Clean bench cleanliness testing method and system

Abstract

The present application discloses a clean bench cleanliness test method and system, which relates to the field of cleanliness test technology. The method includes dividing the target workbench into a number of regional blocks to obtain a block set, detecting the first and second detection data sets of the block set, predicting the third detection data set, performing difference comparison on the second and third detection data sets, selecting the target, the first and second associated blocks from the obtained block set to be checked, comparing the target block with the parameter sets detected by the first and second associated blocks, respectively, performing corresponding transfer item parameter correction processing on the first associated detection data, performing corresponding transfer item parameter correction processing on the second associated detection data, performing corresponding transfer item parameter correction processing on the target detection data, and performing final cleanliness evaluation on the statistically corrected detection data set to obtain a cleanliness test result. The clean bench cleanliness test method and system provided by the present invention improves the efficiency and accuracy of cleanliness detection work.

Description

Technical Field

[0001] The present application relates to the technical field of cleanliness testing, and in particular to a clean bench cleanliness testing method and system. Background Art

[0002] The purpose of regular performance testing or inspection of clean rooms or clean areas is to certify that the clean rooms or clean areas always meet the requirements of standards and specifications. It is of great significance to ensure product quality, improve production efficiency, monitor equipment performance, and support scientific research and innovation.

[0003] In traditional technology, the cleanliness test of the clean workbench is carried out by performing a series of tests on the whole, without considering that different location areas are very likely to have correlation effects with each other, which leads to certain errors in the parameter data of the initial test and reduces the efficiency of the test work. Summary of the invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides a clean bench cleanliness testing method and system.

[0005] In a first aspect, the present application provides a clean bench cleanliness testing method, the method comprising:

[0006] Acquire historical test record data of the target workstation, and train the evolutionary causal model according to the historical test record data to obtain a trained evolutionary causal model;

[0007] Dividing the target workbench into a number of regional blocks to obtain a block set, detecting a first detection data set and a second detection data set of the block set, predicting a third detection data set according to the first detection data set and the trained evolutionary causal model, and performing a difference comparison between the second detection data set and the third detection data set to screen out a block set to be verified from the block set;

[0008] Selecting a target block, a first associated block, and a second associated block from the set of blocks to be checked, performing comparison processing on the target block and the detected parameter set to which the first associated block belongs to obtain a first preprocessing item parameter, and performing comparison processing on the target block and the detected parameter set to which the second associated block belongs to obtain a second preprocessing item parameter;

[0009] Performing a transfer amount evaluation on the first preprocessing item parameter to obtain a first transfer amount, and performing correction processing on the first associated detection data corresponding to the transfer item parameter according to the first transfer amount to obtain first associated corrected detection data;

[0010] Performing transfer amount evaluation on the second preprocessing item parameter to obtain a second transfer amount, performing correction processing on the second associated detection data according to the second transfer amount, and obtaining second associated corrected detection data; performing correction processing on the target detection data according to the first transfer amount and the second transfer amount, and obtaining target corrected detection data;

[0011] According to the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, statistics are collected on the corrected detection data set to which the block set to be checked belongs, and a final cleanliness evaluation is performed on the corrected detection data set to obtain a cleanliness test result.

[0012] Preferably, historical test record data of cleanliness tests performed on a target workbench in historical periods are obtained, and various parameter test data of several historical periods are selected from the historical test record data, and the correlation influence values ​​of various parameter test data of several historical periods are respectively evaluated to obtain historical correlation degree data, and according to the historical correlation degree data, a historical correlation degree data set is output;

[0013] The historical correlation degree data set is input into the evolutionary causal model for training to obtain a trained evolutionary causal model.

[0014] Preferably, the target workbench is divided into a number of area blocks to obtain a block set, and data detection of various key parameters is performed on each position area in the block set to obtain a first detection data set, and when each parameter in the first detection data set is detected, other parameters are under stable or ideal conditions;

[0015] A second detection data set is obtained by performing data detection of various key parameters on each location area in the block set. When each parameter in the second detection data set is detected, other parameters are under non-ideal conditions;

[0016] Inputting the second detection data set into the trained evolutionary causal model for testing to obtain a predicted correlation degree data set;

[0017] predicting a third detection data set based on the predicted correlation degree data set under the effect of the predicted correlation degree data set on the first detection data set;

[0018] Comparing the difference degree of each parameter of the second test data set and the third test data set to obtain a difference degree comparison result set;

[0019] A difference determination threshold is preset, and blocks corresponding to comparison results having a value greater than the difference determination threshold are selected from the block set in the difference degree comparison result set, and a block set to be checked is output.

