Compensating angle estimation errors caused by magnetic saturation in anisotropy-based sensorless control schemes

In anisotropy-based sensorless control schemes, the influence of magnetic saturation is usually neglected. This causes angle estimation errors, especially under high-load conditions. By applying a more accurate machine model, this error can be effectively compensated. The present work gives a short introduction to the non-linear model applied and analyzes the angle errors in both, open-loop and closed-loop sensorless operation. Additionally, an algorithm is proposed, that uses the model information to compensate the angle estimation error. The effectiveness of this approach is verified by measurements on a synchronous reluctance machine.