This book provides an introduction to many aspects of computer control. It covers techniques or control algorithm design and tuning of controllers; computer communications; parallel processing; and software design and implementation. The theoretical material is supported by case studies co ering power systems control, robot manipulators, liquid natural as vaporisers, batch control of chemical processes; and active control of aircraft. The book is suitable for practising engineers, postgraduate students and third year undergraduates specialising in control systems. It assumes some knowledge of control systems theory and computer hardware.
Inspec keywords: discrete systems; real-time systems; three-term control; multivariable control systems; software fault tolerance; control system synthesis; parallel processing; computerised control
Other keywords: multivariable control system design; software design; PID controllers; computer control; software fault tolerance; parallel processing; real-time computer networking; shop floor control systems; discrete controller design; real-time system; batch process control
Subjects: Discrete control systems; Multivariable control systems; Control engineering computing; Control applications
This chapter discusses the theory and practice of discrete time control (i.e. the sort that is implementable in computers) to achieve closed-loop control of continuous time processes. Thus, we are concerned with the discrete time control of continuous time processes within a hybrid feedback loop. A natural initial question suggests itself: what are the advantages of discrete, as opposed to continuous, controllers? The answer is very disappointing. As far as control theory is concerned, there are no advantages (the proof of this is that every realisable discrete time signal is a continuous time signal, but the converse is false). The reason for studying discrete time control is essentially practical: to allow reliable miniature low-cost digital electronic devices to be used as controllers.
An introduction to multivariable control system design has been presented. Due to lack of space, only a brief overview has been given, but the interested reader can consult the references cited for further discussion.
Several methods have previously been developed that automatically tune PID controllers. The PID algorithm itself is usually essentially simple, but commercial implementations of PID controllers contain a multitude of additional features which embody the experience of many years of application. It is therefore prudent to retain those features when introducing self-tuning; basically sound self-tuning methods have failed because these vital details have not been attended to. To take account of these points, the approach taken in this chapter is to use a standard commercial PID controller (a Eurotherm 820) and use a continuous-time self-tuning PID controller to generate its parameters as opposed to using the self-tuning PID controller to generate a control signal. The self-tuning PID controller is implemented on an Atari PC and communicates via an RS-232 port using the Eurotherm communications protocol.
This chapter details some of the techniques employed in the implementation and application of process control instrumentation for industrial plant. Process engineers have become very familiar with regulators using proportional + integral + derivative (PID) control, as the vast majority of continuous processes under closed-loop control use variants of this algorithm. The reason for the widespread use of the PID algorithm is that it forms a reliable core for very robust control regulators and the tuning parameters are relatively easily understood. Digital implementations of PID algorithms are discussed and practical problems with the associated input and output interfaces are highlighted.
The way in which switching and sequences are generated for practical control in industry are introduced and described. Methods have changed with the introduction of PLCs though programming and methodology are still related to the earlier methods used. The extra facilities built into PLCs are also reviewed. Such facilities make them a new tool in their own right with the ability to compute and control online. Large systems require much checking and testing. Software tools and methodologies to facilitate this are an important way of reducing project costs and demonstrating that the system will do what it should (and nothing that it shouldn't).
An instrumentation system or a control system does not have to exist at one specific location, nor does a system have to exist in a specific cabinet. Nowadays, the idea of centralised systems is being replaced by the concept of distributed systems. Control engineers long ago discovered the frailty of centralised control rooms and took the advent of the microprocessor to distribute control actions to the point of application. Instrumentation engineers working in the same epoch seized the opportunity to incorporate intelligence into the point of measurement, and computer engineers developed the technique of distributed computing by means of networking. This chapter investigates the means of communication between instruments and controllers using digital computers. It deals first with the ideas of distributed systems and discusses serial interfaces. Later, the modern techniques of local and wide area networking are applied, but not in detail.
This chapter presents parallel processing for computer control. The use of general purpose microprocessors in direct digital control is very beneficial, so adopted in many wide and varied applications. These early microprocessors were rather primitive; they were 8-bit machines working with clock rates of the order of 1 MHz and possessed little software support usually the programming had to be performed in the assembly language. The situation is somewhat better now. The capabilities of electronic chip manufacturers have evolved to the stage where very large scale integration (VLSI) is possible and 32-bit microprocessors working at 25 MHz are commonly available; such modern microprocessors are powerful computing devices, which are able to perform many millions of instructions every second. In addition, they can also have on-chip floating-point support and are able to be programmed in standard high-level languages.
