A milk powder production process traceability analysis system and method

Abstract

This invention relates to the field of milk powder production process monitoring and analysis technology, specifically a milk powder production process traceability analysis system and method. The method includes: acquiring historical process data of milk powder production; processing the historical process data to construct an anomaly database; performing anomaly impact propagation analysis on the target process and upstream processes based on the anomaly database to determine the propagation characteristics between processes; and identifying and alerting to novel anomalies generated by processes based on the propagation characteristics. The anomaly database is constructed by dividing samples by timestamp and calculating parameter anomaly coefficients; novel anomaly identification is achieved by integrating anomaly sequences, calculating all-cause and isolated impact coefficients to analyze propagation characteristics. This solves the problems of unclear propagation paths, difficulty in quantifying impacts, and difficulty in detecting novel anomalies in existing technologies for milk powder production, improving the accuracy and timeliness of production anomaly traceability and early warning.

Description

Technical Field

[0001] This invention relates to the field of monitoring and analysis technology for milk powder production processes, specifically to a traceability analysis system and method for milk powder production processes. Background Technology

[0002] The production process of milk powder involves a complex multi-process system, from raw material acceptance, sterilization, ingredient mixing to drying and packaging, with close correlation between the process parameters of each step.

[0003] In traditional production, due to the lack of an effective mechanism for associating abnormal parameter sequences, it is difficult to accurately locate the propagation path of parameters in upstream and downstream processes when an anomaly occurs in a certain stage, and it is impossible to quantify the degree of impact of the anomaly on subsequent stages. There is also a lack of systematic integration and in-depth analysis of historical process data. At the same time, existing technologies have weak capabilities in identifying new anomalies, often only responding passively after the anomaly has spread, leading to the expansion of quality problems, increasing production costs, and potentially posing risks to product safety. Therefore, there is an urgent need for a milk powder production process analysis method that can comprehensively trace the propagation of anomalies, quantify their impact, and promptly identify new anomalies. Summary of the Invention

[0004] The purpose of this invention is to address the problems existing in the background technology by proposing a traceability analysis system and method for milk powder production process.

[0005] The technical solution of this invention: a method for traceability analysis of milk powder production process, comprising the following steps:

[0006] Acquire historical process data of milk powder production and process the historical process data;

[0007] In each sample, identify outlier parameters and calculate the outlier coefficients for those parameters;

[0008] Identify the target process and its corresponding upstream process, and obtain multiple anomaly propagation events;

[0009] Based on the abnormal propagation events, the types of abnormal parameters corresponding to the upstream and target processes are obtained, the corresponding historical abnormal parameter sequences are constructed, and the first abnormal sequence and the second abnormal sequence are marked respectively. The first abnormal sequence and the second abnormal sequence are combined to obtain the first abnormal data. An abnormal database is constructed based on the first abnormal data corresponding to multiple abnormal propagation events.

[0010] Based on the anomaly database, an anomaly impact propagation analysis is performed on the target process and upstream processes to determine the propagation characteristics between processes;

[0011] Based on the propagation characteristics between processes, identify new types of anomalies generated by the processes and provide alerts.

[0012] Preferably, the parameter anomaly coefficient Ai is determined by the formula... In the formula, Pi is the abnormal parameter value; SPi is the standard parameter value, which is the median of the standard parameter range, and the standard parameter range is calibrated based on industry big data; αi is the influence weight of the parameter, which is calibrated based on the weight calculation of industry big data; i is the abnormal parameter number, and i is a positive integer.

[0013] Preferably, the abnormal propagation event satisfies the condition that after the upstream process experiences a parameter abnormality, the target process begins to experience a parameter abnormality.

[0014] Preferably, the analysis of the propagation of the impact of anomalies includes dividing and integrating the first anomalous data in the anomalous database according to the same first anomalous sequence to obtain several first anomalous data of the same type, performing a union operation on the second anomalous sequences of the several first anomalous data of the same type, and using them as the propagation anomalous sequences of the first anomalous sequence.

[0015] Repeat the steps to obtain the propagation anomaly sequence corresponding to each first anomaly sequence; combine the first anomaly sequence with the corresponding propagation anomaly sequence as the first anomaly sequence data of the anomaly propagation event.

[0016] Preferably, the analysis of the propagation of anomalies also includes classifying and integrating the propagation events based on the same first anomaly sequence to obtain several similar event libraries. Based on the first anomaly sequence and the propagation anomaly sequence, the all-cause anomaly coefficient Q of the upstream and target processes for each anomaly propagation event in the similar event library is calculated using the following formula: .

[0017] Preferably, the analysis of the propagation of anomalous impacts also includes labeling the all-cause anomaly coefficients of the upstream and target processes, respectively. and j is the number of the abnormal propagation event obtained, j is a positive integer, j∈[1,n], and n is the total number of abnormal propagation events in the event database;

[0018] Through formula The isolated impact coefficient GY of the upstream process on the target process is calculated; the range of the isolated impact coefficient is determined according to the preset coefficient tolerance ratio.

