给水管网余氯的混沌特性与预测

基于混沌理论，研究了给水管网余氯时间序列的混沌特性．并根据混沌理论的最大Lyapunov指数对其进行了预测的研究．采用混沌理论的特征参数一一关联维数和最大Lyapunov指数分析了余氯时间序列的混沌特征．通过实例分析证明给水管网的余氯时间序列存在混沌特征．基于最大Lyapunov指数提出了给水管网余氯的预测模型，实例研究结果表明不需要管网其他监测点或其他水质监测数据的辅助，该方法能够进行连续多步预测，并且其在最大可预测时间尺度内的预测精度较高而且比较稳定，而在最大可预测时间尺度外的预测精度下降很快并且预测稳定性较差．

Chaotic features and forecasting of residual chlorine in water distribution networks

According to chaotic theory, the chaotic features of time series of residual chlorine in water distribution networks are studied, and an approach is presented for short-term forecasting of time series of residual chlorine based on the largest Lyapunov exponent of the chaos theory. The chaotic characters of the time series are ex- amined by means of the largest Lyapunov exponent and correlation dimension. It is proved that the time series exist chaotic phenomena in case study. Then by use of the largest Lyapunov exponent a short-term forecasting model is developed, which is tested in a case. The result indicates that the model can make multi-step forecasting without the help of other monitor point or other data of water quality monitoring and, furthermore, the predicted precision of the model is higher and steadier within maximum predictable time scale than that beyond it.