中文摘要：

利用替代数据法检验了摇摆条件下自然循环系统不规则复合型脉动的混沌特性,并在此基础上进行混沌预测.关联维数、最大Lyapunov指数等几何不变量计算结果表明不规则复合型脉动具有混沌特性,但是由于计算结果受实验时间序列长度的限制和噪声的影响,可能会出现错误的判断结果.为了避免出现误判,在提取流量脉动的非线性特征的同时,需要用替代数据法进一步检验混沌特性是否来自于确定性的非线性系统.本文用迭代的幅度调节Fourier算法进行混沌检验,在此基础上用加权一阶局域法进行混沌脉动的预测.计算结果表明：不规则复合型脉动是来自于确定性系统的混沌脉动,加权一阶局域法对流量脉动进行混沌预测效果较好,并提出动态预测方法.

英文摘要：

Chaos identification and forecasting of the irregular complex flow oscillations in a two-phase natural circulation system under the rolling motion are performed. The irregular complex flow oscillation has chaotic characteristics by calculating the geometric invariants such as the correlation dimension, Kolmogorov entropy and the largest Lyapunov exponent. But the reliability of calculation result is liable to be influenced by data length and the interference of measurement noise, false judgment results may exist in the direct method. To avoid misjudgment for chaos flow oscillation, both the geometric invariants and chaos identification need to be calculated by surrogate-data method. The chaos is identified by the iterated-amplitude adjusted Fourier transform method. Chaotic forecasting for the irregular complex flow oscillation is carried out by adding weight one-rank local region method. By surrogate-data method, we can confirm that the irregular complex flow oscillation is chaotic oscillation from the deterministic system. Comparisons between the prediction results and experimental data indicate that the chaos forecasting based on adding weight one-rank local region method is an effective way for two-phase natural circulation flow instabilities, and a way of dynamical forecast to monitor flow oscillation is presented.