Motor parameter searching system and motor parameter searching method

Abstract

The invention provides a motor parameter searching system and a motor parameter searching method. The motor parameter searching system comprises a processing device and an input device. The processing device may execute a motor design parameter search engine to iteratively execute a parameter search operation according to a plurality of design parameters, a plurality of optimization objectives, a plurality of restrictions, and a plurality of historical recommended parameter combinations to generate a plurality of design parameter combinations. The plurality of design parameter combinations are sequentially input to the simulation software, so that the simulation software generates a plurality of simulation results. The processing device can search a plurality of historical simulation results in a database according to a plurality of optimization objectives to generate a recommended parameter combination. The motor parameter searching system and the motor parameter searching method can effectively generate the motor design parameters.

Description

Technical Field

[0001] The present invention relates to a data operation technology, and more particularly to a motor parameter search system and a motor parameter search method. Background Art

[0002] In the field of motor design, since motor design involves a large number of physical parameters, designers have to adjust back and forth between different physical fields, consuming a lot of manpower and time. Moreover, there is currently no unified and general motor design theory, making it difficult to conduct various or even new topology designs. More importantly, traditional motor design cannot optimize multiple targets simultaneously. Summary of the Invention

[0003] The present invention provides a motor parameter search system and a motor parameter search method, which can effectively generate motor design parameters.

[0004] The motor parameter search system of the present invention includes a processing device and an input device. The processing device executes a motor design parameter search engine, an interactive interface module, a data access module, simulation software, and an optimal combination recommendation module. The input device is coupled to the processing device and receives a plurality of design parameters, a plurality of optimization targets, and a plurality of constraint conditions. The motor design parameter search engine iteratively performs a parameter search operation according to the plurality of design parameters, the plurality of optimization targets, the plurality of constraint conditions, and a plurality of historical recommended parameter combinations to generate a plurality of design parameter combinations. The interactive interface module sequentially inputs the plurality of design parameter combinations into the simulation software to enable the simulation software to generate a plurality of simulation results. The data access module stores the plurality of design parameter combinations and the plurality of simulation results in a database. The optimal combination recommendation module searches the plurality of historical simulation results in the database according to the plurality of optimization targets to generate a recommended parameter combination.

[0005] The motor parameter search method of the present invention includes the following steps: receiving a plurality of design parameters, a plurality of optimization targets, and a plurality of constraint conditions; iteratively performing a parameter search operation according to the plurality of design parameters, the plurality of optimization targets, the plurality of constraint conditions, and a plurality of historical recommended parameter combinations to generate a plurality of design parameter combinations and sequentially inputting them into the simulation software to enable the simulation software to generate a plurality of simulation results; storing the plurality of design parameter combinations and the plurality of simulation results in a database; and searching the plurality of historical simulation results in the database according to the plurality of optimization targets to generate a recommended parameter combination.

[0006] Based on the above, the motor parameter search system and the motor parameter search method of the present invention can automatically generate a plurality of design parameter combinations and automatically input them into the simulation software to generate a plurality of simulation results, and can generate a recommended parameter combination according to the plurality of historical simulation results.

[0007] To make the above features and advantages of the present invention more obvious and understandable, specific embodiments are given below and described in detail in conjunction with the accompanying drawings as follows. Description of the Drawings

[0008] Figure 1 It is a schematic diagram of a motor parameter search system according to an embodiment of the present invention;

[0009] Figure 2 It is a flowchart of a motor parameter search method according to an embodiment of the present invention;

[0010] Figure 3 It is a schematic diagram of multiple modules of a motor parameter search system according to an embodiment of the present invention;

[0011] Figure 4 It is a clustering schematic diagram according to an embodiment of the present invention;

[0012] Figure 5 It is a probability distribution diagram according to an embodiment of the present invention;

[0013] Figure 6 It is a flowchart of a two-stage iterative search operation according to an embodiment of the present invention.

