A fault diagnosis method and system for a flexible fully automatic riveting device

Abstract

The present invention relates to the technical field of equipment fault diagnosis, and discloses a fault diagnosis method and system for a flexible fully automatic riveting device. The method comprises obtaining a rotation angle and a vibration signal; decomposing the vibration signal into a frequency spectrum; extracting the amplitude peak value to determine if there is a peak in the frequency spectrum; determining wear if the amplitude peak value exceeds an engagement threshold; and determining looseness if the amplitude peak value exceeds a damage threshold; extracting tension and load if there is continuous fluctuation, and establishing a judgment model based on the fluctuation characteristics; outputting a probability value through the model; and determining a structural abnormality if the probability value exceeds a probability threshold; outputting the judgment result as a signal result; calculating the deviation between the rotation angle and the center value; extracting the signal and deformation variable analysis to obtain a deviation result if the deviation value exceeds a deviation range; fusing the signal result and the deviation result to generate a data set; and training and classifying the data set to determine a specific fault output report. This method can improve the accuracy of fault identification.

Description

Technical Field

[0001] The present invention relates to the technical field of automated equipment fault diagnosis, and in particular to a fault diagnosis method and system for a flexible fully automatic riveting device. Background Art

[0002] In the actual operation of flexible fully automatic riveting equipment, vibration frequency analysis is a key component of fault diagnosis. During operation, the repetitive and high-frequency nature of the riveting action causes its internal mechanical components to generate vibration signals in various frequency bands. Under normal circumstances, these vibration signals are primarily concentrated in the low and medium frequency bands, corresponding to the equipment's basic operating status and the periodic variations in the riveting action, respectively. However, when the equipment malfunctions, the vibration signal may exhibit abnormal frequency bands, such as spikes in the high frequency band or persistent fluctuations in the low frequency band. These abnormal frequency bands are often closely associated with mechanical wear, loose parts, or rotation angle deviation. Specifically, spikes in the high frequency band indicate poor gear meshing or bearing damage, while persistent fluctuations in the low frequency band suggest a loose drive chain or uneven motor load. Furthermore, rotation angle deviation is another key issue in the operation of flexible fully automatic riveting equipment. When the equipment's rotation angle deviates, the accuracy of the riveting action is compromised, resulting in reduced riveting quality. This deviation can be caused by encoder failure, sensor drift, or mechanical deformation.

[0003] In summary, the existing technology is unable to fuse the vibration signal and the rotation angle analysis results, resulting in insufficient accuracy and comprehensiveness of fault diagnosis. Summary of the Invention

[0004] The present invention provides a fault diagnosis method and system for a flexible fully automatic riveting device, so as to solve the problem in the prior art that vibration signals and rotation angle analysis results cannot be integrated, resulting in insufficient accuracy and comprehensiveness of fault diagnosis.

[0005] In a first aspect, in order to solve the above technical problems, the present invention provides a fault diagnosis method for a flexible fully automatic riveting device, comprising:

[0006] Obtaining a rotation angle and a vibration signal; decomposing the vibration signal using Fourier transform to obtain a decomposition spectrum;

[0007] Determine whether the decomposed spectrum contains a spike signal or continuous fluctuations; if the decomposed spectrum contains a spike signal, extract the maximum signal amplitude for judgment; if the maximum amplitude exceeds a preset engagement threshold, determine that it is mechanical wear; if the maximum amplitude exceeds a preset damage threshold, determine that it is a loose part; if the decomposed spectrum contains continuous fluctuations, extract the chain tension and motor load, establish an abnormality judgment model based on pre-stored fluctuation characteristics, and output a probability value through the abnormality judgment model for judgment; if the probability value exceeds a preset probability threshold, determine that it is a mechanical structure abnormality; output the fault determined as a signal analysis result;

[0008] Calculating a deviation between the rotation angle and a pre-stored rotation angle center value to obtain a deviation value;

[0009] Determine whether the deviation value exceeds a preset deviation value range; if so, extract the encoder signal and the mechanical deformation amount for fault analysis to obtain a deviation analysis result;

[0010] The signal analysis result and the deviation analysis result are fused to generate a comprehensive feature data set; a decision tree classification model is used to classify and train the comprehensive feature data set to obtain a classification result; a specific fault type is determined based on the classification result and a fault diagnosis report is generated.

[0011] In an implementation manner of the first aspect, decomposing the vibration signal by Fourier transform to obtain a decomposed spectrum includes:

[0012] Converting the vibration signal into a frequency domain using a fast Fourier transform to obtain a frequency domain signal;

[0013] Dividing the frequency domain signal into spectrum segments and extracting the spectrum segments to obtain spectrum segment feature values;

[0014] Judging the characteristic values ​​of the spectrum segments based on a preset segmentation threshold to determine abnormal frequency segments;

[0015] Signal decomposition is performed on the abnormal frequency band to obtain a decomposed spectrum.

[0016] In an implementation manner of the first aspect, determining whether the decomposed spectrum has a peak signal or continuous fluctuation includes:

[0017] Extracting the maximum signal amplitude of the decomposed spectrum and calculating the standard deviation of the fluctuation amplitude;

[0018] respectively judging the maximum signal amplitude and the standard deviation of the fluctuation amplitude against a preset amplitude threshold and a preset standard deviation threshold;

[0019] If the maximum value of the signal amplitude exceeds the preset amplitude threshold, a spike signal exists; if the standard deviation of the fluctuation amplitude is greater than the preset standard deviation threshold, continuous fluctuation exists.

[0020] In an implementation of the first aspect, calculating the deviation between the rotation angle and a pre-stored rotation angle center value to obtain the deviation value includes:

[0021] Performing weight calculation according to the rotation angle to obtain a dynamic weight;

[0022] A deviation is calculated based on the dynamic weight and a pre-stored rotation angle center value to obtain a deviation value.

