James D. Schoeffler

A real-time language for industrial process control

Industrial computer control systems require the economical production of efficient software including executive systems, maintenance programs, and both special and general purpose application programs for direct digital control. Moreover, the hardware configuration varies considerably from the single dedicated control computer to a general purpose multicomputer system. RTL, a real-time language developed cooperatively with industrial suppliers and users specifically for industrial control, is described with emphasis on those features peculiar to applications such as dedicated direct digital control, combined direct and supervisory control, operator interfaces, and interaction with plant management computer systems. The use of RTL for the production of special purpose executive systems and general purpose application programs for direct control, startup, etc., is emphasized. Details of the language discussed with examples include its file structure for communication of data bases between independent programs, and a variety of data types including character codes, strings, labels, lists, peripheral variables, and data structures. Peripheral variables are variables in the language associated with hard-ware features of the central processor and its input-output devices such as registers, interrupts, error indicators, and addresses all of which may be referenced in the language. Regular and peripheral variable data structures--combinations of different types of variables--are included and ease considerably the burden of real-time programming. The organization and performance of the existing compiler for RTL is explained.

A Finite State Model for the Control of Adrenal Cortical Steroid Secretion

The steroid hormone Cortisol is the principal secretory product of the adrenal cortex in man and the carnivores. Its rate of secretion is primarily under the control of adrenocorticotropic hormone, or ACTH, secreted by the pituitary gland. The release of ACTH by the pituitary is in turn controlled by the hypothalamus by means of a neurohormone, corticotropin-releasing factor, or CRF. From the late 1940’s until early 1960’s there was a controversy about the mode of control of ACTH release and Cortisol secretion. One school based its view on the possibility of suppression of ACTH release and induction of adrenal atrophy by exogenous Cortisol, and suggested that there is closed-loop automatic control of the peripheral concentration of Cortisol in the blood, with stimulation of ACTH release when the Cortisol level of the blood fell, and suppression of that release when the Cortisol level was elevated. Accordingly, a stimulus would increase the load and lower Cortisol concentration. This would lead to increased CRF release from the hypothalamus and thus increased ACTH release from the pituitary.

The development of process control software

The use of on-line, real-time computers for control of industrial processes has been increasing rapidly during the past ten years. That the cost of the software necessary to implement such systems exceeds even the hardware costs became clear in the initial installations. As a consequence, much effort was devoted to the development of efficient, economical software and software approaches for use in industrial process control applications. During the past five years, there has emerged from these efforts the realization that process control software is different from software for large scale batch processing, time-sharing, message switching, or any other computer applications. It contains its own requirements and problems and leads to a distinct set of solutions. Moreover, this difference is due to more than the size of the computers involved in industrial process control. The objective of this paper is not to catalog the significant features of various software systems in existence today for this has been done very well in a number of recent survey papers. Rather, the objective is to describe the basic structure of current industrial process control software, emphasizing the unique structure of that software, how it evolved, and its current points of controversy and problems.

Multilevel Control Structures for Three Discrete Manufacturing Processes

Abstract Discrete manufacturing systems represent a large fraction of industrial processes and involve control problems whose dimension precludes straightforward solution. In general, such systems involve transformationfrom raw material to finished product and include a sequence of discrete operations (machining or parts, storage in and retrieval from inventory, etc.). Concornmitant with each of these operations are control problems including monitoring, reporting, scheduling, machine loading, start up, shut down, batch control, emergency actions, security enforcement, etc. The great differences among the disturbances and contingencies which can affect the operation of such processes suggests that the controls associated with each stage of the process and with each type or class of disturbance be handled as separate tasks. This decomposition of the control system, in turn, requires coordination between the subunits in order that the overall process objectives are achieved. Hierarchical control systems have proved to be effective models for the design of complex control systems for continuous processes. The objective of this paper is to demonstrate how a hierarchical structure can also lead to the effective control of discrete processes. Three examples are presented, each illustrating a different aspect of the discrete control problem formulation. It is shown, in each case, how the control tasks organize naturally into a multilayer hierarchical structure. The examples focus on problems of scheduling, control of discrete operations, and security control, the latter being control in the face of catastrophic disturbances. Generalizations are drawn from the experiences of these examples. Various features of the proposed structure are pointed up with specific reference made to problems of hardware and software implementation.

