基于α阶神经网络逆系统孵化过程解耦控制  

Decoupling Control of Poultry Hatching Process Based on α-order Neural Network Inverse System

作　　者：周国雄[1] 李俊超[1] 陈爱斌[1] 

机构地区：[1]中南林业科技大学计算机与信息工程学院,湖南长沙410004

出　　处：《系统仿真学报》2016年第7期1661-1666,1672,共7页Journal of System Simulation

基　　金：国家级大学生创新创业训练计划(601000267);中南林业科技大学青年科学研究基金自主(QJ2012012B)

摘　　要：针对家禽孵化过程是一个复杂的生物过程,温度、湿度和氧气浓度之间耦合严重,常规的控制方法难以保证3个参数之间的稳定,提出一种基于α阶神经网络逆系统家禽孵化过程解耦控制方法。将3个参数进行耦合度分析分组,采用α阶神经网络逆系统控制耦合度较强的2个参数,另一个则通过模糊控制器控制,从而保证3个参数的稳定。实验结果表明了该方法的有效性。The poultry hatching process is a complex biological process, in which coupling between temperature, humidity and oxygen concentration is serious, and conventional control methods can not guarantee the stability of three parameters. An a-order inverse system decoupling control method for poultry hatching process was proposed. The three factors were grouped by coupling degree analysis, using a-order neural network inverse system to control the two parameters whose coupling degree is stronger, another parameter would be controlled by fuzzy controller, thus ensuring the stability of the three parameters. The experimental results show the effectiveness of this method.

关 键 词：孵化过程 逆系统 模糊控制 耦合度 

分 类 号：TP299[自动化与计算机技术—检测技术与自动化装置]