Abstract Based on the monthly flow time series data of main hydrological stations in the middle Huaihe River from 1985 to 2015, the river stability of Zhengyangguan–Bengbu and Bengbu–Fushan reaches is scientifically judged by nonequilibrium thermodynamics. The results show that the river pattern is in a stable state, and there is no transformation possible in a short time; the evolution of riverbed tends to be stable, and the influence of riverbed boundary conditions is greater than that of incoming water and sediment conditions. Based on fully understanding the stability of the middle Huaihe River, to improve the intelligence of hydrological forecasting, the chaotic characteristics of monthly flow time series are identified, then the artificial neural network forecasting model is optimized by chaos theory and optimization algorithms. Our findings suggest that the nonequilibrium thermodynamic analysis methods of river stability can provide a new idea for the study of river characteristics, and the forecasting model combined with chaos theory and optimization methods provide an effective technical means to improve the hydrological forecasting accuracy of the middle Huaihe River.

中文翻译：

---

淮河中游水文混沌预测模型稳定性分析与优化

摘要 基于1985—2015年淮河中游主要水文站月流量时间序列数据，利用非平衡热力学对正阳关—蚌埠和蚌埠—浮山段河流稳定性进行科学判断。结果表明，河流格局处于稳定状态，短时间内不可能发生转变；河床演化趋于稳定，河床边界条件的影响大于来水和泥沙条件的影响。在充分了解淮河中游稳定性的基础上，为提高水文预报的智能化水平，识别月流量时间序列的混沌特征，利用混沌理论和优化算法对人工神经网络预测模型进行优化。