A smart actuator not only needs to perform in a way what is in everyday life called intelligent, but also do this efficiently with a minimum use of energy. This goes hand in hand with a mechatronic design often leading to sophisticated integrated design solutions working at highest quality levels.
While simple control mechanisms react according to a predefined control law, smart actuators can distinguish in their behavior according to the further aspects concerning load situation, changed environmental or wear condition, etc. which are not commonly used as inputs in standard control. This allows the actuators to learn and adopt different patterns of behavior they would not be able to in conventional control.
To be smart one has to be efficient. The goal is to gain as much profitable output in terms of system performance from a system with a minimum of energy input. It can be seen as a very nice side effect that these efforts, next to the technical advantages, move the actuator into the direction of ‘green’ technology.
Sophisticated Mechatronic Design:
Intelligence and efficiency are meaningless if they are not implemented in a system designed to reach the limits of physics. The driving factors here are size, power density, performance and not at least costs. A sophisticated mechatronic design makes best use of the utilized materials, keeping costs low and allows a high degree of integration.
Reliability, repeatability, accuracy, precision and many more criteria defining the quality of a product are inseparably connected to smart actuators. For example a maintenance free lifetime combined with a hermetic design solution has major technological advantages and at the same moves the system toward environmentally friendly solutions.
Electronics and Power Electronics
Power electronics plays a major role for electric drives. It has to supply the power in exactly that way that the actuator, often designed at its physical limit, is able to perform in the desired way, while at the same time a high efficiency has to be reached. This leads to the development of non standard power electronic units for electrical drives.
Electric components used in modular capacitor discharge ignition systems play a similar role in defining the system performance.
Power Electronics for High Performance Electrical Actuators
The main area of research is dedicated to specific power electronics applicable for high dynamic operation of electromagnetic actuators.
Mechatronic System Modeling
Large systems inevitably lead to large models which very soon develop to be no longer manageable. This leads to the desire to keep the models as simply as possible but as precise as necessary to serve the given task. Hence various methods of modeling need to be applied while investigating a system. This leads to the need for different models of subsystems and components during the design and optimization process - all developed to lead to the best results with respect to their objectives.
Methods of Modeling Applied for Components and/or Systems:
Lumped Parameter Models (e.g. Equivalent Circuit Modeling)
Distributed Parameter Models
Finite Element Simulation
Multi Domain Simulation
Modeling of Linear Motors and Actuators
Due to their limited length, linear motors and actuators can exhibit significant end effects.
Based on a parametric geometry model, the electro-magnetic circuit is calculated using FE methods.
For the resulting model lookup tables are calculated that give the machine condition with respect to position an load current.
The calculation is implemented in the development and optimization tool MagOpt, which is developed at the JKU under participation of the JKU HOERBIGER Research Institute for Smart Actuators.
Reaching the physical limits is strongly connected with thermal issues. Thus a sufficiently accurate thermal modeling is necessary without loosing effort in exaggerated detailing. A solution is the approach of a distributed lumped parameter model, developed for estimating the thermal distribution of systems.
Given the example of a rotational geometry, the cross section is descretized and the properties are defined in the corresponding material matrix.
Each element of the matrix is defined using the equivalent electric network representation. Boundary elements are available to include convection, radiation or contact situations to the environment.
This allows calculating good estimations of the transient and static thermal distribution including thermal flow graphs.
As a result different load cases can be tested in simulation, maximum conditions can be derived and the cooling condition can be optimized with regard to the defined load requirements. This is especially important for linear drives or actuators as they are very often used under changing load conditions.
Multi Domain Simulation
The software tool MagOpt allows a parameterized multi domain simulation. An example is given with the multi domain simulation of a punching actuator developed by HOERBIGER.
The involved domains are:
Not all domains are equally important. Simplifications are necessary can individually be applied for different phases of development.
Smart Actuator Optimization and Mechatronic System Optimization
Smart actuators offer an optimal applicability with regard to the objectives they have been optimized for. Ideally these objectives can be directly derived from the system requirements. However if the investigated actuator is applied to different use cases, this is no longer possible. An example is a modern punching machine. It can perform a variety of tasks ranging from punching, nibbling or signing to pressing, bending or laser cutting. Thus one single optimization objective cannot be defined any more.
Such demanding tasks lead to the need of a flexible modeling environment and the application of a multi objective optimization procedure. Here the multi domain optimization software-tool MagOpt is applied. With the results given as Pareto surfaces of the optimization objectives the best solution for the desired use case can easily be selected.
Optimization of an Electric Rotational Actuator
The figure shows the projections of the resulting Pareto surface of the multi objective optimization. The finally selected optimum is a compromise between the weighting of the objectives for the given application.
Adaptive Nonlinear Model Predictive Control
Nowadays for many complex systems simple feedback control may not gain the full potential of the system and therefore the best possible performance. Further the relation between optimal tracking of intermediate references can have an influence on the overlying performance goal.
The concept of model predictive control can provide an alternative:
Optimal control of engines
A frequent industrial requirement:
Advantage of learning control:
Learning control can only be applied on repetitive systems to correct repetitive errors.
Objectives of current research projects: