Food safety risk management method and system based on multi-link data

Abstract

The application discloses a food safety risk management method and system based on multi-link data, relates to the field of risk management, and comprises the following steps: determining a food production chain; connecting a supply management platform of a corresponding supplier, identifying a supply cross node and marking a cross food supply characteristic according to a predetermined cross judgment constraint, and generating a cross mark chain; reading quality inspection parameters of each production node, determining unqualified parameter types that do not meet preset quality inspection indexes; performing unqualified factor discrimination, if the discrimination result is a supply node factor, performing unqualified factor cross contamination risk analysis, generating a cross risk index, and if the cross risk index is greater than a preset threshold, predicting a cross contamination site according to the cross mark chain, and sending the prediction result to a food safety management end. The application solves the technical problem that existing food safety risk management cannot comprehensively and accurately identify potential food safety risks, effectively prevents and controls food safety risks, and achieves the technical effect of guaranteeing food quality and safety.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of risk management, in particular to a food safety risk management method and system based on multi-link data. BACKGROUND

[0002] At present, in the food safety risk management, a single-link-based data monitoring and analysis method is mainly adopted, for example, only the production link or the supply link is independently monitored. However, the current method lacks in-depth mining and analysis of the data correlation of each link in the food production chain due to only focusing on a single link, which makes it difficult to comprehensively and accurately identify potential food safety risks, especially complex risks such as cross-contamination.

[0003] At present, in the related art, the food safety risk management has the technical problem of being difficult to comprehensively and accurately identify potential food safety risks. SUMMARY

[0004] The present application provides a food safety risk management method and system based on multi-link data, which determines the production chain of a specific food, covers the raw material supply chain and the production node chain after entering the production workshop, connects the raw material supply chain with the corresponding supplier's supply management platform according to the predetermined cross-determination constraint, identifies the supply cross-node and marks the cross-food supply characteristics, and then generates a cross-marking chain. The quality inspection parameters of each production node are obtained from the production node chain, and the unqualified parameter types that do not meet the preset quality inspection indicators are found out. For the unqualified parameter types, it is determined whether it is a production node factor or a supply node factor. If it is a supply node factor, cross-contamination risk analysis of unqualified factors is carried out, cross-risk indicators are generated, and if the cross-risk indicators exceed the preset threshold, the cross-contamination place is predicted according to the cross-marking chain, and the prediction result is sent to the food safety management end for reminding. The technical means solve the technical problem of the existing food safety risk management that it is difficult to comprehensively and accurately identify potential food safety risks, and achieve the technical effects of effectively preventing and controlling food safety risks and protecting food quality and safety.

[0005] The application provides a food safety risk management method based on multi-link data, comprising: determining a first production chain of a first food, including a first raw material supply chain and a first production node chain after entering a production workshop; based on the first raw material supply chain, connecting a supply management platform of a corresponding supplier, identifying a supply cross node and marking a cross food supply characteristic according to a predetermined cross judgment constraint, and generating a first cross marking chain; based on the first production node chain, reading quality inspection parameters of each production node, and determining a first unqualified parameter type that does not meet a preset quality inspection index; based on the first unqualified parameter type, discriminating unqualified factors of the production node and the supply node, if the discrimination result is a supply node factor, performing cross contamination risk analysis of the unqualified factors, and generating a first cross risk index; if the first cross risk index is greater than a preset threshold, predicting a cross contamination site according to the first cross marking chain, and sending a prediction result to a food safety management end for reminding.

[0006] In a possible implementation manner, after the first cross marking chain is generated, the following processing is further performed: determining cross food attribute information based on the first cross marking chain; generating a food monitoring library with the cross food attribute information, issuing a safety reminding signal when any cross food in the food monitoring library is monitored to have a safety risk; and performing sampling detection of each batch of the first food based on the safety reminding signal, generating a sampling detection result for risk self-check reminding.

[0007] In a possible implementation manner, based on the first raw material supply chain, the supply management platform of the corresponding supplier is connected, the supply cross node identification and the cross food supply characteristic marking are performed according to the predetermined cross judgment constraint, and the first cross marking chain is generated, and the following processing is performed: extracting each transfer node in the first raw material supply chain; based on the supply management platform, performing supply cross node identification at each transfer node according to the predetermined cross judgment constraint; marking the supply cross node in the first raw material supply chain, and marking the cross food supply characteristic corresponding to the supply cross node, including the food type and the supply path, to generate the first cross marking chain.

[0008] In a possible implementation manner, the following processing is performed: the predetermined cross judgment constraint is used to judge whether raw materials or food exist a cross contamination path, and specifically includes equipment sharing and closed warehouse space sharing characteristics.

[0009] In a possible implementation, based on the first unqualified parameter type, the unqualified factor discrimination of the production node and the supply node is performed, and the following processing is performed: determining an unqualified detection node of the first unqualified parameter type; calling an abnormality traceability network based on the unqualified detection node to trace the unqualified factor, and if the traceability result is a single factor in the production node and the supply node, generating the discrimination result with the single factor; wherein the abnormality traceability network includes each abnormality traceability layer corresponding to each production node, and each abnormality traceability layer is constructed based on a historical unqualified parameter type set and a corresponding historical traceability node.

[0010] In a possible implementation, based on the unqualified detection node, the abnormality traceability network is called to trace the unqualified factor, and the following processing is further performed: if the traceability result is a double factor coexisting in the production node and the supply node; extracting a production node factor in the double factor, performing self-production parameter abnormality verification and quality inspection verification of the production food in the corresponding production node, and generating a production verification result; performing single factor reservation constraint on the double factor with the production verification result, and generating the discrimination result with the reservation result.

