Non-transitory computer-readable storage medium, system, control device

Abstract

To provide a program system controller and electric equipment capable of industrially advantageously evaluating a variable relating to the controller of the electric equipment or the like.SOLUTION: A program for causing a system for outputting a second variable different from a first variable of an electrical device on the basis of the first variable to function as means for evaluating the second variable including one or more variables selected from the group consisting of a voltage, a current, a frequency, and a position of the electrical device when the first variable including one or more variables selected from the group consisting of a voltage, a current, a frequency, and a position of the electrical device is received, the program causing the system to function as means for outputting an evaluation value obtained by predicting the second variable from the first variable of the electrical device.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a program and a system that are useful for controlling electric devices such as motors. [Background technology]

[0002] Non-Patent Document 1 discloses a system that uses finite element analysis to build a motor model for control design. This system analyzes the characteristics of a motor under multiple driving conditions, allowing the construction of a motor control model that can be used for advanced control design.

[0003] However, the system in Non-Patent Document 1 had the problem that the actual motor control state did not necessarily match the system state due to the influence of analytical modeling errors caused by manufacturing tolerances, deviations in material properties, differences in driving conditions, etc.

[0004] Therefore, when constructing a control model for an electric device such as a motor, a system that can evaluate actual operating characteristics in more detail is desired. [Prior art documents] [Non-patent literature]

[0005] [Non-Patent Document 1] JSOL Corporation, "JMAG-RT", [online], [Retrieved July 24, 2024], Internet<URL:https: / / www.jmag-international.com / jp / products / jmag-rt / > Summary of the Invention [Problem to be solved by the invention]

[0006] The present invention aims to provide a program system, a control device, and an electrical device that enable industrially advantageous control design of electrical equipment and the like. [Means for solving the problem]

[0007] As a result of intensive research into achieving the above-mentioned object, the inventors have discovered that a program that causes a system for outputting a second variable, different from a first variable of an electrical device, based on the first variable, to function as a means for evaluating one or more variables selected from the group consisting of the voltage, current, frequency, and position of the electrical device when the first variable includes one or more variables selected from the group consisting of the voltage, current, frequency, and position of the electrical device as the second variable, can implement control design of electrical device in an industrially advantageous manner, and in particular can produce significantly superior control devices for electrical device, and have found that such a program can solve the above-mentioned conventional problems in one fell swoop. Furthermore, after obtaining the above findings, the present inventors conducted further studies and completed the present invention.

[0008] That is, the present invention relates to the following inventions. [1] A program that causes a system for outputting a second variable different from a first variable of an electrical device to function as a means for evaluating a type of variable selected from the group consisting of the voltage, current, frequency, and position of the electrical device as the second variable when the system receives the first variable including one or more variables selected from the group consisting of the voltage, current, frequency, and position of the electrical device. [2] The program according to [1], which learns the relationship between the first variable and the second variable, predicts the second variable from the first variable of the electrical device through the learning, and outputs the predicted evaluation value. [3] The program according to [1], wherein the electrical device is a driving body. [4] The program according to [3], wherein the first variable includes one or more selected from the group consisting of the velocity, angular velocity, acceleration, angular acceleration, force, and torque of the driving body. [5] The program according to [3], wherein the second variable includes one or more selected from the group consisting of the velocity, angular velocity, acceleration, angular acceleration, force, and torque of the driving body. [6] The program according to [1], wherein the first variable includes one or more selected from the group consisting of voltage, frequency, and position. [7] The program according to [1], wherein the second variable includes a current. [8] A system including a program, wherein the program is the program described in [1]. [9] A control device including a program, the program being the program described in [1].

[10] An electrical device including a program, the program being the program described in [1]. [Effects of the Invention]

[0009] The program and system of the present invention can evaluate variables relating to electrical equipment and the like in an industrially advantageous manner. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram showing the configuration of the evaluation system. [Figure 2] FIG. 2 is a perspective view of the motor. [Figure 3] Figure 3 is a diagram of the evaluation system using machine learning. [Figure 4] FIG. 4 is a diagram showing the configuration of an LSTM cell. [Figure 5] Figure 5 shows the configuration of the training data for the machine learning model. [Figure 6] FIG. 6 is a configuration diagram of the control design system. DETAILED DESCRIPTION OF THE INVENTION

[0011] The present invention is characterized in that a system for outputting a second variable, different from a first variable of an electrical device, based on the first variable of the electrical device functions as a means for evaluating one or more variables selected from the group consisting of the voltage, current, frequency, and position of the electrical device as the second variable when the first variable includes one or more variables selected from the group consisting of the voltage, current, frequency, and position of the electrical device.

