Method and system for providing a simulation model of a rotating electric machine

Abstract

The invention relates to a computer-implemented method for providing a simulation model of a rotating electrical machine, wherein the simulation model is determined by parameter values, in which method input data (S1) are obtained, wherein the input data are acquired with the aid of a machine (2) in a state not connected to an operating voltage and characterize the machine (2), parameter values ​​are derived (S2) from the input data with the aid of a trained function (F), and the derived parameter values ​​are provided (S3).

Description

Technical Field

[0001] The invention relates to a method for connecting a rotating electric machine to a platform, preferably a software platform, in particular a cloud platform, and to a method for determining the operating behavior of a rotating electric machine connected to an operating voltage and, for example, connected to a platform.

[0002] Furthermore, the invention relates to a device for determining an operating behavior of a rotating electric machine connected to an operating voltage, wherein the device comprises a sensor device and a computing device.

[0003] The present invention discloses a computer-implemented method for providing a trained function.

[0004] The present invention also discloses a computer-implemented method for providing a simulation model of a rotating electric machine using the trained function, wherein the simulation model is determined by parameter values.

[0005] Furthermore, the present invention discloses a computer program, which includes corresponding instructions and causes the computer to execute one or / and another of the above-mentioned methods when the corresponding program is executed by the computer.

[0006] Furthermore, the present invention discloses a sensor device having a computer-readable medium having another simulation model based on the aforementioned simulation model.

[0007] Furthermore, the invention discloses a data carrier signal which transmits the aforementioned simulation model and / or the aforementioned computer program. Background Art

[0008] In the context of the ongoing digitalization of industry, the introduction of new assets is an important aspect. Field equipment, machines, software, etc. are usually referred to as assets at the level of facility components.

[0009] When digitizing old industrial plants, a problem often encountered is that the information available in the plant about asynchronous machines, for example about rotating electrical machines, in particular about asynchronous machines such as asynchronous motors and generators, is not sufficient to create a sufficiently accurate simulation model of the machine and, for example, to carry out a start-up or to determine the operating point of the machine.

[0010] Even if the equivalent circuit (simulation model) of the asynchronous motor is known ("green field"), it is usually difficult to deduce its operating behavior. A possible solution to the "green field" problem is proposed in EP 20 173 618.8 of the present applicant, the content of which is hereby incorporated into the present application. For this purpose, complex measurements with the aid of expensive sensors are usually required.

[0011] In many motors, for example from older series, the available information is often even less. The information is often limited to the data that can be found on the type plate (rating plate) ("brown area").

[0012] One possible solution to this problem and determine the operating performance is to use simpler methods (such as linear interpolation). However, simpler methods have significant inaccuracies and only provide information about torque without providing more detailed information, such as efficiency and power factor.

[0013] Another method is known, for example, from Chinese patent application CN 110007232 A, the subject of which relates to a method for predicting the operating efficiency of a squirrel cage asynchronous motor. In this method, a simulation model of the squirrel cage asynchronous motor is established, and simulation data is obtained by inputting harmonic voltage into the squirrel cage synchronous motor simulation model. A BP neural network is then trained based on the simulation data to obtain a BP neural network prediction model. In order to obtain the operating efficiency of the current squirrel cage asynchronous motor, the actual operating data of the current squirrel cage asynchronous motor is then captured and input into the BP neural network prediction model.

[0014] In addition, a simulation device and a simulation method for soft starting of a motor are known from Chinese patent application CN 102368672A.

[0015] Furthermore, European patent application EP 3 322 088 A1 discloses a method for monitoring the operation of a rotating electrical machine, which proposes measuring a first physical variable of a machine stator and a second physical variable of a machine rotor outside a machine housing, wherein at least one state variable of the rotating electrical machine is determined from the first physical variable and the second physical variable. Summary of the invention

[0016] The object of the present invention can therefore be seen as: to provide a simulation model of a rotating electrical machine and, for example, to enable the entry of a rotating electrical machine based on only a small amount of available information and to dispense with complex, expensive measurements, i.e., for example, measurements at rest (e.g. cold resistance measurements) in combination with rotational measurements under no-load (voltage series) and under load (load series).

