A harmonic reducer and its condition monitoring method

Abstract

This application relates to a harmonic reducer and its condition monitoring method. The scheme includes a hollow wave generator, one end of which receives power from the robot body via a connecting bevel gear; a flexible bearing, mounted on the outer ring of the hollow wave generator, capable of deforming with the hollow wave generator; a flexible wheel, mounted on the outer ring of the flexible bearing; a moving rigid wheel, its inner ring meshing with a portion of the outer ring of the flexible wheel, and the moving rigid wheel connecting to the inner ring of a cross bearing; a fixed rigid wheel, its inner ring meshing with the remaining portion of the outer ring of the flexible wheel, and the fixed rigid wheel covering the moving rigid wheel and the hollow wave generator and connecting to the outer ring of the cross bearing, for connecting to the end effector of the robot body; and a cross bearing, its inner ring serving as the power output. This application can achieve an inner diameter greater than 55mm and can simultaneously monitor the condition of the harmonic reducer to promptly detect problems and issue alarms.

Description

Technical Field

[0001] This application relates to a speed reducer, specifically a harmonic speed reducer and its condition monitoring method. Background Technology

[0002] A harmonic reducer is a speed reduction device mainly composed of four basic components: a wave generator, a flexible gear, a flexible bearing, and a rigid gear. It is developed based on the principle of planetary gear transmission. By assembling a wave generator with a flexible bearing, the flexible gear produces controllable elastic deformation and meshes with the rigid gear to transmit motion and power.

[0003] The sixth axis of the end effector of 10KG welding robots on the market uses a gearbox structure. The cable guide hole in the middle can have an inner diameter of 55mm, which is very important for the passage of the welding torch and cable. However, the gearbox's reduction device has poor transmission accuracy, with the transmission error only controlled within about 3 arc minutes. Moreover, the noise and vibration are also poor, resulting in a very poor user experience. Harmonic reducers, on the other hand, have the characteristics of high transmission accuracy, low noise, low vibration, and high rigidity, which can solve this problem. However, harmonic reducers cannot be well integrated into the robot's end effector. Furthermore, current technology cannot effectively monitor the status of harmonic reducers, making it impossible to detect problems in a timely manner.

[0004] Therefore, there is an urgent need for an ultra-large inner bore double rigid wheel harmonic reducer capable of hollowing out the wires to solve the problems existing in the current technology. Summary of the Invention

[0005] The purpose of this application is to address the aforementioned problems in the prior art by providing a harmonic reducer and its condition monitoring method.

[0006] To achieve the first objective of the above application, this application adopts the following technical solution: a harmonic reducer for installation at the end of a robot body, comprising:

[0007] A hollow wave generator, one end face of which receives power from the robot body via a bevel gear;

[0008] A flexible bearing is installed on the outer ring of the hollow wave generator, and the flexible bearing can deform with the hollow wave generator.

[0009] The flexible wheel is installed on the outer ring of the flexible bearing;

[0010] A moving rigid wheel, the inner ring of which meshes with part of the outer ring of a flexible wheel, and the moving rigid wheel is connected to the inner ring of a cross bearing;

[0011] The fixed rigid wheel has its inner ring meshing with the remaining outer ring of the flexible wheel, and the fixed rigid wheel covers the moving rigid wheel and the hollow wave generator and is connected to the outer ring of the cross bearing for connecting to the end of the robot body.

[0012] Cross bearings, with the inner ring serving as the power output; and the inner diameter of the cross bearing's inner ring is greater than or equal to 55mm;

[0013] The number of teeth on the flexible wheel differs from that on the moving rigid wheel by 2, which enables misalignment motion, and the number of teeth on the flexible wheel is the same as that on the fixed rigid wheel.

[0014] Working principle and beneficial effects: 1. Compared with the prior art, this application utilizes the zero backlash design of harmonic gears, which significantly reduces the transmission error from 3 arc minutes to 1 arc minute, making the welding pattern of the robot body (welding robot) more accurate.

