CN110533278B - SMT production line detection threshold setting method based on particle swarm optimization algorithm - Google Patents
SMT production line detection threshold setting method based on particle swarm optimization algorithm Download PDFInfo
- Publication number
- CN110533278B CN110533278B CN201910635333.4A CN201910635333A CN110533278B CN 110533278 B CN110533278 B CN 110533278B CN 201910635333 A CN201910635333 A CN 201910635333A CN 110533278 B CN110533278 B CN 110533278B
- Authority
- CN
- China
- Prior art keywords
- data
- value
- data table
- package
- threshold
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 239000002245 particle Substances 0.000 title claims abstract description 65
- 238000001514 detection method Methods 0.000 title claims abstract description 45
- 238000004519 manufacturing process Methods 0.000 title claims abstract description 42
- 238000005457 optimization Methods 0.000 title claims abstract description 22
- 238000000034 method Methods 0.000 title claims abstract description 21
- 229910000679 solder Inorganic materials 0.000 claims abstract description 37
- 239000011159 matrix material Substances 0.000 claims description 35
- 238000004806 packaging method and process Methods 0.000 claims description 14
- 238000003466 welding Methods 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 12
- 230000007547 defect Effects 0.000 claims description 9
- 230000003287 optical effect Effects 0.000 claims description 7
- 238000005516 engineering process Methods 0.000 claims description 4
- 230000001174 ascending effect Effects 0.000 claims description 3
- 238000007689 inspection Methods 0.000 claims description 3
- 230000002093 peripheral effect Effects 0.000 claims description 2
- 238000012797 qualification Methods 0.000 abstract description 11
- ATJFFYVFTNAWJD-UHFFFAOYSA-N Tin Chemical compound [Sn] ATJFFYVFTNAWJD-UHFFFAOYSA-N 0.000 description 2
- 238000005476 soldering Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 229910000831 Steel Inorganic materials 0.000 description 1
- BNPSSFBOAGDEEL-UHFFFAOYSA-N albuterol sulfate Chemical compound OS(O)(=O)=O.CC(C)(C)NCC(O)C1=CC=C(O)C(CO)=C1.CC(C)(C)NCC(O)C1=CC=C(O)C(CO)=C1 BNPSSFBOAGDEEL-UHFFFAOYSA-N 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000002068 genetic effect Effects 0.000 description 1
- YOBAEOGBNPPUQV-UHFFFAOYSA-N iron;trihydrate Chemical compound O.O.O.[Fe].[Fe] YOBAEOGBNPPUQV-UHFFFAOYSA-N 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 239000010959 steel Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/22—Indexing; Data structures therefor; Storage structures
- G06F16/2282—Tablespace storage structures; Management thereof
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06395—Quality analysis or management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/04—Manufacturing
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Entrepreneurship & Innovation (AREA)
- Marketing (AREA)
- Educational Administration (AREA)
- Data Mining & Analysis (AREA)
- Tourism & Hospitality (AREA)
- Software Systems (AREA)
- Primary Health Care (AREA)
- Biophysics (AREA)
- Databases & Information Systems (AREA)
- Game Theory and Decision Science (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Manufacturing & Machinery (AREA)
- Computational Linguistics (AREA)
- Evolutionary Computation (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Electric Connection Of Electric Components To Printed Circuits (AREA)
- General Factory Administration (AREA)
Abstract
The invention provides a particle swarm optimization algorithm-based SMT production line detection threshold setting method, which is used for solving the technical problem of low product percent of pass caused by unreasonable detection threshold setting of an SMT production line in the prior art and comprises the following implementation steps: and acquiring a correlation data table, then acquiring a threshold setting data table, calculating a set threshold by adopting a particle swarm optimization algorithm, and finally acquiring an optimal threshold. According to the invention, the production data of the SMT production line is utilized, the detection qualification rate of the SPI solder paste detector on the SMT production line is used for calculating the individual fitness in the particle swarm optimization algorithm, and after iteration of the particle swarm optimization algorithm, the detection qualification rate of the SPI solder paste detector can be improved by the multiple groups of detection thresholds, so that the product qualification rate of the SMT production line is improved, and the risk that all the multiple groups of detection thresholds do not meet the production process requirements is reduced.
Description
Technical Field
The invention belongs to the technical field of industrial big data, relates to a method for setting a detection threshold of an SMT (surface mount technology) production line, and particularly relates to a method for setting a detection threshold of the SMT production line based on a particle swarm optimization algorithm.
