A system for simultaneous analysis of multiple indexes of dairy products

Abstract

The application discloses a kind of dairy detection multi-index synchronous analysis system, it is related to dairy production detection technical field, the system includes: multimodal data detection module, digital twin modeling module, multimodal data fusion module, quality prediction early warning module and process parameter optimization module;The application realizes the synchronous acquisition and standardization processing of the physical properties, chemical composition, microbial index and environmental data of the key nodes of dairy production whole process by multimodal data detection module, breaks through the limitation of traditional single-point detection, constructs the data network covering raw materials-processing-environment by temperature sensor and other equipment, and inputs the standardized data into digital twin model, presents the whole process production state through visual interface, dynamically monitors the correlation of process parameters and quality state, quantifies dairy quality state in combination with weighted fusion algorithm, immediately triggers early warning when risk index value exceeds threshold value.

Description

Technical Field

[0001] This invention relates to the field of dairy production testing technology, specifically a multi-indicator simultaneous analysis system for dairy product testing. Background Technology

[0002] With the development of the dairy industry towards large-scale and refined production, the demand for quality control in the dairy production process is increasing. Dairy production involves multiple stages such as raw milk receiving, sterilization, fermentation, and filling. Its quality is affected by temperature and viscosity, chemical composition, protein and fat content, microbial indicators, total bacterial count and number of pathogenic bacteria, and environmental factors. Currently, dairy companies generally use decentralized testing equipment to collect data on key process nodes and test the quality of dairy products.

[0003] Traditional dairy product testing technologies mainly rely on single-point, offline testing methods, such as manual timed sampling and testing or local monitoring with independent equipment. This approach has several shortcomings: First, data coverage is incomplete, only acquiring information from localized stages and failing to reflect the dynamic correlation between process parameters and quality status throughout the entire process. Second, response is delayed, as quality issues require offline testing for detection and feedback, easily leading to batch product non-compliance. Third, decision-making relies on experience, with process parameter adjustments depending on manual experience or trial-and-error methods, lacking scientific optimization methods based on multi-dimensional data. Traditional methods struggle to quickly pinpoint acidity deviations caused by abnormal fermenter temperatures or simulate quality change trends under different homogenization pressure combinations, thus limiting production efficiency and cost control effectiveness. Summary of the Invention

[0004] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a multi-indicator synchronous analysis system for dairy product testing. This invention, through a multimodal data detection module, achieves synchronous collection and standardized processing of physical characteristics, chemical composition, microbial indicators, and environmental data at key nodes throughout the entire dairy production process. This overcomes the limitations of traditional single-point detection. By using devices such as temperature sensors, a data network covering raw materials, processing, and the environment is constructed. Standardized data is input into a digital twin model, which updates material status and equipment operating parameters in real time based on physical entity parameters. The entire production process is presented through a visual interface, dynamically monitoring the correlation between process parameters and quality status. Combined with a weighted fusion algorithm, the quality status of dairy products is quantified. When the risk index value exceeds a threshold, an early warning is immediately triggered, shortening the response time to quality incidents.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a multi-index simultaneous analysis system for dairy product testing, the system comprising: Multimodal data detection module: By deploying various detection devices at key nodes throughout the dairy production process, multimodal data on physical properties, chemical composition, microbial indicators, and environment are collected. The collected multimodal data is then preprocessed to obtain standardized data, which is then transmitted. Digital twin modeling module: Receives standardized data, builds a digital twin model based on physical entity parameters, updates material status parameters, equipment operating parameters and environmental parameters in real time, and presents the production status through a visual interface; Multimodal data fusion module: Receives standardized data, performs fusion processing on the standardized data to obtain fused feature values, and combines digital twin models to simulate the changing trends of fused feature values ​​under different combinations of process parameters; Quality prediction and early warning module: Based on the fusion feature value and current process parameters, calculate the quality grade index value of dairy products, classify them into grades, and then calculate the risk index value. When the risk index value exceeds the preset threshold, an early warning is triggered. Process parameter optimization module: Based on the quality grade index value and risk index value, the module optimizes the core process parameters, outputs the final process parameter combination, and generates a comparison report.

