Intelligent ventilation system and method of cage for livestock breeding

Abstract

The invention discloses an intelligent ventilation system and method for a cage for livestock breeding, and belongs to the technical field of livestock breeding, and the system comprises an acquisition processing module, an environment modeling module, a change prediction module, an optimization decision module, a strategy simulation module, a storage safety module, an equipment control module, a monitoring feedback module, a user interaction module and an emergency response module. According to the invention, environmental conditions can be managed more actively, environmental parameters are prevented from deviating from an ideal state due to a hysteresis effect, the stress influence on animals is reduced, specific requirements of different livestock animals are met, the air quality and animal health level in an animal house are improved, the decision precision of the system is improved, and the economic benefit is increased. It is ensured that the cage environment is always kept in the most ideal state, improvement of air quality in a short term and minimization of long-term energy consumption are balanced, and the overall efficiency of the system is effectively optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of animal husbandry, and in particular to an intelligent ventilation system and method for animal husbandry cages. Background Art

[0002] With the growth of the global population and the increasing demand for animal products, the livestock farming industry is facing unprecedented pressure. In order to meet market demand, the scale of farming continues to expand, and intensive and high-density farming models have gradually become the mainstream. However, in this environment, the management of air quality and environmental conditions in farms has become particularly important. Poor ventilation conditions may lead to the accumulation of harmful gases, such as excessive ammonia and carbon dioxide, which not only affects the respiratory health of animals, but also leads to decreased feed conversion rate, disease transmission and increased animal stress response, which seriously affects production efficiency and economic benefits. Traditional ventilation systems mostly rely on manual control or automatic control based on simple rules. Although this method can regulate the environment to a certain extent, it is difficult to maintain the optimal state of the cage environment under various complex and changing conditions due to the lack of real-time response to dynamic changes in the environment.

[0003] After searching, Chinese Patent No. CN202211160201.9 discloses an intelligent ventilation system and method for cages for animal husbandry. Although the invention automatically adjusts the opening and closing of the ventilation system and the sprinkler system through the environmental monitoring data in the cage; dynamically balances the cage environment, reduces manpower input and effectively reduces energy consumption, but cannot actively manage environmental conditions, and is prone to environmental parameters deviating from the ideal state due to the hysteresis effect, which has a greater impact on the stress of animals and cannot meet the specific needs of different livestock and poultry creatures; in addition, the existing intelligent ventilation systems and methods for cages for animal husbandry have low decision-making accuracy, and cannot ensure that the cage environment is always maintained in the most ideal state, which is not conducive to balancing the improvement of air quality in the short term and the minimization of long-term energy consumption. For this reason, we propose an intelligent ventilation system and method for cages for animal husbandry. Summary of the invention

[0004] The purpose of the present invention is to solve the defects in the prior art and to propose an intelligent ventilation system and method for cages used for animal husbandry.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions:

[0006] An intelligent ventilation system for livestock breeding cages, comprising a collection and processing module, an environment modeling module, a change prediction module, an optimization decision module, a strategy simulation module, a storage safety module, an equipment control module, a monitoring feedback module, a user interaction module and an emergency response module;

[0007] The acquisition and processing module is used to acquire and pre-process various environmental data in the cage;

[0008] The environmental modeling module is used to establish a virtual model to simulate the environmental changes of the cage;

[0009] The change prediction module is used to predict the change trend of cage environmental parameters;

[0010] The optimization decision module is used to generate the optimal ventilation strategy within a specified time;

[0011] The strategy simulation module is used to simulate and evaluate the execution process of different ventilation strategies;

[0012] The storage security module is used to store environmental data and decision records;

[0013] The equipment control module is used to control the operation of the ventilation equipment;

[0014] The monitoring feedback module is used to monitor the ventilation effect in real time and adjust the system operating parameters;

[0015] The user interaction module is used to adjust the operating parameters of the ventilation equipment and view real-time environmental data, the execution of ventilation strategies and historical records;

[0016] The emergency response module is used to issue an alarm and execute emergency strategies.