[0011] In a possible implementation, if the discrimination result is a supply node factor, unqualified factor cross-contamination risk analysis is performed, a first cross-contamination risk index is generated, and the following processing is performed: determining a quality inspection molecular attribute in the first unqualified parameter type; determining whether the quality inspection molecular attribute belongs to a preset food diffusion contamination molecule, if yes, extracting a contamination diffusion risk feature; based on the supply node factor, supply cross feature extraction is performed in the first cross-contamination marker chain, and the first cross-contamination risk index is determined by comparing the supply cross feature extraction with the contamination diffusion risk feature.

[0012] In a possible implementation, according to the first cross-contamination marker chain, cross-contamination site prediction is performed, the prediction result is sent to a food safety management end for reminding, and the following processing is performed: mapping the supply node factor in the first cross-contamination marker chain, determining whether the mapping node is a marked supply cross node, if yes, extracting a cross food supply feature marker of the corresponding node, performing cross food supply chain extraction and analysis, and generating the prediction result.

[0013] In a possible implementation, based on the first unqualified parameter type, the unqualified factor discrimination of the production node and the supply node is further performed, and the following processing is performed: if the discrimination result is a production node factor, a production abnormality reminding signal is sent; based on the production abnormality reminding signal, production suspension control is performed, and the first unqualified parameter type is input into a production parameter optimization library for optimization; production restart control is performed according to the optimized parameters, and production batch identification early warning during the abnormal period is performed.

[0014] The application also provides a food safety risk management system based on multi-link data, comprising: a first production chain determination module, configured to determine a first production chain of a first food, including a first raw material supply chain and a first production node chain after entering a production workshop; a first cross-label chain generation module, configured to connect a supply management platform of a corresponding supplier based on the first raw material supply chain, identify a supply cross node and mark a cross food supply characteristic label according to a predetermined cross judgment constraint, and generate a first cross-label chain; a quality inspection parameter analysis module, configured to read quality inspection parameters of each production node based on the first production node chain, and determine a first unqualified parameter type that does not meet a preset quality inspection index; an unqualified factor cross-contamination risk analysis module, configured to identify unqualified factors of a production node and a supply node based on the first unqualified parameter type, if the identification result is a supply node factor, perform unqualified factor cross-contamination risk analysis, and generate a first cross-risk index; and a cross-contamination site prediction module, configured to, if the first cross-risk index is greater than a preset threshold, perform cross-contamination site prediction according to the first cross-label chain, and send a prediction result to a food safety management end for reminding.

[0015] The food safety risk management method and system based on multi-link data provided in the application first determine a first production chain of a first food, including a first raw material supply chain and a first production node chain after entering a production workshop, then connect a supply management platform of a corresponding supplier based on the first raw material supply chain, identify a supply cross node and mark a cross food supply characteristic label according to a predetermined cross judgment constraint, generate a first cross-label chain, then read quality inspection parameters of each production node based on the first production node chain, determine a first unqualified parameter type that does not meet a preset quality inspection index, then identify unqualified factors of a production node and a supply node based on the first unqualified parameter type, if the identification result is a supply node factor, perform unqualified factor cross-contamination risk analysis, generate a first cross-risk index, and finally, if the first cross-risk index is greater than a preset threshold, perform cross-contamination site prediction according to the first cross-label chain, and send a prediction result to a food safety management end for reminding. The technical effect of effectively preventing and controlling food safety risks and guaranteeing food quality and safety is achieved. BRIEF DESCRIPTION OF DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the application, the drawings of the embodiments of the application will be briefly introduced as follows. In the present application, a flowchart is used to illustrate the operations performed by the system according to the embodiments of the application. It should be understood that the foregoing or the following operations are not necessarily performed in sequence. On the contrary, various steps can be processed in reverse order or simultaneously according to needs. Meanwhile, other operations can be added to these processes, or a step or several steps of operations can be removed from these processes.

[0017] Figure 1 A flowchart of a food safety risk management method based on multi-link data provided by an embodiment of the present application is shown.

[0018] Figure 2 A structural diagram of a food safety risk management system based on multi-link data provided by an embodiment of the present application is shown.

[0019] Legend: first production chain determination module 10, first cross-link generation module 20, quality inspection parameter analysis module 30, unqualified factor cross-contamination risk analysis module 40, cross-contamination site prediction module 50. DETAILED DESCRIPTION

[0020] The above description is only a summary of the technical solutions of the present application. In order to make the technical means of the present application more clear, the following specific embodiments of the present application are described in accordance with the content of the description, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application are described.

[0021] In order to make the purposes, technical solutions and advantages of the present application more clear, the following will combine the drawings to make a further detailed description of the present application. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by those skilled in the art without making creative efforts fall within the scope of protection of the present application.

[0022] In the following description, "some embodiments" are described, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subset of all possible embodiments, and can be combined with each other without conflict. The term "first\second" is only to distinguish similar objects, and does not represent a specific order of the objects. The terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those steps or units clearly listed, but can include other steps or modules not clearly listed or inherent to these processes, methods, products or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as understood by those skilled in the art of the technology to which the present application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application.

[0023] The embodiments of the present application provide a food safety risk management method based on multi-link data, as shown in Figure 1 The method comprises the following steps:

[0024] Step S100, determining a first production chain of a first food, including a first raw material supply chain and a first production node chain after entering a production workshop.

[0025] Specifically, the first production chain refers to the entire process of food production from raw material procurement to final product production, including the raw material supply chain and the production node chain. The first raw material supply chain refers to all links from the origin of raw materials to the production workshop warehouse, including picking, transportation, storage, etc. The first production node chain refers to the sequential connection of various processing steps of food in the production workshop.

[0026] Through IoT (Internet of Things) devices and sensors, data from each node in the food production chain is collected, including the origin of raw materials, transportation tools (trucks), storage information (warehouses), and equipment information in the production workshop. A central database is established to store and manage the data collected from each node. For example, a relational database (such as MySQL) or a non-relational database (such as MongoDB) is used to store structured and semi-structured data. Using data mapping technology, data from different sources is associated to form a complete production chain data model. For example, the origin of raw materials, transportation tools, and storage information are associated through a unified identifier (such as batch number, two-dimensional code).

