Pig farm sand table decontamination and pollutant monitoring and early warning system

Abstract

The invention discloses a pig farm sand table decontamination and pollutant monitoring early warning system, and relates to the field of pig farm early warning. Comprising a multi-modal sensor array module which comprises a biological pollution detection unit, a chemical pollution detection unit and a radioactivity detection unit and is used for collecting pig house environment and multi-dimensional pollution data in the decontamination process in real time; and the digital twinborn decontamination deduction module renders a pollution diffusion cloud picture in real time, superposes virtual actions of decontamination equipment and predicts a pollutant diffusion path. According to the pig farm sand table decontamination and pollutant monitoring and early warning system, high-frequency acquisition of biological, chemical and radioactive pollution is realized through a multi-mode sensor array, multi-dimensional pollution feature vectors are generated through wavelet transformation and Kalman filtering preprocessing, pollution risk levels are output in combination with a fuzzy logic algorithm, yellow / orange / red three-level accurate early warning is realized, and the accuracy of pig farm sand table decontamination and pollutant monitoring and early warning is improved. Compared with traditional single-index early warning, the comprehensive early warning accuracy is improved, and meanwhile the missed alarm rate is reduced.

Description

Technical Field

[0001] The present invention relates to the field of pig farm early warning, specifically to a pig farm sand table disinfection and pollutant monitoring early warning system. Background Art

[0002] Digital twin technology has gradually been applied in breeding technology, bringing new opportunities and challenges to the development of the breeding industry. In the prior art, generally, the data of the entire pig farm is monitored, and the monitored data is compared with fixed pig growth data (for example, for pigs in a certain growth cycle, suitable growth environment data and weight data are set). If the monitored data is quite different from the growth data, a warning notice is sent to the management staff. However, the growth laws of each pig may vary, and the fixed pig growth data is only a relatively general reference data with low accuracy, which is likely to affect the accuracy of pig farm early warning; after the warning notice is sent to the management staff, the management staff still needs to manually process the warning information, and there may be a certain time difference between manual processing and warning prompts, resulting in low timeliness of early warning.

[0003] The invention patent with the publication number of CN118014184A discloses a pig farm early warning method, system and storage medium based on digital twin, which includes: obtaining the entity device information in the pig farm, and setting virtual devices corresponding to the entity devices in the digital twin pig farm based on digital twin technology; obtaining the current health status of pigs in each building based on the input data of the pre-trained pig health status detection model; comparing the current health status of pigs with the historical health status of pigs to determine whether the current health status is better than the historical health status of pigs, and obtaining the warning reference data range; iteratively updating the warning reference data range until the best warning reference data range is obtained; based on the best warning reference data range, real-time monitoring of the pig status in each building is carried out, and visual warning prompts are given according to the monitoring results; according to the visual warning prompts, the entity device processing required for warning processing is obtained, and the virtual device is controlled accordingly.

[0004] However, the existing pig farm warning methods and systems mainly rely on single - type sensors (such as only monitoring temperature and humidity), lacking multi - dimensional real - time monitoring of biological pollution (pathogens), chemical pollution (toxic gases, heavy metals), and radioactive pollution. They are unable to comprehensively capture complex pollution risks (such as the combined pollution of pathogens and chemical residues carried by vehicles), resulting in a single warning index and a relatively high false - alarm rate. Moreover, the existing technologies mostly use fixed - threshold warnings (such as only setting the safety line of pathogen concentration), without combining the pollutant diffusion kinetics model (such as Fick's law) to dynamically deduce the disinfection and decontamination process. They are unable to simulate the chemical reactions and diffusion paths of disinfectants and pollutants (such as the influence of disinfection spray pressure and angle on the pollution removal effect), leading to blind adjustment of disinfection and decontamination equipment parameters (such as excessive spray pressure causing waste of water resources and too low pressure resulting in incomplete disinfection and decontamination). The average disinfection and decontamination efficiency is lower than 60%, and the future pollution diffusion trend cannot be predicted. Summary of the Invention

[0005] In view of the deficiencies of the prior art, the present invention provides a pig farm sand table disinfection and decontamination and pollutant monitoring and warning system, which solves the existing problems.

