Intelligent processing system and method for whole process of research and development experiment based on Internet of Things

Abstract

The invention relates to the technical field of industrial intelligent platforms, in particular to a research and development experiment whole-process intelligent processing system and method based on the Internet of Things, and the system comprises a collection module which collects original data in each link; the Internet of Things platform is used for classifying and storing the original data; the local knowledge base is used for acquiring original data and retrieving associated local data; the intelligent processing module performs comparative analysis according to the original data and associated local data to generate a guidance strategy and issue the guidance strategy for process guidance, and is also used for obtaining a feedback result; the intelligent processing module takes the adjusted data as whole process data based on the feedback result, and generates a design assistance strategy according to the feedback result; and the design module is used for uploading design object information, acquiring corresponding whole process data to formulate basic parameters and acquiring a design assistance strategy to assist in comparison. The method has the effect of realizing data connection and association in the whole manufacturing link so as to provide a guide strategy for each link.

Description

Technical Field

[0001] This application relates to the technical field of industrial intelligent platforms, and in particular, to an intelligent processing system and method for the whole process of R & D experiments based on the Internet of Things. Background Art

[0002] In the manufacturing field, for the entire manufacturing process of products such as vacuum pumps, a number of corresponding data need to be collected and recorded, and whether the vacuum pump manufacturing is compliant and whether the product is qualified are analyzed through the data of each link.

[0003] However, the following deficiencies exist in the prior art: Most of the data are paper-based data, which are not convenient for storage and filing, and are also not convenient for statistical analysis; The data in each link of the manufacturing process are all point-like distributed, and cannot be correlated with each other, which is not convenient for subsequent data analysis and difficult to propose guiding strategies for the entire design; The data in each link are only used for verification and traceability in the current stage and to detect whether they meet the design requirements, but it is difficult to associate and combine the data with other links to guide the design source and propose optimization strategies. Summary of the Invention

[0004] In order to achieve data connection and association in the whole manufacturing process to propose guiding strategies for each link, this application provides an intelligent processing system and method for the whole process of R & D experiments based on the Internet of Things.

[0005] In the first aspect, this application provides an intelligent processing system for the whole process of R & D experiments based on the Internet of Things, adopting the following technical solutions: An intelligent processing system for the whole process of R & D experiments based on the Internet of Things, comprising: An acquisition module, configured to acquire the original data in each link and upload it to the Internet of Things platform; The Internet of Things platform, configured to classify a number of the original data and store them in the data sets corresponding to different links; A local knowledge base, configured to obtain the original data in each of the data sets, retrieve the associated local data, and package and send it to the intelligent processing module; An intelligent processing module, configured to perform comparison and analysis based on the original data and the associated local data to generate a guiding strategy and send it to each link for process guidance, and is also configured to obtain the feedback results generated after adjustment based on the guiding strategy in each link; The intelligent processing module uploads the adjusted data as the whole-process data to the local knowledge base based on the feedback results, and the intelligent processing module also generates a design assistance strategy including a number of design key abnormal items according to the feedback results; The design module uploads the information of the design object, obtains the corresponding whole-process data in the local knowledge base to formulate the basic parameters of the design process, and docks with the intelligent processing module to obtain the design assistance strategy to assist in comparing the design process.

[0006] In some of the embodiments, the link includes a quality inspection link, in which: Obtain the collected quality inspection data and upload it to the Internet of Things platform; The local knowledge base retrieves the local stored quality inspection list based on the quality inspection object in the quality inspection data; Fill the quality inspection data into the corresponding quality inspection list and send it to the intelligent processing module. The intelligent processing module verifies each quality inspection data based on the specification requirements in the quality inspection list; Generate risk items based on the verification results, and the verification results at least include out-of-tolerance and data missing.

[0007] In some of the embodiments, the intelligent processing module is further configured to retrieve the corresponding processing opinions and cause analysis in the local knowledge base according to the content, quantity and combination correlation coefficient of the risk items; Wherein, the processing opinions include measures, priorities, responsible objects and completion time limits; The cause analysis includes abnormal events and corresponding compliance probabilities, and the compliance probabilities are obtained based on the following steps: Select the corresponding numbered cause list in the local knowledge base based on the verification results; Generate the basic probabilities of each abnormal event in the cause list based on the content of the risk items; Generate a combined risk set based on the quantity and content of the risk items, calculate the combined correlation coefficient based on the combined risk set and the empirical algorithm, and adjust the basic probability based on the combined correlation coefficient to obtain the compliance probability.

[0008] In some of the embodiments, the link includes an assembly link, in which: The intelligent processing module obtains assembly data in the Internet of Things platform and obtains the matching selection rule document in the local knowledge base; The intelligent processing module combines and arranges several assembly objects with assembly relationships, and calculates the corresponding assembly parameters for each combination based on the selection rule document; Pull the assembly standard based on the selection rule document, classify each assembly parameter according to the assembly standard and issue the corresponding level of evaluation indicators, and allocate each combination to the corresponding application environment according to each evaluation indicator.

[0009] In some of these embodiments, the link includes a testing link, in which: The intelligent processing module obtains test data in the Internet of Things platform and obtains test items matched by the test object in the local knowledge base; The test items are sent down for testing, and the change curves of the test data under different test items are obtained in real time; The test data with abnormal change curves and the test items are packaged to obtain key test items.

[0010] In some of these embodiments, the intelligent processing module is further configured to: Take the risk item, the handling opinion, and the cause analysis as the first guiding strategy, and send them down to the responsible object in the corresponding quality inspection link based on the priority for process guidance; Take the evaluation index as the second guiding strategy, and send it down to the assembly link in the corresponding application environment for assembly guidance; Take each of the key test items as the third guiding strategy, and send it down to the corresponding test object in the test link for test guidance.

