Product family-based assembly sequence design methodology

Efforts taken by manufacturing companies to meet the increasing demand for product variety often lead to a proliferation of subassemblies. In this paper, we show that careful design of the product assembly sequence helps to create generic subassemblies that reduce subassembly proliferation and the cost of offering product variety. This approach of designing the assembly sequence to maximize the benefit from commonality of components and assembly operations, referred to as product family-based assembly sequence design, is the focus of this paper. After introducing the approach with a simple example, we formalize the notion of generic subassemblies, and present an algorithmic approach to identify generic subassemblies. We illustrate the algorithm with an example from the literature of an assembly from industry, and provide computational test results of the complexity and benefits of product family-based assembly sequence design.     

Determination of optimal part delivery dates in a stochastic assembly line

We consider the problem of determining the optimal part delivery dates of a low volume, small lot assembly line which manufactures large and costly products. Parts are assembled at each stage to a subassembly. The arrival time of an assembly part from a vendor is assumed to be a normally distributed random variable. A dynamic programming algorithm is developed to minimize work-in-process inventory of parts and subassemblies. It is implemented on the single job problem. Its further extensions to the multi-job case and to incorporate the cost of setting delivery dates are also discussed.     

Generating significant subassemblies from 3D assembly models for design reuse

Significant subassemblies are defined as the reusable regions of pre-existing 3D assembly models. A significant subassembly has great significances for design reuse as it aggregates abundant knowledge in a vivid 3D CAD model and enables designers to reuse existing mature designs from a high-level perspective. Consequently, this paper contributes to significant subassembly generation from pre-existing 3D assembly models for design reuse. The paper first gives an explicit definition of significant subassemblies and further explores the multilevel knowledge embedded in these significant subassemblies. Based on the above definition and multilevel knowledge, a knowledge-based approach is then proposed for significant subassembly generation, which includes three phases: (1) identifying candidate subassemblies with high cohesion inside and low coupling outside using the Markov clustering process; (2) removing normal candidate subassemblies with low reusability and less information, and generating filtered subassemblies using the proposed assembly frequency – inverse mean subassembly frequency based scheme; and (3) determining significant subassemblies by measuring the complexity of the filtered subassemblies. Finally, a computer numerical control honing machine model is taken as an application example to demonstrate the effectiveness of the proposed approach.     

Change management in modular assembly systems to correspond to product geometry change

Modular assembly systems are a category of changeable manufacturing systems, which can handle the rapid change in customer demands, product design change and market fluctuations. On the operational level, jigs and fixtures are fundamental elements of assembly systems. They are used to hold parts and subassemblies in place, and directly affect assembly cost, quality and time. Therefore, modular fixtures that can adapt to different geometries are becoming a very important enabler for changeable manufacturing. In this paper, two mathematical models are presented to optimise the use of a passive modular assembly fixture plan in an automated assembly system by considering different production scenarios and constraints. These models optimise the changeability plan of the modular fixture by minimising the number of dowel replacements between different part geometries assuming that the candidate dowels locations for each part have been determined using existing methods in the literature by considering different assembly requirements. The first model, LRTE, considers all possible part rotations and translations on the fixture to minimise setup time. In addition, the second model, SLRTE, enables the system to simultaneously optimise job sequence. This paper presents various examples in different sizes, and the results show that the model can effectively reduce the fixture setup time up to %50.     

Discussion. A reply to ‘A note on a “Change to group technology”’‡

In the industrial field, the idea of subassembly must be distinguished from the subset one, when considering subassemblies as the only subsets in which all parts are stable. An original approach using the disassembly method, the logical reverse chronology of assembly, and the introduction of a mathematical model describing the liaisons allow us to determine the problems related to the automation of assembly sequences. To do so, the modelling of functional liaisons between parts helps to distinguish a simple contact from an attachment, and subsets from subassemblies. Liaisons between components are described by matrices of half degrees of liaison. These liaisons are extracted automatically from a B-rep model of a mechanical product on CAD. Subassemblies are determined automatically just in time during the determination of disassembly sequences. A limitation of this approach is that as soon as a relation is no more orthogonal to the basic referential, a new referential has to be introduced. An expert will choose the best sequence between sequences which were generating.     

Synchronized manufacturing at final assembly and feeder shops

Many manufacturing environments consist of a final assembly shop that receives components from various feeder shops. Suppose the final assembly shop assembles a mix of products, each requiring varying amounts of different components. Thus, the sequence in which products are assembled imposes time-dependent demands on the feeder shops. Such problems are often found in the assembly of high-tech telecommunications systems and in the assembly of mixed models of automobiles. We develop a variety of models that link the work-in-process inventory in the feeder shop and the finished component inventory to the component production interval and the final assembly sequence. The manufacturing disciplines considered include push, look-ahead and kanban. Feeder shops are assumed to manufacture components continuously, periodically, or at a fixed lot size. Simple performance characteristics of the feeder shop are derived to provide insight into synchronized manufacturing issues and to help evaluate the effect of different final assembly sequences on feeder shops.     

