Frequency converter and driving system

Abstract

The invention relates to a frequency converter (1) for actuating an electric motor (2), comprising:-sensor elements (3 to 8), which are designed to generate sensor data (x1 to xn); and-a control unit (9) configured to control the operation of the frequency converter (1) as a function of the sensor data (x1 to xn), the control unit (9) being configured to calculate at least one characteristic (y1 to ym) as a function of the sensor data (x1 to xn).

Description

[0001] The task underlying the present application is to provide a frequency converter and a drive system which is able to make as reliable a statement as possible about its state.

[0002] The frequency converter is used to operate an electric motor. The frequency converter has conventional sensor elements or sensors which are configured to generate sensor data. The frequency converter also has a control unit, for example in the form of a microprocessor-based controller, which is configured to control the operation of the frequency converter on the basis of the sensor data. The control unit is configured to calculate at least one feature from the sensor data. A feature here is in particular a scalar value which is derived from the sensor data and characterizes the sensor data. For example, a feature is the maximum value, the minimum value, the time average, the energy in a specific frequency band, etc. of the sensor data.

[0003] In one embodiment, the at least one feature constitutes an input variable for an artificial intelligence (AI) unit separate from the frequency converter.

[0004] In one embodiment, the frequency converter has at least one, in particular all, of the following sensor elements: a current sensor for detecting the drive current output by the frequency converter; a rotational speed sensor for detecting the engine rotational speed of an electric motor operated by means of the frequency converter; a position sensor for detecting the position of a mechanical load moved by means of the electric motor; a sensor for detecting the operating duration of the frequency converter; a voltage sensor for detecting the drive voltage generated by means of the frequency converter; a temperature sensor for detecting the operating temperature of the frequency converter; and a temperature sensor connection for connecting an external temperature sensor, in particular an engine temperature sensor.

[0005] In one embodiment, the control unit is configured to calculate at least one, in particular all, of the following features from the sensor data: the frequency spectrum of the drive current, and / or the frequency spectrum of the motor rotational speed, and / or the frequency spectrum of the load position, and / or the frequency spectrum of the Schleppfehler, and / or the average torque generated by means of the electric motor, and / or the average engine rotational speed, and / or the average electrical or mechanical power, and / or the variation of the torque generated by means of the electric motor, and / or the variation of the engine rotational speed, the operating hours, and / or the engine revolutions when the torque is greater than a threshold torque, and / or the maximum rotational speed, and / or the maximum torque generated by means of the electric motor, and / or the maximum electrical power, and / or the maximum operating temperature, and / or the maximum Schleppfehler.

[0006] In one implementation, the control unit is configured to calculate the at least one feature discontinuously over time based on sensor data; that is, not continuously or intermittently. This makes it possible to generate small amounts of information-rich data, since the at least one feature is generated only when given characteristic and meaningful conditions. During other time periods, no calculations are performed, thereby minimizing computational cost and data volume.

[0007] In one embodiment, the control unit is configured to calculate the at least one feature based on sensor data only during a time period during which the inverter manipulates the motor to execute a pre-given test driving profile or measurement profile. Therefore, the at least one feature is determined only under comparable or identical conditions.

[0008] In one implementation, the pre-defined test driving profile consists of a pre-defined sequence of acceleration and motion at a constant speed. For example, the pre-defined test driving profile may consist of: starting from a standstill, accelerating to a maximum speed, maintaining that maximum speed for a pre-defined duration, and then braking again until coming to a standstill.

[0009] In one embodiment, the control unit is configured to calculate the at least one characteristic solely based on sensor data relating to the torque generated by the electric motor. As can be seen from the invention, torque is highly characteristic of the state of an electric drive system, thus evaluating only the torque may be sufficient.

[0010] In one implementation, the control unit is configured to calculate the at least one feature only based on sensor data that meets predefined criteria, which ensure information content and comparability. Typical criteria selected may be: the speed being within a specified range during measurement, or the torque continuously exceeding a minimum value.

[0011] The drive system has a frequency converter as described above and an AI unit constructed separately from the frequency converter, wherein at least one feature generated by means of the frequency converter is set as an input parameter constituting the AI ​​unit.

[0012] In one implementation, the AI ​​unit is configured to perform predictive maintenance, that is, to determine the state of the driving system based on the at least one feature. This state can be, for example: fault-free, requiring maintenance, faulty, etc.

[0013] To determine the condition of a drive system, such as to determine whether there are defects in the drive system components, AI or machine learning is used according to the present invention. In these methods, unlike deep learning, sensor data or its measurement sequences are not used directly. Instead, the input to the AI ​​unit consists of scalar parameters or so-called features calculated from the sensor data. These features are, for example, measured attributes or individual values. During the training of the AI, a wide variety of features are calculated and used. Part of the learning involves selecting the optimal features. Simple features include, for example, maximum values, average values, the number of values ​​above 80%, etc. Correlational features include, for example, the average of the product of two values, the maximum of the sum of two values, the maximum value, etc. Statistical features include, for example, the width and / or shape of the distribution curve. Features from the frequency domain include, for example, frequency components at a specific frequency, the lowest frequency, the uniformity of the frequency distribution, etc.

