Method for predicting, regulating and controlling winter breeding environment of poultry house

Abstract

The invention discloses a poultry house winter breeding environment prediction and regulation method, which comprises the steps of obtaining data in a poultry house, decomposing the data, taking causal characteristics of temperature and carbon dioxide as prediction input information, predicting trend components, seasonal components and residual components of the temperature and the carbon dioxide, setting a linkage control strategy, and optimizing a target function. And solving by using a genetic algorithm objective function to obtain an optimal regulation and control parameter. The henhouse winter environment regulation and control method is researched from the aspects of environment prediction capability, prediction time, prediction cost, temperature and carbon dioxide synergistic effect, a winter and summer combined control strategy is formulated, the winter and summer combined control strategy and the optimal state are achieved to the maximum extent, solving is conducted in cooperation with a genetic algorithm, optimal regulation and control parameters are obtained, and the henhouse winter environment regulation and control method is suitable for popularization and application. A prediction function can be completed at low cost and within a minute level, the purposes of timely response and generalization can be achieved, and an intelligent in-house environment with timely response and high-precision regulation and control is created for poultry.

Description

Technical Field

[0001] The invention relates to the technical field of poultry breeding, and in particular to a method for predicting and regulating the breeding environment of a poultry house in winter. Background Art

[0002] Modern poultry farming tends to be large-scale and intensive, and most of them are raised in houses. When the breeding environment is suitable, the incidence of poultry is low, the feed intake and feed utilization rate are significantly improved, and the production performance of poultry is better. However, when the environment changes significantly, it may cause stress response in poultry, destroy the body heat balance, and lead to a decline in their production performance or even death. Therefore, in the process of intensive poultry farming, it is necessary to monitor the important environmental indicators of the poultry house. The precise environmental control system is an important part of maintaining a good breeding environment in the house. It mainly collects environmental information through sensors distributed in the poultry house, and then controls the corresponding environmental control equipment through environmental control strategies to keep the sensor site area environment in the house at the set threshold. To alleviate the cold and heat stress of poultry, improve the air quality in the house, and control the light intensity, large-scale farms often use wet curtains, fans, ventilation windows and heating equipment for joint control, and match them with intelligent environmental control systems to create a comfortable house environment for poultry.

[0003] At present, many studies are mainly focused on the temperature control technology of the indoor environment, mainly because the temperature in the house is the main factor affecting the comfort of chickens. However, if only temperature is used as the control standard in winter, due to the reduction of ventilation, the air quality in the house cannot reach a suitable state, which will also affect the production and health of poultry. In winter, temperature and carbon dioxide are coupled, so it is very important to achieve the coordination and trade-off between the two in winter. In addition, due to the time lag of sensors and equipment control in the environmental control system, the timeliness and stability of the environmental control in the poultry house cannot be guaranteed. The current prediction is mainly based on hours, but in actual control, shorter prediction time and prediction accuracy are required to ensure the indoor environment.

[0004] After searching, patent No. CN117032361A discloses a chicken house environment control method, system, equipment and medium, patent No. CN116740805A discloses a chicken house environment control method and system, and patent No. CN115220501A discloses an environment control method integrating SSA-Elman network model and pigeon flock inspiration algorithm, which mainly focuses on regulating the temperature in the house and does not consider the carbon dioxide index. However, the high concentration of carbon dioxide in winter will affect the health of chickens. In addition, the above patents aim to obtain feedback from the temperature sensor before controlling, but there will be a problem of control lag, which will cause the environment in the house to not always be in a comfortable state. Summary of the invention

[0005] The purpose of the present invention is to provide a method for predicting and regulating the winter breeding environment of a poultry house. The method for regulating the winter environment of a poultry house is studied from the aspects of environmental prediction ability, prediction time, prediction cost, and the synergistic effect of temperature and carbon dioxide. The purpose is to predict the temperature and carbon dioxide values ​​in advance, so as to carry out early regulation, so as to ensure that the environment of the chicken house is always in a set state range, so as to solve the shortcomings of the prior art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for predicting and regulating the winter breeding environment of a poultry house comprises the following steps:

[0008] S1: Obtain the temperature, humidity, carbon dioxide, wind speed, ammonia data in the poultry house, the temperature data outside the house and the operating status data of the environmental control equipment, and perform data cleaning and normalization on the various data obtained;

[0009] S2: Perform STL decomposition on temperature and carbon dioxide data to obtain trend component, seasonal component and residual component;

