A Big Data-Based Method for Feed Production and Processing Testing

Abstract

This invention relates to the field of feed production, processing, and testing technology. Specifically, it relates to a feed production, processing, and testing method based on big data. The method includes the following steps: S1, collecting historical feed production data and screening it according to processing stages to obtain historical production data corresponding to each processing stage; S2, obtaining the feed quality corresponding to the historical production data, and simultaneously obtaining the standard process and standard quality corresponding to each processing stage. Through big data modeling, a deep correlation analysis between process and quality is achieved, significantly improving the accuracy of quality control. This not only quantifies the differences in feed quality and process deviations at each stage, but also innovatively introduces modeling of raw material quality grades and processing sequence nodes. By calculating the positive and negative values ​​of the impact of process deviations on quality, it accurately identifies the positive or negative destructive effects of fluctuations in different process parameters on quality.

Description

Technical Field

[0001] This invention relates to the field of feed production, processing and testing technology, and more specifically, to a feed production, processing and testing method based on big data. Background Technology

[0002] Feed production, processing, and testing are key upstream links in the livestock breeding industry chain. Their core purpose is to ensure the nutritional balance, hygiene and safety, and processing suitability of feed products by controlling multiple indicators of raw materials, processing, and finished products, so as to meet the growth needs of livestock, poultry, aquatic animals, and other animals in different breeding scenarios.

[0003] Existing technologies suffer from insufficient timeliness in testing. Manual sampling and laboratory analysis have long cycles and cannot capture the dynamic fluctuations of process parameters in real time, leading to delays in process adjustments and the generation of batches of substandard products, resulting in waste of raw materials and energy. Furthermore, the data correlation analysis capability is weak. Traditional technologies only process parameters or quality data of a single link in isolation, ignoring the chain reaction of raw material quality differences and deviations in preceding processes on subsequent links and final quality. This results in one-sided deviation analysis and an inability to pinpoint the core cause of quality fluctuations. Therefore, a big data-based feed production and processing testing method is proposed. Summary of the Invention

[0004] The purpose of this invention is to provide a feed production, processing and testing method based on big data to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, a big data-based feed production and processing testing method is provided, comprising the following steps:

[0006] S1. Collect historical production data of feed and screen the historical production data according to the processing stage to obtain the historical production data corresponding to each processing stage;

[0007] S2. Obtain the feed quality corresponding to historical production data, and at the same time obtain the standard process and standard quality corresponding to each processing link. Then, perform process analysis on the historical production data corresponding to each processing link to obtain the historical process set of each processing link.

[0008] S3. Perform deviation analysis between the feed quality corresponding to the historical process and the standard quality to obtain the feed quality difference value, and perform deviation analysis between the historical process set and the standard process in the same processing link to obtain the process deviation value.

[0009] S4. Combine the historical process set with the feed quality difference value, and conduct a positive and negative value analysis on the impact of the process deviation value of each historical process on feed quality, to obtain the positive and negative value of the impact of the process deviation value on feed quality under different completion process conditions.

[0010] S5. Obtain the required quality of feed, combine the required quality range with the standard quality to set the required quality score, then extract the real-time process and combine the positive and negative values ​​of the influence to obtain the predicted quality score, and combine the predicted quality score with the required quality score to complete the optimization recommendation.

[0011] As a further improvement to this technical solution, in step S1, a communication connection is established with the feed production end to collect historical production data of the feed and obtain the feed processing steps.

[0012] Historical production data is sifted according to the processing stage to obtain the historical production data corresponding to each processing stage.

[0013] As a further improvement to this technical solution, in step S2, historical production data corresponding to feed quality is obtained from the feed production end;

[0014] At the same time, we can obtain the standard processes corresponding to each processing stage, as well as the standard quality corresponding to feed production.

[0015] The historical production data corresponding to each processing stage is analyzed to obtain the historical processes corresponding to the historical production data. Then, the historical processes corresponding to each processing stage are summarized to form a historical process set.

[0016] As a further improvement to this technical solution, in step S3, the feed quality corresponding to the historical process and the standard quality are subjected to deviation analysis to obtain the feed quality difference value. The feed quality difference value is the percentage of the absolute difference between the measured value of the feed quality corresponding to the historical process and the standard value of the standard quality to the standard value.

[0017] In the same processing stage, the historical processes and standard processes in the historical process set are analyzed for deviation to obtain the process deviation values ​​between each historical process and the corresponding standard process in the processing stage.

