A welding equipment fault early warning method based on edge computing

Abstract

The application discloses a welding equipment fault early warning method based on edge computing and relates to the technical field of welding equipment operation monitoring, which comprises the following steps: an edge computing unit reads the work correlation data of each welding equipment according to a detection period, forms a work detection period range according to the work execution state change, and forms a welding work record used for correlating the work detection period range, the welding equipment, the welded piece, the ventilation path and the welding work position; air component detection values and ventilation migration parameters of each ventilation path are read to form air component migration effective records and air component abnormal detection period numbers. According to the welding work record formed according to the detection period, the welding work record is simultaneously correlated with the work detection period range, the welding equipment, the welded piece, the ventilation path and the welding work position, so that subsequent air component migration analysis, welded piece response analysis and equipment working state analysis can all return to the same detection period system for corresponding analysis.

Description

Technical Field

[0001] This invention relates to the field of welding equipment operation monitoring technology, specifically a welding equipment fault early warning method based on edge computing. Background Technology

[0002] In industrial welding production scenarios, multiple welding machines typically operate together according to workstation layout, welded part flow path, and ventilation path. During the welding process, changes in air composition such as welding fumes, carbon monoxide, and ozone are generated. At the same time, the welded parts exhibit thermal and vibration responses, and the welding equipment itself experiences changes in operating status such as welding current, welding voltage, wire feed speed, and equipment temperature rise. To ensure stable welding operations, the production site typically uses edge computing units to collect welding equipment operation data, ambient air composition data, welded part response data, and equipment operating status data on a detection cycle, and uses this data for operation monitoring and fault warning. In existing welding equipment fault early warning methods, a common approach is to directly read the equipment's operating status data, such as current, voltage, temperature rise, and wire feed speed, of a single welding machine and compare this data with a preset benchmark range. When the equipment's operating status data exceeds the corresponding range, an early warning message is generated. This method can perform basic monitoring of the welding equipment's own operating status and can also meet the needs of identifying equipment anomalies under some stable operating conditions. However, in scenarios where multiple machines, multiple workstations, and multiple ventilation paths operate together, abnormal air composition generated by welding operations may migrate along the ventilation path. There are overlapping relationships between different welding operation records in terms of time range, ventilation path, and the location of the welded part. When judging solely based on the equipment's operating status data, it is difficult to clearly define the correspondence between abnormal air composition, the response of the welded part, and the deviation of the equipment status. Therefore, under edge computing conditions, there is a need for a welding equipment fault early warning method that can generate welding operation records according to the detection cycle and continuously correlate the migration process of air components along the ventilation path, the cycle of abnormal source, the response position change process of the welded part, the corresponding welding operation position process, and the deviation process of the welding equipment working state. Therefore, this invention proposes a welding equipment fault early warning method based on edge computing. Summary of the Invention

[0003] The purpose of this invention is to provide a fault early warning method for welding equipment based on edge computing, so as to solve the problems mentioned in the background art.

[0004] This invention can be achieved through the following technical solution: a welding equipment fault early warning method based on edge computing, comprising: Step 1: The edge computing unit reads the operation-related data of each welding equipment according to the detection cycle, forms the operation detection cycle range based on the changes in operation execution status, and generates welding operation records for associating the operation detection cycle range, welding equipment, welded parts, ventilation path, and welding operation position. Step 2: Read the air composition detection values ​​and ventilation migration parameters for each ventilation path. When the air composition detection value exceeds the preset allowable range and the peak value extends along the ventilation migration direction, a valid record of air composition migration and an abnormal air composition detection cycle number are generated. The number of ventilation arrival cycles is generated based on the ventilation migration parameters. The source detection cycle range is generated by backtracking based on the abnormal air composition detection cycle number and the number of ventilation arrival cycles. Step 3: Read the records from the welding operation records that correspond to the valid records of air component migration and whose operation detection cycle range corresponds to the source detection cycle range, to form the first candidate welding operation record set, and form the second candidate welding operation record set based on the number of ventilation arrival cycles and the cycle overlap relationship; Step 4: Read the response data of the workpiece corresponding to the second candidate welding operation record to form a thermal response position sequence and a vibration response position sequence, and form a third candidate welding operation record set based on the position correspondence between the thermal response position sequence, the vibration response position sequence and the welding operation record. Step 5: Read the equipment working status record of the welding equipment corresponding to the third candidate welding operation record, compare it with the welding operation benchmark status range to form the equipment working deviation record, and determine the fault risk equipment number based on the third candidate welding operation record and the equipment working deviation record.

[0005] A further technical improvement of the present invention is that: when forming the welding operation record in step one, the operation inspection cycle range is segmented according to the changes in the ventilation path within the operation inspection cycle range, so that the operation inspection cycle range in each welding operation record corresponds to the same ventilation path, including: Read the ventilation path corresponding to each inspection cycle within the operation inspection cycle range of the same welding equipment; The current work segment cycle range is opened with the first detection cycle as the starting detection cycle, and the ventilation path corresponding to the first detection cycle is written. Compare the ventilation path corresponding to the next detection cycle with the ventilation path written in the current operation segment cycle range; if the comparison result is consistent, write the next detection cycle into the current operation segment cycle range. If the consistent comparison result is not valid, the previous inspection cycle of the next inspection cycle is determined as the termination inspection cycle of the current operation segment cycle range. Welding operation record is formed based on the current operation segment cycle range, and a new operation segment cycle range is started with the next inspection cycle as the starting inspection cycle. The ventilation path corresponding to the next inspection cycle is written. When creating a welding operation record, if the ventilation path associated with the current operation segment cycle range is consistent with the ventilation path associated with the operation segment cycle range that has formed a termination inspection cycle, and there are operation segment cycle ranges associated with different ventilation paths between the current operation segment cycle range and the operation segment cycle range that has formed a termination inspection cycle, welding operation records are created separately, and the current operation segment cycle range and the operation segment cycle range that has formed a termination inspection cycle are not merged.

[0006] A further technical improvement of the present invention is that the method for determining the peak value's extension along the ventilation migration direction in step two includes: Read the air composition detection values ​​at each air composition sampling location within the same ventilation path according to the ventilation migration direction; The detection period segment where the air composition detection value exceeds the preset allowable range of air composition is identified as a candidate peak segment, and the starting detection period, peak detection period and fallback detection period of the candidate peak segment are written respectively. Read the candidate peak fragment from the first air component sampling position upstream of the ventilation migration direction, and write the earliest candidate peak fragment of the starting detection cycle into the deferred chain; read the candidate peak fragments from the next air component sampling position in the ventilation migration direction in sequence, and take the last candidate peak fragment in the deferred chain as the previous candidate peak fragment. When there is a candidate peak segment in the subsequent air composition sampling position whose initial detection period, peak detection period, and fallback detection period are all later than the detection period corresponding to the previous candidate peak segment, the candidate peak segment that satisfies the detection period delay relationship and has the earliest initial detection period will be written into the delay chain. When the delay chain covers the sampling locations of each air component arranged along the ventilation migration direction within the same ventilation path, the peak value is determined to be delayed along the ventilation migration direction.

[0007] A further technical improvement of the present invention is that the step of regressing to form the source detection cycle range based on the air composition anomaly detection cycle number and the number of ventilation arrival cycles in step two includes: Read the air composition sampling location, initial detection period, fallback detection period, and number of ventilation arrival periods corresponding to each candidate peak segment in the deferred chain; For each candidate peak segment, ventilation migration parameters are read cycle by cycle in descending order of detection cycle number, starting from the initial detection cycle, and the number of backtracking detection cycles is accumulated. When the accumulated number of backtracking detection cycles reaches the corresponding number of ventilation arrival cycles, the current detection cycle is determined as the source starting detection cycle of the candidate peak segment. For the same candidate peak segment, starting from the fallback detection cycle, the ventilation migration parameters are read cycle by cycle in descending order of the detection cycle number, and the number of fallback detection cycles is accumulated. When the accumulated number of fallback detection cycles reaches the corresponding number of ventilation arrival cycles, the current detection cycle is determined as the source of the candidate peak segment and the detection cycle is terminated. Read the source start detection period and source end detection period corresponding to each candidate peak segment in the deferred chain, determine the source start detection period with the latest detection period number as the start boundary of the source detection period range, and determine the source end detection period with the earliest detection period number as the end boundary of the source detection period range. When the starting boundary of the source detection cycle range is no later than the ending boundary of the source detection cycle range, the source detection cycle range is formed based on the starting boundary and the ending boundary of the source detection cycle range.

