An intelligent breeding environment optimization control method and system

Abstract

The application relates to the technical field of environment control, and provides a method and system for optimizing control of a smart breeding environment. The method comprises the following steps: performing digital twin modeling of a site of the smart breeding environment, and distributing environment control equipment; fitting a control intensity field in the digital twin model, combining a behavior heat map to construct a grid topology map; arranging monitoring sensors in each grid area, establishing an environment data set, collecting behavior characteristics to construct a behavior characteristic set; fusing the two sets to evaluate a health risk grade; performing control strategy optimization based on the risk grade to generate a local strategy set; performing global coordination fitting based on a cooperative scheduling identifier, and executing optimization control according to the global coordination fitting. The application solves the technical problems of low matching degree between environment parameter regulation and control and biological behavior demand, resource waste caused by lack of global coordination of local control strategies, and lag of health risk early warning in the existing breeding environment control method, and achieves the technical effect of improving the intelligent level and the fine level of breeding environment control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of environmental control, in particular to an intelligent breeding environment optimization control method and system. BACKGROUND

[0002] In modernized farm management, precise regulation of breeding environment has a decisive influence on animal health, breeding efficiency and economic benefits. The existing technology relies on static threshold regulation of single environmental parameters (such as temperature and humidity, ammonia concentration), lacks deep perception of dynamic demand of breeding biological behavior, and causes disconnection between environmental regulation and biological physiological demand. For example, fixed ventilation strategy may cause local overload or energy waste due to neglecting real-time changes of animal gathering hot area; at the same time, the existing system mostly adopts "single-point monitoring-independent control" mode, and there is lack of cooperation between devices, which easily causes control conflict (such as simultaneous operation of humidification and cooling devices). Although the Internet of Things technology improves the data acquisition capability, the analysis of environmental data and biological behavior data is separated, which is difficult to realize early warning of health risk, and lagging regulation often leads to outbreak of stress diseases.

[0003] In addition, the application of digital twin technology in the field of breeding is still in the initial stage, and the existing model mostly focuses on physical field simulation (such as temperature field distribution), and does not integrate biological behavior characteristics (such as activity heat map and feeding rule) into control decision, which causes deviation between virtual model and actual biological response. Multi-objective optimization in environmental control is often limited to the balance between energy consumption and comfort, and does not include biological indicators such as health risk level and behavior delay response into the optimization target, which is difficult to realize precise intervention. Therefore, it is urgent to develop an intelligent breeding environment optimization control method and system based on digital twin, behavior perception and multi-objective optimization, to realize deep perception of environment and animal state in the farm, dynamic assessment of health risk, adaptive generation of regional control strategy and global coordinated control, so as to improve the intelligent level and overall operation efficiency of breeding. SUMMARY

[0004] The present application provides an intelligent breeding environment optimization control method and system, aiming to solve the technical problems of low matching degree between environmental parameter regulation and biological behavior demand in the existing breeding environment control method, resource waste caused by lack of global coordination of local control strategy, and lagging health risk warning, to achieve fine health risk assessment by fusing behavior perception and environmental perception, to construct local control and global coordination mechanism under multi-objective optimization, and to improve the intelligent and fine level of breeding environment control.

[0005] In a first aspect, the application discloses an optimization control method of a smart breeding environment, which comprises the following steps: performing site digital twin modeling of the smart breeding environment and distributing environment control devices; performing control strength field fitting of the environment control devices in the digital twin model, and constructing a grid topology graph by using fitting results and behavior heat maps of the breeding animals, wherein the grid topology graph is provided with a cooperative scheduling correlation identifier; deploying monitoring sensors in each grid area, performing environment monitoring of the grid area by using the monitoring sensors, establishing a grid environment data set, collecting behavior characteristics of the breeding animals by using biological sensors, and establishing a grid behavior characteristic set; evaluating a health risk level of each grid breeding animal by using the grid behavior characteristic set and the grid environment data set; performing control strategy optimization of each grid area according to the health risk level based on a multi-objective optimization channel, and establishing a local control strategy set; and performing global coordination fitting of the local control strategy set by using the cooperative scheduling correlation identifier of the network topology graph, and performing optimization control by using global coordination fitting results.

[0006] In another aspect, the application discloses an optimization control system of a smart breeding environment, which comprises the following modules: a twin modeling module that performs site digital twin modeling of the smart breeding environment and distributes environment control devices; a topology graph construction module that performs control strength field fitting of the environment control devices in the digital twin model, and constructs a grid topology graph by using fitting results and behavior heat maps of the breeding animals, wherein the grid topology graph is provided with a cooperative scheduling correlation identifier; a characteristic collection module that deploys monitoring sensors in each grid area, performs environment monitoring of the grid area by using the monitoring sensors, establishes a grid environment data set, collects behavior characteristics of the breeding animals by using biological sensors, and establishes a grid behavior characteristic set; a risk level evaluation module that evaluates a health risk level of each grid breeding animal by using the grid behavior characteristic set and the grid environment data set; a strategy optimization module that performs control strategy optimization of each grid area according to the health risk level based on a multi-objective optimization channel, and establishes a local control strategy set; and an optimization control module that performs global coordination fitting of the local control strategy set by using the cooperative scheduling correlation identifier of the network topology graph, and performs optimization control by using global coordination fitting results.

[0007] The one or more technical solutions provided in the application have at least the following technical effects or advantages:

[0008] The above-mentioned optimization control method of the intelligent breeding environment divides a grid space with a cooperative scheduling identifier based on dynamic fusion of a device control intensity field and a breeding animal behavior heat map by constructing a digital twin model of a breeding site, synchronously deploying environment control equipment, and integrating environment monitoring sensors and biological sensors in each grid to collect environment parameters (such as temperature and humidity, gas concentration) and behavior characteristics (such as activity distribution, feeding frequency) of breeding animals to form a multidimensional data set. Then, by correlating the environment and biological data, the health risk level of breeding animals in each grid is evaluated, and a multi-objective optimization algorithm is used to generate a local control strategy that takes into account energy consumption, response speed, and health needs for different risk levels. Finally, by coordinating regional strategies through the cooperative identifier in the grid topology, device conflicts are eliminated and global resource optimization allocation is achieved, thereby dynamically adjusting environment control parameters to achieve the comprehensive goals of precise regulation, risk warning, and efficient resource utilization.

