COB (Chip On Board) splitter fault monitoring system and method based on data analysis

Abstract

The invention discloses a COB board splitter fault monitoring system and method based on data analysis, and relates to the technical field of board splitter fault monitoring, the dynamic reset deviation value of a shaft body is collected in real time in the shaft body reset process, and the static reset deviation value of the shaft body is collected at regular time after reset is completed; obtaining feature data capable of being used for distinguishing fault types; whether the dynamic reset deviation and the static reset deviation are abnormal or not is monitored in real time in the shaft body reset process, a dynamic and static reset deviation correlation prediction mechanism is set, the probability of occurrence of the real-time static reset deviation abnormity is calculated when only the dynamic reset deviation occurs, and the static reset deviation abnormity is predicted; prompting corresponding to the abnormal types of the dynamic reset deviation and the static reset deviation; the reset deviation abnormity collected in real time is analyzed for fault monitoring, so that the efficiency and accuracy of fault monitoring are improved; and a dynamic and static correlation prediction mechanism is introduced, so that the effect of predicting the occurrence of static reset deviation abnormity in advance is achieved.

Description

Technical Field

[0001] This invention relates to the field of PCB depaneling machine fault monitoring technology, specifically a COB depaneling machine fault monitoring system and method based on data analysis. Background Technology

[0002] The working principle of a COB PCB separator is as follows: A whole board of LED COBs placed in the material box is pushed onto the table of the initial bending feeding area using a push rod. The initial bending feeding area accurately pushes the board to the edge of the first bending according to the data of the LED COBs. Simulating human hand action, it bends the board vertically, separating the whole board of COBs into a strip according to the pre-cut position. Then, it is transferred to the table of the second bending feeding area. The second bending feeding area accurately pushes the board to the edge of the second bending according to the data of the LED COBs. Simulating human hand action, it bends the board vertically, separating the strip of COB into individual pieces according to the pre-cut position. However, during the operation of the COB PCB separator, mechanical obstructions such as foreign objects on the guide rail, lead screw jamming, servo motor signal deviation, and limit switch failure can cause the push rod and cylinder to malfunction. The current fault monitoring method for incomplete or unreset troubleshooting mainly involves switching to manual cylinder mode, checking each sensor for activation, and manually moving the push rod by clicking the motor when the sensor in the push rod's original position is not activated, followed by verification through emergency stop reset. This fault monitoring method relies too heavily on subjective human judgment, resulting in low efficiency. Frequent manual adjustments increase downtime and reduce production efficiency. Furthermore, the impact of dynamic reset deviation on static reset deviation is not considered during fault monitoring, and the varying degrees of impact of different types of static reset deviation on static reset deviation lead to untimely monitoring and troubleshooting of static reset deviation, potentially causing damage to the PCB depaneling machine. Summary of the Invention

[0003] The purpose of this invention is to provide a COB depaneling machine fault monitoring system and method based on data analysis, so as to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a COB depaneling machine fault monitoring method based on data analysis, the method comprising the following steps:

[0005] S1. During the shaft reset process, the dynamic reset deviation value of the shaft is collected in real time, and the static reset deviation value of the shaft is collected periodically after the reset is completed.

[0006] S2. Obtain feature data that can be used to distinguish fault types from the collected dynamic reset deviation value and static reset deviation value data;

[0007] S3. During the shaft reset process, monitor in real time whether dynamic reset deviation and static reset deviation are abnormal, and set up a dynamic and static reset deviation correlation prediction mechanism to predict the probability of real-time static reset deviation abnormality when only dynamic reset deviation occurs, and predict the occurrence of static reset deviation abnormality.

[0008] S4. Based on the generated fault repair prompt type, generate fault repair prompt content corresponding to the reset deviation abnormality.

[0009] Furthermore, in step S1: during the shaft reset process, starting from the reset start time t1 and ending at the reset completion time, a dynamic reset deviation value is collected every time interval Δt, and the collected dynamic reset deviation values ​​are recorded as {s1, s2, ..., s...} n After the reset is completed, within the data acquisition duration Δd, m static reset deviations are collected periodically, and the collected static reset deviations are denoted as {v1, v2, ..., v...}. m}

[0010] Furthermore, in step S2: the collected dynamic reset deviation data is analyzed, and the peak value in the dynamic reset deviation data is obtained based on the comparative analysis of the same group of data. And analyze the real-time rate of change of the dynamic reset deviation value. Furthermore, the real-time collected dynamic reset deviation values ​​are compared and analyzed with the mean. Dynamic reset deviation values ​​greater than the mean are marked as fluctuation deviation values, and the frequency q of the fluctuation deviation values ​​is recorded. The real-time rate of change of the dynamic reset deviation values ​​is calculated according to the following formula: ;in This represents the rate of change of the j-th deviation, where j represents the deviation rate of change number, j=1, 2, ..., n-1; and i represents the dynamic reset deviation value number, i=2, 3, ..., n.

[0011] Analyze the collected static reset deviation data and calculate the mean value of the static reset deviation data. And calculate the fluctuation value of the static reset deviation according to the following formula: Where u represents the static reset deviation fluctuation value, and k represents the static reset deviation value number, k=1,2,…,m.

[0012] Furthermore, in step S3: set the peak threshold η and the rate of change threshold ω0 for the dynamic reset deviation value; set the frequency threshold q0 for the fluctuation deviation value; and analyze the anomaly types of the dynamic reset deviation data.

[0013] like but The dynamic reset deviation was determined to be normal during the reset process;

[0014] like but This indicates an abnormal peak value of dynamic reset deviation, indicating a shaft jamming fault during the reset process; a fault repair prompt r1 is generated.

