A deep learning-based method for predicting the remaining life of fuel cells

Abstract

The invention relates to a method for predicting the remaining service life of a fuel cell based on deep learning. The method adopts a deep neural network model to realize accurate prediction of fuel cell RUL, and belongs to the technical field of proton exchange membrane fuel cell health status monitoring. The specific steps are: S1: Obtain the monitoring data of the fuel cell, and perform noise reduction processing; S2: Fit the training data after noise reduction processing, and establish the nonlinear mapping relationship between the current input data and the target data; S3: Establish Loss function and optimize the parameters of the neural network model through the method of feedback derivation to obtain the optimal neural network model; S4: The input data of the prediction starting point is used as the input of the optimal DNN to realize iterative rolling prediction. S4 can be simply summarized as: iteratively use the current prediction output of the neural network as the input of the next prediction, and iteratively predict the voltage or power output, so as to realize the fuel cell RUL prediction.

Description

technical field [0001] The invention relates to a method for predicting the remaining life of a fuel cell based on deep learning, and belongs to the technical field of proton exchange membrane fuel cells. Background technique [0002] As an environmentally friendly and clean energy, fuel cells have a good prospect in the application of new energy fuel cell electric vehicles by virtue of their great advantages in power conversion efficiency. However, compared with other energy devices such as lithium-ion batteries, high operating cost, short life cycle and safe and reliable operation are still the main reasons for the widespread application of fuel cells. Such as the most popular proton exchange membrane fuel cell system (PEMFC), its main characteristics are fast start-up performance, high power density (3.8 ~ 6.5kW / m3) and low operating temperature (50 ~ 80 ℃). System Prediction and Health Management (PHM) is an emerging field of scientific and technological development, wh...

AI Technical Summary

Problems solved by technology

However, the current fusion algorithm largely depends on whether the exponential or logarithmic model it constructs can accurately describe the attenuation of battery life.

Images