Semiconductor package data information analysis processing system
By using a semiconductor packaging data information analysis and processing system, the system can directly analyze and predict the losses and benefits of semiconductor packaging production lines from multiple dimensions. This solves the problem of inaccurate analysis results in existing technologies and maximizes packaging benefits and efficient resource utilization.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- SHENZHEN CHENYUE STORAGE ELECTRONIC TECH CO LTD
- Filing Date
- 2023-01-03
- Publication Date
- 2026-06-19
Smart Images
Figure CN116258326B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of semiconductor packaging technology, and more specifically, to a semiconductor packaging data information analysis and processing system. Background Technology
[0002] With the rapid development of society, integrated circuits are being used more and more widely. They are now used in many fields such as personal computers, mobile phones, and smart home appliances. Semiconductor packaging is a crucial link in integrated circuits. It not only plays a role in protecting the internal chip and enhancing thermal conductivity, but also serves as a bridge connecting the internal world of the integrated circuit with the external circuit system.
[0003] Under the above circumstances, the demand for semiconductor packaging is increasing day by day. In order to meet the current packaging needs, semiconductor packaging has evolved into production line packaging. As we all know, any production line will inevitably suffer losses during its operation, and semiconductor packaging production lines are no exception. Moreover, the degree of loss generated by semiconductor packaging production lines varies depending on the packaging body. In order to maximize the packaging efficiency, there is a need to analyze the degree of loss of packaging production lines under different packaging bodies in order to select suitable packaging bodies in a targeted manner.
[0004] However, current analyses of the loss rate of packaging production lines under different packaging entities use the packaging time of different packaging entities into finished products on the packaging production line as an intermediate conversion. This analysis method is indirect. On the one hand, it cannot directly and intuitively analyze the loss rate of the packaging production line under different packaging entities, making the reliability of the analysis results low. On the other hand, it does not take into account that modern semiconductor packaging production lines are all automated, and the duration of each packaging process is precisely controlled. This means that the packaging time of different packaging entities into finished products on the packaging production line is not significantly different, making it unreasonable to use packaging time as the basis for analyzing packaging loss rate. This further weakens the reliability of the analysis results and makes it difficult to provide a reliable reference for selecting suitable packaging entities. Summary of the Invention
[0005] To address the aforementioned technical problems, this invention provides a semiconductor packaging data information analysis and processing system. Based on packaging data from a semiconductor packaging production line under different packaging entities, it enables direct and intuitive analysis of the loss rate of the packaging production line under different packaging entities. Specifically, this invention is implemented through the following technical solution: A semiconductor packaging data information analysis and processing system, comprising: a semiconductor packaging equipment statistics module, used to count the number of packaging devices owned by the target semiconductor packaging company on its corresponding packaging production line and to obtain the name of each packaging device.
[0006] The feature packaging data extraction module is used to extract packaging parameters from each historical packaging record corresponding to the target semiconductor packaging company and identify feature packaging data.
[0007] The packaging production line uses a loss analysis module to analyze the usage loss of the packaging production line for each packaging entity and corresponding historical packaging records.
[0008] The apparent sustaining package quantity prediction module is used to predict the number of sustaining packages relative to each package body in the current packaging production line in terms of appearance.
[0009] The time-sustainable package quantity prediction module is used to predict the number of packages that the current packaging production line will maintain relative to each package body over time.
[0010] The effective maintenance package quantity determination module is used to determine the effective maintenance package quantity of the current packaging production line relative to each package body.
[0011] The processing information database is used to store the original surface area of each packaging device, the normal allowable wear index of each packaging device on the packaging production line of the target semiconductor packaging company, the degree of influence factor of various apparent defect types, the usage loss degree of the packaging production line corresponding to the unit packaging time, and the pre-startup time of the packaging production line.
[0012] The packaging efficiency evaluation module is used to extract the unit packaging revenue corresponding to each packaging entity from the processing information database, and combine it with the effective maintenance packaging quantity of the current packaging production line relative to each packaging entity to evaluate the packaging efficiency of the current packaging production line relative to each packaging entity.
[0013] The matching packaging body screening and display module is used to select the packaging body with the highest packaging benefit from the packaging benefits of each packaging body in the current packaging production line as the matching packaging body and display it.
[0014] Furthermore, the packaging parameters include wafer type, number of packages, packaging time, and a video of the packaging production line usage process.
