State Model Selection for Energy Disaggregation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing disaggregation systems face poor estimation accuracy due to the use of inappropriate state model structures, leading to increased complexity in estimating appliance states, as fully connected state models do not suit all types of appliances.
Innovation Solution
A state model structure selection apparatus that computes change and repetition characteristics from time series data to select between fully connected and one-way direction state models, reducing complexity and improving estimation accuracy by assigning appropriate models to each appliance based on its waveform patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a fully connected state model is used for all appliances, then the model can handle any state transition pattern, but the complexity of estimating appliance states increases significantly
Solution Approach 1:
The patent segments the single fully connected state model into multiple specialized state models (fully connected type, one-way direction type, and independent type) that can be selectively applied to different appliances based on their characteristics. This segmentation reduces the complexity each appliance model needs to handle while maintaining overall system versatility.
Solution Approach 2:
The patent introduces dynamic selection of state model types based on appliance characteristics and operational data. The system can adaptively choose which state model type to use for each appliance, making the system flexible and reducing unnecessary complexity for appliances that don't require full connectivity.
2Ease of manufacture
If a fixed state model structure is used for all appliances, then the model structure is simple to implement, but the estimation accuracy deteriorates for appliances with different waveform patterns
Solution Approach 1:
The patent applies local quality by allowing different state model structures to be used for different appliances based on their specific characteristics. Instead of a uniform approach, each appliance can have a state model type (fully connected, one-way direction, or independent) that best matches its operational patterns, thereby improving estimation accuracy while maintaining implementation simplicity through clear classification criteria.
3Measurement precision
If the state model structure is selected based on detailed waveform analysis, then the estimation accuracy improves, but the computational complexity and processing time increase
Solution Approach 1:
The patent uses parameter changes by analyzing specific characteristics of appliance waveforms (such as repetition patterns and state transition frequencies) to determine the appropriate state model type. This approach improves accuracy by matching model parameters to appliance characteristics while controlling complexity through focused analysis of key parameters rather than comprehensive waveform examination.
Data Source
AI summary
Provided an apparatus that receives time series data from a data storage unit storing time series of sample data or feature values calculated from the sample data, computes a measure indicating change and repetition characteristics of the time series data, based on sample value distribution thereof, selects a state model structure to be used for model learning and estimation, from state models including a fully connected state model and a one way direction state model, based on the measure and stores the selected state model in a model storage unit.