[0011] In some of these embodiments, the intelligent processing module is further configured to respectively obtain the feedback results in the quality inspection link, the assembly link, and the test link, and judge whether the guiding strategies corresponding to each link are effective based on the feedback results; If it is effective, the intelligent processing module sends the adjusted optimization data to the local knowledge base for storage and adds a project case number; If it is not effective, the intelligent processing module marks the local data in the local knowledge base and re-selects the local data to generate a new guiding strategy until the feedback result is effective; When the project case number corresponds to the whole process link, the local knowledge base takes all the data in the same project case number as the whole process data.

[0012] In some of these embodiments, the design key abnormal items include the data adjusted based on the guiding strategy with the feedback result being effective and the data corresponding to the feedback result being ineffective.

[0013] In some of these embodiments, the intelligent processing module takes the design key abnormal items as key features and selects the corresponding quality inspection data and / or assembly data and / or test data as key data; Obtain a preset offset in the local knowledge base based on the key features; Calculate the influence parameter matrix corresponding to each of the key features based on the key data and the preset bias. The influence parameter matrix contains a number of influence parameters used to characterize the magnitude of the influence of the key data corresponding to each of the key features on the target result in the design process; Generate the design assistance strategy based on the influence parameters.

[0014] In a second aspect, the present application provides an intelligent processing method for the entire process of R & D experiments based on the Internet of Things, adopting the following technical solutions: An intelligent processing method for the entire process of R & D experiments based on the Internet of Things includes the following steps: Collect the original data in each link; Classify a number of the original data and store them in the data sets corresponding to different links; Obtain the original data in each of the data sets and retrieve the associated local data; Perform comparison and analysis based on the original data and the associated local data to generate a guidance strategy and issue it to each link for process guidance. At the same time, obtain the feedback results generated after adjustment based on the guidance strategy in each link; Based on the feedback results, use the adjusted data as the whole-process data, and generate a design assistance strategy including a number of key design anomalies according to the feedback results; Upload the design object information, obtain the corresponding whole-process data to formulate the basic parameters of the design process, and obtain the design assistance strategy to assist in comparing the design process.

[0015] The technical solutions provided by the embodiments of the present application have the following technical effects: First, through the combination of the acquisition module and the Internet of Things platform, different types of data in each link of the entire manufacturing process are uniformly collected and stored, and all key data in the equipment R & D and assembly process can be stably stored for a long time, facilitating subsequent analysis work; Unify the management of the original data in each link, the local reference materials corresponding to the guidance strategy, the data after adjustment after guidance, etc. to build a local knowledge base. Through the accumulation of big data, the local knowledge is continuously enriched, and in the subsequent monitoring and management of each link, the intelligent processing module can directly provide distributed independent guidance for the parameter information, process information, flow information, etc. in each link process; While providing distributed guidance, with the support of the whole-process data composed of multiple links, through the data - feedback results, it is also possible to directly provide guidance and assistance to the design source, realizing the control of the link process with data analysis and the guidance of the design process with the link results. Description of the Drawings

[0016] Figure 1It is a schematic diagram of module connection of an intelligent processing system for the whole process of R & D experiment based on the Internet of Things provided by this embodiment.

[0017] Figure 2 It is a schematic diagram corresponding to the quality inspection list in the quality inspection link of the embodiment of this application.

[0018] Figure 3 It is a schematic diagram of risk item marking in the embodiment of this application.

[0019] Figure 4 It is a schematic diagram of handling opinions and cause analysis in the embodiment of this application.

[0020] Figure 5 It is a schematic diagram of the list in the assembly link of the embodiment of this application.

[0021] Figure 6 It is a schematic diagram corresponding to the design assistance strategy in the embodiment of this application.

[0022] Figure 7 It is a schematic diagram of the steps of the intelligent processing method for the whole process of R & D experiment based on the Internet of Things provided by this embodiment. Detailed implementation manners

[0023] To understand the purpose, technical solutions and advantages of this application more clearly, the following describes and explains this application in combination with the accompanying drawings and embodiments. However, those of ordinary skill in the art should understand that this application can be implemented without these details. In some cases, in order to avoid unnecessary descriptions from making various aspects of this application obscure, well-known methods, processes, systems, components and / or circuits that have been described at a higher level will not be elaborated too much. For those of ordinary skill in the art, it is obvious that various changes can be made to the disclosed embodiments of this application, and without departing from the principles and scope of this application, the general principles defined in this application can be applied to other embodiments and application scenarios. Therefore, this application is not limited to the shown embodiments, but conforms to the broadest scope consistent with the scope required to be protected by this application.

[0024] It should be noted here that the description of these implementation manners is used to help understand the present invention, but does not constitute a limitation to the present invention. In addition, the technical features involved in the various implementation manners of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0025] In the description of this application, "several" means one or more, "multiple" means more than two, "greater than", "less than", "exceeding", etc. are understood not to include the base number, and "above", "below", "within", etc. are understood to include the base number. If "first" and "second" are described, they are only used to distinguish technical features and cannot be understood as indicating or implying relative importance or implicitly indicating the quantity of the indicated technical features or implicitly indicating the sequence relationship of the indicated technical features.

[0026] In the description of this application, descriptions with reference terms such as "one embodiment", "some embodiments", "schematic embodiments", "examples", "specific examples", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. In this specification, the schematic descriptions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a combined manner.

