基于粒子群RBF网络的C_3含量软测量模型

Soft Sensor of C3 Concentration Based on Particle Swarm Optimization RBF Neural Network

鉴于C3含量在线估计对催化裂化生产过程操作意义重大,本文通过建立粒子群RBF神经网络模型、利用软测量技术来实现C3含量的在线估计,并对其仿真结果进行分析和比较。实验结果表明,基于粒子群RBF神经网络的C3含量软测量模型具有较高的精度、较好的性能和良好的应用前景。

Aiming at the difficulty in directly-measuring the C3 Concentration of fluid catalytic cracking unit （FCCU）, soft-sensor is applied to solve this question. A practical soft-sensing model based on particle swarm optimization （PSO） RBF neural network are established, and the simulation results are analyzed and compared. After the soft-sensor was applied in actual chemical plant, good application effect was got and process control was improved.