制冷系统仿真中定量参数的神经网络辨识

尝试用现代人工智能技术来改进现有的制冷系统仿真方法．首先，提炼出与制冷系统仿真结果的量化密切相关的定量参数，然后在已有的定性数值仿真模型的基础上，根据实验数据，采用人工神经网络（ＡＮＮ）方法对仿真模型中的定量参数进行辨识，识别出最佳的定量参数．这不仅有利于提高仿真精度，改善计算稳定性，而且降低了对仿真软件用户的技术要求，有利于仿真技术的实用化．对房间空调器稳态特性仿真的初步结果表明该方法效果良好．

Artificial Neural Networks Identification of Quantitative Coefficients in Refrigeration System Simulation

The modern artificial intelligent (AI) technology was utilized in simulation of the refrigeration system.At first,the quantitative coefficients which are sensitive to the quantitative simulation results were persented.Then,based on the qualitative numerical simulation models,artificial neural network(ANN) was used to identify the quantitative coefficients in the simulation models accorrding to the experimental data.Finally,the simulation software with the identified quantitative coefficients was used to predict the behaviors of the actual refrigeration system.The new method can improve the simulation precision and stability of computation,and lower the professional demands of using the software.The recommended method is more promising than the current simulation methods in practical use.An example simulating the steady state performances of a room air conditioner was given which shows the satisfactory effects.