[0027] For example, sensors are installed at the origin of raw materials to record the picking time, location, and quality indicators of raw materials; GPS and temperature sensors are installed in the transportation truck to record the location and temperature changes during transportation; RFID readers are installed in the warehouse to record the in-out information of goods. The collected data is stored in the central database, as shown in Table 1.

[0028] Table 1: Database storage example

[0029]

[0030] Step S200, based on the first raw material supply chain, connecting the supply management platform of the corresponding supplier, identifying the supply cross-node and marking the cross-food supply characteristics according to the predetermined cross-judgment constraint, and generating the first cross-mark chain.

[0031] Specifically, the connection with the supplier's supply management platform is realized through the API interface (Application Programming Interface). The API interface allows data exchange between systems to obtain information such as the supplier's storage, transportation, and equipment usage. For example, an API interface using the HTTP protocol is used to obtain data from the supplier's management system through a URL request. Assuming that the API interface of supplier A is https:1***, the warehouse information of supplier A can be obtained through a GET request: https:2***. The returned data can be a JSON format list containing warehouse ID, location, usage time, etc.

[0032] The data from the supplier management platform is synchronized into a central database using a data synchronization tool, such as an ETL tool (Extract-Transform-Load). The ETL tool can extract data from the supplier platform on a regular basis, perform necessary transformations and cleaning, and load it into the local database. For example, Apache NiFi can be used as an ETL tool to extract data from the supplier API interface on a regular basis, and store it in a unified format in the local database after transformation.

[0033] Cross-determination rules are defined to identify nodes such as warehouses, trucks, or equipment that are shared by different suppliers in the raw material supply chain. The rules can be based on conditions such as time overlap, spatial location, etc. For example, if two suppliers use the same warehouse within the same time period, the warehouse is considered a supply cross-node. Using database queries and data mining techniques, supply cross-nodes are identified according to the cross-determination rules. For example, the identification of supply cross-nodes can be implemented through SQL queries or Python scripts. For each identified supply cross-node, extract the food supply characteristics related to the node. These characteristics include: food batch information (such as batch number, production date, shelf life, etc.), quality indicators (such as pesticide residue, microbial content, etc.), source information (such as supplier ID, raw material origin, etc.). A unique marker ID is generated for each supply cross-node and its related food supply characteristics, and these information is stored in the database. The marker ID can contain node information and food batch information for quick traceability. All cross markers are connected in chronological or logical order to form a cross marker chain, which is used for risk analysis and prediction.

[0034] In one possible implementation, after generating the first cross marker chain, the method further includes: determining cross food attribute information based on the first cross marker chain; generating a food monitoring library with the cross food attribute information, and issuing a safety reminder signal when any cross food in the food monitoring library is detected to have a safety risk; performing sampling detection of each batch of the first food based on the safety reminder signal, and generating a sampling detection result for risk self-checking reminder.

[0035] Specifically, the relevant attribute information of the cross-food is extracted from the first cross-label chain, including but not limited to food batch number, production date, supplier ID, raw material origin, quality indicators (such as pesticide residue, microbial content, etc.). The extracted cross-food attribute information is stored in a special food monitoring library, which can be a database table or a data file, for real-time monitoring of the safety status of the food. Natural language processing (NLP) technology and data crawling technology are used to collect information related to cross-food from social media, news websites, etc. For example, user feedback and news reports on platforms such as Weibo, WeChat public accounts, etc. are obtained through crawling. A two-way risk monitoring system is developed to monitor the safety status of the company's own food and the safety status of other companies' food. If problems are found in other companies' food, the cross-chain is used to determine the food batches of the company that may be affected. At the same time, if problems are found in the company's own food, the cross-chain is used to determine the food batches of other companies that may be affected. When monitoring other companies' food safety risks, a message push system (such as SMS, email, App push) is used to send safety warning signals to relevant departments and personnel. For example, use the enterprise WeChat API to send reminder messages. According to the safety warning signal, the sampling detection process is automatically triggered, and the system will select the relevant batches of food for sampling detection according to the information in the cross-label chain. Through sampling detection, the safety status of the food is accurately determined to ensure food safety. This implementation method can timely discover potential safety risks by monitoring social information and We Media data, take measures in advance to ensure that problems can be handled in a timely manner, thereby effectively improving the efficiency and accuracy of food safety risk management and enhancing the risk prevention and control capability of enterprises.

[0036] In one possible implementation, based on the first raw material supply chain, the supply management platform of the corresponding supplier is connected, the supply cross-node is identified and the cross-food supply feature label is marked according to the predetermined cross-determination constraint, and the first cross-label chain is generated. Step S200 further includes step S210 of extracting each transfer node in the first raw material supply chain. Specifically, by connecting with the supply management platform of the supplier, detailed information in the first raw material supply chain is obtained, including each transfer node (such as warehouse, truck, transfer station, etc.). The data provided by the supplier management platform is analyzed, and the relevant information of all transfer nodes is extracted, such as node ID, location, use time, etc.

[0037] Step S220, based on the supply management platform, the various transfer nodes identify supply cross-node based on predetermined cross-judgment constraints, which are used to determine whether there is a cross-contamination path for raw materials or food, including equipment sharing and closed warehouse space sharing features. Specifically, rules are defined for determining whether there is a cross-contamination path for raw materials or food. These rules include equipment sharing and closed warehouse space sharing features. According to the extracted transfer node information, combined with the cross-judgment constraints, the nodes with cross-contamination paths are identified.

[0038] Step S230, the supply cross-node in the first raw material supply chain is marked, and the corresponding cross-food supply features of the supply cross-node are marked, including food type and supply path, to generate the first cross-label chain. Specifically, a unique cross-label ID is generated for each identified supply cross-node, and it is associated with the relevant transfer node information. At the same time, mark the cross-food supply features related to the supply cross-node, including food type, supply path, etc. The supply cross-node and the related cross-food supply features are integrated together to form a complete cross-label chain for traceability and management. This implementation can effectively identify potential cross-contamination paths and issue early warning signals by extracting transfer node information, identifying supply cross-nodes and marking related features.