[0014] Description of the Reference Numerals

[0015] 100: Motor parameter search system;

[0016] 110: Processing device;

[0017] 120: Input device;

[0018] 130: Storage device;

[0019] 310: User interface;

[0020] 320: Motor design parameter search engine;

[0021] 321: Sorting module;

[0022] 322: Parameter selection module;

[0023] 323: Parameter range setting module;

[0024] 324: Sampling module;

[0025] 325: Clustering module;

[0026] 326: Parameter combination recommendation module;

[0027] 330: Interaction interface module;

[0028] 340: Simulation software;

[0029] 350: Data access module;

[0030] 360: Database;

[0031] 370: Optimal combination recommendation module;

[0032] 401: First reference point;

[0033] 402: Second reference point;

[0034] 410: First group of parameter combinations;

[0035] 420: Second group of parameter combinations;

[0036] 403: Reference point;

[0037] 501, 502: Probability distributions;

[0038] S210~S240, S610~S660: Steps. Detailed implementation

[0039] Reference will now be made in detail to exemplary embodiments of the present invention, examples of which are illustrated in the accompanying drawings. Wherever possible, the same reference numerals are used in the drawings and the description to refer to the same or like parts.

[0040] Figure 1 is a schematic diagram of a motor parameter search system according to an embodiment of the present invention. Refer to Figure 1 , the motor parameter search system 100 includes a processing device 110, an input device 120, and a storage device 130. The processing device 110 is coupled to the input device 120 and the storage device 130. In this embodiment, a user can input a plurality of design parameters through the input device 120, where the plurality of design parameters can be, for example, a plurality of motor design parameters of an electric motor of an electric vehicle, but the present invention is not limited thereto. The processing device 110 can automatically generate a plurality of design parameter combinations according to the plurality of design parameters and input them into simulation software to automatically generate a plurality of simulation results, where the simulation software can be, for example, simulation design software for an electric motor of an electric vehicle, but the present invention is also not limited thereto. In one embodiment, the motor parameter search system 100 can also be applied to the design of various motor devices, and the plurality of design parameters can also be, for example, related mechanism parameters, electricity-related parameters, magnetism-related parameters, and heat-related parameters, etc.

[0041] In this embodiment, the motor parameter search system 100 can be, for example, a personal computer (PC), a notebook computer, a tablet computer, or other related devices with computing capabilities, but the present invention is not limited thereto. In one embodiment, the motor parameter search system 100 can also be implemented in the form of a cloud server.

[0042] In this embodiment, the processing device 110 can include, for example, a central processing unit (CPU), or other programmable general-purpose or special-purpose microprocessors, image processing units (IPUs), graphics processing units (GPUs), digital signal processors (DSPs), application-specific integrated circuits (ASICs), programmable logic devices (PLDs), other similar computing circuits, or combinations of these circuits.

[0043] In this embodiment, the input device 120 can include, for example, a mouse and a keyboard, but the present invention is not limited thereto. In one embodiment, the input device 120 can be an independent touch panel or integrated with the display panel.

[0044] In this embodiment, the storage device 130 can include non-volatile memories such as read-only memory (ROM), erasable programmable read-only memory (EPROM), volatile memories such as random access memory (RAM), and storage devices such as hard disk drives and semiconductor memories, and can be used to store various parameters, modules, simulation software, and other data mentioned in the present invention.

[0045] Figure 2 is a flowchart of a motor parameter search method according to an embodiment of the present invention. Refer to Figure 1 and Figure 2, the motor parameter search system 100 can execute the following steps S210 to S240. In step S210, the input device 120 can receive a plurality of design parameters, a plurality of optimization objectives, and a plurality of constraint conditions. In step S220, the processing device 110 can iteratively execute a parameter search operation according to the plurality of design parameters, the plurality of optimization objectives, the plurality of constraint conditions, and a plurality of historical recommended parameter combinations to generate a plurality of design parameter combinations and sequentially input them into the simulation software, so that the simulation software generates a plurality of simulation parameter results. In step S230, the processing device 110 can store the plurality of design parameter combinations and the plurality of simulation parameter results in a database, where the database can be built in the storage device 130. In step S240, the processing device 110 can search the plurality of historical simulation results in the database according to the plurality of optimization objectives to generate a recommended parameter combination.

[0046] Specifically, referring to Figure 3 , Figure 3 is a schematic diagram of multiple modules of the motor parameter search system according to an embodiment of the present invention. In this embodiment, the storage device 130 can store, for example, relevant programs and software of a user interface 310, a motor design parameter search engine 320, an interaction interface module 330, a simulation software 340, a data access module 350, a database 360, and an optimal combination recommendation module 370. In this embodiment, the user interface 310 can display relevant operation screens through a display device, for example, and the user can input a plurality of design parameters, a plurality of optimization objectives, and a plurality of constraint conditions according to the relevant operation screens by operating the input device 120.

