Low Frequency Energy Disaggregation Using Fourier Transform
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing energy disaggregation methods fail to provide accurate, near real-time information for users to modify their energy usage effectively, especially with lower resolution data from advanced metering infrastructure (AMI) devices, which limits their ability to identify specific appliance signatures and provide actionable insights.
Innovation Solution
A method that uses a processor to analyze consumption data from AMI devices at 15-minute, 30-minute, or 60-minute intervals, detecting active and inactive signals to estimate appliance usage, including water heating, lighting, refrigeration, and vacation modes, using behavior modeling, data smoothing, and machine learning models to provide itemized appliance consumption.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high sampling rate data (1-1,000,000 samples per second) is used for energy disaggregation, then appliance signature extraction accuracy is improved, but data availability and cost increase significantly
Solution Approach 1:
The patent transforms the problem by changing the temporal parameter of data sampling. Instead of using high-frequency samples (1-1,000,000 samples per second), the invention uses low-frequency aggregated data (15-minute, 30-minute, or 60-minute intervals) from AMI meters. The key innovation is applying Fourier transform and frequency domain analysis to extract appliance signatures from this down-sampled data, thereby maintaining measurement precision while dramatically reducing data volume and cost.
2Quantity of substance
If low resolution data from AMI devices (15-minute, 30-minute, or 60-minute intervals) is used, then data availability and cost are reduced, but appliance signature identification capability deteriorates
Solution Approach 1:
The patent applies frequency domain analysis (Fourier transform) to detect characteristic frequencies and patterns in the aggregated consumption data. By transforming the time-domain low-resolution data into the frequency domain, the system can identify unique appliance signatures based on their operational frequencies and harmonic content, even though the original sampling rate is too low for direct time-domain analysis.
Solution Approach 2:
The patent introduces an intermediary processing layer (Fourier transform and signal processing algorithms) between the raw AMI data and the appliance signature extraction. This intermediary transforms the insufficient low-resolution data into a form that reveals appliance characteristics, effectively bridging the gap between limited input data and the required identification capability.
3Loss of information
If software analysis is performed on past data collected at high sampling rates, then detailed appliance usage patterns can be identified, but near real-time information for immediate user action is not provided
Solution Approach 1:
The patent performs preliminary aggregation of consumption data into meaningful time intervals (15-minute, 30-minute, or 60-minute blocks) before analysis. This preliminary action organizes the data in a way that enables both detailed pattern recognition and timely delivery of insights. The system pre-processes the data structure to facilitate rapid generation of actionable recommendations without requiring analysis of every individual high-frequency sample.
Data Source
AI summary
The present invention teaches methods of performing appliance itemization based on consumption data, including: receiving at a processor the data; determining if the data includes active signals and/or inactive signals; upon detection of an active signal: detecting and estimating active water heating consumption and lighting consumption; upon detection of an inactive signal: detecting and estimating passive water heating consumption, refrigerator consumption; and detecting vacation mode. Methods are disclosed of appliance itemization based whole house consumption data consumption from an advanced metering infrastructure device, the data being at 15, 30, or 60 minute intervals, including: applying disaggregation models to provide detection and estimation of any lighting, water heating, refrigeration, pool pumps, heating, or cooling appliances; applying rule-based models to provide detection and estimation of any cooking, laundry, entertainment, and/or miscellaneous appliances; wherein the disaggregation models and the rule-based models provide for a near complete appliance level itemization and estimation.