高级检索
王鹏飞, 李建平, 丁瑞强, 黄荣辉. Lorenz系统误差方程的吸引子特性研究[J]. 气候与环境研究, 2012, 17(5): 574-582. DOI: 10.3878/j.issn.1006-9585.2012.11037
引用本文: 王鹏飞, 李建平, 丁瑞强, 黄荣辉. Lorenz系统误差方程的吸引子特性研究[J]. 气候与环境研究, 2012, 17(5): 574-582. DOI: 10.3878/j.issn.1006-9585.2012.11037
WANG Pengfei, LI Jianping, DING Ruiqiang, HUANG Ronghui. Studies of the Attractor Property of Error Equations for the Lorenz System[J]. Climatic and Environmental Research, 2012, 17(5): 574-582. DOI: 10.3878/j.issn.1006-9585.2012.11037
Citation: WANG Pengfei, LI Jianping, DING Ruiqiang, HUANG Ronghui. Studies of the Attractor Property of Error Equations for the Lorenz System[J]. Climatic and Environmental Research, 2012, 17(5): 574-582. DOI: 10.3878/j.issn.1006-9585.2012.11037

Lorenz系统误差方程的吸引子特性研究

Studies of the Attractor Property of Error Equations for the Lorenz System

  • 摘要: 将Lorenz方程及其导出的误差方程作为联立方程(即全误差方程)来研究误差的性质,结果表明联立方程可以变换为一个特殊的算子方程,误差轨线将收敛于一个有限的区域;此外联立方程对应的流的散度为负值,因此其在相空间中的体积不断收缩,最终趋向一个低纬曲面;联立方程的这两个性质使得Lorenz系统中初始误差不会无限放大,而是趋于一个吸引子。误差在吸引子上的概率分布是确定的,因此平均的绝对误差趋于常数,这个结果可以用来解释小初始误差经过一段时间的发展之后,趋向饱和的现象。利用稳定性分析方法研究了误差吸引中心的位置和个数,并使用数值试验进行了验证,结果显示误差吸引子的结构与解的吸引子位置、数量和结构均有不同。最后本研究将针对Lorenz方程的误差联立方程方法拓展到一般的常微分动力系统,展示了对一般误差方程的特征矩阵进行分析,研究其特征行列式性质的方法,得到了一般误差系统中稳定点和平衡态性质与原动力系统的稳定点和平衡态性质的关系,这些结果对于认识误差系统长期的动力学行为和性质是有意义的。

     

    Abstract: The simultaneous equation from the Lorenz model and its error equation without linear approximation were obtained in this study. Theoretical analysis revealed that the simultaneous equation can be transformed to a special operator equation with a property such that all the orbits obtained from the equation will converge to a finite volume. In addition, the equation demonstrates that the divergence of the flow has a negative value, which implies that the volume of the orbit in phase space approaches zero; therefore, all points will be attracted to a low-level dimensional curve surface. These two properties of the simultaneous equation make the evolution of error for the Lorenz system approach to an attractor. The mean absolute error is constant due to the certainty of probability density distribution of the error attractor; this property can be used to understand the phenomena of a small error in the Lorenz system saturating after a lengthy integration. The location and the number of attractor’s centers are obtained from the stability analysis method, and the results are validated by numerical experiments. The figure of error attractor shows that it differs from the solution’s attractor in location, number, and structure. Moreover, the authors demonstrated a method of extending this error analysis for the Lorenz equation to general ordinary differential equations (ODEs). The property of error equations for ODEs has been obtained, and the relationships between the locations of stable points and stability properties for error systems and ordinary systems are reported.

     

/

返回文章
返回