【作者】 杜辉；

【导师】 楼佩煌；

【作者基本信息】 南京航空航天大学 ， 机械电子工程， 2013， 博士

【摘要】 飞机装配工装（以下简称工装）是一类结构复杂，零部件众多的机械产品，其生产模式为单件小批生产，计划与调度比较复杂。工装生产存在着生产周期长、计划不准确、订单交货拖期等突出问题。制造执行系统（MES）是面向制造过程的集成化车间生产管理与控制系统，能够及时反映生产现场的状况，有效地提高企业的车间生产管理水平。本文以满足长周期工装的齐套性要求，提高作业计划的可行性和准确性、缩短工装的制造周期以及提高订单准时交货率为目标，深入研究了飞机工装MES中的计划与调度，提出了工装多级网络计划模型及其编制方法、基于矩阵元的物料信息集成模型、基于混合粒子群算法的资源受限多项目调度方法以及事件与数据混合驱动的动态调度方法；并将上述理论与方法应用到飞机工装的生产管理中，取得了良好的效果。本文针对长周期复杂工装的生产计划准确性差的问题，研究了一种面向工装订单的多级网络计划模型（HNPM），从而将复杂的工装生产计划分解为相对简单的多级网络计划。HNPM包括四个层次的生产计划：①项目计划②总装配网络计划③成套物料需求计划④工序计划。这四个计划是从上到下由粗到细逐层分解的关系，保证在订单总的时间节点能够按时、成套交付。研究了基于作业任务分解的装配作业单元的划分方法、固定提前期与可变零件提前期压缩率相结合的提前期设置方法以及多级网络计划的编制与调整等方法。此外，为提高工时估算的精确性，提出了一种基于自组织映射神经网络(SOM)与实例推理技术(CBR)的工时估算方法。以工装订单实例说明了多级网络计划模型及其方法。工装生产中的频繁工程更改造成物料数据的不完整和不一致，影响计划的准确性。本文针对已有BOM模型对工装工程更改的物料更改信息表达不全面的问题，提出了一种基于矩阵元的物料信息集成模型（BOMII），实现了基于单一BOM模型的物料多维信息集成。BOMII不仅包含了产品结构的当前信息，而且包含了产品结构的更改信息和更改的历史信息，为计划与调度提供了较为全面和准确的产品物料信息。研究了BOMII模型构造方法，提出了基于BOMII矩阵元的更改信息的表达方法以及基于有效性矩阵的更改的历史信息的记录方法，并以接头定位器的工程更改为例说明了BOMII的具体应用。飞机工装是以项目的形式组织生产，生产中多项目并存。本文针对工装多项目之间存在资源竞争的问题，构造了面向飞机工装的资源受限多项目调度问题模型（RCMPSP），并针对已有求解算法的不足，研究了一种改进粒子群与模拟退火相结合的混合算法（IPSOSA）来实现RCMPSP的求解。在IPSOSA算法中，一方面采用周期振荡衰减权重方法来有效平衡粒子群算法的全局寻优与局部寻优；另一方面，采用模拟退火算法的全局寻优能力来克服粒子群算法易于陷入局部最优的缺陷。以工装多项目实例验证了IPSOSA解决RCMPSP的有效性。在MES的实时数据与集成数据环境下，深入研究了一种事件与数据混合驱动的动态调度方法，提高了调度系统对动态干扰事件的有效处理能力，又使调度系统具备一定的预调度与主动调度的能力。研究应对干扰事件的分类调度规则，采用事件驱动的项目级与车间级的分层动态调度，降低了工装车间调度的复杂度。将车间加工作业与装配作业集成优化，提出了一种包括装配约束的车间调度问题模型（AJSSP），并使用免疫粒子群算法求解AJSSP,既能优化调度工序作业任务，又能为紧急订单任务插入的重调度决策提供参考。提出了一种基于混合触发器的数据驱动的车间动态调度方法，实现对已发生的或未来可能发生的异常事件的监测，弥补人工监测异常事件的不足，并通过联级触发器的方法对异常事件进行及时响应与跟踪处理。针对生产过程中信息共享的需求，提出了一种包括信息发布、反馈、处理结果评价的信息双闭环控制方法，实现了信息发布与跟踪管理，为动态调度提供信息传递与共享的平台。本文在理论研究的基础上，将取得的研究成果应用于飞机工装制造执行系统的计划与调度之中，开发完成的飞机工装制造执行系统在某飞机工装制造公司进行了测试和应用。实际应用验证了本文所提的计划与调度理论和方法的正确性和可行性。更多还原