[0047] For example, the plurality of design parameters can be a plurality of motor design parameters. The plurality of motor design parameters can include stator core length, number of turns, wire diameter size, fin height, rotational speed, current, etc. The plurality of optimization objectives can include cost reduction (i.e., design parameters corresponding to the stator core length and fin height), power greater than 6000 watts (i.e., design parameters corresponding to the rated power), torque increase (i.e., design parameters corresponding to the maximum torque), and efficiency greater than 92% (i.e., design parameters corresponding to the system efficiency), etc. The plurality of constraint conditions can include back electromotive force lower than 5, magnet temperature lower than 150 degrees, torque ripple less than 0.5, and slot fill factor lower than 85%, etc.

[0048] In this embodiment, the processing device 110 can execute the motor design parameter search engine 320, so that the motor design parameter search engine 320 can iteratively execute parameter search operations according to multiple design parameters, multiple optimization objectives, multiple constraint conditions, and multiple historical recommended parameter combinations to generate multiple design parameter combinations, where the multiple historical recommended parameter combinations can be stored in the database 360. Then, the interaction interface module 330 can sequentially input these design parameter combinations into the simulation software 340, so that the simulation software 340 generates multiple simulation results. The data access module 350 can store these design parameter combinations and these simulation results in the database 360. The optimal combination recommendation module 370 can search the multiple historical simulation results in the database 360 according to these optimization objectives to generate a recommended parameter combination. In this embodiment, the motor design parameter search engine 320 can further include a sorting module 321, a parameter selection module 322, a parameter range setting module 323, a sampling module 324, a clustering module 325, and a parameter combination recommendation module 326.

[0049] In this embodiment, the sorting module 321 can sort these optimization objectives. The sorting module 321 can determine the optimization stage stratification and the optimization objectives used in different optimization stages according to the importance of these optimization objectives input by the user and in accordance with the sorting result. For this, the sorting module 321 can use the analytical model method or the expert method to sort the importance of these optimization objectives. The sorting module 321 can determine multiple first-stage optimization objectives and second-stage optimization objectives according to the sorting result of the multiple optimization objectives, so that the motor design parameter search engine 320 can iteratively execute the first-stage and second-stage parameter search operations according to these first-stage optimization objectives and second-stage optimization objectives respectively.

[0050] For example, the multiple optimization objectives input by the user can include, for example, cost, efficiency, or power. If power is the most important objective, then the sorting module 321 can divide the parameter search operation into two stages. The first stage performs a wide and rough parameter search operation with multi-objective optimization, and the second stage performs a single-objective search with the most important objective (such as power) to effectively perform parameter optimization and can strengthen the parameter search for the most important objective.

[0051] In this embodiment, the parameter selection module 322 can select multiple important design parameters from these design parameters according to these optimization objectives. In this regard, since the number of parameters to be searched may be too large, excessive time may be consumed for parameter search due to the large amount of parameter data. Therefore, the motor design parameter search engine 320 can use historical reference points to train a regression model and perform parameter importance calculation. The motor design parameter search engine 320 can analyze historical parameter combinations to select important design parameters that affect the optimization objectives for parameter search (for example, select the design parameters with the top few importance levels for subsequent parameter search), thereby improving the search efficiency.

[0052] In this embodiment, the parameter range setting module 323 can set multiple search ranges for these important design parameters. The parameter range setting module 323 can be determined by analyzing historical parameter search data, or can also be determined according to manual settings by the user. In this embodiment, the sampling module 324 can sample multiple corresponding historical recommended parameter combinations according to these optimization objectives and these search ranges for use by subsequent modules.

