Modeling errors in parts supply processes for assembly lines feeding
| Published date | 10 July 2017 |
| Date | 10 July 2017 |
| DOI | https://doi.org/10.1108/IMDS-08-2016-0333 |
| Pages | 1263-1294 |
| Author | Antonio Casimiro Caputo,Pacifico Marcello Pelagagge,Paolo Salini |
| Subject Matter | Information & knowledge management,Information systems,Data management systems,Knowledge management,Knowledge sharing,Management science & operations,Supply chain management,Supply chain information systems,Logistics,Quality management/systems |
Modeling errors in parts
supply processes for
assembly lines feeding
Antonio Casimiro Caputo
Department of Mechanical and Industrial Engineering,
University of Roma Tre, Roma, Italy, and
Pacifico Marcello Pelagagge and Paolo Salini
Department of Industrial Engineering, Information and Economics,
University of L’Aquila, L’Aquila, Italy
Abstract
Purpose –The purpose of this paper is to develop a quantitative model to assess probability of errors and
errors correction costs in parts feeding systems for assembly lines.
Design/methodology/approach –Event trees are adopted to model errors in the picking-handling-
delivery-utilization of materials containers from the warehouse to assembly stations. Error probabilities and
quality costs functions are developed to compare alternative feeding policies including kitting, line stocking
and just-in-time delivery. A numerical case study is included.
Findings –This paper confirms with quantitative evidence the economic relevance of logistic errors (LEs) in
parts feeding processes, a problem neglected in the existing literature. It also points out the most frequent or
relevant error types and identifies specific corrective measures.
Research limitations/implications –While the model is general purpose, conclusions are specific to each
applicative case and are not generalizable, and some modifications may be required to adapt it to specific
industrial cases. When no experimental data are available, human error analysis should be used to estimate
event probabilities based on underlying modes and causes of human error.
Practical implications –Production managers are given a quantitative decision tool to assess errors
probability and errors correction costs in assembly lines parts feeding systems. This allows better comparing
of alternative parts feeding policies and identifying corrective measures.
Originality/value –This is the first paper to develop quantitative models for estimating LEs and related
quality cost, allowing a comparison between alternative parts feeding policies.
Keywords Assembly line, Human error, Material-handling error, Parts feeding, Quality cost
Paper type Research paper
Nomenclature
A(item/yr) annual end product
demand
CDE (€/container) substitution cost for
wrongly delivered stock ke eping
unit
C
oh
(€/h) operator hourly cost
CRQ (€/error) correction of undetected
assembler’s error cost
d
avg
(m) warehouse to line mean distance
EP (–) percentage of material utilization
L(m) distance
N(–) number of parts
NCA (€/error) non-conformity unit cost for
an entire assembly
n
i
(–) multiplicity of ith part type into
one end product
p(–) probability
QC# (–) quality cost of type # quality
problem
Q
i
(item/container) container capacity
QP# (–) quality problem of type #
T
a
(h) reassembly time
T
d
(h) disassembly time
TECC (€/yr) total annual error correction
cost
t
fr
(h/carton) carton fractioning time
T
i
(h) inspection time
t
L/U
(h) load/unload time
Industrial Management & Data
Systems
Vol. 117 No. 6, 2017
pp. 1263-1294
© Emerald PublishingLimited
0263-5577
DOI 10.1108/IMDS-08-2016-0333
Received 21 August 2016
Revised 7 November 2016
Accepted 14 December 2016
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0263-5577.htm
1263
Modeling
errors in parts
supply
processes
T
p
(h) item picking time
t
r/s
(h) warehouse picking time
v
tp
(m/h) operator walking velocity
V
v
(m/h) vehicle velocity
Additional subscripts
DE error detection
Eerror outcome
EA assembly error
Ggood outcome
Iitem included as a single unit into a
kit
II item included in more than one unit
into a kit
Jjust-in-time policy
JI just-in-time policy, for wrongly
delivered container being the
correct one in the supermarket
JII just-in-time policy, for wrongly
delivered container being the
correct one in the warehouse
Kkitting policy
LS line storage policy
QC# quality cost of type #
QP# quality problem of type #
SM supermarket to line
SW warehouse to supermarket
TOT total
WS warehouse to line
Greek symbols
αcorrect kit order probability
βcorrect delivery probability
γdetection probability of wrong kit
delivery
δprobability of kit right preparation
εprobability of kit error awareness by
line operator
ζprobability of correct assembling by
line operator
ηprobability of part inclusion into
a kit
θprobability of correct part inclusion
into a kit
ιprobability of part multiplicity less
than required
κprobability of part damage
λprobability of part positional
mistake into a kit
μprobability of part error detection at
kitting station
νprobability of correct order
ξprobability of right pallet loaded on
forklift
οprobability of right delivery
πprobability of operato r detecting
wrong delivery
ρprobability of picking an item from
the right pallet by line operator (LS)
σprobability of correct order
τprobability of right material at the
supermarket storage location
ϕprobability of right delivery
χprobability of right container loaded
on tugger train
ψprobability that operator detects
wrong delivery
ωprobability of picking an item from the
right container by line operator ( JIT)
1. Introduction
In assembly plants, internal material-handling systems are needed to periodically replenish
stock along the line according to production plans, thus ensuring uninterrupted supply of parts
at the workstations and a continuous production flow. This usually implies moving parts from
an internal warehouse to a dedicated storage area at the workstations (Caputo and Pelagagge,
2011; Kilic and Durmusoglu, 2015). Continuous material supply is one of the most common parts
feeding methods, where each different part type is supplied in an individual container to the
assembly line (Caputo et al., 2015a). This can be done, in turn, by adopting two different
practices. First, small-sized containers may be moved in by adopting just-in-time ( JIT) delivery,
for instance, a Kanban-based policy. Second, component containers holding bulk quantities are
simply stored along the line and periodically replenished (line storage LS). Parts kitting (K) is
1264
IMDS
117,6
another frequently adopted method to deliver parts to assembly lines. In a kitting policy
(Brynzèr and Johansson, 1995; Bozer and McGinnis 1992; Caputo and Pelagagge, 2011;
Caputo et al., 2015b; Hanson, 2012; Hanson and Brolin, 2012), all parts required to assemble one
unit of the end product are grouped together and placed into one or more kit containers.
