Building Energy Lag Optimization for Demand Response Dispatch
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional weather normalization techniques fail to accurately estimate energy consumption as a function of temperature for periods less than 24 hours, leading to unreliable energy use predictions and inefficient demand response programs.
Innovation Solution
A demand response dispatch prediction system that includes a building lag optimizer, a dispatch prediction element, and a dispatch control element, which shifts energy use data relative to outside temperature values to determine a building's energy lag and generate accurate machine learning model parameters for predicting energy consumption and optimizing demand response events.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional weather normalization techniques are used, then energy consumption estimation is simplified, but prediction accuracy for periods less than 24 hours deteriorates
Solution Approach 1:
The patent segments the energy consumption prediction into multiple lag components (first lag, second lag, third lag, etc.), where each lag represents a different temporal response of the building to temperature changes. This segmentation allows the model to capture transient energy consumption patterns at finer granularities while maintaining operational simplicity through a structured multi-lag framework.
Solution Approach 2:
The patent introduces dynamic lag parameters that adapt to different building types and weather conditions. The energy consumption is modeled as a dynamic function of temperature with multiple time lags, allowing the system to respond to changing conditions while maintaining a manageable computational structure through standardized lag calculations.
2Measurement precision
If energy consumption data is analyzed at finer granularities, then prediction accuracy improves, but data processing complexity increases
Solution Approach 1:
The patent changes the parameter representation from raw time-series data to lagged temperature differences. By transforming the input parameters into lagged components (e.g., temperature at time t-1, t-2, t-3), the system can analyze fine-grained energy consumption patterns while reducing processing complexity through parameter transformation and standardized lag calculations.
Solution Approach 2:
The patent replaces complex mechanical data processing with a mathematical modeling approach using multiple linear regression. Instead of processing raw fine-grained data through complex algorithms, the system uses a regression model with lagged temperature parameters, substituting mechanical processing complexity with statistical modeling simplicity.
3Reliability
If transient energy consumption patterns are accounted for, then demand response program accuracy improves, but model complexity increases
Solution Approach 1:
The patent segments the transient energy consumption response into distinct lag components, where each lag represents a specific temporal phase of the building's thermal response. This segmentation allows the model to capture transient patterns accurately while maintaining manageable complexity through a structured multi-lag framework that can be implemented using standard regression techniques.
Solution Approach 2:
The patent introduces lagged temperature differences as intermediary variables that mediate between raw temperature data and energy consumption. These intermediary lag parameters simplify the complex transient response by breaking it down into manageable temporal components, making the model both accurate and implementable.
Data Source
AI summary
A method for dispatching buildings, including: generating data sets, each having energy values along with corresponding time and outside temperature values, wherein the energy values are shifted by one of a plurality of lag values relative to the corresponding time and outside temperature values; performing a machine learning model analysis on the each of the data sets; determining a least valued residual that indicates a corresponding energy lag for each of the buildings, the corresponding energy lag describes a transient energy consumption period preceding a change in outside temperature; using outside temperatures, model parameters, and energy lags for all of the buildings to estimate a cumulative energy consumption for the buildings, and to predict a dispatch order reception time for the demand response program event; and employing the dispatch order reception time to prepare actions required to control the each of the buildings to optimally shed energy specified in a dispatch order.