[0023] In an implementation manner of the first aspect, calculating the weight according to the rotation angle to obtain the dynamic weight includes:

[0024] The dynamic weight is calculated using the following formula:

[0025]

[0026] in, represents the dynamic weight of the rotation angle collected at the i-th time point, is the adjustable attenuation coefficient, N is the number of discrete time points at which the sensor collects the rotation angle, is the timestamp corresponding to the rotation angle collected at the i-th time point, is the current time.

[0027] In an implementation manner of the first aspect, performing a deviation calculation based on the dynamic weight and a pre-stored rotation angle center value to obtain a deviation value includes:

[0028] The deviation value is calculated using the following formula:

[0029]

[0030] in, Indicates the deviation value of the rotation angle collected at the i-th time point, represents the dynamic weight of the rotation angle collected at the i-th time point, represents the rotation angle collected at the i-th time point, Indicates the pre-stored rotation angle center value.

[0031] In one implementation of the first aspect, extracting the encoder signal and the mechanical deformation amount for fault analysis to obtain the deviation value analysis result includes:

[0032] Obtain encoder signals and mechanical deformation through sensors;

[0033] Judging based on the encoder signal, if the encoder signal has waveform distortion or signal loss, it is determined that there is a measurement system failure;

[0034] Performing a judgment based on the mechanical deformation amount and a pre-stored deformation threshold, and determining that a mechanical failure occurs if the mechanical deformation amount exceeds the pre-stored deformation threshold;

[0035] The deviation analysis results include measurement system failures and mechanical failures.

[0036] In an implementation manner of the first aspect, the fusing the signal analysis result and the deviation analysis result to generate a comprehensive feature dataset includes:

[0037] The signal analysis results and the deviation analysis results are regarded as nodes of a graph, and the correlation between features is used as the weight of the edge to construct a feature graph;

[0038] The feature graph is divided using a community discovery algorithm in graph theory, and features with high correlation are divided into the same community to obtain a community set;

[0039] For each community in the community set, features within the community are selected and fused to generate a comprehensive feature dataset.

[0040] In a second aspect, the present invention provides a fault diagnosis system for a flexible fully automatic riveting device, comprising:

[0041] A data acquisition module is used to acquire the rotation angle and vibration signal; the vibration signal is decomposed by Fourier transform to obtain a decomposed spectrum;

[0042] A signal judgment module is used to judge whether there is a peak signal or continuous fluctuation in the decomposed spectrum, and output the fault obtained by the judgment and analysis as a signal analysis result;

[0043] a deviation calculation module, configured to calculate a deviation between the rotation angle and a pre-stored rotation angle center value to obtain a deviation value;

[0044] a fault judgment module, configured to judge whether the deviation value exceeds a preset deviation value range; if so, extracting the encoder signal and the mechanical deformation amount for fault analysis to obtain a deviation value analysis result;

[0045] The diagnostic output module is used to generate a comprehensive feature data set by fusing the signal analysis results and the deviation analysis results; perform classification training on the comprehensive feature data set using a decision tree classification model to obtain a classification result; determine the specific fault type based on the classification result and generate a fault diagnosis report.

[0046] In a third aspect, the present invention also provides an electronic device comprising a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the fault diagnosis method for the flexible fully automatic riveting device described in any one of the above items is implemented.

[0047] In a fourth aspect, the present invention also provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned fault diagnosis methods for the flexible fully-automatic riveting device.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The present invention obtains the rotation angle and vibration signal and decomposes the vibration signal using Fourier transform to obtain a decomposed spectrum. By using Fourier transform to decompose the spectrum, complex vibration signals can be analyzed in detail, clearly showing the amplitude distribution of different frequency components. This allows for more precise identification of high-frequency spikes and low-frequency sustained fluctuations, corresponding to different fault types, and improves the accuracy and comprehensiveness of fault diagnosis.

[0050] The present invention corresponds to different fault types by identifying high-frequency peak signals and low-frequency continuous fluctuations, and has many beneficial effects. When diagnosing equipment faults, this method can quickly and accurately locate potential faults; mechanical wear or loose parts indicated by high-frequency peak signals, and mechanical structural abnormalities implied by low-frequency continuous fluctuations, can be detected in a timely manner. This avoids further deterioration of the fault, reduces the probability of sudden equipment failure, and reduces the production interruption time caused by equipment failure. At the same time, based on accurate fault judgment, enterprises can arrange maintenance work in a targeted manner, prepare the parts and tools required for maintenance in advance, improve maintenance efficiency, and reduce maintenance costs.

[0051] By analyzing the rotation angle deviation and determining whether there is an encoder failure or mechanical deformation, the present invention can quickly locate the root cause of the rotation angle deviation when the equipment is running, discover the hidden dangers of encoder failure or mechanical deformation in advance, avoid product quality problems caused by the angle deviation affecting the riveting accuracy, and arrange targeted maintenance in time to reduce the risk of equipment damage.

[0052] This invention uses a decision-tree-based fault classification model to fuse vibration signals and rotation angle analysis results. By integrating multi-source data for comprehensive judgment, the accuracy of fault diagnosis is effectively improved, avoiding the one-sidedness of single-source judgments. It can accurately determine the fault type, provide a reliable basis for equipment maintenance, reduce misjudgments and missed judgments, and ensure the stable operation of the flexible fully automatic riveting device.

