摘要
化工过程中的温度对象往往具有惯性大、滞后时间长的特征,严重影响常规PID控制算法的性能。神经网络算法能够拟合操作人员的操作过程,往往能够取得常规控制难以达到的效果。本文针对反应过程,根据现场操作经验离线训练1个3层神经网络模型,在中控ECS-700集散控制系统(DCS)的VF软件平台上实现,并在维生素生产的一个反应中投运,结果验证了神经网络算法在温度控制上的实际有效性。
The temperature in the process of chemical objects often has the characteristicsof large inertia and lag time, which Seriously affect the performance of the conventiosnal PID control algorithm. Neural network algorithm simulates the operate process of operators,and always has the result which conventional control methods are unable to experiences. This paper, according to the site operation experience and reaction process, build a three-layer neural network model by off line training. This model is then implemented on the VF software platform, which belongs to SUPCON ECS- 700 Distributed Control System （DCS）. The result of the application of neural network algorithm in temperature control of chemical reactor that confirm the effectiveness of the model.
出处
《中国仪器仪表》
2016年第8期60-62,共3页
China Instrumentation
