SM-PM

For both scenarios, here is the list of thermal settings needed to initialize the test.

Table 2. List of thermal settings

1	Dialog box to define thermal settings. Iterative solving mode is selecting.
2	With the second solving mode (i.e. with the iterative thermal solving mode) the coolant temperatures must be defined. Definition of coolant temperatures (if there is a cooling circuit) to be considered: External fluid temperature (1) Cooling circuit fluid temperature (2)
3	Temperature of magnets to be considered must be defined: Range of temperatures between the minimum and the maximum ones Number of computations to be considered within the previous range of temperatures. The minimum allowed value is 2. Higher is the number, better is the quality of results, but higher is the computation time. The default number of computations is 3. It is a good compromise

1. Flow chart of the thermal solving mode test with iterative process.

   Table 3. Definition of the temperature list of magnets (see the list in the previous section) and working point inputs

   1	Electromagnetic solving (Finite Element computation) for a list of magnet temperatures between the min. and max. values given in the setting.
   2	Loss terms are interpolation regarding to the magnet temperature given by the thermal computation in the next step.
   3	Based on the speed and losses computed in the electromagnetic working point for the first magnet temperature, the thermal characterization is performed to define the temperature distribution inside the machine. The materials physical properties corresponding to the machine active components are then updated.
   4	Iterative process = loop between the interpolation of losses and the thermal computations until the convergence on the temperature is reached (See below where the convergency criteria can be adjusted, if needed).
   5	A last electromagnetic solving (Finite Element computation) with final set of temperatures is performed once the convergence on the temperature is reached (1). Electromagnetic performance and chart of temperatures are computed and displayed. The test outputs are illustrated in the sections dedicated to the considered tests.

   Note: (1) The temperatures which are considered for computing the final machine performance (step 4 in the previous flow chart) can be read in the table dealing with “Winding and Magnet characteristics” of the test configuration at the beginning of result report. The temperatures are also displayed in the temperature chart and table after the electromagnetic results.
2. Thermal characterization in steady state – Flow chart

   This section illustrates how the thermal characterization is performed to define the temperature distribution inside the machine from a set of losses and a working point speed

   This process corresponds to what is performed to make the thermal characterization of the machine in the test Characterization / Thermal / Steady state

   It corresponds to the second step of the previous general flow chart.

   
   

   

   
   

   The inputs of the internal process are the parameters of: Geometry, Winding, Internal cooling, External cooling, Materials, Test settings and inputs.

   Note: A 2D Finite Element model is solved to identify a thermal network which corresponds accurately to any kind of rotor or stator parts, including user parts.

   Then, the resulting network is extended with analytical computations to consider the 3D effect of the geometry.

   The solving allows to get and to display the whole chart of temperatures of the machines.

3. Iterative process – Adjustment of the convergence criteria

   Table 4. Iterative process – Adjustment of the convergence criteria

   1	Selection of the thermal solving Iterative process in thermal settings
   2	Convergence criteria can be adjusted for reaching steady state behavior. The iterative solving is stopped once the temperature variation in the machine between two iterations is lower than the convergence criteria. A percentage close to zero gives more accurate results but takes a higher computation time. A high percentage can make the convergence shorter but decreases the accuracy of the results.