[0017] As a further solution of the present invention, the specific steps of pre-processing the environmental data in the cage by the acquisition and processing module are as follows:

[0018] S1.1: Receive various environmental data collected by various high-precision sensors, construct corresponding environmental data sets, and calculate the first quartile Q1 and the third quartile Q3 of the data, and then calculate the interquartile range IQR based on Q1 and Q3;

[0019] S1.2: Obtain the upper and lower edge ranges based on Q1, Q3, and IQR. If the environmental data point is greater than the upper edge value, or less than the lower edge value, the environmental data is judged to be abnormal data, and the detected abnormal value is deleted. Then, the missing values ​​in each environmental data in the environmental data set are detected, and the missing values ​​are filled by linear interpolation.

[0020] S1.3: Calculate the average value of the neighboring data points of each data point and replace them to smooth the data. After all the data are smoothed, scale each environmental data to the range of [0,1] using the Min-Max normalization method.

[0021] As a further solution of the present invention, the high-precision sensor described in S1.1 specifically includes a thermocouple sensor, a capacitive humidity sensor, an infrared carbon dioxide sensor, an electrochemical ammonia sensor, a photoresistor, a thermal wind speed sensor, a laser dust sensor and an electrochemical oxygen sensor;

[0022] The environmental data described in S1.1 specifically include temperature, humidity, carbon dioxide concentration, ammonia concentration, light intensity, wind speed and direction, air quality and oxygen concentration.

[0023] As a further solution of the present invention, the change prediction module predicts the change trend of the cage environment parameters in the following specific steps:

[0024] S2.1: Collect and preprocess historical environmental data and weather forecast information, align data of different frequencies to the same time scale, use the mean filling method to handle missing data points, extract time features and lag features from each set of data, and construct a feature data set;

[0025] S2.2: Divide the feature data set into training set, validation set and test set according to the time sequence, input the training set into the LSTM model, obtain the prediction output of the corresponding time step through the forward propagation of the LSTM model layer by layer, and calculate the loss value between the predicted value and the actual value through the root mean square error function;

[0026] S2.3: The LSTM model back-propagates the loss value and calculates the gradient information of the loss value for each network layer of the model. Based on the gradient information of each layer, the parameter information of each layer is updated through the Adam optimizer. After each round of training, the model performance is evaluated using the validation set, and the model parameters are tuned through grid search based on the evaluation results.

[0027] S2.4: Use the test set to evaluate the final model and calculate various performance indicators. Then deploy the trained model to the production environment, receive new environmental data and weather forecast data in real time, and output future environmental change predictions.

[0028] As a further solution of the present invention, the optimization decision module generates the optimal ventilation strategy within a specified time in the following specific steps:

[0029] S3.1: Based on the outputs of the environmental modeling module and the change prediction module, obtain the environmental conditions at different time points to construct the corresponding state space, collect different control strategies that the ventilation system can adopt, and construct the corresponding action space;

[0030] S3.2: Calculate the probability of transitioning to the next state for any state in the state space and any action in the action space, and evaluate the benefits of taking different actions in different states based on factors such as animal comfort, energy consumption, and air quality;

[0031] S3.3: Initialize the value function and policy function of each action in the action space under each state in different state spaces, recursively calculate each value function and policy function through the Bellman optimal equation, and stop recursive updating when the value function change value is within the preset threshold;

[0032] S3.4: Extract the optimal strategy of the final recursive result, that is, the optimal action taken under each set of states, and then use the environmental modeling module to simulate the environmental changes under the optimal strategy, evaluate the actual effect of the strategy, and execute the optimal strategy in the actual environment, while adjusting the strategy based on real-time feedback.

[0033] An intelligent ventilation method for livestock breeding cages, the specific steps of the ventilation method are as follows:

[0034] Ⅰ. Collect and process cage environment data and extract required feature data;

[0035] Ⅱ. Predict the future trend of cage environmental parameters and generate the optimal ventilation strategy;

[0036] III. Control the operation of ventilation equipment and conduct abnormal monitoring based on the optimal ventilation strategy;

[0037] IV. Real-time monitoring of environmental data in the cage and dynamic adjustment of ventilation strategies;

[0038] V. Record and store environmental data and decisions and visualize various cage data;

[0039] VI. Issue alerts and execute emergency strategies in response to detected anomalies or emergencies.