[0053] In this embodiment, the clustering module 325 can cluster these historical recommended parameter combinations to generate a first group of parameter combinations and a second group of parameter combinations. The clustering module 325 can cluster these historical recommended parameter combinations according to the multiple constraint conditions and search objectives input by the user. With reference to Figure 4 , Figure 4 is a schematic diagram of clustering according to an embodiment of the present invention. Taking the distribution results of multiple first reference points 401 of the first group of parameter combinations 410 and multiple second reference points 402 of the second group of parameter combinations 420 on the distribution plane of the first condition (for example, the first design parameter) and the second condition (for example, the second design parameter) as an example, the reference point 403 of the constraint condition can be the intersection point of the requirements of the first condition and the second condition. In this regard, as Figure 4 shown, the distance between multiple first reference points 401 of the first group of parameter combinations 410 and the reference point 403 of the constraint condition is less than the distance between multiple second reference points 402 of the second group of parameter combinations 420 and the reference point 403 of the constraint condition. And, the target values of these first reference points 401 are greater than the target values of these second reference points 402. These first reference points 401 respectively correspond to good and poor parameter combination samples, and each group has multiple parameter combination samples. The quality of the parameter combination samples can be determined, for example, by the distance from the reference point 403 of the constraint condition.

[0054] In this embodiment, the parameter combination recommendation module 326 generates a plurality of design parameter combinations based on the plurality of optimization objectives, the first group of parameter combinations 410, and the second group of parameter combinations 420. To this end, the parameter combination recommendation module 326 can respectively establish distribution surrogate models for the first group of parameter combinations and the second group of parameter combinations, and calculate the acquisition function to generate the next design parameter combination for iterative calculation. With reference Figure 5 , Figure 5 is the probability distribution diagram of an embodiment of the present invention. The parameter combination recommendation module 326 can use the Parzen window density estimation method to calculate Figure 4 the probability distributions of a plurality of first reference points 401 of the first group of parameter combinations 410 and a plurality of second reference points 402 of the second group of parameter combinations 420 to generate probability distributions 501 and 502 as shown in Figure 5 . Figure 4 The plurality of first reference points 401 of the first group of parameter combinations 410 in Figure 4 can correspond to the probability distribution 501, and the plurality of second reference points 402 of the second group of parameter combinations 420 in Figure 4 can correspond to the probability distribution 502. The parameter combination recommendation module 326 can use the formula of the Expected Improvement method as the acquisition function so that the expected sampling points have a higher probability in the probability distribution 501 and a lower probability in the probability distribution 502.

[0055] In this embodiment, the parameter combination recommendation module 326 can output the design parameter combinations generated each time in the iterative search to the interaction interface module 330, so that the interaction interface module 330 can respectively input the design parameter combinations generated each time in the iterative search into the simulation software 340. The simulation software 340 can automatically perform design simulations to generate a plurality of simulation results and return them to the interaction interface module 330. Then, the interaction interface module 330 can store these simulation results and the corresponding design parameter combinations in the database 360 through the data access module 350. And when the iterative search ends, the data access module 350 can provide all the historical simulation results and the corresponding design parameter combinations to the optimal combination recommendation module 370, so that the optimal combination recommendation module 370 can select one of the plurality of historical recommended parameter combinations that match the design requirements and optimization objectives as the (optimal) recommended parameter combination. Therefore, the optimal combination recommendation module 370 can output the recommended parameter combination to the user so that the user can perform motor design and manufacturing according to the recommended parameter combination and effectively obtain a motor device that meets the requirements.

[0056] For example, in an exemplary embodiment of the present invention, a user can input 7 design parameters as shown in Table 1 below according to a relevant operation screen, and include the initial values and search ranges of these 7 design parameters. These 7 design parameters can include stack length (Stator Lam Length / Magnet Length / Rotor Lam Length), number of turns (number of coils), parallel paths, wire diameter, fin extension, rotational speed, and current. Moreover, multiple optimization objectives input by the user can, for example, include cost (the smaller the better), efficiency (the larger the better), and power (the larger the better), where the constraints can include a recommended power greater than or equal to 6000 watts (W) and a recommended torque greater than or equal to 20 (Nm).

[0057] Design Parameters Initial Value Search Range Laminated Length 50 (mm) 40～60 Number of Turns (Coil Turns) 15 (turns) 10～30 Parallel Branches 1 2、4、7 Wire Diameter 25 (mm) 23～48 Heat Sink Length 3 (mm) 1～10 Rotational Speed 5500 (rpm) 5000～6000 Current 110 (amperes) 90～300

[0058] Table 1

[0059] In the multi-objective search of the first stage, the sorting module 321 can perform sorting according to the expert method. For example, the sorting module 321 can make judgments based on power, system efficiency, and cost.