[0010] S3: Use Granger causality test to conduct causal test and analysis on other environmental data, temperature and carbon dioxide, find the causal characteristics of temperature and carbon dioxide as the input information for prediction, and get the optimal lag order at the same time;

[0011] S4: Take the obtained causal features of temperature and carbon dioxide as input and use LSTM to predict the trend component, seasonal component and residual component of temperature and carbon dioxide at the next time step respectively, using MSE as the loss function;

[0012] S5: Take the trend component, seasonal component and residual component of the next time step as input, and use XGBoost to make a final prediction of the temperature and carbon dioxide data to obtain the predicted temperature T pred (t) and predicted carbon dioxide concentration C pred (t);

[0013] S6: Environmental control sets the target temperature T according to the farm requirements set and the maximum permissible carbon dioxide concentration C max , where T set Is the ideal temperature for livestock and poultry house environment, C max It is the upper limit that the carbon dioxide concentration in the livestock and poultry house cannot exceed;

[0014] S7: The predicted temperature T pred (t) and temperature setting value T set Comparison to obtain the temperature error: e T (t) = T set -T pred(t), the PID control output is Where K PT , K iT and K dT are the proportional, integral and differential coefficients of temperature control respectively;

[0015] S8: The predicted carbon dioxide concentration C pred (t) and the maximum value of carbon dioxide C max Comparison of the carbon dioxide error: e CO2 (t) = C max -C pred (t), the PID control output is Where K Pco2 , K iCO2 and K dCO2 are the proportional, integral, and differential coefficients of CO2 control, respectively;

[0016] S9: Set linkage control strategies for different temperatures and carbon dioxide concentrations in winter and summer to adapt to the indoor environmental requirements;

[0017] S10: Setting an optimization objective function to minimize the control error of temperature and carbon dioxide concentration, which is used to balance the control of temperature and carbon dioxide, so that the actual temperature and carbon dioxide concentration are closer to the set value;

[0018] S11: Use genetic algorithm to solve the S10 objective function and obtain the optimal control parameters.

[0019] Furthermore, the method for setting the linkage control strategy in S9 is as follows:

[0020] If T pred (t) <T set And C pred (t)>C max , turn on the heater in advance and reduce ventilation. When the temperature approaches the target value, gradually increase the ventilation;

[0021] If T pred (t)∈[T set -0.5, T set -0.5] and C pred (t)<C max , then maintain the current control measures;

[0022] If T pred (t)>T set And C pred (t)>C max ax, then increase ventilation first to reduce carbon dioxide concentration;

[0023] If T pred (t)∈[Tset -0.5, T set -0.5] and C pred (t)<<C max , then maintain the current control measures;

[0024] If T pred (t)∈[T set -0.5, T set -0.5] and C pred (t)>C max , then gradually increase ventilation until T pred (t+1)<C max , ensuring that the temperature and CO2 are at the appropriate state at the next time step.

[0025] Furthermore, the optimization objective function set in S10 is as follows:

[0026] Where α and β are weight parameters used to balance the priority of temperature and CO2 control. α controls the impact of temperature error in the overall target, and β controls the impact of CO2 concentration error in the overall target. It represents the average relative error between the actual temperature and the set temperature within the time T. It represents the average relative error between the actual carbon dioxide concentration and the set concentration within time T.

[0027] Furthermore, the solution process of the objective function in S11 is as follows:

[0028] S110: First, randomly generate a set of initial PID parameter sets, an initial population including multiple individuals, each individual represents a set of PID parameters x = [K PT , K iT , K dT , K PCO2 , K iCO2 , K dco2 ];

[0029] S111: Use the objective function to calculate the fitness value of each individual:

[0030]

[0031] S112: Randomly select a certain number of individuals, select the individual with the best fitness, then exchange parameters between the selected individual and a randomly selected point, and randomly adjust the PID parameters of the newly generated individual;

[0032] S113: Replace the worst individual in the original population with this newly generated individual. If the maximum number of iterations is reached and the change in fitness value is less than the threshold, the inheritance is terminated, and finally the individual with the best fitness in the new population is selected as the optimization result, and the new x=[K PT , K iT , K dT , K PCO2 , K iCO2 , K dCO2 ].

[0033] Furthermore, the environmental control equipment in S1 includes wet curtains, fans, ventilation windows and heating equipment.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] 1. The poultry house winter breeding environment prediction and regulation method of the present invention studies the chicken house winter environment regulation method from the aspects of environmental prediction ability, prediction time, prediction cost, temperature and the synergistic effect of carbon dioxide, and formulates a joint control strategy for winter and summer, so that the two can reach the optimal state to the greatest extent, and effectively avoid the problem of temperature drop caused by excessive operation of ventilation equipment.

