Assembly quality control method, system and medium based on industrial Internet of Things

Abstract

The present invention provides an assembly quality control method, system and medium based on industrial Internet of Things. The method is executed by an industrial Internet of Things management platform of an assembly quality control system based on industrial Internet of Things. The method includes: obtaining quality inspection information of multiple parts to be assembled, and determining the assembly risk value of at least one assembly process through an assembly database based on the quality inspection information and assembly information; generating assembly parameters in response to the assembly risk value meeting the risk condition; obtaining assembly process data of the assembly operation, and determining the assembly quality of the completed process based on the assembly process data; determining subsequent assembly parameters based on the assembly quality in response to the assembly quality not meeting the quality requirements, and generating quality warning information based on the assembly quality and subsequent assembly parameters; generating quality update data based on the assembly quality. The possibility of unqualified assembly quality can be predicted, and an early warning or assembly parameter adjustment can be issued in time according to the situation of unqualified assembly quality.

Description

Technical Field

[0001] This specification relates to the field of assembly manufacturing, and in particular to an assembly quality control method, system and medium based on industrial Internet of Things. Background Art

[0002] In modern manufacturing, assembly lines are the key to achieving efficient and mass production of products. In the actual parts assembly process, due to factors such as fluctuations in parts quality, assembly quality is difficult to control. Existing assembly quality control methods are still insufficient in predicting and warning abnormal risks.

[0003] Therefore, it is hoped to provide an assembly quality control method, system and medium based on the Industrial Internet of Things, which can efficiently analyze and predict abnormal risks that may occur in the assembly process, and give early warnings or safety recommendations in a timely manner to achieve efficient and safe production of products. Summary of the invention

[0004] One or more embodiments of the present specification provide an assembly quality control method based on the Industrial Internet of Things, the method being executed by an Industrial Internet of Things management platform of an assembly quality control system based on the Industrial Internet of Things, the method comprising: obtaining quality inspection information of a plurality of parts to be assembled through quality inspection equipment of an Industrial Internet of Things perception control platform based on an Industrial Internet of Things sensor network platform; determining an assembly risk value of at least one assembly process through an assembly database based on the quality inspection information and assembly information, the assembly database being configured in a data center of the Industrial Internet of Things management platform; generating assembly parameters in response to the assembly risk value satisfying a risk condition; sending the assembly parameters to an operating device of the Industrial Internet of Things perception control platform, and controlling the operating device to perform an assembly operation based on the assembly parameters. Performing an assembly operation on multiple parts to be assembled; obtaining assembly process data of the assembly operation based on the monitoring equipment of the industrial Internet of Things perception control platform; determining the assembly quality of the completed process based on the assembly process data; in response to the assembly quality not meeting the quality requirements, determining subsequent assembly parameters based on the assembly quality, and sending the subsequent assembly parameters to the subsequent operation equipment of the industrial Internet of Things perception control platform to update the assembly parameters of the subsequent operation equipment; generating quality warning information based on the assembly quality and the subsequent assembly parameters, and sending the quality warning information to the industrial Internet of Things user platform based on the industrial Internet of Things service platform; and generating quality update data based on the completed process to update the assembly database.

[0005] One or more embodiments of the present specification provide an assembly quality control system based on the Industrial Internet of Things, the system comprising an Industrial Internet of Things user platform, an Industrial Internet of Things service platform, an Industrial Internet of Things management platform, an Industrial Internet of Things sensor network platform and an Industrial Internet of Things perception control platform, wherein the Industrial Internet of Things management platform is configured to: obtain quality inspection information of a plurality of parts to be assembled through the quality inspection equipment of the Industrial Internet of Things perception control platform based on the Industrial Internet of Things sensor network platform; determine an assembly risk value of at least one assembly process through an assembly database based on the quality inspection information and assembly information, the assembly database being configured in a data center of the Industrial Internet of Things management platform; generate assembly parameters in response to the assembly risk value satisfying a risk condition; and send the assembly parameters to the operating equipment of the Industrial Internet of Things perception control platform, Control the operating equipment to perform an assembly operation on the multiple parts to be assembled based on the assembly parameters; obtain assembly process data of the assembly operation based on the monitoring equipment of the Industrial Internet of Things perception control platform; determine the assembly quality of the completed process based on the assembly process data; in response to the assembly quality not meeting the quality requirements, determine subsequent assembly parameters based on the assembly quality, and send the subsequent assembly parameters to the subsequent operating equipment of the Industrial Internet of Things perception control platform to update the assembly parameters of the subsequent operating equipment; generate quality warning information based on the assembly quality and the subsequent assembly parameters, and send the quality warning information to the Industrial Internet of Things user platform based on the Industrial Internet of Things service platform; and generate quality update data based on the completed process to update the assembly database.

[0006] One or more embodiments of the present specification provide a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the method described in the above embodiments.

[0007] The beneficial effects brought about by one or more embodiments of the present specification include but are not limited to: (1) being able to predict the possibility of unqualified assembly quality, thereby determining assembly parameters with higher assembly quality, and timely issuing warnings or adjusting assembly parameters according to the situation of unqualified assembly quality, thereby achieving efficient and safe production of products; (2) by clustering, parts of the same type are divided according to quality inspection information to obtain multiple part subsets, and multiple candidate assembly groups are obtained based on the arrangement and combination of multiple part subsets of multiple different types, thereby reducing the amount of data and improving the efficiency of determining candidate assembly groups; (3) the estimated assembly quality is determined by the quality inspection information of multiple parts to be assembled in multiple candidate assembly groups, the assembly quality of each candidate assembly group can be estimated, and the assembly risk value is determined based on the estimated assembly quality of each candidate assembly group, thereby obtaining an assembly risk value with wider applicability and greater possibility, thereby accurately judging whether the assembly risk value meets the risk condition. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] This specification will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0009] Figure 1 It is a schematic diagram of the platform structure of an assembly quality control system based on the industrial Internet of Things according to some embodiments of this specification;

[0010] Figure 2 is an exemplary flow chart of an assembly quality control method based on the industrial Internet of Things according to some embodiments of this specification;

[0011] Figure 3 is an exemplary schematic diagram of determining an assembly risk value according to some embodiments of this specification;

[0012] Figure 4 is an exemplary schematic diagram of a matching model according to some embodiments of the present specification. DETAILED DESCRIPTION

[0013] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following briefly introduces the drawings required for describing the embodiments. The drawings do not represent all implementation methods.

[0014] It should be understood that the "system", "device", "unit" and / or "module" used herein are a method for distinguishing different components, elements, parts, portions or assemblies at different levels. If other words can achieve the same purpose, the words can be replaced by other expressions.

[0015] Unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0016] When the operations are performed according to the step description in the embodiments of this specification, unless otherwise specified, the order of the steps is interchangeable, the steps may be omitted, and other steps may be included in the operation process.

[0017] Figure 1 It is a schematic diagram of the platform structure of an assembly quality control system based on the industrial Internet of Things shown in some embodiments of this specification.

[0018] like Figure 1As shown, the assembly quality control system 100 based on the industrial Internet of Things includes an industrial Internet of Things user platform 110, an industrial Internet of Things service platform 120, an industrial Internet of Things management platform 130, an industrial Internet of Things sensor network platform 140 and an industrial Internet of Things perception control platform 150.

[0019] The industrial Internet of Things user platform refers to a platform for interacting with users. In some embodiments, the industrial Internet of Things user platform is configured as a terminal device and / or a terminal interface, etc. The terminal device includes a mobile device or a tablet computer, etc. The terminal interface includes a terminal device interface and a web page, etc.

