摘要
选用径向基函数神经网络建立pH值和浓度的模型,克服了在白炭黑生产过程中采用负梯度下降法调节反应釜溶液pH值和浓度时存在收敛速度慢和局部极小等缺点。实际应用表明:其速度和精度完全达到了工艺上对pH值和浓度的控制要求。
The RBF neural network was chosen to build pH value and concentration models and to adjust reactor solution so that slow convergence speed and relative minimum and other weakness of negative gradient descent method can be controlled. Actual application shows that both speed and precision can meet the requirement of pH value and concentration control.
出处
《化工自动化及仪表》
CAS
2012年第1期40-43,共4页
Control and Instruments in Chemical Industry