[0020] Preferably, a target block is selected from the set of blocks to be checked, and according to the target block, a first associated block and a second associated block adjacent to the target block are screened from the set of blocks to be checked, and the positions of the first associated block and the second associated block are not adjacent;

[0021] Extracting corresponding target detection data, first associated detection data, and second associated detection data from the detection data set according to the target block, the first associated block, and the second associated block;

[0022] Comparing the target detection data and the first associated detection data for differences in various parameters to obtain a first difference parameter set;

[0023] Extracting transferable parameter item data from the first difference item parameter set to obtain first preprocessing item parameters;

[0024] Comparing the target detection data with the second associated detection data for differences in various parameters to obtain a second difference parameter set;

[0025] Transferable parameter item data is extracted from the second difference item parameter set to obtain second preprocessing item parameters.

[0026] Preferably, two difference item parameters of the same category in the first preprocessing item parameters are differentiated to obtain a first preprocessing difference value;

[0027] According to the first preprocessing difference, parameter data of the same category as the first preprocessing difference is removed from the first associated detection data to obtain first screened associated detection data;

[0028] Inputting the first preprocessing difference and the first screening association detection data into the trained evolutionary causal model for testing to obtain a first transfer association influence degree data set;

[0029] evaluating a corresponding first transfer amount according to the first transfer association influence degree data set and the first preprocessing difference;

[0030] According to the first transfer amount, the first associated detection data is corrected by performing a correction process on the corresponding transfer item parameter to obtain the first associated corrected detection data.

[0031] Preferably, two difference item parameters of the same category in the second preprocessing item parameters are differentiated to obtain a second preprocessing difference value;

[0032] According to the second preprocessing difference, parameter data of the same category as the second preprocessing difference is removed from the second associated detection data to obtain second screened associated detection data;

[0033] Inputting the second preprocessing difference and the second screening association detection data into the trained evolutionary causal model for testing to obtain a second transfer association influence degree data set;

[0034] Evaluate a corresponding second transfer amount according to the second transfer association influence degree data set and the second preprocessing difference;

[0035] According to the second transfer amount, the second associated detection data is corrected by a corresponding transfer item parameter to obtain second associated corrected detection data;

[0036] According to the first transfer amount and the second transfer amount, the target detection data is corrected by performing correction processing on the corresponding transfer item parameters to obtain target corrected detection data.

[0037] Preferably, according to the target block, the first associated block and the second associated block, the remaining blocks in the block set to be checked are preprocessed to obtain a preprocessed remaining block set;

[0038] According to the preprocessing remaining block set, a corresponding preprocessing remaining detection data set is screened out from the detection data set;

[0039] According to the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, a data correction process is performed on the preprocessed remaining detection data set to obtain remaining corrected detection data;

[0040] The target corrected detection data, the first associated corrected detection data, the second associated corrected detection data and the remaining corrected detection data are combined into a corrected detection data set;

[0041] A final cleanliness evaluation is performed on each corrected detection data in the corrected detection data set to obtain a cleanliness test result.

[0042] In a second aspect, a clean bench cleanliness testing system comprises:

[0043] A model training unit is used to obtain historical test record data of a target workbench, and train an evolutionary causal model according to the historical test record data to obtain a trained evolutionary causal model;

[0044] A block screening unit to be tested is used to divide the target workbench into a number of regional blocks to obtain a block set, detect a first test data set and a second test data set of the block set, predict a third test data set based on the first test data set and the trained evolutionary causal model, and compare the second test data set with the third test data set to screen out the block set to be tested from the block set;

[0045] a difference item parameter extraction unit, configured to select a target block, a first associated block, and a second associated block from the block set to be checked, compare the detected parameter set of the target block and the first associated block to obtain a first preprocessing item parameter, and compare the detected parameter set of the target block and the second associated block to obtain a second preprocessing item parameter;

[0046] The first correlation block parameter correction unit is used to perform a transfer amount evaluation on the first preprocessing item parameter to obtain a first transfer amount, and perform a correction process on the first correlation detection data corresponding to the transfer item parameter according to the first transfer amount to obtain the first correlation correction detection data;

[0047] The second associated block parameter correction unit is used to perform transfer amount evaluation on the second preprocessing item parameter to obtain a second transfer amount, perform correction processing on the second associated detection data according to the second transfer amount, and obtain second associated corrected detection data; perform correction processing on the target detection data according to the first transfer amount and the second transfer amount, and obtain target corrected detection data;

[0048] The cleanliness final evaluation unit is used to count the corrected detection data set to which the block set to be checked belongs according to the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, and to perform a final cleanliness evaluation on the corrected detection data set to obtain a cleanliness test result.