[0017] According to the invention, this object is achieved by means of a computer-implemented method for providing a simulation model of a rotating electric machine of the type mentioned above, wherein the simulation model is determined by parameter values, namely

[0018] - obtaining input data, wherein the input data can be obtained from the machine when it is not connected to an operating voltage and the input data are characteristic for the machine and include shaft height, number of pole pairs and nameplate specifications of the rotating electrical machine,

[0019] - Derivation of parameter values ​​from the input data by means of a trained function, wherein the trained function (F) is provided according to the above method, wherein the trained function determines (S20) the loss and the spread coefficient based on the axis height and the number of pole pairs, wherein

[0020] The parameter values ​​are derived from the derived loss and dispersion coefficients and the nameplate specifications (LSA),

[0021] - Provide the derived parameter value.

[0022] In the context of the present invention, the term "by means of" is understood to mean that the derivation of the parameter value can include a plurality of sub-steps, wherein it is not necessary to use the trained function F in each sub-step. For example, the method can have sub-steps that can be performed by means of analytical mathematical formulas. This can be advantageous in terms of computing time.

[0023] In the context of the present invention, the term “connected to operating voltage” is understood to mean the state of a rotating electric machine in which the machine is connected to an operating voltage of fixed height and frequency (rigid network) or variable (inverter operation).

[0024] Preferably, the control method of the inverter is known, whereby the unknown inverter output voltage can be determined from the frequency f1 determined by means of a sensor.

[0025] In one embodiment, it can be provided that the simulation model is designed as an equivalent circuit diagram.

[0026] It is also conceivable that the simulation model is in the form of dependent equations or in table form.

[0027] In one specific embodiment, it can be provided that the trained function is based on a neural network, on regression, preferably on multivariate regression.

[0028] Furthermore, it can advantageously be provided that the input data are subjected to a plausibility check.

[0029] In one specific embodiment, it can be provided that the input data are formed by rating plate specifications.

[0030] It is possible to specify that some parameter values ​​are defined by the data on the nameplate. This can simplify the process by calculating the values ​​of certain parameters directly from the data on the nameplate.

[0031] In this case, provision can be made for a portion of the parameter values ​​to be determined by the rating plate specifications. The method can thus be simplified by calculating the values ​​of certain parameters directly from the rating plate specifications.

[0032] It is entirely conceivable that the input data consists only of the shaft height and the number of pole pairs. Good results can be achieved with such input data.

[0033] The advantage of multivariate regression and in particular also of neural networks is that the training input data do not have to be restricted and do not have to be found in a targeted manner. This is particularly advantageous if the input data include more than two input variables (e.g., shaft height and pole pair number), for example, a plurality of input variables, and the neural network is supplied with the input variables in order to determine the simulation model parameters.

[0034] Furthermore, according to the invention, the object is achieved by means of a computer-implemented method for providing a trained function of the above-mentioned type in that:

[0035] - obtaining training input data, the training input data being characteristic / representative for the rotating electrical machine and being obtainable from the machine when not connected to an operating voltage, wherein the training input data comprises shaft height and number of pole pairs,

[0036] - obtaining training output data, the training output data representing parameter values ​​and comprising losses and spread coefficients, wherein the parameter values ​​determine a simulation model of the rotating electric machine,

[0037] - train the function based on the training input data and the training output data,

[0038] - Provides trained functions.

[0039] In one embodiment, it can be provided that: before training the function on the training data set, the training data set including the training input data and the training output data is preprocessed. The preprocessing can, for example, include placing the data in a relative and standardized form.

[0040] It can be advantageous if the function is based on regression, preferably on multivariate regression.

[0041] This function can also be based on a neural network, so that the trained function described above can also be based on a neural network.

[0042] According to the invention, this object is also achieved by means of a first computer program in that the first computer program comprises instructions which, when the program is executed by a computer, cause the computer to derive a simulation model parameter value.

[0043] The computer program thus receives input data as input and outputs simulation model parameter values ​​as output.Since the simulation model is determined by the simulation model parameter values, it is also possible to speak of providing the simulation model or determining the simulation model parameter values.

[0044] According to the invention, this object is also achieved by means of a second computer program in that the second computer program comprises instructions which, when the program is executed by a computer, cause the computer to provide a trained function.

[0045] The first and / or second computer program can be stored on a computer-readable medium or transmitted by means of a data transmission signal, for example in coded form.