[0020] The Industrial Internet of Things service platform refers to a platform that provides information and services to users. In some embodiments, the Industrial Internet of Things service platform is configured on a local server. The Industrial Internet of Things service platform can interact with the Industrial Internet of Things user platform and the Industrial Internet of Things management platform.

[0021] The industrial Internet of Things management platform refers to a platform for managing information and / or data related to the assembly quality control system 100 based on the industrial Internet of Things. In some embodiments, the industrial Internet of Things management platform includes a processor and a storage device, etc. The storage device stores a data center, and the data center includes a variety of databases. The database refers to a database management system that supports high concurrent access. The processor includes a central processing unit (CPU), an application-specific integrated circuit (ASIC), an application-specific instruction processor (ASIP), a graphics processing unit (GPU), etc. or any combination thereof.

[0022] The IIoT sensor network platform refers to a platform for integrated management of sensor information. In some embodiments, the IIoT sensor network platform is configured as a communication network or gateway and a network interface, etc. The IIoT sensor network platform can interact with the IIoT management platform and the IIoT perception control platform.

[0023] The industrial Internet of Things perception control platform refers to a functional platform for generating perception information and executing control information. In some embodiments, the industrial Internet of Things perception control platform includes operating equipment, quality inspection equipment, and monitoring equipment.

[0024] Work equipment refers to equipment used to complete assembly work, such as automatic assembly machines.

[0025] Quality inspection equipment refers to equipment used to obtain quality inspection information of parts to be assembled, such as a three-dimensional coordinate measuring machine or an electric profiler.

[0026] Parts to be assembled are parts that are used to assemble to obtain a finished product.

[0027] Monitoring equipment refers to equipment that obtains assembly process data. In some embodiments, the monitoring equipment includes temperature sensors, humidity sensors, RFID identification devices, barcode scanners, torque sensors, displacement sensors, and machine vision detection equipment.

[0028] RFID identification devices or barcode scanners are used to read the electronic tags of the parts to be assembled and obtain the quality inspection information of the parts to be assembled. Torque sensors are used to obtain the torque applied by the operating equipment during assembly operations. Displacement sensors are used to obtain the relative displacement between the parts to be assembled during assembly operations. Machine vision inspection equipment is used to obtain the dimensional data of finished or semi-finished products after assembly.

[0029] For detailed description of the above, please refer to Figures 2 to 4 Related description.

[0030] In some embodiments of the present specification, through the assembly quality control system 100 based on the industrial Internet of Things, an information operation closed loop can be formed between various functional platforms, and coordinated and regular operations can be performed to realize the informatization and intelligence of assembly quality control.

[0031] Figure 2 is an exemplary flow chart of an assembly quality control method based on the Industrial Internet of Things according to some embodiments of this specification. In some embodiments, process 200 is executed by the Industrial Internet of Things management platform 130 (hereinafter referred to as the management platform) of the assembly quality control system 100 based on the Industrial Internet of Things. Figure 2 As shown, process 200 includes the following steps:

[0032] Step 210, based on the industrial Internet of Things sensor network platform, quality inspection information of multiple parts to be assembled is obtained through quality inspection equipment of the industrial Internet of Things perception control platform.

[0033] Quality inspection information refers to data related to the quality inspection of parts to be assembled.

[0034] In some embodiments, the quality inspection information includes processing accuracy, surface roughness, etc. Processing accuracy refers to the difference between the actual size of the parts to be assembled and the design size. The design size refers to the theoretical size of the parts to be assembled.

[0035] The surface roughness is used to characterize the surface roughness of the parts to be assembled, and is expressed by numerical values, etc. The larger the numerical value, the higher the surface roughness. For obtaining the design dimensions, refer to step 220 and its related description.

[0036] In some embodiments, the management platform obtains quality inspection information through the quality inspection equipment of the industrial Internet of Things perception control platform based on the industrial Internet of Things sensor network platform.

[0037] The quality inspection information is obtained by measuring the parts to be assembled by the quality inspection equipment of the IIoT perception control platform. For example, the actual size of the parts to be assembled is measured by a three-dimensional coordinate measuring machine, and the difference between the actual size and the design size is calculated to obtain the processing accuracy. For another example, the surface roughness of the parts to be assembled is measured by an electric profilometer.

[0038] For a description of the quality inspection equipment and the parts to be assembled, see Figure 1 and its related description.

[0039] Step 220: Based on the quality inspection information and the assembly information, determine the assembly risk value of at least one assembly process through the assembly database.

[0040] Assembly information refers to information related to the assembly of parts to be assembled. In some embodiments, the assembly information includes at least one assembly process and the assembly object, the part to be assembled, the interference relationship, the assembly part and the connection relationship corresponding to each assembly process. Among them, the assembly object refers to two or more parts to be assembled that need to be assembled. The part to be assembled refers to the part to which the part to be assembled needs to be assembled. The assembly part refers to the part of the part to be assembled used for assembly. Exemplarily, the part to be assembled is a groove, and the assembly part refers to a bump on the part to be assembled.

[0041] The assembly process refers to a process in which the parts to be assembled are assembled to obtain the finished product.

[0042] Interference relationship refers to the interference relationship caused by the assembly order between assembly objects. For example, the machining accuracy of the parts to be assembled earlier in the assembly order is not enough, which makes it impossible to assemble the parts to be assembled later in the assembly order.

[0043] The connection relationship refers to the connection method used between assembly objects, for example, at least one of a threaded connection, a pin connection, etc.

[0044] In some embodiments, the management platform obtains assembly information or assembly drawings from the industrial Internet of Things user platform through the industrial Internet of Things service platform. The assembly drawings refer to design drawings that include data related to the assembly operation.

[0045] The assembly information or assembly drawings of the industrial Internet of Things user platform are obtained through user input, etc. Among them, the management platform extracts assembly information and / or design dimensions from the assembly drawings through image extraction, etc.

[0046] The assembly risk value refers to the data that characterizes the possibility of risks occurring in the assembly process of the parts to be assembled. The risks that occur include substandard assembly quality, etc.

[0047] It is understandable that assembling a finished product requires at least one assembly process, and the assembly risk value of each assembly process in the at least one assembly process needs to be determined.

[0048] In some embodiments, the management platform determines the assembly risk value of at least one assembly process through the assembly database based on the quality inspection information and the assembly information. For example, for each assembly object corresponding to the assembly process in the assembly information, the management platform counts the search range of the quality inspection information of the assembly object, searches the assembly database for historical assembly records whose historical quality inspection information is within the search range according to the search range, and calculates the ratio of the number of historical assembly records with unqualified historical assembly quality to the total number of retrieved historical assembly records, and uses the obtained ratio as the assembly risk value of the assembly process.

[0049] In some embodiments, the management platform counts the maximum and minimum values ​​in the quality inspection information of the assembly object, and uses the maximum value as the upper limit of the search range and the minimum value as the lower limit of the search range. The value of the quality inspection information is represented by the sum of the value of the processing accuracy and the value of the surface roughness.

[0050] The assembly database refers to a database storing historical assembly records. In some embodiments, the assembly database includes multiple historical assembly records. The historical assembly record refers to data related to a single assembly process performed in a historical time. The historical assembly record includes historical quality inspection information, historical assembly information, historical assembly quality, and historical assembly process data of the parts to be assembled used for a single assembly process performed in a single assembly. The historical assembly quality refers to the assembly quality after a single assembly process is performed in a single assembly and / or the assembly quality of the entire assembly operation. For an explanation of the assembly process data and assembly quality, see steps 250 and 260 and their related descriptions.