[0015] Furthermore, the feature packaging data includes the packaging body, the frequency of use of the packaging production line, and the usage information corresponding to each use of the packaging production line, wherein the usage information is the usage video segment corresponding to each packaging device.
[0016] Furthermore, the specific identification process corresponding to the feature packaging data is as follows: extract the wafer model from the packaging parameters corresponding to each historical packaging record, and then classify the historical packaging records corresponding to the same wafer model. At this time, the same wafer model is recorded as the packaging body, thereby obtaining several historical packaging records corresponding to each packaging body.
[0017] The number of historical packaging records owned by each packaging entity is counted, and this number is used as the usage frequency of the packaging production line corresponding to each packaging entity.
[0018] The video of the packaging production line usage process is extracted from the packaging parameters of each historical packaging record of each packaging entity, and then segmented according to the individual packaging equipment to obtain the usage video segment corresponding to each packaging equipment in each historical packaging record owned by each packaging entity.
[0019] Furthermore, the specific operation steps for analyzing the usage loss of the packaging production line corresponding to each historical packaging record for each packaging entity are as follows: decompose the usage video segments of each packaging device in each historical packaging record for each packaging entity according to video frames to obtain several usage images corresponding to each packaging device, and sort each usage image according to its timestamp in the usage video segment, and then take the first and last usage images, and record them as the initial usage image and the end usage image, respectively.
[0020] The outer contour is extracted from the initial and final usage images corresponding to each packaging device. The final usage contour of each packaging device is then superimposed with the initial usage contour. Based on this, the external usage deformation degree of each packaging device in each historical packaging record is calculated, denoted as σ. i j k, where i represents the number of the packaging entity, i = 1, 2, ..., n, n represents the number of packaging entities, j represents the number of historical packaging records, j = 1, 2, ..., m, m represents the number of historical packaging records owned by each packaging entity, and k represents the number of packaging equipment, k = 1, 2, ..., z, z represents the number of packaging equipment owned by the target semiconductor packaging company on its corresponding packaging production line.
[0021] The end-of-use image of each packaging device is compared with the initial use image to identify abnormal areas. The end-of-use image is then focused on the abnormal areas to extract the appearance defect parameters of each packaging device. The appearance defect parameters include the appearance defect type and appearance defect area.
[0022] Based on the apparent defect type corresponding to each packaging device, the degree of influence factor corresponding to each packaging device is obtained from the processing information database.
[0023] Through formula The apparent usage defect η of each packaging device in each historical packaging record corresponding to each packaging entity was statistically obtained. i j k, where ε i j k, s i jk represents the degree of influence factor and apparent defect area of the k-th packaging device in the j-th historical packaging record corresponding to the i-th packaging body, respectively. k Let represent the original surface area corresponding to the k-th packaged device, and e represent the natural constant.
[0024] η i j k and σ i j Substituting k into the formula The wear and tear index of each packaging device in each historical packaging record corresponding to each packaging entity was statistically obtained. Where A and B represent the predefined proportional coefficients corresponding to shape deformation and appearance defects, respectively.
[0025] The wear and tear indices of each packaging device in each historical packaging record for each packaging entity are summed to obtain the packaging production line wear rate for each historical packaging record for each packaging entity, denoted as φ. i j.
[0026] Furthermore, the specific implementation process for predicting the number of packages maintained by the current packaging production line relative to each packaging entity in terms of appearance is as follows: extract the number of packages corresponding to each historical packaging record for each packaging entity from the packaging parameters.
[0027] The usage loss rate V of the packaging production line for each packaging entity is calculated based on the number of packages and the usage loss rate of the packaging production line for each historical packaging record corresponding to each packaging entity. i Its calculation formula is Where X i j represents the number of packages corresponding to the j-th historical package record for the i-th package entity.
[0028] Extract the normal permissible wear index of each packaging equipment on the packaging production line of the target semiconductor packaging company from the processing information database, and then take the minimum normal permissible wear index as the normal permissible loss degree of the packaging production line, denoted as φ. 正常 .
[0029] The historical packaging records are arranged according to their corresponding numbers for each packaging entity. Then, based on the arrangement of the historical packaging records, the packaging production line usage loss rate corresponding to the last historical packaging record is extracted from the packaging production line usage loss rate of each historical packaging record for each packaging entity. This is taken as the current packaging production line usage loss rate relative to each packaging entity, denoted as φ. i当前 .
[0030] V i φ 正常 and φ i当前 Substitute into the apparent prediction formula The number of packages maintained in the current packaging production line relative to each package body in terms of appearance dimension is obtained, Q. PF i.
