Machine Learning Load Disaggregation for Target Device Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Disaggregating device-specific energy usage within household energy consumption is challenging due to the variety of devices and limited availability of training data, leading to inefficiencies in grid planning and energy management.
Innovation Solution
Implementing a machine learning-based system that uses trained models to predict the presence of target device energy usage by analyzing total energy usage data from households, leveraging deep learning schemes and neural networks to improve accuracy and resource efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional metering devices are used to monitor household energy usage, then general energy consumption can be measured, but device-specific energy usage cannot be reliably discovered
Solution Approach 1:
The patent introduces machine learning models as intermediary components between the metering device and the energy usage data. These models act as mediators that process aggregate energy consumption data and infer device-specific usage patterns without requiring direct measurement of each device, thus improving detection precision while avoiding the complexity of installing separate sensors on every device.
Solution Approach 2:
The patent replaces traditional mechanical/electrical measurement approaches with computational intelligence methods. Instead of using physical sensors and circuitry to directly measure each device's energy consumption, the system uses trained machine learning models to analyze aggregate data and predict device-specific usage, substituting a computational system for a complex physical measurement system.
2Measurement precision
If machine learning models are trained to discover target device energy usage, then prediction accuracy improves, but training data availability becomes limited
Solution Approach 1:
The patent applies preliminary action by pre-training machine learning models using available training data before deployment. The models are trained in advance on datasets that contain labeled energy consumption patterns from various devices, allowing them to learn device-specific signatures beforehand. This preliminary training enables the models to accurately discover target device presence even when limited new data is available during actual operation.
3Reliability
If multiple discovery predictions are generated over time, then overall prediction reliability improves, but processing time increases
Solution Approach 1:
The patent implements periodic action by generating multiple discovery predictions at different time intervals and then aggregating these predictions to form an overall determination. The system periodically processes energy usage data, generates individual predictions, and combines them over time, which improves reliability through multiple observations while managing processing time through structured periodic execution rather than continuous analysis.
Data Source
AI summary
Embodiments generate machine learning predictions to discover target device energy usage. One or more trained machine learning models configured to discover target device energy usage from source location energy usage can be stored. Multiple instances of source location energy usage over a period of time can be received for a given source location. Using the trained machine learning model, multiple discovery predictions for the received instances of source location energy usage can be generated, the discovery predictions comprising a prediction about a presence of target device energy usage within the instances of source location energy usage. And based on the multiple discovery predictions, an overall prediction about a presence of target device energy usage within the given source location's energy usage over the period of time can be generated.