[0027] As Figure 1 shown, the embodiment of this application discloses an intelligent processing system for the whole process of R & D experiments based on the Internet of Things, including: A collection module, which is used to collect the original data in each link and upload it to the Internet of Things platform.

[0028] The collection module can collect the original data in each link of manufacturing R & D. The types of the original data include but are not limited to electronic document data, handwritten paper data, image data, and sensor data. The collection module includes a computer, a mobile tablet, an image recognition device, and various Internet of Things sensors.

[0029] After the collection module collects the original data in each link, it uploads various types of data to the Internet of Things platform for storage.

[0030] Based on this, the Internet of Things platform integrates and collects various types of data in each link and stores them uniformly, which is convenient for subsequent various data processing and the generation of guiding strategies.

[0031] An Internet of Things platform, which is used to classify a number of original data and store them in the data sets corresponding to different links.

[0032] After receiving the original data, the Internet of Things platform first processes the data to convert the original data into structured data. The processing methods include but are not limited to data cleaning, text parsing based on NLTK, semantic understanding, etc.

[0033] Structured data is generally in the form of tables, lists, inventories, etc., and unstructured original data includes text data.

[0034] When storing the original data, both the original data and the corresponding processed structured data need to be stored.

[0035] After storing all the data in each link, based on the link where each data is collected, the data corresponding to each link is integrated into a dataset.

[0036] The local knowledge base is used to obtain the original data in each dataset, retrieve the associated local data, and package and send it to the intelligent processing module; Local data includes original data in the historical process, data adjusted by historical guiding strategies, various types of standardized requirement materials, various types of rule lists, various types of drawing process materials, etc.

[0037] The local knowledge base serves as a fusion database, which stores all the data materials used for data analysis and processing. At the same time, the newly obtained data during the real-time operation process will also be stored in the local knowledge base for updating.

[0038] At the same time, when obtaining the data of a certain link and needing to guide this link later, it is necessary to retrieve in the local knowledge base according to the type, object, feature, link requirements, etc. corresponding to the data of this link. After retrieving the associated auxiliary materials, comparison materials, requirement materials, etc., the corresponding materials are used as the basis for comparing materials to determine whether the data of each link needs guidance and how to conduct guidance subsequently.

[0039] The intelligent processing module is used to compare and analyze the original data with the associated local data to generate a guiding strategy and send it to each link for process guidance. At the same time, it is also used to obtain the feedback results generated after adjustment based on the guiding strategy in each link.

[0040] The intelligent processing module is a big data intelligent AI model integrated and built in the system. It is mainly used to analyze and reason about various types of data, conduct intelligent retrieval in the local knowledge base, and generate guiding strategies for each subsequent link.

[0041] It first compares and analyzes each original data with the matching and associated local data. According to the different links, the objects of comparison, the methods of comparison, and the requirements of comparison are all different. Then, corresponding guiding strategies are generated according to different comparison results and sent to the link corresponding to the original data for process guidance. At the same time, after each link obtains the guiding strategy, it will choose whether to adjust according to the guiding strategy and whether the adjustment is effective, and feedback the corresponding results.

[0042] The intelligent processing module uploads the adjusted data as the whole-process data to the local knowledge base based on the feedback results. The intelligent processing module also generates a design assistance strategy containing several key design abnormal items according to the feedback results.

[0043] Meanwhile, the intelligent processing module also integrates the adjusted data in each link through feedback information to serve as the whole-process data, and uploads the whole-process data to the local knowledge base. It should be noted that all the link data of each model of workpiece corresponds to one whole-process data.

[0044] The intelligent processing module can also infer and judge the abnormal content in each link according to the feedback result, and calibrate the design key abnormal items corresponding to different parameter objects. The design key abnormal items correspond to the design source, which is characterized by analyzing the main guiding details from the beginning of the design source through the parameter situation analyzed in the actual manufacturing steps. The analyzed key abnormal items are used as the design assistance strategy.

[0045] The design module uploads the design object information, obtains the corresponding whole-process data in the local knowledge base to formulate the basic parameters of the design process, and interfaces with the intelligent processing module to obtain the design assistance strategy to assist in comparing the design process.

[0046] The design module is used for designers to design the parameters, processes, and flows of workpieces and components. After uploading the information of the object to be designed, such as the type, parameters, and process flow of the workpiece, it can obtain the corresponding whole-process data of this type of product in the local knowledge base. This whole-process data serves as the basic parameters corresponding to the entire design link. The basic data is used to provide an overall guiding idea based on historical analysis at the beginning of the design, enabling designers to know in advance the approximate parameter sizes, parameter relationships, items to be tested in subsequent links, feedback on whether they meet the link requirements, etc. of this type of product.

[0047] Meanwhile, the intelligent processing module also sends the corresponding key abnormal items to the design module based on the design object for design assistance, so as to prompt designers at the design source which parameter characteristics should be focused on in subsequent links or which characteristics have a greater impact on the performance of the finished product.

[0048] Through the above method, first, the combination of the acquisition module and the Internet of Things platform is used to uniformly collect and store different types of data in each link of the entire manufacturing process, which can stably store all key data in the equipment R & D and assembly process for a long time, facilitating subsequent analysis work; the original data in each link, the local reference materials corresponding to the guiding strategies, the data after adjustment after guidance, etc. are all uniformly managed to build a local knowledge base. Through the accumulation of big data, the local knowledge is continuously enriched. In the subsequent monitoring and management process of each link, the intelligent processing module can directly provide distributed and independent guidance for the parameter information, process information, flow information, etc. in each link process; while providing distributed guidance, with the support of the whole-process data composed of multiple links, the design source can be directly guided and assisted through the results of data-feedback, realizing the control of the link process through data analysis and the guidance of the design process through the link results.