[0053] In summary, the fault diagnosis method provided by the present invention improves the accuracy and reliability of fault identification and effectively supports the maintenance and operation optimization of the flexible fully automatic riveting device. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a flowchart of a fault diagnosis method for a flexible fully automatic riveting device provided by the first embodiment of the present invention;

[0055] Figure 2 It is a structural schematic diagram of a fault diagnosis system for a flexible fully automatic riveting device provided in a second embodiment of the present invention. DETAILED DESCRIPTION

[0056] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0057] Reference Figure 1 The first embodiment of the present invention provides a fault diagnosis method for a flexible fully automatic riveting device, comprising the following steps:

[0058] S1, obtaining a rotation angle and a vibration signal; decomposing the vibration signal by Fourier transform to obtain a decomposed spectrum;

[0059] S2, determining whether the decomposition spectrum contains a spike signal or continuous fluctuation; if the decomposition spectrum contains a spike signal, extracting the maximum signal amplitude for judgment; if the maximum amplitude exceeds a preset engagement threshold, it is judged to be mechanical wear; if the maximum amplitude exceeds a preset damage threshold, it is judged to be a loose part; if the decomposition spectrum contains continuous fluctuation, extracting the chain tension and motor load, combining them with pre-stored fluctuation characteristics to establish an abnormality judgment model, and outputting a probability value through the abnormality judgment model for judgment; if the probability value exceeds a preset probability threshold, it is judged to be a mechanical structure abnormality; the judged fault is output as a signal analysis result;

[0060] S3, calculating a deviation between the rotation angle and a pre-stored rotation angle center value to obtain a deviation value;

[0061] S4, determining whether the deviation value exceeds a preset deviation value range, and if so, extracting the encoder signal and the mechanical deformation amount for fault analysis to obtain a deviation analysis result;

[0062] S5, generating a comprehensive feature data set based on the fusion of the signal analysis result and the deviation analysis result; performing classification training on the comprehensive feature data set using a decision tree classification model to obtain a classification result; determining a specific fault type based on the classification result and generating a fault diagnosis report.

[0063] In step S1, a rotation angle and a vibration signal are obtained; and the vibration signal is decomposed by Fourier transform to obtain a decomposed spectrum.

[0064] In a specific embodiment, the rotation angle is obtained by a sensor to obtain the rotation angle data of the device, and the angle value and the corresponding timestamp are recorded. The vibration signal is obtained by a sensor to obtain the real-time vibration signal of the flexible full-automatic riveting device.

[0065] In one feasible manner, the use of Fourier transform to decompose the vibration signal to obtain a decomposed spectrum specifically includes: using fast Fourier transform to convert the vibration signal into the frequency domain to obtain a frequency domain signal; dividing the frequency domain signal into spectrum segments and extracting them to obtain spectrum segment characteristic values; judging the spectrum segment characteristic values ​​based on a preset segmentation threshold to determine an abnormal frequency band; and performing signal decomposition on the abnormal frequency band to obtain a decomposed spectrum.

[0066] In a specific embodiment, a fast Fourier transform is used to convert the vibration signal into the frequency domain; the frequency domain signal is divided into spectrum segments, and the characteristic values ​​of different frequency bands are extracted; a mapping relationship between the vibration signal and the operating status of the riveting device is established based on the characteristic values ​​of the spectrum segments; the characteristic values ​​of the spectrum segments are judged using a preset threshold to determine the abnormal frequency band; the signal is decomposed for the abnormal frequency band, and the frequency components of the abnormal frequency band are extracted as the decomposed spectrum.

[0067] For example, monitoring the rotation angle of equipment in industrial sites is crucial for safe operation. For example, during high-speed operation, the rotation angle of a centrifuge barrel must be monitored in real time. The sensor uses a magnetic induction angle sensor, which offers high precision and strong anti-interference capabilities, converting analog signals into digital signals for acquisition. For vibration signals, for example, the main shaft of a riveting device, the time domain signal collected by the vibration sensor can exhibit complex waveforms and must be converted to the frequency domain for analysis. Fast Fourier transforms can be used to convert time domain signals into frequency domain signals, displaying the amplitude distribution of each frequency component.

[0068] In step S2, it is determined whether there is a peak signal or continuous fluctuation in the decomposition spectrum; if there is a peak signal in the decomposition spectrum, the maximum value of the signal amplitude is extracted for judgment. If the maximum amplitude exceeds the preset engagement threshold, it is determined to be mechanical wear. If the maximum amplitude exceeds the preset damage threshold, it is determined to be a loose part; if there is continuous fluctuation in the decomposition spectrum, the chain tension and motor load are extracted, and an abnormality judgment model is established in combination with the pre-stored fluctuation characteristics, and a probability value is output by the abnormality judgment model for judgment. If the probability value exceeds the preset probability threshold, it is determined to be a mechanical structure abnormality; the fault obtained by judgment is output as a signal analysis result.

[0069] In the above step S2, the determination of whether the decomposed spectrum has a peak signal or continuous fluctuation specifically further includes the following steps:

[0070] S21, extracting the maximum signal amplitude of the decomposed spectrum and calculating the standard deviation of the fluctuation amplitude;

[0071] S22, respectively comparing the maximum signal amplitude and the standard deviation of the fluctuation amplitude with a preset amplitude threshold and a preset standard deviation threshold;

[0072] S23, if the maximum value of the signal amplitude exceeds the preset amplitude threshold, a spike signal exists; if the standard deviation of the fluctuation amplitude is greater than the preset standard deviation threshold, a continuous fluctuation exists.

[0073] In a specific embodiment, during the above steps S21 to S23, for the decomposed spectrum data, it is determined whether there is a high-frequency peak signal or a low-frequency continuous fluctuation. The original signal is processed by fast Fourier transform to obtain spectrum data. The high-frequency band and the low-frequency band are divided according to the preset frequency range. For the high-frequency band data, the maximum signal amplitude is extracted. If it exceeds the preset threshold, it is determined that there is a peak signal. For the low-frequency band data, the standard deviation of the fluctuation amplitude is calculated. If it is greater than the preset threshold, it is determined that there is a continuous fluctuation. The judgment results of the high-frequency band and the low-frequency band are integrated to generate a spectrum feature report. Based on the spectrum feature report, the support vector machine algorithm is used for classification. According to the classification result, the abnormal type of the spectrum data is determined, and the abnormal type obtained by judgment is output as the signal analysis result.

