Industrial air conditioner fault diagnosis method based on neural network and integrated learning fusion

Abstract

The invention provides an industrial air conditioner fault diagnosis method based on a neural network and an integrated learning algorithm. The method fuses various machine learning algorithms, and is a fault diagnosis model with high accuracy. The neural network automatically extracts output and a 'reasonable rule' between output data through learning, and plays an important role in analyzing time sequence data received by the industrial air conditioner in real time. Therefore, an anomaly detection module based on a neural network algorithm is designed. The integration algorithm has the obvious advantages of avoiding overfitting, improving generalization ability and the like. Therefore, a fault diagnosis module based on an ensemble learning algorithm is designed. An anomaly detection model is combined with a fault diagnosis model, and an attention mechanism is introduced, so that detected anomaly auxiliary fault diagnosis forms an enhanced model which is more accurate than a basic model. According to the invention, accurate fault diagnosis can be carried out on industrial air conditioner data.

Description

technical field [0001] The invention relates to the field of the Internet and the field of deep learning, in particular to an industrial air-conditioning fault diagnosis method based on the fusion of neural network and integrated learning. Background technique [0002] An industrial air-conditioning fault diagnosis method based on the fusion of neural network and ensemble learning, based on neural network technology and ensemble learning technology. LSTM is a neural network algorithm that implements a mapping function from input to output. Mathematical theory proves that a three-layer neural network can approximate any nonlinear continuous function with arbitrary precision. It is suitable for solving problems with complex internal mechanisms. The integrated learning algorithm of LightGBM has the obvious advantages of avoiding overfitting and improving generalization ability. The technologies closest to the present invention in recent years are: [0003] (1) XGBoost-based ...

AI Technical Summary

Problems solved by technology

But a single monitoring method is weak in fault diagnosis

Images