[0049] Compared with the prior art, the present invention has the following characteristics and beneficial effects:

[0050] By dividing the clean workbench into several blocks, it is convenient to perform individual detection of the parameters belonging to each divided block, thereby improving the accuracy of data detection and avoiding the large error caused by the unified data detection of traditional technology, that is, not taking into account the different degrees of differences in the information of the parameters belonging to different location areas, and performing comparative analysis of the parameters under two conditions to determine which blocks in the block set need to be re-detected and evaluated for the secondary parameters, that is, one condition is that when a certain parameter is detected, the other parameters are under stable or ideal conditions, and the other condition is that when a certain parameter is detected, the other parameters are under non-ideal conditions, and using The historical detection record data is used to train the evolutionary causal model, which fully improves the utilization rate of data resources, conducts scientific research on the sustainable development of effective data resource information, and enhances the reliability of subsequent data information processing results. The difference item parameters of the target block are corrected by the parameter transfer influence of the adjacent associated first associated block and the second associated block on the target block, so as to achieve a secondary cleanliness re-evaluation of the target block. Similarly, the same steps are performed on the remaining blocks to be tested in the check block set to achieve the final cleanliness evaluation of the check block set. The entire processing process improves the accuracy and work efficiency of the workbench cleanliness evaluation. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a flowchart of the steps of a clean bench cleanliness test method mainly embodied in this embodiment.

[0052] Figure 2 It is a structural block diagram of a clean bench cleanliness test system mainly embodied in this embodiment. DETAILED DESCRIPTION

[0053] The present invention is further described in detail below in conjunction with the following examples.

[0054] Reference Figure 1 , a clean bench cleanliness testing method and system, the method comprising the following steps:

[0055] S1. Obtain historical test record data of a target workbench, and train an evolutionary causal model based on the historical test record data to obtain a trained evolutionary causal model.

[0056] S2. Divide the target workbench into several regional blocks to obtain a block set, detect a first test data set and a second test data set of the block set, predict a third test data set based on the first test data set and the trained evolutionary causal model, and compare the second test data set with the third test data set to screen out the block set to be verified from the block set.

[0057] S3. Select a target block, a first associated block, and a second associated block from the set of blocks to be checked, compare the detected parameter sets of the target block and the first associated block to obtain first preprocessing item parameters, and compare the detected parameter sets of the target block and the second associated block to obtain second preprocessing item parameters.

[0058] S4. Perform transfer amount evaluation on the first preprocessing item parameter to obtain a first transfer amount, and perform correction processing on the first associated detection data according to the first transfer amount using the corresponding transfer item parameter to obtain first associated corrected detection data.

[0059] S5. Perform transfer amount evaluation on the second preprocessing item parameter to obtain the second transfer amount. Based on the second transfer amount, perform correction processing on the second associated detection data with the corresponding transfer item parameters to obtain the second associated corrected detection data. Based on the first transfer amount and the second transfer amount, perform correction processing on the target detection data with the corresponding transfer item parameters to obtain the target corrected detection data.

[0060] S6. According to the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, statistics are collected on the corrected detection data set to which the block set to be checked belongs, and a final cleanliness evaluation is performed on the corrected detection data set to obtain a cleanliness test result.

[0061] Specifically, the clean workbench is divided into several blocks, so as to facilitate the individual detection of various parameters belonging to each divided block, thereby improving the accuracy of data detection and avoiding the large error caused by the unified data detection of traditional technology, that is, not taking into account that the information of various parameters belonging to different location areas has different degrees of difference, and performing comparative analysis of various parameters under two conditions to determine which blocks in the block set need to be re-detected and evaluated for secondary parameters, that is, one condition is that when a certain parameter is detected, other parameters are under stable or ideal conditions, and the other condition is that when a certain parameter is detected, other parameters are under non-ideal conditions, and The historical detection record data is used to train the evolutionary causal model, which fully improves the utilization rate of data resources, conducts scientific research on the sustainable development of effective data resource information, and enhances the reliability of subsequent data information processing results. The difference item parameters belonging to the target block are corrected by the parameter transfer influence of the adjacent associated first associated block and the second associated block on the target block, so as to achieve a secondary cleanliness re-evaluation of the target block. Similarly, the same steps are performed on the remaining blocks to be tested in the check block set to achieve the final cleanliness evaluation of the check block set. The entire processing process improves the accuracy and work efficiency of the workbench cleanliness evaluation.

