A smart aquaculture management platform driven by big data

Abstract

This invention relates to the field of intelligent aquaculture and facility agriculture automation control technology, specifically a big data-driven intelligent aquaculture management platform. The platform includes a multi-dimensional data sensing module for acquiring target energy supply status data, aquatic environmental physicochemical data, and aquatic organism behavior data; an energy supply and demand prediction module for generating an envelope representing available energy within a preset time window based on energy supply status data and preset micro-meteorological forecast information; a biochemical response analysis module for calculating the cumulative stress value representing the superimposed stress, combined with a preset aquatic biochemical inertia model; a multi-objective game optimization module for calculating a dynamic resource scheduling strategy using a game model based on the available energy envelope, physiological tolerance function, and the cumulative stress value; and an execution control module that responds to the dynamic resource scheduling strategy by generating coordinated control commands for feeding and aeration equipment. This invention effectively reduces the latent mortality rate of aquatic organisms.

Description

Technical Field

[0001] This invention relates to the field of intelligent aquaculture and facility agriculture automation control technology, specifically to an intelligent aquaculture management platform driven by big data. Background Technology

[0002] In the current offshore intensive aquaculture management scenario, the aquaculture platform usually relies on distributed independent energy systems such as photovoltaic and wind power to provide power support for feeding equipment, oxygenation equipment and environmental monitoring sensors. At the same time, changes in aquatic environmental parameters have a significant lag effect on the physiological state of farmed organisms.

[0003] To maintain a stable aquaculture environment, existing management solutions generally adopt a passive response architecture based on a single real-time threshold. This means controlling the start and stop of aeration equipment by monitoring whether the current dissolved oxygen concentration is below a set standard, and implementing mechanized feeding according to a fixed schedule. Although this solution has basic operational capabilities under normal weather and energy-sufficient conditions, the highly random and intermittent nature of natural energy supply, coupled with the rigid and uninterrupted oxygen supply requirement for the survival of farmed organisms, leads to a severe spatiotemporal mismatch between resource supply and consumption. Furthermore, existing solutions lack quantitative analysis of the cumulative effects of biological physiological stress, failing to identify the risk of organisms being in a latent sub-healthy state even when instantaneous indicators meet the standards. This results in power outages and oxygen depletion at critical moments during periods of continuous rain or extreme conditions due to a lack of planning for future energy budgets, or ineffective feeding during biological stress, leading to feed waste and water quality deterioration. Ultimately, it is difficult to achieve the global optimization of aquaculture benefits and risk costs.

[0004] Therefore, how to achieve quantitative prediction of biological physiological tolerance status and dynamic resource scheduling based on energy boundaries through multidimensional data analysis under the constraints of limited energy supply and dynamic environmental changes has become an urgent technical problem to be solved. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a smart aquaculture management platform driven by big data. Specifically, the technical solution of this invention includes:

[0006] The multi-dimensional data sensing module is used to acquire data on energy supply status, water environment physicochemical data, and aquatic organism behavior data of the target aquaculture area;

[0007] The energy supply and demand forecasting module is used to generate an available energy envelope that represents the upper limit sequence of available energy within a future preset time window, based on energy supply status data and preset micro-meteorological forecast information.

[0008] The biochemical response analysis module is used to determine the dissolved oxygen decay recovery curve under different oxygenation powers based on the physicochemical data of the water environment and the preset water biochemical inertial model. Based on the dissolved oxygen decay recovery curve and the behavior data of aquatic organisms, it calculates the physiological tolerance function that characterizes the survival probability of aquatic organisms as environmental parameters change and the stress accumulation value that characterizes the risk over time.

[0009] The multi-objective game optimization module is used to construct a game model with the goal of maximizing biological growth benefits and minimizing mortality risk costs based on the available energy envelope, physiological tolerance function and stress accumulation value, and to calculate dynamic resource scheduling strategies based on the game model.

[0010] The execution control module is used to generate coordinated control commands for feeding equipment and oxygenation equipment in response to dynamic resource scheduling strategies.

[0011] Preferably, the process by which the multidimensional data perception module acquires data includes:

[0012] Collect real-time power of photovoltaic and wind turbines, remaining battery power and battery health as energy supply status data;

[0013] Underwater temperature, salinity, depth, dissolved oxygen concentration, pH value, and hydrodynamic flow velocity are collected as physicochemical data of the aquatic environment.

[0014] The swimming speed and distribution density of organisms were collected through machine vision, and the feeding spectrum data were collected through acoustic sensors, as behavioral data of farmed organisms.

