废轮胎胶粒气化脱硫BP神经网络预测

Scrap tire colloidal particle gasification and desulphurization are forecasted by BP neural network

对废轮胎胶粒使用流化床试验装置进行气化和脱硫,得到样本数据。使用BP神经网络建立气化产气热值和含硫气体排放特性预测模型,较好地实现了对试验数据的再现。在训练过程中,误差快速的单调下降,表明该网络的学习能力很强。模型很好的预测了试验工况,预测值和试验值的误差的数量级均在10^(-6)。

The fluid bed test equipment is used to gasify the scrap tire colloidal particle and desulphurise the produce gas,and then obtains the sampled data. The BP neural network was used to establish model to forecast heat value of produce gas and emissions characteristic of the sour gas,it made the tentative data reappearance very good. In the training process,the erroneous monotonous drop fast,indicated that this network's learning capability is very strong. Model forecast experiment operating modes very well,both of error's magnitude of predicted value and experimental value are in 10^(-6).