Formal DEVS modelling and simulation of a flow-shop relocation method without interrupting the production

https://doi.org/10.1016/j.simpat.2006.01.001Get rights and content

Abstract

This paper presents an organizational method of transfer of a flow-shop to a new site without interrupting the production. This method, which can be used even in case of requirements during the removal period, consists in segmenting the flow-shop removal into successive removals of groups of machines. This method can be successfully applied provided a prime condition is met: there must be sufficient stock upstream of the different groups of machines. The role of these stocks is to ensure continuity of production operations between the old and the new site, while a non-operational group of machines is being removed. The removal of groups of machines goes on until the whole production line is transferred to the new site and operational on this site. To validate this approach, we used a simulation and developed a flow-shop model according to the DEVS formalism. Our model makes it possible to segment a production line. We can therefore simulate a sequential transfer of groups of machines to the new site. The most effective solutions among those proposed suggest starting with the final group of machines (finished products) to finish with the first group. In this paper, we present and discuss some simulation results for an industrial case study. These results demonstrate the compared effectiveness of different removal strategies, and help decide for the appropriate estimated project management.

Introduction

Relocating a production site is a difficult industrial project and companies are often reluctant to get into this kind of trouble, especially small or medium sized companies that have to go on operating with the same machines. A simple solution consists in removing the machines during paid holidays: it is then possible to close the firm for the duration of the removal process. However, a firm working on orders needs not only to be continuously present on the market but also to be highly reactive, and therefore this strategy is not conceivable for this type of firms. Although similar situations are commonplace, this problem has not been dealt with yet. We considered the case of relocating the current facilities of a company to a nearby site. Two companies that were faced with the problem of removing a line of machines from one site to another without interrupting the production twice approached us. The constraints of the removal project were the same for both companies:

  • The production could not be entirely stopped during the removal of the machines.

  • The removal process was of flow-shop type.

  • The budget of the removal operation was limited, which forbade any solution of flash removal type (removal of the whole of the production system overnight or over a weekend).

  • The global production process remained identical, i.e. the logical chain of operations remained the same (the same machines would continue to be used one after the other).

This unusual problem is not dealt with in reference books about production organization [2], [9] or facilities planning [26]. Even the fundamental book on the methods of facilities layout [17] does not deal with the organization of a removal faced with such constraints. Recent papers have focused on facilities relocation. In fact, these papers deal with location or layout strategies, by considering mathematical approaches [3], [12], [15], [5] or by proposing oriented management methods [24], [19], [20], [16].

However, the removal organization problem is not dealt with in these papers.

In this context, and as a consequence, we have elaborated an organizational method that allows production to keep on going during the removal of a flow-shop. We propose here a solution of progressive removal transfer of the resources to the new site by segmenting the whole production line in sub-sets and building up a buffer stock for each elementary sub-set removal.

The true problem lies in the choice of the right parameters to balance the relocation organization. A great number of solutions can be planned to segment the totality of the production line. The use of simulations in the field of manufacturing systems to simplify the complexity and reduce the problem dimension is recognized by scientists and industrial managers. In our case, the simulation aims at determining the best combination of input parameter values, given an output criterion [21].

In this paper, attention is focused on DEVS modelling of a relocation problem. The paper is structured as follows: Section 2 describes the relevance of the DEVS formalism for the simulation of manufacturing systems. Section 3 presents theoretical considerations for an elementary removal of a production workstation. Section 4 is devoted to the description of the relocation organization of the complete production line. We use a case study and different simulations to test various strategies about the removal organization of the complete line. Section 5 describes an industrial case, the production line of printed-circuit boards, working on orders. The DEVS models are provided. Section 6 analyzes the simulation experiment results which the company can use to select the most suitable relocation strategy.

Section snippets

DEVS application for manufacturing systems simulation

In this section, the issue of the simulation of a manufacturing system is considered. First, the characteristics of industrial engineering problems are described. Then, the DEVS formalism concept is introduced and its advantages exposed.

Analysis of the modelling problem

In order to relocate a complete production line, an elementary pattern of how the production system will behave during removal is needed: we describe some of the theoretical facts obtained from the example of a sub-set composed of only one machine. We call Mi the machine number i in a line of n machines. The machine which is being removed is noted k. We also note Cpmini=1nCi the production rate of the line, with Ci the maximal production rate of machine Mi, Si the product stock located

Progressive removal of a production line: actual study

As we have seen in the previous paragraph, the scheduled stopping of a workstation in a production line because of its removal can be perfectly anticipated by creating a temporary buffer stock. We can thereby suppress the impact of this stopping on the production rate. We will now generalize the use of this method to an entire production line without any interruption of activity.

In order to exploit all the above elements, we suggest a removal organization based on the two following principles:

Simulation of a removal: an industrial case

We present a project of removal made necessary by the decrepitude of the initial site and a lack of available surface, due to the accumulation of new equipment over the years.

A new industrial building with more floor space allowed a different organization of the machines, while keeping the same functional logic. This had several advantages: better quality of life for the production staff, facilitated inventory control, and better public image towards the customers.

Presentation of simulation results

For each regrouping strategy, the complete simulation of the removal allowed us to determine the evolution of the production flow and that of the Work-In-Process stocks, supposing five days of work per week.

In this flow-shop, there were a lot of machines whose elementary removal time tMov was longer than a weekend’s time. Therefore it was impossible to consider a grouped removal distributed over several weekends: it was only possible to organize each removal by including weekends. The duration

Conclusion

In this paper, we approached a method dealing with unusual problem thanks to discrete events simulation. We dealt with the relocation of a production line from an old site to a new one. Relocation implies that the pre-existing production line, consisting of several machines, has to be totally dismounted from the old site and rebuilt on the new site.

We thus proposed a relocation method for a situation where the production cannot be completely stopped during the removal of the machines. We wanted

References (29)

  • A. Anglani, A. Grieco, E. Guerriero, F. Nucci, Q. Semeraro, T. Tolio, Object oriented simulation models based on the...
  • N.J. Aquilano et al.

    The Fundamentals of Operations Management

    (1995)
  • G. Bel et al.

    Pilotage assisté par la simulation discrète

  • J. Brimberg et al.

    Facility location with closest rectangular distances

    Naval Research Logistics

    (2000)
  • Cited by (8)

    View all citing articles on Scopus
    View full text