[0049] In some other embodiments, the link includes a quality inspection link. The quality inspection is characterized by performing quality inspections on the parameters, dimensions, performance, etc. of each component before assembling each component. Only the components that pass the quality inspection can be used as qualified components that can enter the subsequent links. In the quality inspection link: Obtain the collected quality inspection data and upload it to the Internet of Things platform.

[0050] The quality inspection data includes component dimensions, component structures, etc., which are generally tabular data or handwritten picture data.

[0051] The local knowledge base retrieves the locally stored quality inspection list based on the quality inspection object in the quality inspection data.

[0052] The local knowledge base first retrieves the locally stored quality inspection list according to the quality inspection object. The quality inspection object is characterized by basic information such as the name and model of the component to be inspected. The quality inspection list is the quality inspection items, measurement methods, and parameter specifications corresponding to different components uploaded in advance.

[0053] When performing data comparison and analysis in the subsequent quality inspection link, corresponding data is obtained from the quality inspection data based on the quality inspection list, and it is analyzed whether it meets the quality inspection results required by the list.

[0054] Fill the quality inspection data into the corresponding quality inspection list and send it to the intelligent processing module. The intelligent processing module verifies each quality inspection data based on the specification requirements in the quality inspection list.

[0055] First, based on the component model in the structural data, each component is filled into the quality inspection list, and then the corresponding data in the quality inspection data is sequentially placed in the quality inspection list according to the measurement items required by the quality inspection list. The intelligent processing module compares and verifies each quality inspection data based on the specification requirements in the quality inspection list.

[0056] Generate risk items based on the verification results, where the verification results include at least out-of-tolerance and data missing.

[0057] Take Figure 2 as an example. The quality inspection list includes the following information: number, measurement item, drawing specification, measurement method, and the models of several components. At the same time, in the item corresponding to "total length", the total length of the component with the model "PM03XM" is 551.320, while the drawing specification of this item is 550±0.5, and "PM03XM" exceeds this range, so it corresponds to the risk item of out-of-tolerance; in the quality inspection item of "outer diameter", the component with the model "PM05XM" is missing the corresponding data, so it corresponds to the risk item of data missing. Among them, the components of each model are specifically rotors.

[0058] There are mainly two types of feedback information for the verification results. One is whether a quality inspection item can find the corresponding data in the quality inspection data. If it cannot be found, then the quality inspection item corresponds to a blank item and cannot perform quality inspection verification; the second is that after verification, it is analyzed that the error between a certain or certain parameters and the standard parameter specifications in the quality inspection is greater than the standard error of this item.

[0059] In the first type of feedback information, it corresponds to the risk item of data missing; in the second type of feedback information, it corresponds to the risk item of out-of-tolerance.

[0060] As shown in Figure 3 the figure, the system will separately classify and mark the parameters of each corresponding risk item in the quality inspection list. When all the items in the quality inspection list are completed with comparison and verification, the quality inspection list will be output as a quality inspection report. Among them, in the results obtained after separate classification and marking, it includes the number corresponding to the risk item, the measurement item, the allowable range of parameters calculated based on the drawing specification, the detected abnormal measurement value (if the risk item is out-of-tolerance, it corresponds to the value; if the risk item is data missing, it corresponds to a null value), the out-of-tolerance range, and the risk level (such as CRITICAL - critical, MAJOR - major).

[0061] In some other embodiments, the intelligent processing module is further configured to retrieve the corresponding processing opinions and cause analysis in the local knowledge base according to the content, quantity, and combined correlation coefficient of the risk items.

[0062] Furthermore, in order to facilitate the personnel in the quality inspection process to know how to optimize and rectify the quality inspection process, the intelligent processing module will also retrieve the processing opinions and cause analysis in the local knowledge base according to the specific content, quantity of the risk items, and the combined relationship between multiple risk items.

[0063] Among them, the handling opinions include measures, priorities, responsible parties, and completion time limits. Measures are characterized by what actions the guiding personnel need to take to resolve the risk item; priorities are characterized by the importance corresponding to each measure, and the higher the priority, the faster the action needs to be taken; responsible parties are characterized by the specific personnel or departments that need to carry out each measure; completion time limits are characterized by the time when different measures need to complete the actions. It should be noted that in some cases, some measures in the complete measures will correspond to other departments, which are used to enable other departments to cooperate with the departments in the quality inspection link to jointly guide and optimize the quality inspection data.

[0064] Through the handling opinions, each person or department in the quality inspection link can be guided to take corresponding measures within the scheduled time to deal with the abnormal risk items in the quality inspection link and inform the user how to conduct quality inspection again.

[0065] As shown in the Figure 4 attachment, there are priorities in the generated handling opinions, namely urgent and important; there are four measures, namely: "Isolate the out-of-tolerance parts with serial number 1", "Start the missing data traceability mechanism", "Implement measurement system analysis", "Establish an automatic out-of-tolerance alarm system"; the responsible persons are the Quality Inspection Department, the Production Department, the Quality Department, and the IT Department respectively; the completion time limits are immediately, 2 hours, 48 hours, and 72 hours respectively.

[0066] The cause analysis includes abnormal events and corresponding compliance probabilities. Abnormal events represent the possible cause events when the quality inspection data of a parameter is abnormal, and the compliance probability represents the occurrence probability corresponding to the cause event.

