Reducing the complexity of thermal models for electric machines via sensitivity analyses - Université de technologie de Compiègne Access content directly
Conference Papers Year : 2017

Reducing the complexity of thermal models for electric machines via sensitivity analyses

K. El Kadri Benkara
G. Friedrich
S. Vivier

Abstract

The aim of the reduction method is to integrate the reduced model in a multi-physical electric machine optimization procedure and use the obtained model in real-time monitoring of machine critical parts. In this paper, an electric machine thermal model reduction method is proposed. Using the lumped parameter thermal network (LPTN) method, a thermal model is developed in order to represent the thermal behavior of an internal permanent magnet synchronous machine. From a thermal viewpoint, this machine is considered as a complex system, widely used as an integrated starter-generator (ISG) in automotive applications. The proposed reduction method is based on two sensitivity analysis techniques in order to obtain a simple and fast implemented thermal model. The first approach is the local sensitivity analysis where the ISG thermal model parameters were varied by ±50% around their nominal values. While the second and the following approach consists on applying one of design of experiment methods (full factorial design) to some of the thermal model parameters. In this case, parameters were varied in much larger boundaries or within their global domain of variation. These two techniques allow us to reduce the complexity of the initial model into a simpler and faster model while retaining a good accuracy. Results of this study are compared to data extracted from thermal tests performed on the ISG.
No file

Dates and versions

hal-01996968 , version 1 (28-01-2019)

Identifiers

Cite

B. Assaad, K. El Kadri Benkara, G. Friedrich, S. Vivier, A. Michon. Reducing the complexity of thermal models for electric machines via sensitivity analyses. 2017 IEEE Energy Conversion Congress and Exposition (ECCE), Oct 2017, Cincinnati, France. pp.4658-4665, ⟨10.1109/ECCE.2017.8096795⟩. ⟨hal-01996968⟩
34 View
0 Download

Altmetric

Share

Gmail Mastodon Facebook X LinkedIn More