[0074] For example, the Fast Fourier Transform (FFT) can convert time-domain signals into frequency-domain signals. This conversion facilitates observation of the signal's energy distribution at different frequencies. In vibration monitoring, high frequencies typically correspond to transient shocks, while low frequencies reflect sustained vibrations. For example, in a flexible, fully automatic riveting device, the vibration signal should remain stable within a specific frequency range during normal operation. High-frequency analysis typically focuses on signals with frequencies above 1,000 Hz. If the amplitude at a certain frequency exceeds a preset threshold (e.g., twice the amplitude during normal operation), this indicates abnormal equipment vibration. For example, if a rivet violently collides with the hole wall during riveting, a significant spike signal can be generated in the high-frequency range, with an amplitude up to three to four times the normal value. Low-frequency analysis primarily focuses on frequencies below 100 Hz. By calculating the standard deviation of the fluctuation amplitude, equipment stability can be assessed. If the standard deviation exceeds a preset threshold (e.g., 50% of the normal operating value), this indicates sustained equipment vibration.

[0075] In a specific embodiment, if a peak signal exists in the decomposed spectrum, the maximum signal amplitude is extracted for judgment. If the maximum amplitude exceeds a preset engagement threshold, it is determined to be mechanical wear. If the maximum amplitude exceeds a preset damage threshold, it is determined to be a loose part. The specific implementation process includes:

[0076] If a peak signal is present in the decomposed spectrum, the sensor extracts the maximum signal amplitude. When the maximum gearbox signal amplitude is detected to be significantly greater than a preset engagement threshold, it indicates abnormal gear wear, and the machine is determined to have mechanical wear. In this embodiment, the preset damage threshold is set to 3 kilohertz. When the maximum signal amplitude exceeds 3 kilohertz, a characteristic signal is generated, indicating that a bearing component is loose, and the machine is determined to have loose parts.

[0077] In a specific embodiment, if there is continuous fluctuation in the decomposition spectrum, the chain tension and motor load are extracted, and an abnormality judgment model is established in combination with pre-stored fluctuation characteristics, and a probability value is output by the abnormality judgment model for judgment. If the probability value exceeds a preset probability threshold, it is determined that the mechanical structure is abnormal. The specific implementation process includes: if there is continuous fluctuation, a sensor is used to obtain a low-frequency vibration signal, and the continuous fluctuation characteristics in the vibration signal are extracted; according to the preset threshold, it is judged whether the continuous fluctuation characteristics exceed the range, and if so, further analysis is performed; the tension data of the transmission chain and the load data of the motor are obtained, and the chain slack and load balance are analyzed; in combination with the pre-stored fluctuation characteristics, chain slack and load balance, a mechanical structure abnormality judgment model is established; the probability value of the mechanical structure abnormality is output through the judgment model, and if the probability value exceeds the preset threshold, the mechanical structure abnormality warning information is determined and output to the monitoring system.

[0078] For example, by placing accelerometers at key points on the equipment, fundamental frequency vibration signals with a frequency range of 0 to 100 Hz can be collected. The persistent fluctuation characteristic of these signals is primarily reflected in continuous changes in amplitude. For example, when bearings are experiencing early wear, they will produce persistent, small fluctuations around 20 Hz. For transmission systems, a 5% amplitude change rate is typically set as the warning threshold. For example, for a conveyor system on a production line, if the amplitude change rate reaches 8%, the system will trigger an abnormality signal. This requires analysis based on transmission chain tension data. Under normal operating conditions, chain tension should be maintained between 70% and 90% of the rated value. If the tension drops below 60% of the rated value and is accompanied by abnormal vibration, it often indicates chain slack. By collecting motor current and speed data, load balancing indicators can be obtained. Ideally, motor load fluctuations should be kept within 10% of the rated power. If the motor load fluctuations of a particular film machine exceed 15% and are accompanied by abnormal low-frequency vibrations, it indicates improper installation or wear of mechanical components. These characteristic parameters are input into a mechanical structure abnormality detection model, and a weighted scoring method is used to calculate the abnormality probability. In the model, the vibration feature weight is 0.4, the chain status weight is 0.3, and the load balancing weight is 0.3. When the combined score exceeds 0.8, it is determined to be a mechanical structural abnormality and a fault warning is triggered. The support vector machine algorithm establishes precise classification boundaries based on historical fault data. Through feature space mapping, different types of mechanical anomalies can be effectively distinguished. For example, in a failure sample of a certain packaging equipment, the feature vectors for bearing failure, chain failure, and improper mechanical installation showed clear clustering characteristics, with a classification accuracy exceeding 90%. The final warning output includes the fault type, failure probability, and recommended measures.

[0079] In step S3, the deviation between the rotation angle and the pre-stored rotation angle center value is calculated to obtain a deviation value.

[0080] In the above step S3, the calculation of the deviation between the rotation angle and the pre-stored rotation angle center value to obtain the deviation value specifically includes the following steps:

[0081] S31, performing weight calculation according to the rotation angle to obtain a dynamic weight;

[0082] It should be noted that the dynamic weight is calculated using the following formula:

[0083]

[0084] in, represents the dynamic weight of the rotation angle collected at the i-th time point, is the adjustable attenuation coefficient, N is the number of discrete time points at which the sensor collects the rotation angle, is the timestamp corresponding to the rotation angle collected at the i-th time point, is the current time.

[0085] The formula used in this embodiment is based on the principle of exponential decay to calculate weights. In the denominator, the exponential terms of all time points are summed to normalize and ensure that the sum of all weights is 1. In the numerator, according to the time interval between the current time and the i-th time point, as this interval increases, the value of the exponential term decays exponentially, and the corresponding weight becomes smaller. This design conforms to the principle that "the closer the data to the current time, the greater the weight", because recent data can better reflect the current status of the device, so it is given a higher weight, while data from a more distant time has a lower weight, which reflects the idea of ​​dynamic weighting.

