This section describes the development of an automated design and analysis tool for IMs and PMSMs by the authors. In a first step, the modeling process that serves as the basis for the tool is explained. Then, a short comparison of the simulation of IMs and PMSMs is made to investigate the potential synergy effects in the calculations. Afterwards, the program structure and the user interface of the tool are presented.
3.2. Comparison of the Simulation Process of PMSM and IM
The aim of the comparison of the simulation process of PMSM and IM is to find synergies between the two machine types and implement them in a holistic model. For this purpose, the individual calculation steps of the design are briefly compared with each other.
3.2.1. Design
The calculation of the main dimensions is similar for both machine types. The main difference lies in the choice of approximated values, which can be found in the specified literature for specific machines [
17,
22].
In the stator design, there are no significant differences that directly affect the design process, because both machine types require a winding to generate the necessary rotating field. Therefore, the possibilities for the stator design optimization [
30,
32] can be implemented identically for both IM and PMSM. The only difference must however be taken into account in the selection of a final winding option, depending on the respective requirements. For example, in the case of an IM, a lower scattering coefficient for the harmonics due to parasitic effects is to be rated higher than in the case of a PMSM, where the harmonics are to be assessed less critically ([
17], pp. 118–120).
The rotor design is fundamentally different for the two machine types. In an IM, a winding is accommodated in the rotor, which in the squirrel-cage machine consists of rods and short-circuit rings. This has to be dimensioned accordingly. In addition, the slot geometry and the magnetic circuit must be designed analogously to the stator design. With a PMSM, the PMs must be dimensioned in the rotor design and arranged accordingly in the rotor. Numerical methods are also required for the precise design and calculation of IPMSMs.
There are also differences between IMs and PMSMs with regard to the recalculation. With PMSMs, the focus is on the determination of the longitudinal and transverse axis inductances and stator resistance. In contrast, the total inductances of the stator and rotor, as well as the stator and rotor resistances are more relevant for the analysis in the IM. Additionally, the recalculation of the magnetic circuit is required to correct the geometry and determine the magnetizing current.
3.2.2. Analysis
In the case of torque control for the motor and generator model, the torque control procedures are combined to form a minimum current torque control for both machine types. The difference lies in the different system equations of the machines. In the IM, for example, currents occur due to the winding in the rotor. Additionally, the rotor rotates asynchronously to the stator rotating field. Accordingly, the optimization problem used to calculate the currents must be solved with the respective system equations depending on the type of machine.
With regard to the loss calculation, the underlying calculations between the two machine types are similar. However, the different characteristics of the loss components must be taken into account.
In summary, there are direct synergy effects between the two machine types with regard to the stator design. The stator design can therefore be implemented similarly for both machine types. For the determination of the main dimensions, the torque control and the calculation of the losses, overlaps in the calculations can be used with regard to some adjustments for the respective machine type. However, the rotor design and the recalculation are rather different, whereas no synergy effects can be exploited there.
3.3. Model Structure and Functions
Both model components contain two main functions for the calculation process of the respective machine type. The main functions are designed in such a way that the overlaps identified in the previous section can be used in the calculation steps in the form of a common code basis and only minor adjustments are necessary for the respective machine type. In addition, the procedures for the three different stator design options are combined into a uniform model structure and integrated into the two main functions. Both the design and the analysis component have separate graphical user interfaces (GUI) that guide the user through the design and analysis process and will be presented in
Section 3.4 and
Section 3.5.
An overview of the model structure and the various functions is shown in
Figure 4. The functions that are only required for the internal processes of the graphical user interface and data handling are not listed. The description and classification of the individual functions is carried out step by step in the following sections.
In addition to the two main components, two libraries are created that can be used by the design and analysis model component. Among other things, the reference values taken from the literature are stored in the parameters library according to the respective machine type. The characteristic values of the materials are stored in the materials library. For the electrical sheets, for example, material parameters such as the B-H and iron loss curves of the manufacturer Vacuumschmelze are integrated [
33].