[0008] A further technical improvement of the present invention is that the step of forming a second candidate welding operation record set based on the overlap relationship between the number of ventilation arrival cycles and the cycles in step three includes: Read the welding operation record corresponding to the valid record of the same air composition migration from the first candidate welding operation record set; Welding operation records with identical welding equipment, identical welded parts, and identical ventilation paths are written into the same periodic through-check set. Read the start and end boundaries of the source detection cycle range, and read the operation detection cycle range of each welding operation record in the cycle-through verification set; The inspection cycle range of each welding operation record in the cycle-through verification set is arranged in chronological order according to the inspection cycle number. When the starting boundary of the source detection cycle range falls within any operation detection cycle range in the cycle penetration verification set, the ending boundary of the source detection cycle range falls within any operation detection cycle range in the cycle penetration verification set, and there is no detection cycle within the source detection cycle range that is not covered by the operation detection cycle range in the cycle penetration verification set, the welding operation record in which the operation detection cycle range in the cycle penetration verification set and the source detection cycle range have a periodic overlap relationship is written into the second candidate welding operation record set.

[0009] A further technical improvement of the present invention is that step four, forming the thermal response position sequence and the vibration response position sequence, includes: Read the thermal response and vibration response values ​​of the welded parts at each response acquisition location according to the detection cycle of the second candidate welding operation record; Within the same testing cycle, the dominant position of thermal response is determined based on the ranking of thermal response values, and the dominant position of vibration response is determined based on the ranking of vibration response values. Read the preset acquisition location adjacency list and write the dominant thermal response location and dominant vibration response location corresponding to the first detection cycle within the operation detection cycle range into the thermal response location sequence and vibration response location sequence, respectively. According to the order of the detection cycle numbers, the dominant positions of thermal response and vibration response in the later detection cycle are compared with the dominant positions of thermal response and vibration response in the previous detection cycle for consistency and adjacency. When the consistent comparison result of the dominant position of thermal response is valid, the dominant position of thermal response corresponding to the next detection cycle is written into the thermal response position sequence. If the consistent comparison result of the dominant thermal response position is not valid but the adjacent comparison result is valid, the dominant thermal response position corresponding to the next detection cycle is written into the thermal response position sequence. When the consistent comparison result of the dominant vibration response position is valid, the dominant vibration response position corresponding to the next detection cycle is written into the vibration response position sequence. If the consistent comparison result of the dominant vibration response position is not valid but the adjacent comparison result is valid, the dominant vibration response position corresponding to the next detection cycle is written into the vibration response position sequence.

[0010] A further technical improvement of the present invention is that step four, forming a third candidate welding operation record set, includes: Read the welding operation position corresponding to each detection cycle within the operation detection cycle range of the second candidate welding operation record; The current work position holding section is started with the first inspection cycle within the work inspection cycle range, and the welding work position corresponding to the first inspection cycle is written into the current work position holding section. The welding positions corresponding to adjacent inspection cycles are compared consistently. If the comparison result is successful, the next inspection cycle is written into the current operation position holding segment. If the comparison result is unsuccessful, the previous inspection cycle of the next inspection cycle is determined as the termination inspection cycle of the current operation position holding segment, and a new operation position holding segment is started in the next inspection cycle. The welding position corresponding to the next inspection cycle is written into the new operation position holding segment. For each work position holding section, the distance value from the welding work position to each response acquisition position is calculated according to the detection cycle, and the work position distance is sorted in ascending order of distance value for the corresponding detection cycle. Read the thermal response dominance position and vibration response dominance position in the working position distance sorting results within the same detection period from the thermal response location sequence and vibration response location sequence; When neither the thermal response distance sequence nor the vibration response distance sequence within the work position holding section advances to the sequence with increasing distance values, the position correspondence result of the work position holding section is formed; When all operation positions in the same second candidate welding operation record form a position correspondence in the same segment, the second candidate welding operation record is written into the third candidate welding operation record set.

[0011] A further technical improvement of the present invention is that the step of forming the equipment operation deviation record in step five includes: Read the equipment operating status record of the welding equipment corresponding to the third candidate welding operation record within the operation detection cycle range; Compare each equipment working status item in the equipment working status record with the corresponding status item reference range in the welding operation reference status range; When the status value of a device's operating status item is higher than the upper limit of the corresponding status item's reference range, the status item boundary deviation value is formed based on the difference between the status value and the upper limit of the corresponding status item's reference range. When the status value of a device's operating status item is lower than the lower limit of the corresponding status item's reference range, the status item boundary deviation value is formed based on the difference between the lower limit of the corresponding status item's reference range and the status value. When the status value of a device's operating status item is within the corresponding status item's reference range, the status item's boundary deviation value is written as zero. The cumulative boundary deviations of each status item within the same detection cycle are used to form the equipment cycle deviation value. When the equipment cycle deviation values ​​corresponding to consecutively arranged testing cycles according to the testing cycle number all increase, an equipment operation deviation record is formed based on the corresponding equipment cycle deviation value.

[0012] Compared with the prior art, the present invention has the following beneficial effects: This invention generates welding operation records according to the detection cycle, and simultaneously associates the welding operation records with the operation detection cycle range, welding equipment, welded parts, ventilation path, and welding operation location. This allows subsequent air composition migration analysis, welded part response analysis, and equipment operating status analysis to all be returned to the same detection cycle system for corresponding analysis. Since the operation detection cycle range in the welding operation records corresponds to the same ventilation path, the correspondence between air composition anomalies and welding operation records is clearer, which can reduce the source judgment bias caused by the mixing of records across ventilation paths. Furthermore, this invention combines the delay process of air composition detection values ​​along the ventilation migration direction with ventilation migration parameters, so that air composition anomalies can be traced back to their source cycle according to the number of ventilation arrival cycles. As a result, air composition anomalies no longer correspond only to the instantaneous detection cycle of a certain sampling location, but can be back-linked to the source cycle range that may cause the anomaly, so that the corresponding records read from the welding operation records have a clear cycle basis. On the other hand, the present invention further correlates the change in the response position of the welded part with the welding operation position, and after the correspondence is established, reads the equipment working status record of the corresponding welding equipment to form an equipment working deviation record, so that the determination of the fault risk equipment number has the data basis of air composition migration, source cycle back, response position correspondence and equipment status deviation. The fault risk equipment number formed in this way is not a direct result of a single equipment status overrun, but is determined by multiple types of detection data converging step by step according to the detection cycle, and the data source and judgment path are clearer. Attached Figure Description

[0013] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0014] Figure 1 This is a schematic diagram of the method logic of the present invention. Detailed Implementation

[0015] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided.

[0016] Please see Figure 1 As shown, the present invention provides a welding equipment fault early warning method based on edge computing, including: Step 1: The edge computing unit reads the operation-related data of each welding equipment according to the detection cycle, forms the operation detection cycle range based on the changes in operation execution status, and generates welding operation records for associating the operation detection cycle range, welding equipment, welded parts, ventilation path, and welding operation position. Specifically, the operation-related data includes welding start / stop switch values, welding current sampling values, valid welding torch movement markers, welding equipment number, welded part number, welding torch endpoint coordinates, weld segment number, weld segment position sequence number, and ventilation path number. Each inspection cycle corresponds to an inspection cycle number, which increases sequentially according to the order in which the inspection cycles occur.

