Method for temperature control of a mammal piglet incubator

Abstract

This invention discloses a temperature control method for a piglet warming box. The invention involves acquiring the current environmental parameters of the piglet warming box; inputting these parameters into a preset environmental parameter prediction model to obtain predicted environmental parameters for the piglet warming box; determining a first control command based on the predicted environmental parameters; determining a second control command based on the current temperature and a preset temperature threshold from the current environmental parameters; and controlling the temperature of the piglet warming box according to the first and second control commands. This invention improves the control efficiency of the piglet warming environment during slaughter.

Description

Technical Field

[0001] This application relates to the field of piglet warming environment control technology, and in particular to a temperature control method for a piglet warming box. Background Technology

[0002] In pig farming, high sow productivity is one of the goals pursued to maximize profits. If sows have excellent reproductive performance and can produce multiple piglets per litter, successfully raising a large number of piglets, then large-scale pig farming can be achieved. Temperature is one of the key factors affecting the mortality rate, weight gain, and energy consumption of suckling piglets. Existing temperature control methods for suckling piglet heat lamps typically use equipment such as "heat lamp + lampshade" or "heated floor + heat lamp box," controlling the on / off status of the heating equipment through manual switching and automatic temperature control systems. This method relies on manual operation and is relatively inefficient. Summary of the Invention

[0003] The main technical problem addressed by this application is to provide a temperature control method for a piglet incubator, which can at least improve the temperature control efficiency.

[0004] To solve the above-mentioned technical problems, the temperature control method for a piglet warming box provided in this application includes the following steps: acquiring the current environmental parameters of the piglet warming box; inputting the current environmental parameters into a preset environmental parameter prediction model to obtain the predicted environmental parameters of the piglet warming box, and determining a first control command based on the predicted environmental parameters; determining a second control command based on the current temperature and a preset temperature threshold in the current environmental parameters; and controlling the temperature of the piglet warming box according to the first control command and the second control command.

[0005] In one embodiment, the preset environmental parameter prediction model is a graph convolution model, in which each insulation box is a node. The graph convolution model includes: constructing a static distance matrix based on the physical straight-line distance between each insulation box and a preset attenuation rate control parameter; constructing a dynamic wind field matrix based on the fan vector of each insulation box and the position vector between each insulation box; and normalizing the static distance matrix and the dynamic wind field matrix to obtain the dynamic adjacency matrix.

[0006] In one embodiment, the preset environmental parameter prediction model is a graph convolutional model, and the method includes: using the predicted values ​​and actual values ​​of the graph convolutional model as model prediction constraints; determining a loss function based on the model prediction constraints, temporal smoothing constraints, energy conservation constraints, and thermal health index constraints; and training the graph convolutional model based on the loss function to obtain the preset environmental parameter prediction model.

[0007] In one embodiment, the method includes: acquiring the temperature change of the piglet warming box within a preset time period, the heating efficiency and heat dissipation efficiency of the piglet warming box; and determining the energy conservation constraint based on the temperature change, the heating efficiency, the heat dissipation efficiency and the preset input power.

[0008] In one embodiment, the predicted environmental parameters include a predicted temperature, and the step of determining the first control command based on the predicted environmental parameters includes: acquiring the heat production of piglets in each piglet warming box; determining the heat dissipation of the piglet warming box based on the predicted temperature and the indoor temperature of the piglet warming box; determining the predicted temperature change based on the heating efficiency of the piglet warming box, the heat production of the piglets, and the heat dissipation of the piglet warming box; and determining the first control command based on the predicted temperature change.

[0009] In one embodiment, the current environmental parameters include the current temperature, and the step of obtaining the heat production of each piglet in the incubator includes: determining the body temperature difference of the piglets based on a preset critical low temperature and the current temperature; and determining the heat production of each piglet in the incubator based on the weight of each piglet and the body temperature difference.

[0010] In one embodiment, the step of determining the second control command based on the current temperature and the preset temperature threshold in the current environmental parameters includes: obtaining the environmental temperature difference between the current temperature and the preset temperature threshold; determining the second heating power based on the environmental temperature difference; and determining the second control command based on the second heating power.

[0011] In one embodiment, the step of controlling the temperature of the piglet warming box according to the first control command and the second control command includes: determining a target heating power according to a first heating power in the first control command and a second heating power in the second control command; heating the piglet warming box based on the target heating power to achieve temperature control of the piglet warming box.

[0012] To solve the above-mentioned technical problems, this application provides a piglet warming box, which includes: a controller for executing the temperature control method of the piglet warming box described above; a sensor group connected to the controller, the sensor group being configured to collect the current environmental parameters of the piglet warming box; and an actuator connected to the controller, the actuator being configured to perform temperature control.