[0062] The specific step S1 includes the following sub-steps:

[0063] The historical test record data of cleanliness test of the target workbench in the historical period are obtained, and the parameter test data of several historical periods are selected from the historical test record data. The correlation influence values ​​between the parameter test data of several historical periods are evaluated respectively to obtain the historical correlation degree data, and the historical correlation degree data set is output according to the historical correlation degree data.

[0064] The historical correlation degree dataset is input into the evolutionary causal model for training to obtain the trained evolutionary causal model.

[0065] Specifically, such as historical test record data (including data categories such as wind speed, suspended particle count, settling bacteria, illumination, high-efficiency filter leak detection, noise, airflow pattern and other key parameters), various parameter test data of several historical periods (such as extracting the historical test record data of the five tests closest to the current period, where the various parameter test data refers to wind speed, suspended particle count, settling bacteria, illumination, high-efficiency filter leak detection, noise, airflow pattern, temperature and humidity and other key parameters, if some of them are given as examples, and A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A110, A120, A130, A140, A150, A160, A170, A180, A190, A200, A210, A220, A330, A49, A50, A610, A710, A80, A90, A10 5. A6 is a logo instead of a description), historical correlation data set (such as evaluating the correlation and influence between A1, A2, A3, A4, A5, and A6 detected in the first historical period, and doing the same for the remaining four times), evolutionary causal model (evolutionary causal model is a method for understanding and predicting system behavior. It is usually based on historical data and builds a model by identifying the causal relationship between variables. This model can help us understand which factors have an impact on the results at different points in time and how these impacts change over time).

[0066] The specific step S2 includes the following sub-steps:

[0067] The target workbench is divided into several area blocks to obtain a block set, and data detection of various key parameters is performed on each position area in the block set to obtain a first detection data set. When each parameter in the first detection data set is detected, other parameters are under stable or ideal conditions.

[0068] Data detection of various key parameters is performed on each location area in the block set to obtain a second detection data set. When each parameter in the second detection data set is detected, other parameters are under non-ideal conditions.

[0069] The second test data set is input into the trained evolutionary causal model for testing to obtain a predicted correlation degree data set.

[0070] According to the predicted correlation degree data set, a third detection data set under the effect of the predicted correlation degree data set is predicted.

[0071] The second test data set and the third test data set are compared for the degree of difference of various parameters to obtain a degree of difference comparison result set.

[0072] A difference determination threshold is preset, and blocks corresponding to comparison results with a difference degree comparison result set greater than the difference determination threshold are screened out from the block set, and the block set to be checked is output.

[0073] Specifically, for example, a block set (if divided into six blocks, namely the first block, the second block, the third block, the fourth block, the fifth block, and the sixth block, and each block is equal), a first detection data set (for example, the first block obtains detection data, and refers to the six parameters A1, B1, C1, D1, E1, and F1, and so on, the other five blocks are also tested), when each parameter in the first detection data set is tested, the other parameters are under stable or ideal conditions (that is, when detecting A 1, if B1, C1, D1 are associated with A1, then B1, C1, D1 need to be kept stable or in an ideal state at this time), the second test data set (the explanation is the same as the first test data set, no further explanation is given here, if it is A2, B2, C2, D2, E2, F2), when each parameter in the second test data set is tested, the other parameters are in non-ideal conditions (that is, when A1 is tested, no matter whether other parameters are associated with A1 or not) , there is no need to process the parameters in a stable or ideal state, that is, to maintain real-time data), predicted correlation data set (that is, to predict the correlation data between the parameters after development and evolution through the trained evolutionary causal model), the third detection data set (such as evaluating the parameters of the first block: A1, B1, C1, D1, E1, F1 under the non-ideal condition processing state after the corresponding predicted correlation data, if it is A3, B3, C3, D3, E3, F3), the difference degree comparison result set (that is, A3, B3, C3, D3, E3, F3 are compared with A2, B2, C2, D2, E2, F2 respectively, if the difference comparison results are: 1%, 3%, 2%, 6%, 8%, 10%, if this is located in the first block), preset difference judgment threshold (if it is 5%), and the block set to be verified (the first block can be judged as the block to be verified, because there is a great possibility of error, such as the mutual influence of parameters between related blocks, so a secondary analysis is required).