[0019] Propagation features are constructed based on the first anomalous sequence, the propagation anomalous sequence, and the range of isolated influence coefficients.

[0020] Preferably, the method for identifying abnormal processes includes extracting abnormal parameter sequences from the selected process based on process data and marking them as the first traceability sequence, and performing real-time traceability on the extraction results; if the first traceability sequence is empty, then continue tracing; if the first traceability sequence is not empty, identify whether the real-time abnormal parameter sequence is the first abnormal sequence that has been discovered.

[0021] If it does not belong to the first discovered anomaly sequence, it indicates the occurrence of a new type of process anomaly;

[0022] If it belongs to the first discovered abnormal sequence, extract the abnormal parameter sequence of the downstream process of the intermediate process and mark it as the second traceability sequence, and retrieve the target propagation feature based on the first sequence; calculate the isolated impact coefficient of the selected process on the downstream process based on the first traceability sequence and the second traceability sequence and mark it as the traceability coefficient; trace the second traceability sequence and the traceability coefficient based on the target propagation feature.

[0023] Preferably, tracing the second tracing sequence includes identifying whether the real-time propagation anomaly sequence is a subsequence of the propagation anomaly sequence in the target propagation feature; if the second tracing sequence is a subsequence of the propagation anomaly sequence in the target propagation feature, then tracing continues; otherwise, a new type of process anomaly is indicated.

[0024] Preferably, the method for tracing the traceability coefficient includes determining whether the traceability coefficient belongs to the range of isolated influence coefficients in the target propagation characteristics; if the traceability coefficient belongs to the range of isolated influence coefficients in the target propagation characteristics, then continue tracing; otherwise, indicate the occurrence of a new type of process anomaly.

[0025] This invention also discloses a milk powder production process traceability analysis system, which applies the above-mentioned milk powder production process traceability analysis method, specifically including the following steps:

[0026] The data acquisition module is used to acquire historical process data of milk powder production;

[0027] The data analysis module is used to process historical process data and build an anomaly database;

[0028] The data analysis module is used to perform anomaly impact propagation analysis on the target process and upstream processes based on the anomaly database, and to determine the propagation characteristics between the upstream process and the target process.

[0029] The process traceability module is used to identify new types of anomalies generated in a process and provide alerts based on the propagation characteristics between processes.

[0030] Compared with the prior art, the above-mentioned technical solution of the present invention has the following beneficial technical effects:

[0031] An anomaly database is constructed by dividing samples according to timestamps and calculating parameter anomaly coefficients; new anomaly identification is achieved by integrating anomaly sequences, calculating all-cause and isolated influence coefficients to analyze propagation characteristics; this solves the problems of unclear propagation paths, difficulty in quantifying impacts, and difficulty in discovering new anomalies in existing technologies for milk powder production, and improves the accuracy and timeliness of production anomaly tracing and early warning. Attached Figure Description

[0032] Figure 1 This is a layer structure diagram of Embodiment 1 of the present invention. Detailed Implementation

[0033] Example 1, as Figure 1 As shown, the present invention proposes a method for traceability analysis of milk powder production process, comprising:

[0034] Obtain historical process data of milk powder production, process the historical process data, and build an anomaly database;

[0035] Acquire historical process data and divide it into multiple samples based on timestamps;

[0036] In each sample, identify outlier parameters and calculate the outlier coefficients for those parameters;

[0037] The abnormal coefficient Ai is obtained through the formula In the formula, Pi is the abnormal parameter value; SPi is the standard parameter value, which is the median of the standard parameter range, and the standard parameter range is calibrated based on industry big data; αi is the influence weight of the parameter, which is calibrated based on industry big data weight calculation; i is the abnormal parameter number, and i is a positive integer; for example, the abnormal parameter types include input-output ratio, energy efficiency ratio, defect rate and cycle ratio for quantification;

[0038] Determine the target process and its corresponding upstream process, and obtain multiple anomaly propagation events; the anomaly propagation events satisfy the condition that the upstream process experiences parameter anomalies, after which the target process begins to experience parameter anomalies.

[0039] Based on the abnormal propagation events, the types of abnormal parameters corresponding to the upstream and target processes are obtained, the corresponding historical abnormal parameter sequences are constructed, and the first abnormal sequence and the second abnormal sequence are marked respectively. The first abnormal sequence and the second abnormal sequence are combined to obtain the first abnormal data. An abnormal database is constructed based on the first abnormal data corresponding to multiple abnormal propagation events.

