空调系统在线辨识算法的研究
Online identification algorithm for air conditioning systems
摘要：
空调系统可以看成是具有时变特性的一阶惯性加延迟的对象，针对递推最小二乘法只能辨识对象的过程参数而无法辨识延迟时间的问题，提出了一种基于最速下降法的在线辨识算法，可以同时辨识出时变情况下空调系统对象的过程参数和延迟时间。以空气焓差法试验台测试室的温度控制系统作为仿真对象对该算法进行了验证，结果表明，该算法具有较高的辨识精度，而且性能稳定。
Abstract：
The air conditioning system can be described as a first-order-plus-delay-time model with time variation. By the recursive least squares (RLS) algorithm, the process parameters of the air conditioning system can be estimated but the delay-time can not be identified. Puts forward an online identification algorithm based on the gradient method which can estimate time-varying process parameters and delay-time simultaneously. Tests the identification algorithm in the temperature control system of the testing room of an air enthalpy-difference test bed. The results show that the algorithm has higher identification accuracy and stable performance.
Keywords:airconditioningsystem,onlineidentification,processparameter,delaytime,gradientmethod