【Abstract】 Aircraft assembly tooling is a kind of mechanical product which consisting of many componentsand complicated structure. The production mode of aircraft assembly tooling is single piece and smallbatch production, and its planning and scheduling is very complicated. There are many problems inthe aircraft assembly tooling manufacturing, especially long production cycle and delivery tardiness.Manufacturing Execution System (MES) is an information management system for workshop, whicheffectively improve workshop production management level and timely delivery. So, the purpose ofthis paper is to improve the enforceability and accuracy of the operating planning, and enhance orderdelivery on-time. The planning and scheduling of MES for aircraft assembly tooling is studied depth.In this paper, theories and methods of planning and scheduling are put forward, such as hierarchicalnetwork planning model based on order, bill of material information Integration model based onmatrix, resource constrained multi-project scheduling based on hybrid particle swarm optimizationand dynamic scheduling driven by event and data, and so on. The above theories and methods areapplied to aircraft assembly tooling production management, and achieved good results.Aim at the poor accuracy of planning for the large and complex aircraft assembly tooling, aHierarchical Network Planning Model based on Order Project (HNPOP) is proposed in the secondpart of this paper. HNPOP includes project planning, total assembly network planning, complete setsmaterial requirements planning, parts planning and routing planning, and so on. The complete setsrequirement is satisfied by layer upon layer decomposition method. HNPOP model ensure that orderitem is completely delivered on-time. And, assembly operations partition method based on the workbreakdown unit, classification setting and variable compression ratio of lead time method andplanning and adjustment method are study. Furthermore, an estimated method using of case basedreasoning technology and self-organizing map is used to estimate the processing time. An example ofaircraft assembly tooling order is taken to illustrate the HNPOP model.Frequent engineering change of aircraft assembly tooling manufacturing causes incomplete andinconsistent data of materials, and affects the accuracy of the planning. However, current Bill ofMaterial (BOM) model lack of comprehensive expression of product material information confrontwith the EC environment. To solve this problem, a Bill of Material Information Integration (BOMII)model based on matrix tuple is proposed. BOMII realizes the effective integration ofmultidimensional information of product material based on single model, and it can fully describe thecurrent and change historical information of product material. So, the proposed BOMII model ensuresthe consistency and traceability of product material data in the engineering change environment.BOMII model construction method, the change information expression method using of matrix tuple,and recording method for the change history information using of validity matrix are studied.Connector locator as a case is studied to explain the BOMII model and its application.There inevitability exist multiple projects coexisting in aircraft assembly tooling manufacturingenterprises, and activities or projects usually compete for the exclusive manufacturing resources,which results in resource conflicts. Aim at the problem, a Resource Constrained Multi-ProjectScheduling (RCMPSP) model of aircraft assembly tooling is constructed. Aiming to overcome the shortcomings of existing algorithms, a hybrid optimization algorithm based on Improved ParticleSwarm and Simulated Annealing (IPSOSA) algorithm is proposed to solve RCMPSP of aircraftassembly tooling. In the IPSOSA algorithm, on the one hand, adaptive inertia weight with cyclicalattenuation strategy is employed to balance the global optimization and local optimization of standardPSO. On the other hand, Simulated Annealing (SA) algorithm is employed to overcome the prematureconvergence of standard PSO. The proposed IPSOSA was applied to aircraft assembly toolingmanufacturing. We compare the results of the IPSOSA with that of GA, SA and standard PSOmethods. The simulation results and algorithm comparison show that the IPSOSA is an effectiveapproach for the RCMPSP, and provide optimized scheduling schemes.In the real-time data and integration data environment of MES system, a Dynamic Schedulingmethod based on the Hybrid Driven by Event and Data (DSHDED) is proposed to cope with thedynamic disruption events. DSHDED not only improve system ability of dealing effectively with thedynamic disruption events, but also have capabilities both pre-scheduling and initiative scheduling. Inorder to realize dynamic scheduling driven by event both in project aspect and in job shop aspect,different scheduling rules are proposed. Considering requirement of assembly kitting, a model ofAssembly Job-Shop Scheduling Problem (AJSSP) is proposed, what is more, artificial immuneparticle swarm optimization algorithm is used to solve AJSSP. Integrated with the processingoperation and assembly operation, the optimal solution of AJSSP can provide an effective schedulingscheme to the adjustments or regenerate scheduling. In order to comprehensive monitoring ofabnormal events, a job shop dynamic scheduling diven by data is proposed and realized by a mixedtriggers method. Events that have occurred and future events that may occur are all comprehensivemonitoring, and by using the production data, dynamic scheduling system can timely response toabnormal events and then track and manage them. Especially, trigger even was used to inspect theassembly kitting for the key assembly units or plan net. It realizes that assembly operations performedautomatically driven by data. Furthermore, in order to meet the requirement of information sharing, adouble closed-loop control method is used in manage system of MES. In this way, it realizesorientation publishes of instant message and efficient information sharing and real-time feed-backespecially for the exceptions message of job shop, and also provides an information transfer andsharing platform for dynamic scheduling.On the basis of theoretical research, the research results are applied to planning and schedulingof MES for aircraft assembly tooling, named MES-GZ. Furthermore, MES-GZ is tested and appliedto an aircraft tooling manufacturing enterprises. The practical application of MES-GZ verifies thecorrectness and feasibility of the proposed theories and methods in this paper.更多还原