Pig house local dynamic ventilation system based on multi-source sensing and control method thereof

Abstract

The invention discloses a pig house local dynamic ventilation system based on multi-source sensing and a control method thereof, and relates to the technical field of agricultural engineering, and a ventilation execution module of the system comprises an attaching ventilation system, a flow guide plate and a trench fan which are arranged in a pig house; the environment sensing module comprises a temperature sensor and a gas concentration sensor which are arranged in a pigsty and a camera arranged in a pig house, and is used for collecting environment parameters and pig images in the pig house; the intelligent analysis module is in communication connection with the environment sensing module and is used for receiving and processing the environment parameters and the pig images and outputting aligned environment and pig information; and the ventilation control module is used for dynamically calculating and outputting a control instruction to the ventilation execution module according to the structured data, a preset pig type, an environment standard and a heat balance and gas balance model so as to adjust the opening degree of an air valve, the swinging angle of a guide plate and the running state of a trench fan. The uniformity and stability of ventilation adjustment in the pig house are improved by taking pigs as the basis.

Description

Technical Field

[0001] This application relates to the field of agricultural engineering technology, and in particular to a local dynamic ventilation system for pigsties based on multi-source sensing and its control method. Background Technology

[0002] Current pigsty ventilation systems mostly use fixed fans or global temperature sensors to regulate the overall ambient temperature through uniform airflow. However, some pigs experience elevated body temperatures due to illness, exercise, or overcrowding. Current technology cannot accurately pinpoint the location of each pig for targeted cooling, leading to problems with effective heat dissipation in corners or densely populated areas. Furthermore, direct cold airflow should be minimized during winter ventilation in low temperatures. In recent years, some pig farms have adopted rake-style ventilation, which improves uniformity compared to longitudinal ventilation. However, this only adjusts the airflow volume and cannot regulate the airflow direction or coordinate with pig herd information (number, distribution, average weight, and body condition, etc.), thus failing to achieve precise ventilation control and energy conservation tailored to the needs of each pig.

[0003] In the context of large-scale pig farming, a precise ventilation and environmental control method that can dynamically adjust to meet the health needs of pig herds is urgently needed to ensure healthy pig farming and reduce mortality caused by heat stress. To maximize the suitability of the microenvironment in the pigsty's living area, multi-source sensors are used to detect environmental parameters inside the pigsty, and cameras installed within the pigsty are used to determine the number, distribution, and average weight of the pigs. These parameters are then integrated to intelligently control the ventilation volume and temperature of the air supply system, achieving on-demand and uniform air delivery. Summary of the Invention

[0004] The purpose of this application is to provide a local dynamic ventilation system for pigsties based on multi-source sensing, which can improve the uniformity and stability of ventilation regulation in pigsties.

[0005] To achieve the above objectives, this application provides the following solution: In a first aspect, this application provides a local dynamic ventilation system for pigsties based on multi-source sensing, comprising: The ventilation execution module includes an attached ventilation system installed in the pigsty, a baffle plate, and a ground ventilation fan located in the manure pit; the attached ventilation system delivers air to the pigpen area through branch pipes, forming an air lake that is attached to the vertical wall and eventually spreads to the ground; the baffle plate is used to adjust the airflow direction; The environmental sensing module includes temperature sensors and gas concentration sensors installed in each pigpen, as well as cameras installed in the pigsty, for collecting environmental parameters and images of pigs in the pigsty. The intelligent analysis module is communicatively connected to the environmental perception module and is used to receive and process the environmental parameters and pig images. The intelligent analysis module has a built-in trained algorithm model for identifying the number, weight and distribution uniformity of pigs based on images, and integrating pig information and environmental parameters into structured data. The ventilation control module is communicatively connected to the intelligent analysis module and the ventilation execution module. The ventilation control module is used to dynamically calculate and output control commands to the ventilation execution module based on the structured data, preset pig type environmental standards, and real-time heat balance and gas balance models, so as to adjust the opening degree of the air valve, the swing angle of the guide plate, and the operating status of the trench fan.

[0006] Optionally, the ventilation execution module further includes an air supply system; the air supply system includes an air compressor, a heat exchanger, a filter, an air tank and a pressure valve, used to prepare constant-temperature fresh air and deliver it to the attached ventilation system through a transmission pipeline; the transmission pipeline includes a main pipe and branch pipes, the diameter of the main pipe gradually decreases to maintain stable air pressure, and the branch pipes extend to the middle of each pig pen and are connected through an air valve.