[0013] In one embodiment, the piglet warming box further includes: a box body; a heater fixed inside the box body; a temperature sensor fixed in a slot on the side panel of the box body; and a sliding plate groove installed on both sides of the box body and connected to the side panel of the box body.

[0014] Beneficial effects: By acquiring the current environmental parameters of the piglet warming box; inputting the current environmental parameters into a preset environmental parameter prediction model to obtain the predicted environmental parameters of the piglet warming box; and determining the first control command based on the predicted environmental parameters; determining the second control command based on the current temperature and a preset temperature threshold in the current environmental parameters; and controlling the temperature of the piglet warming box according to the first and second control commands. Thus, by determining the first control command through predicted environmental parameters and the second control command through the current temperature, the temperature of the piglet warming box is controlled, achieving automatic temperature regulation and improving efficiency. Attached Figure Description

[0015] 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 described 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, wherein: Figure 1 This is a schematic flowchart of an exemplary embodiment of the temperature control method for a piglet incubator shown in this application; Figure 2 This is a schematic flowchart of yet another exemplary embodiment of the temperature control method for a piglet incubator shown in this application; Figure 3 This is a schematic diagram of the structure of a piglet warming box, as shown in an exemplary embodiment of this application. Detailed Implementation

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are only for explaining this application and not for limiting it. Furthermore, it should be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all structures. 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.

[0017] like Figure 1 As shown, a temperature control method for a piglet warming box according to this embodiment includes the following steps: Step S110: Obtain the current environmental parameters of the piglet incubator.

[0018] A piglet warming box is a box used to raise piglets with a controlled temperature.

[0019] Current environmental parameters refer to the environmental parameters of the piglet rearing environment. These parameters may include the current temperature and humidity of the piglet rearing environment, as well as the internal temperature and humidity of the piglet warming box.

[0020] The temperature control device acquires the current environmental parameters of the piglet warming box. Specifically, the temperature control device includes a sensor array that collects data on the temperature and humidity inside the piglet rearing environment and the warming box to obtain the current environmental parameters. The sensor array includes temperature sensors, humidity sensors, etc.

[0021] Step S120: Input the current environmental parameters into the preset environmental parameter prediction model to obtain the predicted environmental parameters of the piglet incubator, and determine the first control command based on the predicted environmental parameters.

[0022] The temperature control device inputs the current environmental parameters into a preset environmental parameter prediction model to obtain the predicted environmental parameters of the piglet incubator.

[0023] Predicted environmental parameters include predicted temperature and humidity in the piglet rearing environment.

[0024] The temperature control device determines a first control command based on predicted environmental parameters. Specifically, the temperature control device obtains the difference between the predicted temperature and a preset temperature threshold; in response to a difference greater than zero, it increases the heating power of the heater in the piglet warming box as the first control command; in response to a difference less than zero, it decreases the heating power of the heater in the piglet warming box as the first control command.

[0025] Step S130: Determine the second control command based on the current temperature and the preset temperature threshold in the current environmental parameters.

[0026] The temperature control device responds to a situation where the difference between the current temperature and the preset temperature threshold in the current environmental parameters is greater than zero by increasing the heating power of the heater in the piglet warming box as a second control command; and responds to a situation where the difference is less than zero by decreasing the heating power of the heater in the piglet warming box as a second control command.

[0027] Step S140: Control the temperature of the piglet incubator according to the first control command and the second control command.

[0028] The temperature control device triggers the actuator of the piglet incubator to execute the first control command, and then executes the second control command.

[0029] After executing the first control command, the temperature control device re-detects the current temperature. If the difference between the re-detected current temperature and the preset temperature threshold is greater than zero, it will increase the heating power of the heater in the piglet warming box as a second control command and execute the second control command. In response to the difference between the re-detected current temperature and the preset temperature threshold being less than zero, it will decrease the heating power of the heater in the piglet warming box as a second control command and execute the second control command.

[0030] As can be seen, by acquiring the current environmental parameters of the piglet warmer, inputting these parameters into a preset environmental parameter prediction model, and obtaining the predicted environmental parameters of the piglet warmer, a first control command is determined based on the predicted environmental parameters. A second control command is then determined based on the current temperature and a preset temperature threshold from the current environmental parameters. Finally, the temperature of the piglet warmer is controlled according to both the first and second control commands. Thus, by determining the first control command through predicted environmental parameters and the second control command through the current temperature, the temperature of the piglet warmer is controlled, achieving automatic temperature regulation and improving efficiency.

[0031] The preset environmental parameter prediction model is a graph convolution model, in which each insulation box is a node, which can be represented as a graph. , where nodes For a nursing piglet incubator.

[0032] The graph convolution model includes: constructing a static distance matrix based on the physical straight-line distance between each insulation box and the preset attenuation rate control parameters; constructing a dynamic wind field matrix based on the fan vector of each insulation box and the position vector between each insulation box; and normalizing the static distance matrix and the dynamic wind field matrix to obtain the dynamic adjacency matrix.

