Stefan Palis

Diskrepanzbasierte Regelung der kontinuierlichen Kristallisation

Zusammenfassung Der Beitrag beschreibt einen neuen Regelungsansatz zur Stabilisierung von Kristallisatoren. Es ist bekannt, dass diese in Verbindung mit einer Feinkornauflösung zu Instabilitäten und dem Auftreten von unerwünschten Oszillationen neigen. Modelle für Kristallisatoren bestehen meist aus einer nichtlinearen partiellen Integrodifferentialgleichung, der Populationsbilanz für die Kristallgrößenverteilung, gekoppelt mit einer gewöhnlichen Differentialgleichung zur Beschreibung der flüssigen Phase, was einen rigorosen Reglerentwurf vor erhebliche Probleme stellt. Deshalb wird im Rahmen dieser Arbeit auf ein Stabilitätskonzept zurückgegriffen, welches Aussagen über die Stabilität eines Prozesses bezüglich zweier verallgemeinerter Abstandmaße, den sogenannten Diskrepanzen, erlaubt. Abstract In this contribution a new control approach for stabilization of crystallizers is presented. It is well known that instabilities and undesired oscillations may occur, when operated with fines dissolution. In general, crystallizer models consist of a nonlinear partial integro-differential equation, the population balance for the crystal size distribution, coupled with an ordinary differential equation for the liquid phase. Therefore, rigorous control design is a challenging issue. In order to overcome this problem in this contribution a stability concept is used, which states stability with respect to two generalized distance measures, the so called discrepancies.

Nonlinear model reduction of a continuous fluidized bed crystallizer

This work considers a system of a fluidized bed crystallizer and an ultrasonic attenuator, which separates an enantiomer from a liquid solution. A population balance model of the system shows autonomous oscillations over a wide range of operation conditions. Proper orthogonal decomposition is applied to obtain nonlinear reduced models of low order. An a posteriori error estimator is used to assess the quality of the reduced model during run time.

Discrepancy Based Control of Continuous Fluidized Bed Spray Granulation with Internal Product Classification

Abstract This paper is concerned with stabilizing open loop unstable fluidized bed spray granulation with internal product classification by means of nonlinear feedback control. The process dynamics are described by a population balance model, i.e. a nonlinear partial integro-differential equation. Therefore a direct stabilization in a classical Lp-norm is at least very difficult. To overcome this problem a stability concept using two generalized distance measures, called discrepancies, is used. In order to show that the discrepancy based control approach presented in this contribution also guarantees stability in the sense of a Lp-norm or L∞-norm additional conditions are stated and verified.

Stabilization of Continuous Fluidized Bed Spray Granulation - a Lyapunov Approach

Abstract This article deals with the stabilization of continuous fluidized bed spray granulation with external product classification. Using Lyapunov stability theory a control law is derived guaranteeing exponential convergence of the error particle size distribution in a norm associated to the second moment. In contrast to other approaches popular for particle process control no model reduction is used.

Design of communication systems for networked control system running on PROFINET

This paper shows a way of modelling the timing of an Ethernet-based communication system and examines the influence of a specific configuration on a closed-loop control. The communication system is modelled as asynchronous processes and a distribution of the delay times is derived. Using these delays, the stability criterion is checked and a Monte Carlo simulation of the step response is performed to examine the behaviour of a real application. Besides stability, performance criteria are evaluated as alternative for the specified control scenarios.

Identification of micro parameters for discrete element simulation of agglomerates

The mechanical behaviour of solid particles like agglomerates, granules or crystals strongly depends on their micro structure, e.g. structural defects and porosity. In order to model the mechanical behaviour of these inhomogeneous media the discrete element method has been proven to be an appropriate tool. The model parameters used are typically micro parameters like bond stiffness, particle-particle contact stiffness, strength of the bonds. Due to the lack of general methods for a direct micro parameter determination, normally laborious parameter adaptation has to be done in order to fit experiment and simulation. In this contribution a systematic and automatic way for parameter adaptation using real experiments is proposed. Due to the fact, that discrete element models are typically systems of differential equations of very high order, gradient based methods are not suitable. Hence, the focus will be on derivative free methods.

Nonlinear adaptive control of magnetic bearings

The presented paper shows how nonlinear adaptive control theory can be applied to magnetic bearings. Here the backstepping formalism is used to derive a stabilizing control law, which guarantees stability in the sense of Lyapunov in the presence of a unknown load. The efficiency of the proposed method comparing to linear standard design (e.g. PID) is demonstrated through experimental investigations.

Anti-Sway System for Slewing Cranes

In the last few years, automation of slewing cranes is gaining more and more importance in transport and handling processes. However, each operation of crane with rope-suspended loads provokes swaying movements of load. Neutralization of these swaying movements represents a key problem when high handling speeds, operational reliability, accurate traveling trajectories and positioning are required. Parameter variations and force interactions between rotary and translation movements in slewing cranes call for a special approach when designing the control scheme. The presented paper deals with optimal control of slewing cranes. It will show that optimal control strategy copes very well with the mentioned problems.

High PerformanceTracking Control of Automated Slewing Cranes

Automation of slewing cranes in handling and transport processes is getting more and more important. However, each crane operation causes swaying movements of load, which are weakly damped. Therefore active damping of swaying movements is a major task in automated crane operation. The main difficulty here, in contrast to overhead travelling cranes, is the coupling between rotary and translation movements, which calls for a special approach when designing the control scheme.

Adaptive discrepancy based control of continuous fluidized bed spray granulation with internal classification

This paper is concerned with adaptive stabilization of open loop unstable uidized bed spray granulation with internal product classication by means of nonlinear feedback control. Since the process model is represented by a nonlinear partial integro-dierenti al equation, direct stabilization of the particle size distribution in a Lp or L1 norm is dicult. To overcome this problem a stability notion using two generalized distance measures, the discrepancies, is used. It is shown that the adaptive version of the resulting discrepancy based control law is able to cope with uncertainties present in industrial applications.