[0015] like but This indicates an abnormal rate of change in the dynamic reset deviation, indicating an intermittent encoder signal interference fault during the reset process; a fault repair prompt r2 is generated.

[0016] like but This indicates that both the peak value and the rate of change of the dynamic reset deviation are abnormal, indicating that the shaft is stuck and the encoder signal is intermittently interfered with during the shaft reset process; a fault repair prompt r3 is generated.

[0017] Furthermore, in step S3: during the shaft reset process, the number of times the dynamic reset deviation value exceeds the threshold is monitored and recorded in real time. The number of occurrences is denoted as x, and the duration of the abnormal peak value of dynamic reset deviation {Δα1, Δα2, ... Δα} is recorded. x} will appear The number of occurrences is denoted as y, and the duration of the abnormal peak value of dynamic reset deviation {Δβ1, Δβ2, ... Δβ} is recorded. y The duration of the dynamic reset deviation anomaly is used to determine whether the monitored dynamic reset deviation anomaly is a false dynamic reset deviation anomaly. A threshold for the duration of the true dynamic reset deviation anomaly is set as Δρ. The duration of the acquired dynamic reset deviation anomaly is compared with the set threshold. Dynamic reset deviation anomalies with a duration higher than the set threshold are marked as true dynamic reset deviation anomalies, and those with a duration lower than or equal to the set threshold are marked as false dynamic reset deviation anomalies. The number of true dynamic reset deviation peak anomalies x' and the number of true dynamic reset deviation change rate anomalies y' are recorded in the comparative analysis.

[0018] A dynamic and static reset deviation correlation prediction mechanism is set, and the correlation prediction period is set to [duration value]. If only dynamic reset deviation anomalies occur within a correlation prediction period, the probability of static reset deviation occurrence is calculated based on the number of monitored dynamic reset deviation anomalies after each reset. The probability of static reset deviation occurrence is then calculated using the following formula:

[0019] ;

[0020] Where p represents the probability that the real-time static reset deviation exceeds the threshold; c is the probability that a static reset deviation anomaly occurs when there is no dynamic reset anomaly; and b represents the influence coefficient of the number of occurrences of dynamic reset deviation anomaly on the probability of static reset deviation occurrence. This represents the average probability that a single deviation peak exceeding the threshold will lead to an abnormal static deviation. This represents the average probability that an abnormal rate of change in the real-time value of a single dynamic reset deviation will lead to an abnormal static reset deviation.

[0021] By sampling and statistically analyzing the historical operating data of the depaneling machine, we can obtain the probability that abnormal dynamic reset deviation peak values ​​lead to abnormal static reset deviation, as well as the probability that abnormal dynamic reset deviation change rate leads to abnormal static reset deviation. Among the abnormal dynamic reset deviation, the risk of failure leading to abnormal dynamic reset deviation peak values ​​is relatively high, and the probability of failure leading to abnormal static reset deviation is also relatively high. The risk of failure leading to abnormal dynamic reset deviation change rate is relatively low, and the probability of failure leading to abnormal static reset deviation is also relatively low. and It can analyze and obtain data based on deviation anomalies collected during the historical operation of the PCB depaneling machine; it collects e valid reset data from the historical operation data of the depaneling machine, records the number of times the dynamic reset deviation peak value anomaly occurs as e1, and records the number of times the dynamic reset deviation change rate anomaly occurs as e2; after the e1 reset processes with dynamic reset deviation peak value anomalies are completed, it detects the number of times the static reset deviation anomaly occurs as e3. If, after a reset process with an abnormal rate of change in dynamic reset deviation is completed, the number of times an abnormal static reset deviation occurs is e4, then... Perform W reset data sampling and analysis and Take the mean of W sampling analyses and Used in the correlation prediction process;

[0022] When only dynamic reset deviation anomalies are detected, an increase in the frequency of dynamic reset deviation anomalies will lead to a higher probability of subsequent static reset deviation anomalies. There is a logical correlation between dynamic and static reset deviations; for example, if an anomaly occurs during shaft reset... The situation indicates that there is localized jamming during the movement of the shaft, which prevents the shaft from accurately reaching its theoretical position. According to statistics from the collected historical operating data of the depaneling machine, when... When the occurrence of deviations greater than 0.01 mm occurs 10 times or more, the probability of subsequent static reset deviation values ​​exceeding the threshold increases from 2% under normal circumstances to over 65%. By storing the number of abnormal dynamic reset deviations and the number of times static deviations exceed the standard in each correlation prediction period through a time series database, a linear regression model is established, for example, p=0.065(x+y)+0.02. Based on the analysis of the collected historical data, the influence coefficient b=0.065 is obtained and used in the subsequent correlation prediction process.

[0023] Further, in step S3: set the grading thresholds p1 and p2 for the static reset deviation abnormal probability; compare and analyze the calculated factual probability with the grading probability thresholds to predict the probability that the static reset deviation exceeds the threshold:

[0024] If p ≤ p1, predict that the probability of the static reset deviation abnormality occurring is low, and prompt the staff to repair the fault causing the dynamic reset deviation abnormality according to the fault repair prompt within the fault handling time limit of ΔT, and check for the upcoming static reset deviation abnormality;

[0025] The reserved fault handling time limit ΔT is a dynamically changing value, and the real-time fault handling time limit is calculated according to the following formula:

[0026] ;

[0027] The peak abnormality in the dynamic reset deviation abnormality has a relatively serious impact on the normal operation of the board splitting machine and is likely to cause damage to the board splitting machine. Therefore, the higher the number of occurrences of the peak abnormality, the shorter the reserved fault handling time limit, reducing the risk of damage to the board splitting machine equipment;

[0028] If p1 < p ≤ p2, predict that the probability of the static reset deviation abnormality occurring is relatively high, then immediately issue a prompt to prompt the staff to repair the fault causing the dynamic reset deviation abnormality according to the fault repair prompt within the fault handling time limit of ΔT / 2, and check for the upcoming static reset deviation abnormality; if no fault check is performed within the reserved fault handling time limit, generate a shutdown instruction to control the board splitting machine to stop and wait for fault check;

[0029] If p > p2, predict that the probability of the static reset deviation abnormality occurring is extremely high, and immediately generate a shutdown instruction to control the board splitting machine to stop and wait for fault check and repair.

