Non-invasive load decomposition method based on residual convolution and attention mechanism
A non-invasive, load decomposition technology, applied in reasoning methods, neural learning methods, character and pattern recognition, etc., can solve the problems of user privacy leakage, high installation and maintenance costs of devices
Pending Publication Date: 2022-07-08
XIANGTAN UNIV
View PDF0 Cites 4 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
The intrusive load monitoring method needs to install a monitoring device for each electric load for real-time monitoring, and then complete the transmission of load data by means of local storage or wireless communication, and analyze the load data on the host computer to realize optimal scheduling of load power consumption; This monitoring method has a high degree of data recovery and accuracy, but the cost of device installation and maintenance is high; in addition, user privacy leakage may occur in the installation and monitoring of various loads, and most users may have resistance
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0020] The present invention proposes a non-intrusive resident load decomposition method based on residual convolution and attention mechanism.
[0021] The method includes the following steps:
[0022] Step1: Collect power data by using non-intrusive load monitoring and decomposition (NILMD) measuring device, perform data preprocessing, and use a large number of sample data to build a database;
[0023] Step2 Non-intrusive load decomposition performs feature extraction on input data through residual convolutional neural network to obtain feature map;
[0024] Step3 introduces an attention mechanism to further process feature data, extract favorable features more effectively, discard useless redundant features, and improve the efficiency of non-intrusive load decomposition;
[0025] Step4 outputs the decomposition result through the fully connected layer;
[0026] Step 5 Repeat Step 2-Step 4 to train the decomposition model with sample data, adjust parameters, and build a no...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a non-intrusive load decomposition method based on residual convolution and an attention mechanism. Load features are extracted through the residual convolutional neural network, and the idea of cross-layer connection is used for reference, so that the defects of gradient dispersion and performance degradation caused by network deepening are well overcome, information transmission before and after the network is smoother, and the problems of gradient disappearance and the like of a network model are avoided. And then, an attention mechanism is introduced, useless redundant data is eliminated, the characterization capability of the data is further enhanced, and a decomposition result is output through a full connection layer. A large amount of sample data is used for training a non-intrusive load decomposition model, model parameters are continuously and finely adjusted, an optimized decomposition model is constructed, and the optimized non-intrusive load decomposition model is used for completing load decomposition.
Description
technical field [0001] The present invention relates to a non-intrusive load decomposition method based on residual convolution and attention mechanism, which can be widely applied to non-intrusive load online monitoring of residential electricity loads of contracted users in urban and rural areas, and has high accuracy. , the decomposition speed is fast and so on. Background technique [0002] Electric energy plays a vital role in the development of society. Human demand for electric energy continues to increase. The rational use of electric energy has a significant impact on the economic development of the entire society. Therefore, effectively improving the utilization rate of electric energy and rationally planning the distribution of electric energy resources are the urgent needs to solve the problem of sustainable social development. Under normal circumstances, according to the electricity meter installed outdoors, the user only knows how much electricity is used in t...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More IPC IPC(8): G06K9/00G06N3/04G06N3/08G06N5/04
CPCG06N3/084G06N5/041G06N3/048G06N3/045G06F2218/08Y04S20/242
Inventor易灵芝许翔翔刘江永廖剑霞刘西蒙罗晓雪
OwnerXIANGTAN UNIV