[0007] Optionally, the air valve of the attached ventilation system includes multiple arc-shaped blades driven by a motor; the blade opening range is 0% to 80%; the guide plate is driven by a servo motor to rotate up and down, and a guardrail is provided on the outside of the guide plate.

[0008] Optionally, the algorithm model in the intelligent analysis module is a convolutional neural network; the convolutional neural network includes a segmentation detection head for pig target segmentation, a posture classification head for determining whether the pig is standing or lying down, and a weight regression head for estimating the pig's weight; the weight regression head performs weight regression calculation by fusing the area information obtained from the segmentation feature map with the posture classification result.

[0009] Optionally, when training the algorithm model, the intelligent analysis module uses an auxiliary loss function to ensure a positive correlation between the segmentation area and the weight. The area consistency loss function is as follows: ; in, i For sample index, This is the actual weight. To predict the segmented area, N This represents the total number of pig samples.

[0010] Optionally, the intelligent analysis module determines the uniformity of pig distribution by calculating the area variation coefficient of the Voronoi diagram generated from the pig herd's location coordinates; when the variation coefficient is greater than the threshold of 0.5, it is determined to be a non-uniform distribution. The formula for calculating the coefficient of variation of area is: ; in, A i This represents the pixel area of ​​each pig.

[0011] Optionally, the ventilation control module adopts a parallel control strategy; in the parallel control strategy, the air volume control is based on the feedback adjustment of the air valve opening using the PID (Proportion Integration Differentiation) algorithm, the temperature control adopts a fuzzy logic controller, the input variables are the temperature deviation and the rate of change of temperature deviation, and the output variable is the number of fans turned on.

[0012] Optionally, the ventilation control module calculates the sensible heat production of the pigs using a hybrid linear model based on the pig type, weight, number, and ambient temperature and humidity, and calculates the required ventilation volume based on the sensible heat production of the pigs and the heat exchange model in the pigsty. The heat exchange model within the pigsty is as follows: ; in, For pigs to show heat, To manage the building's heat dissipation, For ventilation and heat dissipation, This refers to the amount of heat dissipated by water evaporation. The formula for calculating the ventilation volume is: ; in, For the heat load inside the building, air density, The specific heat capacity of air at constant pressure. The target temperature inside the pigsty is set. This refers to the outside air temperature.

[0013] Optionally, the control logic of the ventilation control module includes: When in low airflow mode, the minimum ventilation volume required to ensure basic ventilation and gas control is set as the lower limit, and the decision to start heating or mixed return air is made based on the outside temperature. When in high air volume mode, the maximum ventilation volume is calculated based on the heat balance equation to remove heat, and active cooling is activated when the maximum air volume is still insufficient to cool down. When uneven distribution of pigs is detected, the deflector is driven to swing to improve the uniformity of air delivery.

[0014] Secondly, this application provides a method for controlling local dynamic ventilation in pigsties based on multi-source sensing, characterized in that it is applied to a local dynamic ventilation system for pigsties based on multi-source sensing, comprising: The environmental sensing module collects environmental parameters and images of pigs inside the pigsty. The intelligent analysis module processes the environmental parameters and pig images to identify the number, weight, and distribution uniformity of the pigs, and integrates them into structured data. The ventilation control module dynamically calculates control commands based on the structured data, preset environmental standards, and balance model. According to the control command, the air valve, guide plate and trench fan in the ventilation execution module are driven to work together to achieve dynamic and precise ventilation of the local environment of the pig house.

[0015] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application provides a local dynamic ventilation system for pigsties based on multi-source sensing and its control method. By using multi-source sensors to collect real-time environmental parameters and images of pigs within the pigsty, and combining this with a deep learning algorithm in an intelligent analysis module, accurate identification of pig quantity, weight, and distribution uniformity is achieved. Based on this data and preset environmental standards, the ventilation control module employs a parallel control strategy to dynamically adjust the opening of air valves, the angle of deflectors, and the operating status of the drainage fans. This ensures precise matching of ventilation volume within the pigsty and significantly improves airflow uniformity through the synergistic effect of the attached ventilation system and deflectors. Furthermore, by coupling calculations using a hybrid linear model and a heat exchange model, the system can automatically optimize the ventilation strategy based on pig type, weight, and ambient temperature and humidity, improving the uniformity and stability of ventilation regulation within the pigsty. Attached Figure Description