[0030] Further, in step S3: set the mean threshold of the static reset deviation and the fluctuation value threshold u0 of the static reset deviation; after the end of a reset process, analyze the abnormal type of the static reset deviation:

[0031] If and , judge that the static reset deviation is normal;

[0032] If but , it means that the fluctuation value of the static reset deviation is abnormal, and judge that there is a fault of vibration offset after the shaft body is reset when the reset is completed; generate a fault repair prompt a1;

[0033] If but This indicates that the average static reset deviation is abnormal, indicating that the shaft reset is incomplete when the reset is completed; a fault repair prompt a2 is generated.

[0034] like and This indicates that both the static reset deviation fluctuation value and the static reset deviation mean value are abnormal, indicating that the fault of incomplete shaft reset and vibration offset after shaft reset occurs simultaneously after the reset is completed; a fault repair prompt a3 is generated.

[0035] Furthermore, in step S4: based on the generated fault repair prompts, staff are reminded to troubleshoot and repair the fault.

[0036] If a fault repair prompt r1 is generated, the cause of the shaft jamming may be foreign object stuck in the guide rail, partial wear of the lead screw, or bearing damage; then the staff is prompted to investigate and repair the cause of the shaft jamming fault.

[0037] If fault repair prompt r2 is generated, the cause of the intermittent interference fault in the encoder signal may be that the position loop gain of the servo driver is too high, the reset speed setting is unreasonable, etc.; then the staff is prompted to investigate and repair the cause of the intermittent interference fault in the encoder signal.

[0038] If fault repair prompt r3 is generated, it will prompt the staff to investigate and repair the cause of the shaft jamming and intermittent interference of the encoder signal.

[0039] If fault repair prompt a1 is generated, it will prompt the staff to repair the vibration offset fault after the shaft is reset within the fault handling time limit of ΔT / 2.

[0040] If fault repair prompt a2 is generated, it will prompt the staff to repair the fault of incomplete shaft reset within the fault handling time limit of ΔT / 4.

[0041] If fault repair prompt a3 is generated, a stop command will be generated immediately to stop the splitter, prompting the staff to repair the faults of incomplete shaft reset and vibration offset after shaft reset.

[0042] The COB depaneling machine fault monitoring system based on data analysis includes a shaft data acquisition module, a feature extraction module, an anomaly monitoring module, and a fault diagnosis and prompting module.

[0043] The shaft data acquisition module is used to collect the dynamic reset deviation value of the shaft in real time during the shaft reset process, and to collect the static reset deviation value of the shaft at regular intervals after the reset is completed.

[0044] The feature extraction module is used to obtain feature data that can be used to distinguish fault types from the collected dynamic reset deviation value and static reset deviation value data;

[0045] The anomaly monitoring module is used to monitor whether dynamic and static reset deviations are abnormal in real time during the shaft reset process, and to set up a dynamic and static reset deviation correlation prediction mechanism. When only dynamic reset deviation occurs, the probability of real-time static reset deviation anomaly is calculated, and static reset deviation anomaly is predicted.

[0046] The fault diagnosis and prompting module is used to generate fault repair prompts corresponding to the reset deviation abnormality based on the generated fault repair prompt type.

[0047] Furthermore, the shaft data acquisition module includes a dynamic reset deviation acquisition unit and a static reset deviation acquisition unit; the dynamic reset deviation acquisition unit is used to acquire dynamic reset deviation values ​​at fixed intervals during the shaft reset process; the static reset deviation acquisition unit is used to acquire static reset deviation values ​​within the data acquisition time after the reset is completed.

[0048] The feature extraction module includes a dynamic reset deviation feature analysis unit and a static reset deviation feature analysis unit; the dynamic reset deviation feature analysis unit is used to obtain the peak value of dynamic reset deviation and the rate of change of dynamic reset deviation from the collected dynamic reset deviation values; the static reset deviation unit is used to obtain the fluctuation value of static reset deviation and the average value of static reset deviation from the collected static reset deviation values.

[0049] Furthermore, the anomaly monitoring module includes a static reset deviation anomaly monitoring unit, a dynamic reset deviation monitoring unit, and a dynamic-static joint prediction unit. The static reset deviation anomaly monitoring unit is used to analyze the fault type based on the detected static reset deviation anomaly types and generate corresponding fault troubleshooting prompts. The dynamic reset deviation anomaly monitoring unit is used to analyze the fault type based on the detected dynamic reset deviation anomaly types and generate corresponding fault troubleshooting prompts. The dynamic-static joint prediction unit, by setting a dynamic-static reset deviation correlation prediction mechanism, predicts the probability of a static reset deviation anomaly occurring based on the degree of influence of the dynamic reset deviation anomaly on the probability of a static reset deviation anomaly occurring, and generates corresponding fault troubleshooting prompts. The fault troubleshooting prompt module is used to generate fault repair prompt content corresponding to the reset deviation anomaly based on the generated fault repair prompt type.