[0051] In some embodiments, the management platform stores the quality inspection information, assembly information, assembly quality, and assembly process data of the parts to be assembled used for a single assembly in each assembly process as a historical assembly record in the assembly database. By storing multiple historical assembly records, an assembly database is obtained.

[0052] In some embodiments, the assembly database is configured in a data center of the management platform. For a description of the data center, see Figure 1 and its related description.

[0053] In some embodiments, the management platform determines the assembly risk value based on multiple estimated assembly qualities of multiple candidate assembly groups. Figure 3 and its related description.

[0054] Step 230 , in response to the assembly risk value satisfying the risk condition, generating assembly parameters.

[0055] Risk conditions are conditions used to determine whether to generate assembly parameters.

[0056] In some embodiments, the risk condition includes an assembly risk value exceeding a risk threshold. One assembly process corresponds to one risk threshold. In response to an assembly risk value of an assembly process exceeding the corresponding risk threshold, the management platform generates assembly parameters for the assembly process. In response to an assembly risk value of an assembly process not exceeding the corresponding risk threshold, the assembly parameters of the assembly process continue to use the original assembly parameters. The original assembly parameters refer to the assembly parameters used by the assembly process in the historical assembly process.

[0057] In some embodiments, the risk threshold is pre-set based on the unqualified rate acceptable to the user, for example, the lower the unqualified rate acceptable to the user, the lower the risk threshold. The unqualified rate acceptable to the user is pre-set by the user based on production needs.

[0058] Assembly parameters refer to parameters used to guide assembly operations. In some embodiments, assembly parameters include assembly grouping parameters. Assembly grouping parameters refer to the grouping of parts to be assembled for assembly operations.

[0059] For an explanation of the assembly operation, see step 240 and its related description.

[0060] In some embodiments, the management platform determines assembly parameters in a variety of ways. For example, for a single assembly process, the management platform performs grading based on the quality inspection information of the parts to be assembled in the assembly object, divides the assembly objects of the same level into a group and then obtains multiple groups, and uses the group of multiple parts to be assembled in the assembly object as the assembly grouping parameter corresponding to the assembly process.

[0061] Taking the example of an assembly object including two parts to be assembled, the management platform divides each part to be assembled into three levels based on the size of the processing accuracy of the parts to be assembled (such as the third level for processing accuracy of 1.5-2mm, the second level for processing accuracy of 1-1.5mm, and the first level for processing accuracy of 0.5-1mm), and divides the two parts to be assembled of the same level into one group to obtain multiple groups, and uses the numbers of the multiple groups and the parts to be assembled contained in each group as assembly grouping parameters. The numbers of the parts to be assembled are used to distinguish the parts to be assembled. The numbers are automatically generated by the management platform based on the production sequence of the parts to be assembled.

[0062] In some embodiments, the number of levels of the management platform is related to the assembly risk value. The higher the assembly risk value, the more levels there are.

[0063] The management platform determines the assembly grouping parameters corresponding to each assembly process through the above method.

[0064] In some embodiments, the management platform determines the mating parts based on historical assembly records of multiple parts to be assembled, and generates assembly parameters based on the mating parts.

[0065] For an explanation of the historical assembly record, see step 220 and its related description.

[0066] Mating parts refer to the parts to be assembled that need to be assembled with each other in the assembly operation. The difference between mating parts and assembly objects is that assembly objects correspond to a single assembly process, while mating parts correspond to the entire assembly operation.

[0067] In some embodiments, the management platform determines the mating parts based on the historical assembly records of multiple parts to be assembled. For example, for two or more parts to be assembled that need to be assembled with each other in an assembly operation, the management platform uses the part to be assembled with the highest degree of assembly as the central part, and based on the central part, counts the range of reference matching tolerances between each part to be assembled and the central part based on the historical assembly records, and selects at least one part to be assembled whose matching tolerance with the central part is within the range of the reference matching tolerance from each part to be assembled, and randomly selects one of the at least one parts to be assembled as the mating part of the central part. For example, if the central part is assembled with an m-type part and an n-type part, one part to be assembled is selected from the m-type part and the n-type part as the mating part of the central part.

[0068] The assembly degree refers to the number of types of parts to be assembled that are assembled with a single part to be assembled. For example, in assembly process 1, type m parts are assembled with type s parts; in assembly process 2, type n parts are assembled with type s parts, then the assembly degree of type s parts is 2.

[0069] The fit tolerance refers to the size difference between the assembly part of one part to be assembled and the assembly part of another part to be assembled. For example, if the assembly part of one part to be assembled is 50mm and the assembly part of another part to be assembled is 49.5mm, then the fit tolerance of the two parts to be assembled is 0.5mm. The fit tolerance is a value with a positive or negative sign, which is used to indicate the relative size relationship between the assembly part and the assembly part.

[0070] The range of reference fit tolerance refers to the range of fit tolerance used for reference. In some embodiments, for each part to be assembled, the management platform counts the historical fit tolerances of the part to be assembled and the central part in the historical assembly records with qualified assembly quality based on the historical assembly records, and uses the maximum value of the historical fit tolerance as the upper limit of the range of reference fit tolerance, and uses the minimum value of the historical fit tolerance as the lower limit of the range of reference fit tolerance.

[0071] The management platform can determine the matching parts of multiple central parts in the above manner.

[0072] In some embodiments, the management platform divides the central part and the matching parts corresponding to multiple assembly processes into one group, obtains multiple groups after multiple divisions, and uses the groups of multiple parts to be assembled as assembly grouping parameters to obtain assembly parameters.

[0073] In some embodiments of the present specification, parts with similar tolerances are grouped together for assembly, so that a combination of parts to be assembled that are suitable for mutual assembly can be found among a large number of parts to be assembled, thereby improving the overall assembly quality.

[0074] In some embodiments, the assembly parameters include assembly sequence parameters. The management platform generates multiple candidate assembly sequences based on the assembly information, and determines the assembly parameters based on the estimated fit tolerance corresponding to each candidate assembly sequence. Figure 4 and its related description.

[0075] Step 240: Send the assembly parameters to the operating equipment of the industrial Internet of Things perception control platform, and control the operating equipment to perform assembly operations on multiple parts to be assembled based on the assembly parameters.

[0076] Assembly operation is the process of assembling the parts to be assembled into a finished product by the operating equipment. In some embodiments, the management platform sends assembly parameters to the operating equipment of the industrial Internet of Things perception control platform through the industrial Internet of Things sensor network platform, and the operating equipment performs assembly operations based on the assembly parameters. Among them, the management platform can assign a number to each assembly operation to distinguish multiple assembly operations.

[0077] For information on operating equipment, see Figure 1 and its related description.

[0078] Step 250, obtaining assembly process data of the assembly operation based on the monitoring equipment of the industrial Internet of Things perception control platform.

[0079] The assembly process data refers to data related to the process of the assembly operation. In some embodiments, the assembly process data includes at least one of environmental data, part information, process data, and result data.

[0080] Environmental data includes the temperature and humidity of the environment in which the assembly is performed. Part information includes the model and quality inspection information of the parts to be assembled. Process data includes the torque applied by the operating equipment and the relative displacement between the parts to be assembled. Result data includes the dimensional data of the finished or semi-finished product after assembly.

[0081] In some embodiments, the management platform obtains assembly process data of the assembly operation through the industrial Internet of Things sensor network platform and the monitoring device based on the industrial Internet of Things perception control platform. Figure 1 and its related description.