[0040] The method for processing process data also includes obtaining parameter data for each production process, cleaning and normalizing the parameter data, removing abnormal data and mapping it to the [0, 1] interval;

[0041] Based on the anomaly database, anomaly impact propagation analysis is performed on the target process and upstream processes to determine the propagation characteristics between processes;

[0042] The analysis of the propagation of anomalous impacts includes the following steps:

[0043] In the abnormal database, the first abnormal data is divided and integrated according to the same first abnormal sequence to obtain several first abnormal data of the same type. The second abnormal sequences of the several first abnormal data of the same type are subjected to union operation and used as the propagation abnormal sequence of the first abnormal sequence.

[0044] Repeat the steps to obtain the propagation anomaly sequence corresponding to each first anomaly sequence; combine the first anomaly sequence with the corresponding propagation anomaly sequence as the first anomaly sequence data of the anomaly propagation event; it should be noted that the propagation anomaly sequence is the sequence of anomaly parameter numbers obtained by big data processing, which is affected by the anomaly parameter type of the upstream process and propagates in the target library process, and is calculated and integrated by the anomaly parameter type of the target process.

[0045] Anomaly propagation events are categorized and integrated based on the same first anomaly sequence to obtain several similar event libraries. The all-cause anomaly coefficient Q for the upstream and target processes of each anomaly propagation event in each similar event library is calculated based on the first anomaly sequence and the propagation anomaly sequence. The calculation formula is as follows: ;

[0046] The all-cause anomaly coefficients of the upstream and target processes are respectively labeled as follows: and j is the number of the abnormal propagation event obtained, j is a positive integer, j∈[1,n], and n is the total number of abnormal propagation events in the event database;

[0047] Through formula The isolated impact coefficient GY of the upstream process on the target process is calculated; the range of the isolated impact coefficient is determined according to the preset coefficient tolerance ratio.

[0048] Propagation characteristics are constructed based on the first anomalous sequence, the propagation anomalous sequence, and the range of isolated influence coefficients.

[0049] Based on the propagation characteristics between processes, identify new types of anomalies generated by the processes and provide alerts;

[0050] The method for identifying abnormal processes includes extracting abnormal parameter sequences from the selected process based on process data and marking them as the first traceability sequence, and tracing the extraction results in real time; if the first traceability sequence is empty, tracing continues; if the first traceability sequence is not empty, it is determined whether the real-time abnormal parameter sequence is the first abnormal sequence that has been discovered.

[0051] If it does not belong to the first discovered anomaly sequence, it indicates the occurrence of a new type of process anomaly;

[0052] If it belongs to the first discovered abnormal sequence, extract the abnormal parameter sequence of the downstream process of the intermediate process and mark it as the second traceability sequence, and retrieve the target propagation feature based on the first sequence; calculate the isolation impact coefficient of the selected process on the downstream process based on the first traceability sequence and the second traceability sequence and mark it as the traceability coefficient; trace the second traceability sequence and the traceability coefficient based on the target propagation feature;

[0053] Tracing the second tracing sequence includes identifying whether the real-time propagation anomaly sequence is a subsequence of the propagation anomaly sequence in the target propagation feature; if the second tracing sequence is a subsequence of the propagation anomaly sequence in the target propagation feature, tracing continues; otherwise, a new type of process anomaly is indicated.

[0054] The method for tracing the traceability coefficient includes determining whether the traceability coefficient falls within the range of isolated influence coefficients in the target propagation characteristics; if the traceability coefficient falls within the range of isolated influence coefficients in the target propagation characteristics, then tracing continues; otherwise, a new type of process anomaly is indicated.

[0055] An anomaly database is constructed by dividing samples according to timestamps and calculating parameter anomaly coefficients; new anomaly identification is achieved by integrating anomaly sequences, calculating all-cause and isolated influence coefficients to analyze propagation characteristics; this solves the problems of unclear propagation paths, difficulty in quantifying impacts, and difficulty in discovering new anomalies in existing technologies for milk powder production, and improves the accuracy and timeliness of production anomaly tracing and early warning.

[0056] Example 2: The milk powder production process traceability analysis system proposed in this invention is applied to the milk powder production process traceability analysis method proposed in Example 1, and specifically includes the following steps:

[0057] The data acquisition module is used to acquire historical process data of milk powder production;

[0058] The data analysis module is used to process historical process data and build an anomaly database;

[0059] The data analysis module is used to perform anomaly impact propagation analysis on the target process and upstream processes based on the anomaly database, and to determine the propagation characteristics between the upstream process and the target process.

[0060] The process traceability module is used to identify new types of anomalies generated in a process and provide alerts based on the propagation characteristics between processes.