[0039] Step S300, based on the first production node chain, read the quality inspection parameters of each production node, and determine the first unqualified parameter type that does not meet the preset quality inspection index.

[0040] Specifically, the quality inspection parameters of the production node are read in real time by automatic detection equipment (such as spectrometer, chromatograph), and the data is transmitted to the central database. Among them, the quality inspection parameters refer to various indicators for detecting the quality of food in the production process, such as pesticide residue, microbial content, etc. Using the preset quality inspection index, the data comparison algorithm (such as threshold comparison) is used to judge whether the parameter is qualified. For example, Python script is used to realize the threshold comparison function. The parameter type that does not meet the preset quality inspection index is recorded, for example, a spectrometer is installed on the processing equipment in the production workshop to detect the pesticide residue of raw materials in real time. The detected pesticide residue is compared with the preset threshold (such as 0.1mg / kg), and it is found that the pesticide residue of a batch of raw materials is 0.2mg / kg, which exceeds the threshold. The unqualified parameter type is recorded as "pesticide residue exceeds the standard".

[0041] Step S400, based on the first unqualified parameter type, the unqualified factors of the production node and the supply node are distinguished, if the result of the discrimination is the supply node factor, the unqualified factor cross-contamination risk analysis is carried out, and the first cross-risk index is generated.

[0042] Specifically, a machine learning algorithm (such as decision tree, support vector machine) is used to identify the unqualified factors. For example, a decision tree model is trained according to historical data, inputting unqualified parameter types and outputting possible unqualified factors (such as supplier problems, transportation problems, etc.). If the judgment result shows that it is a supply node factor, the risk assessment model (such as Bayesian network) is used to analyze the cross-node pollution risk of the supply. For example, according to the use frequency and common time of the cross-node, the cross-contamination risk index is calculated. The calculated risk index is stored in the database and associated with the cross-link chain.

[0043] For example, input the unqualified parameter type "excessive pesticide residues", and the decision tree model outputs the unqualified factor "supplier problem". According to the common time and use frequency of warehouse 003 in the cross-link chain, the cross-risk index is calculated as 0.8 (full score is 1, indicating high risk). The cross-risk index 0.8 is stored and associated with warehouse 003 in the cross-link chain.

[0044] In one possible implementation, the unqualified factor identification of the production node and the supply node based on the first unqualified parameter type further includes the step S410 of determining the unqualified detection node of the first unqualified parameter type. Specifically, when the unqualified parameter type is detected, it is first determined in which specific production node the unqualified parameter type is detected, and this step is realized by recording the detection results of each detection point. For example, if a batch of food is detected to have excessive pesticide residues at the packaging link of the production workshop, the packaging link is the unqualified detection node.

[0045] At step S420, the unqualified factor is traced based on the unqualified detection node calling an abnormality tracing network, and if the tracing result is a single factor in the production node and the supply node, the discrimination result is generated based on the single factor. The abnormality tracing network includes each abnormality tracing layer corresponding to each production node, and each abnormality tracing layer is constructed based on a historical unqualified parameter type set and corresponding historical tracing nodes. Specifically, the abnormality tracing network is a network constructed based on historical data, which is used to trace the source of unqualified factors. It includes abnormality tracing layers corresponding to each production node, and each layer is constructed based on a historical unqualified parameter type set and corresponding historical tracing nodes. The abnormality tracing network is called from the unqualified detection node. By analyzing historical data, historical cases similar to the current unqualified parameter type are found, and the tracing results of these cases are found. If the tracing result shows that the unqualified factor comes from only one of the production node or the supply node, the discrimination result is generated based on the single factor. If the tracing result shows that the unqualified factor may involve multiple nodes, the contribution degree of each node needs to be further analyzed, and a comprehensive discrimination result is generated. This implementation realizes accurate tracing and rapid discrimination of unqualified factors by introducing an abnormality tracing network, and improves the efficiency and accuracy of food safety risk management.

[0046] In a possible implementation, the unqualified factor is traced based on the unqualified detection node calling an abnormality tracing network, and step S400 further includes step S430. If the tracing result is a double factor of the coexistence of the production node and the supply node, the production node factor in the double factor is extracted, the corresponding production node is verified for abnormality of its own production parameter and quality inspection of the produced food, a production verification result is generated, the double factor is constrained as a single factor based on the production verification result, and the discrimination result is generated based on the result.

[0047] Specifically, when the traceability result shows that the unqualified factor involves a double factor coexisting in the production node and the supply node, the further verification and analysis steps are as follows: extract the factor involving the production node from the double factor, that is, focus on the specific parameters in the production link that may affect the quality of the food, such as the operating parameters of the production equipment, the production process, the environmental conditions, etc. In the production node determined in the traceability result, the parameters in the production process are verified abnormally, including checking whether the operating data of the production equipment, the process parameters conform to the standard operating procedure (SOP), and whether the environmental monitoring data is normal. For example, if the traceability result shows that the problem may originate from the processing link in the production node chain, and the problem is found in the warehousing link (the unqualified detection node), the production parameters of the processing link need to be verified. The food produced by the corresponding production node is re-inspected and verified, including sampling from the corresponding production node, re-detecting the relevant quality indicators, and confirming whether there are unqualified parameters. According to the above verification process, a production verification result is generated. The verification result clearly indicates whether there are abnormalities in the corresponding production node, and whether these abnormalities are directly related to the unqualified parameter type. According to the production verification result, a single factor reservation constraint is performed on the double factor, that is, if the production verification result shows that there are indeed abnormalities in the corresponding production node, and these abnormalities are the direct cause of the unqualified parameter type, the production node factor is reserved and the supply node factor is removed. For example, if the production verification result shows that the abnormal equipment parameters of the processing link cause the food to be unqualified, it can be determined that the production node is the main cause, and thus the supply node factor is removed. Finally, according to the result of the single factor reservation constraint, a final discrimination result is generated. This result clearly indicates whether the specific factor causing the unqualified parameter type is the production node or the supply node. This implementation can avoid unnecessary measures on the supply node due to misjudgment, reduce interference with the supply chain, and avoid resource waste due to false judgment. That is, through the systematic verification and constraint process, the efficiency and accuracy of risk management are improved, and the delay and cost increase caused by uncertain factors are reduced.

