Dissertation > Excellent graduate degree dissertation topics show

The BP Neural Network Simulation of Multi-span Greenhouse Environmental Factors under Ventilation Conditions in Summer in Central China

Author: ZhenYuCun
Tutor: YuanQiaoXia
School: Huazhong Agricultural University
Course: Facilities for horticulture
Keywords: Greenhouse Environment Temperature Humidity Light intensity CO2 content Wind velocity BP neural network
CLC: S625
Type: Master's thesis
Year: 2009
Downloads: 30
Quote: 0
Read: Download Dissertation

Abstract

Greenhouse Horticulture in mature agricultural production facilities in the world, it can provide a relatively independent system, provide a good environment for the growth of crops, and to overcome the impact of changes in the external environment on crop growth, and the anniversary of the crop production. In order to better achieve the greenhouse production, precision control of greenhouse environment, more and more people's attention. The artificial neural network has a good self-adaptability, self-organization and a strong learning, Lenovo, fault tolerance and anti-jamming capability, flexible modeling of multi-factor complex unknown coefficients. For the traditional mathematical simulation method can not solve these problems, this paper using neural network method, study design, a BP neural network to achieve environmental modeling of multi-span plastic greenhouse, provide a good prerequisite for precise control of the greenhouse environment . Through the greenhouse environmental testing, research greenhouse greenhouse environmental factors change under different conditions, provide a theoretical basis for greenhouse environment control. The results show that: the greenhouse under different working conditions, environmental factors within the greenhouse is very complicated, there is no obvious pattern, can not use a simple mathematical method to do an accurate description of the changes in the environment in the greenhouse, environmental factors The relationship between the complex and volatile, with complex interactions greenhouse mathematical modeling more difficult. Papers of Engineering and Technology, Huazhong Agricultural University greenhouse test base for the study, four factors, two level orthogonal experiment test data, outside the greenhouse environment parameters as input parameters will need to control the greenhouse environment parameters as to be predicted parameters to establish the mathematical model of BP neural network to forecast parameters. The original data preprocessing and comprehensive comparison of the neural network model established in the different network structures, under the conditions of different algorithms to determine the structure of neural network type 9-11-5 the network algorithm traingdx obtained under this condition prediction better, and can guarantee a good generalization ability of the network. And provide a written code using MATLAB neural network toolbox. In order to verify the accuracy and usefulness of the forecast system, the paper selected 80 groups of greenhouse internal and external environment of test data to the simulation model and the generalization processing, will make use of the results obtained in the prediction model and actual test data comparison, indicating that the predicted good results, traingdx good generalization ability. Thus indicating that the forecast system has a certain value in use, China regions greenhouse production guidance and reference value of the greenhouse environment control study.

Related Dissertations

CLC: > Agricultural Sciences > Gardening > Horticultural ( protected cultivation ) > Greenhouse
© 2012 www.DissertationTopic.Net Mobile