<p>As a key component of the fuel cell system, the air compressor plays a vital role in ensuring the stability and reliability of fuel cell systems. To conduct a durability study, a 5000-h durability test of the fuel cell air compressor (FCAC) was performed according to the designed test profile, which reflects the real-world operating conditions of fuel cell vehicles. Based on the collected durability test data, the performance degradation characteristics of the compressor over time were analyzed, and a remaining useful life (RUL) prediction model was developed. The durability test results show that as the rotational speed increases, the degradation of both the exhaust flow rate and pressure ratio becomes more pronounced. The operational range of these two parameters decreased by 6.2 % and 11.1 %, respectively. To mitigate stochastic noise interference in health indicators (HI), a novel feature optimization method called Moving Center SVR (MC-SVR) was proposed. This method effectively reduces the nuisance noise while preserving the inherent trend of the original HI, thereby enhancing its robustness. An RUL prediction model for the FCAC was established by integrating the dynamic exponential regression (DER) model with the MC-SVR method. Compared with other methods, the RUL prediction model trained with the optimized HI using the MC-SVR method achieved the best prediction performance across four evaluation metrics, namely MAE, MAPE, RMSE, and CRA. All of these provide valuable insights and references for the durability and thermophysics studies of FCACs.</p>

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Durability Study of Fuel Cell Air Compressors Based on Experimental Measurement Data and Lifespan Prediction Models

  • Chaozheng Chang,
  • Jianqin Fu,
  • Peng Zhou

摘要

As a key component of the fuel cell system, the air compressor plays a vital role in ensuring the stability and reliability of fuel cell systems. To conduct a durability study, a 5000-h durability test of the fuel cell air compressor (FCAC) was performed according to the designed test profile, which reflects the real-world operating conditions of fuel cell vehicles. Based on the collected durability test data, the performance degradation characteristics of the compressor over time were analyzed, and a remaining useful life (RUL) prediction model was developed. The durability test results show that as the rotational speed increases, the degradation of both the exhaust flow rate and pressure ratio becomes more pronounced. The operational range of these two parameters decreased by 6.2 % and 11.1 %, respectively. To mitigate stochastic noise interference in health indicators (HI), a novel feature optimization method called Moving Center SVR (MC-SVR) was proposed. This method effectively reduces the nuisance noise while preserving the inherent trend of the original HI, thereby enhancing its robustness. An RUL prediction model for the FCAC was established by integrating the dynamic exponential regression (DER) model with the MC-SVR method. Compared with other methods, the RUL prediction model trained with the optimized HI using the MC-SVR method achieved the best prediction performance across four evaluation metrics, namely MAE, MAPE, RMSE, and CRA. All of these provide valuable insights and references for the durability and thermophysics studies of FCACs.