GOALI: A time-efficient analytical framework for optimal electromagnetic, thermal and structural design of switched reluctance motor

开关磁阻电机(SRM)作为永磁电机的低成本替代品和感应电机的高效率替代品，近年来受到了广泛的关注。这些电机可以在四个象限和宽恒速功率范围内运行，这使它们成为从电动汽车(EV)到暖通空调系统再到家用电器等各种应用的良好候选者。开关磁阻电机在定子上有集中的绕组，在转子上没有永久磁铁或绕组，这使得它们本身就是低成本和高效率应用的良好候选者。该电机还具有双凸极、高启动扭矩和低惯性，可以在没有大励磁涌流的情况下实现。它们本身也是容错机器，因为相绕组彼此隔离。尽管有这些优点，但阻碍开关磁阻电机广泛采用的关键问题包括高声噪声、振动和扭矩波动、低效率和功率密度。针对这些挑战，本研究提出了一种新的电机电磁与结构耦合性能同时优化的建模方法。它还将允许对非均匀条件和包括声学噪声和振动在内的三维现象进行建模。成功地完成这两个目标将导致设计技术比传统的数值方法更快。这种方法可以实现对电机模型的实时训练，从而开发出高效、安静的开关磁阻电机，并很容易扩展到其他电机类型。传统的机械设计中预测噪声和振动的方法通常是基于有限元分析(FEA)的确定性方法，其中电机是使用求解偏微分方程组(PDE)的有限元来建模的。虽然有限元分析可以用来处理定义在复杂区域上的问题，但它需要对整个计算区域进行完全离散化，这在数值上是不有效的。例如，要设计一台新机器，必须从基本形状开始，然后反复调整参数以满足所需的性能要求。在计算上，需要对区域进行多次离散化，并迭代求解偏微分方程组，以确定输入参数和输出性能之间的关系，这使得计算密集型。这项研究将使用核自由边界积分法(KF-BIM)来开发一种综合的、时间和计算效率都很高的电机设计方法。无核边界积分法是经典边界积分法的推广。它允许将不规则区域中的变系数椭圆型偏微分方程组转化为边界积分，而不需要格林函数的解析表达式。这种方法将被用作开发快速方法的基本框架，以准确地模拟机器的非均匀条件和3D现象，如噪声、振动和粗糙度(NVH)和热行为。这是一项跨学科的努力，工程学和应用数学为尖端的综合电机驱动设计框架做出了贡献。研究成果将通过学生在课堂和研究实验室中的积极参与、行业合作和研究演示来传播。该团队将继续与EDEC实验室的机电工程本科生合作，完成动手任务和行业级基准测试。将通过本科生跨专业项目(IPRO)课程、暑期研究浸入式项目以及向芝加哥城市学院、女性工程师协会和西班牙裔工程师协会发表演讲来加强女性学生和未被充分代表的少数群体的参与。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Switched reluctance motors (SRMs) have recently gained attention as a low-cost replacement for permanent magnet (PM) machines, and as a high-efficiency replacement for induction motors. These motors can operate in four quadrants and wide speed-constant power range, which makes them a good candidate for applications ranging from electrical vehicles (EVs) to HVAC systems to domestic appliances. SRMs have concentrated windings on the stator and do not have permanent magnets or windings on the rotor, which makes them an inherently low-cost and good candidate for high-efficiency applications. This motor also has double saliency, high starting torque and low inertia, which can be achieved without large inrush currents. They are also inherently fault tolerant machines since the phase windings are isolated from each other. Despite these advantages, critical problems hindering the wider adoption of SRMs include high acoustic noise, vibration and torque ripple, low efficiency, and power density. In response to these challenges, this research proposes a new modeling paradigm for simultaneous optimization of coupled electromagnetic and structural performance in electric machines. It will also allow modeling of non-homogeneous conditions and 3-D phenomena including acoustic noise and vibration. Successful completion of both these objectives will lead to design techniques that are faster than conventional numerical approaches. This method can be used to implement real-time training of the machine model towards the development of high efficiency, quiet switched reluctance motors, which can easily be extended to other motor types.Conventional approaches for prediction of acoustic noise and vibration in machine design are usually based upon a deterministic approach using finite element analysis (FEA), where the motor is modeled using finite elements solving the partial differential equations (PDEs). Although FEA can be used for problems defined on complicated domains, it requires full discretization of the entire computational domain, which is not numerically efficient. For example, to design a new machine, one must start with a base shape and then tune parameters iteratively to meet the desired performance requirements. Computationally, the domain needs to be discretized several times and PDEs are solved iteratively to identify a relationship between the input parameters and output performance, which makes it computationally intensive. The proposed research will use a kernel free boundary integral method (KF-BIM) approach to develop a comprehensive time and computationally efficient design approach for electric machine design. The kernel-free boundary integral method is a generalization of the classical boundary integral method. It allows the formulation of variable coefficient elliptic PDEs in irregular domain into boundary integrals, without requiring the analytical expression of the Green's function. This approach will be used as a fundamental framework to develop a fast approach to accurately model non-homogenous conditions and 3D phenomena such as noise, vibration and harshness (NVH) and thermal behavior of the machine. This is an inter-disciplinary effort with contributions from Engineering and Applied Mathematics towards a cutting-edge integrated motor drive design framework. Research outcomes will be disseminated through active student engagement in the classroom and research lab, industry collaboration and research presentations. The team will continue to work with undergraduate students from Electrical and Mechanical Engineering in the EDEC laboratory on hands-on tasks and industry-grade benchmarking. Participation by female students and under-represented minority groups will be enhanced through undergraduate Inter-professional Project (IPRO) courses, summer research immersion projects and presentations to the City Colleges of Chicago, Society of Women Engineers and Society of Hispanic Engineers.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.