Showing papers on "Model order reduction published in 1985"

The Application of Dynamic Programming in a Discrete Supplementary Control for Transient Stability Enhancement of Multimachine Power Systems

Abstract: This paper considers the use of a discrete supplementary control for transient stability enhancement of multimachine power systems. The principle advantage of the method presented in this paper is that off-line compuations are used. This permits the application of the supplementary control of a ``multimachine'' power system. The designation ``multimachine'' is used in this context to denote systems of up to three machines; for practical systems containing more than three machines, appropriate model order reduction is suggested. Supplementary controls differ from primary controls in that they are only used during disturbances rather than continuously. Attention is focused on the insertion of braking resistors and a specially designed dynamic programming algorithm is described for optimal resistor switching. The switching strategy is computed off-line so that the online computational requirements are minimal. The contribution of this paper is in the area of multimachine control and, in particular, a power system emergency state control. Topics concerning control models, optimization criteria and characteristics of discrete supplementary controls have been addressed. Control strategies are evaluated through digital and analog simulation involving a ten machine test system. A systematic approach to the optimization of multi-stage decision processes, called dynamic programming, was introduced by Bellman in the 1950' s [1-2]. Dynamic programming principles provide a perspective for examining optimization problems which lead to iterative functional equations. These equations are efficiently solved by using a digital computer and employing a dynamic programming computational procedure described by Larson [3].

An Engineering Approach to Model-Order Reduction and Its Application to Controller Design.

Abstract: A new method of approximating a high-order system by a lowerorder model in the frequency domain is developed. A reduced-order model of a transfer function obtained by the new method is compared with a reduced-order model obtained by another existing method to illustrate the power of the technique. Furthermore the approximant constructed by this new method is used to design a control system, which is applicable to the plant, to show its usefulness. The thesis also suggests some mathematical criteria for selecting the order of the approximant. This technique is particularly attractive because of its simplic­ ity and versatility. Furthermore, it is applicable to a large group of practical Single Input-Single Output systems' transfer functions.

Transfer of alignment and calibration of multiple sensors in flexible systems

Abstract: A method for the transfer alignment and calibration of sensors in flexible systems is presented. Flexible systems are represented by high order differential equations and reduced in dimension by model order reduction techniques. Both disturbability of alignment errors by control and disturbance inputs are considered. The reduced order model is decomposed by non-symmetric Riccati decomposition into states that directly and indirectly affect alignment errors. A transfer alignment and calibration filter with control feedthrough is developed and demonstrated for the example of a large flexible aircraft model in turbulence.

Dynamics of flexible spacecraft: an analysis approaches towards mathematical model order reduction

Abstract: Increases in size and mechanical complexity of spacecraft result in increased complexity of the mathematical model which describes the spacecraft dynamics. This results in turn in increased computational effort, in increased difficulties in understanding the characteristics of the spacecraft dynamics, and in increased difficulties in the design as well as the implementation of algorithms suitable for the control of the motion of the spacecraft. Reduction of the mathematical order (i.e., of the mathematical complexity) of the open loop spacecraft dynamics model with minimal loss of model accuracy is therefore paramount importance. Literature contains descriptions of a large number of approach towards open loop model order reduction. These have been analyzed from the point of view of usefulness for application to dynamic models of flexible spacecraft. This report describes five basic approaches that have been identified.These are: (i) parameter optimization, (ii) aggregation, (iii) singular perturbation, (iv) modal dominance, and (v) component cost analysis. The latter two approaches appear to be mosc meaningful and convenient amongst the five. The problem of model order reduction is reviewed. Each of the five approaches mentioned is described and commented upon, related approaches are discussed, and applications are described.