[0018] The process deviation value is the absolute difference between the characteristic value of the historical process and the characteristic value of the standard process for each processing step.

[0019] As a further improvement to this technical solution, in step S4, the standard quality is set to 100 points, and the final score of the feed quality corresponding to the historical process is set in combination with the feed quality difference value to obtain the final score corresponding to the historical process.

[0020] Extract raw material quality data, set initial scores based on raw material quality data, extract the historical process set corresponding to the first processing stage, use the extracted historical process set as the initial node, and sort the nodes of other processing stages in the order of processing to generate the corresponding number of processing nodes.

[0021] Input the corresponding process deviation values ​​of the historical process into the processing node. Then, combine the process deviation values ​​of the processing node with the historical process, the initial score and the final score to calculate the impact of the corresponding process deviation values ​​of each processing node on feed quality. This will allow us to obtain the positive and negative values ​​of the impact of the corresponding process deviation values ​​of each historical process on the final score under different completed processes.

[0022] As a further improvement to this technical solution, the different completion processes are the historical processes of the preceding processing nodes corresponding to the processing nodes;

[0023] Meanwhile, the raw material quality data shows that the higher the raw material quality, the higher the initial score;

[0024] The raw material quality data shows that the lower the raw material quality, the lower the initial score.

[0025] The impact scores are grouped and saved based on the raw material quality data.

[0026] As a further improvement to this technical solution, in step S5, the required quality of feed is obtained at the feed production end, the required quality is combined with the standard quality for score analysis, the required quality score corresponding to the required quality is obtained, and then the required quality score is used as a threshold for optimization and recommendation.

[0027] As a further improvement to this technical solution, in the optimization recommendation based on the required quality score as a threshold, real-time process and raw material quality data are obtained at the feed production end; wherein, the real-time process includes completed real-time processes and incomplete real-time processes;

[0028] Initial scores were set based on raw material quality data, and data with positive and negative impact values ​​corresponding to the raw material quality data were selected for analysis.

[0029] The real-time process input is entered into the processing node, and the predicted quality score is analyzed by combining the positive and negative values ​​of the influence. The predicted quality score corresponding to the real-time process is obtained, and the predicted quality score is compared with the required quality score. When the predicted quality score is less than the required quality score, the process optimization is triggered. In the processing node, based on the completed real-time process, the incomplete real-time process is replaced and optimized by combining the historical process, thereby adjusting the predicted quality score, and outputting the replacement optimization scheme with the highest predicted quality score.

[0030] If the predicted quality score is greater than the required quality score, monitoring should continue.

[0031] As a further improvement to this technical solution, in the replacement optimization scheme of the highest predicted quality score, when the highest predicted quality score is lower than the required quality score, the output is stopped and manual adjustment is triggered.

[0032] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0033] 1. This big data-based feed production and processing testing method achieves in-depth correlation analysis between process and quality through big data modeling, significantly improving the accuracy of quality control. It not only quantifies the differences in feed quality and process deviations at each stage, but also innovatively introduces modeling of raw material quality grades and processing sequence nodes. By calculating the positive and negative values ​​of the impact of process deviations on quality, it accurately identifies the positive or negative impact of fluctuations in different process parameters on quality, overcoming the shortcomings of traditional technologies that ignore antecedent factors and provide one-sided analysis. This provides a clear quantitative basis for process adjustment and effectively reduces quality fluctuations caused by blind parameter adjustments.

[0034] 2. In this big data-based feed production and processing testing method, the constructed real-time prediction-optimization intervention closed-loop mechanism completely changes the passive mode of traditional technology of post-event testing and post-event adjustment. By extracting real-time process data and combining it with historical positive and negative impact data to calculate the predicted quality score, it is possible to predict in advance whether the feed quality meets the requirements. When the predicted score does not meet the standard, based on the completed process, the optimal solution is selected from the historical process set to replace the incomplete process, realizing real-time optimization of the processing process, greatly reducing the probability of producing unqualified products, while reducing raw material waste and energy consumption, and improving production efficiency and resource utilization.