[0015] Preferably, the process by which the energy supply and demand forecasting module generates the available energy envelope includes:

[0016] By analyzing the correlation between historical meteorological data and historical power generation data, an energy conversion efficiency model can be established.

[0017] The preset micro-meteorological forecast information is input into the energy conversion efficiency model to predict the theoretical power generation within the preset time window in the future.

[0018] By combining the remaining battery power in the energy supply status data, the upper limit of the energy budget for each time node within the future preset time window is calculated to form the available energy envelope.

[0019] Preferably, the process by which the biochemical response analysis module determines the physiological tolerance function and cumulative stress value includes:

[0020] Analyze the time lag response characteristics of dissolved oxygen concentration in aquatic environmental physicochemical data under different oxygenation interventions, and generate dissolved oxygen decay recovery curves;

[0021] Establish a nonlinear mapping relationship between multidimensional environmental parameters and biological tolerance indicators;

[0022] The dissolved oxygen decay recovery curve and the behavior data of farmed organisms are input into a nonlinear mapping relationship to calculate the instantaneous mortality probability of farmed organisms under the current environmental combination. This is used to construct a physiological tolerance function, and the instantaneous mortality probability is calculated by time integration to obtain the stress accumulation value.

[0023] Preferably, the process by which the multi-objective game optimization module calculates the dynamic resource scheduling strategy includes:

[0024] Define an objective function, which includes a biological growth benefit term, a mortality risk cost term, and an energy consumption cost term.

[0025] Set constraints, including battery charge safety threshold and dissolved oxygen lethal threshold;

[0026] Under the premise of satisfying the constraints, find the global optimal solution that makes the objective function reach the extreme value within the range of the available energy envelope;

[0027] The combination of feeding amount and oxygenation power corresponding to the global optimal solution is determined as the dynamic resource scheduling strategy.

[0028] Preferably, the multi-objective game optimization module executes the following logic during the solution process:

[0029] When the available energy envelope shows that the future energy supply is lower than the preset shortage threshold, the reduction in metabolic oxygen consumption after reducing the feeding amount is calculated based on the physiological tolerance function.

[0030] If the reduction in metabolic oxygen consumption can maintain dissolved oxygen concentration above the dissolved oxygen lethal threshold, a defensive strategy of reducing feeding and pre-charging the battery is generated.

[0031] If the reduction in metabolic oxygen consumption is insufficient to maintain dissolved oxygen concentration above the dissolved oxygen lethal threshold, a survival-first strategy is generated to prioritize the operation of the aerator.

[0032] Preferably, the process of the execution control module generating cooperative control instructions includes:

[0033] Analyze the dynamic resource scheduling strategy and extract the opening parameters of the feeding valve and the frequency conversion parameters of the aerator;

[0034] The opening parameters of the feeding valve are converted into the execution signal of the feeding equipment; the frequency conversion parameters of the aerator are converted into the motor speed control signal of the aerator.

[0035] By utilizing feedback from the amount of remaining feed in the behavioral data of aquaculture organisms, a closed-loop correction is performed on the opening parameters of the feeding valve.

[0036] Preferred options also include:

[0037] The resilience index assessment module is used to calculate the normalized ratio between the cumulative stress value and the energy supply capacity represented by the available energy envelope, which serves as the aquaculture ecological resilience index.

[0038] The multi-objective game optimization module is also configured to: determine whether the aquaculture ecological resilience index is lower than the preset safety value; if so, force a switch to risk avoidance mode, in which the dynamic resource scheduling strategy takes the minimum dissolved oxygen requirement to maintain the survival of organisms as the sole constraint objective.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] 1. This invention constructs an available energy envelope, transforming random natural energy into a visualized future budget boundary, thus solving the spatiotemporal mismatch between fluctuations in independent energy supply and the rigid needs of biological survival. The system can predict energy gaps within a preset time window and lock in the budget or switch strategies in advance before extreme weather arrives, effectively avoiding the risk of power outages and oxygen deficiency at critical moments in the future due to current overconsumption, and ensuring energy security for offshore aquaculture.

[0041] 2. This invention introduces biochemical response analysis and stress accumulation value calculation, breaking the limitations of traditional methods that only focus on real-time dissolved oxygen indicators. By quantifying the lag effect of the aquatic environment and the latent fatigue of organisms, the system can identify hidden dangers where instantaneous indicators meet the standards but organisms are already on the verge of physiological collapse. This achieves a leap from simple indicator management to in-depth life management, significantly reducing the probability of pond overflow caused by the superposition of chronic stress.

