Showing papers on "Smith predictor published in 1988"

Control strategy for single-input single-output non-linear systems with time delays

Abstract: A new controller design technique is developed for a general class of non-linear systems with delays. The proposed suboptimal controller consists of a generalized proportional-integral (PI) control structure plus a new time-delay compensator that is shown to be superior to the conventional Smith Predictor. Simulation results demonstrate the advantages of the new design method in comparison with conventional PID control and gain scheduling.

Steam temperature control using modified smith predictor

Abstract: A system for controlling steam temperature in a boiler using a time delay feedback controller known as a Smith Predictor to provide control tuning of true boiler parameters which change with load.

Adaptive regulation of arterial gas pressures by using robust adaptive control scheme

Abstract: A robust adaptive control scheme is investigated and implemented to regulate arterial oxygen pressure and carbon dioxide pressure independently at their desired levels by automatically adjusting two control inputs, the inspired oxygen concentration and the respiratory frequency, in accordance with continuously monitored transcutaneous oxygen and carbon dioxide pressures. A least-squares scheme using multiple models with different dead time and adjustable parameters can effectively determine the linearized first-order input-output model, including uncertain parameters and dead time. The proposed adaptive model-matching algorithm includes a supervisory controller and a robust controller with a Smith predictor to minimize the sensitivity to modeling errors, variations, and disturbances in the controlled subject. An animal experiment has shown that the proposed control algorithm can be easily implemented in an ordinary artificial ventilator. >

Robust Deadtime Compensation for Nonlinear Processes

Abstract: Many industrially important processes are nonlinear and also have a process deadtime. Powerful deadtime compensation methods, such as the Smith predictor, are available for linear systems. Additionaly, recent work has shown that state feedback, such as in the Globally Linearizing Control structure, can provide linear input/output behavior for delay-free nonlinear systems. A combination of these two approaches is proposed. A state feedback is found for general SISO systems which makes the nonlinear process with deadtime behave like a linear process with deadtime. An external Smith predictor is then used to compensate for the deadtime of this linear system. Computer simulations for an example system demonstrate the high controller performance that can be obtained using the proposed method.

A discussion of the performance of the Smith delay compensator for systems with an appreciable time delay

Abstract: A discussion is presented of the performance of the Smith delay compensator (SDC) of Smith predictor during the closed-loop operation of continuous linear systems with an appreciable time delay. The basic principle of the SDC is to design the closed-loop system in such a way that the system response is a delayed version of the delay-free system response, where the delay is the plant transport lag. Difficulties inherent in the SDC are briefly mentioned with suggestions to compensate for them. An example of the SDC scheme is illustrated in the closed-loop sodium nitroprusside regulation of postoperative blood pressure. >

Pole-placement design for the time delay systems using a generalised, discrete-time Smith predictor

Abstract: Considers the pole-placement design of single input, single output (SISO) linear, discrete systems with time-delay. Two approaches, are discussed. The first one is based on a proportional-integral-plus (PIP) design for a delay-free system coupled with Smith's principle to handle time delays. The second approach arises naturally as an inherent part of the PIP method. This control scheme is related to the extended Smith predictor (ESP) method and it is shown that, under certain conditions, it will produce exact equivalence. Since the method is more general than ESP, the authors feel that the resulting control scheme might be conveniently termed a 'generalised Smith predictor' (GSP) control system. The paper also shows how to construct a suitable non-minimal state space form for the GSP, such that the state vector consists of only the directly measured input and output variables, thus eliminating the usual requirement for state reconstructor (observer) design. In this manner, closed loop pole assignment follows naturally and easily from the normal state variable feedback design method.

Boiler steam temperature controller

Abstract: A system for controlling steam temperature in a boiler uses a time delay feedback controller known as a Smith Predictor to provide control tuning of true boiler parameters which change with load. More specifically, a feedforward predictor (38) presets an expected secondary superheater inlet temperature with a boiler load, the expected temperature is corrected for firing rate deviation, air flow deviation and reheat temperature control by respective modifiers (42, 44 and 46), a final correction is effected by a feedback controller (50), and a cascade controller (48) responds to the inlet temperature to provide rapid process loop response to predictable intermediate process control points.

Controller implementation using novel processors - Case study III

Abstract: In this chapter the application of a number of quite different processor architectures to the design of digital controllers has been discussed. The Intel 2920 can be used to implement PID, Smith Predictor and Dahlin type algorithms. The FAD based controller requires rather more external logic than a 2920-based design but it is capable of dealing with plants and controllers of essentially unlimited complexity. The external logic is needed both to provide data conversion facilities and to overcome the problems of slowing the circuit down to sampling speeds appropriate for control systems.

Robust closed loop log modulus design of siso smith predictors for a special class of processes

Abstract: Linear open loop stable or prestabilized SISO processes with uncertain or changing time delays are controlled via a Smith Predictor. The processes under consideration have no other nonminimum phase characteristics than the time delay. Further, the process models have an inherent integration constant of zero or one. A controller design procedure is presented which guarantees a robust closed loop system, where the major tuning parameter of the controller is a nonlinear function of the maximum closed loop Log Modulus and the model-plant mismatch.

An Application Of Dynamic Compensation Method To The Smith Predictor

Abstract: In this work, the dynamic compensation method is used in conjuntion with the Smith predictor to control the multivariable proccsses with different multi delays in both input and state variables. T h e mismatch is considered when the difference between the model a n d the Drocess is small.