Health calculation method based on laser welding machine

Abstract

The present invention discloses a method for calculating the health of a laser welder. This method collects historical data on the normal operation of each component of the laser welder, selects data from the historical data showing normal operation of the laser welder under a set laser welder load, and uses it as a feature extraction dataset. Each data item in the dataset is classified as a normal data item or a non-normal data item. The mean and standard deviation of the historical data for each data item in the dataset are calculated, a reasonable fluctuation range for each data item is selected, and real-time operating data streams of each component of the laser welder are collected at intervals. The data under a set laser power load is stored as data to be analyzed. The health of each data item is calculated based on the data to be analyzed, and the health of the laser welder equipment and the overall health of the region are calculated from the health of each data item. This calculation method uses algorithm modeling based on the operating data of the laser welder to provide a data analysis foundation for laser welder maintenance and overhaul, ensuring the normal and reliable operation of the laser welder.

Description

Technical Field

[0001] The present invention relates to the technical field of equipment status monitoring and fault detection, and in particular to a health calculation method based on a laser welder. Background Art

[0002] Laser welders utilize high-energy laser pulses to locally heat a tiny area of a material. The energy from the laser radiation diffuses through heat conduction into the material, melting it to form a specific molten pool. This allows for spot welding, butt welding, stitch welding, and seal welding, and is widely used in various welding production lines. For example, metallurgical enterprises utilize laser welders in key production lines such as pickling and hot-dip galvanizing, where their status directly impacts the smooth operation of these critical lines.

[0003] As a device that integrates laser, machinery, electricity, hydraulics, instruments and automation, the laser welder has a large daily maintenance workload and is technically difficult to troubleshoot. Many hidden faults and signs are difficult to detect through daily inspections, making it difficult to control the status of the laser welder. Due to the cross-influence of multiple disciplines, troubleshooting is more difficult.

[0004] Currently, welding machine manufacturers focus on manufacturing, energy conservation, and automation, but lack research into equipment status monitoring and intelligent diagnostics. Currently, these systems only provide simple real-time fault alarms on the human-machine interface. Maintenance personnel are unable to determine the laser welder's true health status and only address faults when they occur, which is time-consuming and labor-intensive, and also impacts the unit's production schedule. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a health calculation method based on a laser welder. This calculation method performs algorithm modeling based on the operating data of the laser welder, quantitatively presents the health of each key data item of the laser welder, and achieves the purpose of real-time monitoring and continuous tracking of the key parts of the laser welder, providing a data analysis basis for the maintenance and inspection of the laser welder, and ensuring the normal and reliable operation of the laser welder.

[0006] To solve the above technical problems, the present invention provides a method for calculating the health of a laser welder, comprising the following steps:

[0007] Step 1: Collect historical data on the normal operation of the laser welder body, laser source high-frequency cabinet, turbine fan inverter, laser source resonant cavity, preheating device, and annealing device. Set the laser welder load to a laser power ≥ 8000 watts. Filter out the data on the normal operation of the laser welder under the set laser welder load from the historical data as the feature extraction data set.

[0008] Step 2: Use qqnorm and shapiro.test tools to determine the normality of each data item in the data set and classify each data item as normal or non-normal.

[0009] Step 3: Calculate the mean of the historical data of each data item in the data set and standard deviations S1, S2, ..., S i , where i is the number of data items;

[0010] Step 4: Draw the distribution diagram of each data item in the data set, based on the empirical threshold and standard deviation S for judging the status of the equipment i , select the reasonable fluctuation range of each data item, denoted as R1, R2, ..., R i , where R i =n×S i n=1,2,3;

[0011] Step 5: Collect the real-time operation data stream of the laser welder body, laser source high-frequency cabinet, turbine fan inverter, laser source resonant cavity, preheating device, and annealing device, judge the equipment operation status according to the set laser power, and store the data under the set laser power load as the data to be analyzed;

[0012] Step 6: Extract the data to be analyzed every half hour. If the number of times the laser welder is running in the data to be analyzed is ≥10 times, calculate the health level. The calculation rules are as follows:

[0013] If the data item to be analyzed is a normal data item, the Z statistic calculation formula is used, based on R i Calculate the value of the Z statistic and, based on the cumulative probability density function of the normal distribution, calculate the probability that the data item falls within the normal range, which is the health level H;

[0014] If the data item to be analyzed is non-normal, use the step function to calculate. If it is within the range, the health level H is set to 95. If not, the health level H is obtained by deducting points based on the degree of deviation of the data item.

