Physics-Informed Neural Network for Non-Intrusive Load Monitoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing non-intrusive load monitoring methods based on deep learning ignore domain knowledge, limiting the accuracy of load monitoring in smart grids, which is crucial for demand side management and response.
Innovation Solution
A non-intrusive load monitoring method using a physics-informed neural network that incorporates both active and reactive power data, with a physics-constrained learning framework to train a deep learning model, enhancing the model's ability to accurately forecast equipment load and improve interpretability.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional deep learning methods are used for load monitoring, then the model can learn from data, but the accuracy is limited due to ignoring domain knowledge
Solution Approach 1:
The patent merges deep learning models with physics-based power flow equations to create a physics-informed neural network. The neural network learns from data while the physics equations provide domain knowledge constraints, combining the strengths of both approaches to improve load monitoring accuracy without excessive complexity
Solution Approach 2:
The patent changes the parameter representation by using power flow equations that relate load parameters to observable quantities. By transforming the problem into the frequency domain and using impedance matrices, the model can accurately estimate load parameters while incorporating physical constraints
2Measurement precision
If only active power data is used for training, then the model is simpler, but the operating characteristics of nonresistive loads cannot be accurately described
Solution Approach 1:
The patent adds reactive power as another dimension to the input data, moving from single-dimensional active power to two-dimensional power data (active and reactive). This dimensional expansion enables accurate description of nonresistive load characteristics while maintaining manageable data quantity through efficient feature extraction
Data Source
AI summary
The present invention relates to the cross field of smart grid and artificial intelligence, provides a non-intrusive load monitoring method and device based on physics-informed neural network, comprising the following steps: Step 1, obtaining a total load data and an equipment load data of a building in a certain period of time, and using a sliding window method to cut to construct a training data. Step 2, designing a deep learning neural network model to learn the equipment load characteristics contained in the total load data, and outputting the equipment load forecasting. Step 3, based on a physics-constrained learning framework, training the deep learning neural network model by iteratively optimizing the training loss to obtain a trained physics-informed neural network model. Step 4, monitoring the equipment's power consumption in the building according to the output results of the physics-informed neural network model. The present invention can fully extract the operation characteristics of electric equipment, and improve the accuracy of load identification without increasing additional cost.