[0086] It should be noted that the dynamic weight means that when analyzing the rotation angle data to determine the equipment failure, the rotation angle collected at each time point has a different degree of influence on the final result. It will change with the change of time point, so it is called a dynamic weight. For example, when judging whether the current equipment has a failure caused by the rotation angle deviation, the rotation angle data collected close to the current time may better reflect the real-time status of the equipment, and its weight will be relatively large; while the weight of the data collected at an earlier time may be smaller. The adjustable attenuation coefficient is used to control the decay rate of the dynamic weight of the rotation angle over time. Its value will affect the trend of the weight change. If If the value is larger, it means that the weight decays quickly over time, that is, the influence of the rotation angle data collected earlier on the current fault judgment will decrease rapidly; if Small, slow weight decay, past data can still maintain a certain influence in fault judgment. This allows for flexible optimization of rotation angle data analysis based on the actual equipment operation and fault diagnosis requirements. The number of discrete time points at which the sensor collects rotation angle data represents the total number of N discrete time points collected during the rotation angle monitoring and analysis process. For example, if the rotation angle is collected at regular intervals over a period of time, the total number of collections is N. This parameter determines the richness of the data. A larger value for N provides a more detailed description of the rotation angle trend, facilitating a more accurate analysis of rotation angle deviations and, consequently, determining the presence and type of equipment faults. The timestamp associated with each collected rotation angle indicates the specific time at which each rotation angle data point was collected. Timestamps precisely record the moment of data collection. By comparing timestamps at different time points, the order and intervals of rotation angle changes over time can be understood, which is critical for analyzing dynamic rotation angle changes. For example, when analyzing rotation angle deviation, timestamps can help determine whether the angle deviation occurred suddenly or accumulated gradually, providing important clues to determine the cause of the fault. The current time serves as a reference when analyzing rotation angle data.

[0087] S32, performing deviation calculation based on the dynamic weight and the pre-stored rotation angle center value to obtain a deviation value.

[0088] It should be noted that the deviation value is calculated using the following formula:

[0089]

[0090] in, Indicates the deviation value of the rotation angle collected at the i-th time point, represents the dynamic weight of the rotation angle collected at the i-th time point, represents the rotation angle collected at the i-th time point, Indicates the pre-stored rotation angle center value.

[0091] The formula used in this embodiment calculates the distance between the angle value at the i-th time point and the center value of the preset rotation angle range, that is, the degree of deviation of the angle value at that time point from the ideal center state. On this basis, multiplying by the weight makes this deviation value weighted according to the importance of the data. The deviation of the time point with high importance accounts for a greater proportion in the final deviation calculation, and the deviation of the time point with low importance has a smaller impact on the overall situation. This comprehensive consideration of the time factor and the angle deviation factor can more accurately reflect the degree to which the actual state of the device at different time points deviates from the ideal state.

[0092] It should be noted that the pre-stored rotation angle center value refers to the standard reference value of the rotation angle determined through extensive testing and data analysis under normal operation of the flexible fully automatic riveting device. In actual operation, the rotation angle of the device should fluctuate within a certain range around this center value.

[0093] In step S4, it is determined whether the deviation value exceeds a preset deviation value range. If so, the encoder signal and the mechanical deformation are extracted to perform fault analysis and obtain a deviation analysis result.

[0094] In the above step S4, it is determined whether the deviation value exceeds the preset deviation value range. If so, the encoder signal and the mechanical deformation are extracted to perform fault analysis to obtain the deviation analysis result, which specifically includes the following steps:

[0095] S41, obtaining encoder signals and mechanical deformation through sensors;

[0096] S42, judging based on the encoder signal, if the encoder signal has waveform distortion or signal loss, it is determined that there is a measurement system failure;

[0097] S43, judging based on the mechanical deformation amount and a pre-stored deformation threshold, and if the mechanical deformation amount exceeds the pre-stored deformation threshold, determining that it is a mechanical failure;

[0098] In one specific embodiment, during steps S41 to S43, the deviation value is compared with a preset threshold value to determine whether it exceeds the threshold range. If the deviation value exceeds the threshold range, an encoder signal is obtained. Based on the encoder signal, a determination is made as to whether a fault exists. If the encoder is operating normally, mechanical structure deformation data is obtained. Based on the mechanical structure deformation data, a determination is made as to whether mechanical deformation exists. Based on the encoder fault state and the mechanical deformation, the cause of the deviation value is determined and output as a deviation analysis result. The deviation analysis result includes both measurement system faults and mechanical faults.

[0099] For example, analysis of rotation angle deviation primarily involves obtaining the actual equipment angle through sensors and setting thresholds for judgment. For example, when a vertical mixing device is operating normally, the angle between the rotation axis and the vertical should be kept within 0.2 degrees. However, if the angle sensor measures a value of 0.5 degrees at a certain moment, this exceeds the preset threshold and requires further analysis. After determining the deviation value, encoder operating data is first obtained. For example, for a rotary encoder, the operating status is determined by examining the output signal waveform. Under normal circumstances, the waveform should exhibit regular rectangular pulses. Any waveform distortion or signal loss indicates an encoder fault. A decrease in the encoder output signal amplitude to 60% of the normal value, or irregular pulse intervals, indicates a fault. When the encoder is operating normally, further testing is required to determine mechanical structure deformation. For example, for a slewing bearing, strain gauges are used to measure deformation. During normal operation, the radial runout of the bearing ring should be within 0.05 mm. If the measured radial runout reaches 0.15 mm, deformation of the bearing ring is indicated. Mechanical deformation can be determined from multiple perspectives. For example, when infrared thermal imaging is used to detect the temperature distribution of the support ring, the normal operating temperature should be within 40 degrees. If the local temperature exceeds 60 degrees, there is deformation caused by poor lubrication. In addition, the vibration spectrum of the support ring is measured by a vibration sensor. If an abnormal peak appears in a certain frequency band, it is also due to deformation. Finally, the cause of the deviation is determined by analyzing the encoder signal and the mechanical deformation. For example, if the encoder output is normal but the support ring is deformed, it can be determined as a mechanical failure. If the encoder signal is abnormal and the mechanical state is normal, it can be determined as a measurement system failure. When both are abnormal, it is necessary to combine the historical operation data of the equipment to determine the sequence and cause-and-effect relationship of the failure. For example, deformation of the support ring often causes changes in the encoder installation reference, which in turn causes signal abnormalities. In this case, mechanical failure should be the main cause.