[0036] 2. The poultry house winter breeding environment prediction and control method of the present invention minimizes the control errors of temperature and carbon dioxide concentration by setting the optimization objective function, and solves it with the genetic algorithm to obtain the optimal control parameters. It can dynamically optimize the PID parameter adjustment and realize dual-objective optimization. It can generate a set of more adaptable PID parameters, and finally automatically select the control scheme that best suits the current environment. The prediction function can be completed at low cost and within minutes, and the purpose of timely response and generalization can be achieved, creating an intelligent house environment with timely response and high-precision control for poultry. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a component diagram of carbon dioxide after decomposition through STL in the case of the present invention;

[0038] Figure 2 It is a causal characteristic diagram obtained through causal testing in the case of the present invention;

[0039] Figure 3 This is a comparison chart of the predicted value and actual value of STL-GC-LSTM-XGBoost in the case of the present invention;

[0040] Figure 4 is a flow chart of the prediction method of the present invention;

[0041] Figure 5 It is the environmental control flow chart of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] See also Figure 4-5 The embodiment of the present invention provides a method for predicting and regulating the winter breeding environment of a poultry house, comprising the following steps:

[0044] S1: Obtain the temperature, humidity, carbon dioxide, wind speed, ammonia data in the poultry house, the temperature data outside the house and the operating status data of the environmental control equipment (fan, wet curtain, ventilation window, heater), and perform data cleaning and normalization on the various data obtained;

[0045] S2: Perform STL decomposition on temperature and carbon dioxide data to obtain trend components, seasonal components, and residual components, which can reduce data noise, extract useful trends, and increase prediction capabilities;

[0046] S3: Through Granger causality test, causal test and analysis are performed on other environmental data, temperature and carbon dioxide to find the causal characteristics of temperature and carbon dioxide as the input information for prediction, and the optimal lag order can be obtained at the same time; different from the existing correlation analysis method, Granger causality test can obtain the causal characteristics of temperature and carbon dioxide, rather than the mutual relationship;

[0047] S4: Take the obtained causal features of temperature and carbon dioxide as input and use LSTM to predict the trend component, seasonal component and residual component of temperature and carbon dioxide at the next time step respectively, using MSE as the loss function;

[0048] S5: Take the trend component, seasonal component and residual component of the next time step as input, and use XGBoost to make a final prediction of the temperature and carbon dioxide data to obtain the predicted temperature T pred (t) and predicted carbon dioxide concentration C pred (t); This step is to reduce the impact of residuals when the amount of data is large, thereby increasing the prediction ability. The prediction process can complete the prediction function at a low cost and within 2 minutes, and can achieve the purpose of timely response and generalization;

[0049] S6: Environmental control sets the target temperature T according to the farm requirements set and the maximum permissible carbon dioxide concentration C max , where T set Is the ideal temperature for livestock and poultry house environment, Cmax It is the upper limit that the carbon dioxide concentration in the livestock and poultry house cannot exceed;

[0050] S7: The predicted temperature T pred (t) and temperature setting value T set Comparison to obtain the temperature error: e T (t) = T set -T pred (t), the PID control output is Where K PT , K iT and K dT are the proportional, integral and differential coefficients of temperature control respectively;

[0051] S8: The predicted carbon dioxide concentration C pred (t) and the maximum value of carbon dioxide C max Comparison of the carbon dioxide error: e CO2 (t) = C max -C pred (t), the PID control output is Where K Pco2 , K iCO2 and K dCO2 are the proportional, integral, and differential coefficients of CO2 control, respectively;

[0052] S9: Set linkage control strategies for different temperatures and carbon dioxide concentrations in winter and summer to adapt to the needs of the indoor environment; in this step, due to the differences between winter and summer, summer is mainly based on temperature, and ventilation will be increased to reduce temperature, so carbon dioxide is often within a safe range; but in winter, due to the lower temperature outside the house, in order to achieve a suitable temperature in the house, the ventilation volume is often small, resulting in a higher carbon dioxide concentration, so a joint control strategy is needed to maximize the optimal state of the two. Specifically, the linkage control strategy is:

[0053] If T pred (t)<T set And C pred (t)>C max , turn on the heater in advance and reduce ventilation. When the temperature approaches the target value, gradually increase the ventilation;