[0009] The above description is only a summary of the technical solutions of the present application. In order to more clearly understand the technical means of the present application, the following specific embodiments of the present application can be implemented according to the content of the specification, and in order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the following specific embodiments of the present application will be described. BRIEF DESCRIPTION OF DRAWINGS

[0010] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor on the basis of these drawings.

[0011] Figure 1 A flowchart of an embodiment of an optimization control method of an intelligent breeding environment.

[0012] Figure 2 An architecture diagram of an embodiment of an optimization control system of an intelligent breeding environment.

[0013] Legend: Twin modeling module 11, topology graph construction module 12, feature acquisition module 13, risk level evaluation module 14, strategy optimization module 15, optimization control module 16. DETAILED DESCRIPTION

[0014] The embodiment of the present application provides an optimization control method and system of a smart breeding environment, solves the technical problems of low matching degree of environment parameter regulation and biological behavior demand, resource waste caused by lack of global coordination of local control strategy and lag of health risk early warning in the existing breeding environment control method, and achieves the technical effects of realizing fine health risk assessment by fusing behavior perception and environment perception, constructing a local control and global coordination mechanism under multi-objective optimization, and improving the intelligent and fine level of breeding environment control.

[0015] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the scope of protection of the present application.

[0016] It should be noted that the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or server including a series of steps or units does not have to be limited to those clearly listed steps or units, but can include other steps or modules that are not clearly listed or inherent to the process, method, product or device.

[0017] Embodiment one, as shown in the present application provides an optimization control method of a smart breeding environment, the method comprises: Figure 1

[0018] Performing site digital twin modeling of a smart breeding environment, and distributing environment control equipment.

[0019] ​In the embodiments of the present application, first, a virtual site framework is constructed according to the environmental data of the site (such as the terrain of the breeding area, the regional distribution), the structure of the breeding facility (such as the feeding area, the drainage area), a preliminary physical model is formed, the physical model includes the geometric parameters (such as area, shape, layout) of each area, the connection relationship (such as pipeline connection, electrical wiring) between facilities, etc. Subsequently, the boundary conditions and constraint conditions related to the breeding environment (such as environmental temperature range, humidity control target, air flow requirement, etc.) are input, the parameter range of modeling and the control analysis dimension (such as temperature and humidity regulation, air quality control, etc.) are limited through these conditions, so as to provide accurate basis for simulation control. Then, all collected parameters are transmitted into a digital twin simulation platform (such as a BIM platform, a CFD simulation tool, etc.), boundary initialization and physical modeling are performed, and a complete digital twin model is output, which can provide a spatial basis for subsequent digital twin analysis and control strategy. In the process of constructing the model, the environmental control equipment (such as fans, heaters, air conditioners, etc.) will be arranged in the virtual space according to the position information respectively, and the running state and interaction of them under different environmental conditions are simulated. In this way, the prediction of environmental changes and the dynamic monitoring of equipment operation in the digital model can be realized, which provides data support for accurate control of the environment.

[0020] The control intensity field fitting of the environmental control equipment is performed in the digital twin model, and a grid topology graph is constructed by using the fitting result and the behavior heat map of the breeding animal, wherein the grid topology graph is provided with a cooperative scheduling association identifier.

[0021] In one embodiment, in the digital twin model, by simulating and quantifying the control capabilities (e.g. adjustment range and intensity of factors such as temperature, humidity, etc.) of environmental control devices (such as fans, heaters, air conditioners, etc.) at different locations, a control intensity dataset is generated, and a fitting result of the control intensity field is generated according to the control intensity dataset, which describes the control effect of the devices at each location, helping to determine which areas of the environment need more or less adjustment. Subsequently, the behavior heat map of the cultured objects is obtained, which shows the activity frequency and behavior distribution of the cultured objects in different areas. By combining the fitting result of the control intensity field with the behavior heat map, a complementary index of each location is calculated, which can better understand the interaction between the cultured objects and the environment, and thus provide a more accurate basis for the division of each area. Then, based on the calculated complementary index, the farm is divided into multiple grids by clustering reconstruction, and each grid represents an independent control unit. By assembling these grids according to their locations, a grid topology map is constructed, which has a collaborative scheduling correlation identifier established according to the boundary interference sensitivity, indicating the relationship and scheduling priority between different grids, so as to realize the collaborative work and resource optimization allocation between different grids.

[0022] Further, the present application provides the control intensity field fitting of the environmental control device in the digital twin model, comprising:

[0023] obtaining the device control gear of the environmental control device, and establishing a distance decay function mapped with the environmental control device; in the digital twin model, simulating the device control gear of the environmental control device by using the distance decay function, establishing a control intensity dataset of the location point; and generating a fitting result of the control intensity field by using the control intensity dataset.

[0024] Preferably, for each environmental control device (such as fan, heater, air conditioner, etc.), its control gear parameters are obtained, which refer to different power or working intensity levels that the device can adjust, for example, a fan may have low, medium and high gears, an air conditioner may have different temperature setting gears, etc. By combining the device control gear with the physical characteristics, a distance decay function of the control intensity with the spatial distance decay is established, which is as follows: ; wherein, is the control intensity when the distance is d, is the baseline strength of the device in the current gear (such as the maximum wind speed), k is the attenuation coefficient, which is calibrated by the device type and environmental characteristics, and d is the Euclidean distance from the spatial position to the device. Subsequently, in the digital twin model, the established distance attenuation function is used to simulate the impact of different gears of the device in space. That is, according to the position of the device, the position of the target area and the simulated gear, the control strength of each location point is calculated, and then these control strength data are added to a set in sequence to obtain the control strength data set of all location points. Afterwards, the control strength data set is divided according to the gear, and the control strength field of each gear at different positions is used as the fitting result to show the control effect of different areas under different parameters, helping to determine which areas require more control resources (such as increasing the power of the equipment or adjusting the position of the equipment) and which areas do not require too much adjustment, thereby providing a basis for subsequent control strategies.