[0082] Step 260 , determining the assembly quality of the completed process based on the assembly process data.

[0083] A completed process refers to an assembly process that has been completed. In some embodiments, when the monitoring device detects that the assembly process has been completed, the information that the assembly process has been completed is sent to the industrial Internet of Things perception control platform. The industrial Internet of Things perception control platform sends the information that the assembly process has been completed to the management platform through the industrial Internet of Things sensor network platform. The management platform determines the completed process based on the information that the assembly process has been completed.

[0084] Assembly quality refers to data used to characterize the quality of assembly. In some embodiments, assembly quality can be represented by qualified and unqualified or by numerical values. For example, a numerical value of 1 represents qualified quality, and a numerical value of 0 represents unqualified quality.

[0085] In some embodiments, the management platform determines the assembly quality of the completed process in a variety of ways based on the assembly process data. For example, the management platform queries the quality table based on the environment data and the part information to determine the reference process data interval and the reference size interval corresponding to the environment data and the part information, and determines the assembly quality of the completed process based on the relative relationship between the process data and the reference process data interval and the relative relationship between the size data and the reference size interval.

[0086] In some embodiments, the quality table is pre-set by a technician based on historical assembly records. The quality table includes a combination of multiple environmental data and part information and a reference process data interval and a reference size interval corresponding to each combination. For example, a technician screens historical assembly records with qualified assembly quality under multiple sets of environmental data and part information in the historical assembly records, and for multiple historical assembly records under each set of environmental data and part information, counts the intervals of process data and the intervals of size data in the multiple historical assembly records, and uses the obtained intervals of process data and the intervals of size data as the reference process data interval and the reference size interval corresponding to the set of environmental data and part information.

[0087] The relative relationship is used to characterize the relative size of the process data and the reference process data interval and / or the relative size of the size data and the reference size interval. In some embodiments, the relative relationship includes the process data being within the reference process data interval, the process data being less than the lower limit of the reference process data interval, and the process data being greater than the upper limit of the reference process data interval.

[0088] In some embodiments, the relative relationship also includes the size data being within a reference size interval, the size data being smaller than a lower limit of the reference size interval, and the size data being larger than an upper limit of the reference size interval.

[0089] In some embodiments, if the relative relationship is that the process data is within the reference process data interval and the dimension data is within the reference dimension interval, the management platform determines that the assembly quality of the completed process is qualified; if the relative relationship is any other relative relationship, the management platform determines that the assembly quality of the completed process is unqualified.

[0090] Step 270, in response to the assembly quality not meeting the quality requirements, determining subsequent assembly parameters based on the assembly quality, and sending the subsequent assembly parameters to subsequent operation equipment of the industrial Internet of Things perception control platform to update the assembly parameters of the subsequent operation equipment.

[0091] In some embodiments, in response to the assembly quality meeting the quality requirement, the subsequent operation equipment continues to perform the assembly operation using the assembly parameters determined in step 230 .

[0092] The quality requirement refers to the overall quality requirement for the assembly operation. In some embodiments, the quality requirement includes that the proportion of unqualified assembly quality of completed processes to the total number of assembly processes is lower than a preset value. The preset value includes an acceptable unqualified rate preset by the user, etc.

[0093] Subsequent assembly parameters refer to the assembly parameters of subsequent operation equipment. Subsequent operation equipment refers to the operation equipment corresponding to the assembly process that has not been assembled after the completed process.

[0094] In some embodiments, the management platform determines subsequent assembly parameters based on assembly quality. For example, the management platform removes semi-finished products of completed processes with unqualified assembly quality through the industrial Internet of Things perception control platform, and determines subsequent assembly parameters based on the size data of the remaining semi-finished products (similar to the processing accuracy and surface roughness in the quality inspection information) in accordance with the method of generating assembly parameters in step 230. The industrial Internet of Things perception control platform includes a robotic arm for sorting finished products or semi-finished products.

[0095] In some embodiments, the management platform sends subsequent assembly parameters to subsequent operation equipment of the Industrial Internet of Things perception control platform through the Industrial Internet of Things sensor network platform to update the assembly parameters of the subsequent operation equipment.

[0096] Step 280: Generate quality warning information based on assembly quality and subsequent assembly parameters, and send the quality warning information to the industrial Internet of Things user platform based on the industrial Internet of Things service platform.

[0097] Quality warning information refers to information used to remind users that the assembly quality of an assembly process is unqualified. In some embodiments, the quality warning information includes at least one of a warning process, a semi-finished product to be discarded, and subsequent assembly parameters. A warning process refers to a completed process whose assembly quality is unqualified.

[0098] In some embodiments, the management platform generates quality warning information in a variety of ways based on assembly quality and subsequent assembly parameters. For example, based on assembly quality and subsequent assembly parameters, the management platform uses the completed process with unqualified assembly quality as the warning process, the warning process, the number of the semi-finished product to be removed, and the subsequent assembly parameters as quality warning information, and sends the quality warning information to the industrial Internet of Things user platform based on the industrial Internet of Things service platform. Among them, the number of the semi-finished product to be removed is the same as the number of the assembly operation corresponding to the semi-finished product.

[0099] In some embodiments, the user obtains quality warning information through the industrial Internet of Things user platform and determines the subsequent assembly parameters. The user determines the subsequent assembly parameters based on technical experience, and sends the determination information used to indicate the determination to the management platform through the industrial Internet of Things user platform and the industrial Internet of Things service platform.

[0100] Step 290 , based on the completed process, quality update data is generated to update the assembly database.

[0101] The quality update data refers to the historical assembly records of the completed process. In some embodiments, the management platform generates the quality update data based on the completed process. For example, the management platform constructs the historical assembly records based on the assembly process data of the completed process, the quality inspection information of the parts to be assembled, the assembly information, and the assembly quality, and stores the historical assembly records as the quality update data in the assembly database.

[0102] In some embodiments, the management platform may merge data in the assembly database, merging identical or similar historical assembly records, to avoid insufficient space in the assembly database due to continuous storage of quality update data.

[0103] For more information about the various platforms of the assembly quality control system based on the Industrial Internet of Things, please refer to Figure 1 The corresponding description.

[0104] In some embodiments of the present specification, an assembly quality control method based on the industrial Internet of Things can predict the possibility of unqualified assembly quality, thereby determining assembly parameters with higher assembly quality, and promptly issuing warnings or adjusting assembly parameters based on the situation of unqualified assembly quality, thereby achieving efficient and safe production of products.

[0105] Figure 3 is an exemplary flow chart of determining an assembly risk value according to some embodiments of the present specification.

[0106] Step 310 , generating a plurality of candidate assembly groups based on the assembly information.

[0107] For information on assembly, see Figure 2 and its related description.

[0108] The candidate assembly group refers to a combination of multiple parts to be assembled that are to be determined for assembling a finished product.

[0109] In some embodiments, the management platform generates multiple candidate assembly groups based on the assembly information in a variety of ways. For example, based on the preset types and the types of parts to be assembled in the assembly information, the management platform randomly selects a preset number of parts to be assembled from the parts to be assembled of the same type as the preset types for each type in the preset types, and arranges and combines the selected multiple types of parts to be assembled to obtain a candidate assembly group. The management platform generates multiple candidate assembly groups through the above method. Each type of part to be assembled can be pre-stored in the data center of the management platform.

[0110] The preset type refers to at least one type of parts to be assembled required to assemble a finished product. The preset quantity refers to the number of different types of parts to be assembled required to compose a finished product.