[0015] 2. This application utilizes the characteristics of harmonic reducers to reduce noise. For example, when the input speed is 1000 RPM, the noise can be reduced from 70 dB to 62 dB. At the same time, the service life (time to maintain accuracy) is also significantly improved, up to 8000 hours.

[0016] 3. The hollow wave generator of this application receives power from the motor (which can utilize a worm gear transmission structure) on the robot body, driving the flexible bearing to rotate. The flexible bearing forces the flexure wheel to deform. The number of teeth on the flexure wheel differs from that on the moving rigid wheel by 2, generating a misaligned motion to achieve power output. The number of teeth on the flexure wheel is the same as that on the fixed rigid wheel, so it does not move and serves as a fixed end connected to the robot body. It is precisely because of the design of the moving rigid wheel, the fixed rigid wheel, and the crossed bearing that an inner diameter of 55mm or more is possible, allowing for proper installation on the robot's end effector, providing both fixation and power reception.

[0017] Furthermore, a deep groove ball bearing is installed between the hollow wave generator and the fixed rigid wheel, and a deep groove ball bearing is also installed between the moving rigid wheel and the flexible wheel. Thus, the use of deep groove ball bearings helps optimize the performance of the mechanical system, improve transmission efficiency, reduce wear, extend the service life of components, and provide smooth and reliable operation. Preferably, a 61813 deep groove ball bearing is selected between the hollow wave generator and the fixed rigid wheel, and a 61812 deep groove ball bearing is selected between the moving rigid wheel and the flexible wheel.

[0018] Secondly, in order to achieve the second objective of the above application, this application adopts the following technical solution: a harmonic reducer condition monitoring method for detecting the condition of the aforementioned harmonic reducer, comprising the following steps:

[0019] S00: Real-time acquisition of temperature signals at the meshing point of the flexible wheel and the moving rigid wheel, and real-time acquisition of vibration signals of the inner ring of the cross bearing;

[0020] S10. Preprocess the temperature signal and vibration signal to obtain processed data to remove noise; establish a benchmark model based on historical normal operation data. This model represents the feature distribution and relationship under normal operation conditions. The benchmark model is used to compare the input features with the benchmark features contained in the benchmark model and output the feature data that does not conform to the benchmark.

[0021] S20. Extract relevant features of temperature and vibration signals from the processed data;

[0022] Among them, the relevant characteristics of temperature signals include average temperature, rate of temperature change, and temperature gradient; the relevant characteristics of vibration signals include vibration frequency and vibration amplitude.

[0023] S30. Input the obtained relevant features into the benchmark model to obtain the output results. The output results include the relevant features that do not conform to the benchmark features and the relevant feature data that conform to the benchmark features.

[0024] S40. Based on the output, perform anomaly analysis on the features that do not conform to the baseline features to obtain the anomaly type and corresponding severity data.

[0025] S50. Issue an alarm based on the anomaly type and its corresponding severity data.

[0026] Compared with existing technologies, the beneficial effects are as follows:

[0027] This application, by simultaneously acquiring temperature and vibration signals and utilizing information from multiple sensors, helps to gain a more comprehensive understanding of the operating status of the harmonic reducer; preprocessing the temperature and vibration signals to remove noise helps to improve data quality and reduce interference with the model; using historical normal operation data to establish a benchmark model helps to capture the characteristic distribution and relationships under normal operating conditions; extracting relevant features from the temperature and vibration signals helps to describe the system's operating status and improves sensitivity to anomalies; analyzing anomalies through S40 and S50 and issuing alarms based on anomaly type and severity helps to take timely measures to prevent potential problems from worsening.

[0028] Furthermore, in step S10, preprocessing includes filtering and denoising, as well as normalization. This removes noise and ensures signal stability, while normalizing the temperature and vibration data converts them to similar scales to facilitate subsequent feature extraction and modeling processes.