[0040] As a further solution of the present invention, the specific steps of dynamically adjusting the ventilation strategy in step IV are as follows:

[0041] S4.1: Construct the root node of the search tree under the current environment state, take the current state of the cage as the initial state, add the current state data and all available action data to the root node, and then take the state change information after taking different actions as child nodes and add them to the search tree;

[0042] S4.2: Starting from the root node, select the upper confidence interval upper limit selection strategy to select child nodes according to the existing node information in the tree until a leaf node is reached. If the leaf node is not fully expanded, select an unexplored action from it and generate a new child node, which is added to the search tree;

[0043] S4.3: Starting from the newly expanded child node, randomly select actions to perform a complete simulation until the preset simulation time is reached. After the simulation is completed, the cumulative reward is calculated and back-propagated to all nodes on the same path in the search tree, and the number of visits and average reward value of each node are updated at the same time;

[0044] S4.4: Repeat the selection, expansion, simulation, and back-propagation steps to gradually expand the search tree and optimize the node reward value and visit times until the preset iteration time is reached. After the iteration, select the branch with the highest average reward as the optimized optimal strategy and adjust the ventilation equipment.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. The present invention constructs a corresponding state space and a corresponding action space based on the outputs of the environment modeling module and the change prediction module, calculates the probability of transferring to the next state under any state in the state space and any action in the action space, and evaluates the benefits of taking different actions under different states based on the comfort, energy consumption and air quality of the animal, initializes the value function and strategy function of each action in the action space under each state in different state spaces, recursively calculates each value function and strategy function through the Bellman optimal equation, stops recursive updating after the value function change value is within a preset threshold, extracts the optimal strategy of the final recursive result, that is, the optimal action taken under each group of states, and then uses the environment modeling module to simulate the environmental changes under the optimal strategy, evaluates the actual effect of the strategy, and executes the optimal strategy in the actual environment, and adjusts the strategy according to real-time feedback, so as to manage environmental conditions more actively, avoid environmental parameters deviating from the ideal state due to the hysteresis effect, reduce the stress effect on animals, meet the specific needs of different livestock and improve the air quality and animal health level in the barn.

[0047] 2. The present invention takes the current state of the cage as the initial state, and adds the current state data and all optional action data to the root node. Starting from the root node, according to the existing node information in the tree, the upper confidence interval upper limit selection strategy is selected to select the child node until the leaf node is reached. If the leaf node is not fully expanded, an unexplored action is selected from it, and a new child node is generated and added to the search tree. Starting from the newly expanded child node, an action is randomly selected to perform a complete simulation until the preset simulation time is reached. After the simulation, the cumulative reward is calculated, and the cumulative reward obtained by simulation is back-propagated to all nodes on the same path in the search tree, and the number of visits and the average reward value of each node are updated at the same time. The selection, expansion, simulation and back-propagation steps are repeated to gradually expand the search tree and optimize the reward value and the number of visits of the node until the preset iteration time is reached. After the iteration, the branch with the highest average reward is selected as the optimal strategy after optimization, and the ventilation equipment is adjusted to improve the decision-making accuracy of the system, ensure that the cage environment is always kept in the most ideal state, help balance the improvement of air quality in the short term and the minimization of energy consumption in the long term, and effectively optimize the overall performance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0049] Figure 1 This is a system block diagram of an intelligent ventilation system for animal husbandry cages proposed by the present invention;

[0050] Figure 2 This is a flowchart of an intelligent ventilation method for livestock breeding cages proposed by the present invention. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0052] Example 1

[0053] Reference Figure 1 , an intelligent ventilation system for livestock breeding cages, including a collection and processing module, an environment modeling module, a change prediction module, an optimization decision module, a strategy simulation module, a storage safety module, an equipment control module, a monitoring feedback module, a user interaction module and an emergency response module;

[0054] The acquisition and processing module is used to collect and pre-process various environmental data in the cage.