[0048] In a possible implementation, if the discrimination result is the supply node factor, an unqualified factor cross-contamination risk analysis is performed to generate a first cross-risk indicator, and step S400 further includes step S440 of determining a quality inspection molecular attribute in the first unqualified parameter type. Specifically, when the unqualified parameter type is detected, the specific properties of the parameter are further analyzed to determine which type of quality inspection molecular attribute it belongs to. For example, is it microbial contamination (such as excessive coliform bacteria), chemical contamination (such as excessive pesticide residues), or unqualified nutritional parameters of raw materials (such as insufficient protein content). This step can be completed through detailed analysis of the detection data, such as through spectral analysis, chromatographic analysis, or other laboratory detection means to obtain specific molecular properties.

[0049] Step S450, determine whether the quality inspection molecular attribute belongs to a preset food diffusion contamination molecule, if yes, extract the contamination diffusion risk characteristics. Specifically, based on big data and historical data, a preset food diffusion contamination molecule database is constructed. This database contains molecules that can cause cross-contamination and their diffusion characteristics, such as diffusion level risk under common contact tools or closed space. Determine whether the detected quality inspection molecular attribute belongs to the contamination molecules in the preset database. If yes, extract the contamination diffusion risk characteristics of the molecule. For example, if the detected unqualified parameter type is microbial contamination (such as E. coli), and this microorganism is marked as having high diffusion risk in the preset database, then extract its diffusion characteristics, such as survival time in a closed space, propagation speed, etc.

[0050] Step S460, based on the supply node factors, extract supply cross characteristics in the first cross label chain, compare with the contamination diffusion risk characteristics, and determine the first cross risk index. Specifically, in the first cross label chain, extract the supply cross characteristics related to the supply node factors. These characteristics include shared warehouses, trucks, equipment, etc. Compare the extracted supply cross characteristics with the contamination diffusion risk characteristics to determine the risk level of cross contamination. For example, if the supply cross characteristics involve shared warehouses, and the environmental conditions (such as humidity, temperature) in the warehouse meet the conditions for microbial diffusion, then according to the comparison result, a higher cross risk index is generated. This implementation can accurately assess the risk level of cross contamination by analyzing the quality inspection molecular attribute in detail and comparing it with the preset contamination molecule database, avoiding a one-size-fits-all approach to all unqualified factors, and improving the accuracy of risk assessment. At the same time, the food diffusion contamination molecule database can be updated at any time according to actual conditions, ensuring the timeliness and accuracy of the evaluation results.

[0051] In one possible implementation, based on the first unqualified parameter type, the unqualified factors of the production node and the supply node are distinguished, and the method further comprises: if the result of the discrimination is the production node factor, issuing a production abnormality warning signal; based on the production abnormality warning signal, controlling the production to be suspended, inputting the first unqualified parameter type into the production parameter optimization library for optimization, restarting the production according to the optimized parameters, and warning about the production batch identification during the abnormal period.

[0052] Specifically, when the discrimination result is the production node factor, the system will immediately issue a production abnormality warning signal. This signal can be in the form of an alarm, a message, an email, or a system notification, notifying production managers and relevant technical personnel. For example, the system can send a message through WeChat Enterprise or DingTalk: "Production abnormality warning: detected unqualified parameters [parameter type] in production node [specific node], please check immediately."

[0053] Upon receiving the production abnormality alert signal, the system automatically triggers production suspension control. This can be achieved by integrating with the control system of the production equipment, such as suspending the relevant production line through PLC (Programmable Logic Controller) or MES (Manufacturing Execution System). For example, the system can send an instruction to the production equipment: "suspend production line [production line ID], wait for further inspection."

[0054] The production parameter optimization library is a system based on historical data and optimization algorithms for analyzing and optimizing production parameters. By inputting the first unqualified parameter type into this library, the system can propose improvement measures based on historical data and optimization models. For example, if the unqualified parameter type is excessive microbial content, the production parameter optimization library may suggest increasing disinfection frequency or adjusting equipment operating parameters.

[0055] After completing the adjustment of optimized parameters, the system will restart production according to the optimized parameters. This can be achieved through an automated control system to ensure that the production process runs according to the optimized parameters. For example, the system can send an instruction to the production equipment: "restart production line [production line ID] according to optimized parameters [parameter list]." For batches produced during the abnormal period, the system will be identified and warned, which can be achieved by marking these batches in the production management system and paying special attention to them in subsequent quality detection and sales links. For example, the system can mark in the production management system: "batch [batch number] was produced during the abnormal period and requires additional detection." This implementation method can quickly respond to problems in production by immediately issuing a production abnormality alert signal and suspending production, avoiding the production of unqualified products and reducing losses. By inputting the unqualified parameter type into the production parameter optimization library, optimization suggestions can be made based on data analysis to improve the quality control level of the production process and reduce the recurrence of similar problems. By identifying and warning batches produced during the abnormal period, special attention can be paid to these batches in subsequent links to ensure that unqualified products do not enter the market, ensuring consumer safety.

[0056] Step S500, if the first cross-risk index is greater than the preset threshold, cross-contamination site prediction is performed according to the first cross-marking chain, and the prediction result is sent to the food safety management end for warning.