[0029] Furthermore, in step S10, the specific steps for establishing the baseline model include:

[0030] Collect historical temperature and vibration signals of the harmonic reducer under normal operating conditions;

[0031] Historical temperature and vibration signals are preprocessed and merged to obtain training and validation datasets;

[0032] Extract relevant features of historical temperature signals and historical vibration signals from the training dataset and add labels to indicate that they are in normal operating condition;

[0033] The regression model is trained using the labeled training dataset so that it learns the relationship between the relevant features of the two signals and the normal operating state.

[0034] The regression model was evaluated using the validation dataset to obtain the evaluation results;

[0035] Based on the evaluation results, the model is saved after the evaluation is passed, and a baseline model is obtained.

[0036] In this way, by using historical data to build a benchmark model, abnormalities in the operation of harmonic reducers can be quickly and accurately identified.

[0037] Furthermore, if the evaluation result is unsatisfactory, the regression model is trained and evaluated again by adjusting the hyperparameters or reselecting the model until a benchmark model that passes the evaluation is obtained.

[0038] In this way, by continuously debugging until the most suitable regression model is found, the waste of computing power caused by constantly adjusting hyperparameters can be reduced.

[0039] Furthermore, in step S40, feature importance analysis is performed on the relevant features to obtain the anomaly types and corresponding severity data of the features that do not conform to the baseline features.

[0040] Thus, feature importance analysis provides crucial information for system anomaly detection, helping to more accurately and efficiently identify the type and severity of anomalies, thereby optimizing maintenance processes and system reliability. This approach enables the system to respond to potential problems more promptly and intelligently, improving equipment stability and lifespan.

[0041] Furthermore, feature importance analysis includes: using the feature importance scores built into the tree model as the contribution of relevant features to the baseline model; or, randomly arranging the feature importance scores of relevant features and observing the performance changes of the baseline model to obtain the importance of each relevant feature; or, using the absolute value of the coefficients in the linear module as a measure of feature importance.

[0042] Thus, mature methods such as tree models, random permutations, and linear models can be used to quickly determine the contribution of each feature in the model to the model's prediction or classification.

[0043] Furthermore, in step S00, a temperature sensor is installed at the meshing point of the flexible wheel and the moving rigid wheel to obtain a temperature signal, and a vibration sensor is installed on the inner ring of the cross bearing to obtain a vibration signal. The temperature sensor and the vibration sensor are connected to a control module, which analyzes and alarms the signal.

[0044] Thus, the control module can acquire temperature and vibration signals and then run the scheme of this application to perform monitoring and alarm.

[0045] Furthermore, the control module includes a remote detection module, through which remote detection functions are implemented.

[0046] This enables remote monitoring, allowing staff to promptly identify problems. Attached Figure Description

[0047] Figure 1 This is a structural schematic diagram of an embodiment of this application.

[0048] Figure 2 This is a flowchart of an embodiment of this application;

[0049] Figure 3 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0050] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0051] Those skilled in the art should understand that, in the disclosure of this application, the terms "longitudinal," "lateral," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, the above terms should not be construed as limitations on this application.

[0052] Example 1

[0053] like Figure 1 As shown, this harmonic reducer, used for installation at the end of a robot body, includes:

[0054] Hollow wave generator 1, one end face of which receives power from the robot body through a bevel gear;

[0055] In this embodiment, the inner diameter of the hollow wave generator 1 is also 55mm or more. The hollow wave generator 1 is elliptical, just like existing wave generators. Power is transmitted to the hollow wave generator 1 through the motor on the robot body using a worm gear transmission structure.

[0056] A flexible bearing 2 is installed on the outer ring of the hollow wave generator 1. The flexible bearing 2 can deform with the hollow wave generator 1.

[0057] In this embodiment, since the flexible bearing 2 is installed on the hollow wave generator 1, the flexible bearing 2 will also be deformed into an ellipse, and the flexible bearing 2 is installed in the flexible inner hole of the flexible wheel 3.