[0012] The electrical equipment used in the present invention is not particularly limited as long as it does not impede the objectives of the present invention, and may be any known electrical equipment. Examples of such electrical equipment include motors, piezoelectric actuators, high-frequency preheaters, capacitors, magnetic gears, circuit breakers, sensors, electrical heating devices, electromagnetic retarders, electromagnetic relays, transformers / reactors, bus bars, generators, heaters, bearings, magnetic couplings, induction heating devices, induction machines, linear solenoid / linear actuators, and wireless power supply systems. Furthermore, in the present invention, it is preferable that the electrical equipment is a driving body. According to such a preferred embodiment, the electromagnetic operating characteristics of the driving body can be more effectively utilized in the design.

[0013] In addition, in the present invention, it is preferable to learn a relationship between the first variable and the second variable, predict the second variable from the first variable of the electrical device through the learning, and output the predicted evaluation value. According to this preferred embodiment, the second variable can be evaluated from the first variable with higher accuracy. Examples of the learning and prediction means include machine learning techniques such as partial least squares regression, support vector regression, Gaussian process regression, random forest, gradient boosting decision tree, neural network, multilayer perceptron, convolutional neural network, Transformer, Vision Transformer, MLP-Mixer, graph neural network, recurrent neural network (RNN), long short-term memory, gated recurrent unit, large-scale language model, and foundation model.

[0014] In the present invention, it is preferable that the first variable includes one or more selected from the group consisting of velocity, angular velocity, acceleration, angular acceleration, force, and torque of the driver. According to such a preferable embodiment, the mechanical characteristics of the driver can be evaluated with higher accuracy.

[0015] In the present invention, it is preferable that the second variable includes one or more selected from the group consisting of velocity, angular velocity, acceleration, angular acceleration, force, and torque of the driver. According to such a preferable embodiment, the mechanical characteristics of the driver can be evaluated with higher accuracy.

[0016] In the present invention, it is preferable that the first variables include one or more selected from the group consisting of voltage, frequency, and position. According to such a preferable aspect, it is possible to evaluate the electrical characteristics of the electrical device with higher accuracy.

[0017] In the present invention, it is preferable that the second variable includes a current. According to this preferable aspect, the electrical characteristics of the electrical device can be evaluated with higher accuracy.

[0018] The present invention will be specifically described below with reference to the drawings, but the present invention is not limited to these specific examples.

[0019] 1 shows an evaluation system according to a preferred embodiment. Based on a first variable 110 of an electrical device, an evaluation device 100 evaluates and outputs a second variable 120 that is different from the first variable 110. The output second variable 120 may be a variable at a certain time, or may be a variable at multiple times. When the evaluation system 100 receives the first variable 110, which changes over time, it can continuously evaluate and output the second variable 120 in real time.

[0020] The evaluation system 100 may be configured by one or more computers having a processor and a memory. The processor may be a CPU or a GPU. The memory is connected to the processor. The memory may include, for example, a primary storage device and a secondary storage device. The primary storage device may be, for example, a RAM. The secondary storage device may be, for example, a hard disk drive (HDD) or a solid-state drive (SSD). The memory may include a computer program executed by the processor. The processor reads and executes the computer program stored in the memory. The computer program has program code for causing the computer to operate as the evaluation system 100.

[0021] The first variable 110 and the second variable 120 are preferably characteristics of the electrical equipment to be evaluated. The characteristics may be, for example, electromagnetic characteristics or mechanical characteristics. If the electrical equipment is a motor, the characteristics may be motor characteristics. The motor characteristics may include, for example, any one of voltage, current, frequency, and position, or a combination thereof. Examples of the voltage include phase voltage, line voltage, neutral point voltage, induced voltage, primary voltage, and secondary voltage. Examples of the current include phase current, line current, primary current, and secondary current. Examples of the frequency include speed, angular velocity, acceleration, angular acceleration, and rotation speed. Examples of the position include angle. Examples of the mechanical characteristics include thrust force, torque, tangential force, and radial force.

[0022] In the following description, the evaluation system 100 according to the embodiment is used for designing a motor, as an example. Motors are used in a variety of products, such as electric vehicles, aircraft, industrial robots, drones, trains, and home appliances. The motor to be designed by the evaluation system 100 is, as an example, an interior permanent magnet synchronous motor (IPMSM).

[0023] As shown in Fig. 2, interior permanent magnet synchronous motor 10 is a type of motor that uses permanent magnets 15. Interior permanent magnet synchronous motor 10 includes stator 11 and rotor 12 provided inside stator 11. Permanent magnets 15 are embedded in rotor 12. Interior permanent magnet synchronous motor 10 shown in Fig. 2 is, as an example, an 8-pole, 48-slot distributed winding IPMSM.

[0024] The stator 11 is made of electromagnetic steel and has multiple slots 16 formed in the circumferential direction. Windings are present in the slots 16. The rotor 12 is constructed by embedding permanent magnets 15 in a rotor body made of electromagnetic steel. In the rotor 12, regions (holes) where air exists are formed in positions adjacent to the permanent magnets 15, etc. When a three-phase alternating current is applied from an external power source to the windings in the slots 16, a rotating magnetic field is generated, and the torque generated by interaction with the field magnetic flux of the rotor 12 causes the rotor 12 to rotate.

