基于改进自适应广义预测控制的天然气分输站压力控制

Pressure Control in Natural Gas Distribution Station Based on an Improved Adaptive Generalized Predictive Control Method

  • 摘要: 针对天然气长输管网分输站用气量大范围波动以及调节阀死区较大引起的压力控制不平稳问题,常规的机理建模难以建立准确的控制模型,控制参数难以整定.本文采用改进自适应广义预测控制进行分输压力控制,先对系统模型进行在线辨识,然后由辨识模型的阶跃响应系数计算G矩阵和预测向量,避免G矩阵不收敛且无需进行Diophantine方程计算;同时针对阀门死区的问题提出了改善方法,并在实际应用中取得较好效果,说明了该方法的有效性.

     

    Abstract: Aiming at the problem for long-distance gas pipeline distribution stations, pressure control is not smooth, which caused by large-scale flow fluctuations and large dead-zones in the valve. The exact mechanism control models are hard to establish by the conventional method, and tuning the control parameter is difficult. An improved adaptive generalized predictive control method is adopted. The system model is identified first. The G matrix and the predicted vector of step response coefficients of the system model are calculated without solving the Diophantine equation. For the dead-zone problem, an improved approach is proposed which yielded better results in practical application, thereby proving the viability of these methods.

     

/

返回文章
返回