[0035] 3. This big data-based feed production and processing testing method groups and stores the positive and negative value data according to the raw material quality grade, which has strong scenario adaptability and practicality. It can quickly match the corresponding historical analysis data based on the raw material quality data of the current production batch, and output personalized process optimization solutions for different raw material scenarios such as high-quality and low-quality raw materials. It significantly improves the versatility in various feed production scenarios and provides strong support for feed production enterprises to achieve precise quality control and efficient production. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the process of a feed production, processing and testing method based on big data according to the present invention. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] Please see Figure 1As shown, the purpose of this embodiment is to provide a feed production and processing testing method based on big data, including the following steps:

[0039] S1. Collect historical production data of feed and screen the historical production data according to the processing stage to obtain the historical production data corresponding to each processing stage; obtain structured historical data to provide a data source for subsequent process and quality correlation analysis;

[0040] In S1, by establishing a communication connection with the feed production end, historical production data of feed is collected from the feed production end, and the feed processing stage is also obtained.

[0041] By establishing a two-way communication connection with the PLC control system, production execution system (MES), and sensor acquisition terminal at the feed production end through the OPCUA industrial communication protocol, and configuring data acquisition permissions and transmission frequency, the real-time performance and security of data transmission are ensured.

[0042] Two types of historical production data were collected simultaneously from the feed production end, as follows:

[0043] Basic attribute data, such as raw material batch, raw material type, production time, product model, and equipment number;

[0044] Process parameter data, sensor data collected at each processing stage such as temperature, humidity, pressure, speed, time, and material flow, as well as manual inspection data of semi-finished / finished products;

[0045] At the same time, a list of standard processing steps for feed production is obtained from the process document library at the production end, clearly defining seven core processing steps: raw material cleaning, raw material crushing, ingredient mixing, pelleting / expansion, cooling and drying, screening and grading, and post-coating.

[0046] Historical production data is sifted according to the processing stage to obtain the historical production data corresponding to each processing stage.

[0047] A unique feature label is assigned to each core processing step. The label includes the step code and the core parameter type. All historical production data collected are traversed, and the parameter type and production time information of each data point are extracted. These data are then matched with the feature labels of each processing step, and the successfully matched data are classified into the data subset of the corresponding processing step.

[0048] S2. Obtain the feed quality corresponding to historical production data, and at the same time obtain the standard process and standard quality corresponding to each processing link. Perform process analysis on the historical production data corresponding to each processing link to obtain the historical process set of each processing link. Establish a ternary correlation between historical process, feed quality and standard parameters, and clarify the analysis benchmark (standard process, standard quality) and analysis object (historical process, actual quality).

[0049] In S2, historical production data corresponding to feed quality is obtained from the feed production end;

[0050] Historical production data and corresponding feed quality testing data are extracted from the Production Execution System (MES) and Laboratory Information Management System (LIMS) at the feed production end. A mapping relationship between the two is established using the production batch number and production timestamp as the unique association key, ensuring that each piece of historical production data can be accurately matched with the corresponding feed quality test results (such as nutritional components, processing quality, hygiene and safety indicators), and invalid production data without matching quality data is eliminated.

[0051] At the same time, we can obtain the standard processes corresponding to each processing stage, as well as the standard quality corresponding to feed production.

[0052] Standard process parameters for each processing stage (raw material cleaning, crushing, ingredient mixing, etc.) are retrieved from the process document database at the production end to form a structured standard process list of stages, parameters, and threshold ranges. At the same time, based on the company's internal quality control standards, a standard quality indicator system corresponding to feed production is organized to clarify the qualified range and optimal range of each quality indicator, forming a standardized quality parameter table.

[0053] The historical production data corresponding to each processing stage is analyzed to obtain the historical processes corresponding to the historical production data. Then, the historical processes corresponding to each processing stage are summarized to form a historical process set.

[0054] For the historical production data corresponding to each processing stage, the statistical feature extraction method is used for process analysis. First, the core process parameters of each stage are screened (such as the rotation speed, time, and particle size detection value in the crushing stage; the mixing time and stirring speed in the mixing stage). Then, the statistical characteristics (mean, variance, peak value, and trend slope) of each core parameter are calculated. These features are combined to form the process feature vector corresponding to a single historical production data, thus completing the transformation from historical production data to historical process.

[0055] The transformed historical processes are categorized and collected according to the processing stage, with historical processes of the same stage forming a subset. Each subset is then validated, and invalid historical processes with missing process features or abnormal parameters are removed. Finally, all valid subsets from all stages are aggregated to form a historical process set covering the entire processing chain.

