Model-Matching Controller Design With Input-Output Data—A Numerical Approach for Redundantly Actuated Systems

1994 ◽  
Vol 116 (4) ◽  
pp. 800-805
Author(s):  
Jenq-Tzong H. Chan
Keyword(s):  
Explicit Knowledge ◽  
Controller Design ◽  
Numerical Technique ◽  
Open Loop ◽  
Loop System ◽  
Synthesis Process ◽  
Model Matching ◽  
Input Output ◽  
Output Data ◽  
Redundantly Actuated

A numerical technique for control system synthesis based on input-output data is presented. The method is applicable when the system is open-loop stable and redundantly actuated. The major merits of the method are as follows. First, the closed-loop system equation may be arbitrarily assigned. Second, explicit knowledge of an open-loop system model is not needed for controller synthesis. Third, the stability of the synthesized system may be verified during the synthesis process; hence, the workability of the controller is ensured.

Multi-Stage System Identification of a Gas Turbine

2017 ◽  
Author(s):  
Amit Pandey ◽  
Maurício de Oliveira ◽  
Chad M. Holcomb
Keyword(s):  
Gas Turbine ◽  
Closed Loop ◽  
Open Loop ◽  
Loop System ◽  
Closed Loop Control ◽  
Input Output ◽  
Open Loop System ◽  
Loop Control ◽  
Multi Stage ◽  
Identification Techniques

Several techniques have recently been proposed to identify open-loop system models from input-output data obtained while the plant is operating under closed-loop control. So called multi-stage identification techniques are particularly useful in industrial applications where obtaining input-output information in the absence of closed-loop control is often difficult. These open-loop system models can then be employed in the design of more sophisticated closed-loop controllers. This paper introduces a methodology to identify linear open-loop models of gas turbine engines using a multi-stage identification procedure. The procedure utilizes closed-loop data to identify a closed-loop sensitivity function in the first stage and extracts the open-loop plant model in the second stage. The closed-loop data can be obtained by any sufficiently informative experiment from a plant in operation or simulation. We present simulation results here. This is the logical process to follow since using experimentation is often prohibitively expensive and unpractical. Both identification stages use standard open-loop identification techniques. We then propose a series of techniques to validate the accuracy of the identified models against first principles simulations in both the time and frequency domains. Finally, the potential to use these models for control design is discussed.

An LQ Controller With a Prescribed Pole Region—A Data-Based Design Approach

1997 ◽  
Vol 119 (2) ◽  
pp. 271-277 ◽  
Author(s):  
Jenq-Tzong H. Chan
Keyword(s):  
Closed Loop ◽  
Explicit Knowledge ◽  
Controller Design ◽  
Design Approach ◽  
Output Data ◽  
Plant Model ◽  
Modified Method ◽  
Input And Output ◽  
Z Domain

In this paper, we present a modified method of data-based LQ controller design which is distinct in two major aspects: (1) one may prescribe the z-domain region within which the closed-loop poles of the LQ design are to lie, and (2) controller design is completed using only plant input and output data, and does not require explicit knowledge of a parameterized plant model.

Control System Synthesis Based on Plant Test Data

1995 ◽  
Vol 117 (4) ◽  
pp. 484-489
Author(s):  
Jenq-Tzong H. Chan
Keyword(s):  
Control System ◽  
Test Data ◽  
Closed Loop ◽  
Explicit Knowledge ◽  
Linear Time ◽  
Open Loop ◽  
Loop System ◽  
Closed Loop System ◽  
System Synthesis ◽  
Control System Synthesis

A correlation equation is established between open-loop test data and the desired closed-loop system characteristics permitting control system synthesis to be done on the basis of a numerical approach using experimental data. The method is applicable when the system is linear-time-invariant and open-loop stable. The major merits of the algorithm are two-fold: 1) Arbitrary placement of the closed-loop system equation is possible, and 2) explicit knowledge of an open-loop system model is not needed for the controller synthesis.

FUZZY PREDICTIVE CONTROL OF UNCERTAIN CHAOTIC SYSTEMS USING TIME SERIES

1999 ◽  
Vol 09 (04) ◽  
pp. 757-767 ◽  
Author(s):  
LIANG CHEN ◽  
GUANRONG CHEN
Keyword(s):  
Chaotic System ◽  
Controller Design ◽  
Recursive Least Squares ◽  
Input Output ◽  
Output Data ◽  
Lyapunov Stability Criterion ◽  
System Input ◽  
Prediction And Control ◽  
And Control ◽  
Fuzzy Predictive Control

In this paper, a simple fuzzy logic based intelligent mechanism is developed for predicting and controlling a chaotic system to a desired target, using only input–output data obtained from the unknown (or uncertain) underlying chaotic system. In the chaos prediction phase, a fuzzy system approach incorporating with Gaussian type of fuzzy membership functions is used. Only system input–output data are needed for prediction, and a recursive least-squares computational algorithm is employed for the calculation. In the controller design phase, the Lyapunov stability criterion is used, which forms the basis of the main design principle. Some simulation results on the chaotic Sin map and Hénon map are given, for both prediction and control, to illustrate the effectiveness and control performance of the proposed method.

Optimal Output Feedback Regulator—A Numerical Synthesis Approach for Input-Output Data

1996 ◽  
Vol 118 (2) ◽  
pp. 360-366 ◽  
Author(s):  
Jenq-Tzong H. Chan
Keyword(s):  
Output Feedback ◽  
Output Feedback Control ◽  
Numerical Approach ◽  
Open Loop ◽  
Input Output ◽  
Linear Quadratic ◽  
Output Data ◽  
Optimal Output ◽  
Synthesis Approach ◽  
Feedback Regulator

A numerical approach is proposed in this work for computing a linear quadratic optimal regulator from input-output data. The method is applicable whenever the plant is open-loop stable. The major advantages of the method are two-fold. First, it involves an output feedback control law; hence, no state estimation is required for implementation. Second, the computation of this optimal controller can be conducted without explicit identification of the plant model.

Sign in / Sign up

Export Citation Format

Share Document