Background
The product qualification rate of the SMT production line is comprehensively influenced by all links. Solder paste printing is the first process of the SMT production line process flow, the solder paste printing effect of the process is mainly influenced by machine parameters such as scraper pressure, scraper speed, workbench separation speed, solder paste viscosity, steel mesh cleaning and the like, and the setting of the machine parameters mainly depends on the experience of workers of SMT technologists. The SPI solder paste detector can be through the volume, area, height and the shape that detect the printing solder paste whether be located the detection threshold value within range of setting for to judge whether quality of printing solder paste is up to standard. Therefore, a set of excellent detection threshold values are set for the detection threshold values of the SMT production line, the standard reaching rate of the quality of the printed solder paste can be improved, and the qualification rate of products of the SMT production line is improved.
The method for setting the detection threshold of the SMT production line is generally set by an artist according to experience, so that the detection threshold of the SMT production line depends on manual experience, the quality of the printed solder paste is unstable, the range of the upper limit and the lower limit of the detection threshold of the SMT production line set by the artist is large, the set detection threshold is unreasonable, products with quality problems are difficult to intercept, and the product yield of the SMT production line is low.
The patent document "solder paste detection threshold optimization method based on SMT big data" (patent application No. CN201810650120.4, patent publication No. CN108960306A) applied by the university of sienna electronics technology discloses a solder paste detection threshold optimization method based on SMT big data. The method mainly comprises the following steps: and constructing a threshold optimization data packet, estimating the detection result state of the SPI solder paste detector, obtaining the false rate and the missing rate of a certain package according to a Bayesian decision minimum error rate criterion, finally solving the minimum value of the sum of the missing rate and the false rate through a genetic algorithm, and taking the corresponding solution as the upper limit and the lower limit of the SMT production line detection threshold. The method has the following defects: the detection threshold value of the SMT production line with one packaging type can only be optimized, and the product qualification rate of the SMT production line is influenced by multiple packaging types, so that the product qualification rate of the SMT production line cannot be obviously improved. Because the method can only recommend one group of detection threshold values, different detection threshold values are set for different products of the SMT production line according to the production process requirements, and the group of detection threshold values recommended by the method possibly cannot meet the production process requirements.
Disclosure of Invention
The invention aims to provide a particle swarm optimization algorithm-based SMT production line detection threshold setting method aiming at the defects in the prior art, and the method is used for solving the technical problem that the product percent of pass is low due to the fact that the SMT production line detection threshold is set unreasonably in the prior art.
In order to achieve the purpose, the technical scheme adopted by the invention comprises the following steps:
(1) acquiring a relevant data table:
acquiring a solder paste amount data table for recording a detection result of an SPI (serial peripheral interface) solder paste detector, a solder paste welding result data table for recording a detection result of an AOI (automated optical inspection) automatic optical detector and a manual repair defect type data table in the production process of an SMT (surface mount technology) production line, and splicing data with the same common field PCB (printed circuit board) number and pad number in the three data tables into one line to obtain an associated data table;
(2) acquiring a threshold setting data table:
(2a) deleting data with a column value of bridging, poor appearance and sharpening corresponding to a field Name tettresult in the associated data table, and deleting data with a column value of pseudo welding corresponding to a field Name Reason _ Name, data with a column value of polar corresponding to a field Name Alarm and data with a column value not equal to BGA, QFN, SOP, SOT or QFP corresponding to a field Name Classication to obtain an experience data table;
(2b) calculating the average value u and the variance sigma of data in a column corresponding to the field name vol in the empirical data table, and deleting the data with the data value larger than u +3 sigma or smaller than u-3 sigma in the column to obtain a threshold setting data table;
(3) calculating a set threshold based on a particle swarm optimization algorithm:
(3a) initializing parameters of a particle swarm optimization algorithm, including an inertia factor w epsilon (0,1) and a first learning factor c1E (1,3) and a second learning factor c2E (1,3), the dimension n _ dim of the particle variable is 2 × p, the total number of particles pop _ size e (50, ∞), the search space maximum value _ max e (180, ∞), the search space minimum value _ min e(45,70), the iteration number t, the maximum iteration number iter, the qualification rate, the error rate, the current fitness list Q, the historical fitness list W and the result list R are set to be 1, and p represents the category number of the data value of the column corresponding to the field name Classification in the threshold setting data table;