[0017] The edge computing unit reads the welding start / stop switch quantity, welding current sampling value, and welding torch movement validity marker in each detection cycle. The operation start level is written by the effective level configuration table of the welding equipment start / stop control signal. The preset effective welding current range includes a lower limit and an upper limit. When the welding current sampling value is not less than the lower limit and not greater than the upper limit, the welding current sampling value is determined to be within the preset effective welding current range. The valid movement marker is written by the welding torch movement detection module when the change in the welding torch endpoint coordinate is not less than the preset lower limit of movement change. When the welding start / stop switch quantity is at the operation start level, the welding current sampling value is within the preset effective welding current range, and the welding torch movement validity marker is a valid movement marker, the operation execution state corresponding to that detection cycle is determined to be in the execution state; when at least one of the following conditions is met—the welding start / stop switch quantity is not at the operation start level, the welding current sampling value is not within the preset effective welding current range, or the welding torch movement validity marker is not a valid movement marker—the operation execution state corresponding to that detection cycle is determined to be in the non-execution state. Therefore, the operation status is determined by the welding start / stop switch quantity, the welding current sampling value, and the effective marker of the welding torch movement.

[0018] The edge computing unit reads the operation execution status of the same welding equipment within adjacent inspection cycles and compares the status changes of the subsequent inspection cycle with those of the previous inspection cycle. If the operation execution status of the first inspection cycle in the inspection cycle sequence is "executed," that first inspection cycle is designated as the start inspection cycle. If the operation execution status of the previous inspection cycle is "not executed" and the operation execution status of the subsequent inspection cycle is "executed," the subsequent inspection cycle is designated as the start inspection cycle. If the operation execution status of the previous inspection cycle is "executed" and the operation execution status of the subsequent inspection cycle is "not executed," the previous inspection cycle is designated as the end inspection cycle. If the last inspection cycle in the currently read inspection cycle sequence has been read, and the operation execution status corresponding to the last inspection cycle is still "executed," the last inspection cycle is designated as the end inspection cycle. The edge computing unit forms the operation inspection cycle range based on the start and end inspection cycles.

[0019] After establishing the operational inspection cycle range, the edge computing unit reads the welding equipment number, the part to be welded number, the welding torch endpoint coordinates, the weld segment number, the position sequence number within the weld segment, and the ventilation path number within that operational inspection cycle range. The edge computing unit determines the corresponding welding equipment based on the welding equipment number, the corresponding part to be welded based on the part to be welded number, the welding operation position for the corresponding inspection cycle based on the welding torch endpoint coordinates, the weld segment number, and the position sequence number within the weld segment, and the ventilation path for the corresponding inspection cycle based on the ventilation path number.

[0020] The edge computing unit reads the ventilation paths corresponding to each detection cycle within the operation detection cycle range in chronological order according to the detection cycle number. The edge computing unit starts the current operation segment cycle range with the first detection cycle within the operation detection cycle range as the starting detection cycle, and writes the ventilation path corresponding to the first detection cycle into the current operation segment cycle range. The current operation segment cycle range is used to record the detection cycles corresponding to the same ventilation path within the operation detection cycle range.

[0021] The edge computing unit continues to read the ventilation path corresponding to the next detection cycle and compares it with the ventilation path written in the current work segment cycle range. If the comparison result is consistent, the next detection cycle is written into the current work segment cycle range, and the unit continues to read the ventilation path corresponding to the next detection cycle. If the comparison result is inconsistent, the previous detection cycle of the next detection cycle is determined as the termination detection cycle of the current work segment cycle range, and a welding operation record is formed based on the current work segment cycle range. A new work segment cycle range is then started with the next detection cycle as the starting detection cycle, and the ventilation path corresponding to the next detection cycle is written into the new work segment cycle range.

[0022] When forming a welding operation record based on the current operation segment cycle range, the edge computing unit reads the welding equipment, welded parts, ventilation path, and welding operation position corresponding to each detection cycle within the current operation segment cycle range, and writes the current operation segment cycle range as the operation detection cycle range in the welding operation record. The resulting welding operation record is associated with the operation detection cycle range, welding equipment, welded parts, ventilation path, and welding operation position, and the operation detection cycle range in the welding operation record corresponds to the same ventilation path.

[0023] When creating a welding operation record, the edge computing unit reads the ventilation path associated with the current operation segment cycle range and the ventilation path associated with the operation segment cycle range that has already formed a termination inspection cycle. If the ventilation path associated with the current operation segment cycle range is the same as the ventilation path associated with the operation segment cycle range that has already formed a termination inspection cycle, and there are operation segment cycle ranges with different associated ventilation paths between the current operation segment cycle range and the operation segment cycle range that has already formed a termination inspection cycle, then a welding operation record is created separately based on the current operation segment cycle range, while the welding operation record corresponding to the operation segment cycle range that has already formed a termination inspection cycle remains independent, and the current operation segment cycle range and the operation segment cycle range that has already formed a termination inspection cycle are not merged.

[0024] After the last detection cycle within the job detection cycle range is read, if the current job segment cycle range has not yet formed a termination detection cycle, the edge computing unit determines the last detection cycle within the job detection cycle range as the termination detection cycle of the current job segment cycle range, and forms a welding job record based on the current job segment cycle range. Through this process, the specific acquisition fields in the job-related data are first read and used to form the job execution status, welding equipment, welded parts, welding job position, and ventilation path. Then, a welding job record is formed based on the job detection cycle range and the job segment cycle range, ensuring that the job detection cycle range in each welding job record corresponds to the same ventilation path.

[0025] Step 2: Read the air composition detection values ​​and ventilation migration parameters for each ventilation path. When the air composition detection value exceeds the preset allowable range and the peak value extends along the ventilation migration direction, a valid record of air composition migration and an abnormal air composition detection cycle number are generated. The number of ventilation arrival cycles is generated based on the ventilation migration parameters. The source detection cycle range is generated by backtracking based on the abnormal air composition detection cycle number and the number of ventilation arrival cycles. Specifically, the air composition detection values ​​include welding fume particle concentration, carbon monoxide concentration, and ozone concentration. Preset allowable ranges for air composition include allowable ranges for welding fume particle concentration, carbon monoxide concentration, and ozone concentration. The allowable ranges for welding fume particle concentration include both lower and upper limits; the allowable ranges for carbon monoxide concentration include both lower and upper limits; and the allowable ranges for ozone concentration include both lower and upper limits. The lower limits for welding fume particle concentration, carbon monoxide concentration, and ozone concentration are all set to zero. A preset lower limit for ventilation velocity is set based on the minimum effective exhaust velocity of the ventilation path, used to distinguish between detection cycles where air composition migration can occur and those where it cannot. A valid ventilation status means that within the corresponding detection cycle, the ventilation migration direction is consistent with the preset ventilation migration direction of the ventilation path, and the ventilation velocity value is not less than the preset lower limit of ventilation velocity. If at least one of the following conditions is met: the ventilation migration direction is inconsistent with the preset ventilation migration direction of the ventilation path, or the ventilation velocity value is less than the preset lower limit of ventilation velocity, the valid ventilation status is written as invalid. The edge computing unit reads the historical normal welding detection cycles from the historical storage records. The historical normal welding detection cycles are the detection cycles in the historical storage records where the valid ventilation status is valid and no equipment operation deviation records have been written. The edge computing unit reads the preset minimum calculation benchmark values ​​for welding fume particle concentration, carbon monoxide concentration, and ozone concentration. All preset minimum calculation benchmark values ​​for welding fume particle concentration, carbon monoxide concentration, and ozone concentration are positive values. When the number of historical normal welding inspection cycles is zero, the preset minimum calculation benchmark value for welding fume particle concentration is written into the allowable upper limit value for welding fume particle concentration, the preset minimum calculation benchmark value for carbon monoxide concentration is written into the allowable upper limit value for carbon monoxide concentration, and the preset minimum calculation benchmark value for ozone concentration is written into the allowable upper limit value for ozone concentration. When the number of historical normal welding inspection cycles is greater than zero, the edge computing unit reads the welding fume particle concentration value, carbon monoxide concentration value, and ozone concentration value within the historical normal welding inspection cycles. The larger of the maximum value of the welding fume particle concentration value and the preset minimum calculation benchmark value for welding fume particle concentration within the historical normal welding inspection cycles is written into the allowable upper limit value for welding fume particle concentration, the larger of the maximum value of the carbon monoxide concentration value and the preset minimum calculation benchmark value for carbon monoxide concentration within the historical normal welding inspection cycles is written into the allowable upper limit value for carbon monoxide concentration, and the larger of the maximum value of the ozone concentration value and the preset minimum calculation benchmark value for ozone concentration within the historical normal welding inspection cycles is written into the allowable upper limit value for ozone concentration.

