A quality monitoring system for the entire life cycle

Abstract

The present invention discloses a full life cycle quality monitoring system, which relates to the technical field of quality monitoring. The system comprises a network modeling module, a life cycle management module and a display feedback module, and the life cycle management module comprises a process monitoring unit, a quality inspection judgment unit and a tracing unit. The technical key points are as follows: by building a full life cycle monitoring platform, by combining blockchain technology and a monitoring knowledge base, a blockchain network is built with production nodes, quality inspection nodes and buffer nodes, and corresponding calling interfaces are provided between nodes, quality assessment analysis is performed based on the judgment function and the evaluation function in the process determination contract, and a first classification model and a second classification model are designed to accurately classify the feature analysis of each process, and whether the production line implements a defect management mechanism is determined based on the quality inspection planning contract; data between processes are seamlessly connected through a number of smart contracts, so as to facilitate the monitoring and tracing of subsequent process information.

Description

Technical Field

[0001] The present invention relates to the technical field of quality monitoring, and in particular to a full life cycle quality monitoring system. Background Art

[0002] In the whole production line of products, a decentralized monitoring method is usually adopted, that is, independent inspection points are set up at different stages of the production process for quality monitoring; although this method can guarantee the quality of products to a certain extent, there are problems such as discontinuous monitoring and information islands. The data between various links cannot be seamlessly connected, and it is difficult to fully grasp the quality changes of the product throughout its life cycle. The information between the links is difficult to trace, especially when the product has a problem in a certain link, it is impossible to trace the root cause of the problem, resulting in the product continuing to circulate and process, resulting in low monitoring efficiency;

[0003] At the same time, there are still deficiencies in achieving quality monitoring throughout the entire life cycle. Specifically, in multiple processes, quality assessment standards and related characteristics are often different. However, in the actual production process, data are usually processed and analyzed in a unified manner, and accurate analysis is not carried out on the characteristics of each process. There is a lack of in-depth understanding and distinction between the differences between different stages and processes, which makes it impossible to effectively capture potential quality risks in each link, making the assessment results not detailed enough and prone to missing certain key quality issues. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In view of the shortcomings of the prior art, the present invention provides a full life cycle quality monitoring system, including a network modeling module, a life cycle management module and a display feedback module, and the life cycle management module includes a process monitoring unit, a quality inspection judgment unit and a traceability unit; by building a full life cycle monitoring platform, and combining blockchain technology and a monitoring knowledge base, a blockchain network is built with production nodes, quality inspection nodes and buffer nodes, and corresponding calling interfaces are provided between nodes, and based on a number of smart contracts of the blockchain network, the data between each process are seamlessly connected, which facilitates the subsequent monitoring and tracing of the information of each process; in this process, a first classification model and a second classification model are designed, corresponding quality assessment indicators and related features are set for the data of each process, each process is accurately analyzed, and the quality problems existing in each process are effectively captured, thereby solving the problems raised in the background technology.

[0006] (II) Technical solution

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions:

[0008] The present application provides a full life cycle quality monitoring system, the system comprising:

[0009] The network modeling module is used to build a full life cycle monitoring platform. The full life cycle monitoring platform has a preset blockchain network and monitoring knowledge base. Several smart contracts are established on the blockchain network. The smart contracts include process determination contracts, quality inspection planning contracts, and smart interaction contracts. The blockchain network consists of several production nodes, several quality inspection nodes, and a buffer node.

[0010] The life cycle management module has a built-in process monitoring unit, a quality inspection and determination unit, and a traceability unit, which is used to analyze the life cycle information of the entire production line based on the full life cycle monitoring platform. The life cycle information includes the process data, monitoring indicator data, and quality inspection indicator data of the entire production line;

[0011] The process monitoring unit is used to perform quality assessment analysis based on the judgment function and evaluation function pre-configured in the process determination contract, wherein the judgment function extracts the combined features of each process based on the life cycle information, and the evaluation function obtains the quality assessment value based on the combined feature analysis of each process, and returns the quality assessment value to the blockchain network;

[0012] The quality inspection judgment unit is used to judge whether the corresponding production line executes the defect management mechanism based on the quality inspection planning contract, wherein the defect management mechanism triggers an operation after receiving the quality assessment value: judging the state event based on the quality assessment value, when it is judged as the first state event, continuing to execute the quality inspection abnormality monitoring strategy; when it is judged as the second state event, evaluating the abnormality degree of the corresponding process, producing the abnormality assessment coefficient, and executing the corresponding strategy;

[0013] The traceability unit is used to build information traceability relationships based on smart interactive contracts, and each information traceability relationship corresponds to a unique monitoring identification tag, and provides a traceability system for the corresponding monitoring process and monitoring progress;

[0014] Display feedback module, used for feedback and display on the full life cycle monitoring platform.