[0006] Furthermore, the multimodal data detection module collects physical property data of dairy products, such as temperature, viscosity, acidity, and density, through temperature sensors, viscosity sensors, pH meters, and densitometers at key nodes in the entire dairy production process, including the raw milk receiving area, sterilization unit, homogenization equipment, fermentation tank group, finished product filling line, and production workshop environment. It also collects chemical composition data of dairy products, such as protein, fat, lactose, and additive content, through near-infrared spectroscopy and high-performance liquid chromatography. Additionally, it collects microbial index data of dairy products, such as total bacterial count and number of pathogenic bacteria, through colony counters and rapid pathogenic bacteria detectors. Finally, it collects environmental data of temperature, humidity, and cleanliness in the production workshop through temperature and humidity sensors and dust particle counters.

[0007] Furthermore, in the digital twin modeling module, the physical entity parameters include inherent parameters of the equipment, inherent parameters of the production workshop, and inherent logical parameters of the process flow. The module updates the material status, equipment operating parameters, and environmental parameters through real-time received standardized data, and outputs the real-time influence coefficient of each modal data in the current process. The module also presents the entire production process status through a visual interface.

[0008] Furthermore, the multimodal data fusion module receives standardized data and, based on the real-time impact coefficient output by the digital twin model, performs fusion processing on the standardized data through a weighted fusion algorithm to obtain a fusion feature value F, quantifying the current dairy product quality status in the production process. At the same time, it combines the digital twin model to simulate the changing trend of the fusion feature value F under different combinations of process parameters, establishing the correlation between process parameters and quality status.

[0009] Furthermore, in the multimodal data fusion module, the calculation formula for the weighted fusion algorithm is as follows: Where F is the fusion feature value, x i w is the standardized test value of the i-th data item. i The basic weight of the i-th data item is determined through historical data, α. i Let be the real-time impact coefficient of the i-th data item, and n be the total number of detected multimodal data.

[0010] Furthermore, in the quality prediction and early warning module, based on the fused feature values ​​and current process parameters, the quality grade index value of the dairy product is calculated using the finished product quality grade prediction formula, which is as follows: Where Q is the quality grade index value of dairy products, λ is the contribution weight of the fusion feature value, determined based on historical data, F is the fusion feature value, m=1,2,3,4, corresponding to the four core process parameters of sterilization temperature, homogenization pressure, fermentation time, and drying temperature, respectively, and P m P represents the real-time measured value of the m-th process parameter. m，min P m，max These are the lower and upper limits of the normal operating range for the m-th process parameter, respectively, θ m δ is the quality influence coefficient of the m-th process parameter, determined based on historical data, and δ is the process deviation correction term. When 100 ≥ Q ≥ 90, it is judged as excellent; when 90 > Q ≥ 70, it is judged as qualified; when 70 > Q ≥ 0, it is judged as unqualified.

[0011] Furthermore, in the quality prediction and early warning module, a risk index value is calculated using a quality risk index calculation formula to assess quality risk. The quality risk index calculation formula is as follows: Where R is the risk index value, Q is the quality grade index value of the dairy product, and Q 标 This refers to the quality grade standard value for dairy products. The rate of change of the integrated characteristic value F is obtained through a digital twin model, where γ is the rate sensitivity coefficient, determined based on process characteristics; An early warning is triggered when R ≥ 60.

[0012] Furthermore, in the process parameter optimization module, based on the quality grade index value and risk index value, with the goal of optimal quality and lowest risk, the four core process parameters of sterilization temperature, homogenization pressure, fermentation time and drying temperature are optimized through the process parameter optimization fitness function. The final process parameter combination is output, and a comparison report is generated to guide manual adjustment, thus completing the dynamic optimization of the production process.

[0013] Furthermore, in the process parameter optimization module, the calculation formula for the fitness function of process parameter optimization is: Among them, Fintness is the fitness value, which ranges from 0 to 1. The higher the value, the better the process parameters. Q is the quality grade index value of dairy products, and R is the risk index value.

[0014] Compared with existing technologies, this multi-index simultaneous analysis system for dairy product testing has the following advantages: I. This invention, through a multimodal data detection module, achieves simultaneous collection and standardized processing of physical characteristics, chemical composition, microbial indicators, and environmental data at key nodes throughout the entire dairy production process. This overcomes the limitations of traditional single-point detection. By using devices such as temperature sensors, a data network covering raw materials, processing, and the environment is constructed. The standardized data is then input into a digital twin model, which updates material status and equipment operating parameters in real time based on physical entity parameters. The entire production process is presented through a visual interface, dynamically monitoring the correlation between process parameters and quality status. Combined with a weighted fusion algorithm, the quality status of dairy products is quantified. When the risk index value exceeds the threshold, an early warning is immediately triggered, shortening the response time for quality incidents.