[0006] To achieve the above objectives, the present invention is realized through the following technical solutions: The pig farm sand table disinfection and decontamination and pollutant monitoring and warning system includes: A multi - modal sensor array module, including a biological pollution detection unit, a chemical pollution detection unit, and a radioactive detection unit, for real - time collecting multi - dimensional pollution data in the pigsty environment and the disinfection and decontamination process; A digital twin disinfection and decontamination deduction module, constructing a digital twin model of disinfection and decontamination equipment and a physical model of the disinfection and decontamination process, and simulating pollutant diffusion and the effect of disinfectants based on Fick's diffusion law and chemical reaction kinetics; An intelligent warning and closed - loop control module, realizing pollution - level warning through a fuzzy logic algorithm, optimizing the disinfection and decontamination strategy by combining a reinforcement learning algorithm, and driving the physical disinfection and decontamination equipment to execute precise control instructions.

[0007] Preferably, the biological pollution detection unit uses a fluorescence - labeled sensor; the chemical pollution detection unit integrates an ion mobility spectrometry sensor and a pH sensor to real - time monitor the concentration of toxic gases and heavy metal ions; the radioactive detection unit is equipped with a gamma - ray detector.

[0008] Preferably, the digital twin model of the disinfection and decontamination equipment is constructed based on Unity3D, and real - time maps the operating parameters of the physical equipment, including spray pressure, atomized particle diameter, and drying temperature.

[0009] Preferably, the expression of the physical model of the disinfection and decontamination process is:

[0010] Where is the pollutant concentration; is the diffusion coefficient; represents the second derivative of the concentration gradient, reflecting the diffusion of pollutants from high concentration to low concentration; represents the convective term, is the air flow velocity, describing the transport effect of the air flow on pollutants; is the reaction source term of the decontamination agent, representing the chemical reaction rate between the decontamination agent and pollutants; is the time change rate of the pollutant concentration; the spatial resolution ≤ 0.5m × 0.5m.

[0011] Preferably, the reinforcement learning control algorithm adopts a deep Q-network (DQN). The state space includes pollution characteristics (12 dimensions), equipment parameters (8 dimensions), and environmental data (10 dimensions), a total of 30 dimensions. The action space includes 10 control variables such as spray pressure, angle, duration, and atomization mode. The reward function is defined as: ; where represents the -th moment reward value; is the current pollutant concentration; is the decontamination efficiency, is the consumption of the decontamination agent, is the weight coefficient, and the decontamination strategy is optimized through ≥ 500,000 times of simulation training.

[0012] Preferably, the input parameters of the fuzzy logic algorithm include biological pollution concentration, chemical pollution concentration, radioactive pollution level, and decontamination efficiency, triggering three-level warnings of yellow (0.3 ≤ R < 0.6), orange (0.6 ≤ R < 0.8), and red (R ≥ 0.8). Based on the fuzzy logic algorithm, the pollution level is comprehensively judged:

[0013] where is the fuzzy inference function, is the biological pollution concentration, is the chemical pollution concentration, is the radioactive pollution level, is the decontamination efficiency, and the output is the risk level , triggering the corresponding warning.

[0014] Preferably, the multi-modal sensor array module integrates edge computing nodes, preprocesses data using wavelet transform (noise suppression ratio ≥ 20dB) and Kalman filtering algorithm (fusion error ≤ 3%), and uploads it to the digital twin platform through the MQTT protocol.

[0015] Preferably, the digital twin decontamination deduction module renders the pollution diffusion cloud map in real time, superimposes the virtual actions of the decontamination equipment (spray direction, dynamic pressure adjustment), and predicts the pollutant diffusion path in the next 10 minutes.

[0016] Preferably, the intelligent early warning and closed-loop control module supports the automatic calibration of the parameters of the decontamination equipment by comparing the output of the digital twin model with the measured data of the physical equipment.