[0015] Furthermore, the production node is used to store production data, including: identifier, process ID, process number, process name, production progress data, flow sheet data and resource status data; the quality inspection node is used to store quality inspection data, including: identifier, quality inspection ID, quality inspection number, quality inspection name, quality inspection progress data, process quality data and quality description data; the buffer node is used to store data to be updated and modified, and perform process determination work before node storage.

[0016] Furthermore, the smart interactive contract is pre-configured with a data storage structure, a data storage function, an associated access rule, and an associated access function, stores the corresponding data under the production node, the quality inspection node, and the buffer node, and provides a call interface corresponding to the production node, the quality inspection node, and the buffer node; the process determination contract is pre-configured with a judgment function and an evaluation function, and provides a call interface between the smart interactive contract and the monitoring knowledge base; the quality inspection planning contract is pre-configured with a state judgment event and a defect management mechanism, and provides a call interface between the smart interactive contract and the monitoring knowledge base;

[0017] Draw a two-dimensional coordinate system, monitor the calling interface, regularly obtain the interface parameters within the preset monitoring time period, and visualize the interface parameters in the two-dimensional coordinate system to obtain the parameter change curve within the preset monitoring time period; draw a standard change curve in the two-dimensional coordinate system, and overlap the parameter change curve with the standard change curve, obtain the abnormal parameter change curve segment based on the overlapping result, identify the abnormal state of the calling interface of the blockchain network, and execute the preset automatic correction strategy accordingly; wherein, the interface parameters include calling frequency, parameter verification and response time.

[0018] Furthermore, the process of extracting characteristic information of each process includes:

[0019] Pre-construct a first classification model, substitute the collected life cycle information into the first classification model, select features for feature combination, construct new features, and select the required features through feature selection methods; introduce a training layer into the first classification model, combine the selected features with the original features to form a new training data set, and provide randomly selected sample data from the collected life cycle information, train and update the training layer, use a decision tree to evaluate the importance of the corresponding features, and obtain a second classification model;

[0020] The acquired life cycle information is stratified according to the time series, and any time series is marked as a process. The life cycle information under each time series is substituted into the second classification model to determine the first probability that the combined features extracted under each process are useful features and the second probability that they are related features.

[0021] Furthermore, the first probability and the second probability are obtained by analyzing and processing based on the information gain of the feature. When the first probability or the second probability is greater than a preset probability threshold, the combined feature is matched to the corresponding data label; otherwise, the proportion of the first probability and the second probability under the corresponding process is counted, and the weight parameters of the second classification model are adjusted based on the proportion, and retraining and classification are performed.

[0022] Furthermore, the evaluation function obtains a quality evaluation value based on the combined characteristic analysis of each process, including:

[0023] Define the combined feature as a key factor, and the key factor includes several dynamic feature indicators, mark any dynamic feature indicator as i, and the data value marking any dynamic feature indicator i is Li, call the standard interval St[SQ1, SQ2] of the dynamic feature indicator i from the monitoring knowledge base, and the standard interval St corresponds to the dynamic feature indicator i one by one; calculate the difference coefficient of the corresponding combined feature under each process based on the standard interval St, and assign the corresponding weight coefficient based on the difference coefficient;

[0024] The calculation formula of the quality assessment value is: ;

[0025] In the formula, quality represents the quality evaluation value, cvi represents the difference coefficient calculated for a certain dynamic feature index i, βi represents the weight coefficient assigned to a certain dynamic feature index i, and βi is greater than 0.

[0026] Furthermore, judging whether the corresponding production line executes the defect management mechanism based on the quality inspection planning contract includes:

[0027] Call the comparison threshold of the quality assessment value from the monitoring knowledge base, and compare and analyze the quality assessment value with the standard threshold:

[0028] When the quality assessment value is greater than or equal to the comparison threshold, the quality assessment value is marked as q1, and edited as a first-level character, and q1 and the first-level character are combined to generate a first state event; an abnormal monitoring strategy is executed on the first state event, and it is temporarily stored;

[0029] When the quality assessment value is less than the comparison threshold, the quality assessment value is marked as q2, and edited as a secondary character, and q2 and the secondary character are combined to generate a second state event; the abnormality degree of the second state event is evaluated, and the specific process is: obtaining an evaluation parameter data set, including the quality assessment value, the time difference between the time when the second state event occurs and the current time, and the risk level;

[0030] The formula for calculating the abnormal assessment coefficient is: ;

[0031] Where PG represents the abnormal assessment coefficient, fx represents the risk level, cz represents the time difference between the time when the second state event occurs and the current time, and Φ0 is the time constant;

[0032] The standard threshold of the abnormal evaluation coefficient is called from the monitoring knowledge base, and when the abnormal evaluation coefficient exceeds the standard threshold, an alarm signal is generated.