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

[0017] Figure 1 This is a schematic diagram of the structure of an attached ventilation system provided in one embodiment of this application; Figure 2 A flowchart of intelligent analysis provided in one embodiment of this application; Figure 3 This is a diagram of the network regression head structure provided in an embodiment of this application; Figure 4 This is a flowchart of a method for local dynamic ventilation control in pigsties based on multi-source sensing, provided as an embodiment of this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0019] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0020] Example 1 This embodiment provides a local dynamic ventilation system for pigsties based on multi-source sensing, including: The ventilation execution module includes an attached ventilation system installed in the pigsty, a baffle plate, and a ground ventilation fan located in the manure pit; the attached ventilation system delivers air to the pigpen area through branch pipes, forming an air lake that is attached to the vertical wall and eventually spreads to the ground; the baffle plate is used to adjust the airflow direction; The environmental sensing module includes temperature sensors and gas concentration sensors installed in each pigpen, as well as cameras installed in the pigsty, for collecting environmental parameters and images of pigs in the pigsty. The intelligent analysis module is communicatively connected to the environmental perception module and is used to receive and process the environmental parameters and pig images. The intelligent analysis module has a built-in trained algorithm model for recognizing the number, weight, posture and distribution uniformity of pigs based on images, and integrating pig information and environmental parameters into structured data. The ventilation control module is communicatively connected to the intelligent analysis module and the ventilation execution module. The ventilation control module is used to dynamically calculate and output control commands to the ventilation execution module based on the structured data, preset pig type environmental standards, and real-time heat balance and gas balance models, so as to adjust the opening degree of the air valve, the swing angle of the guide plate, and the operating status of the trench fan.

[0021] The pigsty is equipped with a manure pit, several pig pens, and cameras; each pig pen is equipped with a temperature sensor and a gas concentration sensor; and the manure pit is equipped with an air outlet. Among them, such as Figure 1As shown, the attached ventilation system is the main working structure, employing an attached ventilation method. The attached ventilation system includes air valves, anemometers, impact plates, guide plates, and guardrails. The supply airflow is delivered through slotted vents located on the branch pipes. After flowing downwards for a certain distance, it is deflected towards the vertical wall due to the Coanda effect, and then flows downwards while still attached to the vertical wall. Upon reaching the ground, the airflow impacts and deflects again, spreading along the floor to form an "air lake." After exchanging heat with the indoor air, it is discharged through the outlet. Vertical wall attached ventilation allows for adjustment of the supply air volume, ensuring complete overlap between the air lake area and the pigsty floor. The maximum velocity attenuation and jet diffusion speed of the airflow under attached ventilation are much faster than other methods, which helps avoid a drafty feeling.

[0022] The damper frame is screwed onto the branch pipe. The circular air outlet precisely matches the branch pipe opening to ensure smooth airflow. The motor is mounted on the central axis of the air outlet and fixed to it with a special bracket. Specifically, the blade section consists of 10 stainless steel curved blades, which can be scaled by the motor to adjust the opening from 0% to 80%. When the opening is set to the maximum value of 80%, the movable blades will rotate behind the fixed blades.

[0023] The impact plate is used to impact the airflow. The impact plate is adjustable up and down to suit different respiratory target areas of pigs at different stages. Setting it at a higher position creates an air lake that is higher up, resulting in more efficient heat exchange.

[0024] The air outlet is located under the slatted floor, and the trench fan is installed in the manure pit to expel foul air from inside the pit. Negative pressure fans (axial flow fans or centrifugal fans) are typically used to avoid uneven airflow and ensure a stable negative pressure environment inside the shed.

[0025] The airflow deflector is used to control the direction of airflow, and is driven by a motor to swing left and right. The deflector is made of ABS plastic with good impact resistance, and each deflector blade is connected to a fixed bracket to ensure structural stability. A servo motor controls the up-and-down rotation of the deflector blades. A guardrail is installed on the outside of the deflector to prevent foreign objects from entering and coming into contact with the pigs.

[0026] The outer surface of the pigsty is covered with a thermal insulation layer, while the interior contains a manure pit. The pig pens are arranged in an array, with a main pipe installed on one side of the wall and branch pipes extending to each pen. A ventilation fan is located at the outlet of the manure pit to expel gases from inside. Cameras are mounted on the ceiling. In addition, temperature, humidity, and gas concentration sensors (also known as environmental sensors) are installed inside the pig pens.

[0027] Specifically, the pigsty consists of a sunny-facing roof, a shady-facing roof, exterior walls, a suspended ceiling, and a manure pit. The pigsty is oriented east-west, with the angle between the sunny-facing roof and the horizontal plane being the latitude of the pigsty's location or slightly greater than 15°. This effectively blocks sunlight and ensures good insulation. The sunny-facing sections of the pigsty are arranged in an array, and both the roof and exterior walls are covered with high-performance insulation layers. The pens within the pigsty are also arranged in an array.