[0074] The specific step S3 includes the following sub-steps:

[0075] A target block is selected from the set of blocks to be checked, and according to the target block, a first associated block and a second associated block adjacent to the target block are screened from the set of blocks to be checked, wherein the first associated block and the second associated block are not adjacent to each other.

[0076] According to the target block, the first associated block and the second associated block, corresponding target detection data, first associated detection data and second associated detection data are extracted from the detection data set.

[0077] The target detection data and the first associated detection data are compared for differences in various parameters to obtain a first difference parameter set.

[0078] Transferable parameter item data is extracted from the first difference item parameter set to obtain first preprocessing item parameters.

[0079] The target detection data and the second associated detection data are compared for differences in various parameters to obtain a second difference parameter set.

[0080] Transferable parameter item data is extracted from the second difference item parameter set to obtain second preprocessing item parameters.

[0081] Specifically, such as the block set to be verified (if it is determined that the first block, the second block and the fourth block are the blocks to be verified that require secondary parameter detection and analysis), the target block (if it is the second block), the first associated block (if it is the first block) and the second associated block (if it is the fourth block), the target detection data (if it is a1, b1, c1, d1, e1, f1), the first associated detection data (if it is a2, b2, c2, d2, e2, f2) and the second associated detection data (if it is a3, b3, c3, d3, e3, f3), the first difference item parameter set (if it is d, e, f parameters), the first pre-processing item parameter (if the f parameter is very likely to be transferred, and transferable means the transfer of the parameter difference amount with the surrounding adjacent associated blocks), the second difference item parameter set (if it is b, c parameters), and the second pre-processing item parameter (if the b parameter is very likely to be transferred).

[0082] The specific step S4 includes the following sub-steps:

[0083] The difference between two difference item parameters of the same category in the first preprocessing item parameters is calculated to obtain a first preprocessing difference.

[0084] According to the first preprocessing difference, parameter data of the same category as the first preprocessing difference is removed from the first associated detection data to obtain first screened associated detection data.

[0085] The first preprocessing difference and the first screening association detection data are input into the trained evolutionary causal model for testing to obtain a first transfer association influence degree data set.

[0086] A corresponding first transfer amount is evaluated according to the first transfer association influence degree data set and the first preprocessing difference.

[0087] According to the first transfer amount, the first associated detection data is corrected by performing a correction process on the corresponding transfer item parameter to obtain the first associated corrected detection data.

[0088] Specifically, such as the first preprocessing difference (if the difference between the f-item parameters is M1), the first screening associated detection data (a2, b2, c2, d2, e2), the first transfer associated influence degree data set (that is, the association influence degree data between a2, b2, c2, d2, e2, and M1, and what is mainly needed is the association influence degree data between M1 and a2, b2, c2, d2, and e2, if they are Y0, Y1, Y2, Y3, Y4, and Y5 respectively), the first transfer amount (if T1=Y0*M1+Y1*a2+Y2*b2+Y3*c2+Y4*d2+Y5*e2+N, where N is the error term), and the first associated corrected detection data (that is, if the target block is transferred to the first associated block, then f2 in the first associated detection data is corrected to f2+T1, that is, a2, b2, c2, d2, e2, f2+T1).

[0089] The specific step S5 includes the following sub-steps:

[0090] The difference between two difference item parameters of the same category in the second preprocessing item parameters is calculated to obtain a second preprocessing difference.

[0091] According to the second preprocessing difference, parameter data of the same category as the second preprocessing difference is removed from the second associated detection data to obtain second screened associated detection data.

[0092] The second preprocessing difference and the second screening association detection data are input into the trained evolutionary causal model for testing to obtain a second transfer association influence degree data set.

[0093] A corresponding second transfer amount is evaluated according to the second transfer association influence degree data set and the second preprocessing difference.

[0094] According to the second transfer amount, the second associated detection data is corrected by performing a correction process on the corresponding transfer item parameter to obtain the second associated corrected detection data.

[0095] According to the first transfer amount and the second transfer amount, the target detection data is corrected by performing correction processing on the corresponding transfer item parameters to obtain the target corrected detection data.