Realtime Operating Systems For Distributed Process Control Systems

Abstract Conventional process control systems today are organized around a single computer with a realtime operating system as the key element. The function of the realtime operation system is to control resources in the system shared by application tasks including input/output devices, computer memory, and the CPU itself. Two developments forthcoming will affect realtime operating systems for process control. They are first, the trend toward line sharing and remote intelligent terminals; and second, the trend toward the use of multiple computers and microprocessors in semi-dedicated control functions but still integrated into the overall computer control system. Both of these trends indicate that the major impact on the realtime operating system will be in the distribution of the data base and intertask communication since both devices and data and tasks will be resident in possibly different computers. The objective of this paper is to examine techniques for communication between computers which can realize effective interprogram communication even when the tasks and data base are distributed among several process control computers or special purpose controllers. Some experimental results are presented.

Organization of Software for Multicomputer Process Control Systems

The trend toward multiple computers for process control applications is accompanied by the introduction of line sharing input/output systems and microprocessors for general use. As a result, many different structures for multiple computer process control applications become practical. Software costs are still dominant however, and it is important that any multi-computer structure be supported by software which is compatible with process control objectives and economical to implement. Of particular importance is the need for an integrated data base and graceful degradation in case of failure. The organization of process control software for single computer systems is used as the starting point to evaluate the desirability of various software organizations for several multi-computer ring-connected systems. Compatibility with line-sharing I/0 systems is discussed.

Identification of Boolean Mathematical Models

A companion paper in this volume has described the successful application of a mathematical model involving quantized variables to an important, real system, the adrenal.1 The objective of this paper is to consider this class of mathematical models in more detail in order to determine how such modeling problems may be formulated, how the parameters of the model may be identified and in particular, to look at the feasibility of actually carrying out the computations involved for systems of reasonable complexity.

Computer optimization methods applied to medical diagnosis

Abstract In this paper we show how the methods of mathematical programming can be combined with models of biological systems to provide a different approach to the medical diagnosis problem through the use of computers. This is accomplished by tuning a mathematical model to fit a set of patient data by means of a modified version of the Davidon-Fletcher-Powell algorithm. The modification consists of imposing an orthogonality constraint on the direction vector and the gradient at an interpolated minimum. This modification then increases the speed of convergence by making more efficient use of each function evaluation. As an example of how this method is used diagnostically, the system of conducting airways was used. Here, by estimating the physiologic parameters associated with the nonlinear model of the conducting airways, a diagnostic characterization can be made.

Data management software for minicomputer production monitoring and control systems

A two-level software system for minicomputers suitable for use in real-time production monitoring and control applications has been designed and implemented. The first level consists of a general-purpose file system implemented in the form of a voluntary virtual memory which can be added to most minicomputer software systems. The second-level software is a set of data management primitives which permit the definition, creation, maintenance, modification, and referencing of large, interlinked, complex data bases which arise in discrete parts manufacturing applications. The software system is described. The application to a discrete parts manufacturing control system is described along with experimental measures of response time and memory requirements.

Security Control in Glass Manufacturing

ABSTRACT Security control is the application of control systems technology to the prevention of system failures and to the minimization of losses during and after a system emergency. The control of security for an industrial system demands an understanding of the causes and effects of disruptive structural disturbances, called contingencies. To reduce the effect of these disturbances one can implement contingency control, contingency prevention and contingency planning. Of particular interest are the actions required prior to a system emergency, because here it is possible to exercise contingency control, that is, to alter the security of the system to make it less susceptible to certain contingencies. Contingency control has been formulated as a constrained optimization problem in which the imposition (relaxation) of security constraints increases (decreases) the security of the system. This paper discusses some of the contingencies and their control that are applicable to the glass manufacturing system, including the problem of viscosity and flow control in the downcomer preceding the glass feeding machine, in detail.