[0055] Specifically, various types of environmental data collected by various high-precision sensors are received, the corresponding environmental data set is constructed, and the first quartile Q1 and the third quartile Q3 of the data are calculated. Then, the interquartile range IQR is calculated based on Q1 and Q3, and the upper edge and lower edge ranges are obtained according to Q1, Q3 and IQR. If the environmental data point is greater than the upper edge value, or less than the lower edge value, the environmental data is judged to be abnormal data, and the detected abnormal value is deleted. Then, the missing values ​​in each environmental data in the environmental data set are detected, and the missing values ​​are filled by linear interpolation. The average value of the adjacent data points of each data point is calculated and replaced to smooth the data. When all data are smoothed, the Min-Max normalization method is used to scale each environmental data to the range of [0,1].

[0056] In this embodiment, the high-precision sensors specifically include thermocouple sensors, capacitive humidity sensors, infrared carbon dioxide sensors, electrochemical ammonia sensors, photoresistors, thermal wind speed sensors, laser dust sensors and electrochemical oxygen sensors; the environmental data specifically include temperature, humidity, carbon dioxide concentration, ammonia concentration, light intensity, wind speed and direction, air quality and oxygen concentration.

[0057] The environmental modeling module is used to establish a virtual model to simulate the environmental changes of the cage; the change prediction module is used to predict the changing trend of the cage environmental parameters.

[0058] Specifically, collect and preprocess various historical environmental data and weather forecast information, align data of different frequencies to the same time scale, use the mean filling method to process missing data points, extract time features and lag features from each group of data, construct a feature data set, divide the feature data set into training set, validation set and test set according to the time sequence, input the training set into the LSTM model, obtain the predicted output of the corresponding time step through the forward propagation of the LSTM model layer by layer, calculate the loss value between the predicted value and the actual value through the root mean square error function, the LSTM model back-propagates the loss value and calculates the gradient information of the loss value for each network layer of the model, based on the gradient information of each layer, update the parameter information of each layer through the Adam optimizer, after each round of training, use the validation set to evaluate the model performance, and according to the evaluation results, tune the model parameters through grid search, use the test set to evaluate the final model, calculate various performance indicators, and then deploy the trained model to the production environment, receive new environmental data and weather forecast data in real time, and output the prediction results of future environmental changes.

[0059] The optimization decision module is used to generate the optimal ventilation strategy within a specified time.

[0060] Specifically, according to the outputs of the environmental modeling module and the change prediction module, the environmental conditions at different time points are obtained to construct the corresponding state space, the different control strategies that the ventilation system can take are collected, the corresponding action space is constructed, and the probability of transferring to the next state under any state in the state space and taking any action in the action space is calculated. At the same time, based on the factors of animal comfort, energy consumption and air quality, the benefits of taking different actions in different states are evaluated, and the value function and strategy function of each action in the action space are initialized under each state in different state spaces. The value function and strategy function are recursively calculated through the Bellman optimal equation until the change value of the value function is within the preset threshold, then the recursive update is stopped, and the optimal strategy of the final recursive result is extracted, that is, the optimal action taken under each group of states. Then, the environmental modeling module is used to simulate the environmental changes under the optimal strategy, the actual effect of the strategy is evaluated, and the optimal strategy is executed in the actual environment, and the strategy is adjusted according to real-time feedback.

[0061] The strategy simulation module is used to simulate and evaluate the execution process of different ventilation strategies; the storage security module is used to store environmental data and decision records; the equipment control module is used to control the operation of ventilation equipment; the monitoring feedback module is used to monitor the ventilation effect in real time and adjust the system operating parameters; the user interaction module is used to adjust the operating parameters of the ventilation equipment and view real-time environmental data, the execution of ventilation strategies and historical records; the emergency response module is used to issue alarms and execute emergency strategies.

[0062] Example 2

[0063] Reference Figure 2 , an intelligent ventilation method for livestock breeding cages, the specific steps of the ventilation method are as follows:

[0064] Collect and process cage environment data and extract required feature data.