(3b) randomly generating a position matrix X, a velocity matrix V and a recording matrix G which are all pop _ size × n _ dim in size, wherein, the ith row X [ i ] in X represents the position of the ith particle, the ith row V [ i ] in V represents the speed of the ith particle, the ith row G [ i ] in G represents the historical optimal position of the ith particle, the columns of X, V and G represent the upper threshold limit of BGA package, the lower threshold limit of BGA package, the upper threshold limit of QFN package, the lower threshold limit of QFN package, the upper threshold limit of SOP package, the lower limit of SOP package, the upper threshold limit of SOT package, the lower threshold limit of SOT package, the upper threshold limit of QFP package and the lower threshold limit of QFP package respectively from left to right, the value of each element in X is greater than value _ min and less than value _ max, the value of each element in V is greater than the inverse of value _ min and less than value _ min, and G is equal to X;
(3c) dividing data of a column corresponding to a field name Classication in a threshold setting data table according to values to obtain a BGA data table, a QFN data table, an SOP data table, an SOT data table and a QFP data table, judging whether a data value corresponding to the field name vol in each piece of data in the five data tables is located between an upper packaging threshold value limit and a lower packaging threshold value limit corresponding to an ith particle X [ i ] in X, if so, marking the piece of data as 1, otherwise, marking the piece of data as 0, and simultaneously judging whether a data value corresponding to the field name vol in each piece of data in the five data tables is located between an upper packaging threshold value limit and a lower packaging threshold value limit corresponding to the ith particle G [ i ] in G, if so, marking the piece of data as 1, otherwise, marking the piece of data as 0;
(3d) calculating the individual fitness adapt _ value of each particle in X, filling the calculation result into a current fitness list Q to obtain an assigned current fitness list Q ', putting the particles with the adapt _ value smaller than an error _ rate into a result list R to obtain a result list R ' with the amplitude, simultaneously calculating the individual fitness adapt _ value of each particle in G, filling the calculation result into a current fitness list W to obtain an assigned current fitness list W ', wherein the calculation formula of the adapt _ value is as follows:
wherein n _ count represents the total number of data marked as 1 in all of BGA package data, QFN package data, SOP package data, SOT package data, and QFP package data, and n represents the total number of data in the threshold setting data table;
(3e) when the current individual fitness Q 'i of the ith particle in Q' is smaller than the historical individual fitness W 'i of the ith particle corresponding to W', replacing W 'i with Q' i, and replacing G [ i with X [ i ], thereby obtaining a historical fitness list W 'after replacement and a recording matrix G';
(3f) updating each row of the velocity matrix V according to the following formula to obtain an updated velocity matrix V':
V'[i]＝w×V[i]+c1×random()×(G'[i]-X[i])+c2×random()×(gbest-X[i])
wherein random () represents to generate a random number between 0 and 1, gbset represents a variable, gbest ═ G' [ a ], and a represents the position corresponding to the minimum value in the history fitness list W ";
(3g) updating each row of the position matrix X according to the following formula to obtain an updated position matrix X ', and replacing the value smaller than value _ min and the value larger than value _ max in the position matrix X ' with value _ min to obtain a replaced position matrix X ':
X'[i]＝X[i]+V'[i]；
(3h) making X ═ X ", V ═ V ', G ═ G', clear values in Q and W, and determine whether t is equal to iter, if yes, perform step (3i), otherwise, make t ═ t +1, and perform step (3 c);
(3i) judging whether particles exist in the result list R ', splicing the particles in the result list R ' according to rows if the particles exist in the result list R ', obtaining a threshold setting table, otherwise, making iter be 2 × iter, and executing the step (3 c);
(4) obtaining an optimal threshold value:
and calculating the difference between the sum of the upper threshold limits of all the packages and the sum of the lower threshold limits of all the packages in each row in the threshold setting table, and performing ascending arrangement on the calculation results to obtain an optimal threshold table.
Compared with the prior art, the invention has the following advantages:
1) according to the invention, the particle swarm optimization algorithm is applied, and the detection qualification rate of the SPI solder paste detector on the SMT production line is used for calculating the individual fitness in the particle swarm optimization algorithm, so that the obtained detection threshold can improve the detection qualification rate of the SPI solder paste detector, and the product qualification rate of the SMT production line is improved.
2) According to the invention, a particle swarm optimization algorithm is used, the algorithm can obtain multiple groups of detection threshold values of the SMT production line, and the risk that all the obtained detection threshold values of the SMT production line do not meet the production process requirements is reduced.
Drawings
FIG. 1 is a flow chart of an implementation of the present invention.