[0100] In step S5, the signal analysis result and the deviation analysis result are fused to generate a comprehensive feature data set; a decision tree classification model is used to perform classification training on the comprehensive feature data set to obtain a classification result; and a specific fault type is determined based on the classification result and a fault diagnosis report is generated.

[0101] In the above step S5, the signal analysis result and the deviation analysis result are integrated to generate a comprehensive feature data set; a decision tree classification model is used to classify and train the comprehensive feature data set to obtain a classification result; and a specific fault type is determined based on the classification result and a fault diagnosis report is generated. Specifically, the following steps are further included:

[0102] S51, treating the signal analysis results and the deviation analysis results as nodes of a graph, and using the correlations between features as edge weights to construct a feature graph;

[0103] S52, using a community discovery algorithm in graph theory to divide the feature graph, dividing highly correlated features into the same community, and obtaining a community set;

[0104] S53 , for each community in the community set, select features within the community and fuse them to generate a comprehensive feature dataset.

[0105] In one specific embodiment, during steps S51 to S53 above, feature fusion processing is performed based on the vibration signal features and deviation features obtained from the judgment results to generate a comprehensive feature dataset. A pre-established decision tree classification model is used to perform classification training on the comprehensive feature dataset. If the comprehensive feature dataset contains unlabeled samples, the model predicts the fault type of the unlabeled samples. Based on the model classification results, the specific fault type of the equipment is determined, and a fault diagnosis report is generated. If the fault diagnosis report contains multiple fault types, the primary fault type is determined through model weight analysis. Based on the primary fault type, equipment maintenance recommendations are generated and output to the equipment management system.

[0106] For example, an accelerometer can capture vibration time-domain waveforms. For example, during normal operation of a rotating device, the vibration amplitude fluctuates within a variance of 0.5 mm / s. However, when a fault occurs, a noticeable shock waveform is generated, with an amplitude exceeding 2 mm / s. Rotational angle deviation reflects the positioning accuracy of the device. Under normal operating conditions, the angle deviation should be controlled within 0.1 degree. Feature extraction requires extracting effective features from the raw data. Statistical features such as mean, variance, and peak value can be extracted from the vibration signal, while frequency domain features can be obtained through Fourier transform. Angle deviation features include maximum deviation, average deviation, and fluctuation range. For example, in a certain type of processing equipment, a sudden increase in the vibration mean and an angle deviation exceeding 0.5 degrees often indicate a bearing failure. Feature fusion is the process of organically combining features from different sources. Principal component analysis can be used to reduce the dimensionality of vibration features and fuse them with angle deviation features to form a more representative comprehensive feature. For example, in the case of a CNC machine tool, the fused feature can simultaneously reflect the correlation between abnormal spindle vibration and decreased positioning accuracy. The decision tree classification model establishes classification rules based on training data. By setting a reasonable splitting criterion, such as a Gini coefficient threshold less than 0.3, the model can automatically learn the correspondence between fault features and types. In practical applications, when an industrial robot experiences an anomaly, the model can classify the fault into specific types, such as motor failure or mechanical wear, based on comprehensive features. Prediction of unlabeled samples relies on the model's generalization capability. If the vibration amplitude and angular deviation characteristics of a newly collected sample are at least 85% similar to those of historical bearing fault samples, it is predicted to be a bearing fault. When multiple fault types exist, the primary fault is determined by calculating the weighted contribution of each feature. For example, if a piece of equipment exhibits both mechanical wear and electrical fault features, if the weight of the mechanical wear feature reaches 70%, it is considered the primary fault. Maintenance recommendations are generated based on the fault diagnosis results. If a bearing fault is detected, recommendations include bearing replacement and lubrication system adjustment; if mechanical wear is the cause, component repair or replacement are recommended. These recommendations are distributed to maintenance personnel through the equipment management system, enabling rapid fault resolution. This approach significantly improves equipment reliability and reduces maintenance costs.

[0107] In summary, the present invention discloses a fault diagnosis method for a flexible fully automatic riveting device. Compared with the prior art, the present invention has the following beneficial effects:

[0108] The present invention obtains the rotation angle and vibration signal and decomposes the vibration signal using Fourier transform to obtain a decomposed spectrum. By using Fourier transform to decompose the spectrum, complex vibration signals can be analyzed in detail, clearly showing the amplitude distribution of different frequency components. This allows for more precise identification of high-frequency spikes and low-frequency sustained fluctuations, corresponding to different fault types, and improves the accuracy and comprehensiveness of fault diagnosis.

[0109] The present invention corresponds to different fault types by identifying high-frequency peak signals and low-frequency continuous fluctuations, and has many beneficial effects. When diagnosing equipment faults, this method can quickly and accurately locate potential faults; mechanical wear or loose parts indicated by high-frequency peak signals, and mechanical structural abnormalities implied by low-frequency continuous fluctuations, can be detected in a timely manner. This avoids further deterioration of the fault, reduces the probability of sudden equipment failure, and reduces the production interruption time caused by equipment failure. At the same time, based on accurate fault judgment, enterprises can arrange maintenance work in a targeted manner, prepare the parts and tools required for maintenance in advance, improve maintenance efficiency, and reduce maintenance costs.

