论文部分内容阅读
针对炼油厂柴油调合生产这个强非线性、多输入、多输出、多扰动、大纯滞后对象 ,提出了以柴油调合倾点RBF神经网络模型为质量指标约束条件的柴油连续调合生产非线性在线最优化方法 ,提出了基于RBF神经网络预测模型、滚动优化目标函数中带有静态经济指标的柴油连续调合倾点非线性预测控制策略。仿真计算证明本文的非线性优化方法能计算柴油调合生产最优配方 ,在满足生产能力约束和产品倾点质量指标的前提下可实现组分油最优利用而获最大利润 ;预测控制策略在预测模型与实际对象模型失配时仍具有良好的跟踪性能和抗干扰性能 ,展示了其良好的工业应用前景。
Aiming at the strong non-linearity, multi-input, multi-output, multi-disturbance and big-time lag object of diesel blending production in refinery, the diesel continuous blending production based on diesel blending pour point RBF neural network is taken as the quality index constraint Linear on-line optimization method, a continuous pour point non-linear predictive control strategy based on RBF neural network prediction model and rolling optimization objective function with static economic indicators is proposed. The simulation results show that the nonlinear optimization method of this paper can calculate the optimal formulation of diesel blending production and achieve the maximum profit under the premise of meeting production capacity constraints and product pour point quality index. The prediction model and the actual object model mismatch still has good tracking performance and anti-jamming performance, showing its good prospects for industrial applications.