Integrated MMPSO and RBFNN for optimal control of cracking depth in ethylene cracking furnace

The prediction models of cracking productions online are realized by using radical basis functions neural network (RBFNN). Then the building model is optimized by using the proposed multi-modes particle swarms algorithm (MMPSO) to catch the optimal operational conditions. At the same time, the intelligent optimal control method for cracking depth is studied integrated by MMPSO-RBFNN. The optimal function is maximized the sum of ethylene & propylene yields, and cracking depth, which is satisfying to the optimal function, is put into the depth´s controller, which is linked into the advanced process control system of coil outlet temperature (COT), so the depth´s control is realized optimally. The applications are showed that the yields of ethylene and propylene are increased, and the depth´s control is more stable than before. The proposed optimal control method has good adaptability, stability and reliability.