[0111] In some embodiments, the management platform obtains the preset type and preset quantity from the industrial Internet of Things user platform through the industrial Internet of Things service platform. The preset type includes multiple types. The preset quantity includes multiple data, each data corresponds to the required quantity of a type of parts to be assembled. In some embodiments, the user pre-sets the preset type and preset quantity through the industrial Internet of Things user platform.

[0112] In some embodiments, the management platform determines a parts set based on the types of multiple parts to be assembled; determines multiple parts subsets based on quality inspection information of multiple parts to be assembled in the parts set; and determines multiple candidate assembly groups based on the multiple parts subsets.

[0113] A parts collection refers to a collection of parts of the same type to be assembled.

[0114] In some embodiments, the management platform determines the parts set based on the types of the multiple parts to be assembled. For example, the management platform regards the multiple parts to be assembled of the same type as a parts set.

[0115] The part subset refers to a subset obtained by dividing the part set. In some embodiments, different part sets may have different numbers of subsets obtained by dividing.

[0116] In some embodiments, the management platform determines multiple parts subsets based on the quality inspection information of multiple parts to be assembled in the parts set. For example, the management platform clusters the multiple parts to be assembled in the parts set using a clustering algorithm to determine the parts subset based on the quality inspection information of the multiple parts to be assembled in the parts set. The clustering algorithm includes a K-means clustering algorithm, etc.

[0117] Exemplarily, the management platform randomly selects k parts to be assembled in the parts collection as initial clustering centers; constructs quality inspection feature vectors based on the quality inspection information of the parts to be assembled, each quality inspection feature vector corresponds to a part to be assembled; determines the vector distance between each part to be assembled and the initial clustering center based on multiple quality inspection feature vectors, and assigns the parts to be assembled to the cluster where the initial clustering center with the shortest vector distance is located; calculates the vector mean of multiple quality inspection feature vectors within the cluster where each initial clustering center is located, and uses it as the new clustering center; recalculates the vector distance between each part to be assembled and the new clustering center, and assigns the parts to be assembled to the cluster where the new clustering center with the shortest vector distance is located; repeats the above steps until the clustering center no longer changes, and obtains k clusters, each cluster being a part subset.

[0118] Wherein, k is the number of part subsets. In some embodiments, the number of multiple part subsets is related to the quality inspection information of multiple parts to be assembled in the part set. The management platform determines the k value based on the quality inspection information of multiple parts to be assembled in the part set. For example, the management platform determines k by the variance of the quality inspection information of multiple parts to be assembled in the part set. The larger the variance, the more dispersed the quality inspection information is, and the larger the value of k is.

[0119] In some embodiments, the number of multiple part subsets is also related to the matching parts of the parts to be assembled of the type corresponding to the part subset and the reference matching tolerance range of the matching parts. For example, the management platform is based on the number of part subsets determined above. If the reference matching tolerance range of the matching parts is narrower, the management platform will increase the number of corresponding part subsets by a larger margin. If the reference matching tolerance range of the matching parts is wider, the management platform will decrease the number of corresponding part subsets by a larger margin. For an explanation of the matching parts and the reference matching tolerance range, see step 230 and its related description.

[0120] If there are multiple mating parts of the type of parts to be assembled corresponding to the parts subset, the management platform counts the maximum upper limit and the minimum lower limit of the reference fit tolerance range of the multiple mating parts, takes the maximum upper limit and the minimum lower limit as the new reference fit tolerance range, and adjusts the number of corresponding parts subsets based on the new reference fit tolerance range.

[0121] In some embodiments of the present specification, the narrower the reference fit tolerance range of the mating parts, the higher the requirements for assembly, and the more detailed the division of the parts set needs to be. The larger the number of multiple parts subsets, the more conducive it is to subsequently obtain candidate assembly groups that better meet actual needs and ensure the accuracy of subsequent estimated assembly quality.

[0122] In some embodiments, the management platform generates a candidate assembly group based on a subset of parts. For example, based on the preset type and the type of parts to be assembled in the assembly information, the management platform randomly selects a preset number of parts to be assembled from multiple subsets of parts of the same type as the preset type for each type in the preset type, and arranges and combines the selected multiple types of multiple parts to be assembled to obtain a candidate assembly group. The management platform can generate multiple candidate assembly groups through the above method.

[0123] In some embodiments of the present specification, parts of the same type are divided according to quality inspection information through clustering to obtain multiple part subsets, and multiple candidate assembly groups are obtained based on the arrangement and combination of multiple part subsets of multiple different types, which can reduce the amount of data and improve the efficiency of determining candidate assembly groups.

[0124] Step 320 , for each of the plurality of candidate assembly groups, based on the quality inspection information of the plurality of to-be-assembled parts in each of the candidate assembly groups, determine the estimated assembly quality of each of the candidate assembly groups through the assembly database.

[0125] For information on quality inspection, assembly database, and assembly quality, see Figure 2 and its related description.

[0126] The estimated assembly quality refers to the estimated assembly quality of the finished product obtained by assembling based on the candidate assembly group.

[0127] In some embodiments, for each of the multiple candidate assembly groups, the management platform determines the estimated assembly quality of the candidate assembly group through the assembly database based on the quality inspection information of the multiple parts to be assembled in the candidate assembly group. For example, for each candidate assembly group, the management platform forms an accuracy vector based on the machining accuracy of the multiple parts to be assembled in the candidate assembly group, retrieves a complete assembly record that meets the first search condition in the assembly database based on the accuracy vector, and uses the assembly quality of the retrieved complete assembly record as the estimated assembly quality.

[0128] A complete assembly record refers to a combined record of historical assembly records corresponding to multiple historical assembly processes in a historical assembly operation. The assembly quality of a complete assembly record refers to the assembly quality of the finished product in the complete assembly record, including qualified or unqualified. For an explanation of historical assembly records, see Figure 2 and its related description.

[0129] In some embodiments, the management platform constructs a reference precision vector based on multiple historical machining precisions in the complete assembly record. The first search condition is pre-set based on historical experience. For example, the first search condition is the complete assembly record corresponding to the reference precision vector with the highest similarity to the precision vector.

[0130] In some embodiments, the management platform determines multiple reference assembly records through an assembly database based on quality inspection information of multiple parts to be assembled in each candidate assembly group, and determines the estimated assembly quality of each candidate assembly group based on the multiple reference assembly records.

[0131] The reference assembly record refers to a complete assembly record corresponding to the candidate assembly group and satisfying the second search condition.

[0132] The following is an example of a candidate assembly group.

[0133] In some embodiments, the management platform forms an accuracy vector based on the machining accuracy of multiple to-be-assembled parts in the candidate assembly group, retrieves multiple complete assembly records that meet the second search condition in the assembly database based on the accuracy vector, and uses the multiple complete assembly records as multiple reference assembly records. The second search condition is preset based on historical experience. For example, the second search condition is a complete assembly record corresponding to a reference accuracy vector whose similarity to the accuracy vector is greater than a preset similarity threshold.

[0134] In some embodiments, the management platform determines a preset similarity threshold based on the quality inspection information of the multiple parts to be assembled of each type in the candidate assembly group. For example, the management platform calculates the range of the quality inspection information of the parts to be assembled of each type based on the quality inspection information of the multiple parts to be assembled of each type in the candidate assembly group, and counts the mean or median of each range, and determines the preset similarity threshold based on the mean or median. The larger the mean or median, the smaller the preset similarity threshold.

