Miguel Castaño Arranz

A new approach to the dynamic RGA analysis of uncertain systems

This paper deals with DRGA analysis for uncertain systems. Uncertainties in a process model can be translated into uncertainties in the DRGA which might invalidate the decision on the variable paring in decentralized control. Bounds for the uncertainties in the DRGA of a multivariable process are derived for a given nominal process model with known uncertainty region. The resulting uncertainty region for the DRGA is used to assess the validity of decisions based on the nominal DRGA. The methodology for the computation of the bounds is currently restricted to the 2-by-2 case. Beside an explanatory example, a case study on a coal injection vessel is conducted and discussed.

Estimation of gramian-based interaction measures for weakly nonlinear systems

A critical step in the control design of industrial processes is the Control Configuration Selection (CCS), where each actuator is grouped with a set of measurements to be used in the computation of its control action. Tools for CCS include gramian-based Interaction Measures (IMs), initially defined for linear systems. Since a trending research topic is the derivation of IMs for nonlinear systems, a decision of the designer is therefore the approach to the problem in the linear or nonlinear framework. For this end, a method is discussed that determines the degree of nonlinearity of a system based on a specially tailored experiment, and thus enables the selection of the correct framework for the analysis. The novelty is in the estimation of two gramian-based IMs with confidence bounds from the tailored experiment which is applicable if the process is revealed to be weakly nonlinear.

Empirical approach to robust gramian-based analysis of process interactions in control structure selection

This paper deals with the estimation of a gramian-based interaction measure from logged process data, and thereby removing the need of creating parametric models prior to the selection of the significant input-output interconnections. Moreover, the resulting confidence regions of the estimates can be used to perform a robust control structure selection.

Control configuration selection for integrating processes using graphs

Recently, methods for Control Configuration Selection (CCS) using weighted graphs have been introduced. These methods were integrated in a software tool which allows systematic CCS in industry processes by choosing a simplified model with a reduced structure on which control will be based. This paper addresses the extension of these graph-based methods to be applied to integrating systems, which often appear in industry processes, e.g. at storage vessels. The extension is based on restricting the analysis to a set of frequencies of interest which omits the infinite DC-gain. The results are validated and illustrated using a real-life case, namely the secondary heating system of a pulp & paper mill.

New methods for structural and functional analysis of complex processes

The selection of the structure of a controller in large scale industry processes requires extensive process knowledge. The aim of this paper is to present methods which help designers to comprehend a process by representing structural and functional relationships from actuators and process disturbances to measured or estimated variables. This paper describes similarities between brain connectivity theory and interaction analysis in multivariable processes. Results from both theories are combined to create new methods for the analysis of industry processes. The developed methods are applied to an illustrative example.

Impulse response extraction and parametric modelling of reverberating ultrasonic echoes from thin layers

Enhancement of material impulse response buried in reverberating ultrasonic echoes from thin layered materials can be exploited in order to be able to detect possible flaws. One of the methods presented in this study is to enhance the impulse response of a material by training an adaptive filter that promotes an appropriate statistical characteristic such as asymmetry. The other approach is to employ a parametric linear model of reverberations that utilizes Maximum Likelihood Estimation on its parameters, to later suppress the reverberations and reveal possible flaws. Both approaches are investigated and shown to succeed under certain conditions and supported with experiments.

Online automatic and robust control configuration selection

This paper presents a complete method for automatic and robust control configuration selection for linear systems which relies upon acquired process data under Gaussian noise excitation. The selection of the configuration is based on the estimation of the Interaction Measure named Participation Matrix. This estimation is derived with uncertainty bounds, which allows to determine online whether the uncertainty is sufficiently low to derive a robust decision on the control configuration to be used or if the uncertainty should be reduced by e.g. prolonging the experiment to obtain more data.

3D synthetic aperture imaging using a water-jet coupled large aperture single transducer

This paper presents a technique for in-situ non-destructive testing of materials with applications in railway crossings. The novelty is in successfully applying the Virtual Source (VS) concept using water jet coupling for a large transducer. By focusing the sound field at the surface of the sample, the water jet probe can be built with a small nozzle opening, limiting the water consumption and making it viable for field applications. The annular geometry of the large transducer ensures the spherical wavefront assumed in the application of the SAFT algorithm, which usually limits the size of the transducer.