[0028] The air supply system consists of key components such as an air compressor, heat exchanger, and filter. It is responsible for producing constant-temperature fresh air and ensuring that the output air is stable at the set pressure through an air tank and pressure valve.

[0029] Specifically, the air inlet of the supply duct connects to the outdoor environment, and the intake duct is designed with an incline to prevent rainwater backflow. Additionally, an air filter is installed to prevent particulate matter from entering the compressor. The air compressor typically produces gas at a temperature 8 to 10 degrees Celsius higher than the ambient temperature. The heat exchanger uses a shell-and-tube design, with fresh air flowing in the shell side and water circulating in the tube side. A circulating water pump calculates the water flow rate based on the heat load to directly cool or heat the fresh air to the required temperature. The air outlet connects to the air storage tank via a downward-sloping pipe to facilitate condensate drainage, and an automatic drain valve is installed at the bottom to remove condensate. The air-water separator integrates cyclone separation and a filter, removing most of the liquid water. The air-water separator inlet connects to the outlet of the aftercooler, while the outlet connects vertically upwards to the air storage tank via a flange. The air storage tank is a vertical pressure vessel. A fresh air inlet is located at the top of the air storage tank, and a drain valve is located at the bottom for timed venting. A pressure sensor and safety valve are installed on the side of the gas tank; the filter includes a condenser filter (for removing oil mist), a microporous filter, and an activated carbon filter (for adsorbing odors); the pressure valve is a pilot-operated pressure reducing valve to ensure the stability of the output pressure and is connected to the transmission pipeline.

[0030] The transmission pipeline system comprises a main duct and branch ducts forming the airflow backbone network. The main duct's diameter gradually decreases to maintain stable air pressure, while the branch ducts extend to the center of the pigpen. The main duct and branch ducts are connected via adjustable-aperture air valves. This network structure of main and branch ducts ensures that airflow is evenly distributed to each pigpen. The main duct is installed on one side of the wall, and the branch ducts extend to different pigpens, ensuring that each pen receives an adequate amount of fresh air.

[0031] Specifically, the main duct, as the backbone of the entire ventilation system, gradually decreases in diameter to ensure a constant air pressure output as distance increases. The main duct has a slope greater than 0.5% (sloping towards the drainage point) to prevent condensation buildup. The structure consists of an inner stainless steel pipe, a middle layer of 30-50 mm thick polyurethane foam insulation, and an outer layer of aluminum foil protection. The main duct runs along the long side of the pigpen, with an interface at regular intervals to connect to branch pipes, thus achieving effective coverage of the entire pigpen space. The branch pipes use a square cross-section design, extending horizontally forward to the middle of the pigpen, a certain distance from the ground. A static pressure box is added at the end of each branch pipe to ensure uniform airflow.

[0032] The design of attached ventilation includes determining the basic indoor parameters, designing and calculating the exhaust temperature, and verifying the wind speed at the horizontal distance from the air lake.

[0033] ; in, This refers to the exhaust air temperature. The dimensionless temperature rise of the air supplied to the vicinity of the ground is related to the type of indoor heat source. Determine the installation height of the exhaust vent; The vertical temperature gradient of the control area is determined at 1.0~1.5℃ / m.

[0034] Wind speed at a horizontal distance x: ; Where x is the horizontal distance from the attached wall surface; b is the width of the air outlet; and C is the shape factor, which is C=0.0075 for vertical walls. For the attachment method correction factor, for vertical walls .

[0035] The control equipment integrates temperature sensors, gas concentration sensors, cameras, and algorithms to analyze pig density, location, and ambient temperature, and dynamically adjust the opening of the air vents and the swing of the deflector.

[0036] Specifically, the camera is equipped with an infrared supplementary lighting module to ensure detection accuracy at night or in low-light conditions. The camera captures one image frame every hour and transmits it to the processor for processing.

[0037] This ventilation system structure is designed to house data acquisition, sensing, detection, and control systems.

[0038] The algorithm model in the intelligent analysis module is a convolutional neural network; the convolutional neural network includes a segmentation detection head for pig target segmentation, a posture classification head for determining whether the pig is standing or lying down, and a weight regression head for estimating the weight of the pig; the weight regression head performs weight regression calculation by fusing the area information obtained from the segmentation feature map with the posture classification result.