[0046] Furthermore, according to the invention, the object of the invention is achieved by means of a method for coupling a rotating electric machine to a platform, preferably a software platform, in particular a cloud platform of the above-mentioned type, in the following manner:

[0047] - acquiring input data in a state where the machine is not connected to an operating voltage, the input data comprising shaft height, number of pole pairs and nameplate specifications of the machine, wherein the input data are characteristic for the machine and are entered as input for configuring the application software, wherein

[0048] - configuring the application software to transmit input data to the platform, wherein the platform comprises instructions causing the above method to be performed by the platform so as to provide a simulation model of the rotating electric machine, wherein

[0049] The simulation model is preferably transmitted to a configuration application, wherein the configuration application transmits the simulation model to a sensor device which is arranged on the machine, preferably fixed on the machine.

[0050] Furthermore, according to the invention, the object of the invention is achieved by means of a method for determining the operating behavior of a rotating electric machine connected to an operating voltage and, for example, coupled to a platform, in that:

[0051] A) Detect the scattered magnetic field outside the machine,

[0052] B) Calculate the slip frequency f2 or the synchronous frequency or the synchronous speed f1 from the detected scattered magnetic field,

[0053] C) Calculate the slip s from the slip frequency f2 and the synchronous frequency or synchronous speed f1,

[0054] D) providing a simulation model according to the above method,

[0055] E) The operating behavior of the machine and in particular the operating point of the machine are derived from the slip s, the synchronous frequency f1 and the provided simulation model.

[0056] In addition, according to the present invention, the object of the present invention is achieved by means of a device of the proposed type in the following manner: the sensor device is designed to: detect a scattered magnetic field outside the machine, and calculate the slip frequency f2 and the synchronous frequency f1 from the detected scattered magnetic field, and transmit the slip frequency and the synchronous frequency to a computing device, wherein a simulation model provided according to the above method is stored on the computing device, and the computing device is configured to derive the slip s based on the calculated slip frequency f2 and the synchronous frequency f1, and then derive the operating performance of the machine by means of the simulation model. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] The present invention will be described and explained in more detail below based on the embodiments shown in the accompanying drawings. The accompanying drawings show:

[0058] Figure 1 A flow chart showing a computer-implemented method for providing an equivalent circuit diagram,

[0059] Figures 2 to 4 A feasible equivalent circuit diagram is shown.

[0060] Figure 5 a flowchart showing a computer-implemented method for providing a trained function,

[0061] Figure 6 shows the industrial IT environment, and

[0062] Figure 7 A flow chart showing a method for determining operating behavior of a rotating electric machine.

[0063] In the exemplary embodiments and figures, identical or identically acting elements are in each case provided with the same reference symbols. DETAILED DESCRIPTION

[0064] First, refer to Figure 1 .exist Figure 1 An embodiment of a computer-implemented method for providing a simulation model of a rotating electric machine 2 is shown as a flow chart in FIG. 1 . In the following, equivalent circuit diagrams EC1 , EC2 , EC3 are considered as examples of simulation models. Other simulation models, i.e., simulation models in the form of correlation equations or tables, for example, can also be used.

[0065] The equivalent circuit diagrams EC1 , EC2 , EC3 are determined by the values ​​of the predetermined parameters. Figures 2 to 4 By way of example, different equivalent circuit diagrams EC1 , EC2 , EC3 are shown. Figure 2 Involving T equivalent circuit diagram EC1, Figure 3 The T equivalent circuit diagram EC2 with separated rotor resistance is equivalent to Figure 2 The equivalent circuit diagram of T, and Figure 4The extended equivalent circuit diagram EC3 with the following parameters is involved: Main inductance X h (main flux), stator inductance X1 and rotor inductance X'2 (leakage flux), rotor resistance R'2 (rotor winding loss), stator resistance R1 (stator winding loss), R fe (Iron Loss) and R zus (Additional losses).

[0066] Without limiting the scope of the present disclosure and in order to simplify the description, the term “equivalent circuit diagram” is understood hereinafter to mean the extended equivalent circuit diagram EC3.

[0067] In step S1 of the method, input data are obtained. The input data can be obtained from the state in which the machine 2 is not connected to the operating voltage and are characteristic of the machine 2. In this case, the input data can be fed into the computing device, for example, via a first interface, for example as input into a configuration application. The input data are data that characterize the machine 2. The input data include, for example, nameplate specifications LSA—manufacturer specifications, standard specifications, legal regulations, etc., which characterize the machine at the rated point (at 100% load). Other data that can be obtained from the unconnected machine, but do not have to be specified on its rating plate, are, for example, the number of pole pairs, shaft height, etc.