Kits are prepared in a stockroom and delivered to the assembly line, either at the start of the line
(traveling kit concept) or to specific workstations (stationary kits), according to the
production schedule.
Overall, any parts feeding policy requires several material-handling operations (i.e. parts
picking, checking, counting, moving, and delivering) to be correctly executed in order to
provide the right material in the right quantity at the right workstation. Nevertheless,
most of these tasks are manual so that human error is a relevant cause of quality problems
(QPs) in internal material-handling systems, even if scarcely addressed by the literature
(Grosse et al., 2015). Moreover, any material-handling error to be corrected requires
additional costs, which jeopardize the performance and profitability of the business process.
The literature about human errors and QPs in assembly systems is quite rich but is mainly
focused on assembly errors. For instance, Cheldelin and Ishii (2004), Fujimoto et al. (2003) and
Zhu et al. (2008) developed line design methods to prevent human errors. Human errors in
assembly tasks have been correlated to ergonomics (Ecklund, 1995; Falck and Rosenqvist, 2014;
Falck et al., 2014; Lin et al., 2001) and to the complexity of the task. Methods have been, thus,
developed to measure difficulty of the operations (Ben-Arieh, 1993; El Maraghy and Urbanic,
2004; Zaeh et al., 2009) and to predict quality defects (Su et al., 2010). Estrada et al. (2007) develop
a taxonomy of assembly errors. Kern and Refflinghaus (2015) and Yang et al. (2012) adopted
human reliability analysis techniques to assess human error in assembly operations.
Nevertheless, human errors and QPs in parts feeding processes received much less attention.
In fact, while errors in parts feeding systems are recognized as a main problem when planning
and managing a material-handling system, literature aimed at quantifying such errors in
assembly lines feeding systems is scarce (Caputo et al., 2015a,b,c,d; Fager et al., 2014).
Moreover, while models to choose the proper parts feeding policy have been suggested
(Battini et al., 2009; Caputo and Pelagagge, 2011; Caputo et al., 2015e; Faccio, 2014; Hanson and
Brolin, 2012, Limère et al., 2012), no comparison has been attempted between the available parts
feeding policies as far as quality costs and error rates are concerned. Therefore, the issue of
modeling quality issues in parts feeding policies remains an open area of research, with
important managerial implications. With the aim of contributing to a better knowledge of this
important system design and management issue, in this paper, an attempt is made to develop
models to analyze errors during the parts delivery process, including kitting K LS and JIT
delivery, in order to allow an economic assessment of quality costs in internal logistics
operations to feed assembly lines. This would provide production managers some
decision-making tools to explore available alternatives and make cost-effective decisions.
Overall, the aim of this proposed model is as follows: to map error types and develop logic
conditions for errors occurrence, to provide a quantitative assessment of the economic impact of
material-handling errors in parts feeding systems, to identify error types which have thehighest
economic impact, and to allow a parametric analysis assessing how the logistic system responds
to changes in parameters values. Once critical problem areas are identified, resorting to the
proposed models, production managers can develop suitable corrective measures. This paper is
structured as follows. At first, the methodological approach is described, and then issues related
to kitting quality are discussed by developing a taxonomy of errors. Subsequently, error models
for kits preparation and delivery are developed, including estimation of errors correction cost.
The same is carried out for LS and JIT parts feeding. A numerical example with a sensitivity
analysis is then included in order to point out the most relevant causes of handling errors and
quantify the resulting quality cost in a given scenario comparing the aforementioned three
policies. Results, discussion, and suggestions for future research conclude the paper.
1265
Modeling
errors in parts
supply
processes
To continue reading
Request your trial