[0031] Furthermore, the prediction of the number of packages maintained by the current packaging production line relative to each packaging entity over time specifically includes the following steps: using the formula Analysis of the current packaging production line's holding time T for each packaged unit i ξ represents the usage loss of the packaging production line corresponding to a unit packaging time.
[0032] The packaging duration and number of packages are extracted from the packaging parameters of each historical packaging record corresponding to each packaging entity. This data is used to calculate the packaging duration t for a single packaged product for each packaging entity. i , Where F i j represents the encapsulation duration of the j-th historical encapsulation record corresponding to the i-th encapsulation subject.
[0033] Using time prediction formula Predict the number of packages Q that the current packaging production line will maintain relative to each package body over time. TM i, where △t represents the pre-startup time corresponding to the packaging production line.
[0034] Furthermore, determining the effective number of packages to be maintained relative to each package body in the current packaging production line refers to the following steps: Q PF i and Q TM i is compared to calculate the similarity of the current packaging production line to the predicted number of packages maintained by each packaging entity. Where U represents a preset constant, and U>0.
[0035] ψ i Comparing the current packaging production line with the set critical prediction similarity, if the prediction similarity of the current packaging production line relative to the maintained packaging quantity of a certain packaging entity is less than or equal to the set critical prediction similarity, then... As the current packaging production line effectively maintains the number of packages relative to the main packaging unit, Q PF Q TM These represent the number of packages maintained by the current packaging production line relative to the packaged entity in terms of appearance and time, respectively. Conversely, α*Q represents the number of packages maintained. PF +(1-α)*Q TM As the effective number of packages maintained relative to the current packaging production line and the main packaging entity, α represents the trade-off factor corresponding to the appearance dimension.
[0036] Furthermore, the formula for evaluating the packaging efficiency of the current packaging production line relative to each packaging entity is R.i =r i *Q 有效 i, R i Let r represent the packaging efficiency of the current packaging production line relative to the i-th packaging entity. i Let Q represent the unit packaging revenue corresponding to the i-th packaging entity. 有效 i represents the number of packages that the current packaging production line can effectively maintain relative to the i-th packaging entity.
[0037] Compared with the prior art, the present invention has the following advantages: 1. The present invention extracts packaging data of the packaging production line under different packaging subjects from historical packaging records, and analyzes the loss degree of the packaging production line under different packaging subjects. This realizes an intuitive, direct and rational analysis of the loss degree of the packaging production line under different packaging subjects, which not only greatly improves the credibility of the analysis results, but also makes full use of historical packaging information, realizes the high-value utilization of packaging data information, and avoids the waste of packaging resources to a certain extent.
[0038] 2. After analyzing the loss rate of the packaging production line under different packaging entities, this invention further predicts the number of packages that the current packaging production line can maintain relative to each packaging entity based on the current loss status of the packaging production line. This allows for an assessment of the packaging efficiency of the current packaging production line relative to each packaging entity, enabling targeted selection of appropriate packaging entities and providing a strong implementation guarantee for maximizing packaging efficiency.
[0039] 3. When predicting the number of packages to be maintained relative to each package body in the current packaging production line, the present invention starts from two dimensions and realizes multi-dimensional prediction of the number of packages to be maintained. Compared with single-dimensional prediction, this prediction method can minimize prediction errors and improve the accuracy of prediction results. Attached Figure Description
[0040] The present invention will be further described with reference to the accompanying drawings, but the embodiments in the drawings do not constitute any limitation on the present invention. For those skilled in the art, other drawings can be obtained based on the following drawings without creative effort.
[0041] Figure 1 This is a schematic diagram of the system connection of the present invention. Detailed Implementation
[0042] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0043] Reference Figure 1 As shown, this invention proposes a semiconductor packaging data information analysis and processing system, including a semiconductor packaging equipment statistics module, a feature packaging data extraction module, a packaging production line usage loss analysis module, an apparent maintenance packaging quantity prediction module, a time maintenance packaging quantity prediction module, an effective maintenance packaging quantity determination module, a processing information database, a packaging benefit evaluation module, and an adaptive packaging subject screening and display module.