[0015] Second, this invention uses a quality prediction and early warning module to dynamically calculate the quality index based on fused feature values ​​and current process parameters, and accurately classifies the product quality grade into excellent, qualified, and unqualified grades. The process parameter optimization module aims to achieve the best quality and lowest risk, and uses a fitness function to intelligently optimize the core parameters, generating a comparison report to guide manual adjustments, thereby improving the production quality of the product, as well as the production efficiency and cost control effect. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0017] Figure 1 A flowchart of a multi-index simultaneous analysis system for dairy product testing; Figure 2 This is a framework diagram of a multi-indicator simultaneous analysis system for dairy product testing. Detailed Implementation

[0018] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below. Example 1:

[0019] Multimodal data detection module: In the scenario of large-scale production and testing of room-temperature pure milk, to achieve end-to-end quality monitoring from raw milk to finished product, various detection devices are deployed at key nodes such as the raw milk receiving area, UHT sterilization unit, high-pressure homogenization equipment, aseptic filling line, and production workshop environment. Physical property data such as temperature, viscosity, acidity, and density of the dairy products are collected through temperature sensors, viscosity sensors, pH meters, and densitometers. Chemical composition data such as protein, fat, lactose, and additive content of the dairy products are collected through near-infrared spectroscopy and high-performance liquid chromatography. Microbiological indicators such as total bacterial count and pathogenic bacteria count of the dairy products are collected through colony counters and rapid pathogenic bacteria detectors. Simultaneously, environmental data such as temperature, humidity, and cleanliness of the production workshop are collected through temperature and humidity sensors and dust particle counters. These collected multimodal data undergo preprocessing operations such as noise reduction and normalization to transform them into standardized data before transmission. Figure 1 As shown.

[0020] The digital twin modeling module receives standardized data and combines it with the inherent parameters of equipment in a normal-temperature pure milk production scenario, such as the rated power of the UHT sterilizer and the rated pressure range of the high-pressure homogenizer, as well as the inherent parameters of the production workshop, such as the workshop area and zoning, and the inherent logical parameters of the process flow from raw milk acceptance to sterilization, homogenization, and filling. It constructs a digital twin model that highly matches the actual production scenario. By receiving standardized data in real time, it dynamically updates the material status parameters during the production process, such as the remaining amount of raw milk and intermediate product inventory, and equipment operating parameters, such as the real-time temperature of the sterilizer and the real-time pressure of the homogenizer, and environmental parameters such as the real-time temperature and humidity of the workshop. Simultaneously, it outputs the real-time influence coefficients of various modal data, such as physical properties and chemical composition, in the current process, such as the sterilization process. Finally, through a visual interface, such as a production process simulation diagram, it intuitively presents the entire production process status, facilitating real-time monitoring of production dynamics.

[0021] Multimodal data fusion module: Receives standardized data and, based on the real-time influence coefficients of each modality output by the digital twin model, performs fusion processing on the standardized data using a weighted fusion algorithm to obtain fusion feature values ​​that comprehensively reflect the current quality status of dairy products in the production process. The calculation formula for the weighted fusion algorithm is as follows: Where F is the fusion feature value, x i w is the standardized test value of the i-th data item. i The basic weight of the i-th data item is determined through historical data, α. idenoted as the real-time impact coefficient of the i-th data item, and n represents the total number of multimodal data detected. Simultaneously, the digital twin model is used to simulate the changing trends of fused feature values ​​under different combinations of process parameters, such as different sterilization temperatures and homogenization pressures. The impact of process parameter adjustments on quality status is analyzed, thereby establishing a clear correlation between process parameters and dairy product quality status, providing data support for subsequent quality prediction and process optimization.