[0096] Specifically, such as the second preprocessing difference (if the difference between the parameters in item b is M2), the second screening association detection data (a3, c3, d3, e3, f3), the second transfer association influence degree data set (that is, the association influence degree data between a3, M2, c3, d3, e3, f3, and the main need is the association influence degree data between M2 and a3, c3, d3, e3, f3, if they are X0, X1, X2, X3, X4, X5 respectively) , the first transfer amount (if T2=X0*M2+X1*a3+X2*b3+X3*c3+X4*d3+X5*e3+N, where N is the error term), the second associated corrected detection data (that is, if the target block is transferred to the second associated block, then b3 in the second associated detection data is corrected to b3+T2, that is, a3, b3+T2, c3, d3, e3, f3), target corrected detection data (a1, b1-T2, c1, d1, e1, f1).

[0097] The specific step S6 includes the following sub-steps:

[0098] According to the target block, the first associated block and the second associated block, the remaining blocks in the block set to be checked are preprocessed to obtain a preprocessed remaining block set.

[0099] According to the preprocessing residual block set, the corresponding preprocessing residual detection data set is filtered out from the detection data set.

[0100] According to the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, data correction processing is performed on the preprocessed remaining detection data set to obtain remaining corrected detection data.

[0101] The target corrected detection data, the first associated corrected detection data, the second associated corrected detection data and the remaining corrected detection data are combined into a corrected detection data set.

[0102] The final cleanliness evaluation is performed on each corrected test data in the corrected test data set to obtain the cleanliness test result.

[0103] Specifically, such as pre-processing the remaining block set (if it is the fifth block and the sixth block), pre-processing the remaining detection data set (that is, the interpretation is the same as the first associated detection data and the second associated detection data, which will not be explained here), the remaining corrected detection data (that is, the interpretation and processing steps are the same as the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, which will not be explained here), and the cleanliness test results (for example: the target corrected detection data: a1, b1-T2, c1, d1, e1, f1, each parameter is compared with the prescribed judgment standard reference data to evaluate the cleanliness level in the corresponding block, that is, the first block, the second block, the third block, the fifth block and the sixth block are respectively evaluated for cleanliness, that is, finally qualified and unqualified are used as the judgment results, and the unqualified and qualified parameter data of each block are retained, so as to provide the staff with detailed analysis control data, so as to realize the subsequent targeted response and treatment, and finally the cleanliness of the entire clean workbench is comprehensively known).

[0104] A clean bench cleanliness test system is applied to the above-mentioned clean bench cleanliness test method, including a model training unit, a block screening unit to be tested, a difference item parameter extraction unit, a related block parameter correction unit 1, a related block parameter correction unit 2 and a cleanliness final evaluation unit, referring to Figure 2, obtain historical test record data of the target workbench through the model training unit, train the evolutionary causal model according to the historical test record data to obtain the trained evolutionary causal model; divide the target workbench into several regional blocks through the to-be-tested block screening unit to obtain a block set, detect a first test data set and a second test data set of the block set, predict a third test data set according to the first test data set and the trained evolutionary causal model, perform a difference comparison between the second test data set and the third test data set to screen out the to-be-tested block set from the block set; select the target block, the first associated block and the second associated block from the to-be-tested block set through the difference item parameter extraction unit. The target block and the detected parameter set belonging to the first associated block are compared and processed to obtain the first preprocessing item parameter, and the target block and the detected parameter set belonging to the second associated block are compared and processed to obtain the second preprocessing item parameter; the first preprocessing item parameter is evaluated for the transfer amount by the associated block parameter correction unit 1 to obtain the first transfer amount, and the first associated detection data is corrected for the corresponding transfer item parameter according to the first transfer amount to obtain the first associated corrected detection data; the second preprocessing item parameter is evaluated for the transfer amount by the associated block parameter correction unit 2 to obtain the second transfer amount, and the second associated detection data is corrected for the corresponding transfer item parameter according to the second transfer amount to obtain the second associated corrected detection data, and the target detection data is corrected for the corresponding transfer item parameter according to the first transfer amount and the second transfer amount to obtain the target corrected detection data; the corrected detection data set to which the block set to be checked belongs is counted by the cleanliness final evaluation unit according to the target corrected detection data, the first associated corrected detection data, and the second associated corrected detection data, and the corrected detection data set is finally evaluated for cleanliness to obtain the cleanliness test result.