Further components of the implemented model are the data storage system and two interfaces. With the data storage system, the machine design can be saved at any time and continued later. It is also possible to call up an already designed machine and recalculate it with desired alterations. Another advantage is the possibility to easily save and archive the results. The interfaces are an export of geometry data to the CAD data format dxf (fct saveDXF) and an export of the design results to an Excel file (fct export_table, fct saveExcel). The CAD data can then, for example, be imported into an FEM program and used for further calculations. This is particularly relevant for the IPMSM, in order for it to be able to calculate the transverse and longitudinal axis inductances accurately.
The model is initialized by calling the script main and the graphical user interface opens in order to guide the user through the design process.
3.4. Graphical User Interface for Design
In the input tab of the design user interface, the user specifies the input parameters for the desired machine design. The user interface of the input is shown in
Figure 5.
In a first step, the user selects the desired machine type. Depending on the machine type selected, the corresponding input fields required for the design are then activated in the user interface and the reference values and available options are loaded from the libraries. A dynamic access to the libraries makes it easy to adapt and expand the options and reference values, without having to make changes to the user interface.
Once the machine type has been selected, the design parameters must be entered and the machine options selected. By selecting the material-specific options, the respective material properties are loaded from the materials library (fct loadMaterial). The Plot B-H curve button and the Plot iron loss curve button can then be used to display the material properties of the electrical sheet (fct plotBH, fct plotlosses).
If required, the approximated values can be adapted to the desired requirements. The user is supported by the user interface by comparing the inputs with the validity range for the respective reference value stored in the parameter library. If the value entered exceeds or falls below the limits for minimum or maximum allowable values, the background color of the input field changes from green to red (
Figure 5). The initially entered reference values represent an average standard value for each parameter, so that a first valid dimensioning is possible without further input.
With the start design button, the main function fct Design_xy of the selected machine type is initiated and the design process begins. The calculation steps described in the previous subsections are then carried out for the respective machine type.
In order to visualize the different parameter combinations, they are illustrated in the form of a radar chart, in which the criteria are listed in a comparative manner. For illustration purposes, the winding matrix of a single-layer integral-slot winding (using N
1 = 36,
p = 2, m = 3) is shown in Equation (4). The first line contains the groove numbers of strand A, the second line the groove numbers of strand C and the third line the groove numbers of strand B.
The winding matrix can be created from the zone plan by simply assigning the groove numbers to the strands. The function fct plotWindinglayout then visualizes the winding layout from the winding matrix. In the derivation from the Tingley plan, the columns of a strand are combined to form the winding matrix of a layer. If the winding step y
1 is known, the winding matrix of a second layer can be generated. The Tingley plan is created from the winding parameters and described in the following using Meyer ([
22], pp. 115–116). In the first step, a matrix is formed with
(q1,n N1)/(2 p) columns and
2 p rows. Then the slot number 1 is entered in the first field of the matrix. Then,
q1,n−1 fields are left empty and the next number is entered. This is repeated until the groove number
N1 is reached. Using the Tingley plan, it is then possible to assign the coil sides to the grooves and the strands and to derive the winding matrix from it.
Table 1 shows the Tingley plan for a single-layer integral-slot winding derived from Equation (4).
Once the design process is completed, the results tab of the design interface displays selected results and the geometry of the machine.
Figure 6 shows the interface of exemplary design results.
In addition to the most relevant main dimensions, such as stator inner diameter or ideal length, some machine characteristics are also displayed. When visualizing the geometry, the user can choose between the overall view of the machine, the geometry of the stator slot and, if available, the geometry of the rotor slot (fct plotMachine, fct plotSlot1, fct plotSlot2). For a PMSM, further functions are activated. For example, the values of the longitudinal and transverse axis inductance and the interlinked flow of the PM can be adjusted here. This makes it possible to feed the data obtained from a FEM simulation back into the simulation tool and to improve the analytical estimation of the design process. A further function is the adaptation of the geometry of the magnets, as well as the magnet arrangement (fct Update_x_y_Coordinates). It should be noted that values for the longitudinal and transverse axis inductance and the interlinked flux of the PM from an external calculation must always be specified for the adapted geometry. These values can be determined with the exported CAD data of the adapted geometry, for example when using an FEM program. After selecting a parameter combination, the design can be continued.