[0068] Furthermore, the method can include a plausibility check of the input data (not shown here). For example, it can be checked here whether the efficiency and power factor in the rated point (if given) result in sufficiently small errors when calculating the mechanical rated power from the rated voltage and rated current.

[0069] In step S2 , parameter values ​​of the equivalent circuit diagram EC3 are derived from the input data with the aid of a trained function F.

[0070] In the context of the present invention, the term “by means of” is understood to mean that the derivation of the parameter value can include a plurality of sub-steps, wherein the trained function F does not necessarily have to be applied in every sub-step.

[0071] Figure 1 An embodiment is shown, in which step S2 includes two sub-steps (a first sub-step S20 and a second sub-step S21). For example, the first sub-step S20 can be a step of applying the trained function F to at least a portion of the input data. This portion of the input data includes, for example, the shaft height and the number of pole pairs of the machine 2. Here, the trained function F can derive the loss and the spread coefficient from at least a portion of the input data to determine the parameter values ​​of the equivalent circuit diagram EC3 of the machine 2.

[0072] The trained function F can be based on regression, preferably multivariate regression, for example.

[0073] Based on the parameters obtained in the first sub-step S20, such as loss and spread coefficients and the remainder of the input data (nameplate specifications LSA), in the second sub-step S21, parameter values ​​of equivalent circuit diagram parameters can be obtained, for example via other physical variables, preferably calculated with the aid of one or more predetermined mathematical formulas.

[0074] A non-exhaustive list of physical variables that can be calculated in step S2, preferably in the second sub-step S21, preferably with the aid of one or more predetermined mathematical formulas, is: number of pole pairs, synchronous speed, slip, power consumption, parasitic losses, individual stator winding losses, iron losses and friction losses and their sum, stator resistance, total reactive power, maximum possible spreading factor, etc.

[0075] For example, the number of pole pairs, synchronous speed and rated slip can be obtained from the rated frequency f N and rated speed n N The rated frequency and rated speed can be specified on the type plate. In addition, the electrical power consumption at the rated point (operation as a motor) can be calculated from the current, voltage and power factor (respectively at the rated point) or from the mechanical output power and efficiency (also respectively at the rated point).

[0076] For example, the additional losses can be calculated according to a standard or specification, for example according to a method according to IEC 60034-2-1 or EN 60034-2-1.

[0077] In a further step S3, the resulting parameter value is provided: main inductance X h (main flux), stator inductance X1 and rotor inductance X'2 (leakage flux), rotor resistance R'2 (rotor winding loss), stator resistance R1 (stator winding loss), R fe (Iron Loss) and R zus (Additional losses). This can be implemented, for example, by means of an output via the second interface (see, for example Figure 6 ). For example, the equivalent circuit diagram EC3 with the derived parameter values ​​can be stored by storing the parameters in a database 8 , for example in the cloud 4 .

[0078] Figure 5 A possible embodiment of a computer-implemented method for providing a trained function, such as the trained function F described above, is shown.

[0079] In step T1 , training input data are acquired, for example by means of a first training interface. The training input data characterize or represent the rotating electrical machine and can be acquired from the machine when it is not connected to an operating voltage. The training input data include shaft height and number of pole pairs.

[0080] In step T2 , training output data are obtained, for example by means of a first training interface, which are representative for parameter values ​​and include losses and spreading factors, wherein the parameter values ​​determine an equivalent circuit diagram of the rotating electric machine.

[0081] That is, the training data set includes data of machines whose equivalent circuit diagram parameter values ​​are already known. This can be data of asynchronous machines of known types, such as asynchronous motors or generators, for example.

[0082] The training data set including the training input data and the training output data can be preprocessed before training the function on the training data set. The preprocessing can include, for example, placing the data in a relative and standardized form.

[0083] In step T3, the function is trained based on the training input data and the training output data.

[0084] In step T4 , the trained function F is provided, for example, by means of a second training interface.

[0085] The function can be based on, for example, regression, preferably multivariate regression.

[0086] The training input data can include shaft heights (e.g., 63 mm to 450 mm), in particular shaft heights and pole pairs (e.g., 2 to 8), and the training output data can include loss and spread coefficients. It is to be understood that the method is also applicable to other shaft heights (e.g., less than 63 mm and greater than 455 mm) and pole pairs.