[0044] The semiconductor packaging equipment statistics module is connected to the feature packaging data extraction module. The feature packaging data extraction module is connected to the packaging production line usage loss analysis module. The packaging production line usage loss analysis module is connected to the apparent maintenance packaging quantity prediction module and the time maintenance packaging quantity prediction module, respectively. The apparent maintenance packaging quantity prediction module and the time maintenance packaging quantity prediction module are both connected to the effective maintenance packaging quantity determination module. The effective maintenance packaging quantity determination module is connected to the packaging benefit evaluation module. The packaging benefit evaluation module is connected to the compatible packaging subject screening and display module. The processing information database is connected to the packaging production line usage loss analysis module, the apparent maintenance packaging quantity prediction module, and the time maintenance packaging quantity prediction module, respectively.
[0045] The semiconductor packaging equipment acquisition module is used to count the number of packaging equipment on the corresponding packaging production line of the target semiconductor packaging company and to obtain the name of each packaging equipment. Specifically, packaging equipment generally includes solder paste printers, die bonders, reflow solderers, dicing machines, etc., among which the most critical is the die bonder.
[0046] The feature packaging data extraction module is used to extract packaging parameters from each historical packaging record corresponding to the target semiconductor packaging company. The packaging parameters include wafer type, packaging quantity, packaging duration, and video of the packaging production line usage process. It also identifies feature packaging data, which includes the packaging body, the frequency of use of the packaging production line, and usage information corresponding to each use of the packaging production line. The usage information is the usage video segment corresponding to each packaging device.
[0047] Specifically, after extracting the historical encapsulation records, each historical encapsulation record is numbered according to the chronological order of the encapsulation time.
[0048] It should be noted that the packaging production line is designed for packaging various wafers. It should also be noted that the wafer model mentioned above refers to the wafer diameter. For example, the wafer diameters available on the market are mainly 150mm, 200mm, and 300mm, which correspond to 6-inch, 8-inch, and 12-inch wafers, respectively.
[0049] In a specific embodiment of the present invention, the specific identification process corresponding to the feature packaging data is as follows: extract the wafer model from the packaging parameters corresponding to each historical packaging record, and then classify the historical packaging records corresponding to the same wafer model. At this time, the same wafer model is recorded as the packaging body, thereby obtaining several historical packaging records corresponding to each packaging body.
[0050] The number of historical packaging records owned by each packaging entity is counted, and this number is used as the usage frequency of the packaging production line corresponding to each packaging entity.
[0051] The video of the packaging production line usage process is extracted from the packaging parameters of each historical packaging record of each packaging entity, and then segmented according to the individual packaging equipment to obtain the usage video segment corresponding to each packaging equipment in each historical packaging record owned by each packaging entity.
[0052] As a specific implementation method, the specific operation process of segmenting the packaging production line usage video is as follows: the start time stamp and end time stamp of each packaging device are identified from the packaging production line usage video according to the appearance shape of each packaging device, and then the packaging production line usage video is segmented based on the start time stamp and end time stamp.
[0053] The packaging production line usage loss analysis module is used to analyze the usage loss of the packaging production line for each historical packaging record corresponding to each packaging entity.
[0054] In a further optimized scheme, the specific operation steps corresponding to the usage loss degree of each packaging production line for each packaging entity and each historical packaging record are as follows: the usage video segments of each packaging device in each historical packaging record for each packaging entity are decomposed into video frames to obtain several usage images corresponding to each packaging device. The usage images are sorted according to their timestamps in the usage video segments, and the first and last usage images are taken and recorded as the initial usage image and the end usage image, respectively.
[0055] The outer contour is extracted from the initial and final usage images corresponding to each packaging device. The final usage contour of each packaging device is then superimposed with the initial usage contour. Based on this, the external usage deformation degree of each packaging device in each historical packaging record is calculated, denoted as σ. i jk, where i represents the number of the packaging entity, i = 1, 2, ..., n, n represents the number of packaging entities, j represents the number of historical packaging records, j = 1, 2, ..., m, m represents the number of historical packaging records owned by each packaging entity, and k represents the number of packaging equipment, k = 1, 2, ..., z, z represents the number of packaging equipment owned by the target semiconductor packaging company on its corresponding packaging production line.
[0056] The above calculation process for the external shape deformation degree of each packaging device in each historical packaging record corresponding to each packaging entity is as follows: First, based on the external shape contour overlap comparison results, obtain the external shape contour overlap area corresponding to each packaging device in each historical packaging record corresponding to each packaging entity. Then, based on the name of each packaging device, extract the original surface area of each packaging device from the processing information database. Finally, divide the external shape contour overlap area corresponding to each packaging device in each historical packaging record corresponding to each packaging entity by the original surface area of the corresponding packaging device to obtain the external shape deformation degree of each packaging device in each historical packaging record corresponding to each packaging entity.