[0033] For example, the static distance matrix satisfies the following formula:

[0034] In the above formula, Characterizing the static distance matrix, Representation Nodes (Insulated box) ) and nodes (Insulated box) The physical straight-line distance between them can be calculated using the Euclidean distance formula; The preset decay rate control parameter is characterized by the standard deviation of the Gaussian kernel function or a length scale parameter. This preset decay rate control parameter is a hyperparameter that controls the decay rate with distance. The smaller the value, the faster the thermal radiation correlation between insulated boxes that are even slightly far apart will drop to zero. A preset distance threshold is used to conserve computational resources when the distance between two insulated boxes is greater than a certain threshold. At that time, it is forcibly assumed that there is no static heat conduction or radiation effect between them, which is 0.

[0035] According to fluid mechanics principles, upwind nodes have a significant thermal convection influence on downwind nodes. Therefore, the dynamic wind field matrix incorporates the wind turbine state vector. (Including the number of wind turbines in operation and their frequency), let the wind direction vector be... The position vector from node i to j is The dynamic wind field matrix satisfies the following formula:

[0036] In the above formula, Characterized by a linear rectified unit (Rectified Linear Unit), its property is that when the input value... Output as is. The output is In the above formula, the effect is only calculated when the wind direction is towards the node (the included angle is less than 90 degrees, and the dot product is positive), which strictly blocks the heat conducted by the "headwind". The dot sign in the numerator indicates the dot product of two vectors. Combined with the multiplication of the vector magnitudes in the denominator, the calculation within the parentheses represents the cosine of the angle between the wind direction vector and the nodal position vector. ); The magnitude (norm) of a vector; Characterizing the mapping function. The mapping function represents the intensity of the wind turbine state (such as the number of turbines in operation, frequency). The mapping is amplified by the actual convective heat transfer weights.

[0037] The static distance matrix and the dynamic wind field matrix are normalized to obtain the dynamic adjacency matrix, as shown below:

[0038] In the above formula, The dynamic adjacency matrix is ​​represented by γ, the learnable balance coefficient is represented by Norm, and the row normalization is represented by Norm.

[0039] It can be seen that traditional GCNs typically use a fixed adjacency matrix, but in the piglet warming box, the thermal interactions between nodes (edges) Due to the strong influence of the ventilation system, the near-distance heat diffusion caused by thermal radiation and natural convection was captured by the static distance matrix. In the dynamic wind field matrix, when node j is located downwind of node i and the wind speed is high, the weight increases significantly; otherwise, the weight is 0 (ReLU cuts off the influence of the headwind). Therefore, the influence between the various piglet heat preservation boxes is fully considered, which helps to improve the accuracy of the model.

[0040] The graph convolution model also includes a graph convolution operator, which uses Chebyshev polynomial approximation at each time step to achieve efficient graph convolution and avoid complex feature decomposition, as shown below:

[0041] In the above formula, The input feature matrix that represents the node's current state, such as the current temperature and humidity of the insulated box, historical sequence, etc. The upper bound of the order of a Chebyshev polynomial is represented; in graph neural networks, it represents the number of hops in the receptive field, for example... This means that the insulated box is not only affected by the adjacent insulated boxes, but also by the "adjacent adjacent" insulated boxes; Characterization Chebyshev polynomials are a type of polynomial used to simplify the Laplace matrix. A mathematical approximation tool for eigenvalue decomposition. Characterizing the scaled Laplacian matrix, These are the convolution kernel parameters. This step achieves the aggregation of spatial features: each insulated box not only utilizes its own historical data, but also aggregates the temperature and humidity information of the surrounding and upwind insulated boxes.

[0042] The temperature control device embeds graph convolution operations into the gating mechanism of the GRU, replacing the traditional fully connected layer, to obtain a GRU unit integrated with GCN, thereby obtaining a prediction model of preset environmental parameters.

[0043] set up Let F be the feature matrix of the entire network at time t (N×F). This is the hidden state from the previous moment. Then reset the door to , can be represented as ; Update the gate to , can be represented as ; Candidate hidden state is , can be represented as ; The final hidden state is , can be represented as .

[0044] In the above formula, Characterized by the Sigmoid activation function, this activation function is used to compress the output value to... arrive Between them, the opening and closing coefficients, used as "gates," determine which historical information should be forgotten and which new information should be retained; Characterized by the hyperbolic tangent activation function, which is used to compress the output to... arrive Between these points, the feature vector at the current time step is generated; The concatenation operation is used to concatenate data from the current time step. and the hidden state of the previous moment They are pieced together along the feature dimension; Graph convolutional network operations that characterize the reset gate. Characterizes the bias term corresponding to the reset gate; Graph convolutional network operations for characterization update gates Characterize the bias term corresponding to the update gate. Graph convolutional network operations that characterize candidate hidden states. Bias terms characterizing candidate hidden states; The Hadamard product (elemental multiplication) represents the direct multiplication of the elements at corresponding positions of two matrices of the same shape.