[0087] Thus, the training data set can comprise data of rotating electric machines, for example asynchronous machines of the same or similar type, for which the dependency of the equivalent circuit diagram parameters on their shaft height or on their shaft height and their number of pole pairs is known.

[0088] Furthermore, the trained function can be verified. This can be done, for example, in a load range of 50% to 125%.

[0089] The training dataset can contain data of 1000 to 2000 asynchronous machines of different types.

[0090] This function can also be based on a neural network, so that the above-mentioned trained function F can also be based on a neural network.

[0091] Each of the aforementioned computer-implemented methods may be in the form of computer program instructions, wherein when the computer program is executed by a computer or, more generally, by a data processing system, the instructions cause the computer or the data processing system to perform the corresponding method.

[0092] The computer program can be stored on a computer-readable medium or transmitted by means of a data carrier signal. The same applies to the method product, i.e. to the equivalent circuit diagrams EC1, EC2, EC3 with the derived parameter values ​​and to the trained function F, which can be stored on a computer-readable medium or transmitted by means of a data carrier signal.

[0093] Figure 6 An industrial IT environment 1 is shown in which one embodiment of a method for connecting or installing a new asset, here a rotating electrical machine 2, to a platform 3 is implemented. The platform can be designed, for example, as a software platform and in particular as a platform 3 (cloud platform 3) based on a cloud 4. For example, the cloud can be designed as an Internet of Things cloud, in particular as an Industrial Internet of Things cloud 4.

[0094] The cloud platform 3 can include one or more computing units. The cloud platform 3 can include, for example, the above-mentioned computer program, by means of which parameter values ​​of the equivalent circuit diagram of the machine 2 can be provided. For example, the computer program can be stored in the cloud 4 on the backend server 5.

[0095] In the present example, it is assumed that neither the equivalent component diagram nor the parameters of the equivalent circuit diagram of the machine 2 are known. It can be assumed that the machine 2 is described, for example, by an extended equivalent circuit diagram EC3. The parameter values ​​of the equivalent circuit diagram can be derived, for example, in the manner described below with the aid of the above-described embodiments of the method.

[0096] If the equivalent circuit diagram parameters of the machine 2 are known, the IT environment 1 can be used to train the function F, for example in the manner described above. For this purpose, for example, the backend server 5 can have a computer program with corresponding instructions.

[0097] Firstly, input data are acquired from the machine 2 (eg an asynchronous motor or a generator) in a state not connected to an operating voltage, which input data are characteristic of the machine 2 and can be entered as input into a configuration application software.

[0098] For example, this can be done by the user by reading out the specification, for example, from the rating plate LS to measure the shaft height of the asynchronous motor 2, wherein the user, for example, first collects data related to the asset (for example, a machine, in particular an asynchronous motor). In addition, the number of pole pairs can be calculated, for example, from the rating plate specifications using the above-mentioned method.

[0099] When setting up the asynchronous motor or generator 2 , the user can enter the input data obtained into a configuration application on a mobile device 6 (eg a smartphone / handheld device) within the scope of a so-called setup.

[0100] The configuration application can transmit the input data to the IoT cloud 4 , in particular the backend server 5 , via a data connection 7 , such as WLAN or GSM radio.

[0101] The input data can also be stored in a database 8 , wherein the database 8 can be designed as a database of the cloud platform 3 (not shown) or also as a cloud database of a third party (not shown here).

[0102] The database 8 is accessible by a computer program stored on the backend server 5 .

[0103] The computer program then accesses the input data and generates / derives an associated simulation model EC3 , ie parameter values, for the asynchronous motor 2 .

[0104] This method can be repeated for each new asset.

[0105] In an intermediate step, the information content of the simulation model can then be reduced for information security reasons. In this case, the extended equivalent circuit diagram EC3 with the resulting parameter values ​​can be converted into a compressed or simplified equivalent circuit diagram.

[0106] A simplified equivalent circuit diagram has less information content than an extended equivalent circuit diagram. In a simplified equivalent circuit diagram, elements can be combined or completely removed from the equivalent circuit diagram. In this case, the simplified information is at least partially supplemented again in the calculation on the IoT device by making (reasonable) assumptions, which, however, only apply to the individual motor (and not necessarily exactly). As a result, statements about the operating behavior are very limited and are limited to torque and power.

[0107] The simulation model (equivalent circuit diagram EC3 with the resulting parameter values) can then be stored, for example, in the platform database 8 or on the backend server 5 in order to be available to the configuration application during further installation or transmitted to the configuration application without intermediate storage.