[0056] S3. Perform deviation analysis on the feed quality and standard quality corresponding to the historical process to obtain the feed quality difference value, and perform deviation analysis on the historical process set and standard process of the same processing link to obtain the process deviation value; establish a quantitative correlation basis between process deviation and quality deviation by quantifying the deviation between actual quality and standard quality (quality difference value) and the deviation between actual process and standard process (process deviation value).

[0057] Organize the previously associated historical process-feed quality measured value data sets, and at the same time retrieve the standard process parameters and feed standard quality indicators of the corresponding processing links. Establish a unique matching index by processing link + production batch to ensure that each historical process and the corresponding quality measured value can be accurately matched with the standard process of the same link and the standard quality of the same product, avoiding cross-link and cross-batch mismatch that may lead to distorted deviation analysis.

[0058] In S3, a deviation analysis is performed between the feed quality corresponding to the historical process and the standard quality to obtain the feed quality difference value. The feed quality difference value is the percentage of the absolute difference between the measured value of the feed quality corresponding to the historical process and the standard value of the standard quality to the standard value.

[0059] For each feed quality indicator (such as crude protein, moisture, mycotoxin content, etc.) corresponding to the historical process of a single batch, deviation calculation is performed one by one. First, the measured quality value of the indicator and the standard value of the standard quality are obtained, and the absolute difference between the two is calculated. Then, the absolute difference is divided by the standard value and converted into a percentage to obtain the feed quality difference value of the single indicator. After completing the calculation of all quality indicators, a list of quality difference values ​​for the batch is compiled to clarify the degree of deviation of each indicator.

[0060] In the same processing stage, the historical processes and standard processes in the historical process set are analyzed for deviation to obtain the process deviation values ​​between each historical process and the corresponding standard process in the processing stage.

[0061] The process deviation value is the absolute difference between the characteristic value of the historical process and the characteristic value of the standard process for each processing step.

[0062] Within the same processing stage (such as the raw material crushing stage), extract the core process characteristics of the historical process (such as the average crushing speed, particle size D50 characteristic value, etc.) and the corresponding characteristic values ​​of the standard process of that stage. For each core process characteristic, calculate the absolute difference between the two, which is the process deviation value of a single characteristic. After traversing all core process characteristics of the stage and completing the full characteristic deviation calculation, summarize them to form a list of process deviation values ​​for a single historical process, covering the deviation of key process parameters in that stage.

[0063] S4. Combine historical process data with feed quality variation values, and conduct a positive and negative impact analysis on feed quality for the process deviation values ​​of each historical process. Obtain the positive and negative impact of the process deviation values ​​on feed quality under different completion process conditions for each historical process; clarify the positive / negative impact of different process deviations on feed quality, and consider the prior impact of processing sequence (preceding processes) and raw material quality to make the impact analysis more in line with the actual production process.

[0064] In S4, the standard quality is set to 100 points, and the final score of the feed quality corresponding to the historical process is set by combining the feed quality difference value to obtain the final score corresponding to the historical process.

[0065] The standard quality benchmark score is set at 100 points. The previously calculated feed quality variation values ​​are linked to the score deduction rules. The larger the variation value, the greater the corresponding deduction. The final feed quality score for each historical process is calculated through deduction (i.e., 100 points minus the deduction points based on the variation value; the score after deduction is not lower than 0 points), as shown in the following formula:

[0066] ;

[0067] Among them, S final Q represents the final feed quality score corresponding to the historical process (range 0-100 points), where 100 is the standard quality benchmark score, and α is the score deduction coefficient (empirical value, ranging from 1 to 2, which can be adjusted according to the importance of quality indicators). d This represents the previously calculated feed quality variation value;

[0068] Extract raw material quality data and set initial scores based on this data. Extract raw material quality data (such as purity, mold rate, and basic nutritional values) from historical production data at the feed production end, and set initial score ranges according to raw material quality grades (excellent, good, medium, and poor): the better the raw material quality, the higher the initial score (e.g., excellent raw materials: 90-100 points; good: 80-89 points; medium: 70-79 points; poor: below 70 points). This completes the precise mapping between initial scores and raw material quality, using the following formula:

[0069] ;

[0070] Among them, S init Here, k represents the initial score corresponding to a single batch of raw materials, and S represents the raw material quality grade. min，k and S max，k Q represents the minimum and maximum values ​​of the initial score range for the k-th grade raw material, respectively. raw Q represents the actual quality parameters of a single batch of raw materials. min，k and Q max，k These are the minimum and maximum values ​​of the quality parameter for the k-th grade raw material, respectively.

