Using scaled FE solutions for an efficient coupled electromagnetic–thermal induction machine model

Abstract

Purpose This paper aims to use a scaling approach to scale the solutions of a beforehand-simulated finite element (FE) solution of an induction machine (IM). The scaling procedure is coupled to an analytic three-node-lumped parameter thermal network (LPTN) model enabling the possibility to adjust the machine losses in the simulation to the actual calculated temperature. Design/methodology/approach The proposed scaling procedure of IMs allows the possibility to scale the solutions, particularly the losses, of a beforehand-performed FE simulation owing to temperature changes and therefore enables the possibility of a very general multiphysics approach by coupling the FE simulation results of the IM to a thermal model in a very fast and efficient way. The thermal capacities and resistances of the three-node thermal network model are parameterized by analytical formulations and an optimization procedure. For the parameterization of the model, temperature measurements of the IM operated in the 30-min short-time mode are used. Findings This approach allows an efficient calculation of the machine temperature under consideration of temperature-dependent losses. Using the proposed scaling procedure, the time to simulate the thermal behavior of an IM in a continuous operation mode is less than 5 s. The scaling procedure of IMs enables a rapid calculation of the thermal behavior using FE simulation data. Originality/value The approach uses a scaling procedure for the FE solutions of IMs, which results in the possibility to weakly couple a finite element method model and a LPTN model in a very efficient way.

Multi-Stage Optimization of Induction Machines Using Methods for Model and Parameter Selection

“…To model the IM, different electromagnetic machine models with different value ranges and level of detail are used. The thermal behavior of the machine necessary for the evaluation of the thermal constraints is modeled with an equivalent thermal network with four nodes, as presented in [37].…”

Approach for the Model and Parameter Selection for the Calculation of Induction Machines

“…The temperature measurement of the IM was performed on a test bench. The test bench setup is described in [22]. Seven temperature sensors were used to measure the average temperature of the stator winding.…”

“…This implies a high computational effort, especially for the numerical models, since the FE simulation must be performed at each step. On the other hand, for a given stator winding and rotor cage temperature, the machine can first be calculated with the selected model and the influences of the temperature changes on the losses can be determined using the scaling laws of IM described in [8,22,25,26]. This reduces the computational effort while maintaining the quality of the calculated losses, as shown in [8,22,25,26].…”