基于分片线性逼近的聚氯乙烯生产计划优化分解算法

聚氯乙烯全流程生产过程计划优化往往描述为复杂MINLP模型,求解难度非常大,为此引入分片线性技术逼近实际生产中的非线性特征,建立基于HH的MILP模型,进一步提出一种基于离线层级模型的分解算法来加速求解过程:第一层在对生产设备以最优能耗点进行层级划分得到离线层级模型的基础上优化一个等价MILP问题,确定表征设备操作状态的二值变量;第二层以HH模型为基础,在二值变量确定的情况下,代入计划优化模型调整设备的工作点,最终确定模型的最优操作决策方案。最后,以一个实际工厂规模的案例来验证模型和算法的有效性,结果表明本算法在基本不损失优化结果性能的前提下可以大大提高求解效率,缩短求解时间达99%以上。     

石油供应链计划层优化与不确定性风险管理模型

提出了一种基于有限场景的两阶段随机混合整数线性规划(MILP)模型,来优化不确定条件下多周期、多层级的石油工业供应链的计划层管理。模型以碳排放税的形式将减少CO2排放的环境目标融入到经济目标之中。供应链的各级节点均以黑箱的形式存在,使模型得以简化,在时间尺度相对较长的计划层获得优化结果,为供应链的计划与管理提供指导方案。并且分析了算例最优期望收益的风险性,在此基础上引入风险管理约束,得到了带有风险管理约束的供应链计划层优化模型。该模型的结果与原模型相比,期望收益附近的收益风险性降低。     

石油供应链计划层优化与不确定性风险管理模型

提出了一种基于有限场景的两阶段随机混合整数线性规划（MILP）模型,来优化不确定条件下多周期、多层级的石油工业供应链的计划层管理。模型以碳排放税的形式将减少CO₂排放的环境目标融入到经济目标之中。供应链的各级节点均以黑箱的形式存在,使模型得以简化,在时间尺度相对较长的计划层获得优化结果,为供应链的计划与管理提供指导方案。并且分析了算例最优期望收益的风险性,在此基础上引入风险管理约束,得到了带有风险管理约束的供应链计划层优化模型。该模型的结果与原模型相比,期望收益附近的收益风险性降低。     

基于改进MILP算法的甘油发酵产1,3-丙二醇最优可行代谢路径分析

针对混合整数线性规划(MILP)算法求解所有最优可行解时变量搜索范围过大的问题,提出了用通量可变性分析(FVA)改进MILP算法;并以克雷伯氏杆菌发酵甘油产1,3-丙二醇(1,3-PD)为研究对象,以1,3-PD产率最大为优化目标,利用改进MILP算法得到不同氧气消耗下所有最优代谢路径;进一步通过ATP消耗总量分析,确定ATP消耗最小的最优代谢路径。     

印染车间作业计划优化调度

由于基于统一离散时间表示的生产过程优化模型的约束和变量多,基于连续时间的批处理短期调度在近10年得到了广泛的重视和研究。本文简略介绍了生产过程调度模型的情况,通过对生产过程优化调度模型和印染生产工艺的研究,建立了基于连续时间印染生产过程优化调度MILP模型。然后,通过把数学模型转换成IL-OG OPL语言描述的模型,以浙江省某印染企业2个案例为数据,利用ILOG CPLEX进行求解,调度结果以甘特图的形式表达。结果表明印染生产连续时间MILP调度模型的有效性,优化了车间生产资源的配置。     

基于流程模拟的常减压装置过程操作与生产计划集成优化

传统的炼油企业生产计划优化与过程操作优化往往是分离的, 从而造成生产计划优化系统制定出的生产方案可能在实际的生产装置操作上无法实现的情况。为了确保石化企业生产计划制定的方案可行的同时实现过程装置操作优化, 基于流程模拟软件建立了常减压蒸馏装置生产计划与过程操作的集成优化策略, 并提出了该优化策略的有效寻优方法。该方法通过流程模拟软件验证生产计划的可达性, 不断修正生产计划关键变量的优化区间, 在求得生产计划最优解的同时确定装置的工艺操作条件。以某炼油厂常减压蒸馏装置炼油为例验证提出的集成优化方法, 案例证明该集成策略不仅确保生产计划在实际生产的可操作性, 还得到了生产计划与过程装置操作的同步优化。     