[0017] The present invention also discloses a monitoring and early warning method for a pig farm sand table decontamination and pollutant monitoring and early warning system, including the following steps: Step 1: Multi-source data collection and preprocessing: Multimodal sensors collect biological, chemical, and radioactive pollution data at a frequency of ≥2 seconds / time. After preprocessing by the edge computing node through wavelet transform and Kalman filtering, a standardized pollution feature vector is generated and uploaded to the digital twin platform through the MQTT protocol; Step 2: Digital twin dynamic deduction and plan: The digital twin platform, based on the Fick diffusion model and chemical reaction kinetics, combines the real-time parameters of the decontamination equipment, simulates the pollutant diffusion path, renders a 0.5m×0.5m pollution cloud map, and generates 3 sets of decontamination candidate plans; Step 3: Fuzzy logic hierarchical early warning: The intelligent module inputs parameters such as pollution characteristics and decontamination efficiency, and outputs the pollution risk level R through the fuzzy logic algorithm, triggering three-level early warnings of yellow (0.3≤R<0.6), orange (0.6≤R<0.8), and red (R≥0.8); Step 4: Optimize control instructions: For warnings of orange or above, select the optimal control instructions from the candidate plans through the Deep Q Network (DQN). The state space is 30-dimensional, and the action space has 10 control quantities. After ≥500,000 times of simulation training, the decontamination strategy is optimized; Step 5: Instruction execution and closed-loop feedback: Drive the physical decontamination equipment to execute the control instructions (such as spray pressure and angle adjustment), and feedback the decontamination effect to the digital twin platform in real time. If the equipment parameter deviation ≥5%, automatic calibration is triggered (cycle ≤24 hours) to form a "monitoring - deduction - control" closed loop.

[0018] Beneficial effects The present invention provides a pig farm sand table decontamination and pollutant monitoring and early warning system. Compared with the prior art, it has the following beneficial effects: 1. The pig farm sand table decontamination and pollutant monitoring and early warning system realizes the high-frequency collection of biological, chemical, and radioactive pollution through a multimodal sensor array (biological fluorescence labeling, chemical ion mobility spectrometry, radioactive γ-ray detector). After preprocessing by wavelet transform and Kalman filtering, a multi-dimensional pollution feature vector is generated, and the pollution risk level is output by combining the fuzzy logic algorithm, realizing accurate three-level warnings of yellow / orange / red. Compared with the traditional single-index early warning, the comprehensive early warning accuracy rate is improved, and the missed warning rate is reduced at the same time.

[0019] 2. The pig farm sand table disinfection and pollutant monitoring and early warning system constructs a digital twin model of disinfection equipment based on Unity3D, combines the Fick diffusion model to render the pollution diffusion cloud map in real time and predict the future path, optimizes the disinfection strategy through the deep Q network, realizes the dynamic adjustment of parameters such as spray pressure and angle, reduces the consumption of disinfectant, forms a monitoring - deduction - control closed loop, solves the blindness of traditional disinfection adjustment based on experience, and improves the disinfection efficiency in complex pollution scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 It is a schematic diagram of the system architecture of the present invention; Figure 2 It is a schematic diagram of the digital twin deduction process of the disinfection process of the present invention; Figure 3 It is a schematic diagram of the method flow of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0022] Refer to Figures 1-3 , the present invention provides multiple technical solutions: The first embodiment: The pig farm sand table disinfection and pollutant monitoring and early warning system includes: A multi - modal sensor array module, including a biological pollution detection unit, a chemical pollution detection unit, and a radioactive detection unit, for real - time collection of multi - dimensional pollution data in the pig house environment and the disinfection process.

[0023] The biological pollution detection unit uses a fluorescence - labeled sensor; the chemical pollution detection unit integrates an ion mobility spectrometry sensor and a pH sensor to real - time monitor the concentration of toxic gases and heavy metal ions; the radioactive detection unit is equipped with a γ - ray detector.

[0024] A digital twin disinfection deduction module, which constructs a digital twin model of disinfection equipment and a physical model of the disinfection process, and simulates the diffusion of pollutants and the effect of disinfectant based on Fick's diffusion law and chemical reaction kinetics.

[0025] The digital twin disinfection deduction module renders the pollution diffusion cloud map in real time, superimposes the virtual actions of the disinfection equipment (dynamic adjustment of spray direction and pressure), and predicts the pollutant diffusion path in the next 10 minutes.

[0026] The digital twin model of the decontamination equipment is built based on Unity3D, which can map the operating parameters of the physical equipment in real time, including spray pressure, atomized particle diameter, and drying temperature.

[0027] The physical model expression of the decontamination process is:

[0028] Where is the pollutant concentration; is the diffusion coefficient; represents the second derivative of the concentration gradient, reflecting the diffusion of pollutants from high concentration to low concentration; represents the convective term, is the air flow velocity, describing the transport effect of the air flow on pollutants; is the reaction source term of the decontamination agent, representing the chemical reaction rate between the decontamination agent and pollutants; is the time change rate of the pollutant concentration; the spatial resolution ≤ 0.5m × 0.5m.