[0058] Flexible wheel 3 is installed on the outer ring of flexible bearing 2;

[0059] In this embodiment, the outer ring of the flexible wheel 3 is machined with external teeth, which can simultaneously mesh with the internal teeth of the moving rigid wheel 4 and the fixed rigid wheel 5.

[0060] The inner ring of the moving rigid wheel 4 meshes with a portion of the outer ring of the flexible wheel 3, and the moving rigid wheel 4 is connected to the inner ring of the cross bearing 6;

[0061] In this embodiment, the end face of the moving rigid wheel 4 is connected and fixed to the inner ring end face of the cross bearing 6 by a threaded fastener 8.

[0062] The fixed rigid wheel 5 has its inner ring meshing with the remaining outer ring of the flexible wheel 3, and the fixed rigid wheel 5 covers the moving rigid wheel 4 and the hollow wave generator 1 and is connected to the outer ring of the cross bearing 6 for connecting the end of the robot body.

[0063] In this embodiment, the end face of the fixed rigid wheel 5 is connected and fixed to the outer ring end face of the cross bearing 6 by a threaded fastener 8.

[0064] Cross bearing 6, with the inner ring serving as the power output; and the inner diameter of the inner ring of cross bearing 6 is greater than or equal to 55mm;

[0065] Preferably, the inner diameter of the inner ring of the cross bearing 6 is 55.1 mm. This is of great significance for welding torches and cables.

[0066] The number of teeth on the flexible wheel 3 differs from the number of teeth on the moving rigid wheel 4 by 2, so as to generate misaligned motion, and the number of teeth on the flexible wheel 3 is the same as the number of teeth on the fixed rigid wheel 5.

[0067] In this embodiment, a deep groove ball bearing 7 is provided between the hollow wave generator 1 and the fixed rigid wheel 5, and a deep groove ball bearing 7 is also provided between the moving rigid wheel and the flexible wheel 3. Preferably, a 61813 deep groove ball bearing is selected between the hollow wave generator 1 and the fixed rigid wheel 5, and a 61812 deep groove ball bearing is selected between the moving rigid wheel and the flexible wheel 3.

[0068] Example 2

[0069] To achieve state detection of the harmonic reducer in Example 1, such as Figure 2 As shown, this embodiment proposes a harmonic reducer condition monitoring method for detecting the condition of a harmonic reducer as described above, including the following steps:

[0070] S00. The temperature signal at the meshing point of the flexible wheel 3 and the moving rigid wheel 4 is acquired in real time, and the vibration signal of the inner ring of the cross bearing 6 is acquired in real time.

[0071] Temperature signals are acquired through a temperature sensor installed at the meshing point of the flexible wheel 3 and the moving rigid wheel 4, and vibration signals are acquired through a vibration sensor installed on the inner ring of the cross bearing 6. The temperature and vibration sensors are connected to a control module, which analyzes and generates alarms. The control module includes a remote detection module, which enables remote detection functionality.

[0072] S10. Preprocess the temperature signal and vibration signal to obtain processed data to remove noise; establish a benchmark model based on historical normal operation data. This model represents the feature distribution and relationship under normal operation conditions. The benchmark model is used to compare the input features with the benchmark features contained in the benchmark model and output the feature data that does not conform to the benchmark.

[0073] In this embodiment, preprocessing includes filtering and noise reduction, as well as normalization. Noise can be removed to ensure signal stability, while normalization of temperature and vibration data converts them to similar scales to facilitate subsequent feature extraction and modeling processes. Preferably, digital filtering techniques (such as low-pass filters) are used to remove high-frequency noise and ensure signal stability. Noise removal can be achieved using digital filters, wavelet transforms, or other signal processing techniques.