[0039] Specifically, for the measurement of pig numbers and average weight, a detection model is used for pig target segmentation and weight regression. Data sets captured by cameras are collected and trained, and pig outlines and pen bounding boxes are labeled. The purpose of labeling pen targets is to determine the positional relationship between pigs and pens. After training, the detection accuracy is required to be greater than 95%. After detection, based on the positional relationship between pens and pigs, the number and weight of pigs in different pens are detected.

[0040] like Figure 3 As shown, the network is modified to jointly model pose and weight, considering the impact of pose on weight measurement. In addition to the existing segmentation detection head and pose classification head, a weight regression branch head is added. Pose (lying down / standing) is a key prerequisite for weight regression. Furthermore, pixel area is positively correlated with weight, and the area-weight relationship differs under different poses. Therefore, when designing the regression head, spatial features are preserved by utilizing segmentation features (area information) and pose classification results (weight adjustment). First, this embodiment utilizes feature maps obtained from the segmentation task, which contain morphological information about the pigs. Then, this embodiment uses the pose classification results (lying down or standing) as additional input, combining them with the features. The combination method involves embedding the pose label into a vector and then fusing it with the features of the regression head.

[0041] Furthermore, an auxiliary loss is used to ensure a positive correlation between the segmented area and body weight. Let the first... i The segmentation mask prediction for each sample is as follows: (Probability graph after sigmoid function), true weight is This embodiment first calculates the predicted segmentation area: (That is, the sum of the probabilities of all pixels being predicted as pigs). Then, in this embodiment, the area consistency loss is calculated as follows: ; Area and weight were normalized (divided by the maximum value in the batch), and then the mean squared error was calculated. This encourages area and weight to maintain the same relative proportion within the batch.

[0042] The loss of different tasks is balanced by weight parameters. The weights are set according to the actual pose labels, and the loss function is: ; During training, both regression branches are trained. Even if a sample is standing, both the standing and lying branches are updated simultaneously, but through weighting, the standing branch has a higher weight (because the pose classifier assigns a higher probability of standing). During training, only the regression loss of the corresponding branch is calculated based on the real pose, while weighted fusion is used during inference. In the early stages of training, the real pose labels are used to assign regression branches (i.e., one-hot encoding of the real pose is used as weights), and as training progresses, the predicted pose probabilities are gradually transitioned to be used.

[0043] To determine uniformity, the area variance is used to judge the uniformity of the pig herd distribution. After normalization, a Voronoi diagram is generated based on the pig herd coordinates, with each point corresponding to a polygon cell.

[0044] The intelligent analysis module determines the uniformity of pig distribution by calculating the area variation coefficient of the Voronoi diagram generated from the pig herd's location coordinates; when the variation coefficient is greater than the threshold of 0.5, it is determined to be a non-uniform distribution. The formula for calculating the coefficient of variation of area is: ; in, A i This represents the pixel area of ​​each pig.

[0045] Data alignment is a crucial step in integrating pig information (number, location, weight) with environmental parameters (temperature, humidity, gas concentration) into a structured input. This system uses JSON format to define a unified data structure, including fields such as pigpen unit ID, number of pigs, average weight, location coordinates, temperature, humidity, and gas concentration, and achieves unified data management through an IoT platform (such as OneNET).

[0046] The spatial nearest neighbor algorithm is used to match the pig's location with environmental parameters: First, the real-time location of the pig is tracked by a camera to obtain coordinate information; then, the corresponding local environmental parameters are matched according to the coordinate grid (e.g., dividing the area into 10m×10m sections). For cases where the pig's location changes frequently, a sliding time window (e.g., synchronizing every 5 minutes) is used to process data with different sampling frequencies, combined with linear interpolation to compensate for delays.

[0047] In summary, based on the above algorithm model, the number, weight, and uniformity of pigs were obtained.

[0048] like Figure 2 As shown, the ventilation control module adopts a parallel control strategy; in the parallel control strategy, the air volume control is based on the feedback adjustment of the air valve opening using a PID algorithm, and the temperature control adopts a fuzzy logic controller. The input variables are the temperature deviation and the rate of change of temperature deviation, and the output variable is the number of fans turned on.