[0026] The ventilation migration parameters include the ventilation migration direction, air component sampling location number, migration distance between adjacent air component sampling locations, ventilation velocity value, detection cycle duration, and ventilation effectiveness status. The air component sampling location number is written in the order of the air component sampling locations along the ventilation migration direction within the same ventilation path; the migration distance between adjacent air component sampling locations records the distance between adjacent air component sampling locations along the ventilation migration direction, in meters; the ventilation velocity value records the air velocity along the ventilation migration direction within the corresponding detection cycle, in meters per second; the detection cycle duration records the duration of each detection cycle, in seconds; and the ventilation effectiveness status is written based on the comparison result between the ventilation migration direction and the preset ventilation migration direction of the ventilation path, as well as the comparison result between the ventilation velocity value and the preset lower limit of ventilation velocity. The edge computing unit reads the ventilation velocity value and detection cycle duration corresponding to the detection cycle with effective ventilation status, multiplies the ventilation velocity value by the detection cycle duration, and forms the single detection cycle migration distance, in meters.

[0027] The edge computing unit reads the air composition detection values ​​from each air composition sampling location within the same ventilation path according to the detection cycle number. For the same air composition sampling location, if at least one of the following conditions is met: the concentration of welding fume particles is higher than the upper limit of the allowable concentration of welding fume particles, the concentration of carbon monoxide is higher than the upper limit of the allowable concentration of carbon monoxide, or the concentration of ozone is higher than the upper limit of the allowable concentration of ozone, the corresponding detection cycle is written into the out-of-range detection cycle. The edge computing unit reads adjacent out-of-range detection cycles according to the detection cycle number. Out-of-range detection cycles with consecutive detection cycle numbers are written into the same candidate peak segment; out-of-range detection cycles with non-consecutive detection cycle numbers are written into different candidate peak segments.

[0028] After forming candidate peak segments, the edge computing unit reads the earliest out-of-range detection cycle number in the candidate peak segment and writes this out-of-range detection cycle into the starting detection cycle of the candidate peak segment. The edge computing unit calculates the welding fume particle concentration exceedance difference, carbon monoxide concentration exceedance difference, and ozone concentration exceedance difference for each detection cycle within the candidate peak segment. When the corresponding air component detection value is higher than the corresponding allowable upper limit, the difference between the corresponding air component detection value and the corresponding allowable upper limit is written into the corresponding exceedance difference; when the corresponding air component detection value is not higher than the corresponding allowable upper limit, the corresponding exceedance difference is written as zero. The edge computing unit forms a welding fume particle concentration exceedance ratio based on the ratio of the welding fume particle concentration exceedance difference to the welding fume particle concentration allowable upper limit, a carbon monoxide concentration exceedance ratio based on the ratio of the carbon monoxide concentration exceedance difference to the carbon monoxide concentration allowable upper limit, and an ozone concentration exceedance ratio based on the ratio of the ozone concentration exceedance difference to the ozone concentration allowable upper limit. The edge computing unit accumulates the proportions of welding fume particle concentration exceeding the proportional value, carbon monoxide concentration exceeding the proportional value, and ozone concentration exceeding the proportional value within the same detection cycle to form the air composition excess amount for that detection cycle. If there is a detection cycle with the largest air composition excess amount within the candidate peak segment, this detection cycle is written into the peak detection cycle of the candidate peak segment. If there are two or more detection cycles with the largest air composition excess amounts within the candidate peak segment, the detection cycle with the earliest detection cycle number is written into the peak detection cycle of the candidate peak segment.

[0029] The edge computing unit continues to read air composition detection values ​​according to the detection cycle number after the candidate peak segment. When the concentration of welding fume particles is not higher than the upper limit of the allowable concentration of welding fume particles, the concentration of carbon monoxide is not higher than the upper limit of the allowable concentration of carbon monoxide, and the concentration of ozone is not higher than the upper limit of the allowable concentration of ozone, the detection cycle that first meets the above conditions is written into the fallback detection cycle of the candidate peak segment. If a fallback detection cycle has not been formed after the candidate peak segment and before the end of the currently read detection cycle, the candidate peak segment is not written into the deferred chain; the candidate peak segment is read again to participate in the deferred chain judgment after a fallback detection cycle is formed in a subsequent detection cycle.

[0030] The edge computing unit reads candidate peak segments from the first air composition sampling location upstream of the ventilation migration direction and writes the candidate peak segment with the earliest starting detection period and a falling detection period into the delay chain. Subsequently, the edge computing unit reads candidate peak segments from the next air composition sampling location in the ventilation migration direction and uses the last candidate peak segment in the delay chain as the previous candidate peak segment. If there is a candidate peak segment in the next air composition sampling location whose starting detection period, peak detection period, and falling detection period are all later than the detection period corresponding to the previous candidate peak segment, the candidate peak segment that satisfies the detection period delay relationship and has the earliest starting detection period is written into the delay chain; if there is no candidate peak segment that satisfies the detection period delay relationship in the next air composition sampling location, the writing of the delay chain downstream of that ventilation path is stopped. When the delay chain covers all air composition sampling locations arranged along the ventilation migration direction within the same ventilation path, the peak value is determined to be delayed along the ventilation migration direction, and a valid record of air composition migration is formed.

[0031] The effective record of air composition migration includes the corresponding ventilation path, the deferred chain, the air composition sampling location corresponding to each candidate peak segment in the deferred chain, the initial detection period, the peak detection period, and the fallback detection period. The air composition anomaly detection period number includes the initial detection period number, the peak detection period number, and the fallback detection period number corresponding to each candidate peak segment in the deferred chain. The edge computing unit writes the initial detection period, peak detection period, and fallback detection period corresponding to each candidate peak segment in the deferred chain into the air composition anomaly detection period number, so that the air composition anomaly detection period number can correspond to the anomaly start point, peak position, and fallback position of each candidate peak segment in the deferred chain.

[0032] The edge computing unit generates the number of ventilation arrival cycles based on ventilation migration parameters. For each candidate peak segment in the deferred chain, the edge computing unit determines whether the air component sampling position is the first air component sampling position upstream of the ventilation migration direction. If the air component sampling position corresponding to the candidate peak segment is the first air component sampling position upstream of the ventilation migration direction, the number of ventilation arrival cycles corresponding to that air component sampling position is written as zero. If the air component sampling position corresponding to the candidate peak segment is not the first air component sampling position upstream of the ventilation migration direction, the edge computing unit reads the migration distance between adjacent air component sampling positions from the first air component sampling position upstream of the ventilation migration direction to that air component sampling position, and accumulates them to form the position migration distance. The edge computing unit uses the starting detection cycle of the first candidate peak segment upstream in the deferred chain as the starting point for accumulation, and reads the ventilation effective status, ventilation velocity value, and detection cycle duration in ascending order of detection cycle number. When the ventilation effective status is effective, the single detection cycle migration distance is formed based on the ventilation velocity value and detection cycle duration, and the single detection cycle migration distance is accumulated to form the cumulative migration distance. When the cumulative migration distance is not less than the location migration distance for the first time, the number of accumulated effective detection cycles is determined as the number of ventilation arrival cycles corresponding to the air component sampling location.

[0033] After establishing the number of ventilation arrival cycles, the edge computing unit reads the air composition sampling location, initial detection cycle, fallback detection cycle, and number of ventilation arrival cycles corresponding to each candidate peak segment in the deferred chain. When the number of ventilation arrival cycles corresponding to a candidate peak segment is zero, the initial detection cycle of that candidate peak segment is determined as the source initial detection cycle, and the fallback detection cycle of that candidate peak segment is determined as the source termination detection cycle. When the number of ventilation arrival cycles corresponding to a candidate peak segment is not zero, the edge computing unit initializes the cumulative fallback detection cycle count to zero and reads the ventilation migration parameters cycle by cycle in descending order of the detection cycle number, starting from the initial detection cycle. If the ventilation effective status is valid in the read ventilation migration parameters, the cumulative fallback detection cycle count is increased by one detection cycle. When the cumulative fallback detection cycle count reaches the number of ventilation arrival cycles corresponding to the candidate peak segment, the current detection cycle is determined as the source initial detection cycle of that candidate peak segment.