In addition, the winding design offers two possible options for adjustments—the optimization option and the manual option. The user interface for the optimization option is shown in
Figure 7. Here, different predetermined winding types can be selected using literature value. The winding options can then be compared in regard to specific machine parameters and their implications. For illustration purposes, a radar chart for a single-layer integral-slot winding with N
1 = 36, q
1 = 3, W
1,Sp,rel =1 and a
1 =1 was chosen in regard to the outer stator diameter D
1a, the air-gap inductance B
m, the scattering coefficient σ
1, the winding factor of the fundamental wave X
1, the absolute error err
1 and the mass of the conductor material m
1.
Similar to the optimization option, the manual option requires an intermediate step in the design process. In this step, the user can select one of the possible winding combinations or enter his own winding layout in the form of a zone plan. The winding layout is then visualized using a generic machine.
Figure 8 shows this in the user interface. The respective denotations in this step are based on [
30].
When entering the winding layout, a predefined notation must be adhered to. Positive coil sides are marked with an upper case letter and negative coil sides with a lower case letter. The letters A, B and C must be selected for the three strands. The next slot must be delimited with the separator “|”. The input is compatible with the online winding calculation tool
Emetor [
34], so that a simple transfer is possible. In the case of manual input, the winding step shortening or lengthening of the winding must also be specified so that the winding factors can be calculated. The basis for the presentation of the winding layout is the winding matrix M
winding, which can be derived from the zone plan. In case a zone plan is not available, the winding is first distributed with the Tingley plan and then the winding matrix is derived from it (fct TingleyAlg). In the winding matrix, one row contains all slots of one strand and one layer. An element of a row represents the slot number in which a certain side of the coil of the strand is located. The sign of the slot number determines whether the respective coil side is positive or negative. For two-layer windings, one winding matrix is created each for the upper and lower layer.
The conventional stator design method, as well as the optimization methods for the winding design of electric machines used in literature have numerous limitations because they are based on expert knowledge or literature values, limit input parameters or always focus on a specific optimization goal. These disadvantages are eliminated with the novel method for stator design by an iterative procedure improving the conventional calculation steps and the automatic slot generation, according to Meyer [
21]. The aim of this optimized procedure is to enable an automated calculation of all physically possible machine parameter combinations in each step of the stator design process, which correspond to the given application-specific input requirements. Once all parameter combinations have been calculated, they are compared using defined criteria and the optimum winding is selected. This optimized stator design procedure is integrated into the model for the design and analysis of permanent-magnet synchronous and induction machines. With the continue to analysis button, the calculation process is continued and the design variables transferred to analysis.
3.5. Graphical User Interface for Analysis
In the user interface analysis, the user can define the input parameters for the analysis on the one hand and on the other hand the calculated efficiency diagrams are displayed. The options include the selection of the losses and settings for the calculation of the efficiency diagrams, as shown in
Table 2.
In the user interface, the user can select the loss components to be taken into account when calculating the efficiency. If required, the iron loss model can be adapted for the calculation of iron losses. For the options concerning the calculation, the generator mode can be activated and the maximum rotational velocity of the efficiency diagram can be defined. The two input parameters resolution rotational velocity and resolution torque define the distance between two operating points in the diagram. A higher resolution is associated with a higher number of calculations, which immediately increases the computation time.
The two control parameters define the current and voltage limits. Through variation, the machine can be operated with different limits, for example to test a possible overload operation. It is pointed out by the authors that the model does not consider thermal investigations of the calculated machine design and therefore the values should be adjusted with caution. The entered standard values correspond to the nominal values of the machine.
The start analysis button starts the calculation of the characteristic diagrams. For both machine types, the full load characteristics are calculated with the function fct Optimization_M, depending on the machine type. The function fct nonlcon_Optimization_M contains the machine dependent constraints of the optimization problem. The currents for the operating points are then calculated by solving the optimization problem implemented in the functions fct Optimization_i and fct nonlcon_Optimization_i. Both optimization problems are suitable for the calculation of both the motor and the generator operating ranges by simple adaptation of the operating limits.
When calculating the currents, the optimization is accelerated by using the currents determined from the previous steps as start values. After the currents have been calculated, the remaining values and the losses are determined and then displayed in the form of characteristic diagrams in the user interface (fct plotDiagram). A selection menu allows the user to display the desired diagram. In addition, the respective characteristic diagram can be saved using the save plot button in the results folder (fct plot results).