[0022] Quality Prediction and Early Warning Module: Based on the fused feature values ​​and the actual process parameters used in the current production process, the module calculates the quality grade index value of the dairy product using the finished product quality grade prediction formula. The finished product quality grade prediction formula is as follows: Where Q is the quality grade index value of dairy products, λ is the contribution weight of the fusion feature value, determined based on historical data, F is the fusion feature value, m=1,2,3,4, corresponding to the four core process parameters of sterilization temperature, homogenization pressure, fermentation time, and drying temperature, respectively, and P m P represents the real-time measured value of the m-th process parameter. m，min P m，max These are the lower and upper limits of the normal operating range for the m-th process parameter, respectively, θ m Let be the quality influence coefficient of the m-th process parameter, determined based on historical data, and δ be the process deviation correction term. When 100 ≥ Q ≥ 90, it is judged as excellent; when 90 > Q ≥ 70, it is judged as qualified; when 70 > Q ≥ 0, it is judged as unqualified, thus clarifying the current quality grade of the dairy product. Subsequently, the risk index value is calculated using the quality risk index calculation formula to assess the potential quality risks in the current production process. The quality risk index calculation formula is as follows: Where R is the risk index value, Q is the quality grade index value of the dairy product, and Q 标 This refers to the quality grade standard value for dairy products. To integrate the rate of change of the characteristic value F, a digital twin model is used, where γ is the rate sensitivity coefficient determined based on process characteristics. When R≥60, an early warning is triggered to promptly remind staff to investigate problems, prevent unqualified products from entering subsequent processes or flowing into the market, and ensure the production quality of room temperature pure milk.

[0023] Process parameter optimization module: With the core objective of achieving optimal quality and minimum quality risk in the production of room-temperature pure milk, based on quality grade index and risk index values, it calculates the optimality of core process parameters during production using a process parameter optimization fitness function. The formula for calculating the process parameter optimization fitness function is as follows: In this system, Fintness is the fitness value, ranging from 0 to 1. A higher value indicates better process parameters. Q is the quality grade index of the dairy product, and R is the risk index. The system selects the final combination of process parameters with the highest fitness value and generates a comparison report that includes a comparison between the current parameters and optimized parameters, as well as an analysis of quality and risk changes. Staff can use this report to manually adjust the process parameters of the production equipment, thereby achieving dynamic optimization of the room temperature pure milk production process and continuously improving product quality stability and production efficiency.

[0024] In summary, in the scenario of large-scale production and testing of room-temperature pure milk, the multimodal data detection module collects and preprocesses various types of data at key nodes to provide standardized data for subsequent analysis; the digital twin modeling module constructs a realistic digital twin model, updates parameters in real time, and visualizes the production status; the multimodal data fusion module calculates fusion feature values ​​to establish the correlation between process and quality; the quality prediction and early warning module calculates quality grade index and risk index values, triggering an early warning when thresholds are exceeded; and the process parameter optimization module seeks optimization with the goal of achieving optimal quality and minimum risk, outputs parameter combinations, and generates reports to assist in dynamic production optimization and ensure the stable quality of room-temperature pure milk. Example 2:

[0025] Multimodal Data Detection Module: In the production and testing of flavored fermented milk, the fermentation process significantly impacts quality. Comprehensive testing equipment is deployed at key points throughout the production process, including the raw milk receiving area, pasteurization units, stirred fermentation tanks, homogenizing equipment, finished product filling lines, and the production workshop environment. This module collects physical property data of the dairy products at different stages using temperature sensors, viscosity sensors, pH meters, and densitometers. It also collects chemical composition data such as protein, fat, lactose content, and additive content (e.g., sucrose) using near-infrared spectroscopy and high-performance liquid chromatography. Microbiological indicators, such as lactic acid bacteria counts, are collected using a colony counter. Rapid pathogen detection instruments detect the number of pathogens like Staphylococcus aureus. Simultaneously, environmental data such as temperature, humidity, and cleanliness in the production workshop are collected using temperature and humidity sensors and a dust particle counter to ensure the fermentation environment meets requirements. The collected multimodal data is preprocessed and converted into standardized data before transmission.

[0026] The digital twin modeling module receives standardized data and combines it with the inherent parameters of the equipment in the flavored fermented milk production scenario, such as the rated temperature range of the pasteurization unit and the volume and temperature control accuracy of the fermentation tank, as well as the inherent parameters of the production workshop, such as the layout and cleanliness level of the fermentation and filling areas, and the inherent logical parameters of the process flow from raw milk acceptance to sterilization, fermentation, homogenization, and filling, to construct a digital twin model of flavored fermented milk production. It updates material status parameters in real time by receiving standardized data, such as the amount of material in the fermentation tank and the freshness index of the raw milk, and equipment operating parameters, such as the pasteurization temperature and the real-time temperature of the fermentation tank, as well as environmental parameters, such as workshop temperature, humidity, and cleanliness. It also outputs the real-time impact coefficients of each modal data in the current process and presents the entire production status through a visual interface, helping staff monitor the fermentation process in real time and promptly detect equipment abnormalities or environmental fluctuations.