[0042] 3. Based on a multi-objective game model, this invention achieves a dynamic balance between growth benefits and survival risks. Under conditions of energy scarcity, the system can automatically generate defensive strategies, utilizing the biological principle that reducing food intake can reduce metabolic oxygen consumption, and sacrificing short-term growth benefits to gain a long-term survival probability. This adaptive adjustment mechanism based on energy conservation greatly enhances the survival resilience of the aquaculture system under severe sea conditions such as typhoons and continuous rain.

[0043] 4. This invention establishes a closed-loop feedback mechanism based on multi-dimensional visual and acoustic perception, which uses the amount of remaining feed and biological behavior data to correct the feeding strategy in real time. When the system detects that the organism's appetite has decreased due to stress or illness, it immediately blocks ineffective feeding, which not only avoids the waste of expensive feed, but also prevents the rotting of uneaten feed from causing further deterioration of water quality, thus achieving a dual optimization of economic benefits and ecological protection. Attached Figure Description

[0044] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0045] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0047] Example 1:

[0048] Please see Figure 1 A smart aquaculture management platform driven by big data includes: a multi-dimensional data sensing module, used to acquire energy supply status data, water environment physicochemical data, and aquaculture organism behavior data of the target aquaculture area;

[0049] The energy supply and demand forecasting module is used to generate an available energy envelope that represents the upper limit sequence of available energy within a future preset time window, based on energy supply status data and preset micro-meteorological forecast information.

[0050] The biochemical response analysis module is used to determine the dissolved oxygen decay recovery curve under different oxygenation powers based on the physicochemical data of the water environment and the preset water biochemical inertial model. Based on the dissolved oxygen decay recovery curve and the behavior data of aquatic organisms, it calculates the physiological tolerance function that characterizes the survival probability of aquatic organisms as environmental parameters change and the stress accumulation value that characterizes the risk over time.

[0051] The multi-objective game optimization module is used to construct a game model with the goal of maximizing biological growth benefits and minimizing mortality risk costs based on the available energy envelope, physiological tolerance function and stress accumulation value, and to calculate dynamic resource scheduling strategies based on the game model.

[0052] The execution control module is used to generate coordinated control commands for feeding equipment and oxygenation equipment in response to dynamic resource scheduling strategies.

[0053] This embodiment details the architecture and core logic of the platform; the multi-dimensional data perception module acts as the system's sensory organ, responsible for collecting all data from the physical world to provide the original basis for subsequent decision-making; the energy supply and demand forecasting module acts as an energy auditor, defining future resource boundaries based on the acquired data;

[0054] The biochemical response analysis module acts as a biological translator, determining the dissolved oxygen decay recovery curve under different oxygenation powers based on the acquired physicochemical data of the water environment and the preset water biochemical inertia model. This water biochemical inertia model is derived from thermodynamic modeling, specifically based on the two-membrane theory and the oxygen mass transfer differential equation. Its physical meaning is to describe the buffering capacity of the water body to describe the rate of decrease in dissolved oxygen concentration over time after the loss of artificial oxygenation intervention.

[0055] Based on this, using dissolved oxygen decay recovery curves and aquaculture organism behavior data, the physiological tolerance function characterizing the survival probability of aquaculture organisms as environmental parameters changes, and the stress accumulation value characterizing the risk over time are calculated. The physiological tolerance function is derived from biological experimental statistics and is specifically fitted using the Weibull survival analysis model. Its physical meaning is the instantaneous mortality probability density of a specific fish species under different combinations of environmental stress. The stress accumulation value is used to assess the sub-health reserves of the organisms. The multi-objective game optimization module acts as the decision-making brain, searching for the optimal solution within the resource boundary. The execution control module acts as the system's hands and feet, translating decisions into physical actions of the equipment.

[0056] This embodiment introduces biochemical response analysis, breaking the limitation of traditional aquaculture that only focuses on whether dissolved oxygen meets the standard at the current moment. In the offshore aquaculture scenario, the system can identify the hidden risk that although the dissolved oxygen meets the standard, the fish population is on the verge of collapse due to prolonged mild hypoxia. This achieves a leap from simple indicator management to in-depth life management, effectively solving the contradiction between the randomness of energy supply and the rigid needs of biological survival.

[0057] Example 2:

[0058] The process of acquiring data by the multi-dimensional data sensing module includes: collecting real-time power of photovoltaic and wind turbines, remaining battery power and battery health as energy supply status data; collecting underwater temperature, salinity, depth, dissolved oxygen concentration, pH value and hydrodynamic flow velocity as aquatic environmental physicochemical data; collecting the swimming speed and distribution density of organisms identified by machine vision, and the feeding spectrum data collected by acoustic sensors as aquaculture organism behavior data.