[0014] Features are particularly meaningful when they have a technical motivation. For example, speed-related errors occur when a bearing fails. Nonlinearity in force transmission near zero point, for example, indicates a loose timing belt. Combining domain knowledge and AI in feature selection is optimal for machine learning.

[0015] It is generally assumed that the data needed to compute features is available at the location where the AI ​​units are deployed, typically in the so-called cloud. However, this is often not the case for drive systems, as some data is not in the cloud or is not stored. Inverters generate large amounts of sensor data at high sampling rates. The resulting amount of data is typically not feasible to transmit to the cloud in real time. Therefore, features cannot be computed in the frequency domain due to the lack of the corresponding sensor data at the necessary time resolution. Furthermore, transmitting sensor data to the cloud is also highly complex.

[0016] According to the invention, it is also recognized that sensor data generated by directional excitation of the frequency converter is significantly more meaningful than sensor data determined during normal operation. The excitation induced by the frequency converter can be, for example, torque noise. This means that, based on a short training phase, better prediction probabilities can be achieved than over a long period of actual operation. Furthermore, the computational performance required for feature-based evaluation is significantly less than that required for sensor data-based evaluation.

[0017] According to the invention, features are provided in the frequency converter or its control unit. According to the invention, features meaningful to the state of the frequency converter or drive system are determined in this case. Furthermore, suitable excitations, such as torque noise, current surges, characteristic motion, or motion superposition, can be generated by means of the frequency converter; these excitations are used to generate meaningful sensor data, which is then used to generate the characteristic features.

[0018] Several features are described below as examples.

[0019] 1. Characteristics based on frequency analysis of current oscillation, rotational speed oscillation, trailing error, and / or position.

[0020] To address this, the rotational speed is determined using an excitation incorporating torque noise, and the determined speed is then converted to the frequency domain. The spectral components are then summed across different frequency domains.

[0021] Such characteristics make it possible, for example, to determine changes in the natural frequency. Nonlinear changes in the natural frequency can indicate, for example, a loosening of the transmission system, such as a loose timing belt. The correlation between high-amplitude oscillations and rotational speed can indicate, for example, shocks caused by gear failure.

[0022] 2. Based on the characteristics of average torque, speed, or electrical power during constant mechanical motion.

[0023] In this way, for example, once a constant speed is reached, the corresponding static torque can be determined. This method can also be used to determine, for example, friction, bearing condition, and capacitor condition.

[0024] 3. Based on the characteristics of torque and speed variations.

[0025] Extensive variation or distribution of sensor data may indicate localized friction due to jamming and / or contamination, as well as sensor element failure.

[0026] 4. Characteristics of inertia in a mapping transmission system.

[0027] When the drive unit accelerates and the frictional and process forces are known, inertia can be calculated. Inertia, and especially its changes, is a characteristic that provides information about the dynamic characteristics of the drive system.

[0028] 5. Characteristics indicating service life.

[0029] Operating hours, RPMs when torque exceeds a threshold, etc., can provide information about the remaining service life of many components of the drive system, because aging depends on the stress on the components.

[0030] 6. Features related to the maximum value of sensor data.

[0031] The maximum values ​​of speed, torque, position, drag error, power, temperature, acceleration, etc., are usually constant during the cycle. Changes in these maximum values ​​may indicate changes in the state of the drive system.

[0032] This invention provides optimal preparation for machine learning and predictive maintenance in drive systems because standardized datasets are provided in the inverter as features. Furthermore, optimized incentives can improve the effectiveness of the data for all types of AI.

[0033] According to the present invention, information in the frequency converter is provided in the form of meaningful features. Because the information content of the generated features or data is increased, a good description of the state of the drive system can be made using a simple decision tree. Attached Figure Description

[0034] The present invention will now be described in detail with reference to the accompanying drawings. Herein: Figure 1 An electric drive system with a frequency converter and an AI unit for determining the state of the electric drive system is shown.

[0035] Figure 1 shows an electric drive system 100, which has a frequency converter 1 and an AI unit 10 for determining the state of the electric drive system 100.

[0036] Inverter 1 is typically used to control motor 2. Inverter 1 has sensor elements 3 to 8, which are configured to generate corresponding sensor data x1 to xn, where n is a natural number greater than 1. The sensor data x1 to xn are represented, for example, as discrete time sequences of digital values.

[0037] The frequency converter 1 has a control unit 9, which is configured to control the operation of the frequency converter 1 in a conventional manner based on sensor data x1 to xn.