[0105] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A clean bench cleanliness test method, characterized in that: The following steps are involved: Acquire historical test record data of the target workstation, and train the evolutionary causal model according to the historical test record data to obtain a trained evolutionary causal model; Dividing the target workbench into a number of regional blocks to obtain a block set, detecting a first detection data set and a second detection data set of the block set, predicting a third detection data set according to the first detection data set and the trained evolutionary causal model, and performing a difference comparison between the second detection data set and the third detection data set to screen out a block set to be verified from the block set; Selecting a target block, a first associated block, and a second associated block from the set of blocks to be checked, performing comparison processing on the target block and the detected parameter set to which the first associated block belongs to obtain a first preprocessing item parameter, and performing comparison processing on the target block and the detected parameter set to which the second associated block belongs to obtain a second preprocessing item parameter; Performing a transfer amount evaluation on the first preprocessing item parameter to obtain a first transfer amount, and performing correction processing on the first associated detection data corresponding to the transfer item parameter according to the first transfer amount to obtain first associated corrected detection data; Performing transfer amount evaluation on the second preprocessing item parameter to obtain a second transfer amount, performing correction processing on the second associated detection data according to the second transfer amount, and obtaining second associated corrected detection data; performing correction processing on the target detection data according to the first transfer amount and the second transfer amount, and obtaining target corrected detection data; According to the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, statistics are collected on the corrected detection data set to which the block set to be checked belongs, and a final cleanliness evaluation is performed on the corrected detection data set to obtain a cleanliness test result.

2. A clean bench cleanliness testing method according to claim 1, characterized in that: The steps of obtaining historical test record data of a target workstation and training an evolutionary causal model according to the historical test record data to obtain a trained evolutionary causal model are specifically as follows: Obtain historical test record data of cleanliness tests performed on a target workbench in historical periods, select various parameter test data of several historical periods from the historical test record data, respectively evaluate the correlation influence values ​​between various parameter test data of several historical periods to obtain historical correlation degree data, and output a historical correlation degree data set based on the historical correlation degree data; The historical correlation degree data set is input into the evolutionary causal model for training to obtain a trained evolutionary causal model.

3. A clean bench cleanliness testing method according to claim 1, characterized in that: The target workbench is divided into a number of regional blocks to obtain a block set, a first test data set and a second test data set of the block set are tested, a third test data set is predicted according to the first test data set and the trained evolutionary causal model, and the second test data set is compared with the third test data set to screen out the block set to be checked from the block set, specifically: The target workbench is divided into a number of area blocks to obtain a block set, and data detection of various key parameters is performed on each position area in the block set to obtain a first detection data set, and when each parameter in the first detection data set is detected, other parameters are under stable or ideal conditions; A second detection data set is obtained by performing data detection of various key parameters on each location area in the block set. When each parameter in the second detection data set is detected, other parameters are under non-ideal conditions; Inputting the second detection data set into the trained evolutionary causal model for testing to obtain a predicted correlation degree data set; predicting a third detection data set based on the predicted correlation degree data set under the effect of the predicted correlation degree data set on the first detection data set; Comparing the difference degree of each parameter of the second test data set and the third test data set to obtain a difference degree comparison result set; A difference determination threshold is preset, and blocks corresponding to comparison results having a value greater than the difference determination threshold are selected from the block set in the difference degree comparison result set, and a block set to be checked is output.

4. A clean bench cleanliness testing method according to claim 3, characterized in that: The steps of selecting a target block, a first associated block, and a second associated block from the set of blocks to be checked, comparing the detected parameter set of the target block and the first associated block to obtain a first preprocessing item parameter, and comparing the detected parameter set of the target block and the second associated block to obtain a second preprocessing item parameter are specifically as follows: Selecting a target block from the set of blocks to be checked, and selecting a first associated block and a second associated block adjacent to the target block from the set of blocks to be checked according to the target block, wherein the positions of the first associated block and the second associated block are not adjacent; Extracting corresponding target detection data, first associated detection data, and second associated detection data from the detection data set according to the target block, the first associated block, and the second associated block; Comparing the target detection data and the first associated detection data for differences in various parameters to obtain a first difference parameter set; Extracting transferable parameter item data from the first difference item parameter set to obtain first preprocessing item parameters; Comparing the target detection data with the second associated detection data for differences in various parameters to obtain a second difference parameter set; Transferable parameter item data is extracted from the second difference item parameter set to obtain second preprocessing item parameters.

