KR102400012B1 - Method for diagnosing seam weld quality - Google Patents

Abstract

The present invention provides a method for diagnosing the quality of seam welding for a joint, comprising the steps of: (A) updating data of a seam welding process in a DB; (B) calculating DB information for the entire period or the relevant period in response to the selection of periodic inspection or ad hoc inspection; and (C) processing the calculation result and storing it in the DB.
Accordingly, there is an effect of reducing the number of times required for analysis of seam welding, improving the level of quality analysis, and ultimately increasing the efficiency of quality control by rapidly analyzing the cause of the increase in the defect rate.

Description

{Method for diagnosing seam weld quality}

The present invention relates to a welding process for a ship, and more particularly, to a method for diagnosing the quality of seam welding by collecting, analyzing, and diagnosing test results conducted daily for LNGC production.

Seam welding accounts for approximately 90% of the manufacturing process of LNGC cargo holds, and has a large and direct impact on productivity. In general, the condition of the equipment is checked through testing before the actual production of the LNGC cargo hold begins. This includes the process of test welding, specimen inspection, cross-section analysis, and report writing. At this time, it is essential to understand the test analysis quantity, the number of defects, and the defect rate. Nevertheless, it takes a lot of time for repetitive work due to insufficient utilization of massive production data.

As prior art documents that can be referenced in relation to this, the following Korean Patent Publication No. 0237153 (Prior Document 1), Korean Patent Application Laid-Open No. 2015-0007293 (Prior Document 2), etc. are known.

Prior Document 1 measures welding voltage E, welding current I, electrode pressing force P, welding speed S, and lap quantity L, and calculates unit lap quantity and heat input per unit plate width dQ/dx based on this, and calculates the material and thickness of the welded object. The standard set value and the allowable range for the measured and calculated values according to each of the set values and the allowable range are respectively set and compared.

Prior Document 2 monitors various variables during the welding process and weights the variables accordingly, detects welding defects to diagnose causes, quantifies the overall quality of welding, obtains and uses data indicating good quality welding, Improve the production and quality control of the automated welding process, teach the proper welding technique, check the cost reduction of the welding process, and derive the optimal welding settings to be used as presets.

However, according to the above-mentioned prior literature, since it is intended to increase quality and productivity in the individual welding process, there is room for improvement in terms of process management for monitoring the defect trend of the entire welding project.

Korean Patent Publication No. 0237153 "Welding quality evaluation method" (published date: May 28, 1997) Korean Patent Application Laid-Open No. 2015-0007293 "System and method for monitoring welding quality" (published date: January 20, 2015)

An object of the present invention to improve the problems of the prior art is to maintain a predictable quality level through failure probability analysis, equipment failure diagnosis, etc. based on a database accumulating results generated through periodic tests. An object of the present invention is to provide a method for diagnosing the quality of seam welding for

In order to achieve the above object, in a method for diagnosing the quality of seam welding for a joint using a DB that has accumulated data of a seam welding process: (A) La for the sample group nugget data, selecting at least one of Lt, P, P1, P2, and e and calculating at least one of a mean, a standard deviation, and an estimation defect rate assuming it is a normal distribution; (B) selecting the same value as the value selected from the sample group from the population nugget data stored in the DB, and calculating at least one of the mean, standard deviation, and estimation defect rate in the same manner as in the sample group by assuming a normal distribution; and (C) diagnosing the welding quality by comparing the values calculated in the population and the sample group.

As a detailed configuration of the present invention, the estimated defective rate is characterized in that it is based on the area below the acceptance criterion in the probability density function of the normal distribution or the probability value of the acceptance criterion in the cumulative distribution function of the normal distribution.

As a detailed configuration of the present invention, in the step of diagnosing the welding quality, the sample group actual defect rate / population actual defect rate, the sample group estimated defect rate / the population actual defect rate, the sample group mean / population mean, or the sample group standard deviation / population standard deviation If it exceeds the allowable standard, it is characterized in that it is judged as higher than the welding quality.

As a detailed configuration of the present invention, the population nugget data and the sample group nugget data stored in the DB are classified for each welding equipment or welder.

As a detailed configuration of the present invention, when it is determined that the welding quality is abnormal in the step of diagnosing the welding quality, the population nugget and the sample group nugget are classified by welding equipment or welder, and the seam welding quality diagnosis method is re-executed. .