[0050] Compared with the prior art, the beneficial effects of the present invention are:

[0051] This application achieves real-time monitoring of incomplete reset faults in COB depaneling machines by collecting dynamic reset deviations during shaft reset and static reset deviations after each reset. By analyzing real-time collected reset deviation anomalies, fault monitoring efficiency and accuracy are improved. Furthermore, this application introduces a dynamic-static correlation prediction mechanism. Within the correlation prediction period, if only dynamic reset deviation anomalies are detected, the probability of subsequent static reset deviations is predicted based on the frequency of dynamic reset deviation occurrences. Among the detected dynamic reset deviation anomalies, there may be false dynamic reset deviation anomalies caused by encoder cable interference. This application analyzes the duration of dynamic reset deviation anomalies to determine whether the dynamic reset deviation is a false dynamic deviation anomaly. The system eliminates false dynamic reset deviations and introduces different probabilities of static reset deviations caused by different types of dynamic reset deviations, improving the accuracy of dynamic-static correlation prediction probability. Based on the predicted probability of dynamic-static correlation, different troubleshooting schemes are selected, achieving the effect of predicting the occurrence of static reset deviation anomalies in advance. This improves the timeliness and foresight of the PCB depaneling machine's fault monitoring, avoiding the lag in troubleshooting and repair when a fault with a significant impact on the depaneling machine's operation is detected. Peak anomalies in dynamic reset deviations have a serious impact on the normal operation of the depaneling machine and can easily cause damage. A dynamically adjusted fault troubleshooting time limit is set; the higher the frequency of dynamic reset deviation peak anomalies, the shorter the reserved fault handling time limit, reducing the risk of depaneling machine equipment damage. Attached Figure Description

[0052] Figure 1 This is a schematic diagram of the COB depaneling machine fault monitoring system based on data analysis according to the present invention;

[0053] Figure 2 This is a schematic diagram of the dynamic and static joint prediction method of the COB depaneling machine fault monitoring method based on data analysis of the present invention. Detailed Implementation

[0054] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0055] like Figures 1-2 As shown, this invention provides a technical solution: a COB depaneling machine fault monitoring method based on data analysis, the method comprising the following steps:

[0056] S1. During the shaft reset process, the dynamic reset deviation value of the shaft is collected in real time, and the static reset deviation value of the shaft is collected periodically after the reset is completed.

[0057] S2. Obtain feature data that can be used to distinguish fault types from the collected dynamic reset deviation value and static reset deviation value data;

[0058] S3. During the shaft reset process, monitor in real time whether dynamic reset deviation and static reset deviation are abnormal, and set up a dynamic and static reset deviation correlation prediction mechanism to predict the probability of real-time static reset deviation abnormality when only dynamic reset deviation occurs, and predict the occurrence of static reset deviation abnormality.

[0059] S4. Based on the generated fault repair prompt type, generate fault repair prompt content corresponding to the reset deviation abnormality.

[0060] In step S1: During the shaft reset process, starting from the reset start time t1 and ending at the reset completion time, a dynamic reset deviation value is collected every time interval Δt. The collected dynamic reset deviation values ​​are recorded as {s1, s2, ..., s...} n After the reset is completed, within the data acquisition duration Δd, m static reset deviations are collected periodically, and the collected static reset deviations are denoted as {v1, v2, ..., v...}. m}

[0061] In step S2: Analyze the collected dynamic reset deviation data, and obtain the peak value in the dynamic reset deviation data based on the comparative analysis of the same group of data. And analyze the real-time rate of change of the dynamic reset deviation value. Furthermore, the real-time collected dynamic reset deviation values ​​are compared and analyzed with the mean. Dynamic reset deviation values ​​greater than the mean are marked as fluctuation deviation values, and the frequency q of the fluctuation deviation values ​​is recorded. The real-time rate of change of the dynamic reset deviation values ​​is calculated according to the following formula: ;in This represents the rate of change of the j-th deviation, where j represents the deviation rate of change number, j=1, 2, ..., n-1; and i represents the dynamic reset deviation value number, i=2, 3, ..., n.

[0062] Analyze the collected static reset deviation data and calculate the mean value of the static reset deviation data. And calculate the fluctuation value of the static reset deviation according to the following formula: Where u represents the static reset deviation fluctuation value, and k represents the static reset deviation value number, k=1,2,…,m.

[0063] In step S3: Set the peak threshold η and the rate of change threshold ω0 for the dynamic reset deviation value; set the frequency threshold q0 for the fluctuation deviation value; analyze the anomaly types of the dynamic reset deviation data:

[0064] like but The dynamic reset deviation was determined to be normal during the reset process;

[0065] like but This indicates an abnormal peak value of dynamic reset deviation, indicating a shaft jamming fault during the reset process; a fault repair prompt r1 is generated.

[0066] like but This indicates an abnormal rate of change in the dynamic reset deviation, indicating an intermittent encoder signal interference fault during the reset process; a fault repair prompt r2 is generated.

[0067] like but This indicates that both the peak value and the rate of change of the dynamic reset deviation are abnormal, indicating that the shaft is stuck and the encoder signal is intermittently interfered with during the shaft reset process; a fault repair prompt r3 is generated.

