Research On Integrated Modeling Of Planning And Scheduling And Uncertainty Algorithm In Petrochemical Industry

With the increasing demand for the energy consumption and the awareness of the environment protection, the tradition petrochemical enterprises should enhance their operational efficiency and handle the complex risks in supply and demand markets. Thus, the planning and scheduling method, integration method and uncertainty analysis have great significance in the real world application problem. Besides, research on the related solving algorithm and solution framework for the large scale enterprise-wide problem and various uncertainties is the key step. This thesis first presents a review of the researches focus on the problem of planning, scheduling, integration and uncertainty optimization. Based on the background of crude refining and ethylene production, this work studies the optimization problem from upstream to downstream and from the specific unit area to the entire plant. Furthermore, various uncertainties in the scheduling and planning problem are analyzed and incorporated. Different methods of reactive scheduling and preventive scheduling approaches are developed to deal with these uncertainties.The details are listed as follows:(1) Two rescheduling approaches for the crude oil operations are proposed for the discrete-time model and continuous-time model. Abnormal events and uncertain resources are considered and analyzed in this framework to improve the robustness of the final plan. In the discrete time model, the heuristic searching method and a re-optimization model are combined to get the new orders. In the continuous time model, a re-optimization model is developed with reactive scheduling constraints for disruptive events and preventive scheduling constraints for distributional uncertainty. Some managerial experiences and heuristic rules can be configured in the model for different needs and scenarios. The results of the case studies indicate our framework can support dynamic optimization of crude oil operations under complex real-world environments.(2) Considering some of the operational characteristics of the cracking furnace system such as shutting down for decoking and varying running capacity with the processing cost and product values, an improved MINLP model is proposed to achieve the best economic performance of the cracking furnace system. The model is solved by a new heuristic iterative solving method. A numerical case study demonstrate the feasibility and the efficiency of the model.(3) A novel synchronized decision-supporting framework for the long-term planning and scheduling problem of an ethylene plant is developed by simultaneously considering the upstream naphtha inventory management and the downstream ethylene furnace operation. The resulting framework is shown to be robust and seamlessly integrated with a subcooperative model. The upstream and downstream subproblems are solved iteratively and converged to an optimized feasible plan. The efficacy and the economic potential of this proposed integration framework are illustrated through a comprehensive case study from the real world ethylene plant.(4) Considering the importance of ethylene cracking operation, an optimization model which integrates the scheduling problem of upstream cracking furnaces and the operational planning of the downstream units is proposed. Due to different time scales in the two sub models, a global time scale is developed to synchronize these time slots in this model. Moreover, a modified Lagrangean decomposition algorithm is proposed for solving the large-scale mixed integer nonlinear optimization problem. An industrial case study demonstrates the feasibility of the integrated model and the effectiveness of the solution algorithm.(5) The ethylene plant production planning problem is investigated with the consideration of uncertainty in raw material supplies and product demands. To tackle the uncertainty, chance constrained modeling technique is applied with the aim of generating reliable production plan that can satisfy the supply and demand constraints with predefined reliability. Individual and joint chance constrained models are formulated and further approximated through robust optimization. To improve the production profit while satisfying the solution reliability, an improved iterative solution method for uncertainty set size adjustment is proposed. Case studies on a realistic ethylene plant show that the joint chance constrained model is less conservative than the individual chance constrained model. Furthermore, the robust optimization approximation method and the improved iterative set size adjustment algorithm can efficiently solve the ethylene plant planning problem with joint chance constraints.At the end of this thesis, promising future researches on integration of planning and scheduling and uncertainties in petrochemical industry are discussed.