[0034] For the same candidate peak segment where the number of ventilation arrival cycles is not zero, the edge computing unit initializes the cumulative backtracking detection cycle count to zero and reads ventilation migration parameters cycle by cycle in descending order of detection cycle number, starting from the fallback detection cycle. If the ventilation validity status is valid among the read ventilation migration parameters, the cumulative backtracking detection cycle count is increased by one detection cycle. When the cumulative backtracking detection cycle count reaches the number of ventilation arrival cycles corresponding to the candidate peak segment, the current detection cycle is determined as the source termination detection cycle for that candidate peak segment. By backtracking from both the starting detection cycle and the fallback detection cycle, the source start detection cycle and source termination detection cycle corresponding to the same candidate peak segment are formed.

[0035] After each candidate peak segment in the deferred chain forms a source start detection period and a source end detection period, the edge computing unit reads each source start detection period and each source end detection period. The source start detection period with the latest detection period number is determined as the starting boundary of the source detection period range, and the source end detection period with the earliest detection period number is determined as the ending boundary of the source detection period range. When the starting boundary of the source detection period range is not later than the ending boundary, the edge computing unit forms the source detection period range based on the starting and ending boundaries. When the starting boundary of the source detection period range is later than the ending boundary, no source detection period range is formed, and the welding operation record is not read based on the effective record of air composition migration to form the first candidate welding operation record set. The source detection period range is used to subsequently read records corresponding to the operation detection period range and the source detection period range from the welding operation records.

[0036] Step 3: Read the records from the welding operation records that correspond to the valid records of air component migration and whose operation detection cycle range corresponds to the source detection cycle range, to form the first candidate welding operation record set, and form the second candidate welding operation record set based on the number of ventilation arrival cycles and the cycle overlap relationship; Specifically, the effective record of air composition migration includes the corresponding ventilation path, the deferred chain, the air composition sampling location corresponding to each candidate peak segment in the deferred chain, the initial detection cycle, the peak detection cycle, and the fallback detection cycle; the source detection cycle range includes the starting boundary and the ending boundary of the source detection cycle range. The edge computing unit reads the operation detection cycle range, welding equipment, welded parts, ventilation path, and welding operation location from the welding operation record, and performs a consistency comparison between the ventilation path in the welding operation record and the corresponding ventilation path in the effective record of air composition migration.

[0037] After the ventilation path consistency comparison result is valid, the edge computing unit reads the operation detection cycle range from the welding operation record and performs a cycle correspondence comparison with the source detection cycle range. If the start detection cycle of the operation detection cycle range is not later than the end boundary of the source detection cycle range, and the end detection cycle of the operation detection cycle range is not earlier than the start boundary of the source detection cycle range, then the operation detection cycle range of the welding operation record is determined to correspond to the source detection cycle range. The edge computing unit writes the welding operation record for which the ventilation path consistency comparison result is valid and the operation detection cycle range corresponds to the source detection cycle range into the first candidate welding operation record set.

[0038] After forming the first candidate welding operation record set, the edge computing unit reads the welding operation records in the first candidate welding operation record set that correspond to the valid records of the same air component migration. For each read welding operation record, the edge computing unit reads the welding equipment, the welded part, and the ventilation path, and writes welding operation records with the same welding equipment, the same welded part, and the same ventilation path into the same periodic through-validation set. Thus, the welding operation records in the periodic through-validation set correspond to the same welding equipment, the same welded part, and the same ventilation path, and are used to perform through-validation on the source detection period range corresponding to the valid records of the same air component migration.

[0039] The edge computing unit reads the start and end boundaries of the source detection cycle range and the operation detection cycle range of each welding operation record in the cycle-through verification set. The edge computing unit arranges the operation detection cycle ranges of each welding operation record in the cycle-through verification set according to the detection cycle number, ensuring that the temporal arrangement of multiple welding operation records in the same cycle-through verification set is consistent with the detection cycle number order. After arrangement, the edge computing unit reads the start boundary of the source detection cycle range and checks whether the start boundary of each source detection cycle range falls within any operation detection cycle range in the cycle-through verification set; simultaneously, it reads the end boundary of the source detection cycle range and checks whether the end boundary of each source detection cycle range falls within any operation detection cycle range in the cycle-through verification set.

[0040] If the starting boundary of the source detection cycle range falls within any job detection cycle range in the cycle penetration verification set, and the ending boundary of the source detection cycle range also falls within any job detection cycle range in the cycle penetration verification set, the edge computing unit reads from the starting boundary to the ending boundary of the source detection cycle range according to the detection cycle number. For each detection cycle within the source detection cycle range, the edge computing unit determines whether the detection cycle falls within the job detection cycle range of at least one welding job record in the cycle penetration verification set. If every detection cycle within the source detection cycle range falls within at least one job detection cycle range in the cycle penetration verification set, then it is determined that there are no detection cycles within the source detection cycle range that are not covered by the job detection cycle ranges in the cycle penetration verification set.

[0041] When the starting boundary of the source detection cycle range falls within any operation detection cycle range in the cycle penetration verification set, and the ending boundary of the source detection cycle range falls within any operation detection cycle range in the cycle penetration verification set, and there are no detection cycles within the source detection cycle range that are not covered by the operation detection cycle ranges in the cycle penetration verification set, the edge computing unit reads each welding operation record in the cycle penetration verification set and reads each detection cycle within the operation detection cycle range of each welding operation record. The edge computing unit compares each detection cycle within the operation detection cycle range with the starting boundary and ending boundary of the source detection cycle range, respectively. When there is a detection cycle within the operation detection cycle range whose detection cycle number is no earlier than the starting boundary and no later than the ending boundary of the source detection cycle range, it is determined that the operation detection cycle range of the welding operation record has a cycle overlap relationship with the source detection cycle range. The edge computing unit writes the welding operation records in the cycle penetration verification set whose operation detection cycle range has a cycle overlap relationship with the source detection cycle range into the second candidate welding operation record set.

[0042] After the second candidate welding operation record set is formed, the edge computing unit saves the operation detection cycle range, welding equipment, welded parts, ventilation path and welding operation position of each welding operation record in the second candidate welding operation record set according to the detection cycle number, so that the second candidate welding operation record can be used to read the corresponding welded parts response data in subsequent steps.

[0043] Step 4: Read the response data of the workpiece corresponding to the second candidate welding operation record to form a thermal response position sequence and a vibration response position sequence, and form a third candidate welding operation record set based on the position correspondence between the thermal response position sequence, the vibration response position sequence and the welding operation record. Specifically, the response data of the welded part includes the response acquisition location number, response acquisition location coordinates, thermal response value, and vibration response value. The response acquisition location coordinates are used to record the position of the corresponding response acquisition location in the coordinate system of the welded part. The thermal response value is used to record the thermal response intensity of the corresponding response acquisition location within the corresponding detection cycle. The vibration response value is used to record the vibration response intensity of the corresponding response acquisition location within the corresponding detection cycle. The welding operation location in the second candidate welding operation record is determined by the welding torch endpoint coordinates, weld segment number, and position sequence number within the weld segment under the same coordinate system of the welded part. The welding operation location and the response acquisition location coordinates use the same length unit. The edge calculation unit reads the thermal response value and vibration response value of the welded part at each response acquisition location in the second candidate welding operation record according to the detection cycle number, so that each detection cycle has a corresponding response acquisition location, thermal response value, and vibration response value.

[0044] Within the same detection cycle, the edge computing unit arranges the response acquisition locations in descending order of thermal response values, forming a thermal response value sorting result. When two or more response acquisition locations correspond to the same thermal response value, they are arranged in ascending order of their response acquisition location numbers. The edge computing unit determines the response acquisition location that ranks first in the thermal response value sorting result as the dominant thermal response location. Similarly, the edge computing unit arranges the vibration response acquisition locations in descending order of vibration response values, forming a vibration response value sorting result. When two or more response acquisition locations correspond to the same vibration response value, they are arranged in ascending order of their response acquisition location numbers. The edge computing unit determines the response acquisition location that ranks first in the vibration response value sorting result as the dominant vibration response location.

