CN-122004151-A - Intelligent cultivation method and system based on water quality dynamic balance

CN122004151ACN 122004151 ACN122004151 ACN 122004151ACN-122004151-A

Abstract

The invention discloses an intelligent aquaculture method and system based on water quality dynamic balance, belonging to the technical field of aquaculture, comprising the steps of collecting multi-source aquaculture data; the method comprises the steps of constructing a water quality dynamic association model based on multi-source aquaculture data, collecting current aquaculture data of aquaculture in real time, inputting the current aquaculture data into the water quality dynamic association model, solving a water quality parameter evolution trend and a balance imbalance risk index in a future period, generating a collaborative regulation strategy according to the balance imbalance risk index when the balance imbalance risk index is larger than a preset value, generating a control instruction according to the collaborative control strategy, and determining the execution time of the control instruction based on the water quality dynamic association model. The invention realizes prospective and accurate collaborative regulation and control of water quality, effectively avoids unbalance risk and improves the intelligent level and benefit of cultivation.

Inventors

XUE YANG
SHAN HONG
XU MINJIE
SUN HAOBO
YIN KEXIN
WANG QING

Assignees

南京市水产科学研究所(南京市水产技术推广站、南京市水生动物疫病预防控制中心)

Dates

Publication Date: 20260512
Application Date: 20260204

Claims (9)

1. An intelligent cultivation method based on water quality dynamic balance is characterized by comprising the following steps: collecting multi-source aquaculture data of aquaculture; Collecting current aquaculture data of aquaculture in real time, inputting the current aquaculture data into the water quality dynamic correlation model, and resolving a water quality parameter evolution trend and a balance imbalance risk index in a future period; When the imbalance risk index is greater than a preset value, generating a cooperative regulation strategy according to the imbalance risk index; and generating a control instruction according to the cooperative control strategy, and determining the execution time of the control instruction based on the water quality dynamic association model.
2. The intelligent aquaculture method based on water quality dynamic balance according to claim 1, further comprising inputting the bait casting plan into a water quality dynamic correlation model for water quality evolution simulation, and optimizing the bait casting plan when the unbalance risk index of the simulation result exceeds the standard.
3. The intelligent aquaculture method based on water quality dynamic balance according to claim 1, wherein the water quality dynamic correlation model comprises a metabolic load prediction sub-model, a multi-parameter coupling sub-model and a device response delay sub-model which are built based on the multi-source aquaculture data.
4. The intelligent aquaculture method based on water quality dynamic balance according to claim 3, wherein constructing a metabolic load predictor model based on the multi-source aquaculture data comprises: The load prediction sub-model comprises a feature extraction part, a feature splicing part and a load prediction part, wherein the feature extraction part is composed of at least two different network models, an inlet of each network model is an input port of the load prediction model and is used for extracting metabolic load prediction features from different dimensions, the feature splicing part is used for splicing the metabolic load prediction features extracted by the different network models to obtain spliced metabolic load prediction features and inputting the spliced metabolic load prediction features to the load prediction part, and the load prediction part is used for outputting a final metabolic load prediction result.
5. The intelligent aquaculture method based on water quality dynamic balance according to claim 3, wherein constructing a multiparameter coupled submodel based on the multisource aquaculture data comprises: Calibrating dynamic influence relation among water quality parameters according to historical culture data, constructing an adjacent matrix A, wherein element A ij represents influence intensity coefficient of an ith parameter on a jth parameter, and when A ij is not equal to 0, a directed edge is established between corresponding nodes to form a water quality parameter association graph G= (V, E); For each node V i epsilon V, fusing the current monitoring value, the change rate, the historical time sequence statistical characteristics and the association degree with the metabolic load predictor model to construct a multidimensional node initial characteristic vector ; The graph attention network is adopted for multi-layer message transmission, and the node characteristic updating formula of the layer 1 is as follows: Wherein N (i) is the neighbor set of node v i , For the weight coefficient dynamically calculated based on the attention mechanism, W l is a trainable parameter matrix, and sigma is an activation function; After L-layer aggregation, node features are aggregated into deep coupling perception feature vectors through a reading layer: Wherein, the For the mean value of the aggregate, For attention weighted aggregation, W a is the attention weight matrix; dynamically updating the edge weight of the adjacent matrix A through a gating mechanism according to real-time data of the culture varieties, the growth stages and the fish carrying quantity; The deep coupling perception feature vector H is input into the full connection layer, and the multi-parameter coupling state index CCS is output.
6. The intelligent aquaculture method based on water quality dynamic balance according to claim 2, wherein current aquaculture data is collected in real time, the current aquaculture data is input into the water quality dynamic correlation model, and the water quality parameter evolution trend and the imbalance risk index in the future period are calculated, specifically comprising: synchronously acquiring current culture data, unifying the culture data with different sampling frequencies to a preset fixed time step delta t through an interpolation or resampling method, and constructing unified input feature vectors at the current moment ; Performing multi-step iterative solution based on the metabolic load predictor model and the multi-parameter coupling submodel to obtain a future period water quality evolution trend prediction track; calculating a risk index R based on a future period water quality evolution prediction track: Wherein, the Is the predicted value of parameter i at time t, Is a preset safety threshold range, penalty is a penalty function when the predicted value exceeds the range, var (Δdo) is the variance of the change rate of dissolved oxygen in the predicted period, trend (TAN) is the accumulated upward Trend intensity of total ammonia nitrogen in the whole predicted period, and w 1 ,w 2 ,w 3 is the weight coefficient of each component.
7. The intelligent cultivation method based on water quality dynamic balance according to claim 6, wherein the method is characterized in that the future period water quality evolution trend prediction track is obtained by performing multi-step iterative solution based on a metabolic load predictor model and a multi-parameter coupling submodel, and comprises the following steps: setting a predicted total duration T pred , enabling a current resolving time t=T 0 , and initializing a future water quality parameter evolution track sequence to be empty; For the solution time t: a) Inputting a characteristic vector X t corresponding to the moment t into the metabolic load predictor model to predict and obtain the generation rate MP t of the metabolic waste in the [ t, t+delta t ] period; b) Inputting the characteristic vector X t and the generation rate MP t corresponding to the moment t into the multi-parameter coupling submodel to obtain a multi-parameter coupling state index CCS t of the moment t, and deducing the variation delta P t of each core water quality parameter in the [ t, t+delta t ] period based on the built-in dynamic relationship of the multi-parameter coupling submodel; c) According to the variation delta P t , updating to obtain a predicted water quality parameter value at the time of t+delta t, and updating the predicted water quality parameter value and predicted equipment state data to a feature vector X t+Δt ; Let t=t+Δt, repeat the above steps with the updated feature vector X t as a new input until t=t 0 +T pred , thereby obtaining a complete water quality parameter evolution prediction trajectory from T 0 to T 0 +T pred .
8. An intelligent aquaculture system based on water quality dynamic balance, which is characterized by comprising: The acquisition module is used for acquiring multi-source aquaculture data of aquaculture; The model construction module is used for constructing a water quality dynamic association model based on the multi-source culture data; the evolution module is used for collecting current aquaculture data of aquaculture in real time, inputting the current aquaculture data into the water quality dynamic association model, and resolving a water quality parameter evolution trend and a balance imbalance risk index in a future period; the judging module is used for generating a cooperative regulation strategy according to the balance disturbance risk index when the balance disturbance risk index is larger than a preset value; And the control module is used for generating a control instruction according to the cooperative control strategy and determining the execution time of the control instruction based on the water quality dynamic association model.
9. The intelligent aquaculture system according to claim 8, further comprising inputting the bait casting plan into a water quality dynamic correlation model for water quality evolution simulation, and optimizing the bait casting plan when the unbalance risk index of the simulation result exceeds the standard.

