Job van Amerongen

Modelling of physical systems for the design and control of mechatronic systems

Abstract Mechatronic design requires that a mechanical system and its control system be designed as an integrated system. This contribution covers the background and tools for modelling and simulation of physical systems and their controllers, with parameters that are directly related to the real-world system. The theory will be illustrated with examples of typical mechatronic systems such as servo systems and a mobile robot. Hands-on experience is realised by means of exercises with the 20-sim software package (a demo version is freely available on the Internet). In mechatronics, where a controlled system has to be designed as a whole, it is advantageous that model structure and parameters are directly related to physical components. In addition, it is desired that (sub-)models be reusable. Common block-diagram- or equation-based simulation packages hardly support these features. The energy-based approach towards modelling of physical systems allows the construction of reusable and easily extendible models. This contribution starts with an overview of mechatronic design problems and the various ways to solve such problems. A few examples will be discussed that show the use of such a tool in various stages of the design. The examples include a typical mechatronic system with a flexible transmission and a mobile robot. The energy-based approach towards modelling is treated in some detail. This will give the reader sufficient insight in order to exercise it with the aid of modelling and simulation software (20-sim). Such a tool allows high level input of models in the form of iconic diagrams, equations, block diagrams or bond graphs and supports efficient symbolic and numerical analysis as well as simulation and visualisation. Components in various physical domains (e.g. mechanical or electrical) can easily be selected from a library and combined into a process that can be controlled by block-diagram-based (digital) controllers. This contribution is based on object-oriented modelling: each object is determined by constitutive relations at the one hand and its interface, the power and signal ports to and from the outside world, at the other hand. Other realizations of an object may contain different or more detailed descriptions, but as long as the interface (number and type of ports) is identical, they can be exchanged in a straightforward manner. This allows top–down modelling as well as bottom–up modelling. Straightforward interconnection of (empty) submodels supports the actual decision process of modelling, not just model input and output manipulation. Empty submodel types may be filled with specific descriptions with various degrees of complexity (models can be polymorphic) to support evolutionary and iterative modelling and design approaches. Additionally, submodels may be constructed from other submodels in hierarchical structures. An introduction to the design of controllers based on these models is also given. Modelling and controller design as well as the use of 20-sim may be exercised in hands-on experience assignments, available at the Internet ( http://www.ce.utwente.nl/IFACBrief/ ). A demonstration copy of 20-sim that allows the reader to use the ideas presented in this contribution may be downloaded from the Internet ( http://www.20sim.com ).

Modelling, Simulation and Controller Design for Mechatronic Systems with 20-sim 3.0

Abstract Use of the modelling and simulation program 20-sim for analysis and design of mechatronic systems is described. In mechatronics, where a controlled system as a whole has to be designed, it is advantageous that model parameters are directly related to physical components and that models be reusable. Block-diagram-based simulation packages hardly support these features. 20-sim allows input of models in the form of equations, block diagrams, bond graphs and iconic diagrams. This will be illustrated by means of the design of a typical mechatronic system.

Learning Feed-Forward Control: A Survey and Historical Note

Abstract From a mechatronic point of view, the performance of electro-mechanical motion systems can be improved by changing both the mechanical design and the controller. The design of a controller is generally based on a model of the plant. Thus, to improve the controller, a more accurate model of the plant is required. When the structure is not known or when many parameters cannot be determined, learning control may be considered. A simple yet powerful learning control scheme that is suitable for electro-mechanical motion systems is Learning Feed-Forward Control. In this paper an overview is given of applications that have been reported concerning this scheme. Also, relations are listed with alternative learning control schemes that are in some sense alike.

A MRAS-BASED LEARNING FEED-FORWARD CONTROLLER

Inspired by learning feed–forward control structures, this paper considers the adaptation of the parameters of a model–reference based learning feed–forward controller that realizes an inverse model of the process. The actual process response is determined by a setpoint generator. For linear systems it can be proved that the controlled system is asymptotically stable in the sense of Liapunov. Compared with more standard model reference configurations this system has a superior performance. It is fast, robust and relatively insensitive for noisy measurements. Simulations with an arbitrary second–order process and with a model of a typical fourth–ordermechatronics process demonstrate this.

Computer support for mechatronic control system design

This paper discusses the demands for proper tools for computer aided control system design of mechatronic systems and identifies a number of tasks in this design process. Real mechatronic design, involving input from specialists from varying disciplines, requires that the system can be represented in multiple views. Several tools are already available but there are still substantial shortcomings. The paper gives indications about the developments needed to come to better design tools in the future. A specific example is worked out in more detail, i.e., automated performance assessment of mechatronic motion systems during the conceptual design stage.

Mechatronics Education and Research – 15 Years of Experience

Abstract This paper describes the experiences with mechatronic research projects and several educational structures in the University of Twente since 1989. Education took place in a two-year Mechatronic Designer programme, in specialisations in Electrical and Mechanical Engineering and in an (international) MSc programme. There are two-week mechatronic projects in the BSc curricula of EE and ME. Many of the PhD and MSc projects were done in projects sponsored by the industry or by application-oriented research programs. Research topics included modelling and simulation (learning) control, embedded systems and mechatronic design.

Performance Optimisation of Learning Feed Forward Control

The performance of sub-optimal feedback controllers can be improved in several ways. In this paper a learning control strategy is considered. The learning control system consists of the feedback and a feed forward controller. The feed forward controller is implemented as a neural network that is trained during control in order to minimise the tracking error. The type of neural network is a single layer network, in which B-spline basis functions are used to store the input-output mapping. The distribution of the Bsplines on the domain of the input(s) is of influence on the performance of the learning controller. Until recently, the basis functions were distributed by rule of thumb. In this paper fuzzy clustering techniques are used to obtain the distribution in a systematic way. In experiments the learning controller has been used to control a linear motor. Also when the B-splines are chosen by rule of thumb, the learning controller was able to improve the performance of the feedback controller considerably. The tracking error could be reduced further by determining the distribution of the basis functions using fuzzy clustering.

Continuous-Time Modelling in 20-sim

This chapter provides an introduction to continuous-time modelling of physical systems for readers whose main area of expertise is in the discrete-event modelling of computing and software. Fundamental concepts such as bond graphs and differential equations are presented together with graphical representations (block diagrams and iconic diagrams) used in the 20-sim tool. Control architectures in the form of feed forward and feedback controllers are briefly discussed, including sensors and actuators. Modern controllers are mostly implemented in computers. The sampled data controllers in the computer are coupled to the continuous-time plant by analogue-to-digital and digital-to-analogue converters. The roles of the sampling rate, the converter and arithmetic accuracy as well as event handing are briefly discussed. The need for discrete-event modelling of modern supervisory control software is identified.

Case Studies in Co-modelling and Co-simulation

We introduce two case studies in co-modelling and co-simulation, concentrating on the design problems presented by each of them. A simple line-following robot serves to illustrate the principles of co-model construction, fault modelling and design space exploration. An experimental self-balancing scooter provides insights into the problems posed by the industrial practice of developing a safety-related system.

Off-axis modal active vibration control of rotational vibrations

Collocated active vibration control is an effective and robustly stable way of adding damping to the performance limiting vibrations of a plant. Besides the physical parameters of the Active Damping Unit (ADU) containing the collocated actuator and sensor, its location with respect to the compliances in the plant plays an important role in the performance of the active vibration controller (AVC). This paper presents a model based evaluation of the interaction between the AVC and the plant, for two variants of the ADU location in the plant. It is concluded that the ADU should be placed as close as possible to the end effector to achieve optimal damping and exclude the excitation of undesired dynamics by the AVC.