[0065] Predict future cage environmental parameter change trends and generate optimal ventilation strategies.

[0066] Control the operation of ventilation equipment according to the optimal ventilation strategy and perform abnormal monitoring.

[0067] Monitor the environmental data in the cage in real time and adjust the ventilation strategy dynamically.

[0068] Specifically, a root node of the search tree is constructed under the current environmental state, the current state of the cage is used as the initial state, and the current state data and all optional action data are added to the root node. Then, the change information of the state after taking different actions is used as a child node and added to the search tree. Starting from the root node, according to the existing node information in the tree, the upper confidence interval upper limit selection strategy is selected to select the child node until the leaf node is reached. If the leaf node is not fully expanded, an unexplored action is selected from it, and a new child node is generated and added to the search tree. Starting from the newly expanded child node, an action is randomly selected for a complete simulation until the preset simulation time is reached. After the simulation is completed, the cumulative reward is calculated, and the cumulative reward obtained by the simulation is back-propagated to all nodes on the same path in the search tree, and the number of visits and the average reward value of each node are updated at the same time. The selection, expansion, simulation and back-propagation steps are repeated to gradually expand the search tree and optimize the reward value and the number of visits of the node until the preset iteration time is reached. After the iteration is completed, the branch with the highest average reward is selected as the optimized optimal strategy, and the ventilation equipment is adjusted.

[0069] Record and store environmental data and decisions and visualize various cage data.

[0070] Raise alerts and execute emergency strategies for detected anomalies or emergencies.

Claims

1. An intelligent ventilation system for animal husbandry cages, characterized in that: It includes acquisition and processing module, environment modeling module, change prediction module, optimization decision module, strategy simulation module, storage security module, equipment control module, monitoring feedback module, user interaction module and emergency response module; The acquisition and processing module is used to acquire and pre-process various environmental data in the cage; The environmental modeling module is used to establish a virtual model to simulate the environmental changes of the cage; The change prediction module is used to predict the change trend of cage environmental parameters; The optimization decision module is used to generate the optimal ventilation strategy within a specified time; The strategy simulation module is used to simulate and evaluate the execution process of different ventilation strategies; The storage security module is used to store environmental data and decision records; The equipment control module is used to control the operation of the ventilation equipment; The monitoring feedback module is used to monitor the ventilation effect in real time and adjust the system operating parameters; The user interaction module is used to adjust the operating parameters of the ventilation equipment and view real-time environmental data, the execution of ventilation strategies and historical records; The emergency response module is used to issue an alarm and execute emergency strategies.

2. The intelligent ventilation system for animal husbandry cages according to claim 1, characterized in that: The specific steps of pre-processing the environmental data in the cage by the acquisition and processing module are as follows: S1.1: Receive various environmental data collected by various high-precision sensors, construct corresponding environmental data sets, and calculate the first quartile Q1 and the third quartile Q3 of the data, and then calculate the interquartile range IQR based on Q1 and Q3; S1.2: Obtain the upper and lower edge ranges based on Q1, Q3, and IQR. If the environmental data point is greater than the upper edge value, or less than the lower edge value, the environmental data is judged to be abnormal data, and the detected abnormal value is deleted. Then, the missing values ​​in each environmental data in the environmental data set are detected, and the missing values ​​are filled by linear interpolation. S1.3: Calculate the average value of the neighboring data points of each data point and replace them to smooth the data. After all the data are smoothed, scale each environmental data to the range of [0,1] using the Min-Max normalization method.

3. The intelligent ventilation system for animal husbandry cages according to claim 2, characterized in that: The high-precision sensors described in S1.1 specifically include thermocouple sensors, capacitive humidity sensors, infrared carbon dioxide sensors, electrochemical ammonia sensors, photoresistors, thermal wind speed sensors, laser dust sensors, and electrochemical oxygen sensors; The environmental data described in S1.1 specifically include temperature, humidity, carbon dioxide concentration, ammonia concentration, light intensity, wind speed and direction, air quality and oxygen concentration.