[0110] By analyzing the rotation angle deviation and determining whether there is an encoder failure or mechanical deformation, the present invention can quickly locate the root cause of the rotation angle deviation when the equipment is running, discover the hidden dangers of encoder failure or mechanical deformation in advance, avoid product quality problems caused by the angle deviation affecting the riveting accuracy, and arrange targeted maintenance in time to reduce the risk of equipment damage.

[0111] This invention uses a decision-tree-based fault classification model to fuse vibration signals and rotation angle analysis results. By integrating multi-source data for comprehensive judgment, the accuracy of fault diagnosis is effectively improved, avoiding the one-sidedness of single-source judgments. It can accurately determine the fault type, provide a reliable basis for equipment maintenance, reduce misjudgments and missed judgments, and ensure the stable operation of the flexible fully automatic riveting device.

[0112] In summary, the fault diagnosis method provided by the present invention improves the accuracy and reliability of fault identification and effectively supports the maintenance and operation optimization of the flexible fully automatic riveting device.

[0113] Reference Figure 2 The second embodiment of the present invention provides a fault diagnosis system for a flexible fully automatic riveting device, comprising:

[0114] The data acquisition module 101 is used to acquire the rotation angle and the vibration signal; the vibration signal is decomposed by Fourier transform to obtain a decomposed spectrum;

[0115] The signal judgment module 102 is used to judge whether the decomposition spectrum contains a spike signal or continuous fluctuation. If the decomposition spectrum contains a spike signal, the maximum signal amplitude is extracted for judgment. If the maximum amplitude exceeds a preset engagement threshold, it is judged as mechanical wear. If the maximum amplitude exceeds a preset damage threshold, it is judged as a loose part. If the decomposition spectrum contains continuous fluctuation, the chain tension and motor load are extracted, and an abnormality judgment model is established based on pre-stored fluctuation characteristics. The probability value output by the abnormality judgment model is used for judgment. If the probability value exceeds a preset probability threshold, it is judged as a mechanical structure abnormality. The fault determined is output as a signal analysis result.

[0116] a deviation calculation module 103, configured to calculate a deviation between the rotation angle and a pre-stored rotation angle center value to obtain a deviation value;

[0117] The fault judgment module 104 is used to judge whether the deviation value exceeds a preset deviation value range. If so, the encoder signal and the mechanical deformation are extracted to perform fault analysis and obtain a deviation value analysis result.

[0118] The diagnostic output module 105 is used to generate a comprehensive feature data set based on the integration of the signal analysis results and the deviation analysis results; perform classification training on the comprehensive feature data set using a decision tree classification model to obtain a classification result; determine a specific fault type based on the classification result and generate a fault diagnosis report.

[0119] It should be noted that the fault diagnosis system for a flexible fully automatic riveting device provided in an embodiment of the present invention is used to execute all the process steps of the fault diagnosis method for a flexible fully automatic riveting device in the above embodiment. The working principles and beneficial effects of the two correspond one to one, and therefore will not be repeated here.

[0120] An embodiment of the present invention further provides an electronic device. The electronic device includes: a processor, a memory, and a computer program stored in the memory and executable on the processor, such as a fault diagnosis program for a flexible fully automatic riveting device. When the processor executes the computer program, the steps in the above-mentioned embodiments of the fault diagnosis method for a flexible fully automatic riveting device are implemented, such as Figure 1 Alternatively, when the processor executes the computer program, the functions of the modules / units in the above-mentioned device embodiments are realized, such as the data acquisition module.

[0121] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory and executed by the processor to implement the present invention. The one or more modules / units may be a series of computer program instruction segments capable of implementing specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device.

[0122] The electronic device may be a computing device such as a desktop computer, notebook, PDA, or smart tablet. The electronic device may include, but is not limited to, a processor and memory. Those skilled in the art will appreciate that the aforementioned components are merely examples of electronic devices and do not constitute a limitation of the electronic device. The electronic device may include more or fewer components than those described above, or a combination of certain components, or different components. For example, the electronic device may also include input / output devices, network access devices, buses, and the like.

[0123] The processor may be a central processing unit (CPU), other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor. The processor is the control center of the electronic device and connects various parts of the entire electronic device using various interfaces and lines.

[0124] The memory can be used to store the computer programs and / or modules. The processor implements the various functions of the electronic device by running or executing the computer programs and / or modules stored in the memory and accessing the data stored in the memory. The memory may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as a sound playback function or an image playback function); the data storage area may store data generated based on the use of the mobile phone (such as audio data, a phone book, etc.). Furthermore, the memory may include high-speed random access memory and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0125] If the module / unit integrated into the electronic device is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention can implement all or part of the process steps in the above-mentioned method embodiments by using a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard drive, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal, and software distribution medium. It should be noted that the content of the computer-readable medium can be appropriately increased or decreased based on the requirements of legislation and patent practice in a jurisdiction. For example, in some jurisdictions, based on legislation and patent practice, computer-readable media does not include electric carrier signals and telecommunication signals.

[0126] It should be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art can understand and implement the present invention without inventive effort.

[0127] The specific embodiments described above further illustrate the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included within the scope of protection of the present invention for those skilled in the art.

Claims

1. A fault diagnosis method for a flexible fully automatic riveting device, characterized in that: include: Obtaining a rotation angle and a vibration signal; decomposing the vibration signal using Fourier transform to obtain a decomposition spectrum; Determine whether the decomposed spectrum contains a spike signal or continuous fluctuations; if the decomposed spectrum contains a spike signal, extract the maximum signal amplitude for judgment; if the maximum amplitude exceeds a preset engagement threshold, determine that it is mechanical wear; if the maximum amplitude exceeds a preset damage threshold, determine that it is a loose part; if the decomposed spectrum contains continuous fluctuations, extract the chain tension and motor load, establish an abnormality judgment model based on pre-stored fluctuation characteristics, and output a probability value through the abnormality judgment model for judgment; if the probability value exceeds a preset probability threshold, determine that it is a mechanical structure abnormality; output the fault determined as a signal analysis result; Calculating a deviation between the rotation angle and a pre-stored rotation angle center value to obtain a deviation value; Determine whether the deviation value exceeds a preset deviation value range; if so, extract the encoder signal and the mechanical deformation amount for fault analysis to obtain a deviation analysis result; Fusion is performed based on the signal analysis result and the deviation analysis result to generate a comprehensive feature data set; Using a decision tree classification model to perform classification training on the comprehensive feature data set to obtain a classification result; The specific fault type is determined based on the classification result and a fault diagnosis report is generated.