[0025] Table 1: Environmental control equipment parameters:

[0026] ;

[0027] Table 2: Position point control strength calculation table (Example: Position point P(12,16)):

[0028] ;

[0029] Table 1 above is a table of environmental control equipment parameters. It shows the basic parameters of three types of environmental control equipment (fan, heater, and humidifier), including equipment type, control gear, baseline intensity, attenuation coefficient, and the equipment's location coordinates in the aquaculture environment. The control intensity of each device is related to its gear, operating conditions (such as temperature, humidity, etc.), and attenuation coefficient, which in turn affects its control effect in the space.

[0030] Table 2 above is a table for calculating the control strength of a location point. This table shows the distance from the device to the control point (unit: m) and the control strength data at that location. The table lists each device's ID, the distance to the specified point, and the corresponding control strength data (such as wind speed, temperature, and humidity). This is used to calculate the control effect of the device at different locations and optimize the control strategy.

[0031] Furthermore, the present application provides a method for constructing a grid topology map using the fitting results and the behavior heat map of the cultivated object, including:

[0032] Perform a joint analysis of the fitting results and the behavior heat map to calculate the behavior-control complementarity index, which is calculated as follows: ;in, Characterization location The behavior-control complementarity index, is the location activity generated based on the behavior heat map, characterizing a position gradient of the control intensity field generated based on the fitting result; dividing the smart farming environment into regular squares to form a basic mesh set; calculating an internal area mean for each basic mesh using the behavior-control complementarity indicator; and performing cluster reconstruction using the internal area mean to generate a mesh topology map.

[0033] Preferably, after obtaining the fitting result, the behavior heat map is combined with the control intensity field of the environmental control device to calculate the behavior-control complementarity indicator, which is calculated as follows: ; wherein, characterizing a position The behavior-control complementarity indicator is used to measure the matching degree between behavior and control, and reflects the complementarity of environmental control and animal behavior at a specific location. A higher behavior-control complementarity indicator indicates that there is both high animal activity and suitable environmental control at the location, and a lower value indicates insufficient control or mismatch between animal activity and control effect. is the position activity extracted from the behavior heat map, characterizing a position gradient of the control intensity field generated based on the fitting result, which is obtained by calculating the rate of change of control intensity with position, reflecting the change in the action intensity of the environmental control device in space. In the smart farming environment, the entire site is divided into regular mesh units to form a basic mesh set. Each basic mesh in the basic mesh set represents an independent area. The mesh can be evenly rectangular or square, and the size of each mesh can be set according to actual needs (e.g., 1 m or 2 m). The purpose of this step is to discretize the farming environment so that each mesh can be independently analyzed for behavior and control. For each basic mesh, the mean value of the behavior-control complementarity indicator for all positions within the mesh is calculated to obtain the internal area mean of each basic mesh, providing a basis for subsequent mesh clustering and topology map generation. Subsequently, the internal area mean is used to cluster and reconstruct all meshes, i.e., to analyze the internal deviation of each basic mesh based on the internal area mean, evaluate the matching degree of behavior and control within each mesh, and help determine which mesh regions have a large difference between behavior and control, and then divide these regions. Then, according to the division result, the mean value of the divided regions is recalculated, and different regions are clustered and reconstructed to generate a new mesh topology map. Finally, the boundary interference sensitivity of the cluster is also calculated to evaluate the interference degree between adjacent regions, and the sensitivity information is used to construct a collaborative scheduling correlation identifier to ensure the coordination and optimization effect between regions. Through these steps, the generated mesh topology map not only reflects which regions have a high matching degree of environmental control and animal behavior, but also guides the subsequent control strategy and resource allocation.

[0034] Further, the application provides that the clustering reconstruction using the internal region mean value generates a grid topology map, comprising:

[0035] The deviation analysis of the internal region mean value is used to analyze the deviation of the internal region of the basic grid, and the internal split region is constructed. After recalculating the region mean value of the internal split region, the clustering merging evaluation of adjacent regions is performed, the clustering evaluation result is used to complete the clustering reconstruction, and the boundary interference sensitivity of the clustering cluster is calculated. The collaborative scheduling correlation identifier is established based on the calculated boundary interference sensitivity.

[0036] Optionally, for each basic grid, the internal deviation analysis is performed according to the internal region mean value, that is, the deviation of the behavior-control complementarity index of the internal position point of the basic grid from the internal region mean value of the basic grid is calculated, and the calculated deviation is compared with the deviation threshold value to determine which position points have a large deviation from the region mean value. The behavior and control of these position points with large deviation may be mismatched. At this time, these position points are removed from the basic grid, and the internal split region is constructed by using these position points. Subsequently, a new internal region mean value is calculated for each split region, which represents the average value of the behavior-control complementarity of all position points in the split region. If the deviation of the behavior-control complementarity of all position points in the split region from the new internal region mean value does not exceed the deviation threshold value, it indicates that the control of the internal split region is consistent. At this time, the internal split region and the corresponding basic grid are taken as a clustering cluster respectively. Otherwise, the evaluation of adjacent regions (split regions in the internal split region) of the next level is continued until the deviation of the internal region mean value of the split region does not exceed the deviation threshold value. After completing the clustering merging evaluation, the clustering evaluation result is used to reconstruct the basic grid set, that is, the basic grid set is re-divided according to the clustering clusters divided in the clustering evaluation result, and a grid topology map is constructed according to the divided grid to reflect a more refined region division. Subsequently, the calculation of the control intensity field gradient change of adjacent regions is performed for each clustering cluster corresponding to the grid in the grid topology map to quantify the boundary interference sensitivity between adjacent regions (the control intensity field gradient change can be directly used as the boundary interference sensitivity). If the control intensity of two adjacent regions changes dramatically (i.e., the gradient is large), it indicates that the boundary interference between them is sensitive. Therefore, the boundary interference sensitivity between the two adjacent regions is larger, and more collaborative adjustment is required. Finally, based on the calculated boundary interference sensitivity, the collaborative scheduling correlation identifier is established for the boundary region between each clustering cluster. The identifier indicates that the collaborative scheduling between adjacent regions is required to ensure the smooth transition of the control strategy and the optimal allocation of resources, thereby improving the management efficiency of the overall breeding environment.