[0015] Step 7: Calculate the laser welder equipment health and the regional total health through the health of each data item H. Set the weight W according to the importance of the data item. The weight setting rule is that the more important the data item, the higher the corresponding weight W, and it also meets the following requirements: Among them, equ is the component equipment of the laser welding machine, and dataitem is the corresponding data item under each component equipment;

[0016] Laser welding equipment health H equ And the total health of the region Harea The calculation rules are as follows:

[0017]

[0018]

[0019] Among them, H dataitem It indicates the health of the corresponding data item under the component equipment.

[0020] Furthermore, the period for collecting historical data on normal operation of the laser welder's welder body, laser source high-frequency cabinet, turbine fan inverter, laser source resonant cavity, preheating device, and annealing device is 3 to 6 months.

[0021] Furthermore, the calculated laser welding equipment health H equ And the total health of the region H area Data is uploaded to the backend intelligent maintenance platform through the Kafka distributed publish-subscribe messaging system for real-time health monitoring and display of historical health changes.

[0022] The present invention adopts the above-mentioned technical solution based on the health calculation method of the laser welder. The method collects historical data of the normal operation of each component of the laser welder, selects data of the normal operation of the laser welder under a set laser welder load from the historical data as a feature extraction data set, classifies each data item in the data set as a normal data item or a non-normal data item, calculates the mean and standard deviation of the historical data of each data item in the data set, selects a reasonable fluctuation range for each data item, collects the real-time operation data stream of each component of the laser welder at intervals, and stores the data under a set laser power load as the data to be analyzed; calculates the health of each data item based on the data to be analyzed, and calculates the health of each data item from the health of each data item to obtain the health of the laser welder equipment and the overall health of the region. The calculation method uses the operation data of the laser welder to perform algorithm modeling, quantitatively presents the health of each key data item of the laser welder, and achieves the purpose of real-time monitoring and continuous tracking of key parts of the laser welder, providing a data analysis basis for maintenance and repair of the laser welder, and ensuring the normal and reliable operation of the laser welder. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments:

[0024] Figure 1 This is a flowchart of the health calculation method of the laser welder according to the present invention. DETAILED DESCRIPTION

[0025] Implementation example Figure 1 As shown, the health calculation method of the laser welder according to the present invention includes the following steps:

[0026] Step 1: Collect historical data on the normal operation of the laser welder body, laser source high-frequency cabinet, turbine fan inverter, laser source resonant cavity, preheating device, and annealing device. Set the laser welder load to a laser power ≥ 8000 watts. Filter out the data on the normal operation of the laser welder under the set laser welder load from the historical data as the feature extraction data set.

[0027] Step 2: Use qqnorm and shapiro.test tools to determine the normality of each data item in the data set and classify each data item as normal or non-normal.

[0028] Step 3: Calculate the mean of the historical data of each data item in the data set and standard deviations S1, S2, ..., S i , where i is the number of data items;

[0029] Step 4: Draw the distribution diagram of each data item in the data set, based on the empirical threshold and standard deviation S for judging the status of the equipment i , select the reasonable fluctuation range of each data item, denoted as R1, R2, ..., R i , where R i =n×S i n=1,2,3;

[0030] Step 5: Collect the real-time operation data stream of the laser welder body, laser source high-frequency cabinet, turbine fan inverter, laser source resonant cavity, preheating device, and annealing device, judge the equipment operation status according to the set laser power, and store the data under the set laser power load as the data to be analyzed;

[0031] Step 6: Extract the data to be analyzed every half hour. If the number of times the laser welder is running in the data to be analyzed is ≥10 times, calculate the health level. The calculation rules are as follows:

[0032] If the data item to be analyzed is a normal data item, the Z statistic calculation formula is used, based on R i Calculate the value of the Z statistic and, based on the cumulative probability density function of the normal distribution, calculate the probability that the data item falls within the normal range, which is the health level H;

[0033] If the data item to be analyzed is non-normal, use the step function to calculate. If it is within the range, the health level H is set to 95. If not, the health level H is obtained by deducting points based on the degree of deviation of the data item.