[0045] The edge computing unit reads the preset upper limit of the adjacent distance between acquisition positions. This preset upper limit is written based on the installation spacing of adjacent response acquisition positions and is used to determine whether two response acquisition positions can form a positional connection within adjacent detection cycles. The edge computing unit reads the coordinates of any two response acquisition positions and calculates the acquisition position distance between them. If the acquisition position distance is not greater than the preset upper limit of the adjacent distance, the two response acquisition position numbers are written to each other's adjacent response acquisition position numbers. A preset acquisition position adjacency table is formed by the response acquisition position numbers and the adjacent response acquisition position numbers.

[0046] The edge computing unit reads the dominant thermal response position and dominant vibration response position corresponding to the first detection cycle within the job detection cycle range. It writes the dominant thermal response position and the dominant vibration response position of the first detection cycle into the thermal response position sequence and the vibration response position into the vibration response position sequence, respectively. Subsequently, the edge computing unit compares the dominant thermal response position and dominant vibration response position of the subsequent detection cycle with those of the previous detection cycle, according to the order of the detection cycle numbers.

[0047] When the consistent comparison result of the dominant thermal response position is true, the edge computing unit writes the dominant thermal response position corresponding to the next detection cycle into the thermal response position sequence; when the consistent comparison result of the dominant thermal response position is false, the edge computing unit reads the preset acquisition position adjacency table, performs an adjacency comparison between the dominant thermal response position corresponding to the next detection cycle and the dominant thermal response position corresponding to the previous detection cycle. When the adjacency comparison result is true, the dominant thermal response position corresponding to the next detection cycle is written into the thermal response position sequence; when the adjacency comparison result is false, the dominant thermal response position corresponding to the next detection cycle is not written into the thermal response position sequence.

[0048] When the comparison result of the dominant vibration response position is valid, the edge computing unit writes the dominant vibration response position corresponding to the next detection cycle into the vibration response position sequence. When the comparison result of the dominant vibration response position is invalid, the edge computing unit reads the preset acquisition position adjacency table, compares the dominant vibration response position corresponding to the next detection cycle with the dominant vibration response position corresponding to the previous detection cycle, and if the adjacency comparison result is valid, writes the dominant vibration response position corresponding to the next detection cycle into the vibration response position sequence; if the adjacency comparison result is invalid, it does not write the dominant vibration response position corresponding to the next detection cycle into the vibration response position sequence. Through this process, both the thermal response position sequence and the vibration response position sequence are formed by response acquisition positions with a detection cycle succession relationship.

[0049] After forming the thermal response position sequence and vibration response position sequence, the edge computing unit reads the welding operation positions corresponding to each detection cycle within the operation detection cycle range of the second candidate welding operation record. The edge computing unit starts the current operation position holding segment with the first detection cycle within the operation detection cycle range as the starting detection cycle, and writes the welding operation position corresponding to the first detection cycle into the current operation position holding segment. The edge computing unit reads the welding operation positions corresponding to the next detection cycle in the order of the detection cycle numbers, and compares the welding operation positions corresponding to the next detection cycle with those corresponding to the previous detection cycle. If the comparison result is true, the next detection cycle is written into the current operation position holding segment; if the comparison result is false, the previous detection cycle of the next detection cycle is determined as the termination detection cycle of the current operation position holding segment, and a new operation position holding segment is started with the next detection cycle, writing the welding operation position corresponding to the next detection cycle into the new operation position holding segment. After the last detection cycle within the operation detection cycle range is read, if the current operation position holding segment has not yet formed a termination detection cycle, the edge computing unit determines the last detection cycle within the operation detection cycle range as the termination detection cycle of the current operation position holding segment.

[0050] For each work position holding section, the edge computing unit reads the welding work positions within that section according to the detection cycle, and also reads the coordinates of each response acquisition position under the same welded part position coordinate system. The edge computing unit calculates the distance values ​​between the welding work positions and the coordinates of each response acquisition position, and arranges the response acquisition positions in ascending order of distance value to form the work position distance sorting result for the corresponding detection cycle; when the distance values ​​are the same, the corresponding response acquisition positions are arranged in ascending order of response acquisition position number.

[0051] The edge computing unit reads the dominant thermal response position and dominant vibration response position within the same detection cycle from the thermal response position sequence and vibration response position sequence, and reads the order of the dominant thermal response position in the working position distance sorting result to form a thermal response distance order; it also reads the order of the dominant vibration response position in the working position distance sorting result to form a vibration response distance order. If the dominant thermal response position and dominant vibration response position are not read simultaneously in any detection cycle within the working position holding section, no position correspondence result is formed for that working position holding section.

[0052] The edge computing unit reads the thermal response distance sequence and vibration response distance sequence corresponding to adjacent detection cycles within the same work position holding section according to the detection cycle number. A failure to advance the thermal response distance sequence to an increasing distance value means that the thermal response distance sequence corresponding to the later detection cycle is not greater than the thermal response distance sequence corresponding to the earlier detection cycle; similarly, a failure to advance the vibration response distance sequence to an increasing distance value means that the vibration response distance sequence corresponding to the later detection cycle is not greater than the vibration response distance sequence corresponding to the earlier detection cycle. If the work position holding section contains only one detection cycle, and this detection cycle can retrieve the dominant thermal response position from the thermal response position sequence and the dominant vibration response position from the vibration response position sequence, the edge computing unit forms the position correspondence result for this work position holding section based on the thermal response distance sequence and vibration response distance sequence corresponding to that detection cycle.

[0053] If the dominant thermal response position can be read from the thermal response position sequence and the dominant vibration response position can be read from the vibration response position sequence in each detection cycle within the operation position holding section, and if neither the thermal response distance sequence nor the vibration response distance sequence within the operation position holding section progresses to a sequence with increasing distance value, then the edge computing unit forms the position correspondence result for the operation position holding section.

[0054] The edge computing unit reads each job position holding segment from the same second candidate welding job record and determines whether each job position holding segment forms a position correspondence result. When each job position holding segment in the same second candidate welding job record forms a position correspondence result, the edge computing unit writes the second candidate welding job record into the third candidate welding job record set. After the third candidate welding job record set is formed, the edge computing unit retains the welding equipment and job detection cycle range corresponding to each third candidate welding job record for subsequent reading of the corresponding welding equipment's working status record.

[0055] Step 5: Read the equipment working status record of the welding equipment corresponding to the third candidate welding operation record, compare it with the welding operation benchmark status range to form the equipment working deviation record, and determine the fault risk equipment number based on the third candidate welding operation record and the equipment working deviation record. Specifically, the third candidate welding operation record includes the operation inspection cycle range, welding equipment, welded parts, ventilation path, and welding operation location. The edge computing unit reads the welding equipment and operation inspection cycle range from the third candidate welding operation record, and reads the equipment working status record of the welding equipment within the operation inspection cycle range according to the inspection cycle number. The equipment working status record includes equipment working status items and corresponding status values; the equipment working status items include welding current status items, welding voltage status items, wire feed speed status items, and equipment temperature rise status items. The welding current status item corresponds to the welding current status value, the welding voltage status item corresponds to the welding voltage status value, the wire feed speed status item corresponds to the wire feed speed status value, and the equipment temperature rise status item corresponds to the equipment temperature rise status value. The welding current status value is written from the welding current sampling value within the same inspection cycle. The welding current sampling value is used to form the operation execution status, and the welding current status value is used to form the equipment working deviation record.

[0056] The welding operation reference state range includes the reference range of each equipment operating state item. The corresponding state item reference ranges include the welding current state item reference range, welding voltage state item reference range, wire feed speed state item reference range, and equipment temperature rise state item reference range. The edge computing unit reads historical detection cycles from the historical storage records that are earlier than the corresponding operation detection cycle range of the third candidate welding operation record, are consistent with the welding equipment corresponding to the third candidate welding operation record, are consistent with the welded parts, and are in the execution state. It also reads detection cycles from the historical detection cycles that are not written into the equipment operation deviation record as reference detection cycles. The preset number of reference detection cycles is an integer not less than two. When the number of reference detection cycles is not less than the preset number of reference detection cycles, the edge computing unit forms the welding operation reference state range based on the reference detection cycles; when the number of reference detection cycles is less than the preset number of reference detection cycles, no equipment operation deviation record is formed corresponding to the third candidate welding operation record, and the fault risk equipment number is not determined based on the third candidate welding operation record.