基于动态模型的在线反馈优化

针对化工过程动态波动明显、优化模型存在较多的不确定性等特点,提出了一种考虑过程不确定性、基于过程动态模型的在线反馈优化策略。将过程动态模型按一定周期离散化为差分方程,基于差分方程进行动态优化,优化目标函数为优化时域的终端时刻的经济指标,优化变量为过程的操作变量,采用非线性规划作为优化算法;优化结果在实施后根据可测输出进行在线反馈,在优化模型的差分方程中引入误差修正项,将对应时刻的状态变量和相关变量的实际值代入可求出误差修正项,从而实现在线反馈优化。仿真结果表明,与传统的稳态操作优化相比,基于动态模型的反馈优化同样可将过程运行于最优操作点,同时具有很强的实时性,在外界干扰出现时可以立即作出反应,将过程推向最优操作点。     

基于不确定蒸汽需求和设备故障的锅炉系统随机规划设计

公用工程系统中锅炉系统设计不仅要考虑锅炉应对发生的故障的可靠性,还要保证系统具有一定应对蒸汽需求的波动的可操作性。本研究针对满足不确定蒸汽需求和考虑设备故障的锅炉系统设计,提出基于数学规划法的设计模型:对生产过程波动引起的不确定蒸汽需求以概率表达,采用Markov模型分析锅炉故障,表达为以一定概率发生的不确定参数。优化模型采用二阶段随机规划策略对蒸汽需求不确定波动和锅炉故障实现引起的约束违背进行补偿,以降低不确定变量对目标函数和约束条件的影响。以年总费用最小为目标,建立混合整数线性模型(MILP),实现锅炉系统配置,设备模式确定以及应对蒸汽需求波动和设备故障发生的补偿操作的优化设计。     

基于不确定蒸汽需求和设备故障的锅炉系统随机规划设计

公用工程系统中锅炉系统设计不仅要考虑锅炉应对发生的故障的可靠性,还要保证系统具有一定应对蒸汽需求的波动的可操作性。本研究针对满足不确定蒸汽需求和考虑设备故障的锅炉系统设计,提出基于数学规划法的设计模型：对生产过程波动引起的不确定蒸汽需求以概率表达,采用Markov模型分析锅炉故障,表达为以一定概率发生的不确定参数。优化模型采用二阶段随机规划策略对蒸汽需求不确定波动和锅炉故障实现引起的约束违背进行补偿,以降低不确定变量对目标函数和约束条件的影响。以年总费用最小为目标,建立混合整数线性模型（MILP）,实现锅炉系统配置,设备模式确定以及应对蒸汽需求波动和设备故障发生的补偿操作的优化设计。     

芳烃抽提过程多目标优化

芳烃抽提是芳烃生产过程的重要环节,其生产调优对提高整个芳烃联合装置的效益具有重要意义。基于流程模拟及响应面分析方法,得到了芳烃抽提过程的产品纯度模型及能耗模型。建立了以产品纯度最大化及过程能耗最小化的多目标优化模型。提出了一种改进的自适应加权求和算法,并用于多目标优化模型的求解。求解结果表明新算法在Pareto最优解分布的均匀性上与原算法相当,但求解效率要高于原算法。给出了不同产品等级下的最佳操作参数,采用优化后的操作参数可有效地提高产品纯度并降低过程能耗。提出的多目标优化模型及求解算法用于芳烃抽提过程的操作调优,可有效地提高决策的准确性。     

Plantwide operability assessment for nonlinear processes using a microscopic level network analysis

The increase of raw material and energy costs has caused a shift in process design philosophy leading to more complex chemical plants utilising heat integration and material recycles. This warrants plantwide dynamic operability analysis in the process design stage. In our previous work, a networked plantwide operability analysis approach was developed, where the plantwide process is viewed as a network of process units connected via mass and energy flow. Such an analysis is based on the dissipativity of each process unit and the topology of the process network. However, to determine the dissipativity of multivariable nonlinear process units is often extremely difficult. In this work, we take the network approach to a microscopic level and treat each nonlinear multivariable process unit as a network of individual (single state) mass and energy balances (sub-systems). The plantwide process is then viewed as a network of such sub-systems rather than physical process units. The dissipativity of these simple sub-systems can often be determined more easily in comparison to that of multivariable sub-systems. The dissipativity property (in terms of supply rate) of the entire nonlinear process can be parametrised by the dissipativity of individual sub-systems, leading to a cluster of supply rates. The operability of the plantwide nonlinear process can then be determined based on the above parametrised dissipativity which can be much less conservative than existing nonlinear analysis. The effects of interactions caused by the interconnections are considered explicitly based on the network topology. The stability and stabilisability analysis problem is then converted into a feasibility problem with linear matrix inequalities which can be solved numerically. The application of the proposed approach requires successful determination of the dissipativity of nonlinear sub-systems.     

