Apparatus and method for evaluating speed reducer for robot

Abstract

To provide an apparatus and method for evaluating a speed reducer for a robot.SOLUTION: An apparatus for evaluating a speed reducer for a robot of the present invention includes: a motor state input module configured to receive a driving state of a drive motor connected directly to a speed reducer; a storage device; a memory configured to store an executable program; and a processor operatively coupled to the motor state input module, the storage device, and the memory. Therein the processor: receives a drive current and a rotational speed from the motor state input module; establishes big data; stores the big data in the storage device; establishes a learning model for evaluating the speed reducer on the basis of the drive current and the rotational speed stored in the storage device; causes the learning model to perform machine learning; and then diagnoses abnormality of the speed reducer on the basis of the diagnosis learning model by means of the inputted drive current and the inputted rotational speed.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present invention relates to a diagnostic device and method for a robot reducer, and more particularly to a diagnostic device and method for a robot reducer that detects the current and speed applied to a drive motor directly connected to the robot reducer, learns based on machine learning, diagnoses faults, and predicts the service life. [Background technology]

[0002] Generally, an industrial robot is a handling machine that can be programmed with respect to multiple axes of motion, in particular with regard to orientation, position and task progression, and is equipped with tools that serve the purpose of automatically handling objects.

[0003] Such an industrial robot includes a tool for performing a manufacturing process on an object, a wrist on which the tool is mounted, an arm unit and an upper unit for changing the position of the wrist, a rotating unit for rotating around a pivot axis to change the direction in which the wrist faces, a base unit for supporting the rotating unit, and a drive unit for providing a driving force for performing the manufacturing process.

[0004] Here, the driving unit may include a driving motor and a reducer, and the reducer contains lubricating oil for lubrication. The reducer is made of iron, which wears over the operation time of the driving unit and mixes with the lubricating oil. The foreign matter mixed with the lubricating oil may cause a sudden malfunction of the reducer during the manufacturing process, and in order to solve this, all of the industrial robots installed in the process line must be stopped.

[0005] The background art of the present invention is disclosed in Korean Patent Publication No. 10-2016-0083392 (published on July 12, 2016, Industrial Robot System). Summary of the Invention [Problem to be solved by the invention]

[0006] In this way, if a malfunction occurs in the reducer of an industrial robot, the reducer and drive motor are in direct contact with each other, and not only will overheating of the drive motor be transmitted through the reducer housing and affect the reducer, but the high heat of the drive motor itself may damage the encoder of the drive motor. Furthermore, if the high heat of the drive motor is transmitted through the reducer housing, the life of the lubricating grease inside the reducer housing may be shortened.

[0007] The present invention has been made to improve the above-mentioned problems, and an object of the present invention in one aspect is to provide a diagnosis device and method for a robot reducer that detects the current and speed applied to a drive motor directly connected to the robot's reducer, learns based on machine learning, diagnoses faults, and predicts the lifespan. [Means for solving the problem]

[0008] According to one aspect of the present invention, a diagnostic device for a robot reducer includes a motor status input module to which the driving status of a drive motor directly connected to the reducer is input, a storage device, a memory in which an execution program is stored, and a processor operatively coupled to the motor status input module, the storage device, and the memory, wherein the processor receives the driving current and rotational speed from the motor status input module, constructs big data and stores it in the storage device, constructs a learning model for diagnosing the reducer based on the driving current and rotational speed stored in the storage device, performs machine learning, and then diagnoses an abnormality in the reducer based on the diagnostic learning model using the input driving current and rotational speed.

[0009] The present invention is characterized in that it further includes an output module that outputs a diagnosis result, and the processor outputs the diagnosis result of diagnosing an abnormality in the reducer via the output module.

[0010] In the present invention, the motor state input module is characterized in that the drive current and rotation speed are input via a network from a motor driver that drives the drive motor.

[0011] In the present invention, the processor outputs the diagnosis result of the reducer to the motor driver.

[0012] In the present invention, the processor stores the diagnostic results of the reducer in a storage device and periodically improves the diagnostic learning model.