[0057] Specifically, a preset threshold (for example, the value range of the cross-risk index is 0 to 1, and the preset threshold is 0.8) is set to determine whether the current cross-risk is within an acceptable range. If the first cross-risk index is greater than the preset threshold, it indicates that there is a high risk of cross-contamination, and further measures need to be taken.

[0058] The first cross-tag chain contains all possible cross-node information from the raw material supply chain to the production chain, as well as detailed characteristics of these nodes (such as shared warehouses, trucks, equipment, etc.). Cross-contamination site prediction is performed using historical data-based analysis, machine learning models, or rule engines, for example, using prediction algorithms (such as time series analysis, neural networks) to predict where cross-contamination is likely to occur based on the cross-tag chain and cross-risk indicators. The prediction results are sent to the food safety management end through a message push system (such as SMS, email, App push). For example, use the API interface of WeChat Enterprise or DingTalk to send reminder messages. The prediction results are displayed on the visualization interface of the food safety management end, for example, by marking the possible contamination sites on the map.

[0059] For example, according to the cross-tag chain and cross-risk indicators, the LSTM model predicts that warehouse 003 has a high cross-contamination risk in the next 24 hours. Send a reminder message through WeChat Enterprise: "Warning: Warehouse 003 has a high cross-contamination risk in the next 24 hours, please check in time." On the map interface of the food safety management end, mark the location of warehouse 003 with red and display the risk level and recommended measures.

[0060] In one possible implementation, according to the first cross-tag chain, the cross-contamination site prediction is performed, and the prediction results are sent to the food safety management end for reminders, step S500 further includes step S510, mapping the supply node factors in the first cross-tag chain, and determining whether the mapped node is a marked supply cross-node. Specifically, the identified supply node factors (e.g., a certain warehouse or truck) are compared with the nodes in the first cross-tag chain to determine whether the node is a marked supply cross-node. A marked supply cross-node is a node that has been marked as a potential cross-contamination risk in the first cross-tag chain, such as a shared warehouse or truck.

[0061] Step S520, if yes, extract the cross-food supply characteristic tags of the corresponding node, perform cross-food supply chain extraction and analysis, and generate the prediction results. Specifically, if the mapped node is a marked supply cross-node, the system will extract the cross-food supply characteristic tags of the node. These tags include food type, supply path, quality indicators, etc. Based on the extracted characteristic tags, further analyze the food supply chain information related to the node to generate detailed prediction results.

[0062] For example, suppose the following supply cross-node information is recorded in the first cross-label chain: node ID: WH001, type: warehouse, label ID: C001; node ID: TR002, type: truck, label ID: C002. If the identified supply node factor is warehouse WH001, the system will compare it with the nodes in the first cross-label chain and find that warehouse WH001 has been labeled as a supply cross-node (label ID: C001). Suppose the cross-food supply characteristics of warehouse WH001 are as follows: food type: fruit, supply path: origin A -> warehouse WH001 -> production workshop. The system will analyze these information and generate the prediction result: warehouse WH001 (label ID: C001) has cross-contamination risk, involving food type: fruit, supply path: origin A -> warehouse WH001 -> production workshop, and it is recommended to immediately check the warehouse hygiene condition and conduct sampling inspection on the related food batches.

[0063] The embodiment of the present application determines the production chain of a specific food, covering the raw material supply chain and the production node chain after entering the production workshop. According to the raw material supply chain, it is connected with the supply management platform of the corresponding supplier, identifies the supply cross-node according to the predetermined cross-determination constraint, labels the cross-food supply characteristics, and then generates a cross-label chain. The quality inspection parameters of each production node are obtained from the production node chain, and the unqualified parameter types that do not meet the preset quality inspection indicators are found out. For the unqualified parameter types, it is determined whether it is a production node factor or a supply node factor. If it is a supply node factor, the unqualified factor cross-contamination risk analysis is carried out, the cross-risk index is generated, and if the cross-risk index exceeds the preset threshold, the cross-contamination place is predicted according to the cross-label chain, and the prediction result is sent to the food safety management end for reminding and other technical means. The technical problem that the existing food safety risk management cannot comprehensively and accurately identify potential food safety risks is solved, and the technical effects of effectively preventing and controlling food safety risks and ensuring food quality and safety are achieved.

[0064] In the foregoing, the food safety risk management method based on multi-link data according to the embodiments of the present application is described in detail. Next, the food safety risk management system based on multi-link data according to the embodiments of the present application will be described with reference to the accompanying drawings. Figure 1 The food safety risk management system based on multi-link data according to the embodiments of the present application is described in detail. Next, the food safety risk management system based on multi-link data according to the embodiments of the present application will be described with reference to the accompanying drawings. Figure 2 The food safety risk management system based on multi-link data according to the embodiments of the present application is described in detail. Next, the food safety risk management system based on multi-link data according to the embodiments of the present application will be described with reference to the accompanying drawings.

[0065] The food safety risk management system based on multi-link data according to the embodiment of the present application is used to solve the technical problem that the existing food safety risk management cannot comprehensively and accurately identify potential food safety risks, and achieve the technical effect of effectively preventing and controlling food safety risks and protecting food quality and safety. The food safety risk management system based on multi-link data comprises a first production chain determination module 10, a first cross-marked chain generation module 20, a quality inspection parameter analysis module 30, an unqualified factor cross-contamination risk analysis module 40, and a cross-contamination site prediction module 50.

