Dynamic classification method for dangerous operation state of motor

Abstract

Provided is a low-load dynamic classification method capable of determining a mechanical dangerous operation state on the basis of three-axis vibration data collected from a motor element. The dynamic classification method for a dangerous operation state of a motor according to an embodiment of the present invention comprises the steps in which: a communication unit collects three-axis vibration data on a motor; a processor generates a prediction model by training an initial prediction model for predicting the operation state of the motor by using the three-axis vibration data; and the processor predicts the operation state of the motor by using the generated prediction model. Accordingly, a calculation load required for driving the algorithm can be minimized.

Description

Dynamic motor hazardous operating condition classification method

[0001] The present invention relates to a method for predicting (classifying) the operating state of an electric motor, and more particularly, to a low-load dynamic classification method capable of determining a mechanically hazardous operating state under low-load conditions according to the characteristics of operating data based on three-axis vibration data collected for electric motor elements.

[0002] In general, abnormal operating conditions of an electric motor can be determined by analyzing characteristics such as vibration, heat, and current. In particular, mechanical abnormal operating conditions such as bearing damage, imbalance, and alignment problems have prominent classification characteristics in the three-axis vibration data of the electric motor.

[0003] For such 3-axis vibration data, the amount of data collected can be huge depending on the required sampling frequency characteristics, and in this case, in terms of data communication, there is uncertainty in that the time it takes for the data to be transmitted from the sensor to the processor that performs state prediction may be delayed depending on the communication speed and packet capacity constraints.

[0004] In terms of motor condition prediction calculations, if the algorithm must be configured in an embedded format in an environment such as an MCU, there is a possibility that the amount of 3-axis vibration data may increase the computational complexity required for preprocessing, feature analysis, and calculations leading to condition prediction.

[0005] In addition, in the case of a general-purpose electric motor hazardous operating state classification algorithm, the number of characteristics used in the analysis of the three-axis vibration signal may be excessively set, which may actually entail unnecessary calculations in the analysis of the characteristics of the target electric motor.

[0006] Accordingly, a dynamic configuration method is required to optimize the configuration of the state classification algorithm according to the operating characteristics of the target motor to which the state diagnosis algorithm is applied, and in particular, a method to predict (classify) the operating status of the motor by reflecting the main hardware and communication characteristics is required.

[0007] The present invention has been devised to solve the above problems, and the purpose of the present invention is to provide a dynamic motor dangerous operating state classification method capable of optimizing the configuration of a state classification algorithm according to the operating characteristics of a target motor and predicting (classifying) the operating state of the motor by reflecting the main hardware and communication characteristics.

[0008] According to one embodiment of the present invention for achieving the above object, a method for classifying a dangerous driving state of a dynamic motor comprises: a step in which a communication unit collects three-axis vibration data of the motor; a step in which a processor trains an initial prediction model for predicting the driving state of the motor with the three-axis vibration data to generate a prediction model; and a step in which the processor utilizes the generated prediction model to predict the driving state of the motor.

[0009] And, the collecting step can collect 3-axis vibration data according to a predefined communication speed and sampling rate.

[0010] In addition, the step of predicting the operating state of the electric motor may include a filtering step of filtering the characteristics of a three-axis vibration signal when the response time of the prediction model does not meet the preset design response speed requirement in order to predict the operating state by reflecting the operating characteristics and operating environment characteristics of the electric motor; and a response characteristic improvement step of improving the response characteristics by reflecting the hardware characteristics of the electric motor.

[0011] And, the filtering step may perform a first filtering procedure to remove the y-axis or z-axis vibration characteristic when the difference between the vibration characteristic of the y-axis or z-axis signal and the x-axis signal is less than a preset first threshold value based on the x-axis signal of the 3-axis vibration signal.

[0012] Additionally, the filtering step may perform a second filtering procedure to remove a time-domain characteristic if the difference between the preset sub-characteristics in the time-domain is less than or equal to a preset second threshold value based on a predefined time-domain characteristic after the first filtering procedure is performed.

