Intelligent live fish cultivation water quality comprehensive forecasting method

Abstract

The invention relates to an intelligent live fish cultivation water quality comprehensive forecasting method. The basic idea is that a radial base nerve network is used as pond water quality data forecasting base, a recurrence rolling type forecasting method is used as a nerve network forecasting strategy to express dynamic change characteristics of pond water quality environment, potential relation and feedback mechanism between key water quality factors and other water quality factors can be found by analyzing a systematic power model aiming at key factors, such as dissolved oxygen, ammonia nitrogen and nitrite nitrogen, of pond water quality, and different nerve network forecasting models of key water quality factors can be constructed. Values of all water quality factors and values of all key water quality factors can be obtained by building pond water quality comprehensive forecast and analyzing and calculating nerve network forecasting models of different key water quality factors. Key factors can be obtained through adjustment and comprehensive forecasting, values between key factors and key factors forecasted by the key factors can be obtained, and the final forecasting value can be obtained. Thus, water quality can be comprehensively and accurately forecasted so as to bring convenience to pond cultivation, greatly reduce cost and risk of pond cultivation and bring benefits to pond cultivation.

Description

technical field [0001] The invention relates to the field of aquaculture, in particular to an intelligent comprehensive water quality prediction method for live fish cultivation. Background technique [0002] Water quality prediction is the basis of water environment planning, evaluation and management; water quality prediction methods include water quality mathematical simulation prediction, multiple regression model, gray prediction model method and neural network model prediction; however, due to the limitations of hydrology and water quality monitoring conditions, domestic Most of the rivers have not yet established corresponding water quality models, or the results are relatively scattered, and the manifestations are relatively single. In terms of pond aquaculture water quality models, research on this aspect in China is even rarer. The influencing factors include not only the physique of the fish itself, but also the influence of many factors such as season, geographic...

AI Technical Summary

Problems solved by technology

[0004] Document 2 (Based on BP Network Aquaculture Water Ammonia Nitrogen Prediction Model and Realization, Agricultural Mechanization Research, No. 7, July 2008) on the basis of analyzing the factors affecting aquaculture water quality, it is found that the content of ammonia nitrogen in water is not only related to the pH value and temperature of aquaculture water There is a close relationship, and it is also related to the source and consumption of oxygen in the water body, the dissolution and overflow rate of oxygen. Using the good nonlinear mapping characteristics of the BP neural network, a BP network prediction model is established to predict the amount of ammonia nitrogen in the future pond culture. The disadvantage of Document 2 is that it simply expounds the relationship between ammonia nitrogen and other water quality factors without further analysis, and has obtained a deeper hidden relationship. At the same time, it also has the disadvantages of Document 1. It just fails to reflect the dynamics between water quality changes

Images