Design And Implementation Of Recipe Management Component Based On Dynamic Scheduling

Batch production is widely used in food, pharmaceutical, and other industries to produce small batch size, multi-product and high value products. The most important characteristic of batch production is that it could adjust production flexibly according to different recipes to meet the market needs by using same equipments. However, the flexibility of batch production relies largely on the efficiency of production scheduling, recipe management, and recipe switch.Recently, the research of scheduling in chemical batch production is limited and doesn't have common methods of automatically scheduling. The existing methodology is based on building mathematical model in fixed and static condition, which can not solve the imprecise data in batch production system, thus decrease the flexibility of batch production enterprise.Regard recipe management and recipe switch, there are mainly two solutions at present:The first one is to leverage the standard control and related module of DCS, once production plan changed, by re-configuration and re-program the recipe procedure to achieve the respective recipe switch. This method need to configure huge amounts of operation point, meanwhile demand to program huge amounts of recipe procedure to link the related variables, which not only increase the manual work of engineering, but also decrease the flexibility of production line. The second one is to purchase professional recipe management software. The application of this non-standard recipe management software, strongly limited to special industry meanwhile only available for fixed scheduling process, could not automatically solve the dynamic scheduling issue, which encumbered the progress of batch production enterprise.Thus, to achieve dynamic scheduling and automatic recipe management, based on the analysis of flexibility needs of batch production and research of existing recipe management model, this thesis proposed a novel recipe management component which tailored for dynamic scheduling.In the thesis, novel dynamic scheduling based model is designed to integrate the information of batch production and respective recipes, which could support the dynamic scheduling. The dynamic scheduling algorithm thus could leverage these data to optimize target goal. Meanwhile, the model could generate recipe procedure automatically to speed up the process of recipe switch. To achieve automatically dynamic scheduling, a tailored advanced mixed-integer linear programming (MILP) algorithm is designed which builds a mixed-integer linear programming model for multi-purpose batch plants and its dynamic scheduling. The model using continuous time representation and the assignment of tasks and units are handled through two set of binary variables. According to the characteristic of batch process and model, this thesis proposed advanced methods to minimize the model scale, for example, circumscribing dynamic scheduling task sets etc., for speeding up the mathematically calculation process. Simulation results indicated that the speed of novel algorithm is relatively fast to solve the dynamic scheduling problem, especially for dynamic scheduling which have repetitious batch.The novel recipe management model adopted object oriented programming method, used Visual Basic as development tool to achieve the function of novel recipe management component, meanwhile provide complete interface which could interactive with other software easily.In the final part of the paper, specific methods of cooperating novel recipe management component with DeltaV control system are presented, which could integrate the dynamic scheduling, batch control and operation. Thus, they better satisfy the batch production enterprise, release configuration manual work, and increase the flexibility. They could better serve batch production enterprise and have promising market future.