[0013] And, the filtering step may perform a third filtering procedure to remove the frequency domain characteristic if the difference between the preset sub-characteristics in the frequency domain is less than or equal to a preset third threshold value based on the predefined frequency domain characteristic after the second filtering procedure is performed.

[0014] Additionally, the response characteristic improvement step can set design response speed requirements consisting of data transmission time conditions and data operation time conditions prior to improving the response characteristic.

[0015] And, in the response characteristic improvement step, if the preset data transmission time condition is not met due to the actual communication speed, the first response characteristic improvement procedure can be performed to reduce the predefined Sampling Interval until it meets the preset data transmission time condition.

[0016] In addition, in the response characteristic improvement step, if the data operation time for predicting the operating state of the motor does not meet the preset data operation time condition due to the complexity of the hardware characteristics reflected in the operating state prediction, a secondary response characteristic improvement procedure may be performed to reduce the number of hardware characteristics reflected in the operating state prediction until it meets the preset data operation time condition.

[0017] Meanwhile, according to another embodiment of the present invention, a dynamic motor dangerous driving state classification system includes a communication unit that collects three-axis vibration data of the motor; and a processor that trains an initial prediction model for predicting the driving state of the motor with the three-axis vibration data to generate a prediction model, and uses the generated prediction model to predict the driving state of the motor.

[0018] And according to another embodiment of the present invention, a method for classifying a dangerous driving state of a dynamic motor includes a step of generating a prediction model by having a processor train an initial prediction model for predicting the driving state of the motor with three-axis vibration data; and a step of predicting the driving state of the motor by using the generated prediction model; wherein, when the response time of the prediction model does not meet a preset design response speed requirement, the prediction model performs a filtering procedure for filtering the characteristics of a vibration signal and a response characteristic improvement procedure for improving the response characteristics by reflecting the hardware characteristics of the motor.

[0019] In addition, according to another embodiment of the present invention, a dynamic electric motor dangerous driving state classification system includes a processor for generating a prediction model by learning an initial prediction model for predicting the driving state of an electric motor with three-axis vibration data, and predicting the driving state of an electric motor by utilizing the generated prediction model; and a storage unit for storing the generated prediction model; wherein, when the response time of the prediction model does not meet a preset design response speed requirement, the prediction model performs a filtering procedure for filtering the characteristics of a vibration signal and a response characteristic improvement procedure for improving the response characteristics by reflecting the hardware characteristics of the electric motor.

[0020] As described above, according to embodiments of the present invention, by configuring an optimized driving state classification algorithm that reflects the driving characteristics and operating environment characteristics (hardware characteristics, communication standard characteristics) of the target motor, the computational load required for algorithm operation is minimized, and through this, the classification response speed is maximized, thereby enabling real-time driving state diagnosis.

[0021] FIG. 1 is a drawing provided to explain the configuration of a dynamic motor dangerous driving state classification system according to one embodiment of the present invention;

[0022] FIG. 2 is a drawing provided for explaining a method for classifying a dynamic motor dangerous driving state according to one embodiment of the present invention;

[0023] FIG. 3 is a drawing providing a more detailed description of a method for performing a filtering procedure for filtering characteristics of a vibration signal according to one embodiment of the present invention; and

[0024] FIG. 4 is a drawing provided for a more detailed description of a method for performing a response characteristic improvement procedure that improves response characteristics by reflecting hardware characteristics of an electric motor according to one embodiment of the present invention.

[0025] Hereinafter, the present invention will be described in more detail with reference to the drawings.

[0026] FIG. 1 is a drawing provided to explain the configuration of a dynamic motor dangerous driving state classification system according to one embodiment of the present invention.

[0027] The dynamic motor dangerous driving state classification system according to the present embodiment (hereinafter collectively referred to as the 'driving state classification system') is provided to optimize the configuration of a prediction model that predicts the driving state of a motor according to the driving characteristics of the target motor, and to predict (classify) the driving state of the motor by reflecting the main hardware and communication characteristics.

