Pig reproductive and respiratory syndrome purification effect evaluation system based on digital twinning

Abstract

The application discloses a pig breeding and respiratory syndrome purification effect evaluation system based on digital twinning, relates to the pig farm breeding technical field, and comprises four modules of multi-source data sensing and edge correction, digital twinning fusion modeling, intelligent evaluation and dynamic reasoning, and cross-scene business application. Through the construction of a six-layer bidirectional closed-loop twinning architecture, real-time data mapping of a physical pig farm and a virtual twinning scene is realized. The system initiates a latent infection reverse recognition mechanism based on litter positive rate correction of single pig negative detection results, a PRRS triple transmission path coupling simulation model, a four-dimensional layered judgment model of signs-population-detection-time sequence, and a multi-dimensional coupling consistency purification effect quantitative scoring strategy, and supports bidirectional closed-loop regulation and control and automatic block chain acceptance and storage. The application upgrades digital twinning from passive data mirroring to active cognitive intelligent agents, and realizes a paradigm transition of PRRS purification evaluation from a static lag mode to a dynamic real-time closed-loop intelligent management.

Description

Technical Field

[0001] This invention relates to the field of pig farming technology, specifically to a digital twin-based system for evaluating the effectiveness of swine reproductive and respiratory syndrome (PRRS) eradication. Background Technology

[0002] Porcine reproductive and respiratory syndrome (PRRS) is a highly contagious disease caused by PRRSV, characterized by reproductive disorders in sows and respiratory diseases in pigs of all ages, causing huge economic losses to my country's pig industry every year. Current assessments of PRRS eradication effectiveness mainly rely on periodic antigen and antibody testing, production performance statistics, and veterinary experience. These existing assessment methods have significant shortcomings; therefore, a digital twin-based PRRS eradication effectiveness assessment system is needed.

[0003] The existing technology, such as the invention application patent with announcement number CN102599103A, discloses a method for the prevention and control of swine respiratory disease syndrome, which includes the following steps: controlling the occurrence of the main viral pathogens of swine respiratory disease syndrome from the source, firstly conducting an epidemiological survey of the main pathogens of swine respiratory disease syndrome in the pig farm to identify the main pathogens, then purifying the main viral diseases, scientifically using vaccines to protect susceptible pig herds, and for secondary bacterial infections, taking samples for bacterial culture and drug sensitivity testing, using sensitive drugs for treatment, and at the same time strengthening the comprehensive prevention and control of the pig farm to achieve the goal of controlling swine respiratory disease syndrome in self-breeding and self-raising pig farms.

[0004] Regarding the above-mentioned solutions, the inventors of this application have discovered at least the following technical problems: 1. Existing PRRS risk assessment systems are all static post-event assessment frameworks, employing a fixed-weight scoring mechanism to quantitatively assess the PRRS risk of pig farms. This assessment method cannot adaptively adjust the assessment weights based on individual factors such as the pig farm's breeding model, the genotypic characteristics of prevalent PRRSV strains, and the level of biosafety hardware, resulting in a lack of specificity and comparability in the assessment results for different pig farms. More importantly, existing assessment methods can only determine the effectiveness of measures after implementation through testing results. They cannot conduct multi-scheme comparisons before the implementation of a purification plan, nor can they dynamically adjust the assessment strategy and purification plan based on real-time monitoring data during the purification process. The assessment results suffer from a serious time lag, making it difficult to meet the actual needs of dynamic management throughout the entire PRRS purification process.

[0005] 2. Current applications of digital twin technology in the livestock industry are limited to environmental monitoring or individual health monitoring, failing to address the core of PRRS purification effect assessment. These systems focus solely on predicting air quality and providing 3D visualization for livestock and poultry farming, or constructing pollutant diffusion models, treating digital twins as passive monitoring tools for the farming environment rather than scientifically evaluating purification effects. Existing digital twin farming systems are mostly one-way data mapping models from the physical world to virtual models. The virtual scene is only used for data display and cannot act in reverse on the physical entity, lacking a two-way closed-loop mechanism from virtual assessment to physical control. Furthermore, current technologies can only perform single-dimensional environmental CFD simulations or epidemiological model analyses of various transmission pathways of PRRSV, such as aerosol cross-pen diffusion, vertical intergenerational transmission in sows, and immune escape due to strain mutations. They cannot place multiple transmission pathways under the same coupled framework for simultaneous simulation and synergistic effect analysis, making it difficult to accurately predict the virus spread range and recurrence risk under different control scenarios.

[0006] 3. Existing technologies lack effective and accurate methods for identifying latent PRRSV infection. PRRSV infection can result in widespread subclinical and persistent infection states, with some pigs carrying the virus but showing no obvious clinical symptoms. ELISA antibody testing and RT-qPCR antigen testing can both produce false negative results. Relying solely on single laboratory test data cannot accurately identify latent carriers, making these individuals the biggest blind spot and potential source of relapse in purification efforts. Furthermore, the existing system's evaluation results lack automated integration with national purification acceptance standards. Certification of purification effectiveness relies on manual compilation of scoring details, rectification reports, and record-keeping materials, which is time-consuming, costly, and prone to errors. This hinders the digitalization and automation of purification acceptance, severely restricting the large-scale promotion and standardized management of PRRS purification efforts. Summary of the Invention

[0007] To address the aforementioned technical shortcomings, the present invention aims to provide a digital twin-based system for evaluating the effectiveness of porcine reproductive and respiratory syndrome (PRRS) eradication.

[0008] To solve the above technical problems, the present invention adopts the following technical solution: The present invention provides a system for evaluating the purification effect of porcine reproductive and respiratory syndrome based on digital twins, including the following modules: Multi-source data perception and edge correction module: used to collect raw data of the target pig farm across the entire domain, and to perform pre-correction on the raw data to output a dataset of accurate infection status across the entire domain.

[0009] Digital twin fusion modeling module: used to receive corrected multi-source data, build a five-in-one exclusive twin database, configure a unique digital twin ID for each pig, and deeply bind multi-source data with single pig, single pen and whole farm virtual twin models to realize real-time data mapping between physical pig farm and virtual twin scene.