[0049] Specifically, a parallel control method with multiple inputs and dual output variables (temperature and ventilation) was proposed for pig house environmental control. A regulation strategy and control rules conforming to the changing patterns of multiple environmental factors in different seasons were constructed. Based on pig house type (lactating sow house, nursery house) and zoning, and taking into account actual production factors such as pig house structure and size, pig herd size and age, the layout of ventilation and heating equipment was first optimized. Then, based on the synergistic relationship between the internal macro-environment and the local micro-environment, the predicted values ​​of temperature and humidity changes in the pig house generated by the time-series prediction algorithm, as well as the environmental control standards for pigs at different stages, were integrated to determine the optimal thresholds for temperature and humidity environment, ventilation volume, and heating volume in local areas, thus determining the operating status of ventilation or heating equipment.

[0050] Based on this, a multi-input dual-output adaptive closed-loop control system is established to achieve precise control of the local environment of the pigsty.

[0051] Based on existing research, environmental factors can be mainly divided into two types: temperature and humidity environment and gaseous environment (mainly including carbon dioxide, ammonia, and hydrogen sulfide). The external environment affects the temperature and humidity environment inside the pigsty, while carbon dioxide is the main gaseous component inside. A closed pigsty can be considered as a microclimate system. Taking into account the effects of pigsty structure, pig herd, and ventilation, a heat exchange model within the pigsty can be constructed. ; in For pigs to show heat, To manage the building's heat dissipation, For ventilation and heat dissipation, This represents the amount of heat dissipated due to water evaporation.

[0052] Formula for calculating heat loss from pigsty enclosure structure: ; In the formula, e is the heat transfer correction coefficient of the enclosure structure; K is the heat transfer coefficient of the pigsty enclosure structure, W / ( C); A is the area of ​​the outer perimeter of the pigsty. ; t represents the temperature difference between inside and outside the building, in °C.

[0053] For airflow control, a dynamic weighted adjustment method is used to regulate the airflow. An airflow meter measures the airflow, calculates the difference between the desired and actual airflow, and uses a PID algorithm to adjust the vent opening based on feedback. Temperature control employs a fuzzy logic controller, with temperature deviation and its rate of change as input variables, and the number of fans activated as the output variable.

[0054] For regulating the air supply temperature, the heat and moisture production of pigs is taken as the target, and pig type (sow, weaner, finishing pig), weight, number, and ambient temperature and humidity are used as variable factors. Pigs are the main heat source in the barn (divided into sensible heat and latent heat). Heat production is closely related to weight, activity level, health status, and ambient temperature. The heat production per unit body weight of pigs is determined. Heat production per unit body weight is not a constant and varies significantly with the pig's weight and the current barn temperature. A mixed linear model is used to analyze the effect of different factors on the heat production of pigs. Based on the pig's weight, it can be simply expressed as: ; ; in, For sensible heat production, W; The total energy production of the pig is W; m is the weight of the pig; and Y is the type of pig. The temperature in the lactating sow area is controlled at 16℃, humidity within the range of 65%~70%, and airflow within the range of 0.15~0.4 m / s. The temperature for newborn piglets is controlled at 32℃, and then dynamically adjusted downwards to 24℃ according to their age, with temperature fluctuations controlled within 1℃, and humidity within the range of 65%~70%. For the nursery pigsty, the initial temperature is controlled at 24℃ for the first week after weaning, and then dynamically adjusted downwards to a stable temperature of 22℃, with temperature fluctuations controlled within 1℃; humidity is controlled within the range of 65%~70%, and airflow is controlled within the range of 0.2~0.6 m / s.

[0055] To maximize the suitability of the microenvironment in the pigs' living area, ensuring basic ventilation and diluting gases (especially CO2 and NH3) and moisture (latent heat) produced by pig respiration and fecal decomposition, this needs to be maintained even in cold weather. The production rate of specific gases is estimated based on the number and weight of the pigs. By fitting empirical formulas or existing data, and calculating the required airflow to dilute the gas to the permissible concentration, it is necessary to calculate the main gas... The maximum value is taken. When the gas concentration detected by the gas concentration sensor is greater than the lower limit of the air supply volume, the air volume cannot be lower than the ventilation volume required to ensure basic ventilation and gas control.

[0056] ; in, The set gas threshold; This is the concentration of the gas in the outside air (usually very low, approximating to 0).

[0057] When in low airflow mode (when the airflow is set to approximately equal to...) At this time, the air supply temperature is mainly used for ventilation, dehumidification, and cleaning. The ratio of heat load to air volume inside the building is relatively high. If the outside temperature is very low, directly supplying cold air will cause a sudden drop in the indoor temperature. Therefore, it is necessary to increase the air supply temperature to prevent cold stress. When the outside temperature is approximately equal to the air supply temperature (usually in spring and autumn or in high-volume mode), outside air is directly introduced. When the outside temperature is lower than the air supply temperature by a certain degree (common in the minimum ventilation mode in cold weather), some warm indoor return air needs to be mixed with the cold fresh air to reach the set temperature. When the outside air is extremely cold, mixing the return air alone cannot reach the air supply temperature; therefore, a heater (gas, hot water, electric) needs to be activated to heat the fresh air or the mixed air.