[0074] In this embodiment, the specific steps for establishing the baseline model include:

[0075] S01. Collect historical temperature and vibration signals under normal operating conditions of the harmonic reducer;

[0076] S02. Preprocess and merge historical temperature signals and historical vibration signals to obtain training datasets and validation datasets;

[0077] S03. Extract relevant features of historical temperature signals and historical vibration signals from the training dataset and add labels to indicate that they are in normal operating condition.

[0078] S04. Train the regression model using the labeled training dataset so that the regression model learns the relationship between the relevant features of the two signals and the normal operating state.

[0079] S05. Use the validation dataset to evaluate the regression model and obtain the evaluation results;

[0080] In this embodiment, regression models such as linear regression, support vector regression, and decision tree regression can all be used in the solution of this application.

[0081] S06. Based on the evaluation results, save the model after the evaluation is passed to obtain the baseline model; if the evaluation results are not passed, train and evaluate the regression model by adjusting the hyperparameters or reselecting the regression model until a baseline model that passes the evaluation is obtained.

[0082] In this embodiment, the model is evaluated using a validation dataset to examine its generalization performance on new data. Evaluation metrics may include mean squared error, R-squared value, etc.

[0083] Preferably, the model can be built using Python and the scikit-learn library, as shown in the following code:

[0084]

[0085]

[0086] S20. Extract relevant features of temperature and vibration signals from the processed data;

[0087] Among them, the relevant characteristics of temperature signals include average temperature, rate of temperature change, and temperature gradient; the relevant characteristics of vibration signals include vibration frequency and vibration amplitude.

[0088] In this embodiment, when extracting temperature features from time-series data, various statistical methods can be used to capture the trends, periodicity, and anomalies of the data. For example:

[0089] Mean Temperature: This is calculated by summing all temperature values ​​in a time series and then dividing by the total number of data points. It can represent the overall trend of temperature.

[0090] Temperature Rate of Change: Calculates the rate of temperature change between adjacent time points, which can be obtained through a difference operation. It is used to capture rapid temperature changes and helps detect abnormal temperature fluctuations.

[0091] Temperature gradient: Uses linear regression analysis to estimate the trend of temperature change over time. It represents the rate and direction of temperature change and helps to identify potential trends.

[0092] Temperature Fluctuation Amplitude: Calculated by the standard deviation of a temperature time series, or the standard deviation within a sliding window. It represents the magnitude of temperature change and is used to detect abnormal fluctuations.

[0093] Temperature distribution parameters: These parameters allow you to calculate statistical parameters of the temperature distribution, such as skewness and kurtosis. They provide information about the shape of the temperature distribution and help determine if there are any abnormal distributions.

[0094] Temperature periodicity analysis: This method uses frequency domain analysis techniques such as Fourier transform or wavelet analysis to identify periodic components in temperature data. It is used to discover potential seasonal or periodic variations, helping to predict future temperature trends.

[0095] Temperature Lag Effect: This involves analyzing the correlation between temperature and itself at different time lags using cross-correlation or autocorrelation functions. Understanding the time lag effect of temperature helps identify lag patterns in temperature changes.

[0096] In summary, the above methods can be used to obtain the relevant features of temperature signals. Vibration feature extraction can be achieved using methods such as:

[0097] Mean Vibration: Calculates the average value of the vibration signal. It represents the overall level of the vibration signal.

[0098] Vibration Standard Deviation: Calculates the standard deviation of a vibration signal. It represents the intensity of the vibration signal fluctuation and is used to detect the amplitude of vibration changes.

[0099] Peak Value: Calculates the maximum or minimum value of the vibration signal. It represents the maximum amplitude of the vibration signal and is used to detect the extreme values ​​of the vibration signal.

[0100] Impulse Factor: Calculates the ratio of the peak value to the root mean square value of a vibration signal. It describes the pulse nature of a vibration signal and is helpful in detecting impact vibrations.

[0101] Kurtosis: Calculates the kurtosis of a vibration signal, i.e., the shape of the tail of the vibration signal distribution. Used to detect whether there are abnormal peaks in the vibration signal.