[0045] In one embodiment, the preset environmental parameter prediction model is a graph convolution model. The method includes: using the predicted values ​​and actual values ​​of the graph convolution model as model prediction constraints; determining a loss function based on the model prediction constraints, temporal smoothing constraints, energy conservation constraints, and thermal health index constraints; and training the graph convolution model based on the loss function to obtain the preset environmental parameter prediction model.

[0046] In one embodiment, the thermal health index constraint can be a temperature and humidity index, as shown below: =0.8· + ·( -14.4)+46.4 In the above formula, Characterizing temperature and humidity index, Characterizing indoor temperature, Characterizing indoor humidity, Characterizes relative humidity.

[0047] It can be seen that by using the temperature and humidity index specifically for piglets as the target for model optimization, when humidity is detected... An increase in temperature leads to an increase in THI; the system temperature should be appropriately reduced. Conversely, the same applies. Compared to simply considering temperature, this multi-dimensional comfort-based control strategy, which optimizes the model using the temperature and humidity index, is significantly superior to the single-dimensional control of most current control systems.

[0048] In one embodiment, the method includes: acquiring the temperature change of the piglet warming box within a preset time period, the heating efficiency and heat dissipation efficiency of the piglet warming box; and determining energy conservation constraints based on the temperature change, heating efficiency, heat dissipation efficiency and preset input power.

[0049] Based on the simplified heat balance equation, the predicted rate of temperature change should be approximately proportional to the input heating power and the calculated heat dissipation. For example, the energy conservation constraint satisfies the following formula:

[0050] In the above formula, Characterizes the amount of temperature change that occurs within the predicted time period. The time step represented by the model prediction, such as how many minutes have passed; These respectively characterize the simplified terms of the physical constants that the model needs to learn or that are preset. Characterizes heating efficiency, that is, the efficiency of converting heating power into temperature rise; Characterizes heat dissipation efficiency, which is the rate at which heat is dissipated due to the temperature difference between the inside and outside. Characterizes the input electrical power of the heater; Characterizing the predicted chamber temperature, It represents the ambient temperature inside the building, that is, the current temperature.

[0051] Among them, heating efficiency corresponds to Heat dissipation power corresponding .

[0052] The temperature control command applies weighted processing to time-domain smoothing constraints, energy conservation constraints, and thermal health index constraints to obtain a weighted value. The sum of the weighted value and the model prediction constraints is determined as the deterministic loss function.

[0053] For example, the loss function satisfies the following formula:

[0054] In the above formula, Characterizing the loss function, Characterization model prediction constraints The mean squared error is used to ensure prediction accuracy, where N is the total number of samples involved in the calculation. The predicted values ​​of the characterization model Characterizes the true value returned by the sensor; The time-domain smoothing constraint is used to penalize abrupt temperature changes (second derivative), which conforms to the characteristics of thermal inertia. These represent preset weighted hyperparameters, used to balance the proportions of data-driven mean square error, smoothness, energy conservation, and comfort. Characterizes the energy conservation constraint; The THI constraint is used to explicitly incorporate the prediction error of the temperature and humidity index into the optimization objective, ensuring that not only the temperature is accurate, but also the human comfort index is accurate.

[0055] It can be seen that purely data-driven models may produce predictions that violate physical laws, such as a sudden increase in temperature in the absence of external heat sources and light. Therefore, this application provides a Physics-Aware Dual-Stream Spatio-Temporal Graph Gated Recurrent Network (PA-DST-G2N) and designs a loss function that incorporates physical constraints. This model not only considers the data-driven pattern but also integrates physical constraints, thereby improving the model's robustness.

[0056] In one embodiment, the predicted environmental parameters include a predicted temperature. The step of determining a first control command based on the predicted environmental parameters includes: acquiring the heat production of piglets in each piglet warmer; determining the heat dissipation of the piglet warmer based on the predicted temperature and the indoor temperature of the piglet warmer; determining the predicted temperature change based on the heating efficiency of the piglet warmer, the heat production of the piglets, and the heat dissipation of the piglet warmer; and determining the first control command based on the predicted temperature change.

[0057] The temperature control device determines the predicted temperature change based on the heating efficiency of the piglet warming box, the heat production of the piglets, and the heat dissipation of the warming box. Specifically, the temperature control device obtains the sum of the heat provided by the heating efficiency of the piglet warming box and the heat production of the piglets, and determines the difference between the sum and the heat dissipation of the warming box as the predicted temperature change.

