Additive manufacturing has many advantages over traditional manufacturing methods and has been increasingly used in medical, aerospace, and automotive applications. The flexibility of additive manufacturing technologies to fabricate complex geometries from copper, polymer, and ferrous materials presents unique opportunities for new design concepts and improved machine power density without significantly increasing production and prototyping cost. Topology optimization investigates the optimal distribution of single or multiple materials within a defined design space, and can lead to unique geometries not realizable with conventional optimization techniques. As an enabling technology, additive manufacturing provides an opportunity for machine designers to overcome the current manufacturing limitation that inhibit adoption of topology optimization. Successful integration of additive manufacturing and topology optimization for fabricating magnetic components for electrical machines can enable new tools for electrical machine designers. This article presents a comprehensive review of the latest achievements in the application of additive manufacturing, topology optimization, and their integration for electrical machines and their magnetic components.
The electrical machine is considered a key part in electric drives, which account for approximately 50% to 70% of electricity usage in the EU and the United States [1]. Its applications include, but are not limited to, compressors, HVAC systems, power tools, generators, electric and hybrid vehicles, elevators, and MAGLEV trains. In the last decade, there have been consistent efforts from both the US Department of Energy and the EU to advance the design and development of future generations of electrical machines that positively impact the environment and reduce greenhouse gas emissions [1,2]. The next generation electrical machines include designs with high efficiency and power density; however, another important aspect is their environmentally friendly construction, including aspects such as minimal material waste and recyclability.
The shaping of the magnetic structures for electrical machines can be generally categorized into two groups: (1) conventional shaping and optimization techniques, and (2) topology optimization. For the conventional shaping techniques, mathematical models and sensitivity analysis are typically used on a pre-selected machine geometry template [66,67]. The computational and analytical efforts are often intensive to improve the accuracy of the calculation of the airgap flux density, torque components, and the magnetic flux density distribution [68]. Conventional optimization, which is typically based on evolutionary multi-objective optimization algorithms, further refines the shape of the magnetic structures to improve the machine performance. As a result, the derivation of uniquely shaped magnetic structures for electrical machines can be slow.
Emerging from structural optimization, TO is increasingly applied in magnetic devices [69], and subsequently in design of electrical machines, especially at the component level such as magnetic cores [70] and permanent magnets [71]. In contrast to conventional optimization, TO can generate an initial geometry template from scratch with less analytical modeling [72]. In general, TO investigates optimal distribution of single or multiple materials within a defined design space [73]. Compared to conventional optimization shaping techniques, it offers additional flexibility in optimizing the geometry of the magnetic components for attaining the desired performance. Thus, TO can yield unique shapes that are generally not realizable with conventional optimization approach.
, which can be assigned a value of zero or one, as illustrated in Figure 9a. Zero and one indicate the absence and presence of material, respectively. The pattern of the material distribution can be determined via selection of objective functions and use of evolutionary multi-objective or gradient-based algorithm. Thus, in a finalized topology optimized design via on/off method, it typically has an unconventional geometry.
An on/off TO is modified to optimally distribute the soft magnetic material for a rotor core of an interior permanent magnet machine and then to smooth the shape of the design is presented as shown in Figure 10. Here, the TO algorithm in [8] first uses a genetic-based method to find an optimal solution in the global search space. The solution is then smoothed out via the use of a gradient-based method in the local search space. The illustration of the modified algorithm is shown in Figure 11.
The on/off TO method can also be applied to find the optimal distribution of multiple materials such as iron, copper, and permanent magnet. In [74], a multi-material on/off TO algorithm is used to maximize the force acted on the plunger of a permanent magnet linear actuator. The TO optimal design achieves a unique structure compared to the original design and an additional increase of 40% in average force, as shown in Figure 12.
As the on/off method assigns the optimizing density variable to be binary, the density-based method assigns the density variable ρn
As TO is increasingly adopted in developing unique geometries for electrical machines, manufacturability of the unique magnetic core designs is equally important. Design complexities may increase manufacturing cost for electrical steel laminations. Additionally, manufacturing methods including subtractive techniques can compromise the magnetic properties of punched laminations [17], leading to magnetic cores with inferior magnetic performance compared the mother coil. Thus, topology optimized designs in Figure 10 and Figure 14 may not be able to achieve the desired magnetic performance. Powder metallurgy can potentially be used as an alternative approach to fabricate such complex designs, as shown in Figure 15. However, the added cost of molding and tooling may become a concern.
In application where non-homogeneous magnetic core is desired as in [79,80], powder metallurgy manufacturing approach is not a viable solution for production as it may reach the limits in fabricating such composite, non-homogenous structures. Similarly, designs of magnetic cores for electrical machines generated with the TO density-based method as in [9,76], may request material whose properties may not correspond to an available material [81]. Additive manufacturing can potentially overcome difficulties typically observed in conventional manufacturing methods, and in some cases is the only viable manufacturing solution [82].
Recent advancements in AM as well as the proliferation of its application in fabricating magnetic components for electrical machine have revitalized TO as an advanced design tool. The synergy between TO and AM can potentially lead to the development of magnetic components, whose properties and geometries are complex. Investigations in integration of TO toward AM in producing magnetic components for electrical machines have shown very promising results. In [10], a topology optimized design of a rotor core of a surface mount permanent magnet machine is additively manufactured via the SLM process, as shown in Figure 16. Here, the TO algorithm combines both the electromagnetic and structural optimization stages to achieve a rotor core geometry with 50% reduction in weight, at a tradeoff of less than 2% in average torque, while achieving maximum von Mises stress in the optimized rotor core well within the yield strength of the material. The result of this work highlights the exploitation of multi-physics TO as an advanced design tool for AM in developing new, unconventional electrical machines.
The integration of TO into AM is also seen in fabricating permanent magnet with unique shapes and structures. In [71], TO is implemented to generate the design of a permanent magnet such that it can provide magnetic flux density waveform close to the predefined external field. The design is then additively manufactured via the FDM process from a magnetic compound of NdFeB powder, ferrites, and polymers, Figure 17. In [11], permanent magnet made of multiple magnet grades is proposed for a surface mounted machine to reduce manufacturing cost without penalizing machine performance. The multi-grade magnet is optimized, and investigated via finite element analysis. Although TO is not implemented, the analysis suggests the combined use of TO and AM technologies in potentially producing lower cost magnets.
Figure 17. Topology optimized magnet. The design is then additively manufactured via fused deposition modeling. Figure is modified from [71].
This entry is adapted from the peer-reviewed paper 10.3390/en14020283