[0010] Intelligent assessment and dynamic reasoning module: used to configure the PRRS triple propagation path coupling simulation model, the latent infection stratification judgment model and the virtual and real two-way closed-loop control model, to complete the PRRS virus propagation simulation, latent infection risk judgment, purification effect quantitative scoring, and then output intelligent control strategy.

[0011] Cross-scenario business application module: used to realize scenario visualization, dynamic assessment of purification level, risk simulation and pre-drilling, closed-loop management of rectification, digital acceptance and evidence storage, and can issue control commands to link physical pig farm equipment to complete adaptive optimization.

[0012] The beneficial effects of this invention are as follows: 1. This embodiment constructs a system architecture that coordinates four modules: multi-source data perception and edge correction, digital twin fusion modeling, intelligent evaluation and dynamic reasoning, and cross-scenario business applications. This forms a closed-loop chain from data acquisition, edge correction, and twin fusion to intelligent evaluation, dynamic reasoning, and cross-scenario business applications. This architecture abandons the passive mode of unidirectional data mapping in traditional digital twin systems and systematically applies digital twin technology to the dynamic evaluation and closed-loop control of PRRS purification effects, realizing a paradigm shift from static post-event evaluation to dynamic closed-loop intelligent evaluation. The various modules of the system are tightly coupled through standardized data interfaces and control command channels. Perceived data drives the twin model and simulation evaluation, evaluation results drive the output of control strategies, control commands drive the execution of physical devices, and the execution effect is verified through feedback from the perception layer, forming a complete closed-loop logical chain. Each core innovation is accompanied by a quantifiable creative evaluation formula, providing an objective and calculable technical basis for judging the system's creative level, fundamentally overcoming the inherent defects of fragmented module stacking and subjective experience judgment in existing technologies.

[0013] 2. In this embodiment, aerosol cross-pen diffusion simulation based on computational fluid dynamics, vertical intergenerational transmission simulation in sows based on SEIR epidemiology, and strain mutation diffusion simulation based on ORF5 gene mutation rate are placed under the same coupling framework. The three sub-models achieve bidirectional data exchange through a coupling interface. The propagation coupling is creatively quantified and evaluated using a formula combining the harmonic mean and logarithmic gain. This mechanism can simultaneously output the spatiotemporal distribution of viral aerosol concentration, the evolution trajectory of infection status, and the prediction results of re-positive risk. It also supports virtual simulations of various typical scenarios, such as air filter failure and negative pressure disturbances, vertical transmission of the virus from introduced breeds, immune escape due to strain mutation, and improper disinfection of personnel and vehicles. Simultaneously... The four-dimensional hierarchical judgment mechanism of vital signs, population, detection, and time series is used to fuse four-dimensional data through multimodal cross-coupling consistency judgment. It abandons the traditional weighted fusion method and adopts a strategy that combines nonlinear cross-coupling and consistency verification. Extreme abnormal signals in any dimension will be preserved and amplified through product coupling and will not be diluted by normal values ​​in other dimensions. At the same time, the coupling consistency between four-dimensional data serves as a verification signal for the credibility of the judgment, which effectively solves the industry pain point of missed detection and misjudgment of PRRSV latent infection in existing technologies.

[0014] 3. In this embodiment, the system achieves a complete two-way closed-loop control process through a virtual-physical two-way closed-loop control model, encompassing physical pig farm status perception, virtual scene simulation evaluation, intelligent decision output, precise control of physical equipment, and scene status updates and iterations. It can issue control commands to optimize physical pig farm ventilation, spray disinfection frequency, introduction isolation duration, and sampling monitoring frequency. Furthermore, it quantifies the overall system effectiveness through creative formulas for closed-loop response and system-wide integration. This paradigm shift from passive data mirroring to proactive cognitive intelligence upgrades PRRS purification assessment from a static, lagging, and unidirectional traditional model to a dynamic, real-time, and closed-loop intelligent management model. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the system structure connection of the present invention. Detailed Implementation

[0017] 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.

[0018] Examples of embodiments of the present invention Figure 1 As shown, the porcine reproductive and respiratory syndrome (PRRS) purification effect evaluation system based on digital twins includes the following modules: multi-source data perception and edge correction module, digital twin fusion modeling module, intelligent evaluation and dynamic reasoning module, and cross-scenario business application module.

[0019] The digital twin fusion modeling module is connected to the multi-source data perception and edge correction module and the intelligent evaluation and dynamic reasoning module, respectively. The intelligent evaluation and dynamic reasoning module is connected to the cross-scenario business application module.

[0020] Multi-source data perception and edge correction module: used to collect raw data from the target pig farm across the entire domain, and to perform pre-correction on the raw data to output a precise infection status dataset across the entire domain.

[0021] In a specific embodiment, the raw data of the target pig farm is collected in the whole area. The specific collection process is as follows: Based on the functional zoning of the target pig farm, several monitoring areas are divided, and multiple types of Internet of Things (IoT) sensors are deployed in each monitoring area. The IoT sensors include aerosol samplers, temperature and humidity sensors, ventilation volume sensors and ammonia concentration sensors.

[0022] It should be noted that the monitoring area is divided based on the functional zoning of the target pig farm. The specific process is as follows: First, obtain the CAD floor plan and biosecurity level zoning map of the pig farm. Based on the actual physical layout and biosecurity control requirements of the pig farm, divide the pig farm into several functional areas, including office and living areas, external vehicle disinfection areas, personnel and material entry isolation areas, sow mating pens, farrowing pens, nursery pens, fattening pens, isolation pens, piglet loading platforms, and environmental protection areas. Then, according to the biosecurity level and spatial distribution characteristics of each functional area, one or more monitoring sub-areas are separately designated for each functional area. Finally, within each monitoring sub-area, corresponding IoT sensors are deployed according to the area size, spatial layout, and risk points to ensure that the sensors in each monitoring sub-area can effectively cover the environmental parameters and pathogen aerosol collection needs of the area, achieving full-area monitoring coverage.

