A Method of Energy Recognition of Micro Energy Devices Based on BP Neural Network

Abstract

The present invention provides a method for identifying energy of a micro-energy device based on a BP neural network, comprising the following steps: S1, collecting the dynamic voltage of the micro-energy device in an open circuit state, obtaining an original voltage signal, and performing adaptive threshold wavelet on the original voltage signal Transform processing to remove noise; S2, extract the peak R of the voltage signal after denoising, so as to obtain the model input data; S3, establish a BP neural network model, input data to train the model, stop training when the training error is less than a predetermined value, and obtain a qualified BP neural network model; S4, using the BP neural network model obtained in step S3 to distinguish different types of micro-energy devices. The invention can carry out relatively accurate and rapid energy identification and classification, and the classification result is reliable; the method of the invention has strong anti-interference ability; several feature quantities with a higher influence ratio in the comparison of energy signals of micro-energy devices are selected.

Description

technical field [0001] The invention relates to the field of smart micro-energy systems, in particular to an energy identification method for micro-energy devices based on a BP neural network. Background technique [0002] At present, smart micro-energy systems featuring self-sensing, self-awakening, self-learning, and self-adaptation are gradually beginning to replace traditional energy management systems. Among them, accurate identification of input energy is the key to smart micro-energy systems. [0003] Since the voltage signal of the micro-energy device is interfered by external factors such as the material itself and the process technology, there may be a variety of noises with different strengths or frequencies in the voltage signal. It is one of the urgent problems to be solved in device energy identification. [0004] At present, the research directions of machine learning mainly include research on decision trees, random forests, artificial neural networks, and ...

AI Technical Summary

Problems solved by technology

Taking the smart micro-energy system as an example, if a common machine learning algorithm such as the naive Bayesian classifier algorithm is used to classify the output electrical forms of various micro-energy devices, the algorithm will only classify the current training set samples (that is, each The dynamic voltage when the micro energy device is open circuit) is used to classify and extract features, but the feature extraction effect for other data samples is poor

Images