[0059] This embodiment specifies the multi-dimensional data sensing module; on the energy supply side, the module directly reads the controllers of the photovoltaic inverter and wind turbine through the Modbus protocol to obtain real-time power. At the same time, the remaining battery power is read through the battery management system. and battery health In the aquatic environment, a distributed sensor array deployed at different water depths is used to collect underwater temperature data. ,salinity ,depth Dissolved oxygen concentration and pH value During this period, hydrodynamic velocities were collected using a Doppler current profiler. This parameter is designed to correct for water exchange rate;

[0060] In terms of the behavior of farmed organisms, the average swimming speed of fish schools is extracted using a binocular underwater camera combined with computer vision algorithms. and spatial distribution density Simultaneously, underwater hydrophones were used to collect acoustic signals in the frequency range of 20Hz to 20kHz, and feeding spectrum data was extracted using fast Fourier transform. To determine the feeding activity level of the fish population;

[0061] This embodiment achieves holographic perception of the aquaculture environment by integrating visual, acoustic, and physicochemical sensor data. In particular, by introducing battery health and hydrodynamic flow velocity, the system can fully consider the aging and deterioration of the energy storage system and the passive oxygenation effect brought by natural water flow in subsequent predictions, thereby significantly improving the robustness of decision inputs under complex sea conditions.

[0062] Example 3:

[0063] The process of generating the available energy envelope by the energy supply and demand forecasting module includes: calling historical meteorological data and historical power generation data for correlation analysis to establish an energy conversion efficiency model; inputting preset micro-meteorological forecast information into the energy conversion efficiency model to predict the theoretical power generation within the preset time window; and combining the remaining battery power in the energy supply status data to calculate the upper limit of the energy budget for each time node within the preset time window to form the available energy envelope.

[0064] This embodiment details the logic for generating the available energy envelope; the system calls historical meteorological data, such as light intensity and wind speed, and performs correlation analysis with historical power generation data, using a regression algorithm to establish an energy conversion efficiency model; preset micro-meteorological forecast information, i.e., the ultra-local weather forecast for the next 48 hours, is input into the model to predict the future preset time window. The theoretical power generation sequence within; combined with the remaining battery power in the energy supply status data. The energy budget ceiling for each time point within a preset future time window is calculated to form an available energy envelope; the calculation model is as follows:

[0065]

[0066] in, Derived from calculation results, the physical meaning is time. The upper limit of available energy, in kilowatt-hours; Derived from the system clock, its physical meaning is the current time; Derived from model predictions, its physical meaning is a power generation sequence; Derived from preset parameters, its physical meaning is overall charging efficiency; This data is read from the BMS and its physical meaning is the current remaining battery capacity; the unit is kWh. If the BMS output is a percentage, it needs to be multiplied by the total battery capacity for conversion. The data is read from the BMS and its physical meaning is the battery health coefficient.

[0067] This embodiment constructs an available energy envelope, enabling the system to no longer blindly execute control commands but to know the future energy pool thickness in advance. This ensures that the system can lock in the energy budget in advance before extreme weather such as continuous rainy days, effectively avoiding the risk of power outages at critical moments in the future due to current excessive consumption.

[0068] Example 4:

[0069] The process of determining the physiological tolerance function and stress accumulation value in the biochemical response analysis module includes: analyzing the time lag response characteristics of dissolved oxygen concentration in the physicochemical data of the aquatic environment under different aeration interventions, generating dissolved oxygen decay recovery curves; establishing a nonlinear mapping relationship between multidimensional environmental parameters and biological tolerance indicators; inputting the dissolved oxygen decay recovery curves and aquaculture biological behavior data into the nonlinear mapping relationship, calculating the instantaneous mortality probability of aquaculture organisms under the current environmental combination, thereby constructing the physiological tolerance function, and performing time integration calculation on the instantaneous mortality probability to obtain the stress accumulation value.

[0070] This embodiment focuses on quantifying the lagged effects of environmental changes on organisms; based on a water body biochemical inertia model, it analyzes the time-lag response characteristics of dissolved oxygen concentration under different aeration interventions, and generates dissolved oxygen decay recovery curves; the water body biochemical inertia model is specifically expressed as follows:

[0071]

[0072] in, Dissolved oxygen concentration, This represents the saturated dissolved oxygen concentration. This represents the initial dissolved oxygen concentration. The mass transfer coefficient is . This represents respiratory oxygen consumption rate, expressed in units of oxygen. To address the variable tracking problem, the hydrodynamic velocity in Example 2 was clarified. In this embodiment, the mass transfer coefficient plays a specific role in the model. Defined as a function of oxygenation power and water flow velocity, its calculation formula is as follows:

[0073]

[0074] in, This is the fit index; it typically ranges from 0.5 to 1.0 and is determined through tracer experiments based on the geometry of the water tank. The oxygenation effect coefficient, its dimensions depend on The unit ensures The unit is , To increase oxygen input power, The coefficient representing the influence of water flow. This represents the hydrodynamic flow velocity; the curve describes the dissolved oxygen concentration after aeration is stopped. Nonlinear decreasing process over time;

[0075] A nonlinear mapping relationship between multidimensional environmental parameters and biological tolerance indices is constructed; dissolved oxygen decay recovery curves and aquaculture organism behavioral data are input into this function to calculate the instantaneous mortality probability of aquaculture organisms under the current environmental combination. This leads to the construction of a physiological tolerance function, which is modeled as follows.

[0076]

[0077] in Probability of death This represents the overall environmental stress value. For scale parameters, For shape parameters; here, the combined environmental pressure value. The specific calculation formula is defined as follows:

[0078]

[0079] in, For each weight coefficient, its initial recommended value can be set as follows: , , Furthermore, in practical applications, fine-tuning is performed based on the differences in the sensitivity of aquaculture species to dissolved oxygen or temperature. Dissolved oxygen concentration, This represents the saturated dissolved oxygen concentration. For real-time water temperature, For the optimal temperature, For stocking density, The ultimate carrying capacity density was used. The distribution parameters were obtained by conducting gradient hypoxia stress experiments on target farmed fish samples, recording mortality rates at different time points, and fitting the experimental data using the maximum likelihood estimation method to calculate the scale parameters. In this embodiment, the value is 10.5, which is related to the shape parameter. In this embodiment, the value is taken as 2.2; the instantaneous probability of death is calculated by time integration to obtain the cumulative stress value. The calculation model is as follows:

[0080]

[0081] in, Derived from integral calculations, its physical meaning is time interval. The cumulative stress value; Derived from a preset constant, its physical meaning is the basic stress coefficient, and its dimensions are set as follows: ; Derived from experimental calibration, its physical meaning is tolerance attenuation factor; Derived from model parameters, its physical meaning is the restoring force decay index; The data is collected from sensors and its physical meaning is a dissolved oxygen concentration sequence.

[0082] Derived from biological databases, its physical meaning is the biological lethal oxygen threshold. Derived from model parameters, its physical meaning is the density crowding stress coefficient. It is obtained by measuring the rate of change of cortisol concentration in fish through a stress experiment with a gradient density of 10-50 kg / m³, and then calibrating it through linear regression. The value ranges from 0.15 to 0.45. Derived from visual recognition, its physical meaning is biological distribution density; Derived from aquaculture standards, its physical meaning is standard reference density, such as 30 kg / m³, used for dimensional normalization;

[0083] This embodiment not only considers the direct damage caused by low oxygen, but also introduces a density crowding stress coefficient to reflect the risk amplification effect brought about by mutual friction and competition for space among fish in high-density aquaculture. By quantifying the cumulative stress value, the system can accurately assess the latent fatigue of the fish population, providing a scientific quantitative basis for subsequent precise game theory.

[0084] Example 5:

[0085] The process of the multi-objective game optimization module to calculate the dynamic resource scheduling strategy includes: defining an objective function, which includes a biological growth benefit term, a mortality risk cost term, and an energy consumption cost term; setting constraints, which include a battery power safety threshold and a dissolved oxygen lethal threshold; finding the global optimal solution that makes the objective function reach its extreme value within the range limited by the available energy envelope, under the premise of satisfying the constraints; and determining the combination of the feeding amount and oxygenation power corresponding to the global optimal solution as the dynamic resource scheduling strategy.

[0086] The multi-objective game optimization module executes the following logic during the solution process: when the available energy envelope shows that the future energy supply is lower than the preset scarcity threshold, it calculates the reduction in metabolic oxygen consumption after reducing the feeding amount based on the physiological tolerance function; if the reduction in metabolic oxygen consumption can maintain the dissolved oxygen concentration above the dissolved oxygen lethal threshold, a defensive strategy of reducing feeding and pre-charging the battery is generated; if the reduction in metabolic oxygen consumption cannot maintain the dissolved oxygen concentration above the dissolved oxygen lethal threshold, a survival priority strategy that prioritizes the operation of the aerator is generated.