[0023] Simultaneously, it acquires biosafety process data such as access control and disinfection records, material entry records, vehicle washing and disinfection records, and pig transfer records of the target pig farm; it also acquires individual pig identification information, pedigree information, weight and feed intake data at each stage recorded by RFID electronic ear tags; and it acquires time-series data of pig body temperature, respiratory rate data, and video behavior analysis data collected by intelligent monitoring equipment.

[0024] It should be noted that biosafety process data is automatically collected through the pig farm's existing information management system. Access control and disinfection records are automatically recorded by the access control system when personnel enter or exit by swiping their cards or using facial recognition, recording the personnel's identity, entry and exit time, and the operating status of the disinfection channel. Material entry records are generated by gatekeepers or warehouse managers by scanning the material's QR code or entering the material information using a handheld terminal. Vehicle disinfection records are recorded by the sensor-operated gates and disinfection equipment at the disinfection points, recording the vehicle's license plate number, disinfection time, and disinfection pressure and temperature parameters. Pig transfer records are automatically recorded by RFID readers in the transfer channel, recording the ear tag numbers of the pigs that passed through, the transfer time, and the route nodes.

[0025] RFID electronic ear tag data is obtained through fixed RFID readers and handheld RFID scanners at key nodes in the pig farm. Individual pig identification information is entered and bound to the system when piglets are born through electronic ear tags. Weight data at each stage is automatically uploaded to the system after being weighed by a smart weighing station or handheld scale. Feed intake data is automatically recorded for each pig at each meal through feed trough sensors and weighing sensors in the smart feeding station.

[0026] Pig body temperature time-series data is automatically collected every 10 to 30 minutes by miniature body temperature sensors worn on pig ear tags or collars and wirelessly transmitted to the gateway; respiratory rate data is collected non-contactly by a multispectral camera combined with a radar wave sensor at a smart monitoring station or above the pigsty; video behavior analysis data is collected around the clock by high-definition network cameras deployed in the pigsty to capture video streams of pig activities, and AI behavior recognition algorithms deployed on edge computing nodes or servers analyze the video footage in real time to automatically identify and quantify the frequency and duration of behaviors such as lying down, standing, eating, drinking, coughing, and drowsiness.

[0027] All data are synchronized with a unified time and coordinate system, forming a spatiotemporally consistent multi-source original dataset.

[0028] In a specific embodiment, the preliminary correction of the original data is performed as follows: Based on the positive rate of litter groups and the sow abortion and stillbirth signs data, the misjudgment data of single-swine ELISA antibody S / P negative and RT-qPCR antigen negative are corrected, and latent carrier individuals are marked. The latent infection correction is obtained through the following formula: The latent infection correction index was obtained. ,in, Positive rate within the same litter or group; The abnormal production rate of sow herd; Ct values ​​were detected by single-pig RT-qPCR. These are the lower and upper limits of the positive result determination threshold, respectively.

[0029] It should be noted that the specific calculation method for the positivity rate within the same litter or group is as follows: The system imports the Ct value of the PRRSV antigen RT-qPCR test for each sow or piglet and the corresponding qualitative positive / negative determination result from the laboratory testing module. The test results are grouped by litter number or pen number. The number of individuals with positive RT-qPCR test results in the same litter or pen is counted. The percentage obtained by dividing the number of positive individuals by the total number of individuals tested in the group is the positivity rate within the same litter or group. This indicator reflects the intensity of horizontal transmission of PRRSV and the prevalence of infection within a specific group.

[0030] It should also be noted that the specific calculation method for the sow herd production abnormality rate is as follows: The system imports farrowing records for each sow within the statistical period from the pig farm production management software, and calculates three indicators: abortion rate, stillbirth rate, and weak piglet rate. Specifically, the abortion rate is the percentage of sows that aborted within the statistical period out of the total number of mated sows; the stillbirth rate is the percentage of the sum of stillbirths and mummified fetuses out of the total number of piglets born; and the weak piglet rate is the percentage of piglets with a birth weight below 1.0 kg out of the total number of piglets born. The three indicators—abortion rate, stillbirth rate, and weak piglet rate—are coupled using a logarithmic product method, and the sow herd production abnormality rate is calculated using the following formula: .

[0031] Ruodang Greater than the preset correction threshold Furthermore, when a single pig tests negative, the system automatically marks that pig as a latent carrier candidate and triggers a retest warning.

[0032] Digital twin fusion modeling module: used to receive corrected multi-source data, build a five-in-one exclusive twin database, configure a unique digital twin ID for each pig, and deeply bind multi-source data with single pig, single pen and whole farm virtual twin models to realize real-time data mapping between physical pig farm and virtual twin scene.

[0033] In a specific embodiment, the real-time data mapping between the physical pig farm and the virtual twin scene is implemented as follows: First, a five-in-one dedicated twin database is constructed, which includes a pig basic information database, a production performance database, a health monitoring database, a test result database, and an environmental parameter database. Each database record is bound to a unique digital twin ID.

[0034] Secondly, a virtual twin model was constructed based on the CAD drawings of the pig farm, 3D laser scanning data, and BIM model, including the building structure of the pig house, the layout of the ventilation system, the distribution of pens, and the location of equipment.

[0035] Finally, the corrected multi-source data is precisely matched with the corresponding spatial location in the virtual twin model through a spatiotemporal registration algorithm. Sensor data is mapped to the corresponding point in the model according to its deployment coordinates, and pig data is mapped to the corresponding area in the model according to its pen location, so as to achieve real-time synchronization between the physical pig farm and the virtual twin scene.

