Advanced actuator design

This theoretical course provides graduate students with knowledge and tools to design and optimize advanced electromagnetic actuators for various applications. The translation from system requirements into specific actuator requirements is presented by means of practical examples from the industry. Different mathematical electromagnetic modeling techniques for calculation of the important performance parameters of different actuator classes are covered. The student will utilize these models on an example from industry, throughout the course, in the form of a group project, and finally apply his acquired skills on a specific problem during assignments. The student not only learns on a physical level to calculate complex electromagnetic phenomena in actuators but also obtains practical knowledge from the industry covering topics like manufacturability, material properties and insulation aspects. In addition, students will have a chance to evaluate their course knowledge with a course content aware chatbot, this opportunity is made possible thanks to project (EXAMAI) on AI in education within our department.

Content keywords: Hard and soft magnetic materials. Hysteresis, eddy currents and excess losses. Short stroke and long stroke actuators. Voice coils. Reluctance actuator. Linear induction motors. Linear permanent magnet actuators. End effects. Laplace and Poisson equation. Boundary value problems. Fourier analysis. Magnetic charge and current model. Lorentz force. Maxwell stress tensor. Virtual work method. Thermal equivalent circuit, Finite Element Method. Constrained nonlinear minimization. Multi‐physical design. Orthocyclic winding. Rectangular wire winding.

The objective of this course for graduate students of electrical engineering to obtain knowledge about physics of advanced actuators, their design aspects and modelling methods. The student will learn methods to compute magnetic field distribution in air and in combination with hard and soft magnetic materials applicable for a wide variety of actuator topologies. The student will learn to derive and implement different modeling techniques, the magnetic equivalent circuit, Fourier analysis, charge and current models and basics of Finite Element Method. Furthermore, the student will learn several methods for force calculation (Lorentz force, Maxwell stress tensor and virtual work method) and will be able to decide depending on the application, which method should be applied. Finally, the student will learn how to apply analytical models to calculate the temperature distribution based on the thermal equivalent circuit method. All these models together give the student a multi‐physical insight into the behavior and dependencies of the various geometric and material parameters on the performance.

Furthermore, the student will apply these models in a multi‐physical design optimization in which the actuator topology can be optimized in a fast manner to scan various topologies. These learnings, the students will directly use in a hands-on group project concerned with a real world design, i.e. in close collaboration with industry.

On a high level, this course focuses on two main parts which are design and modelling. At this end, the design cycle includes topology, materials selection, model definition, parametrization and finally performance optimization with respect to objectives. In addition to well-established (MOO) optimization methods, the course will offer new paradigms from AI domain, advantageous in optimization, reduced computation complexity, such as, Bayesian optimization and surrogate modeling.

According to Bloom’s taxonomy, the following learning objectives are defined:
Students will be able to:

1. Knowledge related

a. Recall underlying physics behind working principles of actuators, which includes magnetic field behavior in air, soft and hard magnetic materials, as well as thermal principles.
b. Identify Modelling techniques (magnetic equivalent circuits, Fourier analysis, charge/current models, Finite Element Methods), and force computation methods (Lorentz force, maxwell stress tensor and virtual work method)

2. Comprehension related

a. Explain the relationship between geometric and material parameters and their impact on the actuator performance. More specifically, how reluctance force is created, what are the roles of magnetic saturation and other nonlinear effects in actuators.
b. Interpret the role and relevance of multi-physical models (magnetic, thermal and mechanical) in the actuator behavior.

3. Application related

a. Apply appropriate modelling technique to compute magnetic field distribution in air and magnetic materials for various actuator topologies.
b. Apply analytical models (such as thermal equivalent circuit) to calculate temperature distributions.
c. Use force calculation methods (Lorentz force, Maxwell stress tensor, virtual work) in practical scenarios.

4. Analysis related

a. Build the ability to decompose and identify the design cycles: topology selection material choice, model definition parametrization and performance optimization.
b. Compare, contrast and choose modelling techniques (e.g. magnetic equivalent circuits vs FEM) for specific actuator applications.
c. Analyze the trade-offs between computational complexity and accuracy in modelling methods

5. Evaluation related

a. Critically evaluate your choice on modelling method chosen (FEM, surface charge, MEC), force computation, and or loss models, depending on the case.
b. Asses the effectiveness of the optimization paradigms chosen (Bayesian, surrogate modelling or simple parameter sweep)
c. Justify the selection of materials based on performance criteria

6. Creation related

a. Design and optimize actuator topologies using the toolset provided in the course
b. Develop a real-world actuator design in a group project, integrating all relevant aspects of physics and design.
c. Propose novel actuator solutions combining the knowledge gained from the course.

Basic knowledge of electromagnetics and electrical engineering. Familiarity with mathematics and numerical modelling. Experience or motivation to learn with programming, and scientific computing is recommended.

Lectures; guided modelling exercises; group project; independent study; design optimization assignment; oral examination/presentation of project results.

• Permanent magnet and electromechanical devices, E. Furlani