论文部分内容阅读
汽油调合具有非连续批次性的特点,并且调合头产品在罐中存在累积过程。已有汽油调合在线配方优化针对的是瞬时调合头产品而非罐中产品,无法解决罐底补偿问题,造成罐产品性质与实际目标性质存在偏差,调合生产经济效益进一步提高受到限制。为解决这一问题,文中采用具有收缩时域的模型预测全局滚动优化方式,兼顾质量过剩最小与经济效益最大化两方面要求,提出累积调合过程在线控制与优化算法。在瞬时产品配方优化层的基础上,通过增加瞬时产品目标值计算层,形成二级递阶结构。应用案例结果:证明该算法能用于累积调合过程,能兼顾罐产品质量和调合经济效益最大化,能够有效地解决罐底补偿问题。
Gasoline blending has the characteristics of non-continuous batch, and mixing head products in the tank there is a cumulative process. Gasoline Blending Online Formulation Optimization Targeted instantaneous blending head products instead of canned products, can not solve the problem of tank bottom compensation, resulting in the deviation between the properties of the can product and the actual target properties, and the further improvement of blending production economy is limited. In order to solve this problem, this paper proposes a method of online control and optimization of cumulative blending process by using the model with shrinkage time domain to predict the global rolling optimization mode, taking into account both the minimum mass surplus and the maximum economic benefit. On the basis of instantaneous product formulation optimization layer, a second-order hierarchical structure is formed by adding instantaneous product target value calculation layers. The results of application cases show that the algorithm can be used to accumulate blending process, which can balance the product quality of cans and maximize the economic benefits of blending, and can effectively solve the problem of tank bottom compensation.