[0058] In hot weather, the primary task is to remove the sensible heat generated by the pigs and the heat transferred from the outside to prevent overheating. At this time, the required airflow is much greater than the minimum ventilation volume. According to the heat balance equation, in steady state, the heat brought in by the incoming air plus the heat generated inside the pigsty equals the heat removed by the exhaust air.

[0059] ; in, For the heat load inside the building, This refers to air density. This refers to the specific heat capacity of air at constant pressure. The target temperature inside the pigsty is set as the upper limit of the comfortable temperature for pigs during this stage. This refers to the outside air temperature.

[0060] High airflow mode (when the set airflow is approximately equal to...) (At this time): This is mainly used to maximize heat removal. The supply air temperature is usually very close to the outside temperature because there is not enough heat exchange capacity (or the cost is too high) to significantly change the temperature of a large flow of air. The system mainly relies on air volume to regulate temperature. When the supply air volume is at its maximum and still cannot achieve cooling (when the internal temperature is close to or even higher than the outside temperature), and the supply air temperature needs to be lower than the outside temperature, active cooling needs to be activated.

[0061] In addition, the evenness of airflow should be adjusted according to the location of the pigs. For example, at high temperatures, pigs will tend to move towards the ventilation openings. If the pigs stay at the ventilation openings, it will cause uneven distribution of the pigs. In this case, the deflector should be driven to move and the ventilation should be made more even. The deflector should be driven to swing to improve the uniformity of airflow.

[0062] Example 2 like Figure 4 As shown, this embodiment provides a method for local dynamic ventilation control in pigsties based on multi-source sensing, characterized in that it is applied to any of the aforementioned local dynamic ventilation systems for pigsties based on multi-source sensing, comprising: S1: Collect environmental parameters and images of pigs in the pigsty through the environmental sensing module; S2: The environmental parameters and pig images are processed by the intelligent analysis module to identify the number, weight and distribution uniformity of the pigs, and integrate them into structured data; S3: The ventilation control module dynamically calculates control commands based on the structured data, preset environmental standards, and balance model. S4: According to the control command, drive the air valve, guide plate and trench fan in the ventilation execution module to perform coordinated actions to achieve dynamic and precise ventilation of the local environment of the pig house.

[0063] In summary, this application has the following technical effects: This application provides a local dynamic ventilation system for pigsties based on multi-source sensing, focusing on its innovative ventilation system and intelligent control system. The ventilation system achieves effective cooling and purification of air entering the pigsty through a series of efficient components. The intelligent control system combines cameras and temperature sensors, utilizing advanced image recognition technology to identify the number, distribution, and average weight of pigs, optimizing ventilation efficiency and improving pig comfort in different seasons. A specially designed air valve and deflector structure allows for precise control of airflow direction and intensity, ensuring sufficient ventilation in high-density areas, while the suspended ventilation method avoids discomfort caused by direct cold air blowing on the pigs. These features work together to not only improve the uniformity and stability of ventilation regulation within the pigsty but also achieve energy conservation and emission reduction goals.

[0064] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0065] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A local dynamic ventilation system for pigsties based on multi-source sensing, characterized in that, include: The ventilation execution module includes an attached ventilation system installed in the pigsty, a baffle plate, and a ground ventilation fan located in the manure pit; the attached ventilation system delivers air to the pigpen area through branch pipes, forming an air lake that is attached to the vertical wall and eventually spreads to the ground; the baffle plate is used to adjust the airflow direction; The environmental sensing module includes temperature sensors, gas concentration sensors, and cameras installed in each pigpen to collect environmental parameters and images of pigs in the pigpen. The intelligent analysis module is communicatively connected to the environmental perception module and is used to receive and process the environmental parameters and pig images. The intelligent analysis module has a built-in trained algorithm model for identifying the number, weight and distribution uniformity of pigs based on images, and aligning and integrating the pig information with the environmental parameters into structured data. The ventilation control module is communicatively connected to the intelligent analysis module and the ventilation execution module. The ventilation control module is used to dynamically calculate and output control commands to the ventilation execution module based on the structured data, preset pig type environmental standards, and real-time heat balance and gas balance models, so as to adjust the opening degree of the air valve, the swing angle of the guide plate, and the operating status of the trench fan.

