Aller au menu Aller au contenu
Travailler à G-Scop
Laboratoire des Sciences pour la Conception, l'Optimisation et la Production de Grenoble
Travailler à G-Scop
Travailler à G-Scop

> GSCOP_RECRUTEMENT > GSCOP_Thèses

Deciding the Level of Automation during the design phase of Assembly Systems

Title: Deciding the Level of Automation during the design phase of Assembly Systems

Supervisor(s): Pierre DAVID; Eric BLANCO; Joshua SUMMERS

Financing - Context - Partnerships

Expected public fellowship

Thesis in collaboration with Pr. Joshua Summers from CLEMSON University (US)

Description of the subject:   

Nowadays market in various fields as automotive or appliance industry is characterized by a high volatility and variability of the demand and a short lifetime of technological solutions. For many manufacturers, one lever to remain competitive is to rationalize the assembly Lotter and Wiendahl (2006). The rationalization of the assembly covers two main aspects: the organizational/human aspect and the technological/design aspect of the product to be assembled Ross, (2002). Lotter and Wiendahl (2006) assumed that the assembly step represents between 30% and 50% of the overall production time in the automotive industry, and up to 70% of the cost to manufacture the product in all industries. As pointed out in Ross, (2002), the automation in the assembly can help to rationalize its activity. However, companies have different experiences with automation. The statement that an increase of the use of automation leads to an increase of the productivity within the plants was a well regarded and widespread idea. However, as pointed out by Lay and Schirrmeister, (2001), the results of a survey in German companies about automation, have shown that more than a third of the 355 surveyed companies planned to reduce the level of automation (LoA) within their plants after having experienced a high LoA. The most important reason for being disappointed with automation remains the inflexibility of highly automated systems. The inflexibility of most of the automation solutions is the tender point in the above presented context, of continuously and rapidly changing market conditions. Determining the optimal LoA, considering influences like location, volume, type of product, quality required, skill level of operator, demographic changes and so on, still remains difficult for companies [Ross (2002), Frohm (2008), Gorlach et al. (2008), Fasth et al. (2008), Almannai et al. (2008)].

During our first contacts with industrial partners, an important point was that ones can observe different LoA for similar tasks within different plants, without strong rational supporting it. Moreover, as noted by Ross (2002), the discussions on the question to automate or not, are generally not well documented and the path that leads to the final decision is often not traceable. Determining the optimum LoA in an assembly system appears to be a topical subject influencing the competitiveness of companies. One main challenge is to be able to determine and justify the choice of an optimum LoA for the assembly. The LoA may be analyzed at several phases of the assembly system life, from its early design to the analysis of existing assembly systems. The LoA problem may also be tackled with various points of view: technical, tactical, human resource management etc...

In the literature, we can identify technical approach such as Boothroyd et al. (2011) determining estimated assembly time and mixed technical-financial approach in Ross (2002). Both approaches study the LoA question in the design phase of the assembly system. We also found, diagnosis method in Frohm (2008) and Fasth et al. (2008) evaluating existing assembly system to drive the future evolution of LoA. We propose to analyze the LoA decision focusing on the early design phases of the assembly system since it is the phase where decision are cheaper and have the maximum impact on the set up solution. Our objective is also to take into account broader criteria than the one used in the highly technical oriented existing literature. Indeed, we feel that the current tendency to abandon the idea of constantly increasing the automation of assembly tasks is not only driven by technical and financial aspects but also by criteria as production flexibility, human resource management or suppliers available in the plant environment.

The objective of this subject is to contribute to the identification, since the early phases of design, of the optimum LoA in assembly system and to address this issue analyzing numerous aspects of the system. These aspects include plant environment or technical aspects and tactical aspects. The LoA problem is a decision problem with a complex context that shows a lot of variability between cases. The same technical actions placed in different locations or done on product with different maturity levels, shall be executed with really different LoAs. Therefore, we identify a need to clarify and support the decision process used to define workstation in assembly system. Moreover, since defining the right LoA might be seen of less importance compare to finding a feasible technical solution, this part of the decision should be supported to easily fit into the whole workstation definition process. Work may be conducted in complementary directions: supporting the work with all the parameters to study for the decision (attributes / constraints / objectives); defining decision rules on this question; organizing the decision process.