Detailed Description
The invention is described in further detail below with reference to the figures and specific examples.
Referring to fig. 1, the present invention includes the steps of:
step 1) obtaining an associated data table:
obtain the solder paste volume data table of record SPI solder paste detector testing result in SMT produces line production process, the field name that includes in this table has: PCB serial number, pad serial number, Component _ id, vol, testresult and Classication, and a solder paste welding result data table for recording the detection result of the AOI automatic optical detector, wherein the table comprises the following field names: the PCB number, the pad number, the Component _ id and the Alarm, and a manual repair defect type data table is recorded, wherein the field names in the table are as follows: and the PCB number, the pad number, the Component _ id and the Reason _ Name are obtained, and the data with the same common field PCB number and pad number in the three data tables are spliced into a row to obtain an associated data table.
Referring to table 1, table 1 shows a table of solder paste amount data of the SPI solder paste detector test results.
TABLE 1 solder paste amount data sheet of SPI solder paste detector detection result
| PCB numbering | Pad numbering | Component_id | vol | testresult | Classification |
| 2721 | 1 | D9 | 34.1 | 0 | QFP |
| 2721 | 2 | D9 | 140.5 | 0 | QFP |
| 2721 | 3 | D9 | 138.1 | 0 | QFP |
| 2721 | 4 | D9 | 137.5 | 0 | QFP |
| 3102 | 1 | D1 | 114.7 | 0 | BGA |
| 3102 | 2 | D1 | 110.7 | 0 | BGA |
| 3102 | 3 | D1 | 110.1 | 0 | BGA |
| 3102 | 4 | D1 | 112.1 | 0 | BGA |
| … | … | … | … | … | … |
Referring to table 2, table 2 shows a table of solder paste soldering results data of the detection results of the AOI automatic optical inspection apparatus.
Table 2 solder paste welding result data table of detection result of AOI automatic optical detector
Referring to table 3, table 3 gives a manual rework defect type data table.
TABLE 3 Manual repair Defect type datasheet
| PCB numbering | Pad numbering | Component_id | Reason_Name |
| 3521 | 1956 | D24 | False welding |
| 3981 | 5 | D53 | False welding |
| 4483 | 3784 | D1 | Tin connection |
| 4392 | 244 | R74 | Tin connection |
Referring to table 4, table 4 shows an associated data table, where the value corresponding to the field Name of Reason _ Name in the table has a blank part, and the blank part indicates that the corresponding PCB has not undergone the maintenance process.
Table 4 data of completed association
Step 2) obtaining a threshold setting data table:
step 2a) deleting data of bridging, poor appearance and pulling tip of a column corresponding to a field Name of teslesult in the associated data table, wherein when a detection result is normal, the value of teslesult is 0, when bridging is detected, the value of teslesult is 4, when poor appearance is detected, the value of teslesult is 8, when pulling tip is detected, the value of teslesult is 16, when volume of solder paste is small, the value of teslesult is 32, when volume of solder paste is large, the value of teslesult is 64, because three defects of bridging, poor appearance and pulling tip in data corresponding to teslesult are not caused by volume of solder paste printing, data corresponding to the defects are required to be deleted, and simultaneously, the value of a column corresponding to a field Name of Reason _ Name is data of false soldering, because the data corresponding to Reason _ Name has welding data, the welding data is not caused by volume of solder paste, and the welding data of the welding is not caused by the welding paste, therefore, to delete the data, the data of which the column value corresponding to the field name of Alarm is policy is deleted, because policy in the data corresponding to Alarm indicates that a problem occurs in the Polarity of the chip to be attached, the problem is caused by a mounting process and is not related to the volume amount of solder paste printing, and therefore the data needs to be deleted, and the data of which the column value corresponding to the field name of Classification is not equal to BGA, QFN, SOP, SOT or QFP is deleted to obtain an experience data table;
and 2b) calculating the average value u and the variance sigma of the data in the column corresponding to the field name vol in the empirical data table, and deleting the data with the data value larger than u +3 sigma or smaller than u-3 sigma in the column to obtain a threshold setting data table.
Referring to table 5, table 5 shows some of the data that needs to be deleted.
