论文部分内容阅读
将多变量模型预测控制技术应用于催化裂化汽油稳定塔的控制，有助于提高稳定塔的控制品质。 Ｍ Ｐ Ｃ控制器控制的变量包括汽油雷氏蒸气压、液化气中 Ｃ５ 含量和塔顶温度３ 个被控变量，塔底温度和冷回流量 ２ 调节变量，以及包括塔操作压力、回流温度和进料流量等３ 个干扰变量。应用机理模型进行汽油雷氏蒸气压和液化气中 Ｃ５ 含量的在线计算以满足在操作条件大范围变化的控制需要。为建立 Ｍ Ｐ Ｃ 控制器通过现场阶跃实验建立了稳定塔的传递函数动态模型矩阵。该控制器已在工业装置上投用，且通过鉴定。
The application of multivariable model predictive control technology to the control of FCC gasoline stabilizers helps to improve the control quality of the stabilizers. M P C Controller-controlled variables include gasoline Ray’s vapor pressure, C5 content in liquefied gas and overhead temperature, three controlled variables, bottom temperature and cool-back flow 2 tuning variables, and including tower operating pressure, reflux temperature and Feed flow and other three disturbance variables. The application of a mechanistic model for on-line calculation of gasoline Reis vapor pressure and C5 content in liquefied gas to meet a wide range of control needs in operating conditions. In order to establish M P C controller, the dynamic model matrix of transfer function of stabilized column was established by field step experiment. The controller has been put into use on industrial installations and passed certification.