[0060] In this example, the parameter selection module 322 can sort the importance of design parameters according to the optimization objectives (such as cost, power, and system efficiency). The parameter selection module 322 can use historical reference points to train a regression model to perform parameter importance calculations and select design parameters according to the parameter importance. In this regard, since the search has not started, the design parameters remain unchanged and are maintained at 7 as shown in Table 1. The parameter range setting module 323 can determine the design parameters to be used and set the search ranges of these design parameters (as shown in Table 1).

[0061] Next, the parameter combination recommendation module 326 can read historical reference points (i.e., historical recommended combinations) from the database 360. If the number of historical reference points does not meet the minimum required number, the sampling module 324 can generate recommended design parameter combinations. For example, if at least 10 historical reference points are required, in the beginning of the search, the first 10 recommended combinations can be recommended by the sampling module 324. In addition, if there are no historical reference points, the sampling module 324 randomly generates a combination within the design parameter range. Moreover, the clustering module 326 can cluster the historical reference points. In this regard, if the historical reference points have met the minimum required number, the clustering module 326 will cluster the historical reference points. For example, the parameter combination recommendation module 326 can obtain the result of clustering the 10 historical reference points by the clustering module 325. The 10 historical reference points can be divided into two clusters, such as "good" and "bad" for example. The parameter combination recommendation module 326 can generate a new recommended parameter combination based on the clustered historical reference points. Next, in this example, the interaction interface module 330 can transmit the new recommended parameter combination generated by the parameter combination recommendation module 326 to the simulation software 340 and obtain the simulation result generated by the simulation software 340. The system can repeatedly execute the above operations until the maximum number of searches is reached. Finally, the best combination recommendation module 370 can obtain all historical reference points (including all the recommended parameter combinations newly added by the system repeatedly executing the above operations) from the data access module and select and output the parameter optimization result of the best combination as shown in Table 2 below, where the cost, power, system performance, and torque of the optimization objective can all be effectively improved.

[0062] Optimization Objective Design Parameters Initial Value Optimization Results of the First Stage Laminated Length 50 (mm) 60 (mm) Number of Turns (Coil Turns) 15 (turns) 10 (turns) Parallel Branches 1 4 Wire Diameter 25 (mm) 26 (mm) Heat Sink Length 3 (mm) 4.8 (mm) Rotational Speed 5500 (rpm) 5700 (rpm) Current 110 (amperes) 199 (amperes) Cost 100 95 Power 5952.50 (watts) 5832.14 (watts) System Efficiency 92.45％ 92.994％ Torque 12.337 (Nm) 14.102 (Nm)

[0063] Table 2

[0064] In the single-objective search of the second stage, the parameter selection module 322 can rank the importance of design parameters according to a single optimization objective (such as power). The parameter selection module 322 can use historical reference points to train a regression model to perform parameter importance calculation so as to find out the parameters that have an impact on power adjustment. In this regard, as shown in Table 3 below, for example, the top 5 important parameters can be selected according to the importance degree. In addition, the design parameters not selected can adopt the recommended (optimized) values generated in the first stage. The parameter range setting module 323 can determine the design parameters to be used and set the search ranges of these design parameters (as shown in Table 3).

[0065] Design Parameters Initial Value Search Range Laminated Length 60 (mm) 40～60 Number of Turns (Coil Turns) 10 (turns) 10～30 Wire Diameter 26 (mm) 23～48 Rotational Speed 5700 (rpm) 5000～6000 Current 199 (amperes) 90～300

[0066] Table 3

[0067] Next, the parameter combination recommendation module 326 can read the historical reference points (i.e., historical recommended combinations) generated in the first stage from the database 360. And since the single-objective search in the second stage can use the historical reference points in the first stage, the sampling module 324 does not need to be executed. Also, the clustering module 326 can cluster at least a part of the historical reference points in the first stage. The parameter combination recommendation module 326 can generate a new recommended parameter combination based on the clustered historical reference points. Next, in this example, the interaction interface module 330 can transmit the new recommended parameter combination generated by the parameter combination recommendation module 326 to the simulation software 340 and obtain the simulation results generated by the simulation software 340. The interaction interface module 330 can store the simulation results in the database 360 (i.e., as new historical reference points). The system can repeatedly execute the above operations until the maximum search number is reached. Finally, the optimal combination recommendation module 370 can obtain all the historical reference points from the data access module (including all the recommended parameter combinations newly added by the system repeatedly executing the above operations). The optimal combination recommendation module 370 can filter out the simulation results with a power less than 6000 watts (W) and a torque less than 20 (Nm) based on the above limiting conditions. Therefore, the optimal combination recommendation module 370 can select and output the parameter optimization results of the optimal combination as shown in Table 4 below, where the power of the main optimization target (single objective) can be further effectively improved.