[0037] The monitoring sensors are deployed inside each grid area, and the monitoring sensors are used for environmental monitoring of the grid area to establish a grid environment dataset, and the behavior characteristics of the cultured animals are collected by using the biological sensors to establish a grid behavior characteristic set.

[0038] In one embodiment, in the smart breeding environment, monitoring sensors are installed in each grid area to collect environmental data in the area. These sensors can monitor and record key parameters in the environment in real time, such as temperature, humidity, light, air quality, etc. The data collected by the sensors will build a detailed grid environment dataset, which contains the environmental state information in each grid area, to help evaluate the changes in the environment in real time and make corresponding adjustments. At the same time, biological sensors are used to monitor the behavior characteristics of the cultured animals. Biological sensors can capture animal activity information in various ways, such as using video monitoring to track animal movements, using sound sensors to detect animal call frequency, or using pressure sensors to monitor animal activity density. These behavior data will form a grid behavior characteristic set, including activity density, aggregation state, call frequency, etc. in each grid area. Through these behavior characteristics, the behavior patterns of animals in each grid area can be understood in real time, and the environmental control strategy can be dynamically adjusted according to these characteristic data to ensure the health and production efficiency of the cultured animals.

[0039] The grid behavior characteristic set and the grid environment dataset are used to evaluate the health risk level of each grid cultured animal.

[0040] In one embodiment, after obtaining the grid behavior characteristic set and the grid environment dataset, the activity density, aggregation state, and call frequency in the grid behavior characteristic set, and the temperature, humidity, air quality, etc. in the grid environment dataset are standardized to make these data in the same dimension. Then, according to the weight determined by the domain expert in advance, the data is weighted to quantify a health risk indicator. According to these indicators, the health risk of each grid is divided into different levels, such as low risk, medium risk, high risk, etc., to help the system identify areas with higher risk in time and adjust the breeding environment according to the evaluation results, so as to intervene at the earliest stage and ensure effective management of the breeding environment and animal health.

[0041] Further, the application provides the use of the grid behavior characteristic set and the grid environment dataset to evaluate the health risk level of each grid cultured animal, which includes:

[0042] A health risk feature vector is constructed by fusing the grid behavior characteristic set and the grid environment dataset. After normalization processing of the health risk feature vector, a health risk indicator is calculated by a weighted scoring function. The health risk indicator is used to generate the health risk level of each grid cultured animal.

[0043] Preferably, after obtaining the grid behavior feature set and the grid environment data set, each data point in the grid behavior feature set and the grid environment data set is combined to form a unified health risk feature vector, and each element of the health risk feature vector represents health risk related information of the aquaculture animals in the grid. In the health risk feature vector, the dimensions and value ranges of different features may be different, for example, the range of environmental temperature may be from 0 to 50℃, and the behavior feature value of the aquaculture animals may be a standardized percentage, therefore, normalization processing is required to ensure that the influence of each feature is balanced when calculating the health risk, and common normalization methods include minimum-maximum normalization and Z-score standardization. Through normalization processing, all feature values in the health risk feature vector are standardized to the same scale, so that they have comparability in subsequent calculation of the health risk index. After the normalization processing is completed, a preset weighted scoring function is activated, and the weighted scoring function records the weight value corresponding to each feature. Through the weighted scoring function, the normalized features and the corresponding weights are multiplied and summed to obtain the health risk index of each grid, and the higher the health risk index, the greater the health risk of the aquaculture animals in the region, wherein the specific form of the weighted scoring function is: R is the health risk index; n is the dimension of the feature vector, that is, the total number of features; is the weight corresponding to the i th feature, which is usually determined by domain experts (such as veterinarians and environmental control engineers) according to practical experience in the early stage, and can be determined according to the Pearson correlation coefficient between each feature and the historical health risk result when there is a large amount of historical data, that is, the Pearson correlation coefficient of each feature is calculated using the ratio of the total Pearson correlation coefficient; is the normalized i th feature value. Finally, the calculated health risk index is compared with a plurality of preset risk intervals to determine the health risk level of the aquaculture animals in each grid, and exemplarily, the plurality of risk intervals can be low risk: [0, 0.3), medium risk: [0.3, 0.7), and high risk: [0.7, 1]. In summary, through the above process, the system can calculate the health risk index of the aquaculture animals in each grid region based on the behavior features and the environment data, and further divide them into different health risk levels, according to which the health status of the aquaculture animals can be finely managed, and timely measures such as adjusting the environmental control and improving the animal behavior can be taken to ensure the optimization of the aquaculture environment.

[0044] Further, the application provides that the health risk index is used to generate the health risk level of the aquaculture animals in each grid, including:

[0045] configuring a multi-scale sliding window; performing time series analysis on the grid behavior feature set under the multi-scale sliding window to generate an enhanced health risk vector; and using the enhanced health risk vector to compensate for the health risk level.