[0029] The intelligent early warning and closed-loop control module realizes pollution classification early warning through the fuzzy logic algorithm, optimizes the decontamination strategy by combining the reinforcement learning algorithm, and drives the physical decontamination equipment to execute precise control instructions.

[0030] The intelligent early warning and closed-loop control module supports the automatic calibration of the decontamination equipment parameters by comparing the output of the digital twin model with the measured data of the physical equipment.

[0031] The reinforcement learning control algorithm uses the deep Q network (DQN). The state space includes 30 dimensions in total, including pollution characteristics (12 dimensions), equipment parameters (8 dimensions), and environmental data (10 dimensions). The action space includes 10 control variables such as spray pressure, angle, duration, and atomization mode. The reward function is defined as: ; Where represents the reward value at the th moment; is the current pollutant concentration; is the decontamination efficiency, is the consumption of the decontamination agent, is the weight coefficient, and the decontamination strategy is optimized through ≥ 500,000 times of simulation training.

[0032] The input parameters of the fuzzy logic algorithm include biological pollution concentration, chemical pollution concentration, radioactive pollution level, and decontamination efficiency, triggering three-level early warnings of yellow (0.3 ≤ R < 0.6), orange (0.6 ≤ R < 0.8), and red (R ≥ 0.8). The pollution level is comprehensively judged based on the fuzzy logic algorithm:

[0033] Among them is the fuzzy inference function, is the biological contamination concentration, is the chemical contamination concentration, is the radioactive contamination level, is the decontamination efficiency, outputting the risk level , triggering the corresponding warning.

[0034] The second implementation method: The warning method of the pig farm sand table decontamination and pollutant monitoring warning system includes the following steps: Step 1: Multi-source data collection and preprocessing: Multimodal sensors collect biological, chemical, and radioactive contamination data at a frequency of ≥2 seconds / time. After preprocessing by the edge computing node through wavelet transform and Kalman filter, a standardized contamination feature vector is generated and uploaded to the digital twin platform through the MQTT protocol; Step 2: Digital twin dynamic deduction and plan: The digital twin platform is based on the Fick diffusion model and chemical reaction kinetics, combined with the real-time parameters of the decontamination equipment, simulates the pollutant diffusion path and renders a 0.5m×0.5m contamination cloud map, generating 3 sets of decontamination candidate plans; Step 3: Fuzzy logic hierarchical warning: The intelligent module inputs parameters such as contamination features and decontamination efficiency, and outputs the contamination risk level R through the fuzzy logic algorithm, triggering three-level warnings of yellow (0.3≤R<0.6), orange (0.6≤R<0.8), and red (R≥0.8); Step 4: Optimize control instructions: For warnings of orange level and above, select the optimal control instructions from the candidate plans through the Deep Q-Network (DQN). The state space is 30-dimensional, and the action space has 10 control quantities. After ≥500,000 times of simulation training, the decontamination strategy is optimized; Step 5: Instruction execution and closed-loop feedback: Drive the physical decontamination equipment to execute control instructions (such as spray pressure and angle adjustment), and feedback the decontamination effect to the digital twin platform in real time. If the equipment parameter deviation ≥5%, automatic calibration is triggered (cycle ≤24 hours), forming a "monitoring - deduction - control" closed loop.

[0035] The third implementation method: Example Example 1: Intelligent control of the decontamination channel at the entrance of the pig house In large-scale pig farms, the disinfection and washing channel at the entrance of the pig house is a key defense line to prevent external pathogens and pollutants from entering the pig house. Transport vehicles frequently enter and leave the farm, which may carry various pathogenic bacteria, toxic chemicals, and potential radioactive pollutants. In this simulated scenario, a vehicle transporting live pigs is about to enter the pig house, and the system needs to conduct a comprehensive disinfection and washing and pollutant monitoring on it to ensure the health and safety of the pigs in the pig house. At this time, the sensor detects that the concentration of pathogenic bacteria carried by the vehicle is 80 CFU / mL (exceeding the safety threshold of 50 CFU / mL), and the ammonia concentration reaches 20 ppm (the safety threshold is 15 ppm), indicating that the vehicle has a high risk of biological and chemical pollution.