[0067] As shown in the Figure 4 attachment, when the quality inspection measurement value of PM03XM is out of tolerance, the possible cause events include fixture positioning deviation, with an occurrence probability of 80%, temperature compensation failure, with an occurrence probability of 15%, and measurement system error, with an occurrence probability of 5%.

[0068] When the data in No. 9 is missing, the possible cause events include manual recording omission, with an occurrence probability of 70%, and sensor communication interruption, with an occurrence probability of 30%.

[0069] Among them, the compliance probability is obtained based on the following steps: Select the corresponding cause list with the serial number in the local knowledge base based on the verification result.

[0070] First, after obtaining the verification result, obtain the cause list corresponding to the measurement item with this serial number in the local knowledge base. All possible causes that may be triggered when this type of component has this cause are stored in this list.

[0071] Generate the basic probability of each abnormal event in the cause list based on the content of the risk items.

[0072] Secondly, obtain the basic probability of this abnormal event in the cause list according to the content of the risk items. Among them, out-of-tolerance and data missing correspond to different basic probabilities. At the same time, the magnitude of the out-of-tolerance value also corresponds to different basic probabilities. For example, when the out-of-tolerance value is small, it is more likely to correspond to systematic errors, while when the out-of-tolerance value is large, it is more likely to correspond to fixture positioning deviation.

[0073] Generate a combined risk set based on the quantity and content of the risk items, calculate the combined correlation coefficient based on the combined risk set and the empirical algorithm, and adjust the basic probability based on the combined correlation coefficient to obtain the compliance probability.

[0074] After determining the basic probability through the content of a single risk item, when the number of risk items is greater than 1, it is also necessary to generate a combined risk set according to the quantity and content of the risk items, and optimize the specific probability corresponding to each risk item through the combined risk set.

[0075] For example, if there is a large amount of data missing in the risk set, and at the same time, the missing data all correspond to the same model component, then it is more likely that it is due to the omission of manual records after the quality inspection of this model component; if a large amount of missing data corresponds to each model, then it is less likely that the quality inspectors corresponding to each model component have omitted manual records, so the probability corresponding to sensor communication failure is greater.

[0076] For example, if there is a large amount of out-of-tolerance in the risk set, then when the inner diameter, outer diameter, and thread accuracy all show out-of-tolerance, because the measurement methods of each parameter are different, the probability caused by fixture deviation is relatively low, while the probability that the parts are deformed due to temperature compensation failure is relatively high.

[0077] Comprehensively analyze the causes and corresponding probabilities of the risk items through the combination of single analysis and combined analysis, which is convenient for the personnel in the quality inspection link to be intuitively guided to quickly adjust the quality inspection actions.

[0078] In some other embodiments, the link includes an assembly link. The assembly is characterized in that the assembly personnel assemble the parts after quality inspection together according to the assembly requirements. At the same time, the accuracy levels corresponding to the assembly of multiple components of different models and the assembly of components with different quality inspection parameters are different. Therefore, in order to analyze the accuracy advantages and disadvantages of different assembly combinations to obtain the guiding strategy for the assembly link, in the assembly link: The intelligent processing module obtains assembly data in the Internet of Things platform and obtains a matching selection rule document in the local knowledge base for the assembly object.

[0079] The intelligent processing module first retrieves the assembly data from the Internet of Things platform. The assembly data is generally saved in the form of a structured size table document, such as "JSC85 housing size detection data" and "JSC85 rotor size detection data".

[0080] At the same time, based on the type and model of the assembly object, it retrieves the matching selection rules document from the local knowledge base. This document contains the rules for assembling each housing and rotor, such as equal-diameter evaluation and other assembly rules, and also includes the comparison, evaluation, and classification rules for the consideration parameters after assembly.

[0081] As Figure 5 shown, the intelligent processing module permutes and combines several assembly objects with assembly relationships, and calculates the corresponding assembly parameters for each combination based on the selection rules document.

[0082] First, the intelligent processing module permutes and combines several assembly objects with assembly relationships to match each housing with each rotor one by one for assembly, and performs global optimization to select the optimal assembly combination of each housing + each rotor, and generates a list. The above steps are implemented through the LLM algorithm.

[0083] At the same time, based on the rule requirements of the selection rules document, the target parameters to be verified for the assembled assembly are obtained. The target parameters are the decision parameters used to consider the assembly effect after the combination of two or more components. For example, the target parameter after the assembly of the rotor and the housing is the clearance.

[0084] After selection, based on the target parameters, the corresponding assembly parameters for each assembly are calculated. The assembly parameters specifically include the overall parameter size and the sub-parameter size. For example, the overall parameter is the total clearance, and the sub-parameter is the clearance distribution.

[0085] Based on the selection rules document, the assembly standards are retrieved, and the assembly parameters are classified according to the assembly standards to issue the corresponding level of evaluation indicators, and each combination is assigned to the corresponding application environment according to each evaluation indicator.

[0086] The standard value corresponding to the target parameter is used as the assembly standard, and the assembly parameters of each assembly are compared with the assembly standard. Based on the difference value, each assembly combination is sorted. The larger the difference value, the worse the assembly effect is characterized.

[0087] At the same time, based on the size of the difference value, each combination is classified, and after classification, it corresponds to different levels of evaluation indicators. The evaluation indicators include "A+", "A", "A-", "B+", "B", "B-", etc. The higher the level, the higher the performance level corresponding to the rotor and housing of the assembly.

[0088] Finally, according to each evaluation index, each combination result is assigned to the corresponding application environment. For example, for a vacuum pump, the higher the level, the more it can be applied to high-precision application scenarios such as semiconductors, while the lower the level, it can be applied to sub-high-precision scenarios such as photovoltaics.

