Non-intrusive fuel cell fault diagnosis method

Abstract

The invention discloses a non-intrusive fuel cell fault diagnosis method. The method comprises fault diagnosis classifier establishing and fault diagnosis testing. The fault diagnosis classifier establishing comprises the following steps of taking the actual operation data of a fuel cell as an original data set extraction diagnosis variable; preprocessing the diagnosis variable; training preprocessed data through an ELM learning algorithm and screening a clustering result which is the same with an actual fault label as a fault sample set; and through an ELM model, learning the fault sample setand outputting an ELM classification model so as to form a fault diagnosis classifier. The fault diagnosis testing comprises the following steps of detecting the data to be diagnosed of the fuel cell; through the data to be diagnosed, establishing a test sample; and sending the test sample into the ELM classification model for testing and outputting a test result. In the invention, the fault diagnosis of the fuel cell can be performed in a non-intrusive mode, and the timeliness, the reliability and the accuracy of fuel cell fault diagnosis are effectively increased.

Description

technical field [0001] The invention belongs to the technical field of fuel cells, in particular to a non-invasive fuel cell fault diagnosis method. Background technique [0002] Fuel cells (FC) are considered to be one of the most promising power generation systems in the future. Because the proton exchange membrane fuel cell (PEMFC) has the advantages of quick start at room temperature, low operating temperature, high power generation efficiency, no pollution, low noise, and flexible use, it has been widely used in portable power supplies, vehicle power supplies, and household power supplies. Research. Despite the desirable properties of PEMFCs, the major hurdles to commercialization today are reliability and durability. For PEMFC systems, through effective diagnosis, early fault warning can be realized, so as to avoid more serious faults. Based on the diagnostic results, operating conditions may be adjusted to allow the fuel cell to operate efficiently and safely. In ...

AI Technical Summary

Problems solved by technology

[0004] Although a variety of fault diagnosis methods for PEMFC have been proposed, on the one hand, due to the complexity of fuel cells, there is no unified standard for fuel cell performance and faults; on the other hand, these methods are basically in the experimental stage , the experimental database has limitations, no systematic large database has been formed, and experiments are only carried out for their respective experimental platforms, which lacks versatility

And each has different defects, such as invasive diagnosis cannot be widely used commercially, EIS diagnosis requires additional external diagnostic equipment and the cost is currently high

Existing methods have high time cost and complicated diagnosis process

Images