[0068] In step S3: During the shaft reset process, the number of times the dynamic reset deviation value exceeds the threshold is monitored and recorded in real time. The number of occurrences is denoted as x, and the duration of the abnormal peak value of dynamic reset deviation {Δα1, Δα2, ... Δα} is recorded. x} will appear The number of occurrences is denoted as y, and the duration of the abnormal peak value of dynamic reset deviation {Δβ1, Δβ2, ... Δβ} is recorded. y For detected dynamic reset deviation anomalies, the duration of the dynamic reset deviation anomaly is used to determine whether it is a false dynamic reset deviation anomaly. A threshold for the duration of a true dynamic reset deviation anomaly is set as Δρ. The acquired duration of the dynamic reset deviation anomaly is compared with the set threshold. Dynamic reset deviation anomalies with a duration higher than the set threshold are marked as true dynamic reset deviation anomalies, and those with a duration lower than or equal to the set threshold are marked as false dynamic reset deviation anomalies. The number of true dynamic reset deviation peak anomalies x' and the number of true dynamic reset deviation change rate anomalies y' are recorded in the comparative analysis.

[0069] A dynamic and static reset deviation correlation prediction mechanism is set, and the correlation prediction period is set to [duration value]. , within an associated prediction cycle, if only dynamic reset bias anomalies occur, after each reset is completed, calculate the probability of static reset bias occurrence using the number of detected dynamic reset bias anomalies. Calculate the probability of static reset bias occurrence according to the following formula:

[0070] ;

[0071] Among them, p represents the probability that the static reset bias changing in real time exceeds the threshold; c is the probability of static reset bias anomaly occurring without dynamic reset anomalies; b represents the influence coefficient of the occurrence times of dynamic reset bias anomalies on the probability of static reset bias occurrence; represents the average probability of static bias anomaly caused by a single deviation peak exceeding the threshold; represents the average probability of static reset bias anomaly caused by the abnormal real-time change rate of a single dynamic reset bias value.

[0072] In step S3: Set the classification thresholds p1 and p2 for the probability of static reset bias anomaly; compare and analyze the calculated actual probability with the classification probability thresholds to predict the probability that the static reset bias exceeds the threshold:

[0073] If p ≤ p1, predict that the probability of static reset bias anomaly occurrence is low, and prompt the staff to, within the fault handling time limit of , repair the fault causing the dynamic reset bias anomaly according to the fault repair prompt, and conduct a check on the upcoming static reset bias anomaly;

[0074] The reserved fault handling time limit ΔT is a dynamically changing value. Calculate the real-time fault handling time limit according to the following formula:

[0075] ;

[0076] If p1 < p ≤ p2, predict that the probability of static reset bias anomaly occurrence is relatively high, then immediately issue a prompt to prompt the staff to, within the fault handling time limit of ΔT / 2, repair the fault causing the dynamic reset bias anomaly according to the fault repair prompt, and conduct a check on the upcoming static reset bias anomaly; if no fault check is carried out within the reserved fault handling time limit, generate a shutdown instruction to control the sub-board machine to stop and wait for fault check; [[ID=3));

[0077] If p > p2, predict that the probability of static reset bias anomaly occurrence is extremely high, and immediately generate a shutdown instruction to control the sub-board machine to stop and wait for fault check and repair.

[0078] In step S3: Set the static reset bias mean threshold and the fluctuation value threshold u0 of the static reset bias; after the end of a reset process, analyze the abnormal type of the static reset bias:

[0079] like and The static reset deviation was determined to be normal.

[0080] like but This indicates that the static reset deviation fluctuation value is abnormal, and it is determined that a fault of vibration offset after shaft reset occurs when the reset is completed; a fault repair prompt a1 is generated.

[0081] like but This indicates that the average static reset deviation is abnormal, indicating that the shaft reset is incomplete when the reset is completed; a fault repair prompt a2 is generated.

[0082] like and This indicates that both the static reset deviation fluctuation value and the static reset deviation mean value are abnormal, indicating that the fault of incomplete shaft reset and vibration offset after shaft reset occurs simultaneously after the reset is completed; a fault repair prompt a3 is generated.

[0083] In step S4: Based on the generated fault repair prompts, staff are reminded to troubleshoot and repair the fault.

[0084] If a fault repair prompt r1 is generated, it will prompt the staff to investigate and repair the cause of the shaft jamming fault;

[0085] If fault repair prompt r2 is generated, it will prompt the staff to investigate and repair the cause of the intermittent interference fault in the encoder signal;

[0086] If fault repair prompt r3 is generated, it will prompt the staff to investigate and repair the cause of the shaft jamming and intermittent interference of the encoder signal.

[0087] If fault repair prompt a1 is generated, it will prompt the staff to repair the vibration offset fault after the shaft is reset within the fault handling time limit of ΔT / 2.

[0088] If fault repair prompt a2 is generated, it will prompt the staff to repair the fault of incomplete shaft reset within the fault handling time limit of ΔT / 4.

[0089] If fault repair prompt a3 is generated, a stop command will be generated immediately to stop the splitter, prompting the staff to repair the faults of incomplete shaft reset and vibration offset after shaft reset.

[0090] COB PCB splitting machine fault monitoring system based on data analysis, such as Figure 1As shown, the system includes a shaft data acquisition module, a feature extraction module, an anomaly monitoring module, and a fault diagnosis and prompting module;

[0091] The shaft data acquisition module is used to collect the dynamic reset deviation value of the shaft in real time during the shaft reset process, and to collect the static reset deviation value of the shaft at regular intervals after the reset is completed;

[0092] The feature extraction module is used to extract feature data that can be used to distinguish fault types from the collected dynamic reset deviation value and static reset deviation value data;

[0093] The anomaly monitoring module is used to monitor whether dynamic and static reset deviations are abnormal in real time during the shaft reset process, and to set up a dynamic and static reset deviation correlation prediction mechanism. When only dynamic reset deviation occurs, it calculates the probability of real-time static reset deviation anomaly and predicts static reset deviation anomaly.