[0066] The first production chain determination module 10 is used to determine the first production chain of the first food, which comprises a first raw material supply chain and a first production node chain after entering a production workshop. The first cross-marked chain generation module 20 is used to connect the supply management platform of the corresponding supplier based on the first raw material supply chain, identify the supply cross-node and mark the cross-food supply characteristics according to the predetermined cross-determination constraint, and generate the first cross-marked chain. The quality inspection parameter analysis module 30 is used to read the quality inspection parameters of each production node based on the first production node chain, and determine the first unqualified parameter type that does not meet the preset quality inspection index. The unqualified factor cross-contamination risk analysis module 40 is used to determine the unqualified factors of the production node and the supply node based on the first unqualified parameter type. If the determination result is the supply node factor, the unqualified factor cross-contamination risk analysis is performed, and the first cross-risk index is generated. The cross-contamination site prediction module 50 is used to predict the cross-contamination site according to the first cross-marked chain if the first cross-risk index is greater than a preset threshold, and send the prediction result to the food safety management end for reminding.

[0067] After the first cross-marked chain is generated, the system can further comprise a cross-food attribute information determination module for determining the cross-food attribute information based on the first cross-marked chain, a safety reminder signal sending module for generating a food monitoring library with the cross-food attribute information, sending a safety reminder signal when any cross-food in the food monitoring library is monitored to have a safety risk, a sampling detection module for performing sampling detection of each batch of the first food based on the safety reminder signal, and generating a sampling detection result for risk self-checking reminding.

[0068] In the following, the specific configuration of the first cross-label chain generation module 20 will be described in detail. As described above, based on the first raw material supply chain, connecting the supply management platform of the corresponding supplier, identifying the supply cross-node and labeling the cross-food supply characteristics according to the predetermined cross-judgment constraint, the first cross-label chain generation module 20 can further include: a transfer node extraction unit for extracting each transfer node in the first raw material supply chain; a supply cross-node identification unit for identifying the supply cross-node at each transfer node based on the supply management platform according to the predetermined cross-judgment constraint; and a labeling unit for labeling the supply cross-node in the first raw material supply chain and labeling the cross-food supply characteristics corresponding to the supply cross-node, including food type and supply path, to generate the first cross-label chain.

[0069] In the above, the specific configuration of the first cross-label chain generation module 20 will be described in detail. As described above, based on the first raw material supply chain, connecting the supply management platform of the corresponding supplier, identifying the supply cross-node and labeling the cross-food supply characteristics according to the predetermined cross-judgment constraint, the first cross-label chain generation module 20 can further include: a transfer node extraction unit for extracting each transfer node in the first raw material supply chain; a supply cross-node identification unit for identifying the supply cross-node at each transfer node based on the supply management platform according to the predetermined cross-judgment constraint; and a labeling unit for labeling the supply cross-node in the first raw material supply chain and labeling the cross-food supply characteristics corresponding to the supply cross-node, including food type and supply path, to generate the first cross-label chain.

[0070] In the following, the specific configuration of the unqualified factor cross-contamination risk analysis module 40 will be described in detail. As described above, based on the first unqualified parameter type, the unqualified factors of the production node and the supply node are distinguished, and the unqualified factor cross-contamination risk analysis module 40 can further include: an unqualified detection node determination unit for determining the unqualified detection node of the first unqualified parameter type; and an unqualified factor tracing unit for calling an abnormal tracing network based on the unqualified detection node to trace the unqualified factor, if the tracing result is a single factor in the production node and the supply node, generating the discrimination result with the single factor, wherein the abnormal tracing network includes each abnormal tracing layer corresponding to each production node, and each abnormal tracing layer is constructed based on a set of historical unqualified parameter types and corresponding historical tracing nodes.

[0071] In the above, the specific configuration of the unqualified factor cross-contamination risk analysis module 40 will be described in detail. As described above, based on the first unqualified parameter type, the unqualified factors of the production node and the supply node are distinguished, and the unqualified factor cross-contamination risk analysis module 40 can further include: an unqualified detection node determination unit for determining the unqualified detection node of the first unqualified parameter type; and an unqualified factor tracing unit for calling an abnormal tracing network based on the unqualified detection node to trace the unqualified factor, if the tracing result is a single factor in the production node and the supply node, generating the discrimination result with the single factor, wherein the abnormal tracing network includes each abnormal tracing layer corresponding to each production node, and each abnormal tracing layer is constructed based on a set of historical unqualified parameter types and corresponding historical tracing nodes.

[0072] If the determination result is a supply node factor, the unqualified factor cross-contamination risk analysis is performed to generate a first cross risk index. The unqualified factor cross-contamination risk analysis module 40 can further include: a quality inspection molecule attribute determination unit configured to determine a quality inspection molecule attribute in the first unqualified parameter type; a pollution diffusion risk feature extraction unit configured to determine whether the quality inspection molecule attribute belongs to a preset food diffusion pollution molecule, and if yes, extract a pollution diffusion risk feature; and a first cross risk index determination unit configured to perform supply cross feature extraction on the supply node factor in the first cross marker chain, compare the pollution diffusion risk feature, and determine the first cross risk index.

[0073] In the following, the specific configuration of the cross-contamination site prediction module 50 will be described in detail. As described above, the cross-contamination site prediction is performed according to the first cross marker chain, and the prediction result is sent to the food safety management end for reminding. The cross-contamination site prediction module 50 can further include: a mapping determination unit configured to map the supply node factor in the first cross marker chain, and determine whether the mapping node is a marked supply cross node; and a cross food supply chain extraction and analysis unit configured to extract a cross food supply feature marker of a corresponding node, perform cross food supply chain extraction and analysis, and generate the prediction result if the mapping node is the marked supply cross node.

[0074] In the following, the specific configuration of the cross-contamination site prediction module 50 will be described in detail. As described above, the cross-contamination site prediction is performed according to the first cross marker chain, and the prediction result is sent to the food safety management end for reminding. The cross-contamination site prediction module 50 can further include: a mapping determination unit configured to map the supply node factor in the first cross marker chain, and determine whether the mapping node is a marked supply cross node; and a cross food supply chain extraction and analysis unit configured to extract a cross food supply feature marker of a corresponding node, perform cross food supply chain extraction and analysis, and generate the prediction result if the mapping node is the marked supply cross node.

