A method for predicting dissolved oxygen in aquaculture water based on DBO-STN-Informer

By using the DBO-STN-Informer model, the spatial transformation network STN and the dung beetle algorithm DBO optimization attention mechanism are employed to address the shortcomings of existing dissolved oxygen prediction technologies, achieving high-precision and efficient future dissolved oxygen prediction and supporting water quality management in scientific aquaculture.

CN120412814BActive Publication Date: 2025-11-18ZHONGKAI UNIV OF AGRI & ENG
2 Cites 0 Cited by

Patent Information

Application Number
CN202510455665.X
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2025-04-11
Publication Date
2025-11-18
Estimated Expiration
2045-04-11

AI Technical Summary

Technical Problem

Existing technologies are insufficient to efficiently predict future trends in dissolved oxygen in aquaculture water, resulting in a lack of forward-looking guidance for oxygenation equipment control and feed feeding decisions. Furthermore, traditional methods are not capable of fitting nonlinear and non-stationary time series.

Method used

A prediction method based on DBO-STN-Informer is adopted. The input time series is transformed by the Spatial Transformation Network (STN), and the attention mechanism parameters are optimized by combining the Dung Beetle Algorithm (DBO). The STN-Informer model is constructed to extract the dissolved oxygen time series relationship features and make predictions for future periods.

Benefits of technology

It improves the accuracy and efficiency of dissolved oxygen prediction, enhances the model's robustness to noise and local fluctuations, and can more accurately predict future changes in dissolved oxygen concentration, supporting water quality management in scientific aquaculture.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

A kind of aquaculture water dissolved oxygen prediction method based on DBO-STN-Informer, comprising the following steps: S1, dissolved oxygen sensor is used to collect dissolved oxygen data in aquaculture water in real time on-line, and after pretreatment, data is divided into training set and verification set according to time sequence;S2, on the basis of Informer coding mode, introduce space transformation network STN to transform input time series, build STN-Informer prediction model, the space transformation and feature calibration of input time series data are carried out by STN, to help the encoder understand deep dissolved oxygen sequence;The attention mechanism parameters in STN-Informer prediction model are optimized using Dung Beetle Algorithm DBO;Finally, the model is tested and scored and optimized using verification set;S3 based on the model output after training and optimization dissolved oxygen concentration prediction value in future period.The present application can predict the dissolved oxygen concentration in future period according to prior data, and has very high prediction accuracy, precision and efficiency.
Need to check novelty before this filing date? Find Prior Art