[0046] Optionally, a multi-scale sliding window is configured according to actual needs. The multi-scale sliding window is a method of analyzing time series data at different time scales. The window size can be selected as short-term (e.g., 5 minutes), medium-term (e.g., 3 minutes), and long-term (e.g., 1 hour). Subsequently, the grid behavior feature set is sorted according to the time stamp to obtain a grid behavior feature sequence set. Then, according to each window scale defined in the multi-scale sliding window, the behavior data of each sliding window is extracted from the start of the grid behavior feature sequence set, and the mean value of each behavior feature in the window is calculated. After that, the mean values of each behavior feature under the same window scale are combined to form an enhanced health risk vector for each window scale. Then, by weighting each normalized mean value feature in the enhanced health risk vector, the compensation coefficient of each window scale is calculated. Then, the compensation coefficient corresponding to the enhanced health risk vector is used to compensate for the current health risk level. In this process, the preset window weight of each window scale is multiplied by the corresponding compensation coefficient, and then accumulated to obtain a comprehensive compensation coefficient. Then, the compensation coefficient is multiplied by the health risk indicator corresponding to the current health risk level to dynamically adjust the health risk level, making the health risk level assessment more accurate. In summary, this dynamic compensation mechanism can effectively respond to fluctuations in behavior patterns and ensure that health risk assessment can respond to actual changes in a timely manner.

[0047] Based on the multi-objective optimization channel, the control strategy of each grid area is optimized according to the health risk level, and a local control strategy set is established.

[0048] In one embodiment, the optimization of the control strategy is performed according to the health risk level of each grid area, which reflects the health status of the aquatic animals in the area. The higher the risk level, the more refined control strategy is needed for the area. To ensure the effectiveness of environmental control, multiple optimization objectives are balanced through a multi-objective optimization channel, including control energy efficiency, health response matching degree, and response delay. During the optimization process, the health risk level is synchronized to the multi-objective optimization function of the multi-objective optimization channel, and the weight of the multi-objective optimization function is dynamically adjusted. After the adjustment of the multi-objective optimization function is completed, the local control strategy is generated based on the adjusted multi-objective optimization function. For high-risk areas, health response matching degree and fast response are prioritized to ensure timely adjustment of environmental conditions. In low-risk areas, more emphasis is placed on energy efficiency optimization to reduce energy consumption. Based on this optimization process, local control strategies suitable for each grid area are generated and stored in a set to obtain a local control strategy set. In summary, as the health risk level changes, the local control strategy set is dynamically adjusted to ensure that the environment of each area is always in the most suitable state for the health of the aquatic animals, thereby achieving fine management and intelligent optimization of the breeding environment.

[0049] Further, the application provides the multi-objective optimization channel based on the health risk level to optimize the control strategy of each grid area, and establishes a local control strategy set, which includes:

[0050] A multi-objective optimization function is established, and the control objectives of the multi-objective optimization function include control energy consumption target, health response matching degree target, and behavior response delay target. The health risk level is synchronized to the multi-objective optimization function, and the weight factor of the multi-objective optimization function is dynamically updated. The multi-objective optimization function with updated weight factor is used to optimize the control strategy of each grid area, and a local control strategy set is established.

[0051] Preferably, the multi-objective optimization channel has a pre-constructed multi-objective optimization function that can simultaneously optimize multiple control objectives, including control energy consumption target, health response matching degree target, and behavior response delay target, to ensure the best overall effect in the breeding environment. The calculation method of the multi-objective optimization function is to weight the control energy consumption target, health response matching degree target, and behavior response delay target, and the specific form is: wherein, is the final multi-objective optimization function value, and the optimization objective is to minimize the function; is the control energy consumption target, which is used to minimize energy consumption, for example, to reduce the power usage of air conditioners, heaters, fans, and other devices while ensuring that the breeding environment meets the health requirements; The health response matching degree target aims to optimize the breeding environment to match the health needs of the breeding animals as much as possible. The behavior response delay target is used to reduce the response delay of the breeding environment to changes in the behavior of the breeding animals, for example, if the activity pattern of the breeding animals changes (such as a decrease or increase in activity), the environmental conditions (such as temperature, humidity, ventilation, etc.) should be quickly adjusted to respond to the behavior changes in a timely manner. 、 、 The weight coefficients corresponding to the three targets are derived from the risk-weight mapping table (previously constructed by domain experts). Then, the health risk level of each grid area is matched with the risk-weight mapping table to obtain the control target weight corresponding to each health risk level, for example, when the health risk level is high risk, the control target weight can be 0.2 (control energy consumption target), 0.5 (health response matching degree target), and 0.3 (behavior response delay target), and for low risk, the control target weight can be 0.5 (control energy consumption target), 0.3 (health response matching degree target), and 0.2 (behavior response delay target). The matched control target weight is then synchronized to the multi-objective optimization function to replace the original weight factor of each control target in the multi-objective optimization function. After that, using the updated weight factor, the pre-set optimization algorithm such as particle swarm optimization, genetic algorithm, etc. is used to optimize the control strategy of each grid area. The purpose of optimization is to find the most suitable control strategy combination according to the control targets (energy consumption, health response, behavior response delay, etc.) of each grid area. For each control strategy combination obtained, the digital twin model is used for simulation to obtain the corresponding control energy consumption target, health response matching degree target, and behavior response delay target. Then, the multi-objective optimization function is used for evaluation until the result of the multi-objective optimization function is less than or equal to the optimization threshold, thereby obtaining a local control strategy set. This local control strategy set includes adjustment measures suitable for the health risk, environmental conditions, and behavior changes of the region, such as temperature setting of air conditioners, working state of humidity adjustment equipment, etc. In summary, by establishing a multi-objective optimization function, a balance between energy consumption, health response matching degree, and behavior response delay can be achieved to generate the optimal control strategy that adapts to different grid areas. Dynamically adjusting the weight factor of the optimization function enables the system to respond to changes in the health risk level in real time and generate a corresponding local control strategy set for each region, thereby ensuring animal health while achieving energy efficiency optimization and minimizing response delay.