Complex assembly variant design in agile manufacturing. Part I: System architecture and assembly modeling methodology

In the distributed and horizontally integrated manufacturing environment found in agile manufacturing, there is a great demand for new product development methods that are capable of generating new customized assembly designs based on mature component designs that might be dispersed at geographically distributed partner sites. To cater for this demand, this paper addresses the methodology for complex assembly variant design in agile manufacturing. It consists in fundamental research in two parts: (i) assembly modeling; and (ii) assembly variant design methodology. This paper, the first of a two-part series, presents the assembly variant design system architecture and the assembly modeling methodology. First, a complementary assembly modeling concept is proposed with two kinds of assembly models, the hierarchical assembly model and the relational assembly model. The first explicitly captures the hierarchical and functional relationships between constituent components whereas the second explicitly captures the mating relationships at the form-feature-level. These models are complementary in the sense that each of them models only a specific aspect of assembly-related information but together they include the required assembly-related information. They are further specialized to accommodate the features of assembly variant design. As a result, two kinds of assembly models, the assembly variants model and the assembly mating graph are generated. These assembly models serve as the basis for assembly variant design which is discussed in the companion paper.     

A new generic timed Petri net model for design and performance analysis of a dual kanban FMS

A new generic black token timed Petri net (PN) model of FMS based on the process plan is presented. The model allows modelling the FMS configuration, determining the optimal work in process, the number of kanban cards and the required capacity for a given maximum production rate. The lead-time and utilization of each station can be determined. Each station can have multiple identical servers. The approach is based on modelling the FMS as an event graph with part circuits, kanban circuits and sequencing circuits. The part circuits contain tokens that represent parts with attached kanban cards. In the kanban circuits there will be places containing tokens that represent kanban Move Cards (MC) and Production Cards (PC). Circuits called sequencing circuits will model the sequencing of the parts on each station. The tokens in these sequencing circuits represent the number of identical servers. The optimal solution of the model is obtained in three steps. The first step is the initial marking of the PN. The second step involves determination of the invariants of the PN. The third is solving a linear programming model.     

A genetic algorithm for the stochastic mixed-model U-line balancing and sequencing problem

Mixed-model assembly lines are widely used to improve the flexibility to adapt to the changes in market demand, and U-lines have become popular in recent years as an important component of just-in-time production systems. As a consequence of adaptation of just-in-time production principles into the manufacturing environment, mixed-model production is performed on U-lines. This type of a production line is called a mixed-model U-line. In mixed-model U-lines, there are two interrelated problems called line balancing and model sequencing. In real life applications, especially in manual assembly lines, the tasks may have varying execution times defined as a probability distribution. In this paper, the mixed-model U-line balancing and sequencing problem with stochastic task times is considered. For this purpose, a genetic algorithm is developed to solve the problem. To assess the effectiveness of the proposed algorithm, a computational study is conducted for both deterministic and stochastic versions of the problem.     

Subassembly identification based on grey clustering

Nowadays discarded electromechanical products are more and more common, and have done much harm to the ecological environment, human health and natural resources. In order to recycle discarded products effectively, it is necessary to disassemble them properly in an integrated consideration of economic returns and environmental protection. Therefore Disassembly Process Planning has become a key part in Environmentally Conscious Design. Because there may be a combination explosion when the disassembly process of a product with a large number of parts is planned, subassembly identification is often adopted to divide the product into some reasonable subassemblies. This paper uses a method of grey clustering based on grey system theory to perform subassembly identification. The clustered objects are part pairs with adjacency relation in a product, and the clustering indices consist of energy consumption of disassembly, disassembly time, disassemblable direction and diameter of part pair. The indices can be obtained by detailed estimate or direct input from a CAD system via secondary development. After five grey clusters are set up, their whitening weight functions are presented in detail. Several other key problems in subassembly identification based on grey clustering are also expounded, such as obtaining the nondimensional matrix of sample values, determining the weight of clustering index relative to grey cluster. A heat-sealing machine with 80 components is selected as an example to validate the method of subassembly identification based on grey clustering. The identification result of the example is feasible according to experiences, and at the same time it satisfies the requirement of Disassembly Process Planning.     