[0033] Furthermore, the acquisition of the risk level includes:

[0034] The combined features of any process are integrated and marked as a set J = {a 1 , a 2 , ..., a N0}; where a 1 、a 2 and a N0 represents the useful features or relevant features corresponding to the process, and N0 represents the number of useful features or relevant features; draws the difference change curve of each feature within the preset acquisition period, and obtains the maximum peak value and the minimum trough value in each difference change curve, and determines the span value based on the difference between the maximum peak value and the minimum trough value;

[0035] Obtain the maximum value of the corresponding span value in the process in the historical period and mark it as the maximum span value. Count the number of times the span value exceeds the maximum span value and mark it as the risk number. At the same time, calculate the difference between the span value and the corresponding maximum span value and mark it as the risk difference.

[0036] Draw a two-dimensional coordinate system, visualize the risk difference in the two-dimensional coordinate system, and obtain the risk difference change curve within the preset collection period; call the risk difference change threshold curve of the monitoring knowledge base and draw it in the two-dimensional coordinate system, obtain the angle value of the first intersection of the two curves, and mark it as the risk angle; generate the risk degree by combining the risk angle and the number of risks, based on the formula: ;

[0037] In the formula, cs represents the number of risks, jd represents the risk angle, α0 represents the weight correction coefficient, and α0 is greater than 0.

[0038] Furthermore, the information traceability relationship is constructed based on the smart interactive contract, including:

[0039] Based on the calling interface between the smart interaction contract and the process monitoring contract, the monitoring attribute description corresponding to each process in the buffer node is obtained, and the monitoring attribute description corresponds to the quality inspection index data; based on the calling interface between the smart interaction contract and the quality inspection planning contract, the quality inspection attribute description corresponding to each process in the buffer node is obtained, and the quality inspection attribute description corresponds to the quality inspection index data;

[0040] The traceability description is defined based on the monitoring attribute description and the quality inspection attribute description, and the monitoring identification tag is generated in combination with the corresponding node of the blockchain network, and the monitoring identification tag is stored in the blockchain network in the form of a hash value.

[0041] (III) Beneficial effects

[0042] The present invention provides a full life cycle quality monitoring system, which has the following beneficial effects:

[0043] 1. The present invention combines blockchain technology and monitoring knowledge base, builds a blockchain network through production nodes, quality inspection nodes and buffer nodes, and designs corresponding call interfaces between nodes; regularly performs parameter analysis on the call interface, executes automatic correction strategies when an abnormality occurs, and ensures the accuracy and security of data collection during the quality monitoring process of the entire life cycle; at the same time, blockchain technology is used to save the on-chain data to ensure its non-tamperability, and a complete production line mapping is formed through the corresponding data and interface parameters in several processes to achieve information traceability, thereby improving the credibility of the traceability system to a certain extent;

[0044] 2. The present invention introduces several smart contracts, and performs quality assessment analysis based on the judgment function and evaluation function in the process determination contract to obtain a quality assessment value. In this process, the collected life cycle information is classified, different combination features of each process are screened out, useless features are eliminated, important features are retained, and weights are further determined, thereby indirectly improving the classification accuracy of the second classification model. Based on the quality inspection planning contract, it is determined whether the production line implements the defect management mechanism, specifically including determining the status event based on the quality assessment value, and executing corresponding strategies according to different status events. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a module diagram of a full life cycle quality monitoring system according to an exemplary embodiment. DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0047] Example

[0048] An embodiment of the present invention provides a full life cycle quality monitoring system. Figure 1 is a schematic diagram of a full life cycle quality monitoring system module according to an exemplary embodiment. Figure 1 , including: a network modeling module, a life cycle management module and a display feedback module, and the network modeling module, the life cycle management module and the display feedback module are communicatively connected; wherein the life cycle management module includes a process monitoring unit, a quality inspection and determination unit and a traceability unit;

[0049] The following is an explanation or description of each module:

[0050] Network Modeling Module:

[0051] It is used to build a full life cycle monitoring platform. The full life cycle monitoring platform has a preset blockchain network and monitoring knowledge base. Several smart contracts are established on the blockchain network. The smart contracts include process determination contracts, quality inspection planning contracts, and smart interaction contracts. The blockchain network consists of several production nodes, several quality inspection nodes, and a buffer node.

