W.K. Ho

Performance and gain and phase margins of well-known PID tuning formulas

The performance and robustness of well-known PID formulas for process with deadtime to time constant ratio between 0.1 and 1 are discussed in this paper. The Ziegler-Nichols, Cohen-Coon, and tuning formulas that optimize for load disturbance response (integral absolute error, integral squared error, and integral time-weighted absolute error) give gain margins of about 1.5. The phase margins increase from about 30 to 60/spl deg/ as the process deadtime to time constant ratio increases from 0.1 to 1. Tuning formulas that optimize setpoint response give gain margins of about two and phase margins of about 65/spl deg/. These formulas mostly make use of the proportional-integral derivative (PID) controller zeros to cancel the process poles. Approximate analytical formulas to compute gain and phase margins of PID control systems are also derived in this paper to facilitate online computation which would be particularly useful for implementing adaptive control.

Self-tuning PID control of a plant with under-damped response with specifications on gain and phase margins

The effectiveness of the proportional-integral-derivative (PID) controller for plants with under-damped step response is not well understood. In practice, one may have to tune the PID controller manually through trial and error for such a plant. In this paper, a gain and phase margin tuning formula is derived for PID control of a second-order plus dead time plant with under-damped step response. A self-tuning controller incorporating this PID tuning formula has been implemented and tested on a pilot plant which consists of a hinged rectangular plate whose angular deflection from the vertical is controlled by an air stream generated by a variable speed fan. The results are given in the paper.

Self-tuning IMC-PID control with interval gain and phase margins assignment

The idea of pole-region assignment is extended to interval gain and phase margin assignment. The internal model control proportional-integral-derivative (IMC-PID) design is examined from the frequency domain point of view. Equations for typical frequency domain specifications such as gain margin, phase margin and bandwidth are derived for the IMC-PID design. The gain and phase margins are monitored in real time and a self-tuning controller with interval gain and phase margin assignment is proposed. An implementation example in the laboratory is also given.

Tuning of Multiloop PID Controllers Based on Gain and Phase Margins Specifications

Abstract An analytical method of tuning multiloop PID controllers based on gain and phase margins specifications is presented in this paper. The proposed design method can tune the multiloop controllers on-line and in real-time to meet specified system robustness and performance. The design method can be easily combined with existing process identification techniques to implement self-tuning multiloop controllers.

Structured fault-detection and diagnosis using finite-state automaton

Fault detection and diagnosis in large, complex dynamic systems is important for safe operation. Faults may occur in the process, the sensors, the actuators and the measuring instruments independently or simultaneously. Some faults can be detected by direct measurement, but many can only be inferred through indirect means. We propose to address the problem of fault-detection and diagnosis using finite-state automaton models (FAM). These models partition the state-space into finite regions and contain information on system trajectory within these regions. A framework for modelling faults using FAM is discussed. A fault-detection engine is then developed to detect and isolate faults.

Knowledge-based multivariable PID control

Abstract Knowledge-based control has become an important approach towards the realization of intelligent control systems. To date, multivariable control system design is restricted to off-line design aided possibly by a knowledge-based system. A first-generation on-line knowledge-based multivariable PID control system is implemented in this paper. Real-time considerations are taken into account. The integration of the knowledge-based system with conventional process control system software and hardware is achieved through an interfaced approach. Heuristics developed in recent research on intelligent PID control is implemented to attain some of the visionary goals of knowledge-based control. An implementation on a laboratory scale heat exchanger is also given.

Tuning of PID Controllers based on Gain and Phase Margin Specifications

Abstract An analytical method is proposed to tune the PIIPID controller to pass through two design points on the Nyquist curve as specified by the gain and phase margins. They are the appropriate use of pole-zero cancellation and algebraic simplification of certain nonlinear functions. The required process parameters can be simply computed on the basis of the ultimate gain, ultimate period and static gain which may be supplied by the well-known relay-feedback or other auto-tuners. The excellent perfonnance of the tuning formulae has been substantiated by extensive simulation

Intelligent Alarm Management Through Suppressing Nuisance Alarms and Providing Operator Advice

Abstract Nuisance alarms often clutter and obscure an operators view of critical and important information. In this paper, an intelligent alarm management system for a refinery plant is presented. Its function includes nuisance alarm suppression, intelligent advisory information to guide operator response, and result recovery mechanism in case of system failure. This system is implemented on Distributed Control System (DCS) and validated in real online tests. The results show that nuisance alarms are effectively suppressed and the DCS alarm system becomes more helpful.

A Real-Time Realization Of Fault-Detection And Diagnosis Using Finite-State Automaton

Abstract This paper addresses the problem of Fault-detection and Diagnosis using Finite-state Automaton Models (FAM). These models partition the state-space into finite regions and contain information on system trajectory within these regions. A framework for modeling faults using FAM is discussed. A Fault-detection engine is then developed to detect and isolate faults. The algorithm is illustrated with a case study.

Implementation of intelligent PID auto-tuning

The authors present a particular implementation of expert control. Specifically, relay feedback proportional-integral-derivative (PID) autotuning is presented. The main characteristics of the implementation are the blackboard architecture of the knowledge base and the clear separation of heuristic logic from the numerical algorithms. Various heuristics have been implemented to give a number of desired features. The result is an intelligent PID autotuner. This implementation attempts to attain the following goals: to achieve stable control with a minimal amount of prior process knowledge; to improve control with increasing process knowledge; to automatically select appropriate controller structures (PI or PID); to monitor control performance; and to build a system such that the control knowledge and heuristics are easy to modify and extend.<<ETX>>