A first analysis of the LoA problem has been conducted in Lacouture (2012), that shown that the complexity of the decision context oblige to work with rules handling multiples attributes. Few rules can be formulated between one attribute and one dimension of the problem (Solution, objective or constraint). Therefore, to develop the rule-based side of the LoA problem it seems important to find correlation between a set of attributes and its resulting LoA. Within the work done in Lacouture, (2012), we identified that there was poor literature results on this kind of rules except from expressing constraints on the difficulty to automate a task regarding the technical actions to perform (cf. Ross (2002)).

The thesis should raise contributions in several dimensions of the LoA optimization problem. First, more complete criteria will be studied: technical, social, economic, strategic and environmental criteria will be mixed. New technical aspects will be studied as the complexity and the dynamic of the assembly sequence. This will extend the currently used technical criteria that focus only on the unitary step of the assembly sequence and not on its dynamic and full realization. This analysis demands a new way to formalize assembly sequences. This will be made by defining a controlled vocabulary and a modeling method for assembly sequence description. A new design process for assembly system including the LoA dimension shall also be a strong contribution of the work. As well as the decision tool allowing the multi criteria handling, supposed to support the defined process. New types of models able to handle the huge complexity of the LoA problem context and able to break it into analyzable subparts are thus to be proposed. Ones formalizing assembly sequences will be of primary importance and should also open to related subjects as workstation instructions creation and management.

 

The expected PhD. Candidate shall demonstrate a good understanding of product assembly techniques. He shall also understand the field of production management. A good understanding of design processes, decision techniques and design for assembly method would also be useful.

The research will be conducted in collaboration between G-SCOP and CEDAR laboratories including time spent in both locations.

 

Almannai, B., Greenough, R., and Kay, J. (2008) A decision support tool based on QFD and FMEA for the selection of manufacturing automation technologies, Robotics and Computer-Integrated Manufacturing, vol. 24, no. 4, pp. 501-507.

 

Boothroyd  G., Dewhurst P., and Knight W. A. (2011) Product Design for Manufacture and Assembly, Third Edition, CRC Press, Taylor and Francis group. 

Fasth, A., Stahre, J. and Dencker, K. (2008) Analysing changeability and time parameters due to levels of Automation in an assembly system, Proceedings of the 18th conference on Flexible Automation and Intelligent Manufacturing-FAIM, no. 46, pp. 700-707.

Frohm, J. (2008) Levels of Automation in production systems, PhD Thesis, Chalmers University of Technology.

Gorlach, I. and Wessel, O. (2008) Optimal Level of Automation in the Automotive Industry, Engineering Letters, vol. 16, no. 1.

Lacouture E. (2012) Study to determine the optimal level of automation for the assembly. Master Thesis of Karlsruhe University and Grenoble-INP, advised by Pierre DAVID & Eric Blanco.

Lay, G. and Schirrmeister, E. (2001) Sackgasse Hochautomatisierung? Praxis des Abbaus von Overengineering in der Produktion, Mitteilungen aus der Erhebung Modernisierung der Produktion, Fraunhofer ISI, vol. 22, no. 22

Lotter, B. and Wiendahl, H.-P. (2006) Montage in der industriellen Produktion, Berlin Heidelberg: Springer

Mathieson, J., Wallace, B., Summers, J., (2012), Estimating Assembly Time with Connective Complexity Metric Based Surrogate Models, International Journal of Computer Integrated Manufacturing, on-line May 21, 201

Miller, M., Griese, D., Peterson, M., Summers, J., Mocko, G. (2012). Reasoning: Instalation process step instructions as an automated assembly time estimation tool. Procedings of ASME 2012 International Design Engineering Technical Conferences. August 12-15, Chicago, USA.

Ross, P. (2002) Bestimmung des wirtschaftlichen Automatisierungsgrades von Montageprozessen in der frühen Phase der Montageplanung, IWB Forschungsberichte nr. 170, München: Utz.

 

Contact(s)

Pierre DAVID: pierre.david@grenoble-inp.fr

 

mise à jour le 3 mai 2013

  • Tutelle CNRS
  • Tutelle Grenoble INP
  • Université Joseph Fourier
  • Tutelle UMR
Univ. Grenoble Alpes