As described above, according to the present invention, there is an effect of improving the quality analysis level as well as reducing the number of times required for the analysis of seam welding, and ultimately increasing the efficiency of quality control by rapidly analyzing the cause of the increase in the defect rate.

1 is a data exemplifying a dimension item related to the quality of a seam weld;
Figure 2 is a table statistically processed dimensional information for each joint of the seam weld
3 is a table showing the change in the defect rate with the average value and standard deviation of FIG. 2
4 is a flowchart showing the main algorithm of the method according to the present invention;
Fig. 5 is a flowchart showing the macro analysis subroutine of Fig. 4;
Fig. 6 is a flowchart showing the equipment diagnostic subroutine of Fig. 4;
7 is a flowchart showing the skill diagnosis subroutine of the welder of FIG.

Hereinafter, an embodiment of the present invention will be described in detail based on the accompanying drawings.

The present invention proposes a method for diagnosing the quality of seam welding for a joint by using a DB in which data of a seam welding process is accumulated. Although it is intended for the manufacture of LNGC cargo holds in which the proportion of seam welding is absolutely high, it is not necessarily limited thereto. 1A and 1B illustrate a state in which seam welding is performed in a state in which three objects to be welded are overlapped.

Hardware resources for the implementation of the present invention are configured based on a PC for a terminal and a network (intranet, Internet, LAN, etc.) in addition to the DB (server). The server is equipped with a microprocessor, memory, and I/O interface, drives the hardware including DB and communication with the set algorithm, and outputs it on the screen of the PC. The data of the seam welding process is stored in DB, updated, classified and retrieved through PC.

Meanwhile, LNGC welding part daily test data is collected into the DB, and the data collected by the server's operation is normalized.

Step (A) according to the present invention selects at least one value among La, Lt, P, P1, P2, and e for the sample group nugget data, and assuming that it is a normal distribution, the mean, standard deviation, and estimation error rate It proceeds to the process of calculating at least one of

In FIG. 1C, acceptance criteria for La, Lt, e, P1, and P2 are shown for each joint type classified by symbols A to E. The joint thickness of a three-ply seam weld is exemplified as 0.7/0.5/0.7 in A and 1.5/1.5/1.5 in E. The sample population nugget data is obtained through a known macro test. In addition, welding equipment, welder, and defect information are collected in connection with macro dimension data and stored in the DB.

In this case, the macro dimension data includes the welding width (La, Lt) and penetration depth (P, P1, P2) of the nugget, and the penetration thickness (e). Referring to FIG. 1A, which is a schematic representation of the macro test result of FIG. 1B, in the three-ply seam welding of the upper member, the middle member, and the lower member, Lt means the maximum welding width of the entire nugget, whereas La is the upper or lower surface of the intermediate member It means the welding width of the smaller side of the nugget area cut by The height of the nugget with respect to the upper and lower surfaces of the intermediate member means the penetration depth expressed by the symbols P1, P2, or P. As the table in Fig. 1c illustrates the acceptance criteria, in seam welding, the weld width (La, Lt) and the penetration depth (P, P1, P2) are important factors for quality determination. The penetration thickness (e) is also a macro value required by the acceptance criteria and is one of the items included in the quality diagnosis.

2A is a distribution diagram of occurrence probability for La, e, P accumulated through a macro test for a specific joint shape. Since the elements (macro dimensions) accumulated in the DB form a Normal or Gaussian distribution as shown in FIG. 2A, utilization can be expected in the field of quality control of seam welding. If the number of data is sufficient, the actual defective rate and the defective rate estimated through the normal distribution agree very well. As the current acceptance criteria, most of the defects are due to the lack of the P value, which is the penetration depth. FIG. 2B shows that the distribution diagram and estimation defect rate can be calculated using the probability density function (PDF) and the cumulative distribution function (CDF) from the normal distribution of each macro dimension.

