DATA_BLENDING_TRANSFER

The management of the production and transfer, by a 187 km pipeline, of phosphoric merchantable ores (MO) in a pull-flow mode is a formidable stake for the world's largest site of this kind, belonging to the Morocco OCP. MOs are obtained by washing extracted Source Ores (SOs) in parallel lines, using alternative routings that modify their weight and composition, before being mixed in tanks to get the required quality defined by a set of composition constraints. The blending problem is original as it considers a two-step production process of washing before blending, which is more binding than the blending before washing. The pipe transfer program defines batches to transport and their schedule, considering downstream demands (a roughly stable demand from chemistry and irregular export demand), preventive maintenance of pipe or washing lines and prevention of defusing and saturation in the tanks of the delivery pipe station. We propose a daily used DSS that encapsulates two models to meet the expedition program at least cost, considering all technical, availability, capacity, and quality constraints. The production program meets this expedition program at least cost by determining the SO and routing for each line assigned to an order.依托一条187公里长的管道，以牵引式流动模式开展商品磷酸矿石（MO）的生产与转运管理，对于隶属于摩洛哥OCP集团的全球同类规模最大的磷酸盐生产基地而言，是一项极具挑战性的核心任务。商品磷酸矿石（MO）由采出的原矿（Source Ores，SO）经并行洗选生产线加工得到：通过可调整物料重量与成分的可选工艺流程，最终在储罐中混合以满足一系列成分约束所定义的质量要求。本次研究的配矿优化问题具有独特性，其考虑了「先洗选、后混合」的两步生产流程，该流程的约束条件严于「先混合、后洗选」的工艺模式。管道转运调度计划需确定待运输的物料批次及其调度安排，同时需兼顾下游需求（化工行业需求相对平稳，出口需求则波动无规）、管道或洗选生产线的预防性维护，以及防止输送管道站储罐出现放空与满罐饱和工况。我们提出了一套每日可投入使用的决策支持系统（Decision Support System，DSS），该系统集成了两个优化模型，可在兼顾所有技术、设备可用性、产能与质量约束的前提下，以最低成本满足发运计划要求。生产调度计划通过为分配至各订单的生产线确定原矿选型与工艺流程路线，以最低成本达成前述发运计划目标。