[0057] The end-of-use image for each packaging device is compared with the initial-use image to identify abnormal areas. The end-of-use image is then focused on these abnormal areas to extract the appearance defect parameters for each packaging device. These parameters include the appearance defect type and area. As an example, appearance defect types include, but are not limited to, dents, peeling, cracks, warping, etc.
[0058] Based on the apparent defect type corresponding to each packaging device, the degree of influence factor corresponding to each packaging device is obtained from the processing information database. Specifically, the apparent defect type corresponding to each packaging device is matched with the degree of influence factors corresponding to various apparent defect types stored in the processing information database to obtain the degree of influence factor corresponding to each packaging device.
[0059] Through formula The apparent usage defect η of each packaging device in each historical packaging record corresponding to each packaging entity was statistically obtained. i j k, where ε i j k, s i j k represents the degree of influence factor and apparent defect area of the k-th packaging device in the j-th historical packaging record corresponding to the i-th packaging body, respectively. k Let represent the original surface area corresponding to the k-th packaging device, and e represent the natural constant. In this invention, both the degree of influence factor of the packaging device and the apparent defect area have a positive impact on the apparent use defect degree of the packaging device.
[0060] η ij k and σ i j Substituting k into the formula The wear and tear index of each packaging device in each historical packaging record corresponding to each packaging entity was statistically obtained. Where A and B represent the predefined proportional coefficients corresponding to shape deformation and appearance defects, respectively. In the above formula for calculating the wear index, the greater the shape deformation and appearance defects of a certain packaging device, the greater the wear index of that packaging device.
[0061] The wear and tear indices of each packaging device in each historical packaging record for each packaging entity are summed to obtain the packaging production line wear rate for each historical packaging record for each packaging entity, denoted as φ. i j.
[0062] It is particularly important to note that all equipment inevitably experiences wear and tear during use. It is precisely because of this wear and tear on the packaging equipment that the packaging production line suffers losses. Furthermore, the degree of wear on the packaging equipment directly reflects the loss status of the packaging production line. This invention takes this into account and, by analyzing the wear of the packaging equipment on the packaging production line, can achieve the most intuitive analysis of the wear and tear on the packaging production line.
[0063] Furthermore, this invention takes into account that during the operation of packaging equipment, the surface of the equipment undergoes various complex changes due to friction under the action of force, resulting in morphological changes. These morphological changes are manifested in both geometric deformation and surface defects. Therefore, this invention analyzes the wear and tear of packaging equipment from both aspects of shape deformation and surface defects, effectively reflecting the actual wear and tear of the equipment during use. This makes the analysis results more realistic and objective, and helps to improve the accuracy and reliability of the analysis results.
[0064] This invention extracts packaging data from historical packaging records under different packaging entities, thereby analyzing the loss rate of the packaging production line under different packaging entities. This enables a more intuitive, direct, and rational analysis of the loss rate of the packaging production line under different packaging entities, which not only greatly improves the reliability of the analysis results, but also makes full use of historical packaging information, realizing the high-value utilization of packaging data information and avoiding the waste of packaging resources to a certain extent.
[0065] The apparent maintenance package quantity prediction module is used to predict the apparent maintenance package quantity of the current packaging production line relative to each packaging entity. The specific implementation process is as follows: extract the packaging quantity of each packaging entity corresponding to each historical packaging record from the packaging parameters according to the number of each historical packaging record corresponding to each packaging entity.
[0066] The usage loss rate V of the packaging production line for each packaging entity is calculated based on the number of packages and the usage loss rate of the packaging production line for each historical packaging record corresponding to each packaging entity. i Its calculation formula is Where X i j represents the number of packages corresponding to the j-th historical package record for the i-th package entity.
[0067] The aforementioned packaging production line usage loss rate, which needs to be explained, refers to the packaging production line usage loss rate corresponding to a unit number of packages. Since each semiconductor packaging process does not involve packaging just one finished product, but rather batch packaging, meaning that each packaging operation involves a certain number of packages, the packaging production line usage loss rate corresponding to each historical packaging record for each packaging entity is not generated by packaging just one finished product, but rather by the accumulation of a large number of packaged finished products. Therefore, it is necessary to analyze the loss generated by a single package quantity.
[0068] Extract the normal permissible wear index of each packaging equipment on the packaging production line of the target semiconductor packaging company from the processing information database, and then take the minimum normal permissible wear index as the normal permissible loss degree of the packaging production line, denoted as φ. 正常 .