[0058] The thermal environment of a piglet heat lamp is a typical heat capacity system. According to the law of conservation of energy, the temperature inside the lamp... The change depends on the difference between the input heat and the heat loss, as shown in the following formula: ρ·V· = + - - In the above formula, ρ represents the specific heat capacity of air (J / kg·K), ρ represents the density of air (kg / m³), and V represents the volume of the insulated box (m³). Characterizing the heat generated by the effective power provided by the heater. Characterizes heat dissipation through conduction via the box walls and floor. It represents the heat removed by ventilation.

[0059] The heat provided by the heating efficiency of the piglet warming box, i.e., the heat provided by the heater, satisfies the following formula: = η·

[0060] In the above formula, η represents the electrothermal conversion efficiency. Characterizes heating power.

[0061] The heat dissipation of a piglet heat exchanger includes heat conducted through the walls and floor, as well as heat removed by ventilation.

[0062] Heat dissipation through conduction via the enclosure walls and floor satisfies the following formula: =∑(UA)i ( - ) In the above formula, UA represents the preset overall heat transfer coefficient. Characterizes and predicts temperature, that is, the ambient temperature of the piglet rearing scenario. Characterizes the temperature inside the chamber.

[0063] The heat removed by ventilation satisfies the following formula: = ( - ) In the above formula, Characterizes air mass flow rate.

[0064] The temperature control device determines a first control command based on the predicted temperature change. Specifically, the temperature control device determines the target heating efficiency as the ratio between the predicted temperature change and the preset prediction; and adjusts the heater's heating efficiency to the target heating efficiency as the first control command.

[0065] The first control command is feedforward control, the core of which is to calculate in reverse based on predicted disturbances to maintain the target temperature. Ideal power required .

[0066] As an example, predicted environmental parameters include predicted ventilation volume, based on an ideal power approximation of steady-state thermal balance. Satisfy the following formula:

[0067] In the above formula, Characterizing feedforward control in The ideal heating power calculated at all times. Characterizing the electrothermal conversion efficiency of the heater, The effective heat dissipation surface area (SurfaceArea) of the insulated box. This represents the temperature at which the target value is maintained, and this target value changes dynamically with the piglets' age. The time step or lead time for the model's predictions, such as the next 30 minutes. Characterizing air density, The specific heat capacity at constant pressure of air. These are combined to calculate the heat capacity of air. Characterizes the predicted future ventilation volume. Characterizing the future t+ of the GCN-GRU model output The predicted temperature at any given time, that is, the ambient temperature. The characteristic is the prediction of ventilation volume, which can be indirectly inferred from the predicted ammonia concentration. The characterization is based on age-based table lookup for piglet heat production compensation. The overall heat transfer coefficient is characterized by real-time updates via online identification algorithms (such as recursive least squares, RLS) to adapt to aging of the insulation box or changes in the covering.

[0068] In yet another embodiment, ideal power Satisfy the following formula:

[0069] In the above formula, Characterizes the pre-adjusted power output increment or bias. The feedforward gain coefficient is used to linearly convert the temperature difference into a power adjustment factor. The characterization model predicts the chamber temperature that will result from maintaining the current power at future times.

[0070] The above formula, simplified to an incremental form, is more conducive to engineering implementation. The meaning of this formula is: if the model predicts that maintaining the current power will lead to a decrease in the internal temperature of the chamber... Higher than the set value, or predicted ambient temperature If the power output increases, reduce it in advance. .

[0071] In one embodiment, the current environmental parameters include the current temperature, and the step of obtaining the heat production of each piglet in the incubator includes: determining the body temperature difference of the piglets based on the preset critical low temperature and the current temperature; and determining the heat production of each piglet in the incubator based on the weight of each piglet and the body temperature difference.

[0072] The temperature control device determines the piglet's body temperature difference based on a preset critical low temperature and the current temperature. The temperature control device defines the difference between the preset critical low temperature and the current temperature as the piglet's body temperature difference.

[0073] As one example, the temperature control device obtains the weight of each piglet using a weighing scale. As yet another example, the temperature control device obtains the initial weight, average daily weight gain, and age of the piglets, and determines the weight of the piglets based on these information.

[0074] For example, the weight of a piglet satisfies the following formula: W(d) = +ADG·d In the above formula, W(d) represents the weight of the piglet. Initial weight is represented by ADG, average daily weight gain is represented by AD, and age is represented by d, all in days.

[0075] The temperature control device determines the heat production of each piglet in its incubator based on the weight and body temperature difference of each piglet. Specifically, the temperature control device obtains a first product term between the body temperature difference and a preset cold stress coefficient; obtains a second product term between the sum of this product term and a preset heat threshold and a preset index of the piglet; and determines a third product term between the second product term and the preset heat coefficient as the heat production of the piglet.

[0076] For example, the heat production of piglets satisfies the following formula: =5.09· ·[1+α·( - )] In the above formula, The heat production of piglets is represented by α, and the cold stress coefficient is represented by α. Characterizing the preset critical low temperature, Characterizes the current temperature. The preset critical low temperature decreases with increasing age.