[0036] System Response Data Acquisition and Preprocessing Sensor Data Acquisition: The multi-modal sensor array module quickly starts to work. The fluorescence-labeled sensor of the biological pollution detection unit scans the vehicle surface and the surrounding environment with high precision (detection accuracy ≥ 99.9%) to detect the concentration of pathogenic bacteria; the ion mobility spectrometry sensor of the chemical pollution detection unit monitors the concentration of toxic gases such as ammonia in real time, the pH sensor detects the environmental acidity and alkalinity, and the heavy metal detector detects possible heavy metal ions; the γ-ray detector of the radioactive detection unit is also working synchronously to monitor whether the vehicle carries radioactive pollutants. All sensors collect data at a frequency of no less than 2 seconds per time to ensure the real-time and accuracy of the data.

[0037] Edge Computing Node Processing: The collected raw data is timely transmitted to the edge computing node. At the edge computing node, first, wavelet transform is used to denoise the data, and the noise suppression ratio is controlled at no less than 20 dB to effectively remove interference signals and improve the data quality. Then, the Kalman filtering algorithm is used to fuse the multi-modal data, and the fusion error is controlled at no higher than 3% to generate a standardized pollution feature vector. This vector contains multi-dimensional pollution information such as biological, chemical, and radioactive, providing an accurate data basis for subsequent digital twin deduction. Finally, the standardized pollution feature vector is uploaded to the digital twin platform through the MQTT protocol, and the latency of the entire data transmission process does not exceed 100 ms, ensuring the real-time response ability of the system.

[0038] Digital Twin Deduction and Solution Generation Model-driven and Simulation: After receiving the standardized pollution feature vector, the digital twin platform immediately uses it as input to drive the physical model of the decontamination process. This model is based on Fick's diffusion law and chemical reaction kinetics, combined with the real-time operating parameters of physical devices mapped by the digital twin model of decontamination equipment (such as spray pressure, atomized particle diameter, drying temperature, etc.), and dynamically simulates the decontamination process. During the simulation process, factors such as the diffusion of pollutants in the air, adsorption and residue on the vehicle surface, and chemical reactions between decontaminants and pollutants are fully considered.

[0039] Diffusion Prediction and Cloud Map Rendering: Through model simulation, the system can not only render the pollution diffusion cloud map in real time (resolution: 0.5m×0.5m), visually display the distribution of pollutants in the decontamination channel, but also predict the diffusion path of pollutants within the next 10 minutes, with the prediction error controlled within no more than 5%. At the same time, by superimposing the virtual actions of decontamination equipment, such as the dynamic adjustment of spray direction and pressure, operators can clearly see the decontamination effects under different decontamination strategies.

[0040] Candidate Solution Generation: According to the simulation results, the digital twin platform generates 3 groups of candidate control solutions. These 3 groups of solutions comprehensively consider factors such as decontamination efficiency, decontaminant consumption, and the impact on pigs and the environment. The solutions are combined from different aspects such as spray pressure, angle, duration, atomization mode, etc. to meet the decontamination requirements in different pollution scenarios. For example, Solution 1 may focus on high-intensity spraying to quickly reduce the concentration of pathogenic bacteria; Solution 2 may adopt a more refined atomization mode to improve the coverage rate of the decontaminant; Solution 3 may minimize the use of decontaminants while ensuring the decontamination effect. The entire solution generation process is completed within no more than 10 seconds, ensuring that the system can respond quickly.

[0041] Intelligent Early Warning and Decision Making Pollution Classification Early Warning: The intelligent module receives the pollution information output by the digital twin platform, uses the fuzzy logic algorithm, and combines input parameters such as biological pollution concentration, chemical pollution concentration, radioactive pollution level, and decontamination efficiency to calculate the pollution risk level R. After calculation, the pollution risk level R this time has reached 0.7, falling within the orange warning range (0.6≤R<0.8). The system immediately triggers an orange warning, with a response time of no more than 2 seconds, and at the same time displays the warning situation on the digital twin interface with prominent colors and prompt messages to remind operators to pay attention.

[0042] Optimal Solution Selection: For orange alerts, the intelligent module calls the reinforcement learning algorithm (Deep Q-Network, DQN) to select the optimal control instruction from 3 sets of candidate control solutions. The state space of this algorithm contains 30-dimensional information (12 dimensions of pollution characteristics, 8 dimensions of equipment parameters, and 10 dimensions of environmental data), and the action space contains 10 control variables (spray pressure, angle, duration, atomization mode, etc.). After simulation training not less than 500,000 times, the algorithm can accurately evaluate the advantages and disadvantages of each solution and select the optimal control instruction. In this scenario, after algorithm evaluation, a set of control instructions with a spray pressure of 9 MPa, an angle of 30°, and a duration of 40 seconds is selected. This solution can effectively control the usage of the decontamination agent while ensuring the decontamination effect, and the entire decision-making process is completed within no more than 500 ms.