[0075] The food safety risk management system based on multi-link data provided by the embodiment of the application can execute the food safety risk management method based on multi-link data provided by any embodiment of the application, and has the corresponding function modules and beneficial effects of the execution method.

[0076] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, however, any number of different modules can be used and run on the user terminal and / or server, and each unit and module is only divided according to the functional logic, but is not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for the convenience of mutual differentiation, and do not limit the protection scope of the present application.

[0077] The foregoing DETAILED DESCRIPTION, including the above section titled "Detailed Description," is not to be taken as limiting the scope of the application. Various modifications, combinations, and equivalents can be apparent to those skilled in the art and can be made once the nature of the application is understood. Any modification, combination, or equivalent, which falls within the principles and the scope of the present application, is intended to be included in the present application. In some instances, the actions or steps can be performed in different order from those described herein, and still achieve desirable results. Additionally, the process depicted in the figures can not necessarily require the particular order shown or sequential order to achieve the desired results. In certain implementations, multitasking and parallel processing can be advantageous.

Claims

1. A method for food safety risk management based on multi-link data, characterized in that, The method comprises the following steps: determining a first production chain of a first food, including a first raw material supply chain and a first production node chain after entering a production workshop; based on the first raw material supply chain, connecting a supply management platform of a corresponding supplier, identifying a supply cross-node and marking a cross-food supply feature according to a predetermined cross-judgment constraint, and generating a first cross-marking chain, including: extracting each transfer node in the first raw material supply chain; based on the supply management platform, identifying a supply cross-node according to a predetermined cross-judgment constraint at each transfer node, the predetermined cross-judgment constraint is used to judge whether there is a cross-contamination path of raw materials or food, specifically including equipment sharing and closed warehouse space sharing features; marking the supply cross-node in the first raw material supply chain, and marking the corresponding cross-food supply feature of the supply cross-node, including food type and supply path, to generate the first cross-marking chain; based on the first production node chain, reading the quality inspection parameters of each production node to determine a first unqualified parameter type that does not meet the preset quality inspection index; based on the first unqualified parameter type, identifying unqualified factors of the production node and the supply node, including: determining an unqualified detection node of the first unqualified parameter type; calling an abnormal source network based on the unqualified detection node to trace the unqualified factors, if the trace result is a single factor of the production node and the supply node, the single factor is used to generate a discrimination result; wherein the abnormal source network includes each abnormal source layer corresponding to each production node, each abnormal source layer is constructed based on a set of historical unqualified parameter types and corresponding historical trace nodes; if the trace result is a double factor of the production node and the supply node; extract the production node factor in the double factor, and perform self-production parameter abnormal verification and quality inspection verification of the production food in the corresponding production node to generate a production verification result; the production verification result is used to reserve a single factor constraint for the double factor to generate a reservation result as the discrimination result; if the discrimination result is a supply node factor, an unqualified factor cross-contamination risk analysis is performed to generate a first cross-risk index, including: determining a quality inspection molecular attribute in the first unqualified parameter type; judging whether the quality inspection molecular attribute belongs to a preset food diffusion contamination molecule, if yes, extracting a pollution diffusion risk feature; based on the supply node factor, extracting a supply cross feature in the first cross-marking chain, and comparing the supply cross feature with the pollution diffusion risk feature to determine the first cross-risk index; if the first cross-risk index is greater than a preset threshold, a cross-contamination site is predicted according to the first cross-marking chain, and the prediction result is sent to a food safety management end for reminding.

2. The multi-link data-based food safety risk management method of claim 1, wherein, After generating the first cross-marking chain, the method further comprises the following steps: determining cross-food attribute information based on the first cross-marking chain; generating a food monitoring library based on the cross-food attribute information; when any cross-food in the food monitoring library is detected to have a safety risk, a safety warning signal is sent out. Based on the safety warning signal, sampling detection of each batch of the first food is performed, and a sampling detection result is generated for risk self-checking warning.

3. The multi-link data-based food safety risk management method of claim 1, wherein, According to the first cross-label chain, a cross-contamination site is predicted, and a prediction result is sent to a food safety management end for warning, including: The supply node factor is mapped in the first cross-label chain, and it is judged whether the mapped node is a labeled supply cross node; If yes, the cross-food supply feature label of the corresponding node is extracted, cross-food supply chain extraction analysis is performed, and the prediction result is generated.

4. The multi-link data-based food safety risk management method of claim 1, wherein, Based on the first unqualified parameter type, unqualified factors of the production node and the supply node are discriminated, and further including: If the discrimination result is a production node factor, a production abnormality warning signal is sent out; Based on the production abnormality warning signal, production suspension control is performed, and the first unqualified parameter type is input into a production parameter optimization library for optimization. According to the optimized parameters, production restart control is performed, and production batch identification early warning during the abnormal period is performed.

5. A food safety risk management system based on multi-link data, characterized by, The system is used to implement the food safety risk management method based on multi-link data according to any one of claims 1-4, and the system includes: A first production chain determination module is used to determine a first production chain of a first food, including a first raw material supply chain and a first production node chain after entering a production workshop; A first cross-label chain generation module is used to connect a supply management platform of a corresponding supplier based on the first raw material supply chain, identify a supply cross node and a cross-food supply feature label according to a predetermined cross-determination constraint, and generate a first cross-label chain; A quality inspection parameter analysis module is used to read quality inspection parameters of each production node based on the first production node chain, and determine a first unqualified parameter type that does not meet a preset quality inspection index; An unqualified factor cross-contamination risk analysis module is used to discriminate unqualified factors of the production node and the supply node based on the first unqualified parameter type, and if the discrimination result is a supply node factor, unqualified factor cross-contamination risk analysis is performed, and a first cross-risk index is generated; A cross-contamination site prediction module is used to predict a cross-contamination site according to the first cross-label chain if the first cross-risk index is greater than a preset threshold, and a prediction result is sent to a food safety management end for warning.

Images