RICERCA ITALIANA     

Research program

Magneto-elastic materials and optoelectronic sensors: technologies integration for "Smart" actuators and sensors realization.

University Co-ordinator

Seconda Università degli Studi di NAPOLI - INGEGNERIA DELL'INFORMAZIONE - CASERTA(CE)

Research Unit Leader

Alberto CAVALLO

Description

From the above analysis, it follows that the project requires different and specific skills in the fields of engineering and physics, which can be grouped in two highly interconnected topics:
• Physical analysis and modelling of the active material of the actuator. It concerns the experimental measurement of the properties of the active material, as well as the development of physical and phenomenological models of the materials themselves. Moreover, a further electromagnetic and elasto-magnetic analysis of the whole device, including embedded fiber optic for strain measurements, which allows to define the requirements of the control system.
• Definition, based on the requirements previously defined, of the control strategies in terms of the most effective and suitable techniques to realize the control (monovariable or multivariable, local or distributed) of the MST devices. This is accomplished by finalizing the knowledge acquired in the thematic area concerning the modelling of the material behaviour, with specific reference to compensation of nonlinearity and hysteresis of both the material itself and the whole device.
The research unit of the Second University of Naples will contribute to both topics with a number of activities described in detail next. The research topics tackled in each thematic area are summarized in the following. Within the first thematic area, the research unit can contribute with specific competences about the phase of parametric identification of physical and phenomenological models previously set-up and their integration in numerical simulators of control systems adopting MST devices. The second thematic area is of specific competence of the research unit, who can contribute in conceiving and setting up innovative control strategies for smart devices, as well as in implementation of embedded controllers in each device; in particular, new control strategies specially suited for self-sensing devices will be considered. Finally, the research unit will also address implementation of control strategies on single-board microcontrollers.
Summarizing, the unit of the Second University of Naples will be involved in Phases 1, 3, 4 and 5, where the following activities will be carried out.
• Phase 1: preliminary
Beyond the bibliographical research about properties and possible applications of MST materials, the operative unit will contribute on the definition of the characteristics of the devices to be realized (archetype) and of the specifications of a control system for active control of vibrations, which adopts the MST actuator.
• Phase 3: modelling
The first activity of the research unit concerns the modelling of the MST material to be used for development of the actuator archetype. The role in this phase of the unit, in charge of control algorithms development, will be of great importance, since the product of this phase consists in a simulator in MATLAB/SIMULINK environment, which will include the control law designed in the following phase. Moreover, the most suitable models for control design purposes will be selected, which are generally different from the detailed models aimed at describing as accurately as possible the behaviour of the physical process to be controlled. The developed models for control will try to reproduce the hysteretic behaviour of the material and they will be based on Preisach operators, for which the research unit has already produced effective techniques for identification based on neuro-fuzzy systems.
Since one of the objectives of the project is to develop a "fast" actuator, these models should include also dynamical effects which produce frequency dependent behaviours and can affect, not only the performance of the actuator, but also the stability of the closed-loop system. To this purpose the contribution of the other units will be crucial for obtaining a reliable dynamic model of the device, which takes into account also magneto-elastic interaction that can lead to resonant phenomena of great importance in the development of a ultra-fast actuator. In this phase, the models will be set up and tuned both on experimental data provided by the Torino unit in collaboration with the National Electrotechnique Institute Galileo Ferraris, and on measurements performed on a MST actuator by Energen Inc. available at the Automatic Control Laboratory of the Second University of Naples. This actuator is equipped with a high-resolution micro-displacement sensor and with a power unit that allows for measurement of the excitation magnetic field as well as the magnetic flux within the active material of the actuator. Moreover, embedded fiber optic strain sensors will yield high resolution measurements needed in micro-positioning applications, with the use of suitable control strategies to compensate for offset and low repeatability. The measurements will be performed on modern high-resolution data acquisition boards and the data will be elaborated in MATLAB environment, where the simulator will be developed in collaboration with the University of Sannio. The model validation will be carried out also on different MST inertial and resonant actuators available at the Automatic Control Laboratory from past research projects.
• Phase 4: compensation and control
In this phase, the requirements for the control system will be fixed first. Specifically, a position controlled and a force controlled actuator will be referred. Then, the specifications for the power amplifier will be defined and, if necessary, a custom design and production will be performed, so that all the requirements of the benchmark application can be satisfied. After the definition of specifications the selection of sensors and power unit, the design of the control strategy most suitable for each application will be addressed. Anyhow, the first step to be made consists in the realization of the compensation of the non linearity certainly present in the actuator, with the aim of getting an actuator as linear as possible, so as to minimize the causes of the closed-loop bandwidth limitation (possible limit cycles, lag effects, control signal distortion with excitation of high frequency unmodelled dynamics). The conceived control strategies will be tested in simulation on the detailed models available produced in Phase 3 and then they can be experimentally validated by resorting to dSPACE rapid prototyping systems, already available at the Automatic Control Laboratory of the unit. These systems consist of powerful data acquisition boards with high computational power owing to the adoption of DSP or Power-PC processors, fully integrated with the MATLAB/SIMULINK environment through the Real-Time Workshop which allows for automatic code generation optimised for real-time execution.
• Phase 5: demonstration
The first step of this phase will address the problem of possible implementation on commercial hardware of the proposed algorithm, in other words it will be decided how to implement on single-chip devices the compensation algorithm together with the force and/or position control strategies. To this aim, a computational and memory consumption analysis will be carried out, taking into account also performance. In fact, one of the objectives of the project is to direct the MST technology development for realization of smart devices, namely integrated systems comprising sensors, power and signal conditioning electronics, as well as control electronics.
This objective is of key importance for actual development of MST technology for an increasing number of applications in the near future. To make concrete this possibility, it is necessary to adopt single-chip devices, owing to their proved reliability and limited cost. Therefore, this phase will require evaluation of the above elements and, if necessary, will ask for reduction of algorithms complexity to prevent the un-applicability of the proposed approach from a real technological point of view. In this case the simplified algorithms will be first tested in simulation on the MATLAB/SIMULINK simulator previously defined.
These activities will be carried out in collaboration with the unit of Sannio, starting from the most suitable control device for the various benchmark applications, through a comparison among the different models available on the market. The current market offer for microcontrollers is highly diversified. Starting from powerful 32 bit systems, which include signal processing capabilities, large fast access on-board memories, A/D converters with conversion time of the order of microseconds at high resolution, serials or dedicated communication channels, and ending with standard and cheap 8 bit devices, with limited memory resources, low resolution converters and only one type of communication channel. At the same time, the functional and physical architecture of the control system will be defined.
Once the choice has be done, the final implementation phase can will start with a training period on the development system of the chosen control device. Then, the device will be programmed with the designed control algorithm and the real-time control law will be tested on the real actuator, in collaboration with all the other units, by demonstrating its functionality also in an application of active control of vibrations of a beam. A parallel activity, starting from the first results of this phase, will be the design a prototypal controller architecture, based on PC and PLC solutions for the hierarchical control.
All the activities described above, will be documented in periodical technical reports, with the aim of substantiate the performed research and actively exchange the technical information and results among the project consortium.