Cost remains a major obstacle to the widespread adoption of electrified vehicles. Significant research effort has been expended in trying to replace hardware with software in systems in an effort to bring it down. To accomplish this, phase current estimation has emerged as a viable means of removing sensors from electric drives. The plurality of work in the field employs linear current observers, which suffer from a need for machine anisotropy and periodic position-dependent nonobservability, both of which limit applicability. This article proposes an optimization-based methodology that eliminates all estimation constraints. It is shown that the problem is always at least quasiconvex, permitting the estimation of the dq fluxes under a set of operating constraints. By way of an additional term in the cost function, the problem is made strictly convex for all operating conditions, enabling efficient estimation of the unique dq fluxes. Experiments validate the concept and show that the optimization-based observer performs similarly to the linear observer in the steady state with ringing after transients reduced.

中文翻译：

---

基于优化的永磁同步电机简化传感器虚拟磁通观测器


成本仍然是电动汽车广泛采用的主要障碍。人们投入了大量的研究工作来试图用系统中的软件取代硬件，以期将其击垮。为了实现这一目标，相电流估计已成为从电力驱动器中去除传感器的可行方法。现场的多种工作采用线性电流观测器，其需要机器各向异性和周期性位置相关的不可观测性，这两者都限制了适用性。本文提出了一种基于优化的方法，消除了所有估计约束。结果表明，该问题至少总是拟凸的，从而允许在一组操作约束下估计 dq 通量。通过成本函数中的附加项，该问题对于所有操作条件都是严格凸的，从而能够有效估计唯一的 dq 通量。实验验证了这一概念，并表明基于优化的观测器在稳定状态下的表现与线性观测器类似，但瞬态后的振铃减少了。