[0013] A method for diagnosing a robot reducer according to one aspect of the present invention includes the steps of: a processor receiving a drive current and rotational speed of a drive motor via a motor status input module; a processor constructing big data based on the input drive current and rotational speed and storing the big data in a storage device; a processor constructing a learning model for diagnosing the reducer based on the drive current and rotational speed and performing machine learning; and a processor diagnosing an abnormality in the reducer based on the diagnostic learning model using the drive current and rotational speed input from the motor status input module.

[0014] The present invention is characterized by further including a step in which the processor outputs a diagnosis result of diagnosing an abnormality in the reducer via the output module.

[0015] The present invention is characterized in that it further includes a step in which the processor outputs a diagnosis result of the reducer to the motor driver.

[0016] The present invention is characterized in that it further includes a step in which the processor stores the diagnosis results of the reducer in a storage device and periodically improves the diagnostic learning model. Effect of the Invention

[0017] A diagnostic device and method for a robot reducer according to one embodiment of the present invention detects the current and speed applied to a drive motor directly connected to a robot's reducer, and learns based on machine learning to diagnose faults and predict the lifespan, as well as providing the diagnostic results to a motor driver to perform protective operations for the drive motor. [Brief description of the drawings]

[0018] [Figure 1] 1 is a block diagram showing a diagnosis device for a robot reducer according to an embodiment of the present invention; [Diagram 2] 4 is a flowchart illustrating a method for diagnosing a robot reducer according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0019] Hereinafter, a diagnosis device and method for a robot reducer according to the present invention will be described with reference to the accompanying drawings. In this process, the thickness of lines and the size of components shown in the drawings may be exaggerated for clarity and convenience of explanation. In addition, the terms used below are defined in consideration of the functions of the present invention, and may vary depending on the intention or practice of a user or operator. Therefore, the definitions of such terms should be based on the contents of this specification as a whole.

[0020] FIG. 1 is a block diagram showing a diagnosis device for a robot reducer according to an embodiment of the present invention.

[0021] As shown in FIG. 1 , a diagnostic device for a robot reducer according to an embodiment of the present invention may include a motor state input module 10 , a storage device 40 , an output module 50 , a memory 20 , and a processor 30 .

[0022] The motor state input module 10 receives the driving state of a drive motor 70 directly connected to a reducer 80 .

[0023] The driving state of the driving motor 70 is determined by inputting a driving current applied to drive the driving motor 70 and a rotation speed of the driving motor 70 .

[0024] At this time, the motor state input module 10 is connected to a motor driver 60 that controls the driving of a driving motor 70 via a network, and receives the driving current and rotation speed from the motor driver 60 .

[0025] The storage device 40 can construct and store big data for the driving current and rotation speed of the driving motor 70 input via the motor status input module 10.

[0026] At this time, the storage device 40 can collect the steady-state driving current and rotation speed for constructing a diagnostic learning model, and construct and store big data.

[0027] The output module 50 provides an interface for outputting the diagnostic results, and can output the diagnostic results not only to a higher-level control device (not shown), but also via a display (not shown).

[0028] The memory 20 can store data and execution programs for diagnosing abnormalities in the robot reducer, and the stored information can be selected by the processor 30 as needed.

[0029] That is, the memory 20 stores various types of data generated during the execution of an operating system and applications (programs or applets) for driving the robot reducer diagnostic device. At this time, the memory 20 is generally referred to as a non-volatile storage device that continues to hold stored information even when power is not supplied, and a volatile storage device that requires power to hold the stored information. Also, the memory 20 can temporarily or permanently store data processed by the processor 30. Here, the memory 20 can include a magnetic storage medium or a flash storage medium in addition to a volatile storage device that requires power to hold the stored information, but the scope of the present invention is not limited thereto.

[0030] The processor 30 is operatively coupled to the motor status input module 10, the storage device 40, the output module 50, and the memory 20, and can copy various programs stored in the memory 20 to the RAM and execute them to perform various operations in order to control the overall operation of the robot reducer diagnostic device.