[0054] If T pred (t)∈[T set -0.5, T set -0.5] and C pred (t)<C max , then maintain the current control measures;

[0055] If T pred (t)>T setAnd C pred (t)>C max , then increase ventilation first to reduce carbon dioxide concentration;

[0056] If T pred (t)∈[T set -0.5, T set -0.5] and C pred (t)<C max , then maintain the current control measures;

[0057] If T pred (t)∈[T set -0.5, T set -0.5] and C pred (t)>C max , then gradually increase ventilation until T pred (t+1)<C max , ensuring that the temperature and carbon dioxide are in a suitable state at the next time step, can avoid excessive operation of the ventilation equipment causing the temperature to drop.

[0058] S10: Setting an optimization objective function to minimize the control error of temperature and carbon dioxide concentration is used to balance the control of temperature and carbon dioxide so that the actual temperature and carbon dioxide concentration are closer to the set value; specifically, in order to make the actual temperature and carbon dioxide concentration closer to the set value, setting an optimization objective function to minimize the control error of temperature and carbon dioxide concentration: Where α and β are weight parameters used to balance the priority of temperature and CO2 control. α controls the impact of temperature error in the overall target, and β controls the impact of CO2 concentration error in the overall target. It represents the average relative error between the actual temperature and the set temperature within the time T. It represents the average relative error between the actual carbon dioxide concentration and the set concentration within time T. In this step, by minimizing the objective function, the control of temperature and carbon dioxide can be balanced and as close to the set value as possible;

[0059] S11: Use genetic algorithm to solve the S10 objective function to obtain the optimal control parameters; specifically:

[0060] First, a set of initial PID parameter sets is randomly generated, an initial population containing multiple individuals, each individual represents a set of PID parameters x = [K PT , K iT , K dT , K PCO2 , K iCO2 , K dCO2 ];

[0061] Then use the objective function to calculate the fitness value of each individual:

[0062]

[0063] Then randomly select a certain number of individuals, select the individual with the best fitness, then exchange parameters between the selected individual and a randomly selected point, and randomly adjust the PID parameters of the newly generated individual to avoid falling into the local optimal solution; then replace the worst individual in the original population with this newly generated individual. If the maximum number of iterations set and the change in fitness value is less than the threshold, the inheritance is terminated, and finally the individual with the best fitness in the new population is selected as the optimization result, and the new x=[K PT , K iT , K dT , K PCO2 , K iCO2 , K dCO2 ].

[0064] In order to further better explain the above embodiment, the present invention also provides the following specific cases:

[0065] like Figure 1-3 As shown in the figure, the indoor ambient temperature and carbon dioxide prediction and control of a large-scale chicken house in Shandong are analyzed: the selected time interval is 10 minutes, and the components of carbon dioxide after STL decomposition are as follows Figure 1 The causal characteristics obtained through causal testing are shown as Figure 2 As shown, the predicted values ​​and actual values ​​of STL-GC-LSTM-XGBoost in the case are as follows Figure 3 As shown, the MAE is 11.82, the RMSE is 15.9, and the MAPE is 0.55%, which indicates good prediction ability. The prediction time is 86s, which meets the condition of being within 10 minutes.

[0066] In summary: The present invention provides a method for predicting and regulating the winter breeding environment of a poultry house, which studies the winter environmental regulation method of a poultry house from the aspects of environmental prediction capability, prediction time, prediction cost, synergistic effect of temperature and carbon dioxide, and formulates a joint control strategy for winter and summer, so that the two can reach the optimal state to the greatest extent, and effectively avoid the problem of temperature drop caused by excessive operation of ventilation equipment. In addition, the present invention minimizes the control error of temperature and carbon dioxide concentration by setting an optimization objective function, and solves it in conjunction with a genetic algorithm to obtain the optimal regulation parameters, which can dynamically optimize PID parameter adjustment, realize dual-objective optimization, generate a set of more adaptable PID parameters, and finally automatically select the control scheme that best suits the current environment, which can complete the prediction function at low cost and within minutes, and can achieve the purpose of timely response and generalization, creating an intelligent house environment for poultry with timely response and high-precision regulation.