[0057] The edge computing unit reads the welding current status value, welding voltage status value, wire feed speed status value, and equipment temperature rise status value within the reference detection cycle. It writes the minimum value of the welding current status value into the lower limit of the reference range for the welding current status item, and the maximum value into the upper limit; similarly, it writes the minimum value of the welding voltage status value into the lower limit and the maximum value into the upper limit; and it writes the minimum value of the wire feed speed status value into the lower limit and the maximum value into the upper limit; and finally, it writes the minimum value of the equipment temperature rise status value into the lower limit and the maximum value into the upper limit, thus forming the welding operation reference status range.

[0058] The edge computing unit forms the baseline range width of the corresponding state item based on the upper and lower limits of the baseline range. When the difference between the upper and lower limits of the corresponding state item's baseline range is not less than the preset minimum baseline width value for the corresponding state item, this difference is determined as the baseline range width of the state item; when the difference is less than the preset minimum baseline width value, the preset minimum baseline width value is determined as the baseline range width of the state item. The preset minimum baseline width value is a positive value, used to ensure that the denominator of the calculation of the state item boundary deviation remains valid.

[0059] The edge computing unit compares each equipment operating status item in the equipment operating status record with the corresponding status item's reference range in the welding operation reference status range. When the status value of an equipment operating status item is higher than the upper limit of the corresponding status item's reference range, the edge computing unit calculates the difference between the status value and the upper limit of the corresponding status item's reference range, and writes the ratio of this difference to the width of the corresponding status item's reference range into the status item boundary deviation value. When the status value of an equipment operating status item is lower than the lower limit of the corresponding status item's reference range, the edge computing unit calculates the difference between the lower limit of the corresponding status item's reference range and the status value, and writes the ratio of this difference to the width of the corresponding status item's reference range into the status item boundary deviation value. When the status value of an equipment operating status item is within the corresponding status item's reference range, the status item boundary deviation value is written as zero. The resulting status item boundary deviation value is a dimensionless value, allowing the welding current status item, welding voltage status item, wire feed speed status item, and equipment temperature rise status item to be accumulated within the same detection cycle.

[0060] After each equipment operating status item generates a status item boundary deviation value within the same inspection cycle, the edge calculation unit accumulates the boundary deviation values ​​for welding current, welding voltage, wire feed speed, and equipment temperature rise status items to form the equipment cycle deviation value corresponding to that inspection cycle. The equipment cycle deviation value is then correlated with the inspection cycle number, welding equipment, and third candidate welding operation record, ensuring that the same third candidate welding operation record has equipment cycle deviation values ​​arranged by inspection cycle number within the operation inspection cycle range.

[0061] The edge computing unit reads the number of inspection cycles within the inspection cycle range corresponding to the third candidate welding operation record. If the number of inspection cycles within the inspection cycle range is less than two, no equipment deviation record is generated for the third candidate welding operation record, and the fault risk equipment number is not determined based on the third candidate welding operation record. If the number of inspection cycles within the inspection cycle range is not less than two, the edge computing unit reads the equipment cycle deviation values ​​corresponding to adjacent inspection cycles in consecutively arranged inspection cycle numbers, and compares the equipment cycle deviation value corresponding to the later inspection cycle with the equipment cycle deviation value corresponding to the previous inspection cycle in increments. If the equipment cycle deviation value corresponding to the later inspection cycle is greater than the equipment cycle deviation value corresponding to the previous inspection cycle, the corresponding adjacent inspection cycle is written into the deviation increment cycle pair; if the equipment cycle deviation value corresponding to the later inspection cycle is not greater than the equipment cycle deviation value corresponding to the previous inspection cycle, the corresponding adjacent inspection cycle is not written into the deviation increment cycle pair.

[0062] When adjacent inspection cycles arranged consecutively according to the inspection cycle number within the corresponding inspection cycle range of the third candidate welding operation record are all written with deviation incrementing cycle pairs, the edge calculation unit forms an equipment operation deviation record based on the corresponding equipment cycle deviation value. The equipment operation deviation record includes the third candidate welding operation record, welding equipment, inspection cycle range, equipment cycle deviation value corresponding to each inspection cycle, and deviation incrementing cycle pairs. The equipment operation deviation record is formed by the cumulative result of the boundary deviation values ​​of multiple equipment operation status items within the same third candidate welding operation record and the continuous inspection cycle incrementing result.

[0063] After a device operation deviation record is generated, the edge computing unit reads the third candidate welding operation record corresponding to that record and then reads the welding equipment corresponding to that record. If the welding equipment corresponding to the third candidate welding operation record matches the welding equipment corresponding to the device operation deviation record, the edge computing unit identifies the welding equipment number corresponding to that equipment as the fault risk device number. Thus, the welding operation records generated in step one, the source detection cycle range generated in step two, the set of second candidate welding operation records generated in step three, the set of third candidate welding operation records generated in step four, and the device operation deviation record generated in step five form a continuous sequence, ultimately determining the fault risk device number.

[0064] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. An edge-computing-based welding equipment failure early warning method, characterized in that, include: Step 1: The edge computing unit reads the operation-related data of each welding equipment according to the detection cycle, forms the operation detection cycle range based on the changes in operation execution status, and generates welding operation records for associating the operation detection cycle range, welding equipment, welded parts, ventilation path, and welding operation position. Step 2: Read the air composition detection values ​​and ventilation migration parameters for each ventilation path. When the air composition detection value exceeds the preset allowable range and the peak value extends along the ventilation migration direction, a valid record of air composition migration and an abnormal air composition detection cycle number are generated. The number of ventilation arrival cycles is generated based on the ventilation migration parameters. The source detection cycle range is generated by backtracking based on the abnormal air composition detection cycle number and the number of ventilation arrival cycles. The ventilation migration parameters include the ventilation migration direction, air component sampling location number, migration distance between adjacent air component sampling locations, ventilation velocity value, detection cycle duration, and ventilation effectiveness status. The air component sampling location number is written in the order of the air component sampling locations along the ventilation migration direction within the same ventilation path. The ventilation velocity value is used to record the speed of air along the ventilation migration direction within the corresponding detection cycle. The ventilation effectiveness status is effective when the ventilation migration direction is consistent with the preset ventilation migration direction of the ventilation path within the corresponding detection cycle, and the ventilation velocity value is not less than the preset lower limit of ventilation velocity. The detection period segment where the air composition detection value exceeds the preset allowable range of air composition is identified as a candidate peak segment, and the starting detection period, peak detection period and fallback detection period of the candidate peak segment are written respectively. Read the air composition sampling location, initial detection period, fallback detection period, and number of ventilation arrival periods corresponding to each candidate peak segment in the deferred chain; For each candidate peak segment, ventilation migration parameters are read cycle by cycle in descending order of detection cycle number, starting from the initial detection cycle, and the number of backtracking detection cycles is accumulated. When the accumulated number of backtracking detection cycles reaches the corresponding number of ventilation arrival cycles, the current detection cycle is determined as the source starting detection cycle of the candidate peak segment. For the same candidate peak segment, starting from the fallback detection cycle, the ventilation migration parameters are read cycle by cycle in descending order of the detection cycle number, and the number of fallback detection cycles is accumulated. When the accumulated number of fallback detection cycles reaches the corresponding number of ventilation arrival cycles, the current detection cycle is determined as the source of the candidate peak segment and the detection cycle is terminated. Read the source start detection period and source end detection period corresponding to each candidate peak segment in the deferred chain, determine the source start detection period with the latest detection period number as the start boundary of the source detection period range, and determine the source end detection period with the earliest detection period number as the end boundary of the source detection period range. When the starting boundary of the source detection cycle range is no later than the ending boundary of the source detection cycle range, the source detection cycle range is formed based on the starting boundary and the ending boundary of the source detection cycle range. Step 3: Read the records from the welding operation records that correspond to the valid records of air component migration and whose operation detection cycle range corresponds to the source detection cycle range, to form the first candidate welding operation record set, and form the second candidate welding operation record set based on the number of ventilation arrival cycles and the cycle overlap relationship; Step 4: Read the response data of the workpiece corresponding to the second candidate welding operation record to form a thermal response position sequence and a vibration response position sequence, and form a third candidate welding operation record set based on the position correspondence between the thermal response position sequence, the vibration response position sequence and the welding operation record. Step 5: Read the equipment working status record of the welding equipment corresponding to the third candidate welding operation record, compare it with the welding operation benchmark status range to form the equipment working deviation record, and determine the fault risk equipment number based on the third candidate welding operation record and the equipment working deviation record.