[0028] To this end, the driving state classification system may include a communication unit (110), a processor (120), and a storage unit (130).

[0029] The communication unit (110) is a communication means for connecting the driving state classification system to a communication network with a sensor or external server that provides 3-axis vibration data of the motor, and can collect 3-axis vibration data of the motor.

[0030] The storage unit (130) is a storage medium that stores programs and data necessary for the processor (120) to operate. For example, the storage unit (130) can store three-axis vibration data of an electric motor collected through the communication unit (110) and a prediction model generated by the processor (120).

[0031] The processor (120) is provided to process all matters of the driving status classification system.

[0032] Specifically, the processor (120) can generate a prediction model by learning an initial prediction model for predicting the operating state of the electric motor using three-axis vibration data, and can predict the operating state of the electric motor by utilizing the generated prediction model.

[0033] At this time, the prediction model generated by the processor (120) is designed to predict the driving state by reflecting the driving characteristics and operating environment characteristics of the motor, and the response time of the prediction model is set to a preset design response speed requirement (t 설계 ) does not match, a filtering procedure for filtering the characteristics of the vibration signal and a response characteristic improvement procedure for improving the response characteristics by reflecting the hardware characteristics of the motor can be performed.

[0034] FIG. 2 is a drawing provided for explaining a method for classifying a dynamic motor dangerous driving state according to one embodiment of the present invention.

[0035] The dynamic motor dangerous driving state classification method according to the present embodiment can be executed by the driving state classification system described above with reference to FIG. 1.

[0036] Referring to FIG. 2, the driving state classification system can collect 3-axis vibration data of the motor through the communication unit (110) (S210), and generate a prediction model by training an initial prediction model for predicting the driving state of the motor through the processor (120) with the 3-axis vibration data (S220).

[0037] Specifically, the driving state classification system can collect three-axis vibration data of the motor according to a predefined communication speed and sampling rate during the process of collecting the data. Furthermore, the driving state classification system can predefine time-domain and frequency-domain characteristics.

[0038] The processor (120) generates a prediction model by reflecting the time domain characteristics and frequency domain characteristics defined in advance, and predicts the driving state by reflecting the driving characteristics and operating environment characteristics of the motor, so that the response time of the prediction model is set to a preset design response speed requirement (t 설계 ) is met (S230-Yes), the operating status of the motor is predicted using the prediction model (S250), and the response time of the prediction model is set to the design response speed requirement (t 설계 ) does not match (S230-No), after performing a filtering procedure (S235) for filtering the characteristics of the 3-axis vibration signal and a response characteristic improvement procedure (S240) for improving the response characteristics by reflecting the hardware characteristics of the motor, the operating state of the motor can be predicted (S250).

[0039] FIG. 3 is a diagram provided for a more detailed description of a method for performing a filtering procedure for filtering characteristics of a vibration signal according to one embodiment of the present invention.

[0040] The processor (120) may operate as illustrated in FIG. 3 when performing a filtering procedure for filtering the characteristics of a three-axis vibration signal.

[0041] Specifically, the processor (120) can remove the y-axis or z-axis vibration characteristic (S315) if the y-axis and z-axis signal characteristics are similar to the x-axis signal (S310-Yes) based on the x-axis signal (short-circuit judgment based on the x-axis signal) among the collected 3-axis vibration signals.

[0042] That is, the processor (120) determines, based on the x-axis signal of the 3-axis vibration signal, if the vibration characteristic of the y-axis or z-axis signal is less than or equal to a preset first threshold value compared to the x-axis signal ( ), a first filtering procedure can be performed to remove y-axis or z-axis vibration characteristics.

[0043] And, after the first filtering procedure is performed, the processor (120) can remove the time domain characteristic (1st momentum characteristic) if the lower characteristics are similar (S320-Yes) based on the time domain characteristic (1st momentum characteristic) (S325).

