From Simulation-Driven to Optimization-Driven Electric Motor Design

MotorXP - Electric Machine Design and Analysis Software

1. Electric motor design optimization overview

Electric motor automatic optimization is a sophisticated process that leverages advanced software and algorithms to enhance the design of electric motors, maximizing their efficiency and performance. The primary objective is to adjust the motor’s geometry, and potentially other parameters, to achieve optimal performance metrics such as torque density, power density, efficiency, etc., while simultaneously reducing weight, costs, and environmental impact. This approach is based on the principle that even minor adjustments to the motor’s geometry can yield substantial improvements in performance.

Automatic optimization plays a vital role in the pursuit of more efficient, power-dense, and cost-effective electrical motors. Unlike human electric motor designers, optimization algorithms can swiftly evaluate hundreds or even thousands of design variations within a matter of hours, a task that would take weeks or months for manual assessment. The improvement of motor design performance is achieved through the gradual refinement of design candidates on each iteration of the optimization algorithm, which automatically uncovers the intricate relationships between optimization objectives and the varying parameters of the motor, adjusted to achieve optimal performance.

The overarching principle of the multi-objective optimization of an electric motor is exemplified in the pursuit of two optimization objectives: minimizing motor weight and minimizing motor losses while maintaining a specific output power, as shown below.

Optimized designs converge towards the Pareto front, representing the optimal trade-offs between motor weight and losses. As the motor becomes lighter, it tends to generate more losses due to reduced materials such as copper and magnets, and vice versa. In cases with three optimization objectives, the Pareto front transforms into a surface, highlighting the intricate nature of the optimization process.

Traditionally, simulation-driven motor design involves evaluating each design candidate individually, leading to extended development timelines and potentially overlooking optimal design candidates. Conversely, optimization-driven motor design focuses on exploring design spaces, where numerous design candidates share similar motor parameters, such as motor topology, slot and pole numbers, winding arrangement, etc., but vary in geometric dimensions. Optimization algorithm is capable of efficiently exploring these design spaces, enabling motor designers to compare different topologies, slot/pole combinations, winding arrangements, and materials at a higher level without delving into particular geometry details, thereby facilitating the creation of more optimized motor designs and reducing time to market.

2. MotorXP electric motor optimization flowchart

 


 

3. MotorXP features for the development of the electric motor optimization workflows

1. MotorXP-MATLAB API

We utilize MATLAB for running the optimization process. MATLAB stands as the preferred programming language for millions of engineers and researchers worldwide. All our MATLAB codes concerning optimization are open-source, allowing you to freely understand their functionality and customize the optimization workflows to suit your requirements. For instance, the additional MATLAB code incorporated with the optimization algorithm can calculate specific optimization objectives, search for the optimal advance angle to maximize torque in an IPM motor, or implement custom penalty strategies to guide the optimization algorithm in searching for design candidates with preferred performance metrics.

Additionally, MATLAB offers a wide range of data visualization options, enabling customization of optimization result representations (such as Pareto plots) and the incorporation of user-defined context menus, as shown below.

MATLAB also facilitates communication with various third-party models and algorithms, enabling the incorporation of additional aspects such as thermal, mechanical, CFD, NVH, power electronics, and more into the optimization workflow.

2. Parallel processing

Utilize all available CPU cores in your system to process the design candidates generated by the optimization algorithm. This approach significantly accelerates the optimization process.

3. Customizable parameterized geometry templates

In addition to utilizing existing geometry templates, you have the option to create a fully parameterized geometry template with a dedicated user interface using our geometry template API. This allows for the customization of rotor or stator geometry templates of any complexity. The parameters of the template can then be linked to the optimization algorithm for the customized geometry optimization.

4. Automatic detection of geometry conflicts and erroneous design candidates

It is advisable to prevent potential geometry conflicts by properly defining optimization variables (e.g., utilizing ratio-based variables instead of absolute values) and carefully selecting optimization variable ranges. However, when working with intricate geometries, like multi-layer IPM rotors, geometry conflicts may arise unavoidably. In such instances, MotorXP will automatically identify these conflicts during the design assembly stage.

Related articles

AI-powered electric motor design

Practical implementation of the automatic motor design workflow using the automatic optimization features of the MotorXP software. Read article>>

Mystery of the Tesla’s Next-Gen Zero-Rare-Earth Electric Motor: Feasibility Study of Transitioning to Ferrite Magnets for EV Traction Motors

Designing an axial flux permanent magnet motor with ferrite magnets to match the efficiency and power density of the Tesla Model 3 motor using the automatic optimization features of the MotorXP software. Read article>>