Table 5 part of data to be deleted
| PCB numbering | Pad numbering | Component_id | vol | testresult |
| 5252 | 208 | D29 | 34.1 | 32 |
| 5668 | 2379 | D1A | 10.21 | 32 |
| 5677 | 2343 | D1A | 336.3 | 64 |
| 5692 | 2343 | D1A | 325.26 | 64 |
| … | … | … | … | … |
Referring to table 5, the vol value of the first data is 34.1, the mean u and variance of the data corresponding to the field name vol are calculated as σ according to the formula, and the values of u +3 σ and u-3 σ are calculated:
wherein n represents the number of data in the column corresponding to the field name vol in the empirical data table, and xjIndicating the jth data in the column corresponding to the field name vol in the empirical data table.
u-3σ＝109.41-3×24.52＝35.85
u+3σ＝109.41+3×24.52＝182.97
Since 34.1 is not within the interval [35.85,182.97], the piece of data needs to be deleted.
Step 3) calculating a set threshold based on a particle swarm optimization algorithm:
step 3a) initializing parameters of the particle swarm optimization algorithm, wherein the parameters comprise an inertia factor w belonging to (0,1) and a first learning factor c1E (1,3) and a second learning factor c2E (1,3), the dimension n _ dim of the particle variable is 2 × p, the total number of particles pop _ size e (50, ∞), the search space maximum value _ max e (180, ∞), the search space minimum value _ min e (45,70), the iteration number t, the maximum iteration number iter, the pass rate passrate, the error rate _ rate, the current fitness list Q, the historical fitness list W, and the result list R, and let t be 1, p denote the number of categories of data values in the threshold setting data table whose field name is the column corresponding to the Classification, the recommended value of the inertia factor W is between 0.5 and 1.0, let W be 0.6, and the first learning factor c is 0.51And a second learning factor c2Has a recommended value of 2, let c1＝2，c2The total number of particles, pop _ size, is 100, the pass rate is the detection pass rate of the SPI solder paste detector on the SMT production line, and the error rate indicates an error value that can be satisfied by the individual fitness of the particles, and the error rate is 0.01.
Step 3b) randomly generating a position matrix X, a speed matrix V and a recording matrix G, wherein the position matrix X, the speed matrix V and the recording matrix G are all pop _ size × n _ dim in size, the ith row X [ i ] in X represents the position of the ith particle, the ith row V [ i ] in V represents the speed of the ith particle, the ith row G [ i ] in G represents the historical best position of the ith particle, columns of X, V and G represent a BGA package upper threshold limit, a BGA package lower threshold limit, a QFN package upper threshold limit, a QFN package lower threshold limit, an SOP package upper threshold limit, an SOP package lower limit, an SOT package upper threshold limit, an SOT package lower threshold limit, an QFP package upper threshold limit and a QFP package lower threshold limit respectively from left to right, and the value of each element in X is greater than value _ min and less than value _ max, the value of each element in V is greater than the opposite number of value _ min and less than value _ min, the position matrix X, the velocity matrix V is as follows:
step 3c) dividing data of a column corresponding to a field name Classification in a threshold setting data table according to values to obtain a BGA data table, a QFN data table, an SOP data table, an SOT data table and a QFP data table, judging whether a data value corresponding to the field name vol in each of the five data tables is located between an upper limit of a packaging threshold value and a lower limit of the packaging threshold value corresponding to an ith particle X [ i ] in X, if so, marking the data as 1, otherwise, marking the data as 0, and simultaneously judging whether a data value corresponding to the field name vol in each of the five data tables is located between an upper limit of the packaging threshold value and a lower limit of the packaging threshold value corresponding to the ith particle G [ i ] in G, if so, marking the data as 1, otherwise, marking the data as 0;
step 3d) calculating the individual fitness adapt _ value of each particle in X, filling the calculation result into a current fitness list Q to obtain an assigned current fitness list Q ', putting the particles with the adapt _ value smaller than the error _ rate into a result list R to obtain a result list R ' after the amplitude value, simultaneously calculating the individual fitness adapt _ value of each particle in G, filling the calculation result into a current fitness list W to obtain an assigned current fitness list W ', wherein the calculation formula of the adapt _ value is as follows:
wherein n _ count represents the total number of data marked as 1 in all of BGA package data, QFN package data, SOP package data, SOT package data, and QFP package data, and n represents the total number of data in the threshold setting data table;
step 3e) when the current individual fitness Q 'i of the ith particle in Q' is smaller than the historical individual fitness W 'i of the ith particle corresponding to W', replacing W 'i by Q' i, replacing G i by X i, and obtaining a replaced historical fitness list W 'and a record matrix G';
step 3f) updating each row of the velocity matrix V according to the following formula to obtain an updated velocity matrix V':
V'[i]＝w×V[i]+c1×random()×(G'[i]-X[i])+c2×random()×(gbest-X[i])
wherein random () represents to generate a random number between 0 and 1, gbset represents a variable, gbest ═ G' [ a ], and a represents the position corresponding to the minimum value in the history fitness list W ";
step 3g) updating each row of the position matrix X according to the following formula to obtain an updated position matrix X ', replacing the value smaller than value _ min in the position matrix X ' with value _ min, and replacing the value larger than value _ max with value _ max to obtain a replaced position matrix X ':
X'[i]＝X[i]+V'[i]；
step 3h) making X ═ X ", making V ═ V ', making G ═ G', emptying values in Q and W, and judging whether t is equal to iter, if so, executing step 3i), otherwise, making t ═ t +1, and executing step (3 c);
step 3i) judging whether particles exist in the result list R ', if so, splicing the particles in the result list R' according to rows to obtain a threshold setting table, otherwise, making iter be 2 × iter, and executing the step (3 c);
step 4), obtaining an optimal threshold:
and calculating the difference between the sum of the upper threshold limits of all the packages and the sum of the lower threshold limits of all the packages in each row in the threshold setting table, and performing ascending arrangement on the calculation results to obtain an optimal threshold table.