[0077] In one embodiment, the heat demand of piglets changes dynamically with age. To maintain a constant body temperature, the metabolic heat production of piglets must balance the heat loss caused by the environment. According to Kleiber's law and related animal husbandry studies, the heat production of piglets... (W) has an exponential relationship with its body weight W (kg), if the current temperature < At this time, piglets must increase metabolic heat production to maintain their body temperature.

[0078] In one embodiment, the step of determining the second control command based on the current temperature and a preset temperature threshold in the current environmental parameters includes: obtaining the ambient temperature difference between the current temperature and the preset temperature threshold; determining the second heating power based on the ambient temperature difference; and determining the second control command based on the second heating power.

[0079] The second control command is feedback control, used to eliminate transient interference and high-frequency noise. It should be noted that the second heating power can be obtained by running a high-frequency (e.g., 10Hz) PID (Proportional-Integral-Derivative) control algorithm, which can be deployed on a local PLC (Programmable Logic Controller).

[0080] The second heating power satisfies the following formula:

[0081] In the above formula, Characterizing the second heating power of the heater, Characterizing environmental temperature difference, Characterization ratio, Characterization integral, Characterize the differential.

[0082] In one embodiment, the step of controlling the temperature of the piglet warming box according to the first control command and the second control command includes: determining a target heating power according to the first heating power in the first control command and the second heating power in the second control command; heating the piglet warming box based on the target heating power to achieve temperature control of the piglet warming box.

[0083] The temperature control device determines the target heating power as the sum of the first heating power and the second heating power, as shown below:

[0084] In the above formula, Characterizes the target heating power, which is the control quantity or command ultimately applied to the heater; The second heating power is represented by the feedback control quantity calculated by the high-frequency PID algorithm running in the local inner loop (PLC), which is used to eliminate instantaneous interference and high-frequency noise; This represents the first heating power, which is also the feedforward reference power calculated by the low-frequency prediction of the cloud outer loop operation. This represents the future spatiotemporal state vector predicted by the model.

[0085] It should be noted that the first control command can be generated by low-frequency (e.g., 0.05Hz, i.e., every 20 seconds) GCN-GRU prediction and feedforward calculations performed by the outer loop (cloud / edge).

[0086] In one embodiment, the first heating power calculated by the temperature control device is sent to the PLC as a reference power, and the PLC superimposes the PID output onto the PLC. The target heating power is obtained above, and heating is performed by the actuator.

[0087] As can be seen, the temperature is controlled via feedforward control through the first control command, which is responsible for "coarse adjustment" and "pre-adjustment" to handle large environmental changes. The temperature is then controlled via feedback control through the second control command, which is responsible for "fine adjustment" to ensure the final steady-state accuracy. This feedforward-feedback combined strategy perfectly solves the problem of lag in regulation and improves the timeliness of temperature adjustment. This application organically combines deep learning with traditional process control theory and fully considers the actual physical environment of the farm (wind, humidity, heat). Theoretical derivation and comparative analysis show that the improved system achieves a qualitative leap in control accuracy, response speed, energy saving effect, and animal welfare protection, providing solid technical support for realizing unmanned smart pig farms.

[0088] Combination Figure 2 As shown, the temperature control device includes a feedforward control mechanism and a feedback control mechanism. In the feedforward control mechanism, current environmental parameters are collected, and a static distance matrix and a dynamic wind field distance are constructed. A dynamic adjacency matrix is ​​then constructed based on these matrices. This dynamic adjacency matrix integrates distance and wind direction to build a network of relationships between the insulated boxes. Spatial relationship features (GCN) are extracted to allow each insulated box to sense the surrounding and upwind temperatures. Temporal evolution features (GRU) are extracted to infer temperature change trends based on historical patterns. Predicted environmental parameters are output to determine the ambient temperature and the temperature of the insulated boxes in advance. The feedforward control quantity is then calculated based on dynamic target requirements to obtain the first control command, thereby enabling proactive defense by back-calculating the ideal heating power based on the predicted environmental parameters. The first control command is then issued. The feedback control mechanism includes: acquiring the ambient temperature difference between the current temperature and a preset temperature threshold; calculating the PID feedback quantity to obtain the second control command; and outputting the final control command based on the first and second control commands.

[0089] Therefore, feedforward control, based on physical models and predicted data, compensates for temperature disturbances caused by environmental changes in advance, solving the lag problem of traditional PID control; feedback control eliminates model prediction errors and unforeseen interferences in real time, ensuring temperature control accuracy; through collaborative control between the cloud and local systems, it takes into account both environmental adaptability on a large time scale and control accuracy on a small time scale, achieving stable and efficient temperature control of the insulation box.