[0108] After the configuration application has the simulation model, it can transmit the simulation model via a further, for example radio-based data connection 9 (eg Bluetooth or WLAN) to an IoT device 10 which is arranged, preferably fastened, to the asynchronous motor 2 .

[0109] The IoT device 10 can be designed as a sensor device, which is designed and configured to: generate sensor data and use a simulation model (equivalent circuit diagram EC3 or simplified equivalent circuit diagram with derived parameter values) to, for example, calculate data of the current motor operating point or asset KPI (KPI is a key performance indicator) from the sensor data with the help of the simulation model. For this purpose, a computing unit and / or a computer-readable storage medium (not shown here) can be provided in the IoT device 10.

[0110] The IoT device 10 is preferably not coupled to machine sensors inside (or outside) the machine 2. Figure 7 Description of the invention) does not require expensive and maintenance-intensive sensors, which can measure current or voltage, for example. As a result, the IoT device 10 is very cheap and maintenance-friendly. For example, torque and / or power, such as mechanical shaft power and / or power factor and / or efficiency can be derived.

[0111] The calculated asset KPI (operating performance of the asynchronous motor 2) can be transmitted, for example, at adjustable (time) intervals via another preferably radio-based data connection 11 (e.g. WLAN) to the IoT cloud platform 3. There, the asset KPI can be further processed in IT or stored in a database 8 or on a backend server 5. In addition, the asset KPI can be displayed, for example, on a display device arranged in the front end of the IoT cloud platform, for example, on a dashboard 12.

[0112] The data connections 7 , 9 , 11 are preferably all double-sided.

[0113] In one specific embodiment (not shown here), the transmission of the simulation model to the IoT device 10 can be dispensed with. In this case, the above-mentioned compression of the equivalent circuit diagram can also be dispensed with, thereby enabling better results to be achieved. Instead, the calculation of the asset KPI (operating behavior of the asynchronous motor or generator 2) can be transferred to the IoT cloud platform 3.

[0114] Figure 7 A flow chart of one specific embodiment of the method briefly discussed above for determining the operating behavior of a machine 2 when it is put into operation is shown.

[0115] In step A, a scattered magnetic field outside the machine 2 is detected. In step B, the slip frequency f2 and the synchronization frequency f1 are calculated from the detected scattered magnetic field. In step C, the slip s is calculated from the slip frequency f2 and the synchronization frequency f1. For example, this can be implemented with the help of signal processing software. In step D, as described above, the equivalent circuit diagram (derived equivalent circuit diagram parameter values) is provided (for example with the help of data transmission from the mobile device 6). The provision of the equivalent circuit diagram with the help of data transmission from the mobile device 6 preferably does not occur during a measurement or calculation sequence. Instead, the equivalent circuit diagram can be transmitted during the inhabitation and stored on the IoT device.

[0116] In step E, the operating behavior of the machine 2 and in particular its operating point are determined from the slip s, the synchronous frequency f1 and the provided equivalent circuit diagram (derived equivalent circuit diagram parameter values).

[0117] For example, steps A to C and E can all be implemented by means of the Internet of Things device 10 .

[0118] However, it is also possible to consider that the IoT device 10 only performs one or more of steps A to C, and the remaining steps are performed by another computing unit, such as a local server or a backend server 5. For example, the IoT device 10 can transmit the scattered magnetic field data to the computing unit for further processing and analysis.

[0119] Thus, the device for determining the operating behavior of a rotating electric machine 2 connected to an operating voltage can include a sensor device 10 and a computing device, which are configured and cooperate to implement steps A to E of the method. The computing device can include, for example, a backend server 5, a mobile device 6, a database 8 and a display device 12 (see Figure 6 ).

[0120] Although the invention has been described and illustrated in detail by way of example, the invention is not limited to the disclosed examples. Other variants can be derived by a person skilled in the art without departing from the scope of protection of the invention. In particular, the described method for providing an equivalent circuit diagram can be improved by features described or claimed with respect to the method for providing a trained function, and vice versa. This applies, mutatis mutandis, to a method for coupling a rotating electric machine to a platform, to a method for deriving operating behavior of a rotating electric machine connected to an operating voltage, and to a device for deriving operating behavior of a rotating electric machine connected to an operating voltage.

