Veit Hagenmeyer

A new approach to inversion-based feedforward control design for nonlinear systems

The finite-time transition between stationary setpoints of nonlinear SISO systems is considered as a scenario for the presentation of a new design approach for inversion-based feedforward control. Design techniques which are based on a stable system inversion result in input trajectories with pre- and/or post-actuation intervals. The presented approach treats the considered transition task as a two-point boundary value problem (BVP) and yields causal feedforward trajectories, which are constant outside the transition interval. The main idea of this approach is to provide free parameters in the desired output trajectory to solve the BVP of the internal dynamics. Thereby, a standard MATLAB function can be used for the numerical solution of the BVP. Feedforward control design techniques are illustrated by simulation results for a simple example.

Brief Robustness analysis of exact feedforward linearization based on differential flatness

A methodology to analyze robustness with respect to parametric uncertainty for exact feedforward linearization based on differential flatness is presented. The analysis takes into consideration the tracking error equation and thereafter makes use of a stability result by Kelemen coupled with results issued from interval analysis theory.

Continuous-time non-linear flatness-based predictive control: an exact feedforward linearisation setting with an induction drive example

A general flatness-based framework for non-linear continuous-time predictive control is presented. It extends the results of Fliess and Marquez (2000) to the non-linear case. The mathematical setting, which is valid for multivariable systems, is provided by the theory of flatness-based exact feedforward linearisation introduced by the authors (Hagenmeyer and Delaleau 2003b). Thereby differential flatness does not only yield an easy calculation of the predicted trajectories considering the respective system constraints, but allows to use simple linear feedback parts in a two-degree-of-freedom control structure. Moreover, this formalism permits one to handle non-minimum phase systems, and furthermore to deal with parameter uncertainties and exogenous perturbations. Respective robustness analysis tools are available. Finally, an induction drive example is discussed in detail and experimental results for this fast electro-mechanical system are presented.

Flachheitsbasierter Entwurf von linearen und nichtlinearen Vorsteuerungen (Flatness-based Design of Linear and Nonlinear Feedforward Controls)

Abstract Modellbasierte Vorsteuerungen werden zur gezielten Beeinflussung des Führungsverhaltens einer Regelung verwendet. Nominale Vorsteuerungen benötigen das inverse Modell der Regelstrecke. Die damit zusammenhängenden Probleme der Realisierbarkeit einer Vorsteuerung und deren Instabilität im Fall von nichtminimalphasigen Strecken können für flache Systeme gelöst werden. Dabei werden die Solltrajektorien der Steuerung und Regelung in den Koordinaten des flachen Ausgangs geplant. Der flachheitsbasierte Entwurf von linearen und nichtlinearen Vorsteuerungen wird beispielhaft für einen Arbeitspunktwechsel von SISO–Systemen erläutert und an einfachen Beispielen demonstriert.

Internal dynamics of flat nonlinear SISO systems with respect to a non-flat output

In this article, the internal dynamics of flat nonlinear SISO systems with respect to a real non-flat output is investigated in the coordinates of the flat output and its derivatives. The main result constitutes that the parameterization of the real output by the flat output and its derivatives up to a certain order forms a differential equation which is diffeomorphic to the aforementioned internal dynamics.

Robustness Analysis With Respect to Exogenous Perturbations for Flatness-Based Exact Feedforward Linearization

A methodology to analyze robustness with respect to exogenous perturbations for exact feedforward linearization based on differential flatness is presented. The analysis takes into consideration the tracking error equation and makes thereafter use of a stability result by Kelemen coupled with results issued from interval analysis. This turns exact feedforward linearization based on differential flatness into a general control methodology for flat systems.

Deadbeat Control for Electrical Drives: A Robust and Performant Design Based on Differential Flatness

The present contribution introduces a new deadbeat controller design that increases robustness without compromising performance. In conventional deadbeat control, feedback linearization is applied, and the feedback gains are set very high to obtain the minimum-step reference response. This makes the control method highly sensitive to parametric uncertainties. To date, the only remedies have been to tune the deadbeat controller settling time higher and the according disturbance estimator more slowly. Recently proposed remedies based on online parameter estimators show either moderate performance or higher demands on hardware. Therefore, first a feedforward linearization-based controller is introduced to obtain the desired reference response via open-loop control. Thereby, the parametric sensitivity is considerably improved. Then, the new generalized flatness-based controller, a mix between feedback and feedforward linearization, is proposed. The result is a deadbeat controller with high dynamic performance and high robustness with respect to both parameter uncertainties and disturbances. The experimental results on an induction machine demonstrate very fast reference tracking, high robustness to typical parameter uncertainties, and active compensation of time-varying disturbances. The results on a synchronous reluctance machine show that even very large inductance uncertainties can be handled.

Optimal Exact Path-Following for Constrained Differentially Flat Systems

We propose a dynamic optimization approach to calculate optimal feedforward controls for exact path-following problems of differentially flat systems. Besides the derivation of a small dimensional optimal control problem, we provide easily checkable conditions on the existence of inputs guaranteeing that a given path is exactly followable in the presence of constraints on states and inputs. Our approach is based on the projection of the feedforward controlled, nonlinear MIMO dynamics along a geometric path onto a linear single-input system in Brunovsky normal form. The presented results indicate how the computation of admissible trajectories for set-point changes can be simplified by relying on steady state consistent paths. The set-point change of a Van de Vusse reactor is considered as an example.

Data processing of high-rate low-voltage distribution grid recordings for smart grid monitoring and analysis

Power networks will change from a rigid hierarchic architecture to dynamic interconnected smart grids. In traditional power grids, the frequency is the controlled quantity to maintain supply and load power balance. Thereby, high rotating mass inertia ensures for stability. In the future, system stability will have to rely more on real-time measurements and sophisticated control, especially when integrating fluctuating renewable power sources or high-load consumers like electrical vehicles to the low-voltage distribution grid.In the present contribution, we describe a data processing network for the in-house developed low-voltage, high-rate measurement devices called electrical data recorder (EDR). These capture units are capable of sending the full high-rate acquisition data for permanent storage in a large-scale database. The EDR network is specifically designed to serve for reliable and secured transport of large data, live performance monitoring, and deep data mining. We integrate dedicated different interfaces for statistical evaluation, big data queries, comparative analysis, and data integrity tests in order to provide a wide range of useful post-processing methods for smart grid analysis.We implemented the developed EDR network architecture for high-rate measurement data processing and management at different locations in the power grid of our Institute. The system runs stable and successfully collects data since several years. The results of the implemented evaluation functionalities show the feasibility of the implemented methods for signal processing, in view of enhanced smart grid operation.

Design of adaptive feedforward control under input constraints for a benchmark CSTR based on a BVP solver

Abstract The transition between multiple stationary setpoints of a benchmark CSTR model with different minimum-phase characteristics is solved by an adaptive feedforward control scheme. The feedforward control design treats the transition problem as a two-point boundary value problem in the coordinates of the input–output normal form, which is numerically solved with a standard M atlab function. The feedforward trajectories are calculated offline over the specified range of the inlet concentration to account for its strong influence on the system dynamics. The nominal trajectories are stored in lookup tables and an extended Kalman filter is used to adapt the feedforward control to the current inlet concentration. It is shown that the adaptation of the feedforward control acts as an integral feedback and leads to a zero steady state offset. Simulation results for a changing inlet concentration and model uncertainties show the performance and robustness of the adaptive feedforward control scheme.