[0090] This application uses a piglet warming box, which includes: a controller for executing the temperature control method of the piglet warming box described above; a sensor group connected to the controller, configured to collect the current environmental parameters of the piglet warming box; and an actuator connected to the controller, configured to perform temperature control.

[0091] The sensor group consists of multi-dimensional environmental sensing sensors, including: a temperature and humidity sensor for identifying temperature T and relative humidity RH; a gas sensor for monitoring ammonia (NH3) and carbon dioxide (CO2); and an electrical parameter acquisition transformer for monitoring the real-time power of the heater, used for feedback correction and fault diagnosis.

[0092] The actuator can be a thyristor voltage regulator, supporting 0-100% linear power regulation and millisecond-level response time.

[0093] The communication network of the piglet warming box includes: a sensing layer (end to edge) which uses a wired RS485 bus to avoid interference with wireless signals from the metal fence inside the pigsty; and a transmission layer (edge ​​to cloud) where the edge gateway accesses the Internet via a 4G / 5G module or fiber optic cable and uploads data using the MQTT (Message Queuing Telemetry Transport) protocol.

[0094] The controller includes a touch screen, control box, controller, two-phase power supply, leakage current protection device, variable silicon voltage regulator module and temperature monitoring module.

[0095] The touchscreen displays relevant information parameters and equipment status, allowing for direct human-machine interaction and modification of control parameters. The control box houses and stores components such as the controller, two-phase power supply, and leakage current protector. The controller receives and calculates feedback data from the temperature monitoring module, outputs analog signals to the voltage regulation module, and connects to the display screen for information exchange. The two-phase power supply provides power to all modules within the control box. The leakage current protector ensures electrical safety and prevents leakage. The voltage regulation module receives analog signals from the controller and supplies power to the heater in the insulation box. The temperature monitoring module collects data from the temperature sensor and transmits the data back to the controller.

[0096] In one embodiment, the piglet warming box is used in a temperature control system, which includes cloud, edge, and end-side components. The remote component is deployed on a remote server or data center; core components include a high-performance GPU (Graphics Processing Unit) cluster, a time-series database, and an AI (Artificial Intelligence) training platform; its main functions are storing massive amounts of historical data, training and updating the GCN-GRU model, and global energy consumption analysis and strategy optimization.

[0097] The edge is deployed in each birthing ward or unit, and the core component includes an industrial-grade edge gateway; its main functions are to run inference models to generate real-time predictions, dynamically build local graph structures, perform protocol conversion, and clean data.

[0098] The device is deployed at the end of the incubator for suckling piglets. Its core components include a controller, a sensor group, and an actuator. Its main functions are to collect temperature, humidity, ammonia, and current data, receive power commands and drive the heater, and execute low-level safety logic, such as emergency stop for overheating.

[0099] In one embodiment, combined with Figure 3 As shown, the piglet warming box includes a box body 1 and a top cover 3. The top cover 3 is connected to the box body 1 via a hinge and is located above the box body 1. The top cover 3 is a transparent cover. Two heating plates 4 are installed at the bottom of the box body 1. Two sliding plate grooves 5 are installed on one side of the box body 1. An adjusting plate 6 is installed in the sliding plate groove 5. A positioning plate 7 is installed on one side of the adjusting plate 6. The positioning plate 7 is connected to the box body 1, and its surface is provided with several positioning through holes. A positioning rod 9 is installed at the upper end of the adjusting plate 6. The positioning rod 9 passes through the adjusting plate 6 and is adapted to the positioning through holes. An opening is provided on one side of the box body 1. The size of the opening varies according to the height of the adjusting plate 6. The inside of the heat preservation cover 14 is equipped with an electric heating tube (heater) and a temperature probe (temperature sensor, fixed in a slot on the side plate of the box body). One end of the electric heating tube is connected to an electric heating tube power cord 16. Two conduits 18 are installed on each side of the housing 1. Each conduit 18 contains a heating plate power cord 19, which passes through the conduit 18 and connects to the bottom surface of the corresponding heating plate 4. The inner wall of the housing 1 has grooves containing positioning magnetic blocks. Each heating plate 4 has a fixing magnetic block 23 installed on one side, and the positioning magnetic block and the fixing magnetic block 23 attract each other. Each heating plate 4 has a pressure plate 24 installed on the side near the positioning plate 7, and screws penetrate the surface of the pressure plate 24.

[0100] In some embodiments, the sensor samples at a frequency of 1Hz, which is then aggregated by the edge gateway and uploaded to the cloud in 1-minute increments. The cloud performs outlier removal (such as sensor fault data) and Z-score normalization to prevent different dimensions (power 0-220V, temperature 20-40°C) from affecting gradient descent. A preset environmental parameter prediction model reconstructs the dynamic adjacency matrix At every 15 minutes based on the latest fan status, using a data window of the past 60 minutes to predict the temperature and humidity trajectory for the next 30 minutes. Control commands calculated in the cloud are pushed to the edge gateway via MQTT (Message Queuing Telemetry Transport), and after parsing, the gateway writes them to the PLC register via RS485.