[0052] The production node is used to store production data, including: identifier, process ID, process number, process name, production progress data, flow order data, and resource status data;

[0053] The quality inspection node is used to store quality inspection data, including: identifier, quality inspection ID, quality inspection number, quality inspection name, quality inspection progress data, process quality data, and quality description data;

[0054] The buffer node is used to store data to be updated and modified, and to determine the process before node storage;

[0055] The smart interactive contract is pre-configured with data storage structure, data storage function, associated access rules and associated access function, which stores the corresponding data under the production node, quality inspection node and buffer node, and provides the corresponding calling interface of the production node, quality inspection node and buffer node;

[0056] Data storage structure: a method used to organize and store data from different nodes (production nodes, quality inspection nodes, and buffer nodes) to ensure that data can be clearly associated and managed between different nodes;

[0057] Data storage function: a function used to store, modify and update data. The function stores the corresponding node data in the state variable of the contract. The data of each node can be identified and updated by its unique ID;

[0058] Association access rules: by associating the monitoring attributes in the monitoring indicator data with the quality inspection attributes in the quality inspection indicator data, these attributes also correspond to the identifier attribute and process name attribute of the monitoring indicator data, and the identifier attribute and quality inspection name attribute of the quality inspection indicator data;

[0059] The associated access function is used to share data based on identifier attributes, specifically data sharing among production nodes, quality inspection nodes, and buffer nodes. Thus, through the association of identifier attributes, a cross-node data traceability chain can be established in the contract, and a unique traceability relationship can be established.

[0060] The process determination contract is pre-configured with judgment functions and evaluation functions, and provides calling interfaces corresponding to production nodes, quality inspection nodes, and buffer nodes;

[0061] The quality inspection planning contract is pre-configured with status determination events and defect management mechanisms, and provides a calling interface between the smart interactive contract and the monitoring knowledge base;

[0062] By drawing a two-dimensional coordinate system, the calling interface is monitored, the interface parameters within the preset monitoring time period are obtained regularly, and the interface parameters are visualized in the two-dimensional coordinate system to obtain the parameter change curve within the preset monitoring time period; and a standard change curve is drawn in the two-dimensional coordinate system, and the parameter change curve and the standard change curve are overlapped with each other, and the abnormal parameter change curve segment is obtained based on the overlapping result, the abnormal state of the calling interface of the blockchain network is identified, and the preset automatic correction strategy is executed accordingly; wherein, the interface parameters include calling frequency, parameter verification and response time;

[0063] Automatic remediation strategies include:

[0064] Limit frequency, load balance, or enable current limiting; automatically correct requests or return error responses; automatically switch to backup services, adjust system resources, and optimize queries or API calls; when multiple anomalies are detected and the accumulated errors exceed the set threshold, the system's global recovery strategy can be initiated, such as automatically restarting the application, clearing the cache, and rebuilding the index;

[0065] Lifecycle Management Module:

[0066] It is used to analyze the life cycle information of the entire production line based on the full life cycle monitoring platform; and has a built-in process monitoring unit, a quality inspection and determination unit, and a traceability unit; wherein the life cycle information includes the process data, monitoring index data, and quality inspection index data of the entire production line;

[0067] Process monitoring unit:

[0068] It is used to perform quality assessment analysis based on the judgment function and evaluation function pre-configured in the process determination contract, wherein the judgment function extracts the combined features of each process based on the life cycle information, and the evaluation function obtains the quality assessment value based on the combined feature analysis of each process, and returns the quality assessment value to the blockchain network;

[0069] The process of extracting characteristic information of each process includes:

[0070] Pre-construct a first classification model, substitute the collected life cycle information into the first classification model, select features for feature combination, construct new features, and select the required features through feature selection methods; introduce a training layer into the first classification model, combine the selected features with the original features to form a new training data set, and provide randomly selected sample data from the collected life cycle information, train and update the training layer, use a decision tree to evaluate the importance of the corresponding features, and obtain a second classification model;

[0071] The acquired life cycle information is stratified according to the time series, and any time series is marked as a process. The life cycle information under each time series is substituted into the second classification model to determine the first probability that the combined features extracted under each process belong to the useful features and the second probability that the combined features are related features.

[0072] The first probability and the second probability are obtained by analyzing and processing based on the information gain of the feature;

[0073] For example: Based on the life cycle information and process sequence, the data is first divided into multiple different processes according to the time series, and feature extraction is performed for any process; assuming that the extracted combined features include size, shape, position, structure, surface microstructure, and welding microstructure, a set of feature values ​​is obtained under the time series data of each process, and the corresponding data set D is taken as the input information of the model; for the data set D, the XGBoost algorithm can be used to reduce the feature dimension, reduce redundancy and noise features, generate a new feature set, and normalize it; in the subsequent iterative training, the information gain corresponding to the value of each feature is calculated, and the principal component analysis method is used to select the features ranked in the top N1 of the information gain to form a new data set, and the combined features of each process are updated in real time;

[0074] The formula for calculating information gain is: ;

[0075] In the formula, H(D) represents information gain, H(D) is the entropy of the data set D formed in each iterative training, and H(D j ) indicates that the dataset D is divided into subsets D by a feature j The entropy after

[0076] The data labels are also annotated for each process, usually using "useful features" and "relevant features" as labels, and supervised training is performed to make the second classification model more accurate.