[0089] In some other embodiments, the link includes a testing link. The test is characterized in that after the components are assembled, the performance and parameters of the components are tested according to different test indexes. In the testing link: The intelligent processing module obtains test data in the Internet of Things platform and obtains the test items matched by the test object.

[0090] First, pull test data in the Internet of Things platform. The test data includes the type of the test object and the parameters that need to be tested.

[0091] At the same time, based on the test object, obtain the test items matched by the test object in the local database. The test items are characterized as the test items that must be carried out for different types of components, such as performance index test, temperature index test, rigidity index test, flipping index test, etc.

[0092] Send the test items for testing and obtain the change curves of each test data under different test items in real time.

[0093] Perform a preliminary screening of the test data according to the test items and monitor the change curves corresponding to the changes in the test data in real time.

[0094] Package the test data with abnormal change curves and the test items to obtain the key test items.

[0095] Based on the front and back data trends reflected by the change curves and the historical test data of the same type of devices that have passed the test stored in the local database, combine to judge whether there are test data with large deviations under different test items. If so, use these data as the key test items.

[0096] In some other embodiments, after the intelligent processing module compares and analyzes the original data and local data of each link, it can correspondingly generate guidance strategies for each link based on the analysis results and send them to the corresponding links. Specifically: Take the risk items, handling opinions and cause analysis as the first guidance strategy and send them to the responsible objects in the corresponding quality inspection link based on the priority for process guidance.

[0097] First, take the risk items, handling opinions, and cause analysis as the first guiding strategy and issue them to each responsible object in the quality inspection link. After analyzing the corresponding data in the quality inspection link, feedback it to the quality inspection link to inform which quality inspection parameters are at risk, how to handle them, and the reasons for their occurrence. The personnel in the quality inspection link can adjust the current quality inspection work according to the first guiding strategy, or make corresponding adjustments and preventive measures for subsequent quality inspection work based on the first guiding strategy.

[0098] Take the evaluation indicators as the second guiding strategy and issue them to the assembly link in the corresponding application environment for assembly guidance.

[0099] Take the evaluation indicators as the second guiding strategy to feedback to the personnel in the assembly link which matching method corresponds to the highest performance, and guide the personnel in the assembly link to classify each assembled component into which application environment.

[0100] Take each key test item as the third guiding strategy and issue it to the corresponding test object in the test link for test guidance.

[0101] Take the key test items as the third guiding strategy to feedback and guide the testers in the test link to focus on some key data for key testing and / or repeated testing.

[0102] In some other embodiments, the intelligent processing module is further configured to respectively obtain the feedback results in the quality inspection link, the assembly link, and the test link, and judge whether the guiding strategies corresponding to each link are effective based on the feedback results.

[0103] The intelligent processing module obtains the feedback results of each link on the guiding strategy. The feedback results include whether adjustments are made, whether there is improvement after adjustment, the adjusted parameters, etc.

[0104] Specifically analyze whether the guiding strategy in this link is effective according to the feedback results. If adjustments are made in this link after receiving the guiding strategy and the adjusted parameters are optimized, it is characterized that the guiding strategy is effective; if no adjustments are made or the parameters do not improve after receiving the guiding strategy in this link, it is characterized that the guiding strategy is ineffective.

[0105] If it is effective, the intelligent processing module sends the optimized data after adjustment to the local knowledge base for storage and adds the project case number.

[0106] If the guiding strategy is effective, the intelligent processing module sends the optimized data after adjustment of each link to the local knowledge base and adds the project case number. One project case number corresponds to the complete manufacturing process of a model of component.

[0107] The local knowledge base stores the optimized data, its corresponding original data, and local data uniformly to build the whole-process data including the data representing the pre-guidance state, the reference during guidance, and the data after guidance.

[0108] If it is invalid, the intelligent processing module marks the local data in the local knowledge base and re-selects local data to generate a new guidance strategy until the feedback result is valid.

[0109] If the guidance strategy is invalid, the intelligent processing module marks the local data used when analyzing this link. The marked local data can inform the system user to replace, update, and verify.

[0110] At the same time, the intelligent processing module will select a new guidance strategy in the local knowledge base to re-analyze and guide the data in the link until the guidance strategy for this link is valid.

[0111] When the project case number corresponds to the whole-process link, the local knowledge base takes all the data in the same project case number as the whole-process data.

[0112] When all feedback results are valid, all the data in the completed whole-process link are integrated as the whole-process data. The subsequent whole-process data is used to trace the design source and further guide the design link at the source.

[0113] In some other embodiments, the design key abnormal items include the data with a valid feedback result and adjusted based on the guidance strategy, and the data corresponding to an invalid feedback result.

[0114] When guiding the design link, the main guidance content focuses on guiding the design key abnormal items. For example, if certain data is likely to be abnormal in subsequent quality inspection, assembly, and testing links, then these data need to be focused on and analyzed in the design of subsequent other components.

[0115] Then the design key abnormal items include the data with a valid feedback result and adjusted based on the guidance strategy, and the data corresponding to an invalid feedback result. These data generally represent the data detected as abnormal in each link through analysis. The data detected as abnormal needs to be specially marked in subsequent design links for subsequent links to pay attention to this parameter. The data with an invalid feedback result represents the data for which the local knowledge base cannot provide accurate local data for comparison. These data have a high potential for abnormality in subsequent links, so they also need to be paid attention to in advance in the design link.

