Application Research Of Artificial Neural Network In Water-saving Irrigation

At present,China is in the critical period of the transformation of traditional agriculture to modern agriculture,and more and more attention and attention are paid to the information and intelligent agriculture.At the same time,Chinese agricultural experts and scholars have made a great deal of research results in water-saving irrigation,and how to apply the results to the actual agricultural production should be the focus of the current consideration.The artificial neural network can well fit the relationship between the sample input and the output in the data storage and memory,so that the existing research results in the water-saving irrigation project can be considered as the sample data,and the artificial neural network model is established through the training of the sample data,The artificial neural network model with expert knowledge is used to replace the limited experts and help the farmers to solve the problem.For the problems that the farmers often encounter in the water-saving irrigation,such as how to find the cause of the failure in the failure of the agricultural well pump in the well irrigation area,how to select the form of the water-saving irrigation during the irrigation,and how to predict the yield when the irrigation and fertilization scheme is determined.Aiming at these problems,the collected expert knowledge and the scientific research achievements of the research group are used to form sample data,and the neural network model for different problems is formed by using the artificial neural network to train different samples,and the feasibility of such a problem-solving method is verified by an example.The main research results are as follows:(1)The problem of well pump failure at the water source in irrigation and the corresponding fault cause are analyzed.The fault problem with ambiguity is quantitatively described by the membership function value,and the cause of the fault is represented by binary number.The correspondence between the problem and the cause is taken as input.The output is used to train the neural network,and the trained neural network is tested with test samples,and the results are consistent with the expected output.(2)It is analyzed how to choose the form of water-saving irrigation according to the limiting factors.Different conditions are used to select different water-saving irrigation forms.Based on this,the raw data of the learning samples are determined,and the input data is normalized,and the output data is represented by binary numbers.The artificial neural network is used to train the processed data.The actual output after training is different from the expected output,and the values are basically consistent,which satisfies the requirements for the preferred form of water-saving irrigation.(3)The relationship between water and fertilizer-yield of winter wheat in a test station was analyzed.The raw data of water and fertilizer and yield were normalized.As the input and output of artificial neural network,the trained neural network model predicted the test samples and predicted the results.The mean square error between predicted output and actual production obtained after denormalization is 3.55%,which satisfies the accuracy requirement for production forecast.