[0087] This embodiment details the construction of the game theory model and the logic for generating dynamic strategies; it defines the objective function. This function encompasses the benefits of biological growth, the costs of mortality risk, and the costs of energy consumption, aiming to solve the problem of balancing survival and growth under limited energy resources; the objective function is as follows:

[0088]

[0089] in, It originates from optimization calculations and its physical meaning is the cumulative total utility value within a preset time window; Derived from the time window setting, its physical meaning is to optimize the termination time and satisfy... ; Derived from the growth model, its physical meaning is the growth reward function, and its specific form is defined as:

[0090]

[0091] in This refers to the feed conversion ratio; Derived from the strategy variable, its physical meaning is the sequence of feeding amounts; Derived from the biochemistry module, its physical meaning is the instantaneous probability of death; Sourced from environmental data, its physical meaning is dissolved oxygen concentration; Derived from market data, its physical meaning is the value of biological assets; Derived from the energy consumption model, in order to achieve computable optimization, its relationship with the policy variables is clarified. The functional relationship is as follows:

[0092]

[0093] in This is the linear power consumption coefficient, which is dimensionless and typically takes a value of 1. These are nonlinear loss weighting coefficients, in units of This is used to characterize the cost of efficiency degradation of a motor under non-rated operating conditions. The price per unit of energy, in physical terms, is the cost of energy. Derived from strategy variables, its physical meaning is oxygenation power; Derived from preset weights, its physical meaning is weight coefficient;

[0094] The method for determining the weighting coefficients is as follows: Set to the current market price per unit of adult fish. It is set as the sum of the replacement cost of fish fry and the sunk cost of feed already invested. This is set as an energy cost weighting coefficient, a dimensionless constant, typically set to 1, used to balance the magnitudes of different optimization objectives; constraints are set, including energy consumption not exceeding the upper limit of the envelope and dissolved oxygen being above the minimum survival threshold. Based on this, the system searches for the globally optimal solution that maximizes the objective function within the range defined by the available energy envelope, specifically employing a particle swarm optimization algorithm or a genetic algorithm for the nonlinear objective function. Iterative optimization is performed to obtain the optimal resource allocation sequence within a preset time window; and the corresponding combination of feeding amount and oxygenation power is determined as the dynamic resource scheduling strategy.

[0095] During the solution process, the system performs rigorous logical judgments: in response to the available energy envelope. This indicates that future energy supply will be lower than a preset shortage threshold. scarcity threshold The setting logic is as follows:

[0096]

[0097] in, This represents the minimum operating power for the oxygenation equipment. In this embodiment, the emergency response buffer time is set to 4 hours; the system calculates the reduction in metabolic oxygen consumption after reducing the feeding amount based on the physiological tolerance function. ; in response If the dissolved oxygen concentration can be maintained above the dissolved oxygen lethal threshold, the system generates a defensive strategy of reducing feed and pre-charging the battery; conversely, in response to... Unable to maintain dissolved oxygen concentration above the lethal threshold, the system generates a survival-first strategy that prioritizes the operation of oxygenation equipment and forcibly cuts off unnecessary loads.

[0098] This embodiment achieves dynamic switching of strategies through game theory optimization. In particular, the introduction of defensive strategies utilizes the biological principle that reducing feeding can reduce oxygen consumption. In the event of energy shortage, survival is achieved by sacrificing growth, which greatly improves the resilience of offshore aquaculture systems under extreme conditions.

[0099] Example 6:

[0100] The process of generating collaborative control instructions by the execution control module includes: parsing the dynamic resource scheduling strategy and extracting the feeding valve opening parameters and aerator frequency conversion parameters; converting the feeding valve opening parameters into execution signals for the feeding equipment and converting the aerator frequency conversion parameters into motor speed control signals for the aerator; and using feedback from the remaining feed amount in the aquaculture behavior data to perform closed-loop correction of the feeding valve opening parameters.

[0101] This embodiment details the generation and modification process of collaborative control commands; the execution control module receives the dynamic resource scheduling strategy from the game optimization module, parses it, and extracts the feeding valve opening parameters and aerator frequency conversion parameters; the feeding valve opening parameters are converted into execution signals for the feeding equipment, and the aerator frequency conversion parameters are converted into motor speed control signals for the aerator. The system uses the amount of uneaten feed from the behavioral data of farmed organisms as feedback. For the opening parameters of the feeding valve Perform closed-loop correction; specifically, in response to If the threshold is exceeded, it indicates that the fish's feeding desire is lower than expected, and the control module immediately corrects the situation using a PID algorithm. This reduces the need for subsequent feeding.