[0102] Frequency-related features: Using methods such as Fourier transform or wavelet transform, the frequency domain features of the vibration signal, such as the dominant frequency and band energy, are extracted. Revealing the structure of the vibration signal in the frequency domain helps in identifying the frequency components of the vibration.

[0103] Pulse counting performs thresholding on vibration signals and counts the number of pulses exceeding a threshold. It is used to detect pulse events in vibration signals and helps identify abnormal impacts.

[0104] Envelope analysis: This involves analyzing the envelope of a vibration signal to extract its characteristics, such as the mean and standard deviation. It reveals the overall trend of the vibration signal and helps in detecting periodic changes.

[0105] Preferably, different features can be selected according to specific circumstances in practical applications (if different needs exist). Simultaneously, techniques such as sliding windows can be used to extract features from local areas to better capture the dynamic characteristics of vibration and temperature signals.

[0106] S30. Input the obtained relevant features into the benchmark model to obtain the output results. The output results include the relevant features that do not conform to the benchmark features and the relevant feature data that conform to the benchmark features.

[0107] S40. Based on the output, perform anomaly analysis on the features that do not conform to the baseline features to obtain the anomaly type and corresponding severity data.

[0108] In this embodiment, feature importance analysis is performed on relevant features to obtain the anomaly types and corresponding severity data of features that do not conform to the baseline features.

[0109] Preferably, feature importance analysis can be performed using the following scheme:

[0110] 1. Feature Importance Analysis Based on the Model:

[0111] 1.1 For tree models (such as decision trees and random forests), the built-in feature importance scores can be used. These scores represent the degree of contribution of a feature to the model.

[0112] 1.2 For linear models (such as linear regression), the absolute value of the coefficients can be used as a measure of feature importance.

[0113] 1.3 For models such as support vector machines, the weights of features can be used to evaluate their importance.

[0114] 2. Permutation Feature Importance:

[0115] 2.1 Observe the changes in model performance by randomly permuting the values ​​of a certain feature. If random permutation of a certain feature leads to a significant decrease in performance, it indicates that the feature has a significant impact on the model's performance.

[0116] 3. SHAP (SHapley Additive exPlanations) value analysis:

[0117] 3.1 SHAP value is a feature importance analysis method based on game theory, which can be used to interpret model output. The SHAP value represents the contribution of each feature to the model output.

[0118] 4. LASSO regression (L1 regularization):

[0119] 4.1 L1 regularization tends to reduce the coefficients of certain features to zero, thereby automatically selecting important features.

[0120] In summary, the results of multiple methods can be combined to obtain more comprehensive information on feature importance. Once the importance is obtained, the corresponding severity level can be derived.

[0121] Preferably, this application uses a random forest regression model to train the model and calculate the original mean squared error. Then, the `permutation_importance` function is used to perform feature ranking importance analysis, obtaining the importance score for each feature, which is then visualized using a bar chart. The code is as follows:

[0122]

[0123]

[0124] S50. Issue an alarm based on the anomaly type and its corresponding severity data.

[0125] In this embodiment, if the anomaly diagnosis confirms a serious problem, the alarm system is triggered, sending an alert to the operator or maintenance personnel. The alarm information should include key information such as the anomaly type, time of occurrence, and location.

[0126] Example 3

[0127] This embodiment also provides an electronic device, see reference. Figure 3It includes a memory 404 and a processor 402, wherein the memory 404 stores a computer program and the processor 402 is configured to run the computer program to perform the steps in any of the above method embodiments.