Description

Intelligent cultivation method and system based on water quality dynamic balance Technical Field The invention relates to the technical field of aquaculture, in particular to an intelligent aquaculture method and system based on water quality dynamic balance. Background Aquaculture is an important component of the agricultural economy and its sustainable development is highly dependent on the stability and health of the aquaculture water environment. The water quality is a key factor for determining success or failure of cultivation, and complex dynamic coupling relations exist among core parameters such as dissolved oxygen, ammonia nitrogen, nitrite, pH value, temperature and the like, and the complex dynamic coupling relations are influenced by multiple factors such as cultivation organism metabolism, bait feeding, microorganism activity, environmental disturbance and the like. Maintaining the dynamic balance of water quality is the core for guaranteeing healthy growth of cultured organisms, improving the utilization rate of feed and reducing the risk of diseases. At present, most of water quality management control strategies in large-scale aquaculture are independently controlled by single parameters, complex interactions among water quality parameters are ignored, for example, oxygenation can influence pH and microbial nitrification processes, and new unbalance can be caused by single regulation actions. Furthermore, there is a time delay from the detection of the device execution and the device performance is acting on the body of water and producing an effect also requires a process, such a system response delay makes it difficult to precisely match the regulation to the real-time dynamic evolution of the water quality. Furthermore, daily production operations (especially feeding) are one of the largest sources of water quality disturbance, but in existing management systems, production plans and water quality regulation are often split, and potential impacts of production behavior on water quality balance cannot be estimated and optimized prospectively. In summary, how to provide a cultivation management method and system capable of deeply fusing multi-source information, dynamically simulating correlation and evolution among water quality parameters, and performing intelligent collaboration and prospective regulation based on future risks is a problem to be solved by those skilled in the art. Disclosure of Invention In view of the above, the invention provides an intelligent cultivation method and system based on water quality dynamic balance, which can simulate, predict and actively maintain the intelligent closed-loop management system of the water quality dynamic balance of the aquaculture system, thereby remarkably improving the scientificity, safety and production efficiency of the cultivation process. In order to achieve the above purpose, the present invention adopts the following technical scheme: in one aspect, the invention provides an intelligent cultivation method based on water quality dynamic balance, which comprises the following steps: collecting multi-source aquaculture data of aquaculture; constructing a water quality dynamic association model based on the multi-source culture data; Collecting current aquaculture data of aquaculture in real time, inputting the current aquaculture data into the water quality dynamic association model, and calculating a water quality parameter evolution trend and a balance imbalance risk index in a future period; When the imbalance risk index is greater than a preset value, generating a cooperative regulation strategy according to the imbalance risk index; and generating a control instruction according to the cooperative control strategy, and determining the execution time of the control instruction based on the water quality dynamic association model. Preferably, the method further comprises the step of inputting the feeding plan into a water quality dynamic association model to perform water quality evolution simulation, and optimizing the feeding plan when the balance imbalance risk index of the simulation result exceeds the standard. Preferably, the water quality dynamic association model comprises a metabolic load prediction sub-model, a multi-parameter coupling sub-model and a device response delay sub-model which are established based on the multi-source culture data. Preferably, constructing a metabolic load predictor model based on the multi-source cultivation data comprises: The load prediction sub-model comprises a feature extraction part, a feature splicing part and a load prediction part, wherein the feature extraction part is composed of at least two different network models, an inlet of each network model is an input port of the load prediction model and is used for extracting metabolic load prediction features from different dimensions, the feature splicing part is used for splicing the metabolic load prediction features extracted by