[0038] The control unit 9 is configured to calculate m features y1 to ym based on sensor data x1 to xn, where m is a natural number greater than or equal to 1.

[0039] For example, inverter 1 may have the following sensor elements: current sensor 3 for detecting drive current, speed sensor 4 for detecting engine speed, position sensor 5 for detecting load position, sensor 6 for detecting the operating duration of inverter 1, voltage sensor 7 for detecting drive voltage, and temperature sensor 8 for detecting the operating temperature of inverter 1.

[0040] Features y1 to ym are set to form input parameters for AI unit 10, which is separate from inverter 1.

[0041] The control unit 9 is configured to calculate at least the following characteristics y1 to ym based on sensor data: the spectrum of drive current, and / or the spectrum of engine speed, and / or the spectrum of load position, and / or the spectrum of drag error, and / or the average torque generated by means of motor 2, and / or the average engine speed, and / or the average electrical or mechanical power, and / or the variation of torque generated by means of motor 2, and / or the variation of engine speed, and / or the number of operating hours, and / or the number of engine speeds when the torque is greater than the threshold torque, and / or the maximum speed, and / or the maximum torque generated by means of motor 2, and / or the maximum electrical power, and / or the maximum operating temperature, and / or the maximum drag error.

[0042] AI unit 10 is constructed to determine the state of driving system 100 in the form of state parameters Z1 to Zk based on features y1 to ym, where k is a natural number greater than or equal to 1.

Claims

1. A frequency converter (1) for controlling an electric motor (2), comprising: - Sensor elements (3 to 8), configured to generate sensor data (x1 to xn), and - Control unit (9), which is configured to control the operation of frequency converter (1) based on the sensor data (x1 to xn), Its features are, - The control unit (9) is configured to calculate at least one feature (y1 to ym) based on the sensor data (x1 to xn).

2. The frequency converter (1) according to claim 1, characterized in that, - The at least one feature (y1 to ym) is configured to form input parameters for the AI ​​unit (10) separate from the frequency converter.

3. The frequency converter (1) according to any one of the preceding claims, characterized in that, - The frequency converter (1) has at least one of the following sensor elements: - Current sensor used to detect drive current (3) - A speed sensor (4) used to detect engine speed. - Position sensor (5) used to detect the position of the load - Sensor (6) for detecting the operating duration of the frequency converter (1) - Voltage sensor (7) used to detect drive voltage - Temperature sensor (8) for detecting the operating temperature of inverter (1), and - Temperature sensor connector for connecting external temperature sensors, especially engine temperature sensors.

4. The frequency converter (1) according to claim 3, characterized in that, - The control unit (9) is configured to calculate at least one of the following features (y1 to ym) based on sensor data: - The spectrum of drive current, and / or the spectrum of engine speed, and / or the spectrum of load position, and / or the spectrum of drag error. - The average torque generated by the electric motor (2), and / or the average engine speed, and / or the average electrical or mechanical power, - By means of the change in torque generated by the electric motor (2), and / or the change in engine speed, - Operating hours, and / or engine speeds when torque is greater than the threshold torque, and - Maximum speed, and / or maximum torque generated by means of the motor (2), and / or maximum electrical power, and / or maximum operating temperature, and / or maximum drag error.

5. The frequency converter (1) according to any one of the preceding claims, characterized in that, - The control unit (9) is configured to calculate at least one feature (y1 to ym) discontinuously based on sensor data (x1 to xn).

6. The frequency converter (1) according to any one of the preceding claims, characterized in that, - The control unit (9) is configured to calculate at least one feature (y1 to ym) based on sensor data (x1 to xn) only when the inverter (1) controls the motor (2) to execute a pre-given test driving profile.

7. The frequency converter (1) according to claim 6, characterized in that, - The pre-given test driving profile consists of a pre-given sequence of acceleration and motion with constant speed.

8. The frequency converter (1) according to any one of the preceding claims, characterized in that, - The control unit (9) is configured to calculate the at least one feature (y1 to ym) based solely on sensor data (x1 to xn) related to the torque generated by the electric motor (2).

9. The frequency converter (1) according to any one of the preceding claims, characterized in that, The control unit (9) is configured to calculate the at least one feature (y1 to ym) based solely on the selected sensor data (x1 to xn), wherein the operating parameter values, in particular in the form of position values, speed values ​​and torque values, are within specified limits for the selected sensor data (x1 to xn).

10. A drive system (100) comprising: - The frequency converter (1) according to any one of the preceding claims, and - An AI unit (10) constructed separately from the frequency converter (1), wherein at least one feature (y1 to ym) generated by means of the frequency converter (1) is set as an input parameter constituting the AI ​​unit (10).

11. The drive system (100) according to claim 10, characterized in that, - The AI ​​unit (10) is configured to determine the state (Z1 to Zk) of the drive system (100) based on the at least one feature.

Images