Kitting in multi-echelon,multi-product assembly systems with parts substitutable

The kitting problem in multi-echelon, multi-product assembly systems with parts substitutable is to allocate on-hand stock and expected future deliveries to kits to minimize total cost including job earliness, job tardiness, and in-process holding cost while considering shop capacity and subassembly precedence restrictions, and parts being substitutable. This class of problem is NP-hard. When dealing with a large instance encountered in industry, in the worst case, it may not be possible to obtain an optimal solution in a reasonable time. In this paper, rules for using part substitutes along with heuristic procedure are presented. Computational results demonstrate that the proposed heuristic outperforms others tested. It is expected that the new heuristic can be applied in resolving largescale kitting problems encountered in industry to allocate available resources near optimality to enhance schedule performance and to lower the total cost of operating multi-echelon, multi-product assembly systems.     

A multi-class approximation technique for the analysis of kanban-like control systems

Analytical methods have been proposed in the literature for performance evaluation of kanban control systems. Among them, the method presented by Di Mascolo and colleagues appears to be of special interest since it can handle manufacturing stages consisting of any number of machines and it is fairly accurate. This paper presents a new way of deriving the analytical method presented by Di Mascolo et al. The approach is to see the queueing network of the kanban control system as a multiclass queueing network in which each kanban loop is represented by a class of customers. This allows one to use the general technique proposed in Baynat and Dallery for analysing multiclass queueing network using product-form approximation methods. In terms of equations, the new method is equivalent to that previously presented. However, the computational algorithm is much more efficient since it avoids the two levels of iterations involved in the original algorithm. Another major advantage of the new method over that originally proposed is that it provides a general framework for the analysis of more general kanban systems. Indeed, it is shown in this paper how this approach can easily be extended in order to handle kanban systems with multiple consumers and multiple suppliers, kanban-controlled assembly systems and generalized kanban systems.     

拆卸序列规划中子装配体的识别与生成

子装配体识别是解决大规模拆卸序列规划"组合爆炸"问题的有效方法之一.为了准确有效地识别与生成子装配体,在装配体零部件间接触关系、联接关系和向位妨碍关系及其相应图模型、矩阵的基础上,提出了一种基于图模型和判断矩阵的拆卸序列规划子装配体的识别算法.该算法利用图模型对子装配体集合做加法,利用判断矩阵对子装配体集合做减法,首先依据子装配体识别准则对图模型检索生成潜在子装配体集合,然后依据矩阵判断公式从集合中排除不符合连续性原则、稳定性原则和可行性原则的子装配体,并筛选理想个数的符合重量标准、同质标准和价值标准的最终子装配体集合.以减速器实例验证方法的可行性和有效性.     

Simulation optimisation of pull control policies for serial manufacturing lines and assembly manufacturing systems using genetic algorithms

Several efficient pull production control policies for serial lines implementing the lean/JIT manufacturing philosophy can be found in the production management literature. A recent development that is less well-studied than the serial line case is the application of pull-type policies to assembly systems where manufacturing operations take place both sequentially and in parallel. Systems of this type contain assembly stations where two or more parts from lower hierarchical manufacturing stations merge in order to produce a single part of the subsequent stage. In this paper we extend the application of the Base Stock, Kanban, CONWIP, CONWIP/Kanban Hybrid and Extended Kanban production control policies to assembly systems that produce final products of a single type. Discrete-event simulation is utilised in order to evaluate the performance of serial lines and assembly systems. It is essential to determine the best control parameters for each policy when operating in the same environment. The approach that we propose and probe for the problem of control parameter selection is that of a genetic algorithm with resampling, a technique used for the optimisation of stochastic objective functions. Finally, we report our findings from numerical experiments conducted for two serial line simulation scenarios and two assembly system simulation scenarios.     

Stochastic coloured Petri net (SCPN) models of traditional and flexible kanban systems

Modelling and analysis of kanban-controlled (just-in-time) production systems under realistic assumptions presents a number of challenges, including the ability to conduct both qualitative and quantitative analysis, and the ability to model control policies. These challenges are due, in part, to station interdependence, blocking and starvation due to limited buffer spaces, and the necessity of modelling both material and kanban flows. Petri nets (PNs) have recently emerged as a promising approach for modelling manufacturing systems. PNs are a graphical and mathematical technique useful for modelling concurrent, asynchronous, distributed, parallel, non-deterministic and stochastic systems. PN models can be analysed to determine both their qualitative and quantitative properties. In this paper, we develop stochastic, coloured PN (SCPN) models of a JIT system utilizing two different kanban control policies: a traditional kanban system (TKS) policy and a flexible kanban system (FKS) policy. The resulting models can be used to represent JIT systems of arbitrary size, producing single or multiple types of products, with fixed order points 1. The models are shown to be live and bounded, and can be simulated to produce quantitative results. Sample simulation results are presented to illustrate the models' capabilities.     