[0036] It should be noted that, firstly, the construction of the five-in-one dedicated twin database requires a clear understanding of the relationship logic between each database. The pig basic information database uses a unique digital twin ID as the primary key. The production performance database, health monitoring database, and test result database are all linked to the basic information database through this ID. The environmental parameter database is spatiotemporally linked to pig data through pen number and collection time, thereby constructing a vertical full life cycle data chain with the digital twin ID as the core and a horizontal environmental association data chain indexed by pen-time. Secondly, the construction accuracy of the virtual twin model needs to reach the centimeter level. The pig house building structure is modeled based on CAD drawings, the ventilation system layout is modeled based on actual equipment installation drawings and pipeline routing, and the pen distribution and equipment location are accurately located based on 3D laser scanning point cloud data through point cloud registration and feature extraction algorithms, ensuring that the geometric dimensions, spatial location, and topological relationships of the virtual model are highly consistent with the physical pig farm. Finally, the specific implementation package of the spatiotemporal registration algorithm is... The process includes two stages: time synchronization and spatial registration. Time synchronization uses a unified time service on edge computing nodes to align sensor sampling timestamps, data recording timestamps, and model update timestamps to the same time base, with time synchronization accuracy controlled at the millisecond level. Spatial registration involves constructing a coordinate transformation matrix between the virtual model's spatial coordinate system and the physical pig farm's geodetic coordinate system. This transforms and maps the sensor deployment coordinates to the model's spatial coordinates, and maps the physical coordinates of the pig's pen to the corresponding area in the model using spatial indexing. Simultaneously, a data update triggering mechanism is established. When new data arrives at the sensing and acquisition layer, the spatiotemporal label of the data is automatically calculated and matched with the corresponding spatial location in the virtual model, triggering a model state update at that location. This achieves real-time synchronous mapping between the physical pig farm and the virtual twin scene, with a data update delay of no more than 500 milliseconds. This mechanism ensures that the digital twin model can follow the changes in the physical pig farm's state in real time, providing an accurate model foundation for subsequent intelligent evaluation.

[0037] Intelligent assessment and dynamic reasoning module: used to configure the PRRS triple propagation path coupling simulation model, the latent infection stratification judgment model and the virtual and real two-way closed-loop control model, to complete the PRRS virus propagation simulation, latent infection risk judgment, purification effect quantitative scoring, and then output intelligent control strategy.

[0038] In a specific embodiment, the configuration process of the PRRS triple transmission path coupling simulation model, the latent infection stratification judgment model, and the virtual-real two-way closed-loop regulation model is as follows: The configuration process of the PRRS triple transmission path coupling simulation model is as follows: Configure a computational fluid dynamics-based aerosol cross-pen diffusion simulation sub-model, a sow vertical intergenerational transmission simulation sub-model based on SEIR epidemiology, and a strain variation diffusion simulation sub-model based on ORF5 gene mutation rate. The three sub-models achieve bidirectional data exchange through a coupling interface and output the spatiotemporal distribution of viral aerosol concentration, the evolution trajectory of infection status, and the prediction results of re-positive risk.

[0039] It should be noted that the aerosol cross-pen diffusion simulation sub-model takes the three-dimensional geometric model of the target pigsty, the operating parameters of the ventilation system, and the temperature and humidity field distribution inside and outside the pigsty as input. The three-dimensional geometric model is provided by the digital twin fusion modeling module. The ventilation system parameters include the position coordinates of each air inlet and outlet, the fan model and rated air volume, the air supply and exhaust methods, the duct routing, and the valve opening. The temperature and humidity field distribution is generated by spatial interpolation of the temperature and humidity data of each monitoring point collected in real time by the multi-source data sensing and edge correction module. This sub-model solves the Navier-Stokes equations and turbulence model based on computational fluid dynamics methods to simulate the transport path, sedimentation rate, and cross-pen diffusion range of PRRSV virus aerosol particles with a particle size of 1 to 5 micrometers in and between pigsties, and outputs the spatiotemporal distribution of virus aerosol concentration at each spatial location.

[0040] The simulation sub-model of vertical intergenerational transmission in sows is based on the SEIR epidemiological model. It divides the sow population into four categories according to their immune and infection status: susceptible, exposed, infected, and recovered. Key parameters include vertical transmission rate, incubation period duration, and viral shedding period duration. The vertical transmission rate is set based on historical vertical transmission statistics of the pig farm, while the incubation period duration and viral shedding period duration are set based on clinical observation data of the prevalent strains on the farm and epidemiological parameters of the strain subtypes reported in the literature. This sub-model simulates the dynamic process of PRRSV vertical transmission from infected sows to piglets through the placenta and the viral shedding pattern of infected sows throughout the gestation and lactation periods.

[0041] The simulation sub-model for the spread of strain variation is based on the ORF5 gene sequencing data of the prevalent strain in this farm. By comparing and analyzing with the global PRRSV gene database, the gene sequence characteristics of the strain are obtained. Combined with the mutation rate of the PRRSV ORF5 gene and the immune selection pressure model, the model simulates the evolutionary path and immune escape risk of different strains in the pig herd and dynamically adjusts the infection rate and incubation period parameters of the transmission model.

[0042] The coupling interface of the three sub-models enables the following bidirectional data exchange: the viral aerosol concentration at each spatial location output by the aerosol cross-pen diffusion simulation sub-model is input as the exposure probability to the SEIR vertical transmission sub-model; the number of infected pigs and viral shedding output by the SEIR vertical transmission sub-model are fed back as virus source terms to the aerosol diffusion sub-model to update the viral concentration distribution; and the transmissibility parameters of the mutated strain output by the strain mutation diffusion sub-model dynamically adjust the infection rate and incubation period distribution parameters of the above two sub-models. The three sub-models calculate sequentially in each time step iteration and exchange data synchronously through the coupling interface, ultimately synchronously outputting the spatiotemporal distribution of viral aerosol concentration at each spatial location, the evolution trajectory of infection status of each type of pig over time, and the predicted results of re-positive risk at different time points. This coupled simulation mechanism ensures that the synergistic effect of the three transmission paths can be accurately captured, avoiding the technical deficiency of single-dimensional simulation in failing to reflect the interaction between paths.

[0043] The configuration process of the latent infection stratification judgment model is as follows: construct a four-dimensional stratification judgment mechanism of physical signs-population-detection-time series, and configure the time series physical sign data acquisition interface, the sow production abnormal data statistics interface, the positive distribution data association interface of the same pen and litter, and the laboratory test data docking interface respectively. The four-dimensional data are fused through the multimodal cross-coupling consistency judgment mechanism to output the judgment results of four infection states.