Claims

1. A method for coupling a rotating electrical machine (2) to a platform (3), in which method - obtaining input data based on the machine (2) being not connected to an operating voltage, the input data comprising shaft height, number of pole pairs and nameplate specifications (LSA) of the machine (2), wherein: The input data characterize the machine (2) and are entered as input for configuring application software, wherein - the configuration application software transmits the input data to the platform (3), wherein the platform (3) provides a simulation model of the rotating electric machine (2), wherein the platform (3): * Get (S1) input data, * determining (S2) the parameters of the simulation model from the input data by means of a trained function, wherein the trained function (F) is trained with the aid of a training input and training input data, The training input data characterizes an electric rotating machine, and can be obtained from the machine when it is not connected to an operating voltage, wherein the training input data includes a shaft height and a number of pole pairs, wherein the training input data comprises losses and spreading coefficients of the machine and the training input data is representative for parameter values ​​that determine a simulation model of the rotating electric machine, wherein a trained function (F) is obtained (S20) based on the shaft height and pole pair number, loss and spread coefficient of the electric rotating machine (2) The parameter value is obtained from the obtained loss and spread coefficient and the nameplate specification (LSA).

2. The method according to claim 1, wherein: The simulation model is transmitted to the configuration application software, wherein the configuration application software transmits the simulation model to a sensor device (10) arranged on the machine (2), preferably fixed on the machine (2).

3. The method according to claim 1 or 2, wherein: The platform (3) is a software platform.

4. The method according to any one of claims 1 to 3, wherein: The platform (3) is a cloud platform.

5. The method according to any one of claims 1 to 4, wherein: The trained function is based on a neural network, based on regression.

6. The method according to claim 5, wherein: The function is based on multiple regression.

7. The method according to any one of claims 1 to 6, wherein: Before training the function on the training data set, the training data set including the training input data and the training output data is preprocessed.

8. The method according to claim 7, wherein: The pre-processing includes putting the data into a relative and normalized form.

9. The method according to any one of claims 1 to 8, wherein: A plausibility check is performed on the input data.

10. The method according to any one of claims 1 to 9, wherein: The simulation model is designed as an equivalent circuit diagram.

11. The method according to claim 10, wherein: The equivalent circuit diagram is an expanded equivalent circuit diagram.

12. A method for determining the operating behavior of a rotating electrical machine (2) connected to an operating voltage, in which method: A) connecting a rotating electrical machine (2) to the platform (3) according to the method according to any one of claims 1 to 11, B) detecting a scattered magnetic field outside the machine (2), C) calculating the slip frequency and the synchronous frequency from the detected stray magnetic field, D) calculating a slip from the slip frequency and the synchronous frequency, E) Calculating the operating performance of the machine (2) based on the slip, the synchronous frequency and the simulation model provided by the platform (3).

13. The method according to claim 12, wherein: An operating point of the machine is determined based on the slip, the synchronous frequency and the provided simulation model.

14. A device for determining the operating behavior of a rotating electrical machine (2) connected to an operating voltage, wherein: The device comprises a sensor device (10) and a computing device (5, 6, 8, 12), in which: The sensor device (10) is designed to detect a scattered magnetic field outside the machine (2), calculate a slip frequency and a synchronous frequency from the detected scattered magnetic field, and transmit the slip frequency and the synchronous frequency to the calculation device (5, 6, 8, 12), wherein: A simulation model (EC1, EC2, EC3) of a rotating electric machine (2) is stored on the computing device (5, 6, 8, 12), wherein the simulation model is provided by a computer program, wherein the computer program * Get (S1) input data, * determining (S2) the parameters of the simulation model from the input data by means of a trained function (F), wherein the trained function (F) is trained with the aid of a training input and training input data, The training input data characterizes an electric rotating machine, and can be obtained from the machine when it is not connected to an operating voltage, wherein the training input data includes a shaft height and a number of pole pairs, wherein the training input data comprises losses and spreading coefficients of the machine and the training input data is representative for parameter values ​​that determine a simulation model of the rotating electric machine, wherein a trained function (F) is obtained (S20) based on the shaft height and pole pair number, loss and spread coefficient of the electric rotating machine (2) wherein the parameter value is obtained from the obtained loss and spread coefficient and the nameplate specification (LSA), The calculation device (5, 6, 8, 12) is configured to determine the slip based on the calculated slip frequency and the synchronous frequency and then determine the operating behavior of the machine (2) with the aid of the simulation model (EC1, EC2, EC3).

Images