Non-linear model predication control method of tank reactor based on on-line support vector machine

Abstract

The invention discloses a non-linear model predication control method of a tank reactor based on an on-line support vector machine, which belongs to the field of the intelligent control of industrial processes; and the control method mainly comprises system modeling and closed-loop control loop design based on the on-line support vector machine. The on-line support vector machine identifies a non-parameter prediction model of the tank reactor through historical data learning, then predicts the future output state of the model with the past input and output information and the future input and output information and realizes the on-line self-correction of the model through the on-line learning capability of the on-line support vector machine. The output of the model is fed back and then is compared with a reference input track, rolling optimization is carried out through a quadratic performance index, a multi-step prediction analytic control law is worked out by solving in a gradual decline principle, the law acts on the reactor, so that temperature control of the reactor is approximate to a set value as much as possible, and a whole control cycle is completed.

Description

technical field [0001] The invention relates to the field of industrial automatic control, in particular to a nonlinear model predictive control method based on an online support vector machine for a trough reactor. Background technique [0002] The trough reactor (CSTR) is a typical chemical process. Due to its inherent nonlinear characteristics, it is often used as a typical severe nonlinear object to test various control methods. The schematic diagram of the principle of CSTR is as follows: figure 1 . [0003] A single-stage irreversible exothermic reaction A→B (A represents the chemical species entering the reactor, B represents the product after the reaction) is carried out in the reactor, through the heat transfer fluid flowing through the cooling jacket (C represents the heat transfer fluid In, D represents the outflow of the heat transfer fluid) to control the characteristics of the entire chemical reaction. The whole process can be described by a set of nonlinear...

AI Technical Summary

Problems solved by technology

Fuzzy modeling requires a large number of fuzzy rules, relies too much on prior knowledge, has human subjectivity, cannot accurately reflect the dynamic characteristics of nonlinear systems, and cannot obtain analytical solutions in predictive control

Although neural network modeling can wirelessly approximate nonlinear objects, the geometric topology is difficult to determine, the learning speed is slow, it is easy to fall into local minimum and over-learning phenomenon, and the generalization ability is poor

Support vector machine modeling only needs small samples, and has many advantages such as global optimality and strong generalization ability, so it is widely used in model predictive control, but this method also has some disadvantages: (1) all current prediction methods The models are established through offline training, and the model cannot be corrected online, but due to the time-varying nature of the nonlinear process, it will cause model mismatch, and the model established offline cannot adapt to this change; (2) the solution of nonlinear rolling optimization is very Difficult to find, generally can only be solved by numerical optimization

Images