Conditional Factorial Hidden Semi-Markov Model for Power Load Disaggregation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Residential energy conservation efforts are hindered by the lack of cost-effective, appliance-specific energy usage breakdowns provided by smart meters, which limits the effectiveness of energy conservation measures.
Innovation Solution
The use of unsupervised variants of factorial hidden Markov models, such as conditional factorial hidden semi-Markov models, to disaggregate power load and provide per-appliance energy usage information without the need for labeled data or extensive instrumentation, leveraging features like time of day and appliance dependencies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If smart meters provide aggregate energy consumption data, then energy usage information is available, but appliance-specific breakdown is not provided
Solution Approach 1:
The patent replaces physical instrumentation (electrical sensors, meters, and measurement devices) with computational analysis. By using factorial hidden Markov models to analyze aggregate power data, the system extracts appliance-specific information without requiring additional physical measurement devices for each appliance.
Solution Approach 2:
The patent creates a computational model that replicates the function of individual appliance meters. The factorial hidden Markov model generates virtual measurements of each appliance's power consumption by analyzing patterns in the aggregate data, effectively copying what individual meters would provide without the physical infrastructure.
2Loss of energy
If efficiency upgrades are implemented, then energy consumption is reduced, but cost is increased
Solution Approach 1:
The patent implements a feedback mechanism that provides homeowners with real-time or near-real-time information about which specific appliances are consuming the most energy. This actionable feedback enables targeted behavioral changes rather than requiring blanket efficiency upgrades across all appliances, reducing implementation costs while achieving energy savings.
Solution Approach 2:
The system enables homeowners to self-identify energy waste patterns and make informed decisions about which appliances to adjust or upgrade. By providing appliance-specific breakdowns, the system allows residents to take autonomous action on energy conservation without requiring expensive professional assessments or comprehensive appliance replacements.
3Loss of energy
If curtailment is practiced, then energy consumption is reduced, but consumer understanding of energy use is not improved
Solution Approach 1:
The patent provides detailed feedback that breaks down aggregate energy consumption into appliance-specific components. This feedback loop helps consumers understand which appliances contribute most to their energy usage, enabling them to make informed curtailment decisions based on actual usage patterns rather than estimates or general advice.
Solution Approach 2:
The patent segments aggregate energy consumption data into individual appliance contributions using factorial hidden Markov models. This segmentation transforms undifferentiated energy usage information into discrete, actionable insights about each appliance's consumption patterns, helping consumers understand and address energy waste at the appliance level.
Data Source
AI summary
Systems and methods of disaggregating power load are provided. An example of a method is carried out by program code stored on non-transient computer-readable medium and executed by a processor. The method includes receiving time series data representing total energy consumption. The method also includes identifying distinguishing features in the time series data. The method also includes identifying energy consumption constituents of the total energy consumption based on the features.