Data-driven intelligent control system and method embedded in hormone regulation mechanism

Abstract

The invention relates to a data driven intelligent controller with an embedded hormone regulation mechanism and a method thereof, output data of an object controlled by the controller is compared with given data inputted as reference for the controlled object through a feedback sensor group, output error of the controlled object is obtained, and the error is simultaneously inputted into a conventional controller and an error threshold evaluation unit. A data driven computing unit is used for running the data according to system history recorded in a historical data storage unit, the current working state between the conventional optimization state and the collaborative optimization state is selected for determining the controller which needs to be optimized, optimization adjustment is carried out on parameters of the controller, the original control quantity obtained by computation is processed by the error threshold evaluation unit and selectively added into the control quantity increment of a hormone regulation unit, and the output of the controller is finally synthesized and sent to the controlled object for implementing the control. The system and the method can improve sensitivity and accuracy of data driven control, and be conductive to keeping the controlled object in the sustained and stable working state.

Description

technical field [0001] The invention belongs to the technical field of automatic control, and in particular relates to a data-driven intelligent control system embedded in a hormone regulation mechanism and a method thereof. Background technique [0002] Many industrial production processes (such as materials, chemicals, pharmaceuticals, etc.) are a collection of a series of complex process steps. The randomness contained in these process steps and their dynamic changes during operation make it impossible to use the standard "input-output" correspondence to establish a complete mathematical model to meet the control needs; even if a relatively complete model can be obtained, the other Internal structures and parameters may also change or fluctuate as the production process runs. On the other hand, most of the existing advanced control methods need to be applied to precise object models in order to achieve good control effects. The uncertainty and random dynamic change char...

AI Technical Summary

Problems solved by technology

[0004] In practice, one of the main problems with data-driven controllers and their methods is the poor dynamics of closed-loop systems using them

Although the data-driven control strategy solves the problem of not being able to obtain a high-quality object model to a large extent, it is precisely because there is no accurate object model for reference that there are often phenomena such as large fluctuations in controlled variables and increased waste products. Some data The consideration of dynamic performance indicators such as the response time and overshoot of the controlled object by the drive controller is also insufficient.

In many cases, the control effect based on the data-driven controller can achieve a satisfactory control effect (that is, "controllable"), but there is still a lot of room for improvement in the dynamic characteristics of the closed-loop system formed by it.

Images