[0135] In some embodiments, the management platform counts the assembly quality of each reference assembly record in the multiple reference assembly records based on the multiple reference assembly records corresponding to the candidate assembly group, and if the ratio of the number of reference assembly records with qualified assembly quality in the multiple reference assembly records to the total number of reference assembly records is greater than a preset evaluation value, the estimated assembly quality of the candidate assembly group is qualified; if the ratio of the number of reference assembly records with qualified assembly quality in the multiple reference assembly records to the total number of reference assembly records is not greater than the preset evaluation value, the estimated assembly quality of the candidate assembly group is unqualified. The preset evaluation value is preset based on historical experience.

[0136] In some embodiments of the present specification, the estimated assembly quality is determined by using multiple reference assembly records, which can take into account the dispersion of quality inspection information of the parts to be assembled, thereby more comprehensively estimating the assembly quality of the candidate assembly group.

[0137] In some embodiments, for each of a plurality of reference assembly records, the management platform obtains the processing accuracy of a plurality of historical assembly parts in each reference assembly record; based on the processing accuracy and the impact values ​​of the plurality of historical assembly parts on the assembly quality of the finished product, determines the matching degree of each reference assembly record with the corresponding candidate assembly group; based on the matching degree and the historical assembly quality in the plurality of reference assembly records, determines the estimated assembly quality of the candidate assembly group.

[0138] The following is an example of a reference assembly record.

[0139] The impact value refers to the data that characterizes the degree of influence of multiple historical assembly parts in the reference assembly record on the quality of the finished assembly. The larger the impact value, the greater the degree of influence of the historical assembly parts on the quality of the finished assembly. The quality of the finished assembly includes qualified or unqualified.

[0140] In some embodiments, for each reference assembly record, the management platform calculates the ratio between the machining accuracy of multiple historical assembly parts in the reference assembly record and the design size, and uses it as the first ratio. The machining accuracy of multiple historical assembly parts can be represented by the average or median value of the historical machining accuracy of multiple historical assembly parts. For an explanation of the design size, see Figure 2 and its related description.

[0141] In some embodiments, the management platform counts the average of the size data of the historical finished products corresponding to the processing accuracy of the multiple historical assembly parts in the historical assembly records, calculates the difference between the average of the size data and the middle value of the reference size interval, and uses the ratio of the obtained difference to the middle value of the reference size interval as the second ratio. The reference size interval here corresponds to the reference assembly record and is determined by the technician in the manner of determining the reference size interval in step 260.

[0142] In some embodiments, the management platform uses the ratio of the first ratio to the second ratio as an influence value of the plurality of historical assembly parts in the reference assembly record on the assembly quality of the finished product.

[0143] Matching degree refers to the data that characterizes the degree of matching between the reference assembly record and the candidate assembly group.

[0144] In some embodiments, the management platform determines the matching degree between the reference assembly record and the corresponding candidate assembly group in a variety of ways based on the processing accuracy and the impact values ​​of multiple historical assembly parts on the finished product assembly quality. For example, the management platform calculates the ratio of the impact value to the processing accuracy and uses it as the matching degree between the reference assembly record and the corresponding candidate assembly group. For another example, the management platform determines the matching degree between the reference assembly record and the corresponding candidate assembly group based on the corresponding relationship between the impact value, processing accuracy and matching degree. The corresponding relationship includes that the smaller the processing accuracy, the greater the impact value, and the higher the matching degree.

[0145] In some embodiments, the management platform determines the estimated assembly quality of the candidate assembly group based on the matching degree between each reference assembly record and the corresponding candidate assembly group and the historical assembly quality in the multiple reference assembly records. For example, the management platform selects the reference assembly record with the highest matching degree among the multiple reference assembly records, and uses the historical assembly quality of the reference assembly record as the estimated assembly quality of the candidate assembly group.

[0146] In some embodiments of the present specification, the estimated assembly quality is determined by the matching degree, which can take into account the matching degree between the reference assembly record and the candidate assembly group, is closer to the actual assembly situation, and obtains a more accurate estimated assembly quality.

[0147] Step 330 , determining an assembly risk value based on a plurality of estimated assembly qualities of a plurality of candidate assembly groups.

[0148] In some embodiments, the management platform can determine the assembly risk value based on the estimated assembly quality. For example, the management platform uses the ratio of the number of candidate assembly groups with unqualified estimated assembly quality to the total number of candidate assembly groups as the assembly risk value. The assembly risk value here corresponds to the entire assembly operation, and accordingly, the risk threshold corresponding to the assembly risk value determined by this method is set separately.

[0149] In some embodiments of the present specification, the estimated assembly quality is determined by quality inspection information of multiple parts to be assembled in multiple candidate assembly groups, the assembly quality of each candidate assembly group can be estimated, and the assembly risk value is determined based on the estimated assembly quality of each candidate assembly group, so as to obtain an assembly risk value with wider applicability and greater possibility, and then accurately judge whether the assembly risk value meets the risk condition.

[0150] It should be noted that the above description of the process 200 and the process 300 is only for example and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0151] Figure 4 is an exemplary schematic diagram of a matching model according to some embodiments of the present specification.

[0152] In some embodiments, the assembly parameters include assembly sequence parameters, and the management platform generates multiple candidate assembly sequences based on the assembly information; constructs an assembly diagram 430 corresponding to each candidate assembly sequence 420 based on multiple part sets (for example, part set 410-1, ..., part set 410-n), multiple part subsets (for example, multiple part subsets corresponding to part set 410-1, ..., multiple part subsets corresponding to part set 410-n), and each candidate assembly sequence 420 in the multiple candidate assembly sequences; determines an estimated fitting tolerance 450 corresponding to each candidate assembly sequence 420 through a fitting model 440 based on the assembly diagram 430 corresponding to each candidate assembly sequence 420; and determines assembly parameters based on the estimated fitting tolerance corresponding to each candidate assembly sequence.

[0153] For a description of assembly information, assembly parameters, parts collections, and parts sub-collections, see Figure 2 and Figure 3 and its related description.

[0154] The candidate assembly sequence refers to the relevant data to be determined as the assembly sequence parameters. The assembly sequence parameters refer to the relevant data about the assembly sequence of the parts to be assembled.

[0155] In some embodiments, the management platform generates multiple candidate assembly sequences based on the assembly information in multiple ways. For example, the management platform randomly sorts multiple parts to be assembled multiple times to obtain multiple candidate assembly sequences, and screens the multiple candidate assembly sequences based on the interference relationship in the assembly information to exclude candidate assembly sequences with conflicting interference relationships.

[0156] The candidate assembly sequence with a conflicting interference relationship refers to a candidate assembly sequence that cannot be realized due to the interference relationship between the parts to be assembled. For example, if the interference relationship between the parts to be assembled A and B is such that if the assembly sequence of A is before B, B cannot be assembled, then the candidate assembly sequence including the assembly sequence of A before B is a candidate assembly sequence with a conflicting interference relationship.

[0157] Assembly graph 430 refers to a graph structure that represents the matching relationship between multiple part sets (or parts to be assembled). The graph structure is a data structure composed of nodes (such as node 431) and edges (such as edge 432). The edges connect the nodes, and the nodes and edges can have features. For an explanation of the matching relationship, see Figure 2 and its related description.

[0158] In some embodiments, the management platform may construct an assembly graph corresponding to each candidate assembly sequence based on multiple part subsets and each candidate assembly sequence in multiple candidate assembly sequences.

[0159] The nodes of the assembly graph include different types of parts collections (or parts to be assembled). The node features are the relevant data of the parts sub-collection corresponding to the parts collection and the assembly degree of the parts to be assembled in the parts collection. The relevant data includes the number of parts sub-collections, the maximum and minimum values ​​of the quality inspection information in each parts sub-collection, etc. The relevant data is obtained by the management platform through statistics of the parts sub-collections.