[0071] Extract the historical process set corresponding to the first processing step, use the extracted historical process set as the initial node, and sort the nodes of other processing steps in the order of processing to generate the corresponding number of processing nodes.

[0072] Extract the historical process set corresponding to the first processing step in feed production (usually the raw material cleaning step), define it as the initial node, and use it as the starting point of the entire processing chain. Then, according to the actual feed production process, sort the remaining processing steps in sequence to generate a time-series processing node with the same number of processing steps, with each node corresponding to one processing step.

[0073] Input the corresponding process deviation values ​​of historical processes into the processing nodes. Then, combine the process deviation values ​​of the processing nodes with the historical processes, the initial scores, and the final scores to calculate the impact of the process deviation values ​​of each processing node on feed quality. This will allow us to obtain the positive or negative impact of the corresponding process deviation values ​​of each historical process on the final score under different completed processes. The steps are as follows:

[0074] The process deviation values ​​corresponding to each historical process are matched and input to the corresponding time-series processing nodes according to the processing stage, ensuring that the deviation values ​​correspond one-to-one with the nodes. Starting from the initial score, the impact of each node on the feed quality score is calculated in turn, based on the process deviation values ​​of each processing node and the historical processes: First, the change in the score caused by the process deviation value of the initial node (the first processing stage) is calculated, and then the cumulative score after the initial node is obtained based on this change; subsequent nodes are calculated based on the cumulative score of the previous nodes, and the change in score caused by the process deviation value of the current node is calculated.

[0075] If a process deviation at a certain node causes the cumulative score to increase towards the final score (or narrows the gap between the final score and the standard score), then the impact of that deviation is positive.

[0076] If the cumulative score decreases (or the gap with the standard score widens), the impact value is negative. The final impact value of each historical process on its corresponding preceding completed process is obtained using the following formula:

[0077] ;

[0078] Wherein, I represents the positive or negative impact of the process deviation on feed quality (I > 0 indicates a positive impact, meaning the deviation increases the quality score; I < 0 indicates a negative impact, meaning the deviation decreases the quality score; the larger the absolute value of I, the stronger the impact), S after S represents the base score (initial score or score influenced by previous nodes) before the calculation of this processing node. before P is the score after the process deviation at this processing node has taken effect. d This is the process deviation value for this processing node (non-zero, to avoid a denominator of 0; the impact value is 0 when there is zero deviation).

[0079] Different completion processes are the historical processes of the preceding processing nodes corresponding to this processing node;

[0080] The completed process of a certain processing node is the historical process actually executed by all preceding processing nodes (preceding nodes) before this node. The analysis of each node must be based on the completed process status of the preceding nodes.

[0081] Meanwhile, the raw material quality data shows that the higher the raw material quality, the higher the initial score;

[0082] The raw material quality data shows that the lower the raw material quality, the lower the initial score.

[0083] The impact scores are grouped and saved based on the raw material quality data.

[0084] Based on the pre-defined raw material quality grades (excellent, good, medium, poor), the process deviation values, corresponding positive and negative impact values, initial scores, and final scores of each processing node are grouped and categorized. The impact data of the same raw material quality grade are grouped together, and a correlation index of raw material quality-processing node-process deviation-positive and negative impact value is established to facilitate quick matching and retrieval during subsequent real-time process prediction.

[0085] S5. Obtain the required feed quality, combine the required quality range with the standard quality to set the required quality score, then extract the real-time process and combine it with the positive and negative impact values ​​to obtain the predicted quality score, and combine the predicted quality score with the required quality score to complete the optimization recommendation. Based on the impact analysis results of S4, perform quality prediction on the real-time production process, and ensure the final quality meets the standards by optimizing the incomplete process, forming a closed loop of real-time monitoring-prediction-optimization-intervention;

[0086] In S5, the required quality of feed is obtained at the feed production end. The required quality is combined with the standard quality for score analysis to obtain the required quality score. Then, the required quality score is used as a threshold for optimization and recommendation.