2. The fault diagnosis method of the flexible fully automatic riveting device according to claim 1 is characterized in that: Decomposing the vibration signal by Fourier transform to obtain a decomposed spectrum includes: Converting the vibration signal into a frequency domain using a fast Fourier transform to obtain a frequency domain signal; Dividing the frequency domain signal into spectrum segments and extracting the spectrum segments to obtain spectrum segment feature values; Judging the characteristic values ​​of the spectrum segments based on a preset segmentation threshold to determine abnormal frequency segments; Signal decomposition is performed on the abnormal frequency band to obtain a decomposed spectrum.

3. The fault diagnosis method of the flexible fully automatic riveting device according to claim 1 is characterized in that: The determining whether the decomposed spectrum has a peak signal or continuous fluctuation includes: Extracting the maximum signal amplitude of the decomposed spectrum and calculating the standard deviation of the fluctuation amplitude; respectively judging the maximum signal amplitude and the standard deviation of the fluctuation amplitude against a preset amplitude threshold and a preset standard deviation threshold; If the maximum value of the signal amplitude exceeds the preset amplitude threshold, a spike signal exists; if the standard deviation of the fluctuation amplitude is greater than the preset standard deviation threshold, continuous fluctuation exists.

4. The fault diagnosis method of the flexible fully automatic riveting device according to claim 1 is characterized in that: Calculating the deviation between the rotation angle and a pre-stored rotation angle center value to obtain a deviation value includes: Performing weight calculation according to the rotation angle to obtain a dynamic weight; A deviation is calculated based on the dynamic weight and a pre-stored rotation angle center value to obtain a deviation value.

5. The fault diagnosis method of the flexible fully automatic riveting device according to claim 4 is characterized in that: The weight calculation according to the rotation angle to obtain the dynamic weight includes: The dynamic weight is calculated using the following formula: ； in, represents the dynamic weight of the rotation angle collected at the i-th time point, is the adjustable attenuation coefficient, N is the number of discrete time points at which the sensor collects the rotation angle, is the timestamp corresponding to the rotation angle collected at the i-th time point, is the current time.

6. The fault diagnosis method of the flexible fully automatic riveting device according to claim 4 is characterized in that: The calculating the deviation between the dynamic weight and the pre-stored rotation angle center value to obtain the deviation value includes: The deviation value is calculated using the following formula: ； in, Indicates the deviation value of the rotation angle collected at the i-th time point, represents the dynamic weight of the rotation angle collected at the i-th time point, represents the rotation angle collected at the i-th time point, Indicates the pre-stored rotation angle center value.

7. The fault diagnosis method of the flexible fully automatic riveting device according to claim 1 is characterized in that: The extracting of encoder signals and mechanical deformation amounts for fault analysis to obtain deviation analysis results includes: Obtain encoder signals and mechanical deformation through sensors; Judging based on the encoder signal, if the encoder signal has waveform distortion or signal loss, it is determined that there is a measurement system failure; Performing a judgment based on the mechanical deformation amount and a pre-stored deformation threshold, and determining that a mechanical failure occurs if the mechanical deformation amount exceeds the pre-stored deformation threshold; The deviation analysis results include measurement system failures and mechanical failures.

8. The fault diagnosis method of the flexible fully automatic riveting device according to claim 1 is characterized in that: The generating a comprehensive feature data set by fusing the signal analysis result and the deviation analysis result includes: The signal analysis results and the deviation analysis results are regarded as nodes of a graph, and the correlation between features is used as the weight of the edge to construct a feature graph; The feature graph is divided using a community discovery algorithm in graph theory, and features with high correlation are divided into the same community to obtain a community set; For each community in the community set, features within the community are selected and fused to generate a comprehensive feature dataset.

9. A fault diagnosis system for a flexible fully automatic riveting device, characterized in that: include: A data acquisition module is used to acquire the rotation angle and vibration signal; the vibration signal is decomposed by Fourier transform to obtain a decomposed spectrum; a signal judgment module for judging whether the decomposed spectrum contains a spike signal or continuous fluctuation; if the decomposed spectrum contains a spike signal, extracting the maximum signal amplitude for judgment; if the maximum amplitude exceeds a preset engagement threshold, it is judged as mechanical wear; if the maximum amplitude exceeds a preset damage threshold, it is judged as a loose part; if the decomposed spectrum contains continuous fluctuation, extracting the chain tension and motor load, combining them with pre-stored fluctuation characteristics to establish an abnormality judgment model, and outputting a probability value through the abnormality judgment model for judgment; if the probability value exceeds a preset probability threshold, it is judged as a mechanical structure abnormality; and outputting the judged fault as a signal analysis result; a deviation calculation module, configured to calculate a deviation between the rotation angle and a pre-stored rotation angle center value to obtain a deviation value; a fault judgment module, configured to judge whether the deviation value exceeds a preset deviation value range; if so, extracting the encoder signal and the mechanical deformation amount for fault analysis to obtain a deviation analysis result; A diagnostic output module, configured to generate a comprehensive feature data set by fusing the signal analysis result and the deviation analysis result; Using a decision tree classification model to perform classification training on the comprehensive feature data set to obtain a classification result; The specific fault type is determined based on the classification result and a fault diagnosis report is generated.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the fault diagnosis method for the flexible fully-automatic riveting device according to any one of claims 1 to 8.

Images