[0128] Specifically, the processor 402 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0129] The memory 404 may include a mass storage device for data or instructions. For example, and not limitingly, the memory 404 may include a hard disk drive (HDD), a floppy disk drive, a solid-state drive (SSD), flash memory, an optical disk drive, a magneto-optical disk drive, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. Where appropriate, the memory 404 may include removable or non-removable (or fixed) media. Where appropriate, the memory 404 may be internal or external to a data processing device. In a particular embodiment, the memory 404 is non-volatile memory. In a particular embodiment, the memory 404 includes read-only memory (ROM) and random access memory (RAM). Where appropriate, the ROM may be a mask-programmed ROM, a programmable read-only memory (PROM), an erasable read-only memory (EPROM), an electrically erasable read-only memory (EEPROM), an electrically alterable read-only memory (EAROM), or flash memory, or a combination of two or more of these. Where appropriate, the RAM can be Static Random-Access Memory (SRAM) or Dynamic Random-Access Memory (DRAM). DRAM can be Fast Page Mode Dynamic Random-Access Memory (FPMDRAM), Extended Data Out Dynamic Random-Access Memory (EDODRAM), Synchronous Dynamic Random-Access Memory (SDRAM), etc.

[0130] The memory 404 can be used to store or cache various data files that need to be processed and / or communicated, as well as possible computer program instructions executed by the processor 402.

[0131] The processor 402 reads and executes computer program instructions stored in the memory 404 to implement any of the harmonic reducer status monitoring methods in the above embodiments.

[0132] Optionally, the electronic device may further include a transmission device 406 and an input / output device 408, wherein the transmission device 406 is connected to the processor 402, and the input / output device 408 is connected to the processor 402.

[0133] The transmission device 406 can be used to receive or send data via a network. Specific examples of the network described above may include wired or wireless networks provided by the communication provider of the electronic device. In one example, the transmission device includes a Network Interface Controller (NIC), which can connect to other network devices via a base station to communicate with the Internet. In another example, the transmission device 406 may be a Radio Frequency (RF) module used for wireless communication with the Internet.

[0134] Input / output device 408 is used to input or output information. In this embodiment, the input information may be monitoring instructions, etc., and the output information may be abnormal information and alarm information, etc.

[0135] Example 4

[0136] This embodiment also provides a readable storage medium storing a computer program, the computer program including program code for controlling a process to execute the process, the process including the harmonic reducer state monitoring method according to Embodiment 1.

[0137] It should be noted that the specific examples in this embodiment can refer to the examples described in the above embodiments and optional implementations, and will not be repeated here.

[0138] Generally, various embodiments can be implemented in hardware or dedicated circuitry, software, logic, or any combination thereof. Some aspects of the invention can be implemented in hardware, while others can be implemented by firmware or software executed by a controller, microprocessor, or other computing device, but the invention is not limited thereto. Although various aspects of the invention may be shown and described as block diagrams, flowcharts, or using some other graphical representation, it should be understood that, by way of non-limiting example, these blocks, apparatuses, systems, techniques, or methods described herein can be implemented in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0139] Embodiments of the present invention can be implemented by computer software, which may be executable by a data processor of a mobile device, such as a processor entity, or by hardware, or by a combination of software and hardware. Computer software or programs (also referred to as program products), including software routines, applets, and / or macros, can be stored in any device-readable data storage medium, and they include program instructions for performing specific tasks. A computer program product may include one or more computer-executable components configured to perform embodiments when the program is run. One or more computer-executable components may be at least one piece of software code or a portion thereof. Additionally, it should be noted that any block in the logical flow of the figures may represent a program step, or interconnected logic circuits, blocks, and functions, or a combination of program steps and logic circuits, blocks, and functions. The software may be stored on physical media such as memory chips or blocks of storage implemented within a processor, magnetic media such as hard disks or floppy disks, and optical media such as, for example, DVDs and their data variants, CDs, etc. The physical medium is a non-transient medium.

