Construction And Validation Of An Irrigation Model Based On Comprehensive Evaluation Of Watermelon Growth

The total amount of water resources in China is seriously insufficient,with a large amount of water used for agricultural irrigation and extensive irrigation management methods.Irrigation mainly relies on manual experience or a single parameter threshold,resulting in low water efficiency and resource waste.In order to promote sustainable agricultural development and meet the requirements of precise regulation of modern water and fertilizer agronomy,it is urgent to study crop water demand models that adapt to China’s agricultural development to guide agricultural irrigation and improve water efficiency.This study determined the optimal irrigation and nitrogen application levels based on multi-objective comprehensive evaluation method,based on the data of three consecutive spring substrate cultivation watermelon water nitrogen coupling experiments;A backpropagation(BP)neural network was used to analyze greenhouse environmental parameters and the optimal daily irrigation amount,and a watermelon full growth period irrigation model based on environmental decision-making was established;And the model was tested and applied for validation.The main research findings are as follows:(1)Determination of the optimal water and nitrogen supply level for watermelon based on comprehensive evaluation.Based on the data analysis of three consecutive spring trials from 2019 to 2021,a comprehensive evaluation system of water and nitrogen application was established,including seven indicators in three categories of watermelon growth,yield and quality.The analytic hierarchy process(AHP)was used to determine the subjective weights of factors and sub factors,and the entropy weight method was used to give objective weights.The combined weights of each indicator were determined based on game theory,of which the yield was the largest in the three trial seasons,They are 0.396,0.386,and 0.387,respectively.Based on TOPSIS,a comprehensive evaluation was conducted on various water and nitrogen treatments,with an irrigation rate of 120% ETc as the optimal water supply level.Under these conditions,watermelon has the best comprehensive growth.(2)Construction of Watermelon Irrigation Model Based on BP Neural Network.A water irrigation model for the entire growth period of watermelon was established using BP neural network,with three parameters of daily average temperature,daily average relative humidity,and average light intensity during the light period(6:00～19:00)as inputs and the daily irrigation amount at the optimal irrigation level as output.Through experimental analysis,the prediction error was significantly reduced compared to the model with two parameters of environmental temperature and humidity as inputs,And there is no significant change in the prediction error of the indicators as they continue to increase.The irrigation model sample data consists of 270 sets,with weights and thresholds corrected.The relative error of the system is less than 0.15%,and the determination coefficients are all above 0.9.The prediction accuracy is high.(3)Implementation of a model-based automatic irrigation system.The system consists of data collection components,communication components,decision-making components,monitoring components,and irrigation components.RS485 temperature,humidity,and light sensors are used to detect the environmental parameters in the temperature chamber in realtime.A computer is used as the decision-making component to run a neural network model and output prediction results.The STM32 microcontroller is used for overall system control,and the 4G module is used for remote communication to achieve dynamic irrigation control during watermelon growth.(4)Field experiment validation of watermelon irrigation model.Apply an automatic control system based on irrigation models to greenhouse watermelon cultivation,and use conventional manual irrigation as a control to verify the application effect of the model.The results showed that the yield per hectare of watermelon based on model automatic irrigation was 4.13% higher than that of the control treatment,the total soluble sugar was 12.13% higher than that of the control group,and the vitamin C was 10.53% higher than that of the control group.The advantage is significant,and the cumulative irrigation amount saved throughout the entire growth period of watermelon is 11.03%,which can greatly reduce labor costs and achieve automatic precision irrigation.