5. A clean bench cleanliness testing method according to claim 4, characterized in that: The step of performing transfer amount evaluation on the first preprocessing item parameter to obtain a first transfer amount, and performing correction processing on the first associated detection data corresponding to the transfer item parameter according to the first transfer amount to obtain the first associated corrected detection data, specifically comprises: Taking the difference of two difference item parameters of the same category in the first preprocessing item parameters to obtain a first preprocessing difference value; According to the first preprocessing difference, parameter data of the same category as the first preprocessing difference is removed from the first associated detection data to obtain first screened associated detection data; Inputting the first preprocessing difference and the first screening association detection data into the trained evolutionary causal model for testing to obtain a first transfer association influence degree data set; evaluating a corresponding first transfer amount according to the first transfer association influence degree data set and the first preprocessing difference; According to the first transfer amount, the first associated detection data is corrected by performing a correction process on the corresponding transfer item parameter to obtain the first associated corrected detection data.

6. A clean bench cleanliness testing method according to claim 5, characterized in that: The steps of performing transfer amount evaluation on the second preprocessing item parameter to obtain a second transfer amount, performing correction processing on the corresponding transfer item parameter on the second associated detection data according to the second transfer amount to obtain second associated corrected detection data, and performing correction processing on the corresponding transfer item parameter on the target detection data according to the first transfer amount and the second transfer amount to obtain target corrected detection data are specifically as follows: Taking the difference of two difference item parameters of the same category in the second preprocessing item parameters to obtain a second preprocessing difference value; According to the second preprocessing difference, parameter data of the same category as the second preprocessing difference is removed from the second associated detection data to obtain second screened associated detection data; Inputting the second preprocessing difference and the second screening association detection data into the trained evolutionary causal model for testing to obtain a second transfer association influence degree data set; Evaluate a corresponding second transfer amount according to the second transfer association influence degree data set and the second preprocessing difference; According to the second transfer amount, the second associated detection data is corrected by a corresponding transfer item parameter to obtain second associated corrected detection data; According to the first transfer amount and the second transfer amount, the target detection data is corrected by performing correction processing on the corresponding transfer item parameters to obtain target corrected detection data.

7. A clean bench cleanliness testing method according to claim 6, characterized in that: The steps of counting the corrected detection data set to which the block set to be checked belongs according to the target corrected detection data, the first associated corrected detection data, and the second associated corrected detection data, and performing a final cleanliness evaluation on the corrected detection data set to obtain a cleanliness test result are specifically as follows: According to the target block, the first associated block and the second associated block, preprocessing the remaining blocks in the block set to be checked obtains a preprocessed remaining block set; According to the preprocessing remaining block set, a corresponding preprocessing remaining detection data set is screened out from the detection data set; According to the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, a data correction process is performed on the preprocessed remaining detection data set to obtain remaining corrected detection data; The target corrected detection data, the first associated corrected detection data, the second associated corrected detection data and the remaining corrected detection data are combined into a corrected detection data set; A final cleanliness evaluation is performed on each corrected detection data in the corrected detection data set to obtain a cleanliness test result.

8. A clean bench cleanliness test system, characterized in that: The system is used to implement a clean bench cleanliness testing method according to any one of claims 1 to 7, comprising: A model training unit is used to obtain historical test record data of a target workbench, and train an evolutionary causal model according to the historical test record data to obtain a trained evolutionary causal model; A block screening unit to be tested is used to divide the target workbench into a number of regional blocks to obtain a block set, detect a first test data set and a second test data set of the block set, predict a third test data set based on the first test data set and the trained evolutionary causal model, and compare the second test data set with the third test data set to screen out the block set to be tested from the block set; a difference item parameter extraction unit, configured to select a target block, a first associated block, and a second associated block from the block set to be checked, compare the detected parameter set of the target block and the first associated block to obtain a first preprocessing item parameter, and compare the detected parameter set of the target block and the second associated block to obtain a second preprocessing item parameter; The first correlation block parameter correction unit is used to perform a transfer amount evaluation on the first preprocessing item parameter to obtain a first transfer amount, and perform a correction process on the first correlation detection data corresponding to the transfer item parameter according to the first transfer amount to obtain the first correlation correction detection data; The second associated block parameter correction unit is used to perform transfer amount evaluation on the second preprocessing item parameter to obtain a second transfer amount, perform correction processing on the second associated detection data according to the second transfer amount, and obtain second associated corrected detection data; perform correction processing on the target detection data according to the first transfer amount and the second transfer amount, and obtain target corrected detection data; The cleanliness final evaluation unit is used to count the corrected detection data set to which the block set to be checked belongs according to the target corrected detection data, the first associated corrected detection data and the second associated corrected detection data, and to perform a final cleanliness evaluation on the corrected detection data set to obtain a cleanliness test result.

Images