[0116] Through the above method, an advance guidance plan for subsequent newly designed products is generated based on the data analysis and guidance strategy feedback in each historical link, so as to optimize some data, processes, shapes, etc. specifically from the design source, realize a feedback closed-loop circuit from the data in the subsequent link to guide the previous design link, and enable positive feedback guidance for each link during data monitoring and reverse feedback guidance for the design link based on the analysis results and knowledge base.

[0117] In some other embodiments, the intelligent processing module takes the key design abnormal items as key features and selects the corresponding quality inspection data and / or assembly data and / or test data as key data.

[0118] The intelligent processing module takes each key design abnormal item as a key feature, and extracts the corresponding data from the data in each corresponding link as key data to bind the content that needs to guide the design link.

[0119] For the selection of key features, all key abnormal items can be selected, or only some key abnormal items can be selected. At the same time, the data corresponding to each model component is extracted and a table is drawn.

[0120] Obtain the preset bias in the local knowledge base based on the key features.

[0121] In the local knowledge base, obtain the bias corresponding to each key feature. The bias is characterized as a fixed deviation item independent of the characteristics of each data, and is used to adjust the size of the subsequent calculation influence parameters to improve the accuracy and robustness of the model. Through continuous training, the bias is continuously optimized among random values and updated based on the backpropagation algorithm.

[0122] Based on the key data and the preset bias, calculate the influence parameter matrix corresponding to each key feature. The influence parameter matrix contains several influence parameters used to characterize the influence magnitude of the key data corresponding to each key feature on the target result in the design link.

[0123] After obtaining the bias, calculate the influence parameter matrix corresponding to each key feature based on the key data and the bias. The influence magnitude of the data of each feature of each model in the matrix on the final target result is characterized as the influence parameter.

[0124] The larger the influence parameter, the more critical the feature is, and the abnormality of this feature will cause a greater impact on the final target result.

[0125] Generate a design assistance strategy based on the influence parameter.

[0126] Finally, organize the influence parameters of each characteristic corresponding to each model to obtain the corresponding list, and use this list as the design assistance strategy to send it to the design link. Designers in the design link can combine the design parameters in the design process with the influence parameters and put them into the model of the intelligent processing module as input items. Finally, the corresponding target parameters are output through the calculation of the model.

[0127] Specifically, as Figure 6 shown, in the design assistance strategy, there are first wire screw rotors of different models and several key characteristics, such as the root circle radius of the tooth, the suction lead, the exhaust lead, the rotor clearance, the internal pressure ratio, etc. At the same time, it includes the key data parameters corresponding to these key characteristics, and also includes the target result of "gas work" and the numerical values of the target results of each model.

[0128] At the same time, obtain the bias size corresponding to each key characteristic from the local knowledge base, and combine the linear matrix formula to calculate the size of the influence parameter corresponding to the key characteristic value of each model. The specific linear matrix formula is: ; Among them, represents the numerical size corresponding to the nth key characteristic of the mth model, represents the influence parameter corresponding to the mth key characteristic, y represents the predicted gas work size, and b represents the bias.

[0129] Through the above formula, after clarifying the size of the influence parameter of each key characteristic, send the corresponding list to the design link. Designers can know how much a certain key characteristic in different models of components affects the final target data through the size of the influence parameter, and use this as a guide to pay attention to the key data. At the same time, after designers design the size, performance and other data of each characteristic of the component, they can directly predict the gas work by bringing it into this linear matrix formula in combination with the bias and influence coefficient, and compare the predicted gas work with the test numerical value in the final test link. According to the comparison result, judge whether the assistance strategy in the design link matches the guidance strategy in the test link, so as to realize the feedback closed-loop loop once again.

[0130] As Figure 7 shown, the present application also discloses an intelligent processing method for the whole process of R & D experiment based on the Internet of Things, including the following steps: S100, collect the original data in each link.

[0131] S200, classify a number of original data and store them in the data sets corresponding to different links.

[0132] S300, obtain the original data in each data set and retrieve the associated local data.

[0133] S400 compares and analyzes the original data with the associated local data to generate a guidance strategy and distribute it to each link for process guidance. At the same time, it obtains the feedback results generated after adjustment based on the guidance strategy in each link.

[0134] S500 uses the adjusted data as the whole-process data based on the feedback results, and generates a design assistance strategy including several key design abnormal items according to the feedback results.

[0135] S600 uploads the design object information, obtains the corresponding whole-process data to formulate the basic parameters of the design process, and obtains the design assistance strategy to assist in the comparison of the design process.

[0136] The implementation principle is as follows: It should be understood that although the steps in the flowchart of the accompanying drawings are shown sequentially according to the arrows, these steps do not necessarily have to be executed in the order indicated by the arrows. Unless there is a clear description in this article, the execution of these steps has no strict order limit and can be executed in other orders.

[0137] The above are all the preferred embodiments of this application. The protection scope of this application is not limited by this. Therefore, all equivalent changes made according to the structure, shape, and principle of this application should be covered within the protection scope of this application.

Claims

1. An intelligent processing system for the whole process of R & D experiments based on the Internet of Things, characterized in that, It includes: A collection module, which is used to collect the original data in each link and upload it to the Internet of Things platform; The Internet of Things platform, which is used to classify a number of the original data and store them in the data sets corresponding to different links; A local knowledge base, which is used to obtain the original data in each of the data sets, retrieve the associated local data, and package and send them to the intelligent processing module; The intelligent processing module, which is used to compare and analyze the original data with the associated local data to generate a guidance strategy and send it to each link for process guidance. At the same time, it is also used to obtain the feedback results generated after adjustment based on the guidance strategy in each link; The intelligent processing module uploads the adjusted data as the whole-process data to the local knowledge base based on the feedback results. The intelligent processing module also generates a design assistance strategy including a number of key design exception items according to the feedback results; The design module uploads the design object information, obtains the corresponding whole-process data in the local knowledge base to formulate the basic parameters of the design process, and docks with the intelligent processing module to obtain the design assistance strategy to assist in comparing the design process.