[0043] Instruction Execution and Effect Feedback Equipment Execution and Monitoring: After selecting the optimal control instruction, the intelligent module sends it to the physical decontamination equipment. The decontamination equipment quickly adjusts its operating parameters according to the instruction. The high-pressure water gun sprays the vehicle at a pressure of 9 MPa and an angle of 30°, and the atomized spray system evenly sprays the decontamination agent on the vehicle surface in a suitable atomization mode. During the decontamination process, the sensor continuously monitors the decontamination effect and real-time feedbacks the operating status and decontamination effect data to the digital twin platform.

[0044] Closed-loop Feedback and Calibration: The digital twin platform compares and analyzes the feedback data with the simulation results. If a deviation in equipment parameters is found (such as the spray pressure deviation is not less than 5%), the automatic equipment parameter calibration mechanism is triggered. During the calibration process, the system adjusts the equipment parameters according to the actual situation to ensure that the matching accuracy of the equipment parameters is not less than 95%. The calibration period does not exceed 24 hours to ensure that the decontamination equipment is always in the best operating state. After a 40-second decontamination process, the sensor detects that the pathogen concentration has dropped to 10 CFU / mL and the ammonia concentration has dropped to 8 ppm, both meeting the safety standards, indicating that the decontamination operation has achieved good results.

[0045] Example 2: Decontamination and Monitoring of the Whole Pig Farm Area In a large pig farm, there are multiple functional areas such as pig houses, feed storage areas, and manure treatment areas. Due to factors such as the daily activities of pigs, the transportation and storage of feed, and the discharge of manure, the entire pig farm faces varying degrees of biological, chemical, and radioactive pollution risks. To ensure the overall biosafety of the pig farm, regular decontamination and real-time pollutant monitoring of the whole pig farm area are required. In this simulation scenario, the system detects that the pathogen concentration in some pig houses has increased, and at the same time, the ammonia concentration in the manure treatment area exceeds the standard, and targeted decontamination and early warning treatments are required for these areas.

[0046] System Response Comprehensive Data Collection and Analysis Distributed Sensor Network: A large number of multi-modal sensors are deployed throughout the entire pig farm to form a distributed sensor network. These sensors are distributed at key locations such as each pigsty, feed storage area, and manure treatment area, and they collect biological, chemical, and radioactive pollution data in real time. Each sensor node has independent data collection and transmission capabilities and can upload the collected data to the edge computing node in a timely manner.

[0047] Data Fusion and Feature Extraction: The edge computing node fuses and processes the data from each sensor. Through wavelet transform and Kalman filtering algorithms, the noise and interference in the data are removed, and useful pollution feature information is extracted. At the same time, correlation analysis is performed on the data in different regions to identify the sources and propagation paths of pollution. For example, by analyzing the changing trend of the pathogen concentration in the pigsty and the environmental data in the feed storage area, it is judged whether there is a situation where feed pollution leads to the spread of pathogens.

[0048] Overall Pollution Assessment: The digital twin platform conducts an overall assessment of the pollution status of the entire pig farm based on the processed data. By constructing a digital twin model of the entire pig farm, the diffusion and propagation process of pollutants between different regions is simulated, and the pollution risk levels of each region are calculated. According to the assessment results, the pig farm is divided into different pollution areas, such as high-risk areas, medium-risk areas, and low-risk areas, providing a basis for subsequent disinfection and early warning decisions.

[0049] Formulation of Zoned Disinfection Strategies Disinfection Plan for High-Risk Areas: For high-risk areas (such as pigsties with severely exceeded pathogen concentrations and manure treatment areas with extremely high ammonia concentrations), a high-intensity disinfection strategy is adopted. Increase the investment in disinfection equipment, improve the spray pressure and disinfectant concentration, and extend the disinfection time. At the same time, these areas are managed in a closed manner, restricting the entry and exit of personnel and vehicles to prevent further spread of pollutants. For example, in high-risk pigsties, professional disinfectants are used for comprehensive spraying, pigs are isolated and observed, and manure and other pollutants are promptly cleaned and harmlessly treated.