[0160] For details on the degree of fit, see Figure 2 and its related description.

[0161] There is an edge between the parts sets with matching relationship in the assembly diagram. The characteristics of the edge include the order of the parts sets at both ends of the edge in the candidate assembly sequence.

[0162] It is understandable that all parts in a parts set are of the same type of parts to be assembled, so a parts set can be used in an assembly diagram to represent one type of parts to be assembled.

[0163] The assembly graphs corresponding to different candidate assembly orders differ only in the features of the edges.

[0164] The estimated fit tolerance refers to the estimated fit tolerance between two parts to be assembled in the candidate assembly sequence. In some embodiments, the estimated fit tolerance includes multiple data, each of which corresponds to an edge in the assembly diagram. For more information about fit tolerance, see Figure 2 and its related description.

[0165] The fit model refers to a model used to determine the estimated fit tolerance. The fit model may be a machine learning model. For example, the fit model may include any one or a combination of a Graph Neural Network (GNN) model or other custom model structures.

[0166] In some embodiments, an edge output of the fit model represents an estimated fit tolerance between two nodes connected by the edge.

[0167] In some embodiments, the management platform can train the matching model based on a large number of labeled training samples by gradient descent method, etc. The training samples include sample assembly graphs, and the labels include actual matching tolerances corresponding to the sample assembly graphs.

[0168] In some embodiments, training samples and labels are obtained based on complete assembly records. For example, the management platform constructs a sample assembly graph based on the historical parts set and historical assembly sequence in the complete assembly record. The nodes and their features, and the edges and their features of the sample assembly graph are similar to the above description. For a description of the complete assembly record, see Figure 3 and its related description.

[0169] For another example, the management platform screens multiple complete assembly records with the same historical assembly sequence in the complete assembly record, counts the number of reworks and the number of qualified products per unit time under the corresponding historical fit tolerance for each complete assembly record in the multiple complete assembly records, selects complete assembly records whose difference between the number of qualified products and the number of reworks exceeds the qualified number requirement, counts the two historical assembly parts corresponding to the two sample nodes in such complete assembly records, calculates the maximum and minimum values ​​of the historical fit tolerances of the two historical assembly parts, and uses the range composed of the maximum and minimum values ​​as the actual fit tolerance of the edges corresponding to the two sample nodes in the sample assembly graph. A sample node refers to a node in the sample graph. The actual fit tolerance corresponding to each edge in the sample assembly graph is determined in the above manner.

[0170] The number of qualified products refers to the number of parts to be assembled that do not need to be reworked per unit time. The unit time and qualified number requirements are preset based on historical experience. The rework quantity refers to the number of semi-finished products rejected in step 270.

[0171] In some embodiments, the input of the assembly model also includes assembly environment information.

[0172] Assembly environment information refers to information related to the environment of the assembly operation, such as temperature or humidity. For instructions on assembly operations, see Figure 2 and its related description.

[0173] In some embodiments, the management platform obtains assembly environment information from the temperature sensor and / or humidity sensor of the industrial Internet of Things perception control platform through the industrial Internet of Things sensor network platform.

[0174] If the input of the assembly model also includes assembly environment information, the training sample also includes sample assembly environment information. The method of obtaining the sample assembly environment information is similar to the method of obtaining the training sample.

[0175] In some embodiments of the present specification, different assembly environments may cause deformation of the parts to be assembled. By inputting the assembly environment information into the assembly model, an estimated fitting tolerance that is more in line with the actual situation can be obtained.

[0176] In some embodiments, the matching model is trained by: inputting a plurality of labeled training samples into the initial matching model, constructing a loss function through the labels and the prediction results of the initial matching model, updating the initial matching model based on the iteration of the loss function, and completing the matching model training when the loss function of the initial matching model satisfies a preset condition. The preset condition may be that the loss function converges, the number of iterations reaches a set value, etc.

[0177] In some embodiments, the fit model is also incrementally trained based on assembly quality. The incremental training process includes: in response to the assembly quality meeting the preset training condition, determining an incremental training set based on the historical assembly records corresponding to the assembly quality; and incrementally updating the fit model based on the incremental training set. For an explanation of assembly quality, see Figure 2 and its related description.

[0178] The preset training condition refers to the condition for determining whether to perform incremental training. In some embodiments, the preset training condition includes that the proportion of the number of complete assembly records with unqualified assembly quality in the total number of complete assembly records in the preset period exceeds the preset proportion. The preset proportion is pre-set based on the assembly risk value, and the higher the assembly risk value, the higher the preset proportion. The preset period refers to the period from the last time the incremental training set was determined to the current moment.

[0179] In some embodiments, the management platform determines new training data and labels based on historical assembly records within a preset period of time according to the above-mentioned method for determining training samples and labels, thereby forming an incremental training set.

[0180] In some embodiments, the management platform incrementally updates the coordination model based on the incremental training set. During the incremental update, the learning rate of the coordination model is lower than the learning rate of the coordination model before the incremental update. The incremental update refers to retraining the coordination model based on the incremental training set.

[0181] In some embodiments of this specification, the accuracy of the assembly model can be effectively improved through incremental training. At the same time, since incremental training is performed when the assembly quality is not up to standard, reducing the learning rate of the matching model during incremental update can avoid missing the best matching model parameters.

[0182] In some embodiments, the management platform determines assembly parameters based on the estimated fit tolerance corresponding to each candidate assembly sequence. For example, for each candidate assembly sequence, the management platform uses the maximum value among the multiple estimated fit tolerances as the upper limit of the range of the estimated fit tolerance, and uses the minimum value among the multiple estimated fit tolerances as the lower limit of the range of the estimated fit tolerance based on the multiple estimated fit tolerances corresponding to the candidate assembly sequence, and the management platform selects the candidate assembly sequence with the widest range of the estimated fit tolerance as the assembly sequence parameter. The widest range means that the difference between the upper limit and the lower limit of the range of the estimated fit tolerance is the largest.

[0183] In some embodiments of the present specification, if the fitting tolerance is too small, it may cause assembly difficulties, which may lead to easy errors or low assembly efficiency. Therefore, by determining the best estimated fitting tolerance through the assembly model, it is possible to ensure assembly efficiency and reduce assembly risk value.

[0184] Some embodiments of the present specification also provide a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes any of the methods in the above embodiments.

[0185] Furthermore, certain features, structures or characteristics in one or more embodiments of this specification may be appropriately combined.

[0186] In some embodiments, numbers describing the number of components and attributes are used. It should be understood that such numbers used in the description of the embodiments are modified by the modifiers "about", "approximately" or "substantially" in some examples. Unless otherwise specified, "about", "approximately" or "substantially" indicate that the numbers are allowed to vary by ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximate values, which may change according to the required features of individual embodiments. In some embodiments, the numerical parameters should take into account the specified significant digits and adopt the general method of retaining the digits. Although the numerical domains and parameters used to confirm the breadth of the range in some embodiments of this specification are approximate values, in specific embodiments, the setting of such numerical values ​​is as accurate as possible within the feasible range.

[0187] If there is any inconsistency or conflict between the descriptions, definitions, and / or usage of terms in the materials cited in this specification and the contents described in this specification, the descriptions, definitions, and / or usage of terms in this specification shall prevail.