[0068]

[0069] Table 4

[0070] Figure 6 is a flowchart of the two-stage iterative search operation of an embodiment of the present invention. Refer to Figure 1 , Figure 3 and Figure 6 , the motor parameter search system 100 can perform the following steps S610 to S650. In step S610, the processing device 110 can first iteratively execute the multi-objective parameter search operation. For this, the motor design parameter search engine 320 can first perform parameter search according to multiple optimization objectives to generate multiple recommended parameter combinations, and input them to the simulation software 340 to generate corresponding multiple simulation results. In step S620, the database 360 can store the multiple historical recommended parameter combinations generated by the iteratively executed multi-objective parameter search operation to increase the number of samples required for the parameter search in the next stage. In this embodiment, the processing device 110 can iteratively execute the multi-objective parameter search operation according to a fixed number of iterations.

[0071] In step S630, the processing device 110 can iteratively perform a single-objective parameter search operation. In this regard, the motor design parameter search engine 320 can perform a parameter search based on a single objective (the important optimization objective) to generate a recommended parameter combination, and input it to the simulation software 340 to generate a corresponding simulation result. In step S640, the processing device 110 can add the historical recommended parameter combination to the database 360. In step S650, the processing device 110 can determine whether the currently executed number of parameter searches meets the iteration number. If so, in step S660, the best combination recommendation module 370 can generate a recommended parameter combination. If not, the processing device 110 can perform the single-objective parameter search operation again.

[0072] In this regard, the sample of the historical recommended parameter combination newly added by each search operation can be added to the first group of parameter combinations generated by the clustering module 325, and the sample numbers of the first group and the second group of parameter combinations can be incremented. In this way, the reasonable sample rate of the parameter search can be effectively increased, and the relevant values or parameters corresponding to the important optimization objective (i.e., the second-stage optimization objective described above) can be further optimized.

[0073] In summary, the motor parameter search system and the motor parameter search method of the present invention can perform parameter search operations for multiple optimization objectives and single optimization objectives, and the single optimization objective parameter search operation can be performed by selecting important design parameters, so as to effectively improve the reasonable sample rate of the parameter search. Therefore, the motor parameter search system and the motor parameter search method of the present invention can automatically and effectively generate the best recommended parameter combination for motor manufacturing.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A motor parameter search system, characterized in that: include: A processing device for executing a motor design parameter search engine, an interactive interface module, a data access module, a simulation software, and an optimal combination recommendation module; as well as an input device, coupled to the processing device, and receiving a plurality of design parameters, a plurality of optimization objectives, and a plurality of constraints, The motor design parameter search engine iteratively performs parameter search operations according to the multiple design parameters, the multiple optimization objectives, the multiple constraints, and the multiple historical recommended parameter combinations to generate multiple design parameter combinations, and the interactive interface module sequentially inputs the multiple design parameter combinations into the simulation software so that the simulation software generates multiple simulation results. The data access module stores the plurality of design parameter combinations and the plurality of simulation results in a database, and the best combination recommendation module searches for a plurality of historical simulation results in the database according to the plurality of optimization objectives to generate a recommended parameter combination.

2. The motor parameter search system according to claim 1, characterized in that: The motor design parameter search engine includes a sorting module, and the sorting module sorts the plurality of optimization objectives. The sorting module determines multiple first-stage optimization targets and second-stage optimization targets according to the sorting results of the multiple optimization targets, and the motor design parameter search engine iteratively executes the parameter search operations of the first stage and the second stage according to the multiple first-stage optimization targets and the second-stage optimization targets respectively.

