conditions and maintains accurate energy savings verification
2Measurement precision
If manual monitoring of static factors is performed, then changes can be detected, but changes may go unnoticed or be incorrectly calculated
Solution Approach 1:
The patent applies self-service by enabling the system to automatically detect and respond to changes in static factors without human intervention. The recursive regression analysis autonomously identifies when model coefficients drift from expected ranges, automatically triggers recalibration, and adjusts the baseline model accordingly, eliminating the need for manual monitoring while ensuring accurate detection of static factor changes
Solution Approach 2:
The patent replaces manual monitoring mechanisms with automated computational systems. Instead of human operators reviewing static factor data, the system uses recursive regression algorithms and statistical analysis to automatically detect changes in building characteristics, occupancy patterns, and environmental conditions, substituting mechanical human judgment with automated computational detection
3Measurement precision
If changes in static factors are not accounted for, then energy savings may be underestimated, but implementing manual adjustments increases complexity
Solution Approach 1:
The patent merges the detection of static factor changes with the existing baseline modeling process. Instead of adding separate manual adjustment procedures, the system integrates recursive regression analysis into the baseline model itself, combining change detection, model updating, and energy savings calculation into a unified automated framework that maintains accuracy without increasing operational complexity
Data Source
AI summary
A method for operating HVAC equipment includes obtaining building data for each of a plurality of time steps. The building data relates to resource usage of HVAC equipment. The method also includes calculating a recursive residual for each of a plurality of overlapping time periods using the building data. Each overlapping time period includes a subset of the plurality of time steps. The method also includes, for each of the plurality of time steps, calculating a metric based on the recursive residuals for overlapping time periods that end on or before the time step and automatically detecting a change in static factors for one or more buildings served by the HVAC equipment by comparing the metrics for the plurality of time steps based on a statistical property of the metrics.