This chapter reviews some of the issues relating to the development and design of real-time software and briefly outlines two design methodologies. With the increased use of microprocessors in equipment, there is a growing need for software engineers and other engineers to be familiar with techniques for developing real-time software. Implementors are moving away from reliance on monolithic, general-purpose operating systems and are using minimum operating system kernels. The additional operating system features required are then built using a high-level language for a particular application or group of applications. There is also an increasing use of multiprocessor systems with a consequent increased concern with communications, distributed databases and distributed operating systems.
The name 'MASCOT' is an acronym for Modular Approach to Software Construction, Operation and Test and, as acronyms go, this one is accurate and useful without succeeding in telling quite the whole story. It provides a number of significant keywords on which a description can be based. An introductory section attempts to lay the foundation for the details of MASCOT which are then described in later sections.
Methodologies for analysing the reliability of complex systems and techniques for making such systems tolerant of faults, thus increasing the reliability, are well established. In software, the emphasis has been on improved software construction techniques or “software engineering” to reduce latent errors; there has also been work on techniques to introduce redundancy into software systems. The majority of work on fault tolerance has concentrated on “anticipated faults”, i.e. faults which the design can anticipate and hence “design” in tolerance. A much more difficult and insidious problem is that of faults in the design of the system. These are by definition “unanticipated” (and unanticipatable) faults. Design faults can occur both in complex hardware and software, but are more common in software, and much of the effort of software engineering has been directed towards reducing design faults, i.e. unanticipated faults.
This chapter describes how object-oriented design (OOD) was used to produce a high-level design in conjunction with a top-down stepwise refinement methodology. It also describes how OOD was used recursively through several layers of recursion on the design of a large, complex software system.
This chapter presents the modelling, simulation and control of direct-fired liquid-gas vaporizers. The British Gas Isle of Grain Storage Facility was built to meet the requirement for a large volume peak shave storage facility to serve the south east of England. Liquefied natural gas (LNG) was chosen as the most suitable means of providing this storage. The installation consists of four 21,000-tonne LNG storage tanks, two liquefaction plants and a vaporisation/export system. LNG is stored at a temperature of -160°C and a pressure of 100 mbar, and for export into the National Transmission System the gas temperature must be raised to ambient and the pressure to grid conditions, i.e. between 40 and 70 bar.
An overview is given of a number of computer-controlled drug administration schemes in clinical medicine. These range from systems which have been used on many patients in intensive care to experimental schemes which require further clinical evaluation. The control algorithms used vary from simple PI controllers to multi-mode adaptive techniques. Measurement of relevant clinical variables is often a major problem, and the use of extended Kalman filtering is described for the estimation of unmeasurable states. Recent developments in expert control are also described.
Robots offer the opportunity of automating manufacturing processes in a flexible way which has far reaching consequences on the economics of automation. They have made such a great impact because the effort required to reprogramme them, for instance to change the pattern of welds in a spot-welding application, requires only an adjustment to the controlling software. This is relatively simple compared with the mechanical redesign necessary when the functioning of hard automation machinery needs changing. The industrial robot is therefore of great importance in the modern manufacturing scene, and this is reflected in the fact that the application of robots to industrial tasks is growing at the rate of about 50% per year.
Many of the active control technology (ACT) systems discussed in this chapter can be found in modern fighter aircraft. Most have been designed using modern control methods, such as the optimal linear quadratic regulator problem, to determine the appropriate feedback control law. Other methods such as model-following or eigenstructure assignment can be used with equal facility, but the resulting control law will take the same form namely a linear full, state-variable feedback control law. When gust load alleviation (GLA) or bending mode control (BMC) are being considered, where the most significant bending modes are taken into account, the corresponding feedback control will require that the bending displacements and rates are available for measurement. This may be unlikely, and if those methods are to be employed, then recourse to state estimation techniques will be required.
Some of the characteristics of batch control systems are discussed, and illustrated with particular reference to an application from the sugar industry. A comparison of experience with different control equipment is touched on, and related to present and future directions for development.
This chapter reviews the main functions of a process control computer, including measurement and actuation, direct digital control, sequence control, supervisory control and operator communications. The hardware and software aspects of distributed and hierarchical computer control systems and the integrity and economics of computer operations are discussed. The intention is to provide a broad introductory overview of the significant features of computer control systems.