[0077] When the first probability or the second probability is greater than the preset probability threshold, the combined feature is matched to the corresponding data label; otherwise, the proportion of the first probability and the second probability under the corresponding process is counted, and the weight parameter of the second classification model is adjusted based on the proportion, and retraining and classification are performed;

[0078] In this embodiment, a preset probability threshold is defined as 0.5. Assuming that the probability of a feature belonging to the first category calculated by the second classification model is 0.7 and the probability of belonging to the second category is 0.3, and 0.7>0.5, the feature is judged to be a useful feature. According to this step, the situation that other features in the combined feature belong to useful features or relevant features is calculated, and when the second probability of the relevant feature is 0, it indicates that the relevant feature is a useless feature and is deleted.

[0079] Based on the first probability and the second probability, the proportion of the first probability is calculated, and the loss function L(W) of the second classification model is called at the same time for weighted update. The updated weight is:

[0080] ;

[0081] Where, L weight (W) represents the weight of the weighted update, zb represents the proportion value, x and y represent the input feature values ​​and the corresponding data labels, and L(W, x, y) represents the loss of the second classification model;

[0082] The quality assessment value is obtained based on the combined characteristics analysis of each process, including:

[0083] Define the combined feature as a key factor, and the key factor includes several dynamic feature indicators, mark any dynamic feature indicator as i, and the data value marking any dynamic feature indicator i is Li, call the standard interval St[SQ1, SQ2] of the dynamic feature indicator i from the monitoring knowledge base, and the standard interval St corresponds to the dynamic feature indicator i one by one; calculate the difference coefficient of the corresponding combined feature under each process based on the standard interval St, and assign the corresponding weight coefficient based on the difference coefficient;

[0084] The calculation formula of the quality assessment value is: ;

[0085] In the formula, quality represents the quality evaluation value, cvi represents the difference coefficient calculated for a certain dynamic feature index i, βi represents the weight coefficient assigned to a certain dynamic feature index i, and βi is greater than 0;

[0086] For example, suppose there are three dynamic indicator characteristics in the process, namely temperature, pressure and production speed:

[0087] Case 1: The difference coefficients of temperature, pressure and production speed are 0.05, 0.15 and 0.10 respectively; the weight coefficients are 0.4, 0.3 and 0.3 respectively;

[0088] The quality assessment value is: 0.4*0.05+0.3*0.15+0.3*0.10=0.095;

[0089] Case 2: The difference coefficients of temperature, pressure and production speed are 0.1, 0.15 and 0.10 respectively; the weight coefficients are 0.4, 0.3 and 0.3 respectively;

[0090] The quality assessment value is: 0.3*0.1+0.3*0.15+0.3*0.10=0.105;

[0091] It should be noted that 0.095 < 0.105; the abnormal fluctuation in the production process of case 1 is smaller than that of case 2;

[0092] Quality inspection unit:

[0093] Used to determine whether the corresponding production line implements the defect management mechanism based on the quality inspection planning contract, wherein the defect management mechanism triggers the operation after receiving the quality assessment value: determine the state event based on the quality assessment value, and when it is determined to be a first state event, continue to execute the quality inspection abnormality monitoring strategy; when it is determined to be a second state event, evaluate the abnormality degree of the corresponding process, produce the abnormality assessment coefficient, and execute the corresponding strategy;

[0094] The process of implementing the defect management mechanism is:

[0095] Call the comparison threshold of the quality assessment value from the monitoring knowledge base, and compare and analyze the quality assessment value with the standard threshold:

[0096] When the quality assessment value is greater than or equal to the comparison threshold, the quality assessment value is marked as q1, and edited as a first-level character, and q1 and the first-level character are combined to generate a first state event; an abnormal monitoring strategy is executed on the first state event, and it is temporarily stored;

[0097] When the quality assessment value is less than the comparison threshold, the quality assessment value is marked as q2, edited as a secondary character at the same time, and q2 and the secondary character are combined to generate a second state event;

[0098] The abnormality level of the second state event is evaluated. The specific process is as follows:

[0099] Obtaining a parameter data set for evaluation, including a quality evaluation value, a time difference between the time when the second state event occurs and the current time, and a risk level;

