Abstract: Based on inverse transient analysis method, the identification calculation was performed by minimizing the inverse least square problem representing the difference between the measured and predicted transient response in the waxy oil pipeline. And Levenberg-Marquardt method was adopted to optimize the friction factors. The hydraulic transient was modeled using MOC. The Jacobian matrix of the friction factor’s sensitivity was obtained by direct differentiation of the method of characteristic equations. By means of an example, the errors between the measured values and the calculated values of the middle point on the waxy oil pipeline were compared. The calculated values were solved by means of average, formular and identification friction factors. The results show that the identification method has good evaluation for the internal state variables. Identification basing on inverse transient method can be used to improve the accuracy of simulation and control with pipelines.

Key words: Identification ,  Friction factor , Sensitivity ,  Inverse problem ,  , Non-linear distributed parameter system ,  Simulation

摘要： 基于反问题分析方法,以最小二乘准则为终止条件,根据原油管道不稳定过程的压力实测数据,采用
Levenberg-Marquardt方法辨识含蜡原油管道站间摩阻系数。应用特征线法直接计算摩阻系数敏感度Jacobian矩
阵,并通过算例比较了采用平均摩阻系数法、公式计算摩阻系数法及辨识算法计算的原油管道中间点压力值与实测
值的偏差。结果表明,采用辨识方法求解的管道沿线摩阻系数能够对管道沿线压力流量值进行较好的估计。

关键词: 辨识 ,  摩阻系数 ,  敏感度 ,  反问题 , 非线性分布参数系统 , 仿真