[0069] It should be explained that the normal allowable wear index mentioned above refers to the wear index that does not affect normal use. Due to the different precision of each packaging equipment, the normal allowable wear index varies for different packaging equipment. The larger the normal allowable wear index of a packaging equipment, the more wear-resistant the packaging equipment is. Here, the minimum normal allowable wear index is taken as the normal allowable loss degree corresponding to the packaging production line in order to meet the normal use of all packaging equipment.
[0070] The historical packaging records are arranged according to their corresponding numbers for each packaging entity. Then, based on the arrangement of the historical packaging records, the packaging production line usage loss rate corresponding to the last historical packaging record is extracted from the packaging production line usage loss rate of each historical packaging record for each packaging entity. This is taken as the current packaging production line usage loss rate relative to each packaging entity, denoted as φ. i当前 .
[0071] V i φ 正常 and φ i当前 Substitute into the apparent prediction formula The number of packages maintained in the current packaging production line relative to each package body in terms of appearance dimension is obtained, Q. PF i, where the lower the wear rate of the packaging production line corresponding to a certain packaging entity, the more packaging production lines maintain the number of packages relative to that packaging entity in terms of appearance.
[0072] The time-sustaining package quantity prediction module is used to predict the number of packages maintained relative to each package entity in the current packaging production line over time. Specifically, it includes the following steps: using the formula... Analysis of the current packaging production line's holding time T for each packaged unit i ξ represents the usage loss of the packaging production line corresponding to a unit packaging time.
[0073] The packaging duration and number of packages are extracted from the packaging parameters of each historical packaging record corresponding to each packaging entity. This data is used to calculate the packaging duration t for a single packaged product for each packaging entity. i , Where F i j represents the encapsulation duration of the j-th historical encapsulation record corresponding to the i-th encapsulation subject.
[0074] Using time prediction formula Predict the number of packages Q that the current packaging production line will maintain relative to each package body over time. TM i, where △t represents the pre-startup time corresponding to the packaging production line.
[0075] The effective package maintenance quantity determination module is used to determine the effective package maintenance quantity of the current packaging production line relative to each package body. See the following steps for details: Q PF i and Q TM i is compared to calculate the similarity ψ between the current packaging production line and the predicted number of packages to be maintained for each packaging entity. i , Where U represents a preset constant, and U>0, the closer the number of packages maintained by the current packaging production line relative to a certain packaging entity in the time dimension is to the number of packages maintained in the appearance dimension, the greater the similarity of the predicted number of packages maintained by the current packaging production line relative to that packaging entity.
[0076] ψ i Comparing the current packaging production line with the set critical prediction similarity, if the prediction similarity of the current packaging production line relative to the maintained packaging quantity of a certain packaging entity is less than or equal to the set critical prediction similarity, then... As the current packaging production line effectively maintains the number of packages relative to the main packaging unit, Q PF Q TM These represent the number of packages maintained by the current packaging production line relative to the packaged entity in terms of appearance and time, respectively. Conversely, α*Q represents the number of packages maintained. PF +(1-α)*Q TM As the effective number of packages maintained relative to the current packaging production line and the main packaging entity, α represents the trade-off factor corresponding to the appearance dimension.
[0077] This invention predicts the number of packages to be maintained relative to each package body in the current packaging production line from two dimensions, thus achieving multi-dimensional prediction of the number of packages to be maintained. Compared with single-dimensional prediction, this prediction method can minimize prediction errors and improve the accuracy of prediction results.
[0078] The processing information database is used to store the original surface area of each packaging device, the normal allowable wear index of each packaging device on the packaging production line of the target semiconductor packaging company, the degree of influence factor corresponding to various apparent defect types, the usage loss degree of the packaging production line corresponding to the unit packaging time, and the pre-start-up time of the packaging production line.
[0079] The packaging efficiency evaluation module extracts the unit packaging revenue corresponding to each packaging entity from the processing information database, and combines it with the effective maintenance packaging quantity of the current packaging production line relative to each packaging entity to evaluate the packaging efficiency of the current packaging production line relative to each packaging entity. The evaluation formula is R. i =r i *Q 有效 i, R i Let r represent the packaging efficiency of the current packaging production line relative to the i-th packaging entity. i Let Q represent the unit packaging revenue corresponding to the i-th packaging entity. 有效 i represents the effective number of packages maintained relative to the i-th package body in the current packaging production line, where Q 有效 The value of i is or α*Q PF +(1-α)*Q TM .