[0094] The fault diagnosis and prompt module is used to generate fault repair prompts corresponding to the reset deviation abnormality based on the generated fault repair prompt type.

[0095] The shaft data acquisition module includes a dynamic reset deviation acquisition unit and a static reset deviation acquisition unit. The dynamic reset deviation acquisition unit is used to acquire dynamic reset deviation values ​​at fixed intervals during the shaft reset process. The static reset deviation acquisition unit is used to acquire static reset deviation values ​​within the data acquisition time after the reset is completed.

[0096] The feature extraction module includes a dynamic reset deviation feature analysis unit and a static reset deviation feature analysis unit. The dynamic reset deviation feature analysis unit is used to obtain the peak value of the dynamic reset deviation and the rate of change of the dynamic reset deviation from the collected dynamic reset deviation values. The static reset deviation unit is used to obtain the fluctuation value of the static reset deviation and the average value of the static reset deviation from the collected static reset deviation values.

[0097] The anomaly monitoring module includes a static reset deviation anomaly monitoring unit, a dynamic reset deviation monitoring unit, and a dynamic-static joint prediction unit. The static reset deviation anomaly monitoring unit analyzes the fault type based on the detected static reset deviation anomalies and generates corresponding fault troubleshooting prompts. The dynamic reset deviation anomaly monitoring unit analyzes the fault type based on the detected dynamic reset deviation anomalies and generates corresponding fault troubleshooting prompts. The dynamic-static joint prediction unit, by setting a dynamic-static reset deviation correlation prediction mechanism, predicts the probability of a static reset deviation anomaly occurring based on the degree to which the probability of a static reset deviation anomaly is affected by the dynamic reset deviation anomaly, and generates corresponding fault troubleshooting prompts. The fault troubleshooting prompt module generates fault repair prompts corresponding to the reset deviation anomalies based on the generated fault repair prompt type.

[0098] Example 1: In step S1: During the shaft reset process, starting from the reset start time t1 and ending at the reset completion time, a dynamic reset deviation value is collected every 10ms for a duration Δt = 10ms. The collected dynamic reset deviation values ​​are recorded as {s1 = 0.014mm, s2 = 0.008mm, s3 = 0.01mm, s4 = 0.008mm, s5 = 0.008mm, s6 = 0.008mm}. After the reset is completed, within a data acquisition duration Δd = 1s, m = 5 static reset deviations are collected periodically. The collected static reset deviations are recorded as {v1 = 0.016mm, v2 = 0.01mm, v3 = 0.01mm, v4 = 0.016mm, v5 = 0.01mm}.

[0099] In step S2: Analyze the collected dynamic reset deviation data, and obtain the peak value in the dynamic reset deviation data based on the comparative analysis of the same group of data. =0.014mm, and analyze the real-time rate of change of the dynamic reset deviation value. Furthermore, the real-time collected dynamic reset deviation values ​​are compared and analyzed with the mean. Dynamic reset deviation values ​​greater than the mean are marked as fluctuation deviation values, and the frequency q of the fluctuation deviation values ​​is recorded. The real-time rate of change of the dynamic reset deviation values ​​is calculated according to the following formula: ;in This represents the rate of change of the j-th deviation, where j represents the deviation rate of change number, j=1,2,…,n-1; and i represents the dynamic reset deviation value number, i=1,2,…,n. mm / ms; 0; ; mm / ms;

[0100] Analyze the collected static reset deviation data and calculate the mean value of the static reset deviation data. =0.0124mm, and calculate the fluctuation value of the static reset deviation according to the following formula: u=0.03286.

[0101] Set the peak threshold for dynamic reset deviation η = 0.01 mm and the rate of change threshold for dynamic reset deviation ω0 = 0.0005 mm / ms; analyze the anomaly types of the dynamic reset deviation data:

[0102] but but This indicates that the peak value of the dynamic reset deviation is abnormal, indicating that a shaft jamming fault has occurred during the reset process; a fault repair prompt r3 is generated.

[0103] but The static reset deviation was determined to be normal.

[0104] A dynamic and static reset deviation correlation prediction mechanism is set up, with the correlation prediction period set to ΔT0=12h. Within one correlation prediction period, if only dynamic reset deviation anomalies occur, the probability of static reset deviation occurrence is calculated after each reset using the number of monitored dynamic reset deviation anomalies. The probability of static reset deviation occurrence is calculated using the following formula:

[0105] ;

[0106] Set the threshold for the mean static reset deviation. =0.02mm and the fluctuation threshold of static reset deviation u0=0.001mm; After a reset process is completed, the abnormal types of static reset deviation and the causes of the abnormality are analyzed: where c=0.02 is the probability of static reset deviation abnormality when there is no dynamic reset abnormality; b=0.065 represents the influence coefficient of the number of occurrences of dynamic reset deviation abnormality on the probability of static reset deviation occurrence, which is calculated through historical fault monitoring data; f1=0.9, f2=0.3; x'=1, y'=1, p=0.098;

[0107] Set graded thresholds p1=0.4 and p2=0.8 for the abnormal probability of static reset deviation; compare and analyze the calculated factual probability with the graded probability thresholds to predict the probability of static reset deviation exceeding the thresholds:

[0108] If p≤p1, the probability of abnormal static reset deviation is low, suggesting that staff should pay attention to the time interval. Within the fault handling time limit, repair the faults that cause abnormal dynamic reset deviation according to the fault repair prompts, and investigate the abnormal static reset deviation that is about to occur.