3. The motor parameter search system according to claim 1, characterized in that: The motor design parameter search engine includes a parameter selection module, and the parameter selection module selects a plurality of important design parameters from the plurality of design parameters according to the plurality of optimization objectives. The motor design parameter search engine further includes a parameter range setting module, and the parameter range setting module sets multiple search ranges of the multiple important design parameters. The motor design parameter search engine further includes a sampling module, and the sampling module samples the multiple historically recommended parameter combinations according to the multiple optimization targets and the multiple search ranges.

4. The motor parameter search system according to claim 3, characterized in that: The motor design parameter search engine further includes a grouping module, and the grouping module groups the plurality of historically recommended parameter combinations to generate a first group parameter combination and a second group parameter combination. The motor design parameter search engine further includes a parameter combination recommendation module, and the parameter combination recommendation module generates the plurality of design parameter combinations according to the plurality of optimization objectives, the first group of parameter combinations, and the second group of parameter combinations. The grouping module groups the plurality of historical recommendation parameter combinations according to the plurality of restriction conditions and the search target.

5. The motor parameter search system according to claim 4, characterized in that: The distance between the first reference points of the first group parameter combination and the reference point of the constraint condition in the grouping graph is smaller than the distance between the second reference points of the second group parameter combination and the reference point of the constraint condition in the grouping graph.

6. The motor parameter search system according to claim 4, characterized in that: The parameter combination recommendation module establishes distribution agent models for the first group of parameter combinations and the second group of parameter combinations respectively, and calculates acquisition functions to generate the next design parameter combination for iterative calculation.

7. A motor parameter search method, characterized in that: include: Receiving multiple design parameters, multiple optimization objectives, and multiple constraints; Iteratively performing a parameter search operation according to the plurality of design parameters, the plurality of optimization objectives, the plurality of constraints, and a plurality of historically recommended parameter combinations to generate a plurality of design parameter combinations and sequentially inputting the combinations into a simulation software so that the simulation software generates a plurality of simulation results; Storing the plurality of design parameter combinations and the plurality of simulation results in a database; as well as A plurality of historical simulation results in the database are searched according to the plurality of optimization objectives to generate a recommended parameter combination.

8. The motor parameter search method according to claim 7, characterized in that: The step of iteratively performing the parameter search operation according to the plurality of design parameters, the plurality of optimization objectives, the plurality of constraints and the plurality of historically recommended parameter combinations comprises: Determining a plurality of first-stage optimization objectives and a second-stage optimization objective according to the ranking results of the plurality of optimization objectives; and The parameter search operations of the first stage and the second stage are iteratively performed according to the plurality of first stage optimization objectives and the second stage optimization objectives respectively.

9. The motor parameter search method according to claim 7, characterized in that: The step of iteratively performing the parameter search operation according to the plurality of design parameters, the plurality of optimization objectives, the plurality of constraints and the plurality of historically recommended parameter combinations comprises: selecting a plurality of important design parameters from the plurality of design parameters according to the plurality of optimization objectives; Setting a plurality of search ranges for the plurality of important design parameters; and The plurality of historically recommended parameter combinations are sampled according to the plurality of optimization objectives and the plurality of search ranges.

10. The motor parameter search method according to claim 9, characterized in that: The step of iteratively performing the parameter search operation according to the plurality of design parameters, the plurality of optimization objectives, the plurality of constraints and the plurality of historically recommended parameter combinations further comprises: Grouping the plurality of historically recommended parameter combinations to generate a first group of parameter combinations and a second group of parameter combinations; and generating the plurality of design parameter combinations according to the plurality of optimization objectives, the first group of parameter combinations, and the second group of parameter combinations, The step of grouping the plurality of historical recommendation parameter combinations comprises: The plurality of historical recommendation parameter combinations are grouped according to the plurality of restriction conditions and the search target.

11. The motor parameter search method according to claim 10, characterized in that: The distance between the first reference points of the first group parameter combination and the reference point of the constraint condition in the grouping graph is smaller than the distance between the second reference points of the second group parameter combination and the reference point of the constraint condition in the grouping graph.

12. The motor parameter search method according to claim 10, characterized in that: The step of iteratively performing the parameter search operation according to the plurality of design parameters, the plurality of optimization objectives, the plurality of constraints and the plurality of historically recommended parameter combinations further comprises: Establishing distributed proxy models for the first group parameter combination and the second group parameter combination respectively; and The acquisition function is calculated to generate the next design parameter combination for iterative calculation.

Images