Claims

1. An assembly quality control method based on industrial Internet of Things, characterized in that: The method is performed by an industrial Internet of Things management platform of an assembly quality control system based on the industrial Internet of Things, and the method includes: Based on the IIoT sensor network platform, the quality inspection equipment of the IIoT perception control platform is used to obtain the quality inspection information of multiple parts to be assembled; Based on the quality inspection information and the assembly information, determining an assembly risk value of at least one assembly process through an assembly database, wherein the assembly database is configured in a data center of the industrial Internet of Things management platform; In response to the assembly risk value satisfying the risk condition, generating assembly parameters, including: The parts to be assembled with the highest assembly degree are taken as the central parts, and the range of reference matching tolerances between each part to be assembled and the central part is counted based on historical assembly records, and at least one part to be assembled whose matching tolerance with the central part is within the range of the reference matching tolerance is selected from each part to be assembled, and one of the at least one part to be assembled is randomly selected as the matching part of the central part, wherein the assembly degree refers to the number of types of parts to be assembled that are assembled with a single part to be assembled; and generating the assembly parameters based on the mating parts; The assembly parameters include assembly sequence parameters, and the method further includes: Determining a parts set based on types of the multiple parts to be assembled; Determining a plurality of part subsets based on the quality inspection information of the plurality of parts to be assembled in the part set; Based on the assembly information, generating a plurality of candidate assembly orders; Based on the plurality of part sets, the plurality of part subsets and each candidate assembly sequence in the plurality of candidate assembly sequences, construct an assembly graph corresponding to each candidate assembly sequence; Based on the assembly diagram corresponding to each candidate assembly sequence, determining the estimated fitting tolerance corresponding to each candidate assembly sequence through a fitting model; Determining the assembly parameters based on the estimated fit tolerance corresponding to each candidate assembly sequence; Sending the assembly parameters to the operating equipment of the industrial Internet of Things perception control platform, and controlling the operating equipment to perform assembly operations on the multiple parts to be assembled based on the assembly parameters; Acquiring assembly process data of the assembly operation based on the monitoring equipment of the industrial Internet of Things perception control platform; Based on the assembly process data, determining the assembly quality of the completed process; In response to the assembly quality not meeting the quality requirement, determining subsequent assembly parameters based on the assembly quality, and sending the subsequent assembly parameters to subsequent operation equipment of the industrial Internet of Things perception control platform to update the assembly parameters of the subsequent operation equipment; Based on the assembly quality and the subsequent assembly parameters, generate quality warning information, and send the quality warning information to the industrial Internet of Things user platform based on the industrial Internet of Things service platform; and Based on the completed process, quality update data is generated to update the assembly database.

2. The method according to claim 1, characterized in that The step of determining the assembly risk value of at least one assembly process through an assembly database based on the quality inspection information and the assembly information comprises: Based on the assembly information, generating a plurality of candidate assembly groups; For each of the plurality of candidate assembly groups, based on the quality inspection information of the plurality of to-be-assembled parts in the each candidate assembly group, determining the estimated assembly quality of the each candidate assembly group through the assembly database; and The assembly risk value is determined based on a plurality of the estimated assembly qualities of the plurality of candidate assembly groups.

3. The method according to claim 2, characterized in that The number of the plurality of part subsets is related to the quality inspection information of the plurality of parts to be assembled in the part set, the quality inspection information including machining accuracy and surface roughness, and the generating of a plurality of candidate assembly groups based on the assembly information includes: Based on the plurality of subsets of parts, the plurality of candidate assembly groups are determined.

4. The method according to claim 2, characterized in that For each of the plurality of candidate assembly groups, determining the estimated assembly quality of each candidate assembly group through the assembly database based on the quality inspection information of the plurality of to-be-assembled parts in each candidate assembly group includes: Based on the quality inspection information of the plurality of parts to be assembled in each candidate assembly group, determining a plurality of reference assembly records through the assembly database; and The estimated assembly quality of each candidate assembly group is determined based on the plurality of reference assembly records.

5. An assembly quality control system based on industrial Internet of Things, characterized in that: The system includes an industrial Internet of Things user platform, an industrial Internet of Things service platform, an industrial Internet of Things management platform, an industrial Internet of Things sensor network platform and an industrial Internet of Things perception control platform. The industrial Internet of Things management platform is configured as follows: Based on the industrial Internet of Things sensor network platform, quality inspection information of multiple parts to be assembled is obtained through the quality inspection equipment of the industrial Internet of Things perception control platform; Based on the quality inspection information and the assembly information, determining an assembly risk value of at least one assembly process through an assembly database, wherein the assembly database is configured in a data center of the industrial Internet of Things management platform; In response to the assembly risk value satisfying a risk condition, generating assembly parameters, including: The parts to be assembled with the highest assembly degree are taken as the central parts, and the range of reference matching tolerances between each part to be assembled and the central part is counted based on historical assembly records, and at least one part to be assembled whose matching tolerance with the central part is within the range of the reference matching tolerance is selected from each part to be assembled, and one of the at least one part to be assembled is randomly selected as the matching part of the central part, wherein the assembly degree refers to the number of types of parts to be assembled that are assembled with a single part to be assembled; and generating the assembly parameters based on the mating parts; The assembly parameters include assembly sequence parameters, and the industrial Internet of Things management platform is further configured as follows: Determining a parts set based on types of the multiple parts to be assembled; Determining a plurality of part subsets based on the quality inspection information of the plurality of parts to be assembled in the part set; Based on the assembly information, generating a plurality of candidate assembly orders; Based on the plurality of part sets, the plurality of part subsets and each candidate assembly sequence in the plurality of candidate assembly sequences, construct an assembly graph corresponding to each candidate assembly sequence; Based on the assembly diagram corresponding to each candidate assembly sequence, determining the estimated fitting tolerance corresponding to each candidate assembly sequence through a fitting model; Determining the assembly parameters based on the estimated fit tolerance corresponding to each candidate assembly sequence; Sending the assembly parameters to the operating equipment of the industrial Internet of Things perception control platform, and controlling the operating equipment to perform assembly operations on the multiple parts to be assembled based on the assembly parameters; Acquiring assembly process data of the assembly operation based on the monitoring equipment of the industrial Internet of Things perception control platform; Based on the assembly process data, determining the assembly quality of the completed process; In response to the assembly quality not meeting the quality requirement, determining subsequent assembly parameters based on the assembly quality, and sending the subsequent assembly parameters to subsequent operation equipment of the industrial Internet of Things perception control platform to update the assembly parameters of the subsequent operation equipment; as well as Based on the assembly quality and the subsequent assembly parameters, quality warning information is generated, and the quality warning information is sent to the industrial Internet of Things user platform based on the industrial Internet of Things service platform; as well as Based on the completed process, quality update data is generated to update the assembly database.

6. The system according to claim 5, characterized in that The industrial Internet of Things management platform is further configured to: Based on the assembly information, generating a plurality of candidate assembly groups; For each of the plurality of candidate assembly groups, based on the quality inspection information of the plurality of to-be-assembled parts in each of the candidate assembly groups, determining the estimated assembly quality of each of the candidate assembly groups through the assembly database; as well as The assembly risk value is determined based on a plurality of the estimated assembly qualities of the plurality of candidate assembly groups.

7. The system according to claim 6, characterized in that The industrial Internet of Things management platform is further configured to: Based on the quality inspection information of the plurality of parts to be assembled in each candidate assembly group, determining a plurality of reference assembly records through the assembly database; as well as The estimated assembly quality of each candidate assembly group is determined based on the plurality of reference assembly records.

8. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions. When a computer reads the computer instructions in the storage medium, the computer executes the method as claimed in claim 1.

Images