非线性和时变参数时间序列模型及其在水质分析中的应用

时间序列分析是根据观测值建立数学模型，研究数据的内在规律，现有文献主要是介绍线性时间序列模型。研究表明化学分析数据很多涉及非线性时间序列，且具有时变参数特性。本文在研究线性、非线性时间序列模型基础上，提出一种具有时变参数特性的非线性时间序列模型。该模型用于加酸调ｐＨ的循环冷却水系统进行预报，可使ｐＨ值极差降低３～６倍，对保证水质稳定具有重要意义。

Non-linear/Time Varying Parameter Time Series Models and Their Applications in Water Quality Analysis

The modeling approach of time series analysis is established based on the obser-vation values with the aim of investigating the internal regularities of the data. However, present references mainly deal with linear time series models. Experimental results show that the data of chemical analysis are usually concerned with non-linear data system. In this paper, on the basis of research into linear/non-linear time series models, a non-linear time series model with time varying parameters is proposed. When applied this model to forecast-ing control of water quality of cyclic cooling water system, the upper and lower pH specifica-tions limit may be decreased to 1/3~1/6 of that before using forecasting control. This is of importance to the water treatment technology ensuring stable water quality.