Maximising manufacturing system efficiency for multi-characteristic linear assembly by using particle swarm optimisation in batch selective assembly

Quality of an assembly is mainly based on the quality of mating parts. Due to random variation in sources such as materials, machines, operators and measurements, even those mating parts manufactured by the same process vary in their dimensions. When mating parts are assembled linearly, the resulting variation will be the sum of the mating part tolerances. Many assemblies are not able to meet the assembly specification in the available assembly methods. This will decrease the manufacturing system efficiency. Batch selective assembly is helpful to keep the assembly requirement and also to increase the manufacturing system efficiency. In traditional selective assembly, the mating part population is partitioned to form selective groups, and the parts of corresponding selective groups are assembled interchangeably. After the invention of advanced dimension measuring devices and the computer, today batch selective assembly plays a vital role in the manufacturing system. In batch selective assembly, all dimensions of a batch of mating parts are measured and stored in a computer. Instead of forming selective groups, each and every part is assigned to its best matching part. In this work, a particle swarm optimisation based algorithm is proposed by applying the batch selective assembly methodology to a multi-characteristic assembly environment, to maximise the assembly efficiency and thereby maximising the manufacturing system efficiency. The proposed algorithm is tested with a set of experimental problem data sets and is found to outperform the traditional selective assembly and sequential assembly methods, in producing solutions with higher manufacturing system efficiency.     

On the trade-offs between risk pooling and logistics costs in a multi-plant network with commonality

This paper examines the benefits and costs of two alternative manufacturing network configurations in the presence of component commonality. We evaluate the trade-off between the decreased logistics costs and loss of risk-pooling benefits in plant networks which spread component manufacturing over each plant (product network) as compared to those that consolidate component manufacturing in a single plant (process network). We examine for conditions that mean that a product network would be chosen instead of a process network and vice-versa. We find that the risk-pooling benefit obtained by consolidating common subassembly production is reduced when the cost of acquiring common component capacity is sufficiently high or low. A post-optimality sensitivity analysis for the process network provides insights into subtle substitution effects, which are a direct outcome of cost mix differentials and network structure and complementarity effects, which are induced by the considered sequential assembly system. Our results suggest that the impact of operational cost parameters on strategic decisions can often be non-intuitive. Overall, our analysis provides a link between strategic and operational decision-making in supply chain management, in the context of multi-plant configuration.     

Release policies for assembly systems

We consider a production system consisting of several fabrication lines feeding an assembly machine. The machines in the fabrication lines and at assembly are assumed to have general processing time distributions. Releases to the system are governed by the kanban release mechanism. We first derive a heuristic for the throughput of this system by replacing the assembly system under kanban release by an equivalent system under CONWIP release and by making use of an approximation for the throughput of a CONWIP assembly system (Duenyas, 1992). Comparisons with simulation show that this heuristic is robust over a wide range of conditions.

We also address the issue of which release mechanism is more effective in obtaining a desired throughput level with the minimum possible work-in-process inventory level in the system. We present both analytical and simulation results to demonstrate that the CONWIP release policy seems to be a more effective release policy for assembly systems.     

Selection of a Set of Part Delivery Dates in a Multi-Job Stochastic Assembly System

The problem of determining part delivery dates in a low-volume assembly line manufacturing large and costly products is studied. At each workstation a part is assembled to the subassembly. These parts are supplied by external vendors with unreliable delivery schedules. The arrival date of a part is assumed to follow a normal distribution whose mean is to be specified. A dynamic programming approach in combination with a heuristic procedure is used to prescribe the vendor delivery dates for a batch of products. Simulation results show the procedure to be significantly superior to a deterministic approach to the problem. Comparisons are also made with a safety buffer method.     

Determination of unit load sizes of AGV in multi-product multi-line assembly production systems

This study deals with the problem of determining unit load sizes of automated guided vehicles (AGVs) in multi-product and multi-line assembly production systems. It is assumed that an AGV delivers assembly parts to each workstation of assembly lines from a miniload automated storage/retrieval system (AS/RS), and conveyor lines transport subassemblies between workstations. In this system, determination of the unit load sizes of parts carried by the AGV becomes an important decision factor due to its effects on the work-in-process inventory level and the operating cost of the system. Under a proposed AGV dispatching policy, a nonlinear mathematical model which allows unequal unit load sizes among assembly lines is formulated. Based on the characteristics of the objective function and feasible regions, a heuristic solution algorithm is developed. To illustrate the model, problems are chosen and solved.