[0061] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for tracing and analyzing the milk powder production process, characterized in that, Includes the following steps: Acquire historical process data of milk powder production and process the historical process data; In each sample, identify outlier parameters and calculate the outlier coefficients for those parameters; The abnormal coefficient Ai is obtained through the formula In the formula, Pi is the abnormal parameter value; SPi is the standard parameter value, which is the median of the standard parameter range, and the standard parameter range is calibrated based on industry big data; αi is the influence weight of the parameter function, which is calibrated based on the weight calculation of industry big data. i is the abnormal parameter number, and i is a positive integer; Identify the target process and its corresponding upstream process, and obtain multiple anomaly propagation events; Based on the abnormal propagation event, the types of abnormal parameters corresponding to the upstream process and the target process are obtained, the corresponding historical abnormal parameter sequences are constructed, and the first abnormal sequence and the second abnormal sequence are marked respectively. The first abnormal sequence and the second abnormal sequence are combined to obtain the first abnormal data. An anomaly database is constructed based on the first anomaly data corresponding to multiple anomaly propagation events; based on the anomaly database, an anomaly impact propagation analysis is performed on the target process and upstream processes to determine the propagation characteristics between processes; The analysis of the propagation of anomalies includes dividing and integrating the first anomalous data in the anomalous database according to the same first anomalous sequence to obtain several first anomalous data of the same type, performing a union operation on the second anomalous sequences of the several first anomalous data of the same type, and using them as the propagation anomalous sequences of the first anomalous sequence. Repeat the steps to obtain the propagation anomaly sequence corresponding to each first anomaly sequence; combine the first anomaly sequence with the corresponding propagation anomaly sequence to obtain the first anomaly sequence data of the anomaly propagation event; The analysis of the propagation of anomalies also includes classifying and integrating propagating events based on the same first anomaly sequence to obtain several similar event libraries. Based on the first anomaly sequence and the propagating anomaly sequence, the all-cause anomaly coefficient Q for the upstream and target processes of each propagating event in the similar event library is calculated using the following formula: ; The analysis of the propagation of anomalous impacts also includes labeling the all-cause anomaly coefficients of the upstream and target processes, respectively. and j is the number of the abnormal propagation event obtained, j is a positive integer, j∈[1,n], and n is the total number of abnormal propagation events in the event database; Through formula The isolated impact coefficient GY of the upstream process on the target process was calculated. The range of isolated influence coefficients is determined based on the preset coefficient tolerance ratio; Propagation characteristics are constructed based on the first anomalous sequence, the propagation anomalous sequence, and the range of isolated influence coefficients; Based on the propagation characteristics between processes, identify new types of anomalies generated by the processes and provide alerts.

2. The method for traceability analysis of milk powder production process according to claim 1, characterized in that, An abnormal propagation event is defined as follows: after an abnormal parameter occurs in the upstream process, the target process begins to experience an abnormal parameter.

3. The method for traceability analysis of milk powder production process according to claim 1, characterized in that, The method for identifying abnormal processes includes extracting abnormal parameter sequences from the selected process based on process data and marking them as the first traceability sequence, and tracing the extraction results in real time; if the first traceability sequence is empty, tracing continues. If the first traceability sequence is not empty, identify whether the real-time abnormal parameter sequence is the first abnormal sequence that has been discovered; If it does not belong to the first discovered anomaly sequence, it indicates the occurrence of a new type of process anomaly; If it belongs to the first discovered abnormal sequence, then extract the abnormal parameter sequence of the downstream process in the middle process and mark it as the second traceability sequence, and retrieve the target propagation features based on the first sequence; The isolated impact coefficient of the selected process on the downstream process is calculated based on the first and second traceability sequences and marked as the traceability coefficient; the second traceability sequence and traceability coefficient are traced based on the target propagation characteristics.

4. The method for traceability analysis of milk powder production process according to claim 3, characterized in that, Tracing the second tracing sequence includes identifying whether the real-time propagation anomaly sequence is a subsequence of the propagation anomaly sequence in the target propagation feature; if the second tracing sequence is a subsequence of the propagation anomaly sequence in the target propagation feature, then tracing continues. Otherwise, a new type of process anomaly will be indicated.

5. The method for traceability analysis of milk powder production process according to claim 3, characterized in that, Methods for tracing the tracing coefficient include determining whether the tracing coefficient falls within the range of isolated influence coefficients in the target propagation characteristics; If the tracing coefficient falls within the range of isolated influence coefficients in the target propagation characteristics, then tracing continues. Otherwise, a new type of process anomaly will be indicated.

6. A milk powder production process traceability analysis system, applied to the milk powder production process traceability analysis method described in any one of claims 1 to 5, characterized in that, Specifically, the following steps are included: The data acquisition module is used to acquire historical process data of milk powder production; The data analysis module is used to process historical process data and build an anomaly database; The data analysis module is used to perform anomaly impact propagation analysis on the target process and upstream processes based on the anomaly database, and to determine the propagation characteristics between the upstream process and the target process. The process traceability module is used to identify new types of anomalies generated in a process and provide alerts based on the propagation characteristics between processes.

Images