基于软测量的化工精馏过程推断控制策略

针对化工精馏过程产品成分无法在线检测及其用温度间接控制产品成分的常规控制策略存在着控制精度低的问题,提出基于软测量的精馏过程成分非线性串级推断控制策略。该控制策略首先提出核岭回归的实时软测量方法,即利用满足Mercer条件的核函数改进线性岭回归算法,实现精馏过程产品成分的在线检测;然后在此基础上,提出一种新的非线性串级推断控制策略,即副环采用常规的温度间接控制,主环采用基于核岭回归软测量的推断控制策略。通过Mejedell等建立的精馏塔动态模型分别对单端和双端成分非线性串级推断控制策略性能进行分析,仿真结果表明,与传统控制方案比较,新控制策略的控制质量有了较大提高,控制结构简单,易于实施。

A Nonlinear Cascade Inferential Control Strategy Based on Soft Sensing of the Composition of Chemical Industry Distillation Process

Few hardware sensors are available on-line to the composition of distillation products.Moreover,as taking the general control strategy such as indirectly temperature control to control the distillation compositions,the control performance is always bad.A nonlinear cascade inferential control strategy based on soft sensor is proposed.Firstly,in the control strategy,a kernel ridge regression to implement on-line soft sensing of distillation compositions is proposed.This means that kernel function which conforms to Mercer condition is used to improve linear ridge regression algorithm in order to realize on-line measurement value of the distillation compositions.Then,a novel nonlinear cascade inferential control strategy is proposed on the basis of the first step.In other words,the general indirect temperature control strategy is introduced in auxiliary loop and the inferential control strategy based on kernel ridge regression soft sensor is adopted in main loop.The ability of single-end and double-end nonlinear cascade inferential control strategy based on the dynamic model of the distillation tower established by Mejedell is analyzed.Compared with traditional control method,the simulation results show that the novel control strategy is efficiency.Furthermore,the new control strategy is easy to apply to practice because of its simple composition.