[0067] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A method for predicting and regulating the winter breeding environment of a poultry house, characterized in that: The following steps are involved: S1: Obtain the temperature, humidity, carbon dioxide, wind speed, ammonia data in the poultry house, the temperature data outside the house and the operating status data of the environmental control equipment, and perform data cleaning and normalization on the various data obtained; S2: Perform STL decomposition on temperature and carbon dioxide data to obtain trend component, seasonal component and residual component; S3: Use Granger causality test to conduct causal test and analysis on other environmental data, temperature and carbon dioxide, find the causal characteristics of temperature and carbon dioxide as the input information for prediction, and get the optimal lag order at the same time; S4: Take the obtained causal features of temperature and carbon dioxide as input and use LSTM to predict the trend component, seasonal component and residual component of temperature and carbon dioxide at the next time step respectively, using MSE as the loss function; S5: Take the trend component, seasonal component and residual component of the next time step as input, and use XGBoost to make a final prediction of the temperature and carbon dioxide data to obtain the predicted temperature T pred (t) and predicted carbon dioxide concentration C pred (t); S6: Environmental control sets the target temperature T according to the farm requirements set and the maximum permissible carbon dioxide concentration C max , where T set Is the ideal temperature for livestock and poultry house environment, C max It is the upper limit that the carbon dioxide concentration in the livestock and poultry house cannot exceed; S7: The predicted temperature T pred (t) and temperature setting value T set Comparison to obtain the temperature error: e T (t) = T set -T pred (t), the PID control output is Where K PT , K iT and K dT are the proportional, integral and differential coefficients of temperature control respectively; S8: The predicted carbon dioxide concentration C pred (t) and the maximum value of carbon dioxide C max Comparison of the carbon dioxide error: e CO2 (t) = C max -C pred (t), the PID control output is Where K PCO2 , K iCO2 and K dCO2 are the proportional, integral, and differential coefficients of CO2 control, respectively; S9: Set linkage control strategies for different temperatures and carbon dioxide concentrations in winter and summer to adapt to the indoor environmental requirements; S10: Setting an optimization objective function to minimize the control error of temperature and carbon dioxide concentration, which is used to balance the control of temperature and carbon dioxide, so that the actual temperature and carbon dioxide concentration are closer to the set value; S11: Use genetic algorithm to solve the S10 objective function and obtain the optimal control parameters.

2. A method for predicting and controlling the poultry house winter breeding environment as claimed in claim 1, characterized in that: The method for setting the linkage control strategy in S9 is as follows: If T pred (t)<T set And C pred (t)>C max , turn on the heater in advance and reduce ventilation. When the temperature approaches the target value, gradually increase the ventilation; If T pred (t)∈[T set -0.5, T set -0.5] and C pred (t)<C max , then maintain the current control measures; If T pred (t)>T set And C pred (t)>C max , then increase ventilation first to reduce carbon dioxide concentration; If T pred (t)∈[T set -0.5, T set -0.5] and C pred (t)<C max , then maintain the current control measures; If T pred (t)∈[T set -0.5, T set -0.5] and C pred (t)>C max , then gradually increase ventilation until T pred (t+1)<C max , ensuring that the temperature and CO2 are at the appropriate state at the next time step.

3. A method for predicting and controlling the poultry house winter breeding environment as claimed in claim 1, characterized in that: The optimization objective function set in S10 is as follows: Where α and β are weight parameters used to balance the priority of temperature and CO2 control. α controls the impact of temperature error in the overall target, and β controls the impact of CO2 concentration error in the overall target. It represents the average relative error between the actual temperature and the set temperature within the time T. It represents the average relative error between the actual carbon dioxide concentration and the set concentration within time T.

4. A method for predicting and controlling the poultry house winter breeding environment as claimed in claim 3, characterized in that: The solution process of the objective function in S11 is as follows: S110: First, randomly generate a set of initial PID parameter sets, an initial population including multiple individuals, each individual represents a set of PID parameters x = [K PT , K iT , K dT , K PCO2 , K iCO2 , K dCO2 ]; S111: Use the objective function to calculate the fitness value of each individual: S112: Randomly select a certain number of individuals, select the individual with the best fitness, then exchange parameters between the selected individual and a randomly selected point, and randomly adjust the PID parameters of the newly generated individual; S113: Replace the worst individual in the original population with this newly generated individual. If the maximum number of iterations is reached and the change in fitness value is less than the threshold, the inheritance is terminated, and finally the individual with the best fitness in the new population is selected as the optimization result, and the new = [K PT , K iT , K dT , K PCO2 , K iCO2 , K dCO2 ].

5. A method for predicting and controlling the poultry house winter breeding environment as claimed in claim 1, characterized in that: The environmental control equipment in S1 includes wet curtains, fans, ventilation windows and heating equipment.

Images