Tactical capacity planning for semiconductor manufacturing: MILP models and scalable distributed parallel algorithms

A multiperiod stochastic mixed‐integer linear programming model is developed to address the tactical capacity planning of semiconductor manufacturing with considerations of complex routing of material flows, in‐process inventory, demand and capacity variability, multisite production, capacity utilization rate, and downside risk management. Both planning level decisions (i.e., capacity allocation and customer service level decisions) as well as operational level decisions (i.e., production, inventory, and shipment decisions) can be simultaneously determined based on the two proposed multiobjective optimization models. To address the huge number of scenarios needed to characterize the uncertainty and the large number of first‐stage integer variables in industrial scale applications, two novel scalable distributed parallel optimization algorithms are developed to mitigate the computational burden. The proposed mathematical models and algorithms are illustrated through two case studies from a major US semiconductor manufacturer. Results from these case studies provide key decision support for capacity expansion in semiconductor industry. © 2016 American Institute of Chemical Engineers AIChE J, 62: 3930–3946, 2016     

基于分解算法的功热交换网络多目标优化

功热网络设计问题指在流程设计中对变压和换热过程进行耦合优化设计的问题,以此来提高整体系统的能效并降低成本。前人工作中一般采用数学规划法对功热网络建模优化。然而,由于存在变压过程和换热器面积计算的非线性约束,以及换热匹配的二元变量,整体模型往往是一个高度非凸的混合整数非线性规划模型,难以求解。本文提出一种高效的功热网络优化方法。模型中分别用透平压缩机和换热器实现功热网络中轴功和热的交换。求解过程采用分解算法,主问题中用随机算法对关键变量优化,功和热两个子网络问题中用确定性算法求解。目标函数考虑了经济和环境影响。案例测试对比了不同优化目标得到的结果以及多目标Pareto曲线,验证了所提出方法的高效性。     

不确定条件下炼化企业计划与调度整合策略

A strategy for the integration of production planning and scheduling in refineries is proposed.This strategy relies on rolling horizon strategy and a two-level decomposition strategy.This strategy involves an upper level multiperiod mixed integer linear programming(MILP) model and a lower level simulation system,which is extended from our previous framework for short-term scheduling problems [Luo,C.P.,Rong,G.,"Hierarchical approach for short-term scheduling in refineries",Ind.Eng.Chem.Res.,46,3656-3668(2007)].The main purpose of this extended framework is to reduce the number of variables and the size of the optimization model and,to quickly find the optimal solution for the integrated planning/scheduling problem in refineries.Uncertainties are also considered in this article.An integrated robust optimization approach is introduced to cope with uncertain parameters with both continuous and discrete probability distribution.     

Product tri-section based crude distillation unit model for refinery production planning and refinery optimization

Accuracy of a crude distillation unit (CDU) model has a significant impact on refinery production planning. High accuracy is typically accomplished via nonlinear models which causes convergence difficulties when the entire refinery model is optimized. CDU model presented in this work is a mixed-integer linear model with a modest number of binary variables; its accuracy is on par with rigorous tray to tray CDU models. The model relies on the observation¹² that a line through the middle of the product true boiling point (TBP) curve depends on the crude feed properties and the yields of the adjacent products. Novelty of the product tri-section CDU model is that it does not require models of individual distillation towers comprising the CDU, thereby leading to a much simpler model structure. Significant reduction in the computational effort required for the optimization of nonlinear refinery models is illustrated by comparison with previous work.     

An MILP Model for Optimal Scheduling of the Lubricant Production Plant

In this article we are developing a model that can be used for determining the optimal production schedule in a lubricant production plant. The model includes all the main stages in the lubricant production process, contains both continuous and binary variables, and results in the formulation of a mixed integer linear programming (MILP) problem that is solved using standard optimization techniques. The model can be easily adapted to any lube production facility, thus providing a valuable tool to refineries in their effort to automate the production scheduling process. The proposed tool can save valuable time and resources by eliminating the time-consuming search for a feasible production plan that production engineers go through in order to meet production demands.     