Meanwhile, according to the present invention, DB information for the entire period or the corresponding period is calculated in response to selection of a periodic inspection or an occasional inspection. In FIG. 4 , the flow on the left illustrates the progress state of the periodic inspection, and the flow on the right illustrates the progress state of the regular inspection. Periodic inspection is conducted for regular diagnosis of LNGC production quality, and occasional inspection is conducted to prepare for concerns about deterioration of welding quality due to excessive quantity. In any case, it is possible to detect a change in a specific recent period (eg, 1 month, 3 months, quarters, etc.) compared to the entire period (previous period) accumulated in the DB. In addition, regular quality diagnosis by the manager's judgment can be added. Accordingly, quality monitoring can be performed in response to the introduction of new equipment and changes in construction methods.

Step (B) according to the present invention selects the same value as the value selected from the sample group from the population nugget data stored in the DB, and assuming that it is a normal distribution, at least one of the mean, standard deviation, and estimated defective rate as in the sample group. It goes through the process of calculating one. It appears that the estimated defective rate for the entire period calculated by a normal distribution according to the mean and standard deviation of the population calculated from the DB and the actual defective rate for the period are consistent. The comparison of the estimated defective rate calculated by normalizing the population and the sample group is the first method to find the defective rate in the sample group. Although there is a slight difference between the population and the sample group in the normalization operation and additional verification is possible through reliability analysis, it is omitted because it does not fall under the gist of the present invention.

As a detailed configuration of the present invention, the estimated defective rate is characterized in that it is based on the area below the acceptance criterion in the probability density function of the normal distribution or the probability value of the acceptance criterion in the cumulative distribution function of the normal distribution. When the probability density function (PDF) mentioned in FIG. 2(b) is integrated, a cumulative distribution function (CDF) is generated. The probability density function (PDF) represents the probability of being included in a certain section, and the cumulative distribution function (CDF) represents the probabilistic distribution of the random variable X. In the graph on the left of FIG. 2(b), the area below the allowable standard indicated by “fail” can be calculated as the estimated defective rate, and the portion marked with “estimated defective rate” in the right graph of FIG. 2(b) is the acceptance criterion It can be calculated as a probability value of .

3A shows an increase in the defective rate according to a change in the average value. For example, when the average value decreases by 5% (0.365 mm -> 0.347 mm), the estimated probability of failure increases from 0.6% to 1.3%. 3B shows an increase in the defective rate according to a change in standard deviation. For a 10% increase in standard deviation (0.066 -> 0.073), the estimated probability of defectiveness increases from 0.6% to 1.2%. In conclusion, it can be used as a variable for diagnosing the mean and standard deviation by assuming that the weld macro dimensions are normally distributed. Accordingly, it is possible to determine whether the quality is abnormal by detecting changes in the mean and standard deviation.

As a detailed configuration of the present invention, in the step of diagnosing the welding quality, the sample group actual defect rate / population actual defect rate, the sample group estimated defect rate / the population actual defect rate, the sample group mean / population mean, or the sample group standard deviation / population standard deviation If it exceeds the allowable standard, it is characterized in that it is judged as higher than the welding quality. The fact that the mean of the sample group is lowered or the standard deviation is increased can also be used as the second and third methods to find a quality problem.

As a detailed configuration of the present invention, the population nugget data and the sample group nugget data stored in the DB are classified for each welding equipment or welder. Through regular inspection, the calculated value of the judgment condition in the macro analysis subroutine is compared with the reference value to proceed with the equipment diagnosis subroutine and the welder skill diagnosis subroutine. Referring to FIG. 5 , a report on periodic diagnosis is created by calculating the actual defect rate, average, standard deviation, and estimated defect rate, and inspection/repair of specific welding equipment or skill of a specific welder is determined according to predetermined determination conditions to be described later. do.

At this time, the determination conditions include the calculated values of the actual defective rate for the relevant period/the total period's actual defect rate, the corresponding period's estimated defective rate/the total period's estimated defective rate, the average for the period/the average for the entire period, and the standard deviation for the period/the standard deviation for the entire period. do. If the actual defective rate for the period/the total period exceeds the first reference value (R1), or the estimated defective rate for the period/the estimated defective rate for the entire period exceeds the second reference value (R2), or the average for the period/the average for the entire period exceeds the third It is determined whether the standard value (R3) is less than the reference value (R3) or whether the standard deviation/standard deviation of the entire period exceeds the fourth reference value (R4). If any one of the above judgment conditions is yes, it proceeds to the equipment diagnosis subroutine and the welder skill diagnosis subroutine; otherwise, the macro analysis subroutine program ends. Accordingly. It is possible to accurately respond to changes in the welding environment and to determine whether to perform quality inspection.

