摘要
湿空气中温度和湿度的关系通常通过水蒸气性质表进行查取,通过温度和饱和水蒸气分压力的关系计算相对湿度值,但其为离散的非线性关系,不便于实时控制。提出了一种简单、有效的广义回归神经网络(GRNN)用于中央空调控制中的湿度调节,GRNN可以对离散的非线性关系进行拟和,不同于数值分析中的插值和拟和,也不同于常用的BP网络,GRNN易于实现,仅需要一个参数,结构简单,便于编程,可以在较少数据中较好地工作。
Generally, the relationship between temperature and humidity of air can be obtained by viewing the chat of the character of vapor. And then, the relative humidity can be calculated in term of the relationship between temperature and the pressure of vapor. But it's difficult to be used in realtime control for the discreteness and nonlinear. A simple yet effective general regression neural network(GRNN) paradigm is suggested for adjusting humidity in HVAC control system. Unlike interpolation or curvefitting and BP paradigm, the proposed GRNN is simple to implement, requiring only one parameter and can work well with sparse data. And it's easy to realize in programming for its simple construction.
出处
《控制工程》
CSCD
2003年第2期162-164,共3页
Control Engineering of China