Planning and scheduling of flexible process networks under uncertainty with stochastic inventory: MINLP models and algorithm

The tactical planning and scheduling of chemical process networks consisting of both dedicated and flexible processes under demand and supply uncertainty is addressed. To integrate the stochastic inventory control decisions with the production planning and scheduling, a mixed‐integer nonlinear programming (MINLP) model is proposed that captures the stochastic nature of the demand variations and supply delays using the guaranteed‐service approach. The model takes into account multiple tradeoffs and simultaneously determines the optimal selection of production schemes, purchase amounts of raw materials, sales of final products, production levels of processes, detailed cyclic production schedules for flexible processes, and working inventory and safety stock levels of all chemicals involved in the process network. To globally optimize the resulting nonconvex MINLP problems with modest computational times, the model properties are exploited and a tailored branch‐and‐refine algorithm based on the successive piecewise linear approximation is proposed. To handle the degeneracy of alternative optima in assignment configurations of production scheduling, three symmetry breaking cuts are further developed to accelerate the solution process. The application of the model and the performance of the proposed algorithm are illustrated through three examples with up to 25 chemicals and 16 processes including at most 8 production schemes for each flexible process. © 2012 American Institute of Chemical Engineers AIChE J, 59: 1511–1532, 2013     

Robust design and operations of hydrocarbon biofuel supply chain integrating with existing petroleum refineries considering unit cost objective

This paper addresses the optimal design and planning of the advanced hydrocarbon biofuel supply chain with the unit cost objective. Benefited from the drop-in properties of advanced hydrocarbon biofuels, the supply chain takes advantage of the existing petroleum infrastructure, which may lead to significant capital and transportation savings. A mixed-integer linear programming model is proposed to simultaneously consider the supply chain design, integration strategy selection, and production planning. A robust optimization approach which tradeoffs the performance and conservatism is adopted to deal with the demand and supply uncertainty. Moreover, the unit cost objective makes the final products more cost-competitive. The resulting mixed-integer linear fractional programming model is solved by tailored optimization algorithm. County level cases in Illinois are analyzed and compared to show the advantage of the proposed optimization framework. The results show that the preconversion to petroleum-upgrading pathway is more economical when applying the unit cost objective.     

Short-term scheduling and recipe optimization of blending processes

The main objective of this paper is to develop an integrated approach to coordinate short-term scheduling of multi-product blending facilities with nonlinear recipe optimization. The proposed strategy is based on a hierarchical concept consisting of three business levels: Long-range planning, short-term scheduling and process control. Long-range planning is accomplished by solving a large-scale nonlinear recipe optimization problem (multi-blend problem). Resulting blending recipes and production volumes are provided as goals for the scheduling level. The scheduling problem is formulated as a mixed-integer linear program derived from a resource-task network representation. The scheduling model permits recipe changeovers in order to utilize an additional degree of freedom for optimization. By interpreting the solution of the scheduling problem, new constraints can be imposed on the previous multi-blend problem. Thus bottlenecks arising during scheduling are considered already on the topmost long-range planning level. Based on the outlined approach a commercial software system has been designed to optimize the operation of in-line blending and batch blending processes. The application of the strategy and software is demonstrated by a detailed case study.     

Nonlinear scheduling with time-variable electricity prices using sensitivity-based truncations of wavelet transforms

We propose an algorithm for scheduling subject to time-variable electricity prices using nonlinear process models that enables long planning horizons with fine discretizations. The algorithm relies on a reduced-space formulation and enhances our previous work (Schäfer et al., Comput Chem Eng, 2020;132:106598) by a sensitivity-based refinement procedure. We therein expose the coefficients of the wavelet transform of the time series of independent process variables to the optimizer. The problem size is reduced by truncating the transform and iteratively adjusted using Lagrangian multipliers. We apply the algorithm to the scheduling of a multi-product air separation unit. The nonlinear power consumption characteristic is replaced by an artificial neural network trained on data from a rigorous model. We demonstrate that the proposed algorithm reduces the number of optimization variables by more than one order of magnitude, whilst furnishing feasible schedules with insignificant losses in objective values compared to solutions considering the full dimensionality.