[0100] The combined features of any process are integrated and marked as a set J = {a 1 , a 2 , ..., a N0}; where a 1 、a 2 and a N0represents the useful features or relevant features corresponding to the process, and N0 represents the number of useful features or relevant features; draws the difference change curve of each feature within the preset acquisition period, and obtains the maximum peak value and the minimum trough value in each difference change curve, and determines the span value based on the difference between the maximum peak value and the minimum trough value;

[0101] Obtain the maximum value of the corresponding span value in the process in the historical period and mark it as the maximum span value. Count the number of times the span value exceeds the maximum span value and mark it as the risk number. At the same time, calculate the difference between the span value and the corresponding maximum span value and mark it as the risk difference.

[0102] Draw a two-dimensional coordinate system, visualize the risk difference in the two-dimensional coordinate system, and obtain the risk difference change curve within the preset collection period; call the risk difference change threshold curve of the monitoring knowledge base and draw it in the two-dimensional coordinate system, obtain the angle value of the first intersection of the two curves, and mark it as the risk angle; generate the risk degree by combining the risk angle and the number of risks, based on the formula: ;

[0103] In the formula, fx represents the risk degree, cs represents the number of risk events, jd represents the risk angle, α0 represents the weight correction coefficient, and α0 is greater than 0;

[0104] The formula for calculating the abnormal assessment coefficient is: ;

[0105] Among them, PG represents the abnormal assessment coefficient, and cz represents the time difference between the time when the second state event occurs and the current time. The longer the time when the second state event occurs, the weaker its influence, so time decay is required. represents the time attenuation factor, Φ0 is the time constant, which can be adjusted according to the actual situation. It is a common technical means for those skilled in the art, so it will not be elaborated here. It should be noted that the larger the corresponding value of the abnormal assessment coefficient is, the more unqualified products appear on the production line, which greatly reduces the production efficiency.

[0106] The standard threshold of the abnormality assessment coefficient is called from the monitoring knowledge base. When the abnormality assessment coefficient exceeds the standard threshold, an alarm signal is generated. When the alarm signal is received, the text is immediately edited and displayed on the full life cycle monitoring platform. The text content is "quality inspection abnormality occurs". Professionals take corresponding actions based on the text content, such as batch rework or batch discard of the process.

[0107] Traceability unit:

[0108] It is used to build information traceability relationships based on smart interactive contracts, and each information traceability relationship corresponds to a unique monitoring identification tag, and provides a traceability system for the corresponding monitoring process and monitoring progress;

[0109] Building information traceability relationships based on smart interactive contracts, including:

[0110] Based on the calling interface between the smart interaction contract and the process monitoring contract, the monitoring attribute description corresponding to each process in the buffer node is obtained, and the monitoring attribute description corresponds to the quality inspection index data; based on the calling interface between the smart interaction contract and the quality inspection planning contract, the quality inspection attribute description corresponding to each process in the buffer node is obtained, and the quality inspection attribute description corresponds to the quality inspection index data;

[0111] Based on the monitoring attribute description and the quality inspection attribute description, a traceability description is defined, and a monitoring identification tag is generated in combination with the corresponding node of the blockchain network, and the monitoring identification tag is stored in the blockchain network in the form of a hash value;

[0112] For example: Assume that there are several processes on the production line, including process 1, process 2 and process 3; each product will be set with a unique monitoring identification tag during the production process and bound to the product's monitoring database to ensure the data is tamper-proof and traceable; for process 1, assuming that it involves product welding, the traceability description is = {monitoring attribute description, quality inspection attribute description}, then the monitoring attribute description can be {welding temperature, welding current, welding time}, then the quality inspection attribute description can be {welding joint strength, welding appearance, corresponding quality assessment value}, when entering process 2 and process 3, new monitoring indicator data and quality inspection indicator data will be generated, and new monitoring attribute descriptions and quality inspection attribute descriptions will be generated accordingly. By combining the corresponding production nodes and quality inspection nodes, a hash value is generated and marked on the corresponding blockchain for storage with a monitoring identification tag;

[0113] Display Feedback Module:

[0114] Used to provide feedback and display on the full life cycle monitoring platform.

[0115] In summary of the above technical solutions: the present invention includes a network modeling module, a life cycle management module and a display feedback module, and the life cycle management module includes a process monitoring unit, a quality inspection judgment unit and a traceability unit; its technical points are: by building a full life cycle monitoring platform, by combining blockchain technology and a monitoring knowledge base, a blockchain network is built with production nodes, quality inspection nodes and buffer nodes, and corresponding calling interfaces are provided between nodes, and quality assessment analysis is performed based on the judgment function and evaluation function in the process determination contract, and a first classification model and a second classification model are designed to accurately classify the feature analysis of each process, and whether the production line implements the defect management mechanism is determined based on the quality inspection planning contract; the data between the processes are seamlessly connected through a number of smart contracts, which facilitates the subsequent monitoring and tracing of the information of each process.