[0044] It should be noted that the temporal abnormality index is obtained by continuously collecting data on individual pig body temperature fluctuations, feed intake changes, and the frequency of coughing and lethargy through intelligent monitoring equipment, and then normalizing the data using a temporal abnormality detection algorithm. The sow production abnormality index is calculated by using a logarithmic product coupling method on three indicators: abortion rate, stillbirth rate, and weak piglet rate, abandoning the traditional weighted summation mechanism. An abnormally high index for any one indicator will lead to a significant increase in the index. The herd infection distribution index is calculated by comprehensively considering the positive rate in the same pen, the positive rate in the same litter, and the spatial clustering index. The laboratory test abnormality index is obtained by nonlinear mapping of ELISA antibody S / P value and RT-qPCR antigen Ct value. After the system performs consistency verification and outlier identification on the data of each dimension, it is fused through a mechanism combining product coupling and consistency verification. The multiplication of the abnormal indices of the four dimensions is used as the basic judgment value. Extreme abnormal signals in any dimension will not be diluted by normal values ​​in other dimensions. At the same time, the variance of the standardized scores of the four-dimensional data and the sign conflict rate are calculated. When there is a significant logical conflict, a stronger penalty is automatically applied, and the individual is marked as having a judgment conflict and requiring re-examination. The system outputs four determination states based on the comparison results of the four-dimensional coupling determination index and three preset thresholds. If the index is greater than the first threshold, it is determined to be an overt infection; if it is between the first and second thresholds, it is determined to be a low-load latent infection; if it is between the second and third thresholds, it is determined to be a borderline latent carrier; and if it is less than the third threshold, it is determined to be completely negative. The determination result is output as individual targeting information to the virtual and real two-way closed-loop control model.

[0045] The configuration process of the virtual-real two-way closed-loop control model is as follows: configure the first input interface for receiving the risk prediction information, the second input interface for receiving individual target information, and the third input interface for receiving the macroscopic positioning information, and configure the output interface for linking physical pig farm equipment through standard industrial protocols.

[0046] In a specific embodiment, the process of completing the PRRS virus transmission simulation is as follows: First, obtain the spatiotemporal distribution of virus aerosol concentration at various spatial locations in the target pigsty.

[0047] Secondly, the sow population was divided into four categories: susceptible, exposed, infected, and recovered, to simulate the vertical transmission dynamics of PRRSV from sows to piglets and the virus shedding patterns of infected sows.

[0048] Next, a sub-model of strain mutation and diffusion was constructed to simulate the evolutionary path and immune escape risk of different strains in pig herds.

[0049] It should be noted that the aerosol cross-pen diffusion simulation sub-model takes the three-dimensional geometric model of the pigsty, the operating parameters of the ventilation system, and the temperature and humidity field distribution as inputs, and outputs the spatiotemporal distribution of viral aerosol concentration by solving the Navier-Stokes equations and the turbulence model; the SEIR vertical intergenerational transmission sub-model divides sows into four categories: susceptible, exposed, infected, and recovered, and uses the vertical transmission rate and incubation period as core parameters. It simulates the vertical transmission dynamics of PRRSV from sows to piglets and the virus shedding pattern of infected sows through a system of differential equations; the strain mutation diffusion sub-model establishes a Markov chain evolution model based on ORF5 gene sequencing data, and combines immune selection pressure to predict the direction of strain mutation and the risk of immune escape. When the predicted protective efficacy is lower than the threshold, the transmissibility parameters of the mutated strain change and are fed back to the first two sub-models. The three sub-models calculate sequentially within each time step and exchange data bidirectionally through the coupling interface. Aerosol concentration is used as the exposure probability input to the SEIR model, and the viral shedding output of the SEIR model is used as the virus source term to feed back to the aerosol diffusion model. The transmissibility parameter after mutation dynamically adjusts the infection rate and incubation period parameters. The simulation is iterated until it terminates. The spatiotemporal distribution of viral aerosol concentration at each spatial location is output simultaneously to identify high-risk areas, the evolution trajectory of infection status is used to judge the epidemic trend, and the prediction result of re-positive risk is used to trigger early warning. All three serve as risk prediction information and are transmitted to the virtual and real two-way closed-loop control model.

[0050] Finally, the three propagation sub-models described above are coupled through a coupling interface to achieve bidirectional data exchange. The propagation coupling is obtained using the following formula: The PRRS virus transmission coupling index was obtained, where, Contribution to the risk of aerosol diffusion across hurdles; Contribution to the risk of vertical intergenerational transmission in sows; Contribution to the risk of virus mutation and spread.

[0051] Output the spatiotemporal distribution of viral aerosol concentration, the evolution trajectory of infection status, and the prediction results of the risk of recurrence.

[0052] In a specific embodiment, the process of determining the risk of latent infection is as follows: a four-dimensional hierarchical determination mechanism is constructed, consisting of physical signs, herd, detection, and time series. The time series dimension collects the time series physical sign abnormality index of pigs; the physical sign dimension counts the sow production abnormality index; the herd dimension correlates with the herd infection distribution index; and the detection dimension connects with the laboratory antigen and antibody detection values.

[0053] Four-dimensional data is fused through a multimodal cross-coupling consistency determination mechanism, and the four-dimensional coupling determination is obtained by the following formula: The four-dimensional coupling decision index is obtained, where, For the time-series abnormality index of pigs; This refers to the sow production abnormality index; This is the distribution index of cluster infections; Abnormal index in laboratory testing; The variance of the standardized scores for the four-dimensional data. This is the conflict penalty coefficient. The sign conflict rate between dimensions, It is a natural constant.

[0054] when Greater than the first preset threshold When it is determined to be an symptomatic infection, when Between the first threshold With the second threshold When the infection is in the range of low viral load, it is considered a latent infection. Between the second threshold With the third threshold When the interval is between, it is determined to be borderline latent carrier, when Less than the third threshold The result was determined to be completely negative, and .

[0055] In one specific embodiment, the purification effect is quantified and scored, and the specific scoring process is as follows: The original scores for each dimension are obtained, and the evaluation dimensions include the biosafety hardware dimension. Biosafety management system dimension Dimensions of pig introduction and herd management PRRS monitoring and in-group control dimensions .

