摘要
采用人工神经网络与传统理论计算模型相结合的方式,建立了智能型的制冷压缩机热力性能计算模型,利用人工神经网络的自学和泛化能力提高了制冷压缩机容积效率和电效率的计算模型精度。仿真结果表明,智能型制冷压缩机模型很好地改善了传统理论计算模型精度,而且适应能力更强;并且人工神经网络和传统理论模型的融合程度,对于制冷装置仿真系统的优化和仿真精度的提高起到了至关重要的作用。
Aintelligent computing model for the refrigeration compressor which combine artificial neural network with traditional theoretical model is proposed. The function of self-learning and generalization of ANN is used to improve precision of the traditional model. The Multi-Layer Perception （MLP） network is adopted , and the BP algorithm is used to train the MLP efficiently. It shows that the new model based on neural model bring more precise results than the traditional model. Furthermore, the new computing model is of better flexibility in a large scale.
出处
《流体机械》
CSCD
北大核心
2008年第7期41-43,共3页
Fluid Machinery
关键词
制冷压缩机
热力学性能
人工神经网络
refrigeration compressor
thermodynamic performance
neural network