[0116] The size of the interval and threshold is set for the convenience of comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technicians in this field for each group of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value;

[0117] In the application, the several formulas involved are all calculated by removing dimensions and taking their numerical values, and the formula is a formula obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The formula is set by technical personnel in this field according to actual conditions.

[0118] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented using software, the above embodiments may be implemented in whole or in part in the form of a computer program product. A person of ordinary skill in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution.

[0119] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0120] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A full life cycle quality monitoring system, characterized in that: include: The network modeling module is used to build a full life cycle monitoring platform. The full life cycle monitoring platform has a preset blockchain network and monitoring knowledge base. Several smart contracts are established on the blockchain network. The smart contracts include process determination contracts, quality inspection planning contracts, and smart interaction contracts. The blockchain network consists of several production nodes, several quality inspection nodes, and a buffer node. The life cycle management module has a built-in process monitoring unit, a quality inspection and determination unit, and a traceability unit, which is used to analyze the life cycle information of the entire production line based on the full life cycle monitoring platform. The life cycle information includes the process data, monitoring indicator data, and quality inspection indicator data of the entire production line; The process monitoring unit is used to perform quality assessment analysis based on the judgment function and evaluation function pre-configured in the process determination contract, wherein the judgment function extracts the combined features of each process based on the life cycle information, and the evaluation function obtains the quality assessment value based on the combined feature analysis of each process, and returns the quality assessment value to the blockchain network; The process of extracting characteristic information of each process includes: Pre-construct a first classification model based on the XGBoost algorithm, substitute the collected life cycle information into the first classification model, select features for feature combination, construct new features, and select the required features through feature selection methods; introduce a training layer into the first classification model, combine the selected features with the original features to form a new training data set, and provide randomly selected sample data from the collected life cycle information, train and update the training layer, use a decision tree to evaluate the importance of the corresponding features, and obtain a second classification model; The acquired life cycle information is stratified according to the time series, and any time series is marked as a process. The life cycle information under each time series is substituted into the second classification model to determine the first probability that the combined features extracted under each process belong to the useful features and the second probability that the combined features are related features. Among them, the quality assessment value is obtained based on the combined characteristics analysis of each process, including: Define the combined feature as a key factor, and the key factor includes several dynamic feature indicators, mark any dynamic feature indicator as i, and the data value marking any dynamic feature indicator i is Li, call the standard interval St[SQ1, SQ2] of the dynamic feature indicator i from the monitoring knowledge base, and the standard interval St corresponds to the dynamic feature indicator i one by one; calculate the difference coefficient of the corresponding combined feature under each process based on the standard interval St, and assign the corresponding weight coefficient based on the difference coefficient; The calculation formula of the quality assessment value is: ; In the formula, quality represents the quality evaluation value, cvi represents the difference coefficient calculated by a certain dynamic feature index i, βi represents the weight coefficient assigned to a certain dynamic feature index i, and βi is greater than 0; The quality inspection judgment unit is used to judge whether the corresponding production line executes the defect management mechanism based on the quality inspection planning contract, wherein the defect management mechanism triggers an operation after receiving the quality assessment value: judging the state event based on the quality assessment value, when it is judged as the first state event, continuing to execute the quality inspection abnormality monitoring strategy; when it is judged as the second state event, evaluating the abnormality degree of the corresponding process, producing the abnormality assessment coefficient, and executing the corresponding strategy; The abnormality degree assessment is performed on the second state event, and the specific process is: obtaining the parameter data set for the assessment, including the quality assessment value, the time difference between the time when the second state event occurs and the current time, and the risk degree; The formula for calculating the abnormal assessment coefficient is: ; Where PG represents the abnormal assessment coefficient, fx represents the risk level, cz represents the time difference between the time when the second state event occurs and the current time, and Φ0 is the time constant; The acquisition of risk level includes: The combined features of any process are integrated and marked as a set J = {a1, a2, ..., a N0 }; Among them, a1, a2 and a N0 represents the useful features or relevant features corresponding to the process, and N0 represents the number of useful features or relevant features; draws the difference change curve of each feature within the preset acquisition period, and obtains the maximum peak value and the minimum trough value in each difference change curve, and determines the span value based on the difference between the maximum peak value and the minimum trough value; Obtain the maximum value of the corresponding span value in the process in the historical period and mark it as the maximum span value. Count the number of times the span value exceeds the maximum span value and mark it as the risk number. At the same time, calculate the difference between the span value and the corresponding maximum span value and mark it as the risk difference. Draw a two-dimensional coordinate system, visualize the risk difference in the two-dimensional coordinate system, and obtain the risk difference change curve within the preset collection period; call the risk difference change threshold curve of the monitoring knowledge base and draw it in the two-dimensional coordinate system, obtain the angle value of the first intersection of the two curves, and mark it as the risk angle; generate the risk degree by combining the risk angle and the number of risks, based on the formula: ; In the formula, cs represents the number of risks, jd represents the risk angle, α0 represents the weight correction coefficient, and α0 is greater than 0; The traceability unit is used to build information traceability relationships based on smart interactive contracts, and each information traceability relationship corresponds to a unique monitoring identification tag, and provides a traceability system for the corresponding monitoring process and monitoring progress; Display feedback module, used for feedback and display on the full life cycle monitoring platform.

