【摘要】 针对燃料电池/燃气轮机(SOFC/GT)混合动力系统在变工况运行过程中极易出现的电堆超温和重整器碳沉积等故障，提出一种基于反向传播(Back Propagation,BP)神经网络的故障诊断与性能预测方法。通过搭建混合动力系统动态模型，进行了动态特性分析及验证，设计并优化了BP神经网络结构，实现了对超温与碳沉积故障的准确诊断，以及对电堆温度和系统输出功率的预测。结果表明:所设计的混合动力系统输出功率为388.4 k W，发电效率为61.8%，满足系统设计要求;优化后的BP神经网络模型对超温故障的诊断准确率为97.5%，对碳沉积故障的诊断准确率为92.5%;在空气流量分别阶跃降低12.5%和17.5%的工况下，BP神经网络模型对电堆温度及电堆输出功率的动态预测准确率分别达到97.3%和98.8%。研究结果为未来发展长寿命、绿色高效发电技术提供技术支撑。更多还原

【Abstract】 To address faults such as stack overheating and carbon deposition in reformers,which frequently occur in fuel cell/gas turbine( SOFC/GT) hybrid power systems under variable operating conditions,a fault diagnosis and performance prediction method was proposed based on a back propagation( BP) neural network. A dynamic model of the hybrid system was developed,and dynamic characteristics were analyzed and validated. The structure of the BP neural network was designed and optimized to achieve accurate diagnosis of overheating and carbon deposition faults,as well as prediction of stack temperature and system output power. The results indicate that the designed hybrid system achieves an output power of 388. 4 k W with a power generation efficiency of 61. 8%,meeting the system design requirements. The optimized BP neural network model demonstrates a fault diagnosis accuracy of 97. 5% for overheating and 92. 5% for carbon deposition. Under conditions where the air flow rate is step-reduced by12. 5% and 17. 5%,the dynamic prediction accuracy for stack temperature and output power reaches97. 3% and 98. 8%,respectively. These findings can provide technical support for the future development of green and efficient power generation technologies with long service life.更多还原