用RBF神经网络建立航空煤油密度软测量模型

Development of Soft Sensor Model for Density of Aviation Kerosene Using RBF Neural Networks

在航空煤油的生产过程中，由于航空煤油密度的检测有很大的滞后而难以实现直接质量控制，降低了航空煤油产品的质量和产量，影响了生产效益。针对实际的生产过程，利用径向基函数（ＲＢＦ）神经网络开发了一个航空煤油密度的软测量模型，并用遗忘因子递推算法（ＲＦＦ）对其做了实时自适应校正，实现了航空煤油密度的在线估计。

In the process of aviation kerosene production because of the delay of measuring the density of aviation kerosene it is difficult to control quality online.It causes the quality and quantity to drop and benefit to be influenced.A soft sensor based on radial function (RBF) neural networks is developed and it adopted recursive forgetting factor (RFF) algorithm to make realtime adaptive correction.The online estimation of quality of aviation kerosene is realized,meantime the efficient soft sensor means is provided for quality control.