2. A full life cycle quality monitoring system according to claim 1, characterized in that: The production node is used to store production data, including: identifier, process ID, process number, process name, production progress data, flow order data and resource status data; The quality inspection node is used to store quality inspection data, including: identifier, quality inspection ID, quality inspection number, quality inspection name, quality inspection progress data, process quality data and quality description data; The buffer node is used to store data to be updated and modified, and to perform process determination work before node storage.

3. A full life cycle quality monitoring system according to claim 2, characterized in that: The smart interactive contract is pre-configured with a data storage structure, a data storage function, an associated access rule, and an associated access function, stores the corresponding data under the production node, the quality inspection node, and the buffer node, and provides a call interface corresponding to the production node, the quality inspection node, and the buffer node; the process determination contract is pre-configured with a judgment function and an evaluation function, and provides a call interface between the smart interactive contract and the monitoring knowledge base; The quality inspection planning contract is pre-configured with status determination events and defect management mechanisms, and provides a calling interface between the smart interactive contract and the monitoring knowledge base; Draw a two-dimensional coordinate system, monitor the calling interface, regularly obtain the interface parameters within the preset monitoring time period, and visualize the interface parameters in the two-dimensional coordinate system to obtain the parameter change curve within the preset monitoring time period; A standard change curve is drawn in the two-dimensional coordinate system, and the parameter change curve is overlapped with the standard change curve. Based on the overlapping result, the abnormal parameter change curve segment is obtained, the abnormal state of the calling interface of the blockchain network is identified, and the preset automatic correction strategy is executed accordingly; among which, the interface parameters include calling frequency, parameter verification and response time.

4. A full life cycle quality monitoring system according to claim 1, characterized in that: The first probability and the second probability are obtained by analyzing and processing based on the information gain of the feature. When the first probability or the second probability is greater than the preset probability threshold, the combined feature is matched to the corresponding data label; otherwise, the proportion of the first probability and the second probability under the corresponding process is counted, and the weight parameters of the second classification model are adjusted based on the proportion, and retraining and classification are performed.

5. A full life cycle quality monitoring system according to claim 1, characterized in that: The determining whether the corresponding production line implements the defect management mechanism based on the quality inspection planning contract includes: Call the comparison threshold of the quality assessment value from the monitoring knowledge base, and compare and analyze the quality assessment value with the standard threshold: When the quality assessment value is greater than or equal to the comparison threshold, the quality assessment value is marked as q1, and edited as a first-level character, and q1 and the first-level character are combined to generate a first state event; an abnormal monitoring strategy is executed on the first state event, and it is temporarily stored; When the quality assessment value is less than the comparison threshold, the quality assessment value is marked as q2, edited as a secondary character at the same time, and q2 and the secondary character are combined to generate a second state event; The standard threshold of the abnormal evaluation coefficient is called from the monitoring knowledge base, and when the abnormal evaluation coefficient exceeds the standard threshold, an alarm signal is generated.

6. A full life cycle quality monitoring system according to claim 1, characterized in that: The information traceability relationship is constructed based on the smart interactive contract, including: Based on the calling interface between the smart interaction contract and the process monitoring contract, the monitoring attribute description corresponding to each process in the buffer node is obtained, and the monitoring attribute description corresponds to the quality inspection index data; based on the calling interface between the smart interaction contract and the quality inspection planning contract, the quality inspection attribute description corresponding to each process in the buffer node is obtained, and the quality inspection attribute description corresponds to the quality inspection index data; The traceability description is defined based on the monitoring attribute description and the quality inspection attribute description, and the monitoring identification tag is generated in combination with the corresponding node of the blockchain network, and the monitoring identification tag is stored in the blockchain network in the form of a hash value.

Images