[0109] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

Claims

1. A COB paneling machine fault monitoring method based on data analysis, characterized in that: The method comprises the following steps: S1, collecting the dynamic reset deviation value of the shaft body in real time during the shaft body reset process, and collecting the static reset deviation value of the shaft body at regular intervals after the reset is completed; S2, obtaining characteristic data capable of distinguishing fault types from the collected dynamic reset deviation value and static reset deviation value data; S3, monitoring whether the dynamic reset deviation and the static reset deviation are abnormal in real time during the shaft body reset process, and setting a dynamic and static reset deviation correlation prediction mechanism to predict the probability of occurrence of real-time static reset deviation abnormality when only the dynamic reset deviation is present, and to predict the occurrence of static reset deviation abnormality; S4, generating fault repair prompt content corresponding to the reset deviation abnormality according to the generated fault repair prompt type.

2. The data analysis based COB panelizer fault monitoring method according to claim 1, characterized in that: In step S1: during the shaft body resetting process, from the resetting starting time t1 to the resetting completion time t2, the dynamic resetting deviation value is collected every time interval Δt, and the collected dynamic resetting deviation value is recorded as {s1, s2, …, s n m}; after the resetting is completed, m static resetting deviations are collected in the data collection time interval Δd, and the collected static resetting deviations are recorded as {v1, v2, …, v m m}. In step S2: analyzing the collected dynamic reset bias data, and obtaining the peak value in the dynamic reset bias data according to the comparison and analysis of the same group data ; And analyze the real-time change rate of dynamic reset deviation value ; And compare and analyze the real-time collected dynamic reset deviation value with the average value, mark the dynamic reset deviation value greater than the average value as fluctuation deviation value, and record the fluctuation deviation value frequency q; The real-time change rate of the dynamic reset bias value is calculated according to the following formula: ; wherein represents the jth bias change rate, j represents the bias change rate number, j = 1, 2, …, n-1; i represents the dynamic reset bias value number, i = 2, 3, …, n; The collected static reset deviation data is analyzed, and the mean value of the static reset deviation data is calculated , and the fluctuation value of the static reset deviation is calculated according to the following formula: ; wherein u represents the static reset deviation fluctuation value, k represents the static reset deviation value number, k = 1, 2…, m.

3. The data analysis based COB panelizer fault monitoring method according to claim 2, characterized in that: In step S3: setting a dynamic reset offset value peak threshold value η and a dynamic reset offset value change rate threshold value ; Set the fluctuation deviation value frequency threshold q0; analyze the abnormal type of the dynamic reset deviation data: If But , the dynamic reset deviation in the reset process is determined to be normal; If But , indicating a dynamic reset deviation peak anomaly, judging that the shaft body jam fault occurs during the reset process; Generate fault repair prompt r1; If But , indicating that the dynamic reset deviation change rate is abnormal, it is judged that there is an intermittent interference fault of the encoder signal in the reset process; a fault repair prompt r2 is generated; If But , indicating that the dynamic reset deviation peak value and the dynamic reset deviation change rate are both abnormal, judging that the shaft body reset process simultaneously appears shaft body jam and intermittent interference fault of encoder signal; generating fault repair prompt r3.

4. The data analysis based COB panelizer fault monitoring method according to claim 3, characterized in that: In step S3: During the shaft reset process, the number of times the dynamic reset deviation value exceeds the threshold is monitored and recorded in real time. The number of occurrences is denoted as x, and the duration of the abnormal peak value of dynamic reset deviation {Δα1, Δα2, ... Δα} is recorded. x } will appear The number of occurrences is denoted as y, and the duration of the abnormal peak value of dynamic reset deviation {Δβ1, Δβ2, ... Δβ} is recorded. y The duration of the dynamic reset deviation anomaly is used to determine whether the detected dynamic reset deviation anomaly is a false dynamic reset deviation anomaly. A threshold for the duration of the true dynamic reset deviation anomaly is set as Δρ. The duration of the acquired dynamic reset deviation anomaly is compared with the set threshold. Dynamic reset deviation anomalies with a duration higher than the set threshold are marked as true dynamic reset deviation anomalies, and those with a duration lower than or equal to the set threshold are marked as false dynamic reset deviation anomalies. The number of true dynamic reset deviation peak anomalies x' and the number of true dynamic reset deviation change rate anomalies y' are recorded in the comparative analysis. The dynamic static reset deviation correlation prediction mechanism is set, and the length of the correlation prediction period is set as In one correlation prediction period, if only dynamic reset deviation anomalies occur, the probability of static reset deviation occurrence is calculated after each reset is completed by using the number of monitored dynamic reset deviation anomalies, and the probability of static reset deviation occurrence is calculated according to the following formula: ； Wherein, p represents the probability of real-time change of static reset deviation exceeding threshold value; c is the probability of static reset deviation anomaly when there is no dynamic reset anomaly; b represents the influence coefficient of the number of dynamic reset deviation anomalies on the probability of static reset deviation; Indicates the average probability of static deviation anomaly caused by single deviation peak value exceeding threshold value; Indicates the average probability of static reset deviation anomaly caused by real-time change rate anomaly of single dynamic reset deviation value.

