Automating the kiln process - updates on Fuzzy logic and shell scanners.

Automating the kiln process -- updates on Fuzzy logic and shell scanners

Introduction

Advanced techniques have been successfully used to help control cement kilns for over a decade. As more operating experience was gained, several improvements followed. This paper outlines the state of the art in high level kiln control, focusing on recent advances in Fuzzy logic and high speed shell temperature scanners.

Fuzzy II(*) -- The state of the art

As the number of installations of high level kiln control (Fuzzy and others) increased, the results fell into one of three categories:

* Highly successful -- exceeded expectations * Successful -- met expectations * Marginal -- did not meet expectations. The marginal cases, while very much in the minority, are of most concern as they offer clear examples where the control technique could be improved. In examining these cases, several different reasons for the marginal performance emerged:

* Controller not suited to the task -- |simple solutions' * Controller not implemented properly * Plant conditions changed after controller implemented * Controller not supported by plant people * Plant presented difficult or unusual problems.

The first two reasons were not frequently found with Fuzzy systems. The final three indicated improvements should take two different directions -- make the controller simpler to increase support from plant people; make the controller more complex to allow it to handle more difficult situations.

As development work continued to improve the controller and meet these conflicting requirements, a new generation of controller was developed: Fuzzy II.

Objectives and priority management

A fresh look at the control problem indicated a successful plant operation is driven by a series of objectives, sometimes conflicting, and prioritised at different levels of plant staff. Successful management of the plant equals successful management of these priorities.

Basic concepts

Most high-level control systems, including the first Fuzzy systems, are based on implementing a series of control rules, either by direct programming or a rule-block interpreter. The collection of rules and program instructions for a particular application has been termed the control strategy. Fuzzy II also includes both rule-block definition and direct programming facilities, but more importantly it includes a priority management design tool to enable advanced control strategies to be developed and easily maintained.

Objectives

The first step in developing a Fuzzy II control strategy is to define the objectives. Objectives can be simple -- those that are easily defined and measured, or complex -- those based on calculation results or complex relationships. The following table shows some common objectives included in a kiln control strategy:

 
                         Objectives 
Simple                       Complex 
No CO                        Stable operation 
Maintain oxygen setpoint     Maintain …