- Author:
- Valentine Charles Matula
Since the 1950's the advantages of using a digital computer in an industrial plant has been recognized. More recently, the declining cost and size of digital computers has made feasible implementing distributed computer control systems. One of the problems in the design of such systems is the assignment of control programs, or "tasks" to processors, and the assignment schedule for resources to tasks on each processor.
By separating the computer control system issues from the process control issues through the common denominator of task execution time and frequency, the modeling and design analysis procedures proceed along the traditional disciplinary lines of computer engineering and control engineering.
This work develops a model of the relationship between the control tasks and the controlled process that permits the estimation of the overall process performance given the time duration and frequency of execution (the effective task time) for each control task. A second model is developed of the hardware and software that comprise a distributed computer control system. This second model is used to estimate the range of execution time duration and execution frequency values that are possible for each control task, given a particular computer control system configuration to be analyzed.
A procedure is given that uses the two models to provide an estimate of overall process performance given a hardware and software control system configuration. A set of programs to implement the procedure is developed and described. A sample system consisting of a pair of 2-tank fluid processes is examined using the models and analysis procedures.
The methods developed in this research differ from methods described in the literature that minimizes intertask communication, overall task execution time, or other computer control system variables that are internal to the computer system. The drawback of these methods is that for process control systems, optimization with respect to internal computer control system variables does not lead to optimal performance of the controlled process.
Recommendations
A Simulation Method for Computer Control Systems
A computer control system simulation method utilizing two computers is presented. One is a control computer operating under almost actual conditions. The other is a multiprogrammed computer that simulates the plant. The time taken by the control ...
Control of modular and distributed discrete-event systems
FMCO'05: Proceedings of the 4th international conference on Formal Methods for Components and ObjectsControl of modular and distributed discrete-event systems appears as an approach to handle computational complexity of synthesizing supervisory controllers for large scale systems. For both modular and distributed discrete-event systems sufficient and ...
Computer control systems in organizations
The purpose of this paper is to present an overview of control systems in organizations, with the emphasis on computer control systems. Specifically, the paper will present a model of organizational control systems, discuss the responsibility for ...