Fuel cell fault prediction method based on self-updating of impedance prediction model

Abstract

The invention relates to a fuel cell fault prediction method based on self-updating of an impedance prediction model, and the method comprises the steps of collecting the operation condition information and impedance information of a fuel cell in real time, carrying out the preprocessing, and loading the information into an impedance prediction model, thereby obtaining a predicted impedance value; acquiring predicted impedance values of at least three frequency points, extracting characteristic parameters reflecting a water content state, and then acquiring a prediction result of a water content fault through a classifier, wherein the impedance prediction model is an LSTM network, the input of the impedance prediction model is an actually measured impedance value and a predicted impedance value at a previous moment, and the output of the impedance prediction model is a predicted impedance value at a next moment. Compared with the prior art, the invention has the advantages that water content fault prediction is achieved, then prediction control over the water content state is achieved, the problems that the change response of the water content in the fuel cell is lagged, and control is not timely are solved, input and output of the impedance prediction model are data capable of being directly measured, self-updating of the model is achieved, and the problem that the model prediction effect becomes poor due to aging of the fuel cell is solved.

Description

technical field [0001] The invention relates to the field of fuel cell failure prediction, in particular to a fuel cell failure prediction method based on self-updating impedance prediction model. Background technique [0002] In recent years, energy and environmental issues have received widespread attention from the society, and fuel cells have become a current research hotspot because of their advantages such as high efficiency, no pollution, and renewable energy sources, especially in the field of new energy vehicles. . [0003] During the operation of the fuel cell, the content, transmission, distribution and physical state of the internal water will directly affect the health of the fuel cell. If the operation is not done properly, it will easily cause water flooding or membrane drying, which will lead to performance degradation and life of the fuel cell system. attenuation. With the improvement of fuel cell power level and power density, the water management problem...

AI Technical Summary

Problems solved by technology

However, the inability to accurately and quickly identify the water state inside the vehicle-mounted fuel cell has always been an important reason for the poor water management of the fuel cell. As a result, the reliability and environmental adaptability of the fuel cell are directly affected, and it is a bottleneck that hinders the commercialization of fuel cells. technology

[0004] At the same time, the response to the change of the water content state inside the vehicle fuel cell is relatively lagging. Even if the water content state can be accurately and quickly identified, it is difficult to restore it to the normal operating state immediately through the control strategy.

Images