2. The local dynamic ventilation system for pigsties based on multi-source sensing according to claim 1, characterized in that, The ventilation execution module also includes an air supply system; the air supply system includes an air compressor, a heat exchanger, a filter, an air tank and a pressure valve, used to prepare air at a preset pressure and temperature and deliver it to the attached ventilation system through a transmission pipeline; the transmission pipeline includes a main pipe and branch pipes, the diameter of the main pipe gradually decreases to maintain stable air pressure, and the branch pipes extend to the middle of each pig pen and are connected through an air valve.

3. The local dynamic ventilation system for pigsties based on multi-source sensing according to claim 2, characterized in that, The air valve of the attached ventilation system includes multiple arc-shaped blades driven by a motor; the guide plate is driven by a servo motor to rotate up and down.

4. A local dynamic ventilation system for pigsties based on multi-source sensing according to claim 3, characterized in that, The algorithm model in the intelligent analysis module is a convolutional neural network; the convolutional neural network includes a segmentation detection head for pig target segmentation, a posture classification head for determining whether the pig is standing or lying down, and a weight regression head for estimating the weight of the pig; the weight regression head performs weight regression calculation by fusing the area information obtained from the segmentation feature map with the posture classification result.

5. A local dynamic ventilation system for pigsties based on multi-source sensing according to claim 4, characterized in that, When training the algorithm model, the intelligent analysis module uses an auxiliary loss function to ensure a positive correlation between the segmentation area and the weight. The area consistency loss function is as follows: ； in, i For sample index, This is the actual weight. A i To predict the segmented area, N This represents the total number of pig samples.

6. A local dynamic ventilation system for pigsties based on multi-source sensing according to claim 5, characterized in that, The intelligent analysis module determines the uniformity of pig distribution by calculating the area variation coefficient of the Voronoi diagram generated from the pig herd's location coordinates; when the variation coefficient is greater than the threshold of 0.5, it is determined to be a non-uniform distribution. The formula for calculating the coefficient of variation of area is: ； in, A i This represents the pixel area of ​​each pig.

7. A local dynamic ventilation system for pigsties based on multi-source sensing according to claim 6, characterized in that, The ventilation control module adopts a parallel control strategy. In the parallel control strategy, the air volume control is based on the feedback adjustment of the air valve opening using a PID algorithm, and the temperature control adopts a fuzzy logic controller. The input variables are the temperature deviation and the rate of change of the temperature deviation, and the output variable is the number of fans turned on.

8. A local dynamic ventilation system for pigsties based on multi-source sensing according to claim 1, characterized in that, The ventilation control module calculates the sensible heat production of pigs based on pig type, weight, number, and ambient temperature and humidity using a hybrid linear model, and calculates the required ventilation volume based on the sensible heat production of pigs and the heat exchange model in the pigsty. The heat exchange model within the pigsty is as follows: ； in, For pigs to show heat, To manage the building's heat dissipation, For ventilation and heat dissipation, This refers to the amount of heat dissipated by water evaporation. The formula for calculating the ventilation volume is: ； in, For the heat load inside the building, air density, The specific heat capacity of air at constant pressure. The target temperature inside the pigsty is set. This refers to the outside air temperature.

9. A local dynamic ventilation system for pigsties based on multi-source sensing according to claim 1, characterized in that, The control logic of the ventilation control module includes: When in low airflow mode, the minimum ventilation volume required to ensure basic ventilation and gas control is set as the lower limit, and the decision to start heating or mixed return air is made based on the outside temperature. When in high air volume mode, the maximum ventilation volume is calculated based on the heat balance equation to remove heat, and active cooling is activated when the maximum air volume is still insufficient to cool down. When uneven distribution of pigs is detected, the deflector is driven to swing to improve the uniformity of air delivery.

10. A method for local dynamic ventilation control in pigsties based on multi-source sensing, characterized in that, An application to a multi-source sensing-based local dynamic ventilation system for pigsties as described in any one of claims 1-9, comprising: The environmental sensing module collects environmental parameters and images of pigs inside the pigsty. The intelligent analysis module processes the environmental parameters and pig images to identify the number, weight, and distribution uniformity of the pigs, and integrates them into structured data. The ventilation control module dynamically calculates control commands based on the structured data, preset environmental standards, and balance model. According to the control command, the air valve, guide plate and trench fan in the ventilation execution module are driven to work together to achieve dynamic and precise ventilation of the local environment of the pig house.

Images