[0027] The multimodal data fusion module receives standardized data and, based on the real-time influence coefficients of each modality, uses a weighted fusion algorithm to process the standardized data to obtain fusion feature values. This quantitatively assesses the quality status of the flavored fermented milk in the current production stage. The calculation formula for the weighted fusion algorithm is as follows: Simultaneously, a digital twin model was used to simulate the changing trends of fusion feature values ​​under different combinations of process parameters. Different combinations of process parameters, such as different fermentation times and sterilization temperatures, were analyzed to determine the impact of process parameter adjustments on the quality of flavored fermented milk. The correlation between process parameters and quality status was established, providing a basis for subsequent quality prediction and process optimization.

[0028] Quality Prediction and Early Warning Module: Based on the fusion feature values ​​and current production process parameters, the module calculates the quality grade index value of the flavored fermented milk using the finished product quality grade prediction formula. The finished product quality grade prediction formula is as follows: When 100 ≥ Q ≥ 90, it is judged as excellent; when 90 > Q ≥ 70, it is judged as acceptable; when 70 > Q ≥ 0, it is judged as unacceptable, thus clarifying the current product quality level. Then, the risk index value is calculated using the quality risk index calculation formula to assess the potential quality risks in the current production process. The quality risk index calculation formula is as follows: When R≥60, an early warning is triggered, reminding staff to take timely intervention measures, such as adjusting the fermentation temperature, to avoid production losses due to quality problems and ensure the product quality and safety of flavored fermented milk.

[0029] Process parameter optimization module: Focusing on the production goal of achieving optimal quality and lowest risk in flavored fermented milk, based on quality grade index values ​​and risk index values, such as... Figure 2 As shown, the fitness function for optimizing core process parameters, such as sterilization temperature, homogenization pressure, and fermentation time, is used to find the optimal parameters. The formula for calculating the fitness function for these core process parameters is as follows: The system selects the final combination of process parameters with the highest adaptability and generates a comparative report that includes parameter adjustment suggestions and expected improvements in quality and risk. Staff adjust the production equipment parameters based on the report to achieve dynamic optimization of the flavored fermented milk production process, improve the consistency and stability of product taste and flavor, and meet consumers' quality requirements for flavored fermented milk.

[0030] In summary, in the production and testing scenario of flavored fermented milk, the multimodal data detection module collects and preprocesses multi-dimensional data from the entire fermentation process to meet the needs of the fermentation stage; the digital twin modeling module constructs a digital twin model by combining fermented milk production equipment, workshop, and process parameters, updates data in real time, and outputs influence coefficients; the multimodal data fusion module fuses data to obtain fusion feature values, simulating the impact of process parameters on quality; the quality prediction and early warning module assesses quality level and risk, and provides timely warnings; and the process parameter optimization module optimizes core parameters, generates reports to guide adjustments, and dynamically optimizes the production process to ensure that the quality and safety of flavored fermented milk meet the requirements.

[0031] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A multi-index simultaneous analysis system for dairy product testing, characterized in that, The system includes: Multimodal data detection module: By deploying various detection devices at key nodes throughout the dairy production process, multimodal data on physical properties, chemical composition, microbial indicators, and environment are collected. The collected multimodal data is then preprocessed to obtain standardized data, which is then transmitted. Digital twin modeling module: Receives standardized data, builds a digital twin model based on physical entity parameters, updates material status parameters, equipment operating parameters and environmental parameters in real time, and presents the production status through a visual interface; Multimodal data fusion module: Receives standardized data, performs fusion processing on the standardized data to obtain fused feature values, and combines digital twin models to simulate the changing trends of fused feature values ​​under different combinations of process parameters; Quality prediction and early warning module: Based on the fusion feature value and current process parameters, calculate the quality grade index value of dairy products, classify them into grades, and then calculate the risk index value. When the risk index value exceeds the preset threshold, an early warning is triggered. Process parameter optimization module: Based on the quality grade index value and risk index value, the module optimizes the core process parameters, outputs the final process parameter combination, and generates a comparison report.