2. The fault early warning method for welding equipment based on edge computing according to claim 1, characterized in that, In step one, when generating welding operation records, the operation inspection cycle is segmented based on the changes in ventilation paths within the operation inspection cycle, so that the operation inspection cycle range in each welding operation record corresponds to the same ventilation path, including: Read the ventilation path corresponding to each inspection cycle within the operation inspection cycle range of the same welding equipment; The current work segment cycle range is opened with the first detection cycle as the starting detection cycle, and the ventilation path corresponding to the first detection cycle is written. Compare the ventilation path corresponding to the next detection cycle with the ventilation path written in the current operation segment cycle range; if the comparison result is consistent, write the next detection cycle into the current operation segment cycle range. If the consistent comparison result is not valid, the previous inspection cycle of the next inspection cycle is determined as the termination inspection cycle of the current operation segment cycle range. Welding operation record is formed based on the current operation segment cycle range, and a new operation segment cycle range is started with the next inspection cycle as the starting inspection cycle. The ventilation path corresponding to the next inspection cycle is written. When creating a welding operation record, if the ventilation path associated with the current operation segment cycle range is consistent with the ventilation path associated with the operation segment cycle range that has already formed a termination inspection cycle, and there are operation segment cycle ranges associated with different ventilation paths between the current operation segment cycle range and the operation segment cycle range that has already formed a termination inspection cycle, welding operation records are created separately, and the current operation segment cycle range and the operation segment cycle range that has already formed a termination inspection cycle are not merged.

3. The fault early warning method for welding equipment based on edge computing according to claim 1, characterized in that, The method for determining the peak value's extension along the ventilation migration direction in step two includes: Read the air composition detection values ​​at each air composition sampling location within the same ventilation path according to the ventilation migration direction; Read the candidate peak fragment from the first air component sampling position upstream of the ventilation migration direction, and write the earliest candidate peak fragment of the starting detection cycle into the deferred chain; read the candidate peak fragments from the next air component sampling position in the ventilation migration direction in sequence, and take the last candidate peak fragment in the deferred chain as the previous candidate peak fragment. When there is a candidate peak segment in the subsequent air composition sampling position whose initial detection period, peak detection period, and fallback detection period are all later than the detection period corresponding to the previous candidate peak segment, the candidate peak segment that satisfies the detection period delay relationship and has the earliest initial detection period will be written into the delay chain. When the delay chain covers the sampling locations of each air component arranged along the ventilation migration direction within the same ventilation path, the peak value is determined to be delayed along the ventilation migration direction.

4. The welding equipment fault early warning method based on edge computing according to claim 2, characterized in that, Step three, which involves forming a second set of candidate welding operation records based on the overlap between the number of ventilation arrival cycles and the cycles, includes: Read the welding operation record corresponding to the valid record of the same air composition migration from the first candidate welding operation record set; Welding operation records with identical welding equipment, identical welded parts, and identical ventilation paths are written into the same periodic through-check set. Read the start and end boundaries of the source detection cycle range, and read the operation detection cycle range of each welding operation record in the cycle-through verification set; The cycle of each welding operation record in the cycle verification set is arranged in chronological order according to the inspection cycle number; When the starting boundary of the source detection cycle range falls within any operation detection cycle range in the cycle penetration verification set, the ending boundary of the source detection cycle range falls within any operation detection cycle range in the cycle penetration verification set, and there is no detection cycle within the source detection cycle range that is not covered by the operation detection cycle range in the cycle penetration verification set, the welding operation record in which the operation detection cycle range in the cycle penetration verification set and the source detection cycle range have a periodic overlap relationship is written into the second candidate welding operation record set.

5. The fault early warning method for welding equipment based on edge computing according to claim 1, characterized in that, Step four, which involves forming the thermal response location sequence and the vibration response location sequence, includes: Read the thermal response and vibration response values ​​of the welded parts at each response acquisition location in the second candidate welding operation record according to the detection cycle; Within the same testing cycle, the dominant position of thermal response is determined based on the ranking of thermal response values, and the dominant position of vibration response is determined based on the ranking of vibration response values. Read the preset acquisition location adjacency list and write the dominant thermal response location and dominant vibration response location corresponding to the first detection cycle within the operation detection cycle range into the thermal response location sequence and vibration response location sequence, respectively. According to the order of the detection cycle numbers, the dominant positions of thermal response and vibration response in the later detection cycle are compared with the dominant positions of thermal response and vibration response in the previous detection cycle for consistency and adjacency. When the consistent comparison result of the dominant position of thermal response is valid, the dominant position of thermal response corresponding to the next detection cycle is written into the thermal response position sequence. If the consistent comparison result of the dominant thermal response position is not valid but the adjacent comparison result is valid, the dominant thermal response position corresponding to the next detection cycle is written into the thermal response position sequence. When the consistent comparison result of the dominant vibration response position is valid, the dominant vibration response position corresponding to the next detection cycle is written into the vibration response position sequence. If the consistent comparison result of the dominant vibration response position is not valid but the adjacent comparison result is valid, the dominant vibration response position corresponding to the next detection cycle is written into the vibration response position sequence.

6. The welding equipment fault early warning method based on edge computing according to claim 5, characterized in that, Step four, which involves forming the third candidate welding job record set, includes: Read the welding operation position corresponding to each detection cycle within the operation detection cycle range of the second candidate welding operation record; The current work position holding section is started with the first inspection cycle within the work inspection cycle range, and the welding work position corresponding to the first inspection cycle is written into the current work position holding section. The welding positions corresponding to adjacent inspection cycles are compared consistently. If the comparison result is successful, the next inspection cycle is written into the current operation position holding segment. If the comparison result is unsuccessful, the previous inspection cycle of the next inspection cycle is determined as the termination inspection cycle of the current operation position holding segment, and a new operation position holding segment is started in the next inspection cycle. The welding position corresponding to the next inspection cycle is written into the new operation position holding segment. For each work position holding section, the distance value from the welding work position to each response acquisition position is calculated according to the detection cycle, and the work position distance is sorted in ascending order of distance value for the corresponding detection cycle. Read the thermal response dominance position and vibration response dominance position in the working position distance sorting results within the same detection period from the thermal response location sequence and vibration response location sequence; When neither the thermal response distance sequence nor the vibration response distance sequence within the work position holding section advances to the sequence with increasing distance values, the position correspondence result of the work position holding section is formed; When all operation positions in the same second candidate welding operation record form a position correspondence in the same segment, the second candidate welding operation record is written into the third candidate welding operation record set.

7. The welding equipment fault early warning method based on edge computing according to claim 1, characterized in that, Step five, which involves creating a record of equipment operational deviations, includes: Read the equipment operating status record of the welding equipment corresponding to the third candidate welding operation record within the operation detection cycle range; Compare each equipment working status item in the equipment working status record with the corresponding status item reference range in the welding operation reference status range; When the status value of a device's operating status item is higher than the upper limit of the corresponding status item's reference range, the status item boundary deviation value is formed based on the difference between the status value and the upper limit of the corresponding status item's reference range. When the status value of a device's operating status item is lower than the lower limit of the corresponding status item's reference range, the status item boundary deviation value is formed based on the difference between the lower limit of the corresponding status item's reference range and the status value. When the status value of a device's operating status item is within the corresponding status item's reference range, the status item's boundary deviation value is written as zero. The cumulative boundary deviations of each status item within the same detection cycle are used to form the equipment cycle deviation value. When the equipment cycle deviation values ​​corresponding to consecutively arranged testing cycles according to the testing cycle number all increase, an equipment operation deviation record is formed based on the corresponding equipment cycle deviation value.

Images