[0052] The global coordination fitting of the local control strategy set is performed using the cooperative scheduling association identifier of the network topology graph, and the optimization control is performed using the global coordination fitting result.

[0053] In one embodiment, after obtaining the local control policy set, the local control policy of each grid area is evaluated, the adjustment cost is calculated, and the adjustment penalty term is generated, which reflects the cost that may be brought in the adjustment process, aiming to avoid excessive adjustment or invalid intervention, so as to ensure the rationality of the control policy adjustment. Subsequently, through the cooperative scheduling association identifier, the consistency between the control policies of different grid areas is analyzed, which helps to identify which areas of the control policy need to be adjusted synchronously to avoid conflicts or inconsistent effects between local policies. On this basis, the results of the control consistency analysis and the adjustment penalty term are used to perform adjustment optimization from a global perspective, which aims to optimize the global policy and ensure that the control effects between different grid areas are consistent, maximizing the performance of the overall system. Finally, according to the global coordination fitting result, the optimized control policy is executed, which not only considers the requirements of each grid area, but also adjusts the environmental control policy of each area according to the requirements of coordination and optimization, ensuring the overall efficiency and stability of environmental regulation.

[0054] Further, the application provides the global coordination fitting of the local control policy set using the cooperative scheduling association identifier of the network topology graph, and the execution of the optimized control using the global coordination fitting result, comprising:

[0055] The adjustment cost of the corresponding local control policy set for each grid area is calculated, and the adjustment penalty term is generated. The control consistency analysis of the local control policy set is performed using the cooperative scheduling association identifier, and the adjustment optimization from a global perspective is performed based on the control consistency analysis result and the adjustment penalty term, and the global coordination fitting result is established.

[0056] Preferably, after obtaining the set of local control strategies, the digital twin model is used to simulate each local control strategy for each grid area, and record the energy efficiency cost, health response cost and response delay cost during the simulation, wherein the energy efficiency cost is the product of the actual power consumption in the simulation of the control device (which is affected by the adjacent grid area) and the time of device operation, the health response cost is the weighted result of the deviation of the actual temperature in the simulation from the optimal temperature required for the health of the livestock, and the deviation of the actual humidity from the optimal humidity required for the health of the livestock, and the response delay cost is the actual time delay between the occurrence of the behavior change in the simulation and the response of the control strategy. Then, the energy efficiency cost, the health response cost and the response delay cost are normalized and weighted summed to obtain the adjustment penalty term of each grid area. Then, the relationship between adjacent areas is determined by using the cooperative scheduling association identifier, i.e., whether there is a large interference between adjacent areas, by comparing the cooperative scheduling association identifier of adjacent areas with the cooperative threshold value. If it is greater than or equal to the cooperative threshold value, it means that the two adjacent areas need to be controlled synchronously, otherwise it means that they can be controlled independently. Then, according to the result of the consistency analysis, the adjustment penalty term of the grid area that needs to be cooperatively controlled is introduced for adjustment optimization from a global perspective. In this process, it is determined whether the adjustment cost of the local control strategy involved is too high. If there is a penalty tolerance greater than or equal to the penalty tolerance, the adjustment amplitude of the strategy is reduced to reduce the negative impact on the system as a whole. After adjustment, simulation is performed again until there is no local control strategy with a penalty tolerance greater than or equal to the penalty tolerance. Finally, the adjusted local control strategies are summarized to establish a global coordination fitting result, which contains the control strategy of the entire breeding environment, ensuring that the control strategy of each grid area not only achieves the optimal state, but also coordinates with the control strategy of the adjacent area.

[0057] Further, the application provides that after the optimization control is performed using the global coordination fitting result, the following steps are included:

[0058] A response cycle is established, in which the environmental response monitoring of the intelligent breeding environment is performed, and a cycle response result is generated; the cycle response result and the global coordination fitting result are subjected to control anomaly identification, and an abnormality warning is issued based on the anomaly identification.

[0059] Preferably, a response cycle is established for periodically monitoring the environmental response in the smart breeding environment. The response cycle can be set to a certain time period according to the needs, for example, once an hour or once a day. In each response cycle, various data in the breeding environment, such as temperature, humidity, air flow, light intensity, etc. are collected and analyzed. Through these data, the periodic response results are generated, which reflect the actual environmental conditions in the cycle. Then, the periodic response results are compared with the expected control effect in the global coordination fitting results to determine whether there are environmental parameters with large fluctuations, i.e., whether the deviation of a certain environmental parameter exceeds the fluctuation tolerance range. If significant deviations are found between the actual environment and the expected effect during the comparison process, it indicates that the current environment is abnormal, such as equipment failure, environmental control failure, etc. Subsequently, based on the abnormal identification result, an abnormal warning is immediately issued to remind the management personnel. In this way, the control effect of the breeding environment can be monitored in real time, potential problems can be found in time, and the health of the breeding animals and the stability of the environment can be ensured.

[0060] In summary, the embodiments of the present application have at least the following technical effects:

[0061] The embodiments of the present application first perform site digital twin modeling of the smart breeding environment and distribute environmental control equipment; then, control strength field fitting of the environmental control equipment is performed in the digital twin model, and a grid topology graph is constructed using the fitting results and the behavior heat map of the breeding animals, wherein the grid topology graph is provided with a cooperative scheduling association identifier; thereafter, monitoring sensors are deployed inside each grid region, and the monitoring sensors are used for environmental monitoring of the grid region to establish a grid environment data set, and biological sensors are used to collect behavior characteristics of the breeding animals to establish a grid behavior characteristic set; further, the health risk level of each grid breeding animal is evaluated using the grid behavior characteristic set and the grid environment data set; then, control strategy optimization of each grid region is performed based on the health risk level according to the multi-objective optimization channel to establish a local control strategy set; finally, global coordination fitting of the local control strategy set is performed using the cooperative scheduling association identifier of the network topology graph, and the optimization control is performed using the global coordination fitting results. These technical effects collectively solve the technical problems of low matching degree of environmental parameter regulation and biological behavior demand, resource waste caused by lack of global coordination of local control strategies, and lag of health risk early warning in the existing breeding environment control method, and achieve the technical effects of realizing fine health risk evaluation by fusing behavior perception and environment perception, constructing a local control and global coordination mechanism under multi-objective optimization, and improving the intelligent and fine level of breeding environment control.