5. The data analysis based COB panelizer fault monitoring method according to claim 4, characterized in that: In step S3: set the hierarchical threshold p1 and p2 of the probability of static reset deviation abnormality; compare and analyze the actual probability obtained by calculation with the hierarchical probability threshold, and predict the probability of static reset deviation exceeding the threshold: If p≤p1, the probability of occurrence of static reset deviation abnormality is low, prompting the staff to repair the fault causing the dynamic reset deviation abnormality within the fault handling time limit of ΔT, and to investigate the static reset deviation abnormality that will occur; The reserved fault handling time limit ΔT is a dynamic value, which is calculated in real time according to the following formula: ； If p1<p≤p2, the probability of occurrence of static reset deviation abnormality is high, so an immediate prompt is sent to prompt the staff to repair the fault causing the dynamic reset deviation abnormality within the fault handling time limit of ΔT / 2, and to investigate the static reset deviation abnormality that will occur; If no fault investigation is performed within the reserved fault handling time limit, a shutdown instruction is generated to control the splitter to shut down and wait for fault investigation and repair; If p>p2, the probability of occurrence of static reset deviation abnormality is extremely high, and a shutdown instruction is immediately generated to control the splitter to shut down and wait for fault investigation and repair.

6. The data analysis based COB panelizer fault monitoring method according to claim 5, characterized in that: In step S3: setting a static reset offset mean value threshold and a static reset offset fluctuation value threshold u0; after the end of one reset procedure, analyzing the static reset offset for abnormal types: If and , it is determined that the static reset deviation is normal. If But , indicating that the static reset deviation fluctuation value is abnormal, judging that the fault of vibration offset after the shaft body reset occurs when the reset is completed; generating a fault repair prompt a1; If But , indicating that the static reset deviation mean is abnormal, judging that the axis body reset is not complete when the reset is completed; generating a fault repair prompt a2; If and , it indicates that both the static reset deviation fluctuation value and the static reset deviation mean value are abnormal, it is judged that the shaft body reset is not complete and the vibration offset after the shaft body reset occurs at the same time after the reset is completed; a fault repair prompt a3 is generated.

7. The data analysis based COB panelizer fault monitoring method according to claim 6, characterized in that: In step S4: according to the generated fault repair prompt, remind the staff to investigate and repair the fault: If the fault repair prompt r1 is generated, prompt the staff to investigate and repair the reason causing the shaft body jam fault; If the fault repair prompt r2 is generated, prompt the staff to investigate and repair the reason causing the intermittent interference fault of the encoder signal; If the fault repair prompt r3 is generated, prompt the staff to investigate and repair the reason causing the shaft body jam and the intermittent interference fault of the encoder signal; If the fault repair prompt a1 is generated, prompt the staff to repair the fault of the shaft body vibration offset after reset within the fault handling time limit of ΔT / 2. If the fault repair prompt a2 is generated, the staff is prompted to repair the fault of incomplete shaft body reset within the fault handling time limit of ΔT / 4; If the fault repair prompt a3 is generated, a shutdown instruction is immediately generated to control the board machine to shut down, prompting the staff to repair the fault of incomplete shaft body reset and vibration offset after shaft body reset.

8. A COB paneling machine fault monitoring system based on data analysis, characterized in that: The system comprises a shaft body data acquisition module, a feature extraction module, an abnormality monitoring module, and a fault troubleshooting prompt module. The shaft body data acquisition module is configured to acquire dynamic reset deviation values of the shaft body in real time during the shaft body reset process, and to acquire static reset deviation values of the shaft body at regular intervals after the reset is completed. The feature extraction module is configured to obtain feature data capable of distinguishing fault types from the acquired dynamic reset deviation values and static reset deviation value data. The abnormality monitoring module is configured to monitor whether the dynamic reset deviation and the static reset deviation are abnormal in real time during the shaft body reset process, and to set a dynamic-static reset deviation correlation prediction mechanism to calculate the probability of occurrence of the static reset deviation abnormality when only the dynamic reset deviation is abnormal, and to predict the static reset deviation abnormality. The fault troubleshooting prompt module is configured to generate fault repair prompt contents corresponding to the reset deviation abnormality according to the generated fault repair prompt types.

9. The data analysis based COB panelizer fault monitoring system of claim 8, wherein: The shaft body data acquisition module comprises a dynamic reset deviation acquisition unit and a static reset deviation acquisition unit. The dynamic reset deviation acquisition unit is configured to acquire dynamic reset deviation values at fixed interval time lengths during the shaft body reset process. The static reset deviation acquisition unit is configured to acquire static reset deviation values within a data acquisition time length after the reset is completed. The feature extraction module comprises a dynamic reset deviation feature analysis unit and a static reset deviation feature analysis unit. The dynamic reset deviation feature analysis unit is configured to obtain a dynamic reset deviation peak value and a dynamic reset deviation change rate from the acquired dynamic reset deviation values. The static reset deviation unit is configured to obtain a static reset deviation fluctuation value and a static reset deviation mean value from the acquired static reset deviation values.

10. The data analysis based COB panelizer fault monitoring system of claim 8, wherein: The abnormality monitoring module comprises a static reset deviation abnormality monitoring unit, a dynamic reset deviation monitoring unit, and a dynamic-static joint prediction unit. The static reset deviation abnormality monitoring unit is configured to analyze the generated fault type according to the monitored static reset deviation abnormality type, and to generate a corresponding fault troubleshooting prompt. The dynamic reset deviation abnormality monitoring unit is configured to analyze the generated fault type according to the monitored dynamic reset deviation abnormality type, and to generate a corresponding fault troubleshooting prompt. The dynamic-static joint prediction unit predicts the probability of occurrence of the static reset deviation abnormality by setting a dynamic-static reset deviation correlation prediction mechanism when detecting the dynamic reset deviation abnormality, and generates a corresponding fault troubleshooting prompt. The fault troubleshooting prompt module is configured to generate fault repair prompt contents corresponding to the reset deviation abnormality according to the generated fault repair prompt types.

Images