Referring to table 6, table 6 gives an optimum threshold value table.
TABLE 6 optimal threshold table
Claims (2)
1. A SMT production line detection threshold setting method based on a particle swarm optimization algorithm is characterized by comprising the following steps:
(1) acquiring a relevant data table:
acquiring a solder paste amount data table for recording a detection result of an SPI (serial peripheral interface) solder paste detector, a solder paste welding result data table for recording a detection result of an AOI (automated optical inspection) automatic optical detector and a manual repair defect type data table in the production process of an SMT (surface mount technology) production line, and splicing data with the same common field PCB (printed circuit board) number and pad number in the three data tables into one line to obtain an associated data table;
(2) acquiring a threshold setting data table:
(2a) deleting data with a column value of bridging, poor appearance and sharpening corresponding to a field Name tettresult in the associated data table, and deleting data with a column value of pseudo welding corresponding to a field Name Reason _ Name, data with a column value of polar corresponding to a field Name Alarm and data with a column value not equal to BGA, QFN, SOP, SOT or QFP corresponding to a field Name Classication to obtain an experience data table;
(2b) calculating the average value u and the variance sigma of data in a column corresponding to the field name vol in the empirical data table, and deleting the data with the data value larger than u +3 sigma or smaller than u-3 sigma in the column to obtain a threshold setting data table;
(3) calculating a set threshold based on a particle swarm optimization algorithm:
(3a) initializing particle swarm optimization algorithm parametersA number comprising an inertia factor w ∈ (0,1), a first learning factor c1E (1,3) and a second learning factor c2E (1,3), the dimension n _ dim of the particle variable is 2 × p, the total number of particles pop _ size e (50, ∞), the search space maximum value _ max e (180, ∞), the search space minimum value _ min e (45,70), the number of iterations t, the maximum number of iterations iter, the pass rate, the error rate _ rate, the current fitness list Q, the historical fitness list W, and the result list R, and let t be 1, p denote the number of categories of data values in the threshold setting data table whose field name is the column corresponding to the Classification; the pass rate is the detection pass rate of an SPI solder paste detector on an SMT production line;
(3b) randomly generating a position matrix X, a velocity matrix V and a recording matrix G which are all pop _ size × n _ dim in size, wherein, the ith row X [ i ] in X represents the position of the ith particle, the ith row V [ i ] in V represents the speed of the ith particle, the ith row G [ i ] in G represents the historical optimal position of the ith particle, the columns of X, V and G represent the upper threshold limit of BGA package, the lower threshold limit of BGA package, the upper threshold limit of QFN package, the lower threshold limit of QFN package, the upper threshold limit of SOP package, the lower threshold limit of SOP package, the upper threshold limit of SOT package, the lower threshold limit of SOT package, the upper threshold limit of QFP package and the lower threshold limit of QFP package respectively from left to right, the value of each element in X is greater than value _ min and less than value _ max, the value of each element in V is greater than the inverse of value _ min and less than value _ min, and G is equal to X;
(3c) dividing data of a column corresponding to a field name Classication in a threshold setting data table according to values to obtain a BGA data table, a QFN data table, an SOP data table, an SOT data table and a QFP data table, judging whether a data value corresponding to the field name vol in each piece of data in the five data tables is located between an upper packaging threshold limit and a lower packaging threshold limit corresponding to an ith particle X [ i ] in X, if so, marking the piece of data as 1, otherwise, marking the piece of data as 0, and simultaneously judging whether a data value corresponding to the field name vol in each piece of data in the five data tables is located between an upper packaging threshold limit and a lower packaging threshold limit corresponding to the ith particle G [ i ] in G, if so, marking the piece of data as 1, otherwise, marking the piece of data as 0;