[0080] The adaptive packaging body screening and display module is used to select the packaging body with the maximum packaging benefit from the packaging benefits of each packaging body in the current packaging production line as the adaptive packaging body, and display it.
[0081] After analyzing the loss rate of the packaging production line under different packaging entities, this invention further predicts the number of packages that the current packaging production line can maintain relative to each packaging entity based on the current loss status of the packaging production line. This allows for an assessment of the packaging efficiency of the current packaging production line relative to each packaging entity, enabling targeted selection of appropriate packaging entities and providing a strong implementation guarantee for maximizing packaging efficiency.
[0082] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.
Claims
1. A semiconductor packaging data information analysis and processing system, characterized in that, include: The semiconductor packaging equipment statistics module is used to count the number of packaging equipment on the corresponding packaging production line of the target semiconductor packaging company and obtain the name of each packaging equipment. The feature packaging data extraction module is used to extract packaging parameters from each historical packaging record corresponding to the target semiconductor packaging company and identify feature packaging data. Feature packaging data includes the packaging body, the frequency of use of the packaging production line and the usage information corresponding to each use of the packaging production line, wherein the usage information is the usage video segment corresponding to each packaging equipment. The packaging production line uses a loss analysis module to analyze the usage loss of the packaging production line corresponding to each historical packaging record for each packaging entity. The apparent sustaining package quantity prediction module is used to predict the current packaging production line's sustaining package quantity relative to each package body in the apparent dimension. The time-sustaining package quantity prediction module is used to predict the number of packages that the current packaging production line will maintain relative to each package body over time. The effective package quantity determination module is used to determine the effective package quantity to be maintained relative to each package body in the current packaging production line. The information database is used to store the original surface area of each packaging device, the normal allowable wear index of each packaging device on the packaging production line of the target semiconductor packaging company, the degree of influence factor of various apparent defect types, the usage loss degree of the packaging production line corresponding to the unit packaging time, and the pre-startup time of the packaging production line. The packaging efficiency evaluation module is used to extract the unit packaging revenue corresponding to each packaging entity from the processing information database, and combine it with the effective maintenance packaging quantity of the current packaging production line relative to each packaging entity to evaluate the packaging efficiency of the current packaging production line relative to each packaging entity. The matching packaging body screening and display module is used to select the packaging body with the highest packaging benefit from the packaging benefits of each packaging body in the current packaging production line as the matching packaging body and display it.
2. The semiconductor packaging data information analysis and processing system according to claim 1, characterized in that: The packaging parameters include wafer type, number of packages, packaging time, and a video of the packaging production line process.
3. The semiconductor packaging data information analysis and processing system according to claim 1, characterized in that: The specific recognition process corresponding to the feature encapsulation data is as follows: The wafer model is extracted from the packaging parameters corresponding to each historical packaging record. Then, the historical packaging records corresponding to the same wafer model are classified. At this time, the same wafer model is recorded as the packaging body, thus obtaining several historical packaging records corresponding to each packaging body. The number of historical packaging records owned by each packaging entity is counted and used as the usage frequency of the packaging production line corresponding to each packaging entity. The video of the packaging production line usage process is extracted from the packaging parameters of each historical packaging record of each packaging entity, and then segmented according to the individual packaging equipment to obtain the usage video segment corresponding to each packaging equipment in each historical packaging record owned by each packaging entity.
4. The semiconductor packaging data information analysis and processing system according to claim 1, characterized in that: The specific operational steps for analyzing the usage loss rate of each packaging body corresponding to each historical packaging record of the packaging production line are as follows: The usage video segments of each packaging device in each historical packaging record corresponding to each packaging entity are decomposed according to video frames to obtain several usage images corresponding to each packaging device. Each usage image is sorted according to its timestamp in the usage video segment, and then the first and last usage images are taken and recorded as the initial usage image and the end usage image, respectively. The outer contour is extracted from the initial and final usage images corresponding to each packaging device. The final usage contour of each packaging device is then superimposed with the initial usage contour. Based on this, the external usage deformation degree of each packaging device in each historical packaging record is calculated, denoted as [missing information]. Where i represents the number of the packaged entity, n represents the number of packaged entities, and j represents the number of the historical package record. , m represents the number of historical packaging records owned by each packaging entity, and k represents the packaging device number. z represents the number of packaging devices on the corresponding packaging production line of the target semiconductor packaging company; The end-of-use image of each packaging device is compared with the initial use image to identify abnormal areas. The end-of-use image is then focused on the abnormal areas to extract the appearance defect parameters of each packaging device. The appearance defect parameters include the appearance defect type and the appearance defect area. Based on the apparent defect type corresponding to each packaging device, the degree of influence factor corresponding to each packaging device is obtained by matching from the processing information database; Through formula The apparent usage defect rate of each packaging device in each historical packaging record corresponding to each packaging entity was statistically obtained. ,in , These represent the degree of influence factor and apparent defect area of the k-th packaging device in the j-th historical packaging record corresponding to the i-th packaging body, respectively. Let be the original surface area corresponding to the k-th packaging device, and e be the natural constant. Will and Substitute into the formula The wear and tear index of each packaging device in each historical packaging record corresponding to each packaging entity was statistically obtained. , where A and B represent the predefined proportional coefficients corresponding to shape deformation and appearance defects, respectively; The wear and tear indices of each packaging device in each historical packaging record for each packaging entity are summed to obtain the packaging production line wear rate for each historical packaging record for each packaging entity, denoted as . .