[0102] To verify the technical effectiveness of this system, the applicant conducted verification tests at a deep-sea cage base in a certain sea area. Test example: Two adjacent cages were selected; Group A used the system of this invention, while Group B used traditional timed control. During a 30-day typhoon season test, Group A triggered defensive strategies three times in advance by predicting energy shortages. Experimental data showed that Group A's average dissolved oxygen fluctuation variance was 45% lower than Group B's, and the average cortisol level of the fish population was reduced by 32%. Furthermore, during periods of continuous rain, Group A did not experience any fish kills, while Group B showed slight surface surfacing. Ultimately, Group A's feed conversion ratio decreased by 0.15, and overall energy consumption decreased by 18%.

[0103] The closed-loop mechanism constructed in this embodiment ensures the accuracy of strategy implementation. In particular, it uses the remaining feed amount for real-time correction, effectively avoiding the need to continue feeding when fish are not eating due to illness or stress. This saves on expensive feed costs and prevents the water quality from deteriorating further due to the decay of uneaten feed.

[0104] Example 7:

[0105] It also includes: a resilience index assessment module, used to calculate the normalized ratio between the cumulative stress value and the energy supply capacity represented by the available energy envelope, as the aquaculture ecological resilience index; the multi-objective game optimization module is also configured to: determine whether the aquaculture ecological resilience index is lower than the preset safety value; if so, force a switch to risk avoidance mode, in which the dynamic resource scheduling strategy is based solely on the minimum dissolved oxygen requirement to maintain biological survival.

[0106] This embodiment introduces a resilience index assessment module, which serves as the highest priority criterion for system switching operating modes. It is used to calculate the normalized ratio between the energy supply capacity represented by the available energy envelope and the cumulative stress value, serving as the aquaculture ecological resilience index. The calculation formula is as follows:

[0107]

[0108] in, It is derived from ratio calculations and its physical meaning is the aquaculture ecological resilience index; Originating from the prediction module, its physical meaning is the total amount of available energy within the predicted time window; Derived from a basal metabolic model, its specific calculations are based on the principles of bioenergetics:

[0109]

[0110] in, The total mass of cultured organisms, This is the standard basal metabolic rate, expressed in units of . , For temperature sensitivity coefficient, For real-time water temperature, For reference temperature, In physical terms, the time window is the minimum energy required to maintain basic survival.

[0111] Derived from biochemical analysis, its physical meaning is the cumulative stress value at the current moment. Based on the dimensionless definition in Example 4, this value is a dimensionless number, ensuring... It is a dimensionless comprehensive index; Derived from mathematical processing, its physical meaning is a small constant to prevent the denominator from being zero; the multi-objective game optimization module monitors this index in real time; it responds to the aquaculture ecological resilience index. Below the preset safety value Preset safety value The setting is based on the system's safety margin. In this embodiment, the value is 1.2, which means that the energy supply capacity must be at least 1.2 times the demand for biological stress consumption in order to be considered safe.

[0112] This indicates that the system is in a vulnerable state with low energy and high pressure, and the system is forced to switch to a risk-avoidance mode. In this mode, the dynamic resource scheduling strategy takes the minimum dissolved oxygen requirement to maintain the survival of organisms as the sole constraint objective, ignores all growth benefits, and prohibits any feeding actions.

[0113] The resilience index proposed in this embodiment is a dimensionless comprehensive index that cleverly combines the energy supply capacity of the physical layer with the physiological health status of the biological layer. By monitoring this index, the system can enter a dormant state in advance to save itself before disasters such as typhoons occur, thereby minimizing the risk of flooding.

[0114] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A smart aquaculture management platform driven by big data, characterized in that, include: The multi-dimensional data sensing module is used to acquire energy supply status data, water environment physicochemical data, and aquatic organism behavior data of the target aquaculture area. The energy supply and demand forecasting module is used to generate an available energy envelope that represents the upper limit sequence of available energy within a future preset time window, based on the energy supply status data and preset micro-meteorological forecast information. The biochemical response analysis module is used to determine the dissolved oxygen decay recovery curve under different oxygenation powers based on the physicochemical data of the water environment and the preset water biochemical inertia model. Based on the dissolved oxygen decay recovery curve and the behavior data of the cultured organisms, it calculates the physiological tolerance function that characterizes the survival probability of the cultured organisms as environmental parameters change and the stress accumulation value that characterizes the risk over time. The multi-objective game optimization module is used to construct a game model with the objectives of maximizing biological growth benefits and minimizing mortality risk costs based on the available energy envelope, the physiological tolerance function, and the stress accumulation value, and to calculate a dynamic resource scheduling strategy based on the game model. The execution control module is used to generate coordinated control commands for the feeding equipment and the oxygenation equipment in response to the dynamic resource scheduling strategy. The specific expression of the biochemical inertial model of water bodies is as follows: ； in, Dissolved oxygen concentration, This represents the saturated dissolved oxygen concentration. This represents the initial dissolved oxygen concentration. The mass transfer coefficient is . This refers to respiratory oxygen consumption rate.

