摘要
直接从汽轮机转子的一维热传导方程出发,研究了汽轮机转子径向上的危险截面的一维温度场与Hopfield神经网络的相似机理,提出了应用连续Hopfield神经网络实现转子一维温度场的实时监测方法。通过仿真可以看出,此种方法测量速度快,测量精度高,转子热时间常数与Hopfield神经网络的电时间常数近似相等,Hopfield神经网络模拟测量结果与转子的温度变化规律较一致。
According to the theory of heat exchanging, the co mparability between one dimensional temperature field of danger section of turbi ne rotor and Hopfield Neural Network is analyzed. Real-time measurement method of one dimensional temperature field with continual Hopfield Neural Network is given. The method has high measurement rate and is precise. The heat time consta nt is approximately equal to the time constant of Hopfield Neural Network, and t he change of temperature of danger section rotor is approximately consistent wit h the calculating result of Hopfield neural network.
出处
《热力透平》
2003年第2期78-81,94,共5页
Thermal Turbine