[0050] Disinfection Plan for Medium-Risk Areas: Medium-risk areas (such as pigsties with slightly increased pathogen concentrations and areas with ammonia concentrations close to the exceeded value) adopt a medium-intensity disinfection strategy. Adjust the parameters of the disinfection equipment, appropriately increase the spray frequency and disinfectant dosage, and strengthen the monitoring and management of these areas. For example, health checks are carried out on the pigs in medium-risk pigsties, and the environment is disinfected regularly to ensure that the pollution situation is effectively controlled.

[0051] Disinfection plan for low-risk areas: Conventional disinfection strategies are used in low-risk areas (such as feed storage areas and personnel activity areas with low pollution levels). Regular disinfection is carried out according to the established disinfection plan to keep the environment clean and sanitary. At the same time, ventilation in these areas is strengthened to reduce the accumulation of pollutants. For example, feed storage areas are cleaned and disinfected regularly, and personnel activity areas are cleaned and disinfected daily.

[0052] Intelligent early warning and emergency response Real-time warning mechanism: The intelligent module issues warning information in real time according to the pollution status and risk level of the entire pig farm. When the pollution index in a certain area exceeds the safety threshold, the system immediately triggers the corresponding level of warning (yellow, orange or red) and notifies the relevant management personnel through SMS, email, sound and light alarm, etc. For example, when the ammonia concentration in the manure treatment area exceeds the safety threshold, the system issues an orange warning to remind the management personnel to take timely measures to reduce the ammonia concentration.

[0053] Emergency response plan: formulate corresponding emergency response plans for different levels of warnings. When a red warning is triggered, initiate emergency disinfection and isolation measures, fully block and disinfect the contaminated area, and urgently transfer and treat the pigs. At the same time, notify the veterinarian and environmental protection department for on-site guidance and treatment. When an orange warning is triggered, strengthen the disinfection and monitoring of the contaminated area, adjust the disinfection strategy, and ensure that the pollution situation is under timely control. When a yellow warning is triggered, pay close attention to the contaminated area, appropriately increase the disinfection frequency, and prevent further deterioration of the pollution.

[0054] Data analysis and continuous optimization: During the disinfection and early warning process, the system continuously collects and analyzes data to evaluate the disinfection effect and early warning accuracy. Through the mining and analysis of historical data, potential pollution risks and problems are discovered, and the disinfection strategy and early warning mechanism are continuously optimized. For example, by analyzing multiple disinfection data, it is found that the disinfection effect in some areas is not ideal, which may be due to unreasonable layout of disinfection equipment or improper selection of disinfectant. Adjustments and improvements are made to these problems to improve the efficiency and effect of disinfection.

[0055] It can be seen from the above two embodiments that the pig farm sand table disinfection and pollutant monitoring and early warning system of the present invention can realize comprehensive, real-time and accurate disinfection and monitoring of the pig farm, effectively reduce the pollution risk of the pig farm, ensure the health and safety of pigs, and improve the production efficiency and management level of the pig farm.

[0056] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art, and the model parameters of each electrical appliance are not specifically limited, and conventional equipment can be used.

[0057] It should be noted that in this text, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or apparatus comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or apparatus.

[0058] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. Pig farm sand table cleaning and pollutant monitoring and early warning system, characterized by: include: Multimodal sensor array module, including biological contamination detection unit, chemical contamination detection unit, and radioactive detection unit, used to collect multi-dimensional contamination data of pig house environment and disinfection process in real time; The digital twin decontamination simulation module builds a digital twin model of the decontamination equipment and a physical model of the decontamination process, and simulates the diffusion of pollutants and the effect of decontamination agents based on Fick's diffusion law and chemical reaction kinetics; The intelligent early warning and closed-loop control module uses fuzzy logic algorithms to implement pollution classification early warnings, combines reinforcement learning algorithms to optimize disinfection strategies, and drives physical disinfection equipment to execute precise control instructions.

2. The pig farm sand table cleaning and pollutant monitoring and early warning system according to claim 1 is characterized by: The biological pollution detection unit adopts a fluorescent labeling sensor; the chemical pollution detection unit integrates an ion mobility spectrometer sensor and a pH sensor to monitor the concentration of toxic gases and heavy metal ions in real time; and the radioactive detection unit is equipped with a gamma-ray detector.