[0044] That is, after the first filtering procedure is performed, the processor (120) determines that, based on the predefined time domain characteristics, if the predetermined sub-characteristic difference in the time domain is less than or equal to the predetermined second threshold value ( ), the corresponding time domain characteristics ( A secondary filtering procedure can be performed to remove time domain features.

[0045] In addition, the processor (120) can remove the frequency domain characteristic (S335) if the lower characteristic is similar to the rotation frequency (1x frequency) characteristic after the secondary filtering procedure is performed (S330-Yes).

[0046] Specifically, the processor (120) determines, after the secondary filtering procedure is performed, that the difference in the preset sub-characteristics in the frequency domain is less than or equal to a preset third threshold value based on the predefined frequency domain characteristics. ), a third-order filtering procedure can be performed to remove the corresponding frequency domain characteristics (nx frequency domain characteristics).

[0047] Thereafter, the processor (120) can perform a response characteristic improvement procedure by utilizing the results obtained up to the third filtering procedure (S340).

[0048] FIG. 4 is a drawing provided for a more detailed description of a method for performing a response characteristic improvement procedure that improves response characteristics by reflecting hardware characteristics of an electric motor according to one embodiment of the present invention.

[0049] When the processor (120) performs a response characteristic improvement procedure that improves the response characteristic by reflecting the hardware characteristics of the motor, it can operate as illustrated in FIG. 4.

[0050] Specifically, the processor (120) can set design response speed requirements consisting of data transmission time conditions and data operation time conditions before improving the response characteristics (S410)( ).

[0051] That is, the processor (120) sets the design response speed requirement ( ), a response characteristic improvement procedure is performed to improve the response characteristics.

[0052] To this end, the processor (120) sets a preset data transmission time condition (t) due to the actual communication speed. 전송 ) does not meet the conditions (S420-No), a first improvement procedure can be performed to reduce the predefined Sampling Interval until it meets the preset data transmission time conditions (S425).

[0053] In addition, the processor (120) has a preset data operation time condition (t) for predicting the operating state of the motor due to the complexity of the hardware characteristics reflected in the operating state prediction. 연산 ) does not meet the condition (S430-No), a secondary improvement procedure is performed to reduce the number of hardware characteristics reflected in the driving state prediction until it meets the preset data operation time condition (S435), and the driving state of the motor can be predicted based on the result of the response characteristic improvement procedure performed thereafter (S440).

[0054] Here, the processor (120) can reduce the number of hardware characteristics by setting a rank for contribution by hardware characteristic and removing the hardware characteristic with the lowest rank of contribution.

[0055] Meanwhile, it goes without saying that the technical idea of ​​the present invention can also be applied to a computer-readable recording medium containing a computer program that performs the functions of the device and method according to the present embodiment. In addition, the technical idea according to various embodiments of the present invention can be implemented in the form of computer-readable code recorded on a computer-readable recording medium. The computer-readable recording medium can be any data storage device that can be read by a computer and store data. For example, the computer-readable recording medium can be a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, an optical disk, a hard disk drive, etc. In addition, the computer-readable code or program stored on the computer-readable recording medium can be transmitted through a network connected between computers.

[0056] In addition, although the preferred embodiments of the present invention have been illustrated and described above, the present invention is not limited to the specific embodiments described above, and various modifications can be made by a person having ordinary skill in the art to which the present invention pertains without departing from the gist of the present invention as claimed in the claims, and such modifications should not be understood individually from the technical idea or prospect of the present invention.

Claims

1. The communication department collects three-axis vibration data of the electric motor; A step of the processor generating a prediction model by learning an initial prediction model that predicts the operating state of the electric motor using three-axis vibration data; and A method for classifying a dynamic electric motor dangerous operating state, comprising: a step of a processor predicting an operating state of an electric motor by utilizing a generated predictive model; 2. In claim 1, The collecting steps are: A method for classifying a dynamic electric motor hazardous driving state, characterized by collecting three-axis vibration data according to a predefined communication speed and sampling rate.