[0140] Those skilled in the art should understand that the technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments have been described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0141] The above embodiments are merely illustrative of several implementation methods of this application, and their descriptions are relatively specific and detailed, but they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A harmonic reducer condition monitoring method, used to perform condition monitoring on a harmonic reducer, the harmonic reducer being installed at the end of a robot body, comprising: A hollow wave generator, one end face of which receives power from the robot body via a bevel gear; a flexible bearing, mounted on the outer ring of the hollow wave generator, capable of deforming with the hollow wave generator; a flexible wheel, mounted on the outer ring of the flexible bearing; a moving rigid wheel, the inner ring of which meshes with a portion of the outer ring of the flexible wheel, and the moving rigid wheel is connected to the inner ring of the cross bearing; a fixed rigid wheel, the inner ring of which meshes with the remaining portion of the outer ring of the flexible wheel, and the fixed rigid wheel covers the moving rigid wheel and the hollow wave generator and is connected to the outer ring of the cross bearing. A ring connection is used to connect the end of the robot body; a cross bearing is used, with the inner ring serving as the power output; and the inner diameter of the cross bearing's inner ring is greater than or equal to 55mm; wherein, the number of teeth on the flexible wheel differs from the number of teeth on the moving rigid wheel by 2, so as to generate misalignment movement, and the number of teeth on the flexible wheel is the same as the number of teeth on the fixed rigid wheel; a deep groove ball bearing is provided between the hollow wave generator and the fixed rigid wheel, and a deep groove ball bearing is also provided between the moving rigid wheel and the flexible wheel; characterized in that the state monitoring method includes the following steps: S00. The temperature signal at the meshing point of the flexible wheel and the moving rigid wheel is acquired in real time, and the vibration signal of the inner ring of the cross bearing is acquired in real time. The temperature signal is acquired by setting a temperature sensor at the meshing point of the flexible wheel and the moving rigid wheel, and the vibration signal is acquired by setting a vibration sensor on the inner ring of the cross bearing. The temperature sensor and the vibration sensor are connected to a control module, and the control module performs analysis and alarm. S10. The temperature signal and the vibration signal are preprocessed to obtain processed data to remove noise. The preprocessing includes filtering and normalization. A benchmark model is established based on historical normal operation data. This model represents the feature distribution and relationship under normal operation conditions. The benchmark model is used to compare the input features with the benchmark features contained in the benchmark model and output the feature data that does not conform to the benchmark. The specific steps for establishing the benchmark model include: Collect historical temperature and vibration signals of the harmonic reducer under normal operating conditions; Historical temperature and vibration signals are preprocessed and merged to obtain training and validation datasets; Extract relevant features of historical temperature signals and historical vibration signals from the training dataset and add labels to indicate that they are in normal operating condition; The regression model is trained using the labeled training dataset so that it learns the relationship between the relevant features of two signals and the normal operating state. The regression model is evaluated using the validation dataset to obtain the evaluation results; Based on the evaluation results, the model is saved after the evaluation is passed, thus obtaining the benchmark model; S20. Extract the relevant features of the temperature signal and the vibration signal from the processed data; The temperature signal's relevant features include average temperature, temperature change rate, and temperature gradient; the vibration signal's relevant features include vibration frequency and vibration amplitude. S30. Input the obtained relevant features into the benchmark model to obtain the output result. The output result includes the relevant features that do not conform to the benchmark features and the relevant feature data that conform to the benchmark features. S40. Based on the output, perform anomaly analysis on the features that do not conform to the baseline features to obtain the anomaly type and corresponding severity data. S50. Issue an alarm based on the anomaly type and the corresponding severity data.

2. The harmonic reducer condition monitoring method according to claim 1, characterized in that, In step S10, if the evaluation result fails, the regression model is trained and evaluated again by adjusting the hyperparameters or reselecting the model until a benchmark model that passes the evaluation is obtained.

3. The harmonic reducer condition monitoring method according to claim 1, characterized in that, In step S40, feature importance analysis is performed on relevant features to obtain the anomaly types and corresponding severity data of features that do not conform to the baseline features.

4. The harmonic reducer condition monitoring method according to claim 1, characterized in that, The control module includes a remote detection module, which enables remote detection functionality.

Images