[0056] A multi-dimensional coupled consistency judgment strategy is adopted, which comprehensively evaluates the purification effect by considering the synergistic effect and consistency of the original scores of each dimension. The quantitative score of the purification effect is obtained by the following formula: The purification effect was quantitatively scored, among which, The standard deviation of the original scores in the four dimensions after normalization. This represents the range ratio between dimensions.

[0057] when Greater than the first preset threshold When it is determined to be purification maintenance, when Between the first threshold With the second threshold When the time is between, it is judged that the purification standard has been met. Between the second threshold With the third threshold When it is determined to be in the process of purification, when Less than the third threshold It was determined to be unpurified at that time, and .

[0058] In a specific embodiment, the output of the intelligent control strategy is carried out in the following specific process: receiving the prediction results output by the PRRS triple propagation path coupling simulation model as risk prediction information; receiving the judgment results output by the latent infection stratification judgment model as individual targeting information; and receiving the purification effect quantification score index output by the purification effect quantification scoring model as farm-level macro positioning information.

[0059] When risk prediction information indicates that the viral load of aerosols in a certain area exceeds the standard and individual targeting information indicates that there are latent carriers in that area, the system determines that there is a risk of aerosol transmission and spread in that area, and outputs enhanced ventilation and regional isolation control strategies.

[0060] It should be noted that the enhanced ventilation and regional isolation control strategy involves the system issuing instructions through standard industrial protocols to increase the fan speed in the target area to the preset enhanced ventilation mode and continue running until the aerosol sampling test results in the area are below the risk threshold twice in a row. At the same time, physical isolation is implemented in the target area by closing the passage doors between the area and adjacent pigsties, suspending the transfer of pigs in and out of the area, and increasing the frequency of spray disinfection in the area. During the isolation period, aerosol sampling monitoring is conducted daily at the boundary of the isolation area to verify the isolation effect.

[0061] When the risk prediction information shows that the risk of re-positive results exceeds the preset re-positive threshold, the enhanced monitoring plan and the extended control strategy of group closure management will be activated.

[0062] It should be noted that the enhanced monitoring plan and extended control strategy of herd closure management are as follows: The system will automatically increase the frequency of PRRSV antigen testing for the entire herd from the current set value to twice a week, covering pig houses at all production stages. At the same time, the time window of herd closure management will be extended, the introduction of replacement gilts and the movement of pigs across houses will be suspended. The extended period of herd closure management will be at least one longest incubation period after the original end date of herd closure. During this period, serological sampling tests of the entire herd will be conducted every two weeks to assess the effectiveness of herd closure.

[0063] When individual target information indicates the presence of individuals with symptomatic infection, a group-wide screening instruction is output.

[0064] It should be noted that the whole-group screening instruction is as follows: The system automatically generates a screening plan centered on the pen where the symptomatic infected individual is located. The pen where the individual is located, along with the pens on the left and right, and all pigs in the same litter and batch that have been in contact with the infected individual, are designated as the first screening pen. All pigs in the building are designated as the second screening pen. All pigs in the first screening pen are immediately subjected to RT-qPCR antigen testing and ELISA antibody testing. All pigs in the second screening pen are subjected to RT-qPCR antigen testing. The test results are entered into the system within 24 hours, and the system automatically generates a screening report.

[0065] When individual target information shows that there are multiple cases of low-viral latent infection with spatial clustering, an encrypted sampling monitoring and adjustment elimination strategy is output.

[0066] It should be noted that the encrypted sampling monitoring and adjustment culling strategy is as follows: The system automatically increases the sampling monitoring frequency of the spatial cluster area and its adjacent pens to once a day, and continuously monitors for at least one virus shedding cycle (usually 4 to 6 weeks). At the same time, it dynamically generates a culling list based on the daily monitoring results. Individuals with two consecutive positive test results and Ct values ​​below the preset culling threshold are automatically included in the culling list. After the culling list is reviewed and confirmed by the veterinarian, it is pushed to the execution personnel. After the herd is cleared, the area is cleaned and disinfected, and the environment is verified to be negative before pigs can be reintroduced.

[0067] When macro-location information shows that the pig farm is at a medium level of purification and individual target information shows that the proportion of critical latent carriers exceeds the preset proportion threshold, the strategy of strengthening the introduction of new breeding stock and screening and controlling the entire herd will be output.

[0068] It should be noted that the strategy of strengthening the quarantine of introduced pigs and screening and controlling the entire herd is as follows: The system automatically extends the quarantine period of the subsequently introduced pigs from the standard quarantine period to an extended quarantine period. During the quarantine period, PRRSV antigen and antibody tests are conducted once on the day of entry into the quarantine pen, on the 7th day of quarantine, on the 14th day of quarantine, and before the end of quarantine. Quarantine can only be lifted if all four test results are negative. At the same time, a full coverage PRRSV antigen test is conducted on all pigs in all production stages of the entire herd. The test results are summarized by pen to form a herd infection distribution map, which is used to guide subsequent targeted culling and zoned management.

[0069] It should also be noted that the comprehensive rectification and reformulation of the purification plan and control strategy are as follows: The system automatically generates a comprehensive rectification task list based on the weak links identified by the purification effect quantitative scoring model in each dimension. This list covers three aspects: hardware facility renovation, management system revision, and monitoring plan re-editing. The rectification task list is pushed to the pig farm management after being sorted by priority. After the rectification is completed, the system re-collects the original scores of each dimension and recalculates the purification effect quantitative scoring index. If the recalculated index still does not reach the preset threshold, the rectification plan is automatically iterated and adjusted until the target is met.

[0070] Cross-scenario business application module: used to realize scenario visualization, dynamic assessment of purification level, risk simulation and pre-drilling, closed-loop management of rectification, digital acceptance and evidence storage, and can issue control commands to link physical pig farm equipment to complete adaptive optimization.

[0071] In a specific embodiment, the implementation process of scene visualization, dynamic assessment of purification level, risk simulation and pre-drilling, closed-loop management of rectification, and digital acceptance and evidence storage is as follows: The digital twin model is presented on the terminal device, supporting free roaming, perspective switching and layer control to achieve scene visualization.