2. The intelligent processing system for the whole process of R & D experiment based on the Internet of Things according to claim 1, characterized in that, The link includes a quality inspection link. In the quality inspection link: Obtain the collected quality inspection data and upload it to the Internet of Things platform; The local knowledge base retrieves the local stored quality inspection list based on the quality inspection object in the quality inspection data; Fill the quality inspection data into the corresponding quality inspection list and send it to the intelligent processing module. The intelligent processing module verifies each quality inspection data based on the specification requirements in the quality inspection list; Generate risk items based on the verification results. The verification results at least include out-of-tolerance and data missing.

3. The intelligent processing system for the whole process of R & D experiment based on the Internet of Things according to claim 2, wherein The intelligent processing module is also used to retrieve the corresponding handling opinions and cause analyses in the local knowledge base according to the content, quantity, and combination correlation coefficient of the risk items; Among them, the handling opinions include measures, priorities, responsible objects, and completion time limits; The cause analysis includes abnormal events and corresponding compliance probabilities. The compliance probabilities are obtained based on the following steps: Select the corresponding numbered cause list in the local knowledge base based on the verification results; Generate the basic probabilities of each abnormal event in the cause list based on the content of the risk items; Generate a combined risk set based on the quantity and content of the risk items. Calculate the combined correlation coefficient based on the combined risk set and the empirical algorithm, and adjust the basic probability based on the combined correlation coefficient to obtain the compliance probability.

4. The intelligent processing system for the whole process of R & D experiments based on the Internet of Things according to claim 2, wherein The link includes an assembly link. In the assembly link: The intelligent processing module obtains the assembly data in the Internet of Things platform and obtains the matching selection rule document for the assembly object in the local knowledge base; The intelligent processing module combines and arranges a number of the assembly objects with assembly relationships, and calculates the corresponding assembly parameters for each combination based on the selection rule document; Pull the assembly standard based on the selection rule document, classify each assembly parameter according to the assembly standard, and send the corresponding level of evaluation indicators to allocate each combination to the corresponding application environment.

5. The intelligent processing system for the whole process of R & D experiment based on the Internet of Things according to claim 4, characterized in that, The link includes a test link, in which: The intelligent processing module obtains test data in the Internet of Things platform and obtains test items matched by the test object in the local knowledge base; The test items are sent down for testing, and the change curves of the test data under different test items are obtained in real time; The test data with abnormal change curves and the test items are packaged to obtain key test items.

6. The intelligent processing system for the whole process of R & D experiments based on the Internet of Things according to claim 5, characterized in that, The intelligent processing module is also used for: Taking the risk items, the processing opinions and the cause analysis as the first guidance strategy and sending them down to the responsible objects in the corresponding quality inspection link based on the priority for process guidance; Taking the evaluation indicators as the second guidance strategy and sending them down to the assembly link in the corresponding application environment for assembly guidance; Taking each key test item as the third guidance strategy and sending it down to the corresponding test object in the test link for test guidance.

7. The intelligent processing system for the whole process of R & D experiment based on the Internet of Things according to claim 5, characterized in that The intelligent processing module is also used to obtain the feedback results in the quality inspection link, the assembly link and the test link respectively, and judge whether the guidance strategies corresponding to each link are effective based on the feedback results; If it is effective, the intelligent processing module sends the optimized data after adjustment to the local knowledge base for storage and adds a project case number; If it is invalid, the intelligent processing module marks the local data in the local knowledge base and reselects the local data to generate a new guidance strategy until the feedback result is effective; When the project case number corresponds to the whole process link, the local knowledge base takes all the data in the same project case number as the whole process data.

8. The intelligent processing system for the whole process of R & D experiment based on the Internet of Things according to claim 7, characterized in that, The design key abnormal items include the data adjusted based on the guidance strategy with the feedback result being effective and the data corresponding to the invalid feedback result.

9. The intelligent processing system for the whole process of R & D experiment based on the Internet of Things according to claim 8, characterized in that, The intelligent processing module takes the design key abnormal items as key features and selects corresponding quality inspection data and / or assembly data and / or test data as key data; Obtain a preset offset in the local knowledge base based on the key features; Calculate the influence parameter matrix corresponding to each key feature based on the key data and the preset offset, and the influence parameter matrix contains several influence parameters used to characterize the influence degree of the key data corresponding to each key feature on the target result in the design link; Generate the design assistance strategy based on the influence parameters.

10. An intelligent processing method for the whole process of R & D experiments based on the Internet of Things, characterized in that, It includes the following steps: Collect the original data in each link; Classify several of the original data and store them in the data sets corresponding to different links; Obtain the original data in each data set and retrieve the associated local data; Compare and analyze the original data with the associated local data to generate a guidance strategy and send it down to each link for process guidance, and at the same time obtain the feedback results generated after adjustment based on the guidance strategy in each link; Take the adjusted data as the whole process data based on the feedback results, and generate a design assistance strategy including several design key abnormal items according to the feedback results; Upload the design object information, obtain the corresponding whole-process data to formulate the basic parameters of the design process, and obtain the design assistance strategy to assist and compare the design process.

Images