5. The semiconductor packaging data information analysis and processing system according to claim 4, characterized in that: The specific implementation process for predicting the number of packages maintained by the current packaging production line relative to each packaging entity in terms of appearance is as follows: Extract the number of packages corresponding to each historical package record for each package body from the package parameters; The usage loss rate of the packaging production line for each packaging entity is calculated based on the number of packages and the usage loss rate of the packaging production line for each historical packaging record corresponding to each packaging entity. Its calculation formula is ,in This represents the number of packages corresponding to the j-th historical package record for the i-th package entity; Extract the normal permissible wear index of each packaging equipment on the packaging production line of the target semiconductor packaging company from the processing information database. Then, take the minimum normal permissible wear index as the normal permissible loss rate of the packaging production line, denoted as . ; The historical packaging records are arranged according to their corresponding numbers for each packaging entity. Then, based on this arrangement, the packaging production line usage loss rate of the last-ranked historical packaging record is extracted from the packaging production line usage loss rate of each historical packaging record for each packaging entity. This is taken as the current packaging production line's usage loss rate relative to each packaging entity, denoted as . ; Will , and Substitute into the apparent prediction formula This yields the number of packages maintained by the current packaging production line relative to each packaging entity in terms of appearance. .
6. The semiconductor packaging data information analysis and processing system according to claim 5, characterized in that: The prediction of the number of packages that the current packaging production line will maintain relative to each packaging entity over time specifically includes the following steps: Through formula Analysis of the current packaging production line's maintenance packaging time relative to each packaged unit ,in This represents the usage loss of the packaging production line corresponding to a unit packaging time. The packaging duration and number of packages are extracted from the packaging parameters of each historical packaging record corresponding to each packaging entity, and the packaging duration of a single packaged product corresponding to each packaging entity is calculated accordingly. , ,in This represents the encapsulation duration of the j-th historical encapsulation record corresponding to the i-th encapsulation entity; Using time prediction formula Predict the number of packages that the current packaging production line will maintain relative to each packaging entity over time. ,in This represents the pre-startup time for the packaging production line.
7. A semiconductor packaging data information analysis and processing system according to claim 6, characterized in that: Determining the effective number of packages to be maintained relative to each package body on the current packaging production line is described in the following steps: Will and By comparing the current packaging production line with the predicted number of packages to be maintained for each packaging entity, the similarity can be calculated. , Where U represents a preset constant, and U>0; Will Comparing the current packaging production line with the set critical prediction similarity, if the prediction similarity of the current packaging production line relative to the maintained packaging quantity of a certain packaging entity is less than or equal to the set critical prediction similarity, then... As the current packaging production line effectively maintains the number of packages relative to the main packaging unit, among which , These represent the number of packages maintained by the current packaging production line relative to the packaged entity in terms of appearance and time, respectively; conversely, they are represented by... As a means to effectively maintain the number of packages relative to the current packaging production line and the main packaging unit. This is represented as the trade-off factor corresponding to the apparent dimension.
8. The semiconductor packaging data information analysis and processing system according to claim 7, characterized in that: The formula for evaluating the packaging efficiency of the current packaging production line relative to each packaging entity is as follows: , This represents the packaging efficiency of the current packaging production line relative to the i-th packaging entity. This represents the unit packaging revenue corresponding to the i-th packaging entity. This represents the number of packages that the current packaging production line can effectively maintain relative to the i-th packaging entity.