(3d) calculating the individual fitness adapt _ value of each particle in X, filling the calculation result into a current fitness list Q to obtain an assigned current fitness list Q ', putting the particles with the adapt _ value smaller than an error _ rate into a result list R to obtain an assigned result list R ', simultaneously calculating the individual fitness adapt _ value of each particle in G, filling the calculation result into a historical fitness list W to obtain an assigned historical fitness list W ', wherein the calculation formula of the adapt _ value is as follows:
wherein n _ count represents the total number of data marked as 1 in all of BGA package data, QFN package data, SOP package data, SOT package data, and QFP package data, and n represents the total number of data in the threshold setting data table;
(3e) when the current individual fitness Q 'i of the ith particle in Q' is smaller than the historical individual fitness W 'i of the corresponding ith particle in W', replacing W 'i with Q' i, and replacing G i with X i to obtain a replaced historical fitness list W 'and a record matrix G';
(3f) updating each row of the velocity matrix V according to the following formula to obtain an updated velocity matrix V':
V'[i]＝w×V[i]+c1×random()×(G'[i]-X[i])+c2×random()×(gbest-X[i])
wherein random () represents to generate a random number between 0 and 1, gbset represents a variable, gbest ═ G' [ a ], and a represents the position corresponding to the minimum value in the history fitness list W ";
(3g) updating each row of the position matrix X according to the following formula to obtain an updated position matrix X ', and replacing the value smaller than value _ min and the value larger than value _ max in the position matrix X ' with value _ min to obtain a replaced position matrix X ': x '[ i ] + V' [ i ];
(3h) making X ═ X ", V ═ V ', G ═ G', clear values in Q and W, and determine whether t is equal to iter, if yes, perform step (3i), otherwise, make t ═ t +1, and perform step (3 c);
(3i) judging whether particles exist in the result list R ', splicing the particles in the result list R ' according to rows if the particles exist in the result list R ', obtaining a threshold setting table, otherwise, making iter be 2 × iter, and executing the step (3 c);
(4) obtaining an optimal threshold value:
and calculating the difference between the sum of the upper threshold limits of all the packages and the sum of the lower threshold limits of all the packages in each row in the threshold setting table, and performing ascending arrangement on the calculation results to obtain an optimal threshold table.
2. The particle swarm optimization algorithm-based SMT production line detection threshold setting method according to claim 1, wherein the mean value u and the variance σ of the data in the column corresponding to the field name vol in the empirical data table calculated in step (2b) are respectively calculated according to the following formula:
wherein n represents the number of data in the column corresponding to the field name vol in the empirical data table, and xjIndicating the jth data in the column corresponding to the field name vol in the empirical data table.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910635333.4A CN110533278B (en) | 2019-07-15 | 2019-07-15 | SMT production line detection threshold setting method based on particle swarm optimization algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910635333.4A CN110533278B (en) | 2019-07-15 | 2019-07-15 | SMT production line detection threshold setting method based on particle swarm optimization algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110533278A CN110533278A (en) | 2019-12-03 |
| CN110533278B true CN110533278B (en) | 2022-03-22 |
Family
ID=68660209
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910635333.4A Active CN110533278B (en) | 2019-07-15 | 2019-07-15 | SMT production line detection threshold setting method based on particle swarm optimization algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110533278B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113158401B (en) * | 2020-12-20 | 2022-10-04 | 河海大学常州校区 | Multi-granularity SMT production line energy consumption modeling method |
| CN118111377A (en) * | 2022-11-24 | 2024-05-31 | 中兴通讯股份有限公司 | Solder paste detection threshold adjustment method, device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105590143A (en) * | 2015-12-17 | 2016-05-18 | 兰州理工大学 | Optimization method for balancing loads of multiple chip mounters in assembly line in PCB assembling technology |
| CN108960306A (en) * | 2018-06-22 | 2018-12-07 | 西安电子科技大学 | Tin cream detection threshold value optimization method based on SMT big data |
| CN109214561A (en) * | 2018-08-21 | 2019-01-15 | 上海电力学院 | Consider the distributed generation resource configuration method of active distribution system dynamic path optimization |
