Air handling unit fault detection and diagnosis method combining dual-channel convolutional neural network and light gradient elevator

Abstract

The invention discloses an air handling unit fault detection and diagnosis method combining a dual-channel convolutional neural network (DCCNN) and a light gradient elevator (LGBM), and the method employs a hybrid model DCCNN-LGBM combining two advanced classifiers: the DCCNN and the LGBM, and accurate fault detection and diagnosis are carried out on the air handling unit (AHU). According to the method, the residual network is used in the convolutional neural network, the feature extraction capability of the dual-channel convolutional neural network is improved, the overall fault detection anddiagnosis precision of the model is improved, and the method can be effectively applied to accurate fault detection and diagnosis of the air handling unit in actual engineering. Moreover, the portability of the model is very good, and the model can be easily transplanted to other fields, such as the fields of fault detection and diagnosis of the water chilling unit, only by changing some parameters of the model.

Description

technical field [0001] The invention relates to the field of fault detection, and more specifically, relates to a fault detection and diagnosis method of an air handling unit combined with a dual-channel convolutional neural network and an optical gradient lifter. Background technique [0002] Tunnel heating, ventilation and air conditioning (HIVC) systems are an important component in industrial and domestic buildings. The system plays a leading role in regulating the indoor environment, providing people with a comfortable and safe working and living environment. In order to meet the increasing requirements of modern buildings for indoor environmental quality, HVAC systems are becoming more and more complex, and their energy consumption is also increasing, accounting for more than 40% of the total global building energy consumption. The operation of HIVC system in a fault state will not only shorten the service life of the equipment, but also cause additional loss, which a...

AI Technical Summary

Problems solved by technology

First, the above-mentioned feature extraction methods all have fixed modes, which are effective for simple systems, because their state signals often have obvious frequency characteristics

However, for a highly nonlinear complex system such as AHU, these feature extraction methods are relatively weak, and can only extract one-sided features within the scope of their own algorithm settings.

Second, the feature extraction process and the subsequent fault detection and diagnosis process are independent of each other, and there is no joint optimization between the two

The classifier can only use the features extracted by the feature extractor, and it is impossible to determine whether the extracted features are defective or redundant, which is not conducive to the fault detection and diagnosis of the system

Images