[0087] At the feed production end, obtain clear feed quality requirements (such as nutrient accuracy, hygiene and safety thresholds, processing quality standards, etc.), refer to the standard quality 100-point quantitative system, and set weights according to the importance of quality indicators (key indicators such as mycotoxins and crude protein weight 0.3-0.4, common indicators such as particle size uniformity weight 0.1-0.2), and convert the required quality into corresponding required quality scores (i.e., optimization thresholds), as shown in the following formula:

[0088] ;

[0089] Among them, S req To achieve the desired quality score (optimized threshold), w i Q represents the weight of the i-th quality requirement indicator. req，i The allowable deviation score corresponding to the i-th quality requirement is deducted, where n is the number of quality requirement indicators.

[0090] In the optimization recommendation based on the required quality score as a threshold, real-time process and raw material quality data are obtained at the feed production end; among them, real-time process includes completed real-time process and incomplete real-time process.

[0091] Two types of data are collected in real time from the production end:

[0092] Real-time process data clearly distinguishes between completed real-time processes (such as finished raw material cleaning and crushing processes) and incomplete real-time processes (such as pending batching, mixing, and granulation processes).

[0093] The current batch of raw material quality data (such as purity, mold rate, and nutritional baseline value) is matched with the corresponding initial score (using the previous raw material quality grade-initial score mapping rule). At the same time, historical positive and negative impact analysis data consistent with the current raw material quality grade are selected to ensure that the prediction basis matches the actual production conditions.

[0094] Initial scores were set based on raw material quality data, and data with positive and negative impact values ​​corresponding to the raw material quality data were selected for analysis.

[0095] The real-time process input is entered into the processing node, and the predicted quality score is analyzed by combining the positive and negative values ​​of the influence. The predicted quality score corresponding to the real-time process is obtained, and the predicted quality score is compared with the required quality score. When the predicted quality score is less than the required quality score, the process optimization is triggered. In the processing node, based on the completed real-time process, the incomplete real-time process is replaced and optimized by combining the historical process, thereby adjusting the predicted quality score, and outputting the replacement optimization scheme with the highest predicted quality score.

[0096] If the predicted quality score is greater than the required quality score, continue monitoring as follows:

[0097] Completed and incomplete real-time processes are input into the corresponding time-series processing nodes according to the processing order. Combined with the filtered positive and negative impact data, the predicted quality score is calculated starting from the initial score and according to the logic of superimposing the impact of the preceding nodes. First, the impact values ​​of the nodes corresponding to the completed processes are superimposed to obtain the current cumulative score. Then, based on the current parameters of the incomplete processes and combined with historical impact patterns, the impact values ​​of subsequent nodes are predicted. Finally, the complete real-time process predicted quality score is obtained by summarizing.

[0098] If the predicted quality score is greater than the required quality score, it means that the current real-time process can meet the quality requirements and no optimization is needed. Just continue to monitor the subsequent unfinished processes in real time.

[0099] If the predicted quality score is less than the required quality score, the process optimization process is immediately triggered, entering the historical process replacement stage. Using the completed real-time process as a fixed basis (which cannot be retrospectively modified), incomplete historical processes that match the current scenario (same raw material quality grade, same previously completed process) are selected from the historical process set. These incomplete processes are then replaced one by one, and the predicted quality score for each replacement scheme is recalculated. The replacement scheme with the highest predicted quality score is then selected as the optimal optimization scheme, as shown in the following formula:

[0100] ;

[0101] Among them, S pred S represents the predicted quality score corresponding to the real-time process. init The initial score for matching the current raw material quality, m is the number of processing nodes corresponding to the completed process, and I j P represents the positive or negative value of the influence of the j-th completed node. real，j Let I be the real-time process deviation value of the j-th node, p be the number of processing nodes corresponding to incomplete processes, and I be the value of the process deviation value of the j-th node. v P represents the positive or negative value of the influence of the v-th incomplete node. real，v This represents the deviation value of the currently planned process parameters for the v-th node.

[0102] In the replacement optimization scheme that outputs the highest predicted quality score, if the highest predicted quality score is lower than the required quality score, the output will stop and manual adjustment will be triggered.