[0062] Embodiment two, based on the same inventive concept as the optimization control method of the smart breeding environment in the foregoing embodiments, such as Figure 2As shown, the present application provides an optimization control system for a smart farming environment, the system comprising: a twin modeling module 11: executing digital twin modeling of the site of the smart farming environment and distributing environmental control equipment; a topology map construction module 12: executing control intensity field fitting of the environmental control equipment in the digital twin model, and constructing a grid topology map using the fitting results and the behavior heat map of the farmed objects, wherein the grid topology map is provided with a collaborative scheduling association identifier; a feature acquisition module 13: deploying monitoring sensors in each grid area, using the monitoring sensors to perform environmental monitoring of the grid area, establishing a grid environment data set, and using biosensors to collect behavioral characteristics of farmed objects to establish a grid behavior feature set; a risk level assessment module 14: using the grid behavior feature set and the grid environment data set to assess the health risk level of each grid farmed object; a strategy optimization module 15: performing control strategy optimization for each grid area according to the health risk level based on a multi-objective optimization channel, and establishing a local control strategy set; an optimization control module 16: using the collaborative scheduling association identifier of the network topology map to perform global coordinated fitting of the local control strategy set, and performing optimization control using the global coordinated fitting result.

[0063] Furthermore, the topology map construction module 12 is further configured to execute the following method:

[0064] Obtain the device control gear of the environmental control device and establish a distance attenuation function mapped to the environmental control device; in the digital twin model, use the distance attenuation function to perform device control gear simulation of the environmental control device and establish a control intensity data set of the location point; use the control intensity data set to generate a fitting result of the control intensity field.

[0065] Furthermore, the topology map construction module 12 is further configured to execute the following method:

[0066] Perform a joint analysis of the fitting results and the behavior heat map to calculate the behavior-control complementarity index, which is calculated as follows: ;in, Characterization location The behavior-control complementarity index, is the location activity generated based on the behavior heat map, Characterize the position gradient of the control intensity field generated based on the fitting result; divide the smart farming environment into regular grids to form a basic grid set; calculate the internal area mean of each basic grid using the behavior-control complementarity index; and perform cluster reconstruction using the internal area mean to generate a grid topology map.

[0067] Furthermore, the topology map construction module 12 is further configured to execute the following method:

[0068] The internal region mean is used for deviation analysis of the internal base grid, and an internal split region is constructed. After recalculating the region mean of the internal split region, the clustering and merging evaluation of adjacent regions is performed, the clustering evaluation result is used to complete the clustering reconstruction, the boundary interference sensitivity of the clustering cluster is calculated, and the collaborative scheduling correlation identifier is established by using the boundary interference sensitivity.

[0069] Further, the risk level evaluation module 14 is also used to execute the following method:

[0070] A health risk feature vector is constructed by fusing the grid behavior feature set and the grid environment data set. After normalizing the health risk feature vector, the health risk index is calculated by a weighted scoring function. The health risk index is used to generate the health risk level of each grid aquaculture object.

[0071] Further, the risk level evaluation module 14 is also used to execute the following method:

[0072] A multi-scale sliding window is configured. The grid behavior feature set is analyzed by time series analysis under the multi-scale sliding window to generate an enhanced health risk vector. The enhanced health risk vector is used for health risk level compensation.

[0073] Further, the strategy optimization module 15 is also used to execute the following method:

[0074] A multi-objective optimization function is established, and the control targets of the multi-objective optimization function include control energy consumption targets, health response matching degree targets, and behavior response delay targets. The health risk level is synchronized to the multi-objective optimization function, and the weight factor of the multi-objective optimization function is dynamically updated. The multi-objective optimization function after updating the weight factor is used for control strategy optimization of each grid region to establish a local control strategy set.

[0075] Further, the optimization control module 16 is also used to execute the following method:

[0076] The adjustment cost calculation of the corresponding local control strategy set is performed for each grid region to generate an adjustment penalty term. The control consistency analysis of the local control strategy set is performed by using the collaborative scheduling correlation identifier. The adjustment optimization is performed based on the control consistency analysis result and the adjustment penalty term from a global perspective to establish a global coordination fitting result.

[0077] Further, the optimization control module 16 is also used to execute the following method:

[0078] A response cycle is established, and the environment response monitoring of the intelligent aquaculture environment is performed in the response cycle to generate a cycle response result. The control abnormality identification is performed on the cycle response result and the global coordination fitting result, and the abnormality warning is reported based on the abnormality identification.

[0079] It should be noted that the above-mentioned order of the embodiments of the present application is merely for description, and does not represent the advantages and disadvantages of the embodiments. And the above describes the specific embodiments of the present application. The processes depicted in the drawings do not necessarily require the specific order and continuous order shown to achieve the desired results. In some embodiments, multi-task processing and parallel processing are also possible or can be advantageous.

[0080] The above only describes the preferred embodiments of the present application, and does not limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

[0081] The present application and the drawings are merely exemplary descriptions of the present application, and are considered to cover any and all modifications, changes, combinations or equivalents within the scope of the present application. Obviously, those skilled in the art can make various modifications and changes to the present application without departing from the scope of the present application. Thus, if these modifications and changes of the present application belong to the scope of the present application and its equivalent technology, the present application intends to include these modifications and changes.