[0072] Based on the purification effect quantitative scoring index output by the purification effect quantitative scoring model, the pig farm is automatically divided into four levels: unpurified, medium-purified, purified to standard, and purified to maintain, so as to realize dynamic evaluation of purification level.

[0073] In response to user-defined control scenario parameters, the system automatically invokes the PRRS triple propagation path coupling simulation model to perform virtual simulations, outputting risk evolution trajectories and intervention effect predictions, thereby achieving risk simulation and pre-playing.

[0074] The system automatically generates rectification task orders for risk points identified in the assessment and pushes them to the responsible persons. It tracks the progress of rectification and verifies the effectiveness, forming a complete closed loop of discovery, task assignment, rectification and acceptance, and realizing closed-loop management of rectification.

[0075] Based on the dynamic evaluation results, the system automatically generates scoring details, rectification reports, and ledger files that conform to the evaluation standards for disease purification farms set by the Ministry of Agriculture and Rural Affairs, thus achieving digital acceptance and evidence preservation.

[0076] The examples described in this invention are not limited to the specific embodiments listed above. The examples are merely illustrative to facilitate understanding of the invention and do not constitute a limitation on the scope of protection of this invention. Any modifications, equivalent substitutions, etc., made within the spirit and principles of this invention should be included within the scope of protection.

[0077] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. A digital twin-based system for evaluating the effectiveness of porcine reproductive and respiratory syndrome (PRRS) eradication, characterized in that, Includes the following modules: Multi-source data perception and edge correction module: used to collect raw data from the target pig farm across the entire domain, and to perform pre-correction on the raw data to output a precise infection status dataset across the entire domain; Digital twin fusion modeling module: used to receive corrected multi-source data, build a five-in-one exclusive twin database, configure a unique digital twin ID for each pig, and deeply bind multi-source data with single pig, single pen and whole farm virtual twin models to realize real-time data mapping between physical pig farm and virtual twin scene; Intelligent assessment and dynamic reasoning module: used to configure the PRRS triple propagation path coupling simulation model, the latent infection stratification judgment model and the virtual and real two-way closed-loop control model, to complete the PRRS virus propagation simulation, latent infection risk judgment, purification effect quantitative scoring, and then output intelligent control strategy. Cross-scenario business application module: used to realize scenario visualization, dynamic assessment of purification level, risk simulation and pre-drilling, closed-loop management of rectification, digital acceptance and evidence storage, and can issue control commands to link physical pig farm equipment to complete adaptive optimization.

2. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 1, characterized in that, The raw data from the target pig farms is collected across the entire region. The specific collection process is as follows: Based on the functional zoning of the target pig farm, several monitoring areas are divided, and various types of IoT sensors are deployed in each monitoring area. The IoT sensors include aerosol samplers, temperature and humidity sensors, ventilation volume sensors and ammonia concentration sensors. Simultaneously, it acquires biosafety process data such as access control and disinfection records, material entry records, vehicle washing and disinfection records, and pig transfer records of the target pig farm; it also acquires individual pig identification information, pedigree information, weight and feed intake data at each stage recorded by RFID electronic ear tags; and it acquires pig body temperature time series data, respiratory rate data, and video behavior analysis data collected by intelligent monitoring equipment. All data are synchronized with a unified time and coordinate system, forming a spatiotemporally consistent multi-source original dataset.

3. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 2, characterized in that, The preliminary correction of the original data is performed as follows: Based on the positive rate of litter groups and the signs of abortion and stillbirth in sows, misjudgment data for single-swine ELISA antibody S / P negative and RT-qPCR antigen negative were corrected, and latent carriers were marked. The correction for latent infection was obtained using the following formula: The latent infection correction index was obtained. ,in, Positive rate within the same litter or group; The abnormal production rate of sow herd; Ct values ​​were detected by single-pig RT-qPCR. These are the lower and upper limits of the positive result determination threshold, respectively. Ruodang Greater than the preset correction threshold Furthermore, when a single pig tests negative, the system automatically marks that pig as a latent carrier candidate and triggers a retest warning.

4. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 3, characterized in that, The specific implementation process for achieving real-time data mapping between the physical pig farm and the virtual twin scenario is as follows: First, a five-in-one dedicated twin database is constructed, which includes a pig basic information database, a production performance database, a health monitoring database, a test result database, and an environmental parameter database. Each database record is bound to a unique digital twin ID. Secondly, a virtual twin model was constructed based on the pig farm's CAD drawings, 3D laser scanning data, and BIM model, including the pig house's building structure, ventilation system layout, pen distribution, and equipment location. Finally, the corrected multi-source data is precisely matched with the corresponding spatial location in the virtual twin model through a spatiotemporal registration algorithm. Sensor data is mapped to the corresponding point in the model according to its deployment coordinates, and pig data is mapped to the corresponding area in the model according to its pen location, so as to achieve real-time synchronization between the physical pig farm and the virtual twin scene.

5. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 4, characterized in that, The configuration process for the PRRS triple propagation path coupling simulation model, the latent infection stratification determination model, and the virtual-real two-way closed-loop control model is as follows: The configuration process of the PRRS triple transmission path coupling simulation model is as follows: configure a simulation sub-model of aerosol cross-pen diffusion based on computational fluid dynamics, a simulation sub-model of vertical intergenerational transmission in sows based on SEIR epidemiology, and a simulation sub-model of strain variation diffusion based on ORF5 gene mutation rate. The three sub-models achieve bidirectional data exchange through the coupling interface and output the spatiotemporal distribution of viral aerosol concentration, the evolution trajectory of infection status, and the prediction results of re-positive risk. The configuration process of the latent infection stratification judgment model is as follows: construct a four-dimensional stratification judgment mechanism of physical signs-population-detection-time series, configure the time series physical sign data acquisition interface, the sow production abnormal data statistics interface, the positive distribution data association interface of the same pen and litter, and the laboratory test data docking interface respectively. The four-dimensional data are fused through the multimodal cross-coupling consistency judgment mechanism to output the judgment results of four infection states. The configuration process of the virtual-real two-way closed-loop control model is as follows: configure the first input interface for receiving the risk prediction information, the second input interface for receiving individual target information, and the third input interface for receiving the macroscopic positioning information, and configure the output interface for linking physical pig farm equipment through standard industrial protocols.

6. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 5, characterized in that, The specific process for completing the PRRS virus propagation simulation is as follows: First, obtain the spatiotemporal distribution of viral aerosol concentration in various spatial locations of the target pigsty; Secondly, the sow population was divided into four categories: susceptible, exposed, infected, and recovered, to simulate the vertical transmission dynamics of PRRSV from sows to piglets and the virus shedding patterns of infected sows. Next, a sub-model of strain mutation and diffusion was constructed to simulate the evolutionary path and immune escape risk of different strains in pig herds. Finally, the three propagation sub-models described above are coupled through a coupling interface to achieve bidirectional data exchange. The propagation coupling is obtained using the following formula: The PRRS virus transmission coupling index was obtained, where, Contribution to the risk of aerosol diffusion across hurdles; Contribution to the risk of vertical intergenerational transmission in sows; Contribution to the risk of strain mutation and spread; Output the spatiotemporal distribution of viral aerosol concentration, the evolution trajectory of infection status, and the prediction results of the risk of recurrence.

7. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 6, characterized in that, The process for determining the risk of asymptomatic infection is as follows: A four-dimensional hierarchical judgment mechanism is constructed, encompassing physical signs, herd, detection, and time series. The time series dimension collects the time-series abnormal physical signs index of pigs; the physical signs dimension statistically analyzes the abnormal production index of sows; the herd dimension correlates with the herd infection distribution index; and the detection dimension connects with laboratory antigen and antibody detection values. Four-dimensional data is fused through a multimodal cross-coupling consistency determination mechanism, and the four-dimensional coupling determination is obtained by the following formula: The four-dimensional coupling decision index is obtained, where, For the time-series abnormality index of pigs; This refers to the sow production abnormality index; This is the distribution index of cluster infections; Abnormal index in laboratory testing; The variance of the standardized scores for the four-dimensional data. This is the conflict penalty coefficient. The sign conflict rate between dimensions, It is a natural constant; when Greater than the first preset threshold When it is determined to be an symptomatic infection, when Between the first threshold With the second threshold When the infection is in the range of low viral load, it is considered a latent infection. Between the second threshold With the third threshold When the interval is between, it is determined to be borderline latent carrier, when Less than the third threshold The result was determined to be completely negative, and .

8. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 7, characterized in that, The quantitative scoring of the purification effect is carried out through the following process: Obtain the original scores for each dimension, including the biosafety hardware dimension. Biosafety management system dimension Dimensions of pig introduction and herd management PRRS monitoring and in-group control dimensions ; A multi-dimensional coupled consistency judgment strategy is adopted to comprehensively evaluate the purification effect by considering the synergistic effect and consistency of the original scores of each dimension. The quantitative score of the purification effect is obtained by the following formula: The purification effect was quantitatively scored, among which, The standard deviation of the original scores in the four dimensions after normalization. The ratio of the range between dimensions; when Greater than the first preset threshold When it is determined to be purification maintenance, when Between the first threshold With the second threshold When the time is between, it is judged that the purification standard has been met. Between the second threshold With the third threshold When it is determined to be in the process of purification, when Less than the third threshold It was determined to be unpurified at that time, and .

9. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 8, characterized in that, The specific output process for the intelligent control strategy is as follows: The system receives prediction results from the PRRS triple propagation path coupling simulation model as risk prediction information; and receives judgment results from the latent infection stratification judgment model as individual targeting information. The purification effect quantitative scoring index output by the purification effect quantitative scoring model is received as macro-level positioning information for pig farms. When risk prediction information shows that the viral load of aerosols in a certain area exceeds the standard and individual targeting information shows that there are asymptomatic carriers in the area, the system determines that there is a risk of aerosol transmission and spread in the area and outputs enhanced ventilation and regional isolation control strategies. When the risk prediction information shows that the risk of re-positive results exceeds the preset re-positive threshold, the enhanced monitoring plan and the group closure management extended control strategy will be activated. When individual target information indicates the presence of individuals with symptomatic infection, a group-wide screening instruction is output; When individual target information shows that there are multiple cases of low-viral latent infection with spatial clustering, an encrypted sampling monitoring and adjustment elimination strategy is output. When macro-location information shows that the pig farm is at the medium level of purification and individual target information shows that the proportion of critical latent carriers exceeds the preset proportion threshold, the strategy of strengthening the introduction isolation and whole-herd screening and control will be output. When macro-location information indicates that the pig farm is in an unpurified level, a comprehensive rectification and re-formulation of purification plan and control strategy will be implemented.

10. The porcine reproductive and respiratory syndrome (PRRS) eradication effect evaluation system based on digital twins according to claim 9, characterized in that, The specific implementation process for achieving scenario visualization, dynamic assessment of cleanliness levels, risk simulation and pre-drilling, closed-loop management of rectification, and digital acceptance and evidence storage is as follows: Presenting digital twin models on terminal devices allows for free roaming, perspective switching, and layer control, enabling scene visualization. Based on the purification effect quantitative scoring index output by the purification effect quantitative scoring model, the pig farm is automatically divided into four levels: unpurified, purified, purified to standard, and purified to maintain, so as to realize dynamic evaluation of purification level. In response to user-defined control scenario parameters, the system automatically invokes the PRRS triple propagation path coupling simulation model to perform virtual simulations, outputting risk evolution trajectories and intervention effect predictions, thereby achieving risk simulation and pre-playing. The risk points identified in the assessment are automatically generated into rectification task orders and pushed to the responsible persons. The progress of rectification and the verification of the effect are tracked, forming a complete closed loop of discovery-assignment-rectification-acceptance, and realizing closed-loop management of rectification. Based on the dynamic evaluation results, the system automatically generates scoring details, rectification reports, and ledger files that conform to the evaluation standards for disease purification farms set by the Ministry of Agriculture and Rural Affairs, thus achieving digital acceptance and evidence preservation.

Images