[0031] In various embodiments, the processor 30 may be implemented as a digital signal processor (DSP) for processing digital signals, a microprocessor, or a time controller (TCON). However, without being limited thereto, the processor 30 may include or be defined as one or more of a central processing unit (CPU), a micro controller unit (MCU), a micro processing unit (MPU), a controller, an application processor (AP), a communication processor (CP), or an ARM processor. Furthermore, the processor 30 may be implemented as a system on chip (SoC) or large scale integration (LSI) having a built-in processing algorithm, or may be implemented in the form of a field programmable gate array (FPGA).

[0032] That is, the processor 30 runs an execution program for diagnosis stored in the memory 20, receives the driving current and rotation speed from the motor status input module 10, constructs big data, and stores it in the storage device 40.

[0033] Thereafter, the processor 30 constructs a learning model for diagnosing the reducer 80 based on an artificial neural network based on the driving current and rotational speed stored in the storage device 40, performs machine learning, and then diagnoses abnormalities in the reducer based on the diagnostic learning model using the input driving current and rotational speed, and predicts its lifespan.

[0034] Additionally, the processor 30 can store the diagnostic results of the reducer 80 in the storage device 40 and periodically improve the diagnostic learning model.

[0035] Here, an artificial neural network (ANN) is a machine learning algorithm that mathematically models the learning method of the brain, in which information is stored through a process of communication via neurons and synapses that connect neurons. In an artificial neural network, neurons are defined as nodes and synapses as weights.

[0036] The artificial neural network includes any one of a deep neural network (DNN), a recurrent neural network (RNN) and a convolutional neural network (CNN). In addition, the artificial neural network is generated by interconnecting at least one of a deep neural network (DNN) and a recurrent neural network (RNN). In addition to the above-mentioned artificial neural network, various deep learning algorithms can be applied. A deep neural network (DNN) is an artificial neural network composed of multiple hidden layers, and a recurrent neural network (RNN) is a neural network having a cyclic structure different from a deep neural network (DNN), and stores data of the hidden layer and uses it recurrently for the next learning, thereby memorizing past information and reflecting it in the learning, and therefore can be a neural network suitable for analyzing time series data.

[0037] In this way, the processor 30 can diagnose abnormalities in the reducer 80 based on a diagnostic learning model using the input driving current and rotational speed, and output the diagnosis results via the output module 50 to a higher-level control device or a display.

[0038] Furthermore, the processor 30 can output the diagnosis results of the reducer 80 to the motor driver 60, so that the motor driver 60 can control the drive motor 70 depending on the abnormal state and lifespan of the reducer 80 to perform protective operation.

[0039] As described above, the robot reducer diagnosis device according to the embodiment of the present invention detects the current and speed applied to the drive motor directly connected to the robot's reducer, learns based on machine learning, diagnoses faults, predicts the life span, and not only provides the diagnosis results to the motor driver to perform protective operations for the drive motor.

[0040] FIG. 2 is a flowchart illustrating a method for diagnosing a robot reducer according to an embodiment of the present invention.

[0041] As shown in FIG. 2, in the diagnostic method for a robot reducer according to one embodiment of the present invention, first, the drive current and rotation speed of the drive motor 70 are input to the processor 30 via the motor state input module 10 (S10).

[0042] At this time, the motor state input module 10 is connected to a motor driver 60 that controls the driving of a driving motor 70 via a network, and receives the driving current and rotation speed from the motor driver 60 .

[0043] When the driving current and rotation speed of the driving motor 70 are input in step S10, the processor 30 constructs big data based on the input driving current and rotation speed and stores the big data in the storage device 40 (S20).

[0044] After constructing big data based on the driving current and rotational speed in step S20, the processor 30 constructs a learning model for diagnosing the reducer 80 based on an artificial neural network and performs machine learning on the big data stored in the storage device 40 (S30).

[0045] Here, an artificial neural network (ANN) is a machine learning algorithm that mathematically models the learning method of the brain, in which information is stored through a process of communication via neurons and synapses that connect neurons. In an artificial neural network, neurons are defined as nodes and synapses as weights.