Step (C) according to the present invention proceeds to the process of diagnosing the welding quality by comparing the values calculated in the population and the sample group. If the number of recent tests in the relevant period/average number of tests in the entire period does not exceed the standard value through frequent inspection, the equipment diagnosis subroutine and the welder skill diagnosis subroutine are carried out. If you are concerned about the deterioration of welding quality due to excessive quantity, request on the PC screen to conduct a self-inspection. If not, proceed with the equipment diagnosis subroutine and welder skill diagnosis subroutine described above. In other words, if the increase in the defect rate is not due to excessive quantity, it predicts a defect in welding equipment or a change in the skill of the welder.

As a detailed configuration of the present invention, when it is determined that the welding quality is abnormal in the step of diagnosing the welding quality, the population nugget and the sample group nugget are classified by welding equipment or welder, and the seam welding quality diagnosis method is re-executed. .

According to the equipment diagnosis subroutine, information for each welding equipment is retrieved from the DB and an equipment inspection message is generated if the judgment condition is met. Referring to FIG. 6 , the DB for each welding equipment is searched to calculate the actual defective rate, average, standard deviation, and estimated defective rate for the entire period and the corresponding period, and determine whether a specific welding equipment is abnormal according to a predetermined determination condition. If any one of the criteria is yes by applying the same determination conditions as in FIG. 5, it is determined that the welding equipment is abnormal and a message requesting inspection is generated.

According to the skill diagnosis subroutine, a warning message to the welder is generated when the information for each welder is retrieved from the DB and the judgment condition is met. Referring to FIG. 7 , the DB is searched for each welder to calculate the actual defect rate, average, standard deviation, and estimated defect rate for the entire period and the period, and determine whether a specific welder's skill is abnormal according to a predetermined determination condition. If any one of the criteria is yes by applying the same determination conditions as in FIG. 5, it is judged as an abnormality in the skill of the corresponding welder and a message to call attention is generated.

After the present invention step (C), the operation result is processed and stored in the DB. Information related to changes in the welding environment, information on repair or replacement of welding equipment, and information on changes in the skill of the welder are updated in the DB. When data accumulation proceeds in this way, it is possible to accurately detect changes in the trends of welding equipment and welders, so that the reliability for quantifying the quality level of seam welding can be improved.

The present invention is not limited to the described embodiments, and it is apparent to those skilled in the art that various modifications and variations can be made without departing from the spirit and scope of the present invention. Accordingly, it should be said that such variations or modifications fall within the scope of the claims of the present invention.

R1~R4: Reference value La, Lt: Weld width
P, P1, P2: penetration depth

Claims (5)

In the method of diagnosing the quality of seam welding for a joint by using a DB that has accumulated data of the seam welding process:
(A) At least one of La, Lt, P, P1, P2, and e is selected for the sample group nugget data, and at least one of the mean, standard deviation, and estimation error rate is calculated assuming a normal distribution. step;
(B) selecting the same value as the value selected from the sample group from the population nugget data stored in the DB, and calculating at least one of the mean, standard deviation, and estimation defect rate in the same manner as in the sample group by assuming a normal distribution; and
(C) diagnosing the welding quality by comparing the values calculated in the population and the sample group;
In the step of diagnosing the welding quality, the sample group actual defect rate / population actual defect rate, the sample group estimated defect rate / the population actual defect rate, the sample group mean / population mean, or the sample group standard deviation / population standard deviation is greater than or equal to the allowable standard. judged by
The population nugget data and the sample group nugget data stored in the DB are seam welding quality diagnosis method, characterized in that classified by welding equipment or welder. The method according to claim 1,
The estimated defective rate is a seam welding quality diagnosis method, characterized in that it is based on the area less than the allowable standard in the probability density function of the normal distribution or the probability value of the acceptance criterion in the cumulative distribution function of the normal distribution. delete delete The method according to claim 1,
When it is determined that the welding quality is abnormal in the step of diagnosing the welding quality, a seam welding quality diagnosis method, characterized in that by classifying the population nuggets and the sample group nuggets by welding equipment or welders, and re-performing the seam welding quality diagnosis method.

Images