[0025] 3 is a diagram showing the configuration of a machine learning model 300, which is a suitable example of the evaluation system 100. The machine learning model 300 includes an LSTM layer 310 and a fully connected layer 320. Two or more LSTM layers 310 or fully connected layers 320 may be connected in series.

[0026] 4 is a diagram showing the configuration of a cell in the LSTM layer 310. The LSTM cell 400 includes a first sigmoid layer 401 and a first multiplier 402 that form a forget gate, a second sigmoid layer 403 and a second multiplier 404 that form an input gate, a first tanh layer 405 and an adder 406, a third sigmoid layer 407 and a third multiplier 408 that form an output gate, and a second tanh layer 409.

[0027] In Figure 4, Xt is the input vector at the current time t, which is the combined voltage, angular velocity, and angle data at the current time t. Ct is the state vector at the current time t. Ct-1 is the state vector at time t-1. ht is the hidden state vector at the current time t. ht-1 is the hidden state vector at time t-1.

[0028] When inputting information at multiple times t, the LSTM layer 310 is configured from multiple LSTM cells 400. For example, when inputting voltage, angular velocity, and angle data at multiple times t to t-4, the LSTM layer 310 is configured from five LSTM cells 400.

[0029] There are various variations in the configuration of LSTM other than those described above, but the technology disclosed herein does not exclude other LSTM configurations. Furthermore, other models may be adopted as long as they can handle time-series data. For example, a recurrent neural network (RNN) may be adopted.

[0030] 5 is a diagram schematically showing the configuration of training data 500 used when training the machine learning model 300. The training data 500 is composed of input data 501 and supervised data 502. The input data 501 includes voltage, angular velocity, and angle data at multiple times. The supervised data 502 includes current data at multiple times.

[0031] 6 shows a preferred example of a control design system 600 using the evaluation system 100. The control design system 600 is configured with a controller 610 and the evaluation system 100, and exchanges a control input 620 corresponding to the first variable 110 and a control amount 630 corresponding to the second variable 120.

[0032] The present invention is not limited to the above-described embodiment and can be modified in various ways. In the above-described embodiment, the present invention is applied to a motor as one suitable example. However, the present invention can also be applied to other applications. For example, the present invention can be applied to piezoelectric actuators, high-frequency preheaters, capacitors, magnetic gears, circuit breakers, sensors, electrical heating devices, electromagnetic retarders, electromagnetic relays, transformers and reactors, bus bars, generators, heaters, bearings, magnetic couplings, induction heating devices, induction machines, linear solenoids and linear actuators, and wireless power transfer systems. When the present invention is applied to a motor, it has the effect of being able to accurately represent nonlinear characteristic information of the motor under various driving conditions. [Industrial Applicability]

[0033] The system and program of the present invention are particularly useful for control design of motors and the like. [Explanation of symbols]

[0034] 10: Interior permanent magnet synchronous motor 11: Stator 12: Rotor 15: Permanent magnet 16: Slot 100: Evaluation device 110: First variable 120: Second variable 300: Machine learning models 310 :LSTM layer 320 :Fully connected layer 400: LSTM cells 401: First sigmoid layer 402: First multiplier 403: Second sigmoid layer 404: Second multiplier 405: First tanh layer 406: Adder 407: Third sigmoid layer 408: Third multiplier 409: Second tanh layer 500: Teacher data 501: Input data 502: Correct data 600: Control Design System 610: Controller 620: Control input 630: Control amount

Claims

1. A program that causes a system for outputting a second variable different from a first variable of an electrical device based on the first variable to function as a means for evaluating one or more variables selected from the group consisting of the voltage, current, frequency, and position of the electrical device as the second variable when the system receives the first variable including one or more variables selected from the group consisting of the voltage, current, frequency, and position of the electrical device.

2. 2. The program according to claim 1, further comprising: learning a relationship between the first variable and the second variable; predicting the second variable from the first variable of the electrical device through the learning; and outputting the predicted evaluation value.

3. The program according to claim 1 , wherein the electrical device is a driving body.

4. 4. The program according to claim 3, wherein the first variable includes at least one selected from the group consisting of velocity, angular velocity, acceleration, angular acceleration, force, and torque of the driving body.

5. 4. The program according to claim 3, wherein the second variable includes at least one selected from the group consisting of velocity, angular velocity, acceleration, angular acceleration, force, and torque of the driving body.

6. 2. The program according to claim 1, wherein the first variable includes at least one selected from the group consisting of voltage, frequency, and position.

7. 2. The program of claim 1, wherein the second variable comprises a current.

8. A system including a program, said program being the program of claim 1.

9. A control device including a program, wherein the program is the program according to claim 1.

10. 10. An electrical device including a program, said program being the program according to claim 1.

Images