4. The intelligent ventilation system for animal husbandry cages according to claim 2, characterized in that: The specific steps of predicting the change trend of the cage environment parameters by the change prediction module are as follows: S2.1: Collect and preprocess historical environmental data and weather forecast information, align data of different frequencies to the same time scale, use the mean filling method to handle missing data points, extract time features and lag features from each set of data, and construct a feature data set; S2.2: Divide the feature data set into training set, validation set and test set according to the time sequence, input the training set into the LSTM model, obtain the prediction output of the corresponding time step through the forward propagation of the LSTM model layer by layer, and calculate the loss value between the predicted value and the actual value through the root mean square error function; S2.3: The LSTM model back-propagates the loss value and calculates the gradient information of the loss value for each network layer of the model. Based on the gradient information of each layer, the parameter information of each layer is updated through the Adam optimizer. After each round of training, the model performance is evaluated using the validation set, and the model parameters are tuned through grid search based on the evaluation results. S2.4: Use the test set to evaluate the final model and calculate various performance indicators. Then deploy the trained model to the production environment, receive new environmental data and weather forecast data in real time, and output future environmental change predictions.

5. The intelligent ventilation system for animal husbandry cages according to claim 4, characterized in that: The specific steps of the optimization decision module to generate the optimal ventilation strategy within a specified time are as follows: S3.1: Based on the outputs of the environmental modeling module and the change prediction module, obtain the environmental conditions at different time points to construct the corresponding state space, collect different control strategies that the ventilation system can adopt, and construct the corresponding action space; S3.2: Calculate the probability of transitioning to the next state for any state in the state space and any action in the action space, and evaluate the benefits of taking different actions in different states based on factors such as animal comfort, energy consumption, and air quality; S3.3: Initialize the value function and policy function of each action in the action space under each state in different state spaces, recursively calculate each value function and policy function through the Bellman optimal equation, and stop recursive updating when the value function change value is within the preset threshold; S3.4: Extract the optimal strategy of the final recursive result, that is, the optimal action taken under each set of states, and then use the environmental modeling module to simulate the environmental changes under the optimal strategy, evaluate the actual effect of the strategy, and execute the optimal strategy in the actual environment, while adjusting the strategy based on real-time feedback.

6. An intelligent ventilation method for livestock breeding cages, used to realize the intelligent ventilation system function of livestock breeding cages according to any one of claims 1 to 5, characterized in that: The specific steps of the ventilation method are as follows: Ⅰ. Collect and process cage environment data and extract required feature data; Ⅱ. Predict the future trend of cage environmental parameters and generate the optimal ventilation strategy; III. Control the operation of ventilation equipment and conduct abnormal monitoring based on the optimal ventilation strategy; IV. Real-time monitoring of environmental data in the cage and dynamic adjustment of ventilation strategies; V. Record and store environmental data and decisions and visualize various cage data; VI. Issue alerts and execute emergency strategies in response to detected anomalies or emergencies.

7. The intelligent ventilation method for animal husbandry cages according to claim 6, characterized in that: The specific steps of dynamically adjusting the ventilation strategy described in step IV are as follows: S4.1: Construct the root node of the search tree under the current environment state, take the current state of the cage as the initial state, add the current state data and all available action data to the root node, and then take the state change information after taking different actions as child nodes and add them to the search tree; S4.2: Starting from the root node, select the upper confidence interval upper limit selection strategy to select child nodes according to the existing node information in the tree until a leaf node is reached. If the leaf node is not fully expanded, select an unexplored action from it and generate a new child node, which is added to the search tree; S4.3: Starting from the newly expanded child node, randomly select actions to perform a complete simulation until the preset simulation time is reached. After the simulation is completed, the cumulative reward is calculated and back-propagated to all nodes on the same path in the search tree, and the number of visits and average reward value of each node are updated at the same time; S4.4: Repeat the selection, expansion, simulation, and back-propagation steps to gradually expand the search tree and optimize the node reward value and visit times until the preset iteration time is reached. After the iteration, select the branch with the highest average reward as the optimized optimal strategy and adjust the ventilation equipment.

Images