| CN109255465A (en) * | 2018-07-19 | 2019-01-22 | 中国矿业大学 | A kind of welding production line work class distribution system and method based on particle swarm algorithm |
| CN109597968A (en) * | 2018-12-29 | 2019-04-09 | 西安电子科技大学 | Paste solder printing Performance Influence Factor analysis method based on SMT big data |
| CN109657718A (en) * | 2018-12-19 | 2019-04-19 | 广东省智能机器人研究院 | SPI defect classification intelligent identification Method on a kind of SMT production line of data-driven |
-
2019
- 2019-07-15 CN CN201910635333.4A patent/CN110533278B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105590143A (en) * | 2015-12-17 | 2016-05-18 | 兰州理工大学 | Optimization method for balancing loads of multiple chip mounters in assembly line in PCB assembling technology |
| CN108960306A (en) * | 2018-06-22 | 2018-12-07 | 西安电子科技大学 | Tin cream detection threshold value optimization method based on SMT big data |
| CN109255465A (en) * | 2018-07-19 | 2019-01-22 | 中国矿业大学 | A kind of welding production line work class distribution system and method based on particle swarm algorithm |
| CN109214561A (en) * | 2018-08-21 | 2019-01-15 | 上海电力学院 | Consider the distributed generation resource configuration method of active distribution system dynamic path optimization |
| CN109657718A (en) * | 2018-12-19 | 2019-04-19 | 广东省智能机器人研究院 | SPI defect classification intelligent identification Method on a kind of SMT production line of data-driven |
| CN109597968A (en) * | 2018-12-29 | 2019-04-09 | 西安电子科技大学 | Paste solder printing Performance Influence Factor analysis method based on SMT big data |
Non-Patent Citations (2)
| Title |
|---|
| Research on real time feature extraction method for complex manufacturing big data;Xianguang Kong et al;《The International Journal of Advanced Manufacturing Technology》;20161214;第99卷;第1101-1108页 * |
| 基于SMT大数据的产品质量预测方法研究与实现;冯尔磊;《中国优秀硕士学位论文全文数据库工程科技I辑》;20190215(第02期);摘要，第1-62页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110533278A (en) | 2019-12-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108960306B (en) | Solder paste detection threshold value optimization method based on SMT big data | |
| CN110533278B (en) | SMT production line detection threshold setting method based on particle swarm optimization algorithm | |
| CN110992317B (en) | PCB defect detection method based on semantic segmentation | |
| CN106079892B (en) | A kind of PCB paste solder printing procedure quality intelligent monitor system and method | |
| CN106934800B (en) | Metal plate strip surface defect detection method and device based on YOLO9000 network | |
| CN106952250A (en) | A kind of metal plate and belt detection method of surface flaw and device based on Faster R CNN networks | |
| CN111080622A (en) | Neural network training method, workpiece surface defect classification and detection method and device | |
| CN114216915B (en) | Method for detecting and classifying class levels based on stains and defect types of circuit board | |
| KR20190075707A (en) | Method for sorting products using deep learning | |
| CN112329860B (en) | Mixed deep learning visual detection method, device, equipment and storage medium | |
| US20080046210A1 (en) | Method, device and program for setting a reference value for substrate inspection | |
| JP5959444B2 (en) | Defect cause estimation (classification) method by printing inspection equipment | |
| CN111311576A (en) | Point cloud information-based defect detection method | |
| CN102522350B (en) | Method and device for detecting failed production machine platforms | |
| CN115082401B (en) | SMT production line chip mounter fault prediction method based on improved YOLOX and PNN | |
| CN113822889A (en) | Method for detecting surface defects of hot-rolled steel plate | |
| CN116337868B (en) | Surface defect detection method and detection system | |
| CN109903711A (en) | Liquid crystal display defect automatic checkout system | |
| CN114489528A (en) | Printing equipment fault monitoring method and system | |
| KR20200006028A (en) | Printed circuit board inspecting apparatus, method for determining fault type of screen printer and computer readable recording medium | |
| CN110856437A (en) | SMT production process control chart pattern recognition method | |
| CN115719326A (en) | PCB defect detection method and device | |
| CN108548472B (en) | Automatic measuring device of six-equal-division window hole ball cage retainer | |
| CN102118920A (en) | Patching and assessing system of circuit boards and method | |
| JP5253745B2 (en) | Product manufacturing method and process management program |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |