Osvaldo Agamennoni

Robust controller design under highly structured uncertainty

A robust controller design method is presented for general delayed systems (including systems with delays in the denominator) with highly structured uncertainties. The methodology makes use of the critical direction concept and a controller structure previously presented by the author.

Advanced controller design for a distillation column

Abstract This paper considers the design of an advanced (relative to conventionally tuned multiloop PI control) control scheme for an industrial C3-splitter column using a range of recently proposed analytic tools. Due to product specifications, this column must be kept under tight control despite the effects of large disturbances and slow system response. Model predictive control was shown to provide a substantial improvement over a conventionally tuned multiloop controller but required a level of computation beyond that possible in a basic distributed control system. Controlled column behaviour comparable with that possible with model predictive control was shown to be achievable using a multiloop controller specifically tuned so as to incorporate disturbance rejection capabilities. The value of such controller design tools and their ready implementation within commercial software (here Matlab) should see their more widespread acceptance and use by the industrial control community.

Applying Continuous Piecewise Linear Approximations to Affine Non-Linear Control systems

Abstract In this paper continuous piecewise linear (CPWL) approximations of non-linear affine control systems are considered. An optimal control context allows to determine asymptotic error bounds between real and approximate control systems. Then it is possible to apply the controller designed using the CPWL approximation to the real vector field with guarantee of stability. An example is presented to show the potential applications of these ideas.

Robustness analysis of Wiener systems

Abstract As reported in the literature, Wiener models have emerge as an appealing proposal for nonlinear processes representation due to their simplicity and the property of being valid over a larger operating region than a LTI model. In this article, we propose a methodology to analyzed the robustness of a typical control scheme. To perform this analysis, we use a parametric description for the Wiener system. This model allows to describe the uncertainty as a set of parameters for the linear and the nonlinear blocks. Then, the linear block uncertainty is considered as a parameter-affine-dependent model and the nonlinear block uncertainty is studied as a conic-sector. The robustness analysis is then performed using μ -theory.

IDENTIFICATION OF UNCERTAIN WIENER SYSTEMS

Abstract A significant research work has been carried out on modeling, identification and control of processes represented by ***Wiener models. These models include a cascade connection of a linear time invariant system and a static nonlinearity. Several approaches can be found in the literature to perform the identification process. In this article, we describe a parametric description for the system, that allows to describe the uncertainty as a set of parameters. The proposed algorithm is illustrated through a pH neutralization process.

An Industrial Implementation of a Model Based Control Strategy

This paper considers the design, analysis and implementation within a commercial distributed control system of a model-based control strategy for an industrial C3splitter column. Due to product specifications and integration with other parts of the production complex, this column must to kept under tight control despite the effects of large, slow disturbances and slow system response.

A Generalized Loop-Shaping Technique for Design of Multivariable Feedback Control Systems

In this paper, we develop a loop-shaping method to deal with various uncertainty descriptions for robust control design. The applicability of the method to the case of structured perturbations is discussed. The block sensitivity in the M-¿ structure is exploited, yielding bounds on the structured singular value. These bounds help us develop an interactive design strategy for various controller modifications to shape the blocks of M matrix that have the largest impact on ¿.