3. The pig farm sand table cleaning and pollutant monitoring and early warning system according to claim 1 is characterized by: The digital twin model of the disinfection equipment is built based on Unity3D, and maps the operating parameters of the physical equipment in real time, including spray pressure, atomized particle diameter and drying temperature.

4. The pig farm sand table cleaning and pollutant monitoring and early warning system according to claim 1 is characterized by: The physical model expression of the decontamination process is: in is the pollutant concentration; is the diffusion coefficient; It represents the second derivative of the concentration gradient, reflecting the diffusion of pollutants from high concentration to low concentration; represents the convection term, is the airflow velocity, which describes the transport effect of airflow on pollutants; is the disinfectant reaction source term, representing the chemical reaction rate between the disinfectant and the pollutant; It is the time rate of change of pollutant concentration; spatial resolution ≤ 0.5m×0.5m.

5. The pig farm sand table cleaning and pollutant monitoring and early warning system according to claim 1 is characterized by: The reinforcement learning control algorithm adopts a deep Q network. The state space contains pollution characteristics, equipment parameters and environmental data, and the action space contains control quantities such as spray pressure, angle, duration and atomization mode. The reward function is defined as: ； in Indicates The reward value at the moment; is the current pollutant concentration; For decontamination efficiency, is the consumption of disinfectant, The weight coefficient is used to optimize the disinfection strategy through ≥ 500,000 simulation trainings.

6. The pig farm sand table cleaning and pollutant monitoring and early warning system according to claim 1 is characterized by: The input parameters of the fuzzy logic algorithm include biological contamination concentration, chemical contamination concentration, radioactive contamination level, and decontamination efficiency, triggering yellow, orange, and red level warnings, and comprehensively judging the contamination level based on the fuzzy logic algorithm: in is the fuzzy inference function, is the biocontamination concentration, is the concentration of chemical pollution, is the level of radioactive contamination, Output risk level for decontamination efficiency , triggering the corresponding warning.

7. The pig farm sand table cleaning and pollutant monitoring and early warning system according to claim 1 is characterized by: The multimodal sensor array module integrates edge computing nodes, uses wavelet transform and Kalman filter algorithm to preprocess data, and uploads it to the digital twin platform through the MQTT protocol.

8. The pig farm sand table cleaning and pollutant monitoring and early warning system according to claim 1 is characterized by: The digital twin disinfection simulation module renders the pollution diffusion cloud map in real time, superimposes the virtual movements of the disinfection equipment, and predicts the diffusion path of pollutants.

9. The pig farm sand table cleaning and pollutant monitoring and early warning system according to claim 1 is characterized by: The intelligent early warning and closed-loop control module supports automatic calibration of disinfection equipment parameters by comparing the digital twin model output with the actual measured data of the physical equipment.

10. A monitoring and early warning method based on the pig farm sand table disinfection and pollutant monitoring and early warning system according to any one of claims 1 to 9, characterized in that: The following steps are involved: Step 1: Multi-source data collection and preprocessing: Multimodal sensors collect biological, chemical, and radioactive pollution data at a frequency of ≥ 2 seconds / time, and the edge computing node uses wavelet transform and Kalman filter preprocessing to generate standardized pollution feature vectors, which are uploaded to the digital twin platform through the MQTT protocol; Step 2: Digital twin dynamic deduction and solution: Based on the Fick diffusion model and chemical reaction kinetics, the digital twin platform combines the real-time parameters of the decontamination equipment to simulate the diffusion path of pollutants and render a 0.5m×0.5m pollution cloud map to generate three groups of decontamination candidate solutions; Step 3: Fuzzy logic graded warning: The intelligent module inputs the pollution characteristics and disinfection efficiency parameters, outputs the pollution risk level R through the fuzzy logic algorithm, and triggers yellow, orange, and red level warnings; Step 4: Optimize control instructions: For orange and above warnings, select the optimal control instructions from the candidate solutions through a deep Q network, with 30 dimensions in the state space and 10 control quantities in the action space, and optimize the decontamination strategy through simulation training; Step 5. Command execution and closed-loop feedback: Drive the physical disinfection equipment to execute control commands and provide real-time feedback of the disinfection effect to the digital twin platform. If the equipment parameter deviation is ≥5%, automatic calibration is triggered to form a monitoring-deduction-control closed loop.

Images