Hybrid Resolver Model Generalization for Fault Condition Modeling: A Promising Tool for Reliability Study

TL;DR

This paper introduces a generalized hybrid resolver model that efficiently simulates various fault conditions, aiding reliability analysis and fault diagnosis with high accuracy and reduced computational time.

Contribution

A novel hybrid modeling approach for resolvers that accurately simulates multiple fault types and operational conditions, facilitating reliability studies and fault diagnosis.

Findings

Model closely matches Time-Stepping Finite Element Analysis results.

Significant reduction in simulation time compared to traditional methods.

Experimental verification confirms model accuracy.

Abstract

Resolvers, like all electromagnetic devices, are constantly under investigation, both operationally and structurally. In this regard, proposing a modeling methodology that can save significant time without compromising accuracy is a big honor. In this study, a generalized hybrid model is suggested that, in addition to the above benefits, has sufficient capability to ease reliability study in the field of resolvers, where a large number of faulty conditions must be investigated under different operating conditions, including changes in angular velocity, voltage, and frequency of excitation; all of which are highlighted in the context of fault coverage. This model also serves as a promising tool for generating large datasets, which is advantageous for fault diagnosis. A resolver with a non-uniform air gap is chosen as a case study to challenge the suggested model, particularly in relation to eccentricity faults. We generalize the suggested model to account for the most common faulty conditions of resolvers: in-turn short circuits in signal and excitation windings, as well as static and dynamic eccentricity faults. The close agreement between the results of the suggested model and those from Time-Stepping Finite Element Analysis (TS-FEA), along with significant time savings in both healthy and faulty conditions, highlights the generality and proficiency of the suggested model. Finally, the case study is prototyped, and we verify the accuracy of the suggested model experimentally.