2. The multi-index simultaneous analysis system for dairy product testing according to claim 1, characterized in that, The multimodal data detection module collects physical property data of dairy products, including temperature, viscosity, acidity, and density, at key nodes throughout the dairy production process, such as the raw milk receiving area, sterilization unit, homogenization equipment, fermentation tank group, finished product filling line, and production workshop environment. It also collects chemical composition data of dairy products, including protein, fat, lactose, and additive content, using temperature sensors, viscosity sensors, pH meters, and densitometers. Furthermore, it collects microbial index data of dairy products, including total bacterial count and pathogenic bacteria count, using colony counters and rapid pathogenic bacteria detectors. Finally, it collects environmental data of temperature, humidity, and cleanliness in the production workshop using temperature and humidity sensors and dust particle counters.

3. The multi-index simultaneous analysis system for dairy product testing according to claim 1, characterized in that, In the digital twin modeling module, the physical entity parameters include inherent parameters of the equipment, inherent parameters of the production workshop, and inherent logical parameters of the process flow. The module updates the material status, equipment operating parameters, and environmental parameters by receiving standardized data in real time, and outputs the real-time influence coefficient of each modal data in the current process. The module also presents the entire production process status through a visual interface.

4. The multi-index simultaneous analysis system for dairy product testing according to claim 1, characterized in that, In the multimodal data fusion module, standardized data is received, and based on the real-time influence coefficient output by the digital twin model, the standardized data is fused using a weighted fusion algorithm to obtain a fusion feature value F, which quantifies the current dairy product quality status in the production process. At the same time, the fusion feature value F is simulated under different combinations of process parameters using the digital twin model to establish the correlation between process parameters and quality status.

5. The multi-index simultaneous analysis system for dairy product testing according to claim 4, characterized in that, In the multimodal data fusion module, the calculation formula for the weighted fusion algorithm is as follows: Where F is the fusion feature value, x i w is the standardized test value of the i-th data item. i The basic weight of the i-th data item is determined through historical data, α. i Let be the real-time impact coefficient of the i-th data item, and n be the total number of detected multimodal data.

6. The multi-index simultaneous analysis system for dairy product testing according to claim 1, characterized in that, In the quality prediction and early warning module, based on the fused feature values ​​and current process parameters, the quality grade index value of the dairy product is calculated using the finished product quality grade prediction formula, which is as follows: Where Q is the quality grade index value of dairy products, λ is the contribution weight of the fusion feature value, determined based on historical data, F is the fusion feature value, m=1,2,3,4, corresponding to the four core process parameters of sterilization temperature, homogenization pressure, fermentation time, and drying temperature, respectively, and P m P represents the real-time measured value of the m-th process parameter. m，min P m，max These are the lower and upper limits of the normal operating range for the m-th process parameter, respectively, θ m δ is the quality influence coefficient of the m-th process parameter, determined based on historical data, and δ is the process deviation correction term. When 100 ≥ Q ≥ 90, it is judged as excellent; when 90 > Q ≥ 70, it is judged as qualified; when 70 > Q ≥ 0, it is judged as unqualified.

7. The multi-index simultaneous analysis system for dairy product testing according to claim 6, characterized in that, In the quality prediction and early warning module, a risk index value is calculated using a quality risk index calculation formula to assess quality risk. The quality risk index calculation formula is as follows: Where R is the risk index value, Q is the quality grade index value of the dairy product, and Q 标 This refers to the quality grade standard value for dairy products. The rate of change of the integrated characteristic value F is obtained through a digital twin model, where γ is the rate sensitivity coefficient, determined based on process characteristics; An early warning is triggered when R ≥ 60.

8. The multi-index simultaneous analysis system for dairy product testing according to claim 1, characterized in that, In the process parameter optimization module, based on the quality grade index value and risk index value, with the goal of optimal quality and lowest risk, the four core process parameters of sterilization temperature, homogenization pressure, fermentation time and drying temperature are optimized through the process parameter optimization fitness function. The final process parameter combination is output and a comparison report is generated to guide manual adjustment, thus completing the dynamic optimization of the production process.

9. The multi-index simultaneous analysis system for dairy product testing according to claim 8, characterized in that, In the process parameter optimization module, the formula for calculating the fitness function of process parameter optimization is: Among them, Fintness is the fitness value, which ranges from 0 to 1. The higher the value, the better the process parameters. Q is the quality grade index value of dairy products, and R is the risk index value.

Images