2. The smart aquaculture management platform based on big data as described in claim 1, characterized in that, The process by which the multidimensional data perception module acquires data includes: Real-time power of photovoltaic and wind turbines, remaining battery power and battery health are collected as energy supply status data; Underwater temperature, salinity, depth, dissolved oxygen concentration, pH value, and hydrodynamic flow velocity are collected as the physicochemical data of the water body environment. The swimming speed and distribution density of the organisms, identified by machine vision, and the feeding spectrum data, collected by acoustic sensors, are used as the behavioral data of the cultured organisms.

3. The smart aquaculture management platform based on big data as described in claim 1, characterized in that, The process by which the energy supply and demand forecasting module generates the available energy envelope includes: By analyzing the correlation between historical meteorological data and historical power generation data, an energy conversion efficiency model can be established. The preset micro-meteorological forecast information is input into the energy conversion efficiency model to predict the theoretical power generation within the preset time window in the future. By combining the remaining battery power in the energy supply status data, the upper limit of the energy budget for each time node within the future preset time window is calculated to form the available energy envelope.

4. The smart aquaculture management platform based on big data as described in claim 1, characterized in that, The process by which the biochemical response analysis module determines the physiological tolerance function and cumulative stress value includes: Analyze the time lag response characteristics of dissolved oxygen concentration in the physicochemical data of the water body under different oxygenation interventions, and generate the dissolved oxygen decay recovery curve; Establish a nonlinear mapping relationship between multidimensional environmental parameters and biological tolerance indicators; The dissolved oxygen decay recovery curve and the behavior data of the cultured organisms are input into the nonlinear mapping relationship to calculate the instantaneous mortality probability of the cultured organisms under the current environmental combination. The physiological tolerance function is constructed in this way, and the instantaneous mortality probability is integrated over time to obtain the stress cumulative value.

5. The smart aquaculture management platform based on big data as described in claim 1, characterized in that, The process by which the multi-objective game optimization module calculates the dynamic resource scheduling strategy includes: Define an objective function, which includes a biological growth benefit item, a mortality risk cost item, and an energy consumption cost item; Set constraints, including a battery charge safety threshold and a dissolved oxygen lethal threshold; Under the premise of satisfying the above constraints, find the global optimal solution that makes the objective function reach the extreme value within the range defined by the available energy envelope; The combination of the feeding amount and oxygenation power corresponding to the global optimal solution is determined as the dynamic resource scheduling strategy.

6. The smart aquaculture management platform based on big data as described in claim 5, characterized in that, The multi-objective game optimization module executes the following logic during the solution process: When the available energy envelope indicates that the future energy supply is lower than the preset shortage threshold, the reduction in metabolic oxygen consumption after reducing the feeding amount is calculated based on the physiological tolerance function. If the reduction in metabolic oxygen consumption can maintain the dissolved oxygen concentration above the dissolved oxygen lethal threshold, a defensive strategy of reducing feeding and pre-charging the battery is generated. If the reduction in metabolic oxygen consumption cannot maintain a dissolved oxygen concentration above the dissolved oxygen lethal threshold, a survival-first strategy is generated to prioritize the operation of the aerator.

7. The smart aquaculture management platform based on big data as described in claim 1, characterized in that, The process by which the execution control module generates collaborative control instructions includes: The dynamic resource scheduling strategy is analyzed to extract the feeding valve opening parameters and the aerator frequency conversion parameters; The opening parameters of the feeding valve are converted into execution signals for the feeding equipment, and the frequency conversion parameters of the aerator are converted into motor speed control signals for the aerator. The remaining feed amount in the behavioral data of the farmed organisms is used as feedback to perform closed-loop correction on the opening parameter of the feeding valve.

8. The smart aquaculture management platform based on big data as described in claim 1, characterized in that, Also includes: The resilience index assessment module is used to calculate the normalized ratio between the cumulative stress value and the energy supply capacity represented by the available energy envelope, as the aquaculture ecological resilience index. The multi-objective game optimization module is further configured to: determine whether the aquaculture ecological resilience index is lower than a preset safety value; if so, force a switch to risk avoidance mode, wherein the dynamic resource scheduling strategy in the risk avoidance mode takes the minimum dissolved oxygen requirement for maintaining biological survival as the sole constraint objective.

Images