Claims

1. A method for optimizing and controlling a smart breeding environment, characterized in that: The method comprises: Implement site digital twin modeling of smart farming environments and distribute environmental control equipment; Performing control intensity field fitting of the environmental control device in the digital twin model, and constructing a grid topology map using the fitting results and the behavior heat map of the cultivated animals, wherein the grid topology map is provided with a collaborative scheduling association identifier; The method of constructing a grid topology map using the fitting results and the behavior heat map of the cultivated animals includes: Perform a joint analysis of the fitting results and the behavior heat map to calculate the behavior-control complementarity index, which is calculated as follows: ； in, Characterization location The behavior-control complementarity index, is the location activity generated based on the behavior heat map, characterize the position gradient of the control intensity field generated based on the fitting results; Dividing the smart farming environment into regular grids to form a basic grid set; Calculating the internal region mean for each basic grid using the behavior-control complementarity index; Performing cluster reconstruction using the internal region mean to generate a grid topology map; The clustering reconstruction using the internal area mean to generate a grid topology map includes: Using the internal region mean value to perform deviation analysis inside the basic grid and construct an internal split region; After recalculating the regional mean of the internal split region, perform cluster merging evaluation of adjacent regions, use the cluster evaluation results to complete cluster reconstruction, calculate the boundary interference sensitivity of the cluster cluster, and use the boundary interference sensitivity to establish the collaborative scheduling association identifier; Deploy monitoring sensors in each grid area, use the monitoring sensors to monitor the environment of the grid area, establish a grid environment data set, and use biosensors to collect behavioral characteristics of farmed animals to establish a grid behavioral feature set; Using the grid behavior feature set and the grid environment data set, the health risk level of the cultured animals in each grid is evaluated; Based on the multi-objective optimization channel, the control strategy of each grid area is optimized according to the health risk level, and a local control strategy set is established; The collaborative scheduling association identifier of the network topology graph is used to perform global coordinated fitting of the local control strategy set, and the global coordinated fitting result is used to perform optimal control.

2. The optimization control method of a smart breeding environment according to claim 1, characterized in that: The performing of control intensity field fitting of the environmental control device in the digital twin model includes: Obtaining a device control gear of an environmental control device and establishing a distance attenuation function mapped to the environmental control device; In the digital twin model, the distance decay function is used to perform a device control gear simulation of the environmental control device to establish a control intensity data set of the location point; The control intensity data set is used to generate a fitting result of the control intensity field.

3. The optimization control method of a smart breeding environment according to claim 1, characterized in that: The method of evaluating the health risk level of each grid cultured animal using the grid behavior feature set and the grid environment data set includes: Constructing a health risk feature vector that integrates the grid behavior feature set and the grid environment data set; After normalizing the health risk feature vector, a health risk index is calculated using a weighted scoring function; The health risk index is used to generate a health risk level for each grid culture.

4. The optimization control method of a smart breeding environment according to claim 3, characterized in that: The step of generating the health risk level of each grid cultured animal using the health risk indicator includes: Configure multi-scale sliding windows; Performing a time series analysis on the grid behavior feature set under a multi-scale sliding window to generate an enhanced health risk vector; The enhanced health risk vector is used to perform health risk level compensation.

5. The optimization control method of a smart breeding environment according to claim 1, characterized in that: The multi-objective optimization channel is used to optimize the control strategy for each grid area according to the health risk level, and establish a local control strategy set, including: Establishing a multi-objective optimization function, wherein the control objectives of the multi-objective optimization function include controlling energy consumption objectives, health response matching objectives, and behavior response delay objectives; Synchronizing the health risk level to the multi-objective optimization function, and dynamically updating the weight factor of the multi-objective optimization function; The multi-objective optimization function after updating the weight factors is used to optimize the control strategy of each grid area and establish a local control strategy set.

6. The optimization control method of a smart breeding environment according to claim 1, characterized in that: The method of performing global coordinated fitting of a local control strategy set using the collaborative scheduling association identifier of the network topology graph and performing optimized control using the global coordinated fitting result includes: Calculate the adjustment cost of the corresponding local control strategy set for each grid area and generate the adjustment penalty term; The collaborative scheduling association identifier is used to perform control consistency analysis of the local control strategy set, and based on the control consistency analysis result and the adjustment penalty item, an adjustment optimization is performed from a global perspective to establish a global coordinated fitting result.

7. The optimization control method for a smart breeding environment according to claim 1, characterized in that: After the optimization control is performed using the global coordinated fitting result, the method includes: Establishing a response cycle, performing environmental response monitoring of the smart farming environment during the response cycle, and generating a cycle response result; Anomaly identification is performed on the periodic response result and the global coordinated fitting result, and an anomaly warning is issued based on the anomaly identification.

8. An optimization control system for a smart breeding environment, characterized in that: The system is used to execute the optimization control method of a smart farming environment according to any one of claims 1 to 7, and the system comprises: Twin modeling module: performs site digital twin modeling of the smart farming environment and distributes environmental control equipment; Topology map construction module: performs control intensity field fitting of environmental control equipment in the digital twin model, and uses the fitting results and the behavior heat map of the farmed animals to construct a grid topology map, wherein the grid topology map is provided with a collaborative scheduling association identifier; Feature collection module: Deploy monitoring sensors in each grid area, use the monitoring sensors to monitor the environment of the grid area, establish a grid environment data set, and use biosensors to collect behavioral characteristics of farmed animals to establish a grid behavior feature set; Risk level assessment module: using the grid behavior feature set and the grid environment data set to assess the health risk level of each grid culture; Strategy optimization module: Based on the multi-objective optimization channel, the control strategy of each grid area is optimized according to the health risk level, and a local control strategy set is established; Optimization control module: Use the collaborative scheduling association identifier of the network topology graph to perform global coordination fitting of the local control strategy set, and use the global coordination fitting results to perform optimization control.

Images