[0034] Step 7: Calculate the laser welder equipment health and the regional total health through the health of each data item H. Set the weight W according to the importance of the data item. The weight setting rule is that the more important the data item, the higher the corresponding weight W, and it also meets the following requirements: Among them, equ is the component equipment of the laser welding machine, and dataitem is the corresponding data item under each component equipment;

[0035] Laser welding equipment health H equ And the total health of the region H area The calculation rules are as follows:

[0036]

[0037]

[0038] Among them, H dataitem It indicates the health of the corresponding data item under the component equipment.

[0039] Preferably, the period for collecting historical data on normal operation of the laser welder's welder body, laser source high-frequency cabinet, turbine blower frequency converter, laser source resonant cavity, preheating device, and annealing device is 3 to 6 months.

[0040] Preferably, the calculated laser welding equipment health H equ And the total health of the region H area Data is uploaded to the backend intelligent maintenance platform through the Kafka distributed publish-subscribe messaging system for real-time health monitoring and display of historical health changes.

[0041] This method monitors the status of multiple components and multiple data items of a laser welding machine group, and quantifies and presents them in the form of health. It can be used to display the status of welding equipment, making it easier for maintenance personnel to locate abnormal equipment and data items, so as to formulate a reasonable preventive maintenance plan, thereby ensuring the normal and reliable operation of the laser welding machine.

Claims

1. A method for calculating the health of a laser welder, characterized in that This calculation method includes the following steps: Step 1: Collect historical data on the normal operation of the laser welder body, laser source high-frequency cabinet, turbine fan inverter, laser source resonant cavity, preheating device, and annealing device. Set the laser welder load to a laser power ≥ 8000 watts. Filter out the data on the normal operation of the laser welder under the set laser welder load from the historical data as the feature extraction data set. Step 2: Use qqnorm and shapiro.test tools to determine the normality of each data item in the data set and classify each data item as normal or non-normal. Step 3: Calculate the mean of the historical data of each data item in the data set and standard deviations S1, S2, ..., S i , where i is the number of data items; Step 4: Draw the distribution diagram of each data item in the data set, based on the empirical threshold and standard deviation S for judging the status of the equipment i , select the reasonable fluctuation range of each data item, denoted as R1, R2, ..., R i , where R i =n×S i n=1,2,3; Step 5: Collect the real-time operation data stream of the laser welder body, laser source high-frequency cabinet, turbine fan inverter, laser source resonant cavity, preheating device, and annealing device, judge the equipment operation status according to the set laser power, and store the data under the set laser power load as the data to be analyzed; Step 6: Extract the data to be analyzed every half hour. If the number of times the laser welder is running in the data to be analyzed is ≥10 times, calculate the health level. The calculation rules are as follows: If the data item to be analyzed is a normal data item, the Z statistic calculation formula is used, based on R i Calculate the value of the Z statistic and, based on the cumulative probability density function of the normal distribution, calculate the probability that the data item falls within the normal range, which is the health level H; If the data item to be analyzed is non-normal, use the step function to calculate. If it is within the range, the health level H is set to 95. If not, the health level H is obtained by deducting points based on the degree of deviation of the data item. Step 7: Calculate the laser welder equipment health and the regional total health through the health of each data item H. Set the weight W according to the importance of the data item. The weight setting rule is that the more important the data item, the higher the corresponding weight W, and it also meets the following requirements: Among them, equ is the component equipment of the laser welding machine, and dataitem is the corresponding data item under each component equipment; Laser welding equipment health H equ And the total health of the region H area The calculation rules are as follows: Among them, H dataitem It indicates the health of the corresponding data item under the component equipment.

2. The method for calculating the health of a laser welder according to claim 1, wherein: The period for collecting the historical data of the normal operation of the laser welder body, laser source high frequency cabinet, turbine fan inverter, laser source resonant cavity, preheating device, and annealing device is 3 to 6 months.

3. The method for calculating the health of a laser welder according to claim 1, wherein: Calculated laser welding equipment health H equ And the total health of the region H area Data is uploaded to the backend intelligent maintenance platform through the Kafka distributed publish-subscribe messaging system for real-time health monitoring and display of historical health changes.

Images