3. In claim 1, The steps to predict the operating status of the motor are: In order to predict the driving state by reflecting the driving characteristics and operating environment characteristics of the motor, a filtering step for filtering the characteristics of the three-axis vibration signal when the response time of the prediction model does not meet the preset design response speed requirements; and A method for classifying a dynamic motor dangerous driving state, characterized by including a response characteristic improvement step for improving the response characteristic by reflecting the hardware characteristics of the motor.

4. In claim 3, The filtering step is, A method for classifying a dangerous driving state of a dynamic motor, characterized in that a first filtering procedure is performed to remove a y-axis or z-axis vibration characteristic when the difference between the vibration characteristic of the y-axis or z-axis signal and the x-axis signal of a three-axis vibration signal is less than a preset first threshold value.

5. In claim 4, The filtering step is, A method for classifying a dynamic electric motor dangerous driving state, characterized in that after a first filtering procedure is performed, a second filtering procedure is performed to remove a time-domain characteristic if a difference between preset sub-characteristics in the time domain is less than or equal to a preset second threshold value based on a predefined time-domain characteristic.

6. In claim 5, The filtering step is, A method for classifying a dynamic motor dangerous driving state, characterized in that after a second filtering procedure is performed, a third filtering procedure is performed to remove the frequency domain characteristic if the difference between preset lower characteristics in the frequency domain is less than or equal to a preset third threshold value based on the frequency domain characteristic defined in advance.

7. In claim 3, The response characteristic improvement step is, A method for classifying a dynamic electric motor hazardous driving state, characterized in that prior to improving the response characteristics, a design response speed requirement condition consisting of a data transmission time condition and a data operation time condition is set.

8. In claim 7, The response characteristic improvement step is, A method for classifying a dynamic motor dangerous driving state, characterized in that, when the actual communication speed does not meet the preset data transmission time condition, a first response characteristic improvement procedure is performed to reduce the predefined Sampling Interval until the preset data transmission time condition is met.

9. In claim 8, The response characteristic improvement step is, A method for classifying a dynamic motor dangerous driving state, characterized in that, when the data calculation time for predicting the driving state of the motor does not meet the preset data calculation time condition due to the complexity of the hardware characteristics reflected in the driving state prediction, a secondary response characteristic improvement procedure is performed to reduce the number of hardware characteristics reflected in the driving state prediction until it meets the preset data calculation time condition.

10. A communication unit that collects three-axis vibration data of the motor; and A dynamic electric motor risk driving state classification system, comprising: a processor for generating a prediction model by learning an initial prediction model predicting the driving state of an electric motor using three-axis vibration data, and predicting the driving state of an electric motor by utilizing the generated prediction model; 11. A step in which the processor learns an initial prediction model predicting the operating state of the electric motor using 3-axis vibration data to create a prediction model; and A processor comprises a step of predicting the operating state of the electric motor by utilizing the generated prediction model; The prediction model is, A method for classifying a dynamic motor dangerous operating state, characterized in that, when the response time of a prediction model does not meet a preset design response speed requirement, a filtering procedure for filtering the characteristics of a vibration signal and a response characteristic improvement procedure for improving the response characteristics by reflecting the hardware characteristics of the motor are performed.

12. A processor that learns an initial prediction model for predicting the operating state of an electric motor using three-axis vibration data to create a prediction model, and uses the created prediction model to predict the operating state of the electric motor; and A storage unit for storing the generated prediction model; The prediction model is, A dynamic motor dangerous driving state classification system characterized in that, when the response time of the prediction model does not meet the preset design response speed requirement, a filtering procedure for filtering the characteristics of a vibration signal and a response characteristic improvement procedure for improving the response characteristics by reflecting the hardware characteristics of the motor are performed.

Images