[0046] After performing machine learning in step S30, the processor 30 diagnoses an abnormality in the reducer 80 based on the diagnostic learning model for the drive current and rotation speed input via the motor state input module 10 (S40).

[0047] After diagnosing the abnormality of the reducer 80 in step S40, the processor 30 outputs the diagnosis result of the abnormality of the reducer 80 to a higher-level control device (not shown) via the output module 50, or outputs the diagnosis result via a display (not shown) (S50).

[0048] The processor 50 may also periodically improve the diagnostic learning model after storing the diagnostic results in the storage device 40 .

[0049] On the other hand, the processor 50 outputs the diagnosis result of the reducer 80 to the motor driver 60, and the motor driver 60 controls the drive motor 70 depending on the abnormal state and lifespan of the reducer 80 to perform protective operation.

[0050] As described above, according to the diagnostic method for a robot reducer according to an embodiment of the present invention, the current and speed applied to the drive motor directly connected to the robot reducer are detected, and the method learns based on machine learning to diagnose faults and predict the life span, and also provides the diagnosis results to the motor driver so that protective operations can be performed for the drive motor.

[0051] The implementations described herein may be implemented, for example, as a method or process, an apparatus, a software program, a data stream, or a signal. Even if discussed only in the context of a single implementation (e.g., discussed only as a method), the implementation of the discussed features may also be implemented in other forms (e.g., apparatus or program). An apparatus may be implemented in appropriate hardware, software, firmware, and the like. A method may be implemented in an apparatus, such as, for example, a processor, which generally refers to a processing device including a computer, a microprocessor, an integrated circuit, or a programmable logic device. Processors further include communication devices, such as computers, mobile phones, portable / personal digital assistants ("PDAs"), and other devices that facilitate communication of information between end users.

[0052] Although the present invention has been described with reference to the embodiments shown in the drawings, this is by way of example only, and those having ordinary skill in the art will recognize that various modifications and equivalent alternative embodiments are possible.

[0053] Therefore, the true technical scope of the present invention should be defined by the following claims.

Claims

1. a motor status input module to which a driving status of a drive motor directly connected to the reducer is input; A storage device; A memory in which an executable program is stored; a processor operatively coupled to the motor status input module, the storage device, and the memory; The processor receives the drive current and rotation speed from the motor state input module, constructs big data and stores it in the storage device, constructs a learning model for diagnosing the reducer based on the drive current and the rotation speed stored in the storage device, performs machine learning, and then diagnoses an abnormality in the reducer based on the diagnostic learning model using the input drive current and rotation speed.

2. and an output module for outputting the diagnosis result. The diagnostic device for a robot reducer according to claim 1 , wherein the processor outputs a diagnosis result of diagnosing an abnormality in the reducer via the output module.

3. 2. The diagnostic device for a robot reducer according to claim 1, wherein the motor state input module receives the drive current and the rotation speed from a motor driver that drives the drive motor via a network.

4. The diagnostic device for a robot reducer according to claim 1 , wherein the processor outputs a diagnosis result of the reducer to a motor driver.

5. The diagnostic device for a robot reducer according to claim 1 , wherein the processor stores a diagnostic result of the reducer in the storage device and periodically improves a diagnostic learning model.

6. A step in which a drive current and a rotation speed of a drive motor are input to a processor via a motor state input module; The processor constructs big data based on the input driving current and the rotation speed, and stores the big data in a storage device; The processor constructs a learning model for diagnosing the reducer based on the drive current and the rotation speed, and performs machine learning; A method for diagnosing a robot reducer, comprising: a step in which the processor diagnoses an abnormality in the reducer based on a diagnostic learning model using the drive current and the rotational speed input from the motor state input module.

7. The method for diagnosing a robot reducer according to claim 6, further comprising a step of the processor outputting a diagnosis result of diagnosing an abnormality in the reducer via an output module.

8. The method for diagnosing a robot reducer according to claim 6, further comprising a step of the processor outputting a diagnosis result of the reducer to a motor driver.

9. The method for diagnosing a robot reducer according to claim 6, further comprising a step of: the processor storing a diagnosis result of the reducer in the storage device and periodically improving a diagnostic learning model.

Images