[0101] like Figure 3 As shown, the piglet warming box includes a heater, box body, transparent cover, stainless steel hinges, temperature sensor, slide groove and sliding plate, and pull rope module.

[0102] The enclosure consists of a front panel, a rear panel, side panels, and a transparent cover. The transparent cover is connected and fixed to the rear panel with stainless steel hinges. This combination allows staff to easily open the transparent cover manually and automatically maintain it in its fully open position for easy handling of the piglets. At the same time, when the transparent cover is closed, staff can observe the growth of the piglets inside the enclosure without any additional operation.

[0103] The heater is fixed inside the insulated box near the front baffle; the temperature sensor is fixed in the slot on the side panel, and the signal line is connected to the control box; the sliding plate is installed on both sides of the insulated box and connected to the side panel; the sliding plate is equipped with a sliding plate hook and a quick-locking buckle.

[0104] The methods involved in the temperature control method embodiments of the piglet warming box of this application, when implemented as software functional units and sold or used as independent products, can be stored in a device, such as a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0105] The above description is merely an embodiment of this application and does not limit the patent scope of this application. Any equivalent structural or procedural transformations made using the content of this application's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of this application.

[0106] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The term "and / or" is merely a description of the association of related objects, indicating that three relationships can exist; for example, A and / or B can represent: A alone, A and B simultaneously, and B alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects are in an "or" relationship. Furthermore, "many" in this document means two or more. In addition, the term "at least one" in this document means any combination of at least two of any one or more of a plurality of elements, such as including at least one of A, B, and C, and may mean including any one or more elements selected from the set consisting of A, B, and C.

Claims

1. A temperature control method for a mammal piglet incubator, characterized by, The method includes: Obtain the current environmental parameters of the piglet warming box; The current environmental parameters are input into a preset environmental parameter prediction model to obtain the predicted environmental parameters of the piglet incubator, and the first control command is determined based on the predicted environmental parameters. The second control command is determined based on the current temperature and the preset temperature threshold in the current environmental parameters. The temperature of the piglet incubator is controlled according to the first control command and the second control command.

2. The method of claim 1, wherein, The preset environmental parameter prediction model is a graph convolution model, in which each insulated box is used as a node. The graph convolution model includes: A static distance matrix is ​​constructed based on the physical straight-line distance between each insulation box and the preset attenuation rate control parameters. A dynamic wind field matrix is ​​constructed based on the fan vector of the insulation box and the position vector between each insulation box; The static distance matrix and the dynamic wind field matrix are normalized to obtain the dynamic adjacency matrix.

3. The method of claim 1, wherein, The preset environmental parameter prediction model is a graph convolution model, and the method includes: The predicted and true values ​​of the graph convolution model are used as model prediction constraints. The loss function is determined based on the model prediction constraints, time-domain smoothing constraints, energy conservation constraints, and thermal health index constraints. The graph convolution model is trained based on the loss function to obtain the preset environment parameter prediction model.

4. The method of claim 3, wherein, The method includes: The temperature change of the piglet warming box within a preset time period, the heating efficiency and heat dissipation efficiency of the piglet warming box are obtained. The energy conservation constraint is determined based on the temperature change, the heating efficiency, the heat dissipation efficiency, and the preset input electrical power.

5. The method according to claim 1, characterized in that, The predicted environmental parameters include a predicted temperature, and the step of determining the first control command based on the predicted environmental parameters includes: Obtain the heat production of piglets in each piglet heat incubator; The heat dissipation of the piglet incubator is determined based on the predicted temperature and the indoor temperature of the piglet incubator. The predicted temperature change is determined based on the heating efficiency of the piglet warming box, the heat generated by the piglets, and the heat dissipation of the piglet warming box. The first control command is determined based on the predicted temperature change.

6. The method according to claim 5, characterized in that, The current environmental parameters include the current temperature, and the step of obtaining the heat production of piglets in each piglet heat incubator includes: The body temperature difference of the piglets is determined based on the preset critical low temperature and the current temperature; The heat production of each piglet in the incubator is determined based on the weight of each piglet and the difference in body temperature.

7. The method according to claim 1, characterized in that, The step of determining the second control command based on the current temperature and the preset temperature threshold in the current environmental parameters includes: Obtain the ambient temperature difference between the current temperature and the preset temperature threshold; The second heating power is determined based on the ambient temperature difference; The second control command is determined based on the second heating power.

8. The method according to claim 1, characterized in that, The step of controlling the temperature of the piglet warming box according to the first control command and the second control command includes: The target heating power is determined based on the first heating power in the first control command and the second heating power in the second control command; The target heating power is used to heat the piglet warming box, thereby achieving temperature control of the piglet warming box.

Images