[0103] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A feed production and processing detection method based on big data, characterized in that: Includes the following steps: S1. Collect historical production data of feed and screen the historical production data according to the processing stage to obtain the historical production data corresponding to each processing stage; S2. Obtain the feed quality corresponding to historical production data, and at the same time obtain the standard process and standard quality corresponding to each processing link. Then, perform process analysis on the historical production data corresponding to each processing link to obtain the historical process set of each processing link. S3. Perform deviation analysis between the feed quality corresponding to the historical process and the standard quality to obtain the feed quality difference value, and perform deviation analysis between the historical process set and the standard process in the same processing link to obtain the process deviation value. S4. Combine the historical process set with the feed quality difference value, and conduct a positive and negative value analysis on the impact of the process deviation value of each historical process on feed quality, to obtain the positive and negative value of the impact of the process deviation value on feed quality under different completion process conditions. In S4, the standard quality is set to 100 points, and the final score of the feed quality corresponding to the historical process is set in combination with the feed quality difference value to obtain the final score corresponding to the historical process. Extract raw material quality data, set initial scores based on raw material quality data, extract the historical process set corresponding to the first processing stage, use the extracted historical process set as the initial node, and sort the nodes of other processing stages in the order of processing to generate the corresponding number of processing nodes. Input the corresponding process deviation value of the historical process into the processing node. Then, combine the process deviation value of the processing node with the historical process, the initial score and the final score to calculate the impact value of the corresponding process deviation value of each processing node on the feed quality. This will obtain the positive and negative values ​​of the impact of the corresponding process deviation value of each historical process on the final score under different completed processes. S5. Obtain the required quality of feed, combine the required quality range with the standard quality to set the required quality score, then extract the real-time process and combine the positive and negative values ​​of the influence to obtain the predicted quality score, and combine the predicted quality score with the required quality score to complete the optimization recommendation. In S5, the required quality of feed is obtained at the feed production end, and the required quality is combined with the standard quality for score analysis to obtain the required quality score. Then, the required quality score is used as a threshold for optimization and recommendation. In the optimization recommendation based on the required quality score as a threshold, real-time process and raw material quality data are obtained at the feed production end; wherein, the real-time process includes completed real-time processes and incomplete real-time processes; Initial scores were set based on raw material quality data, and data with positive and negative impact values ​​corresponding to the raw material quality data were selected for analysis. The real-time process input is entered into the processing node, and the predicted quality score is analyzed by combining the positive and negative values ​​of the influence. The predicted quality score corresponding to the real-time process is obtained, and the predicted quality score is compared with the required quality score. When the predicted quality score is less than the required quality score, the process optimization is triggered. In the processing node, based on the completed real-time process, the incomplete real-time process is replaced and optimized by combining the historical process, thereby adjusting the predicted quality score and outputting the replacement optimization scheme with the highest predicted quality score. If the predicted quality score is greater than the required quality score, monitoring should continue.

2. The feed production, processing, and testing method based on big data according to claim 1, characterized in that: In S1, a communication connection is established with the feed production end to collect historical production data of the feed and obtain the feed processing steps. Historical production data is sifted according to the processing stage to obtain the historical production data corresponding to each processing stage.

3. The feed production, processing, and testing method based on big data according to claim 1, characterized in that: In step S2, historical production data corresponding to feed quality is obtained from the feed production end. At the same time, we can obtain the standard processes corresponding to each processing stage, as well as the standard quality corresponding to feed production. The historical production data corresponding to each processing stage is analyzed to obtain the historical processes corresponding to the historical production data. Then, the historical processes corresponding to each processing stage are summarized to form a historical process set.

4. The feed production, processing, and testing method based on big data according to claim 1, characterized in that: In step S3, a deviation analysis is performed on the feed quality corresponding to the historical process and the standard quality to obtain the feed quality difference value. The feed quality difference value is the percentage of the absolute difference between the measured value of the feed quality corresponding to the historical process and the standard value of the standard quality to the standard value. In the same processing stage, the historical processes and standard processes in the historical process set are analyzed for deviation to obtain the process deviation values ​​between each historical process and the corresponding standard process in the processing stage. The process deviation value is the absolute difference between the characteristic value of the historical process and the characteristic value of the standard process for each processing step.

5. The feed production, processing, and testing method based on big data according to claim 1, characterized in that: The different completion processes are the historical processes of the preceding processing nodes corresponding to this processing node; Meanwhile, the raw material quality data shows that the higher the raw material quality, the higher the initial score; The raw material quality data shows that the lower the raw material quality, the lower the initial score. The data on the positive and negative impacts of raw material quality are grouped and stored.

6. The feed production, processing, and testing method based on big data according to claim 1, characterized in that: In the replacement optimization scheme of the highest predicted quality score, if the highest predicted quality score is lower than the required quality score, the output is stopped and manual adjustment is triggered.

Images