Journal of Cleaner Production xxx (2016) 1e9

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Reliability and cost optimization for remanufacturing process

planning
Zhigang Jiang a, *, Tingting Zhou a, Hua Zhang a, Yan Wang b, Huajun Cao c,
Guangdong Tian d
a
College of Machinery and Automation, Wuhan University of Science & Technology, Wuhan 430081, China
b
Department of Computing, Engineering and Mathematics, University of Brighton, Brighton BN2 4GJ, United Kingdom
c
State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing 400044, China
d
Transportation College, Northeast Forestry University, Harbin 150040, China

a r t i c l e i n f o a b s t r a c t

Article history: Remanufacturing is a practice of growing importance as it returns the end of life products back to
Received 18 March 2015 conditions that is as good as or better than new ones. The increasingly stringent environmental legis-
Received in revised form lation and economic demands has led to the rapid development of remanufacturing industry in the
3 September 2015
world. Remanufacturing is still at its infantry. One of the major challenges faced by remanufacturing is to
Accepted 16 November 2015
guarantee the reliability of remanufactured products since they came from cores with varying condition.
Available online xxx
Process planning plays a critical role in realizing a successful remanufacturing strategy since it directly
affects the success rate of remanufacturing as well as reliability and cost. To do so, this work presents an
Keywords:
Remanufacturing
optimization method for remanufacturing process planning in which reliability and cost are taken into
Process planning consideration. In this method, reliability is represented by failure rate of remanufacturing operations
Reliability which is influenced by the quality of returned used products (cores), whilst process cost includes ma-
Genetic algorithm chine cost and tool cost. The multi-objective optimization problem is solved by a genetic algorithm. To
assess the usefulness and practicality of the proposed method, an illustrative example is given to
illustrate the proposed models and the effectiveness of the proposed algorithm. The results showed that
the proposed method is effective for improving reliability and reducing cost.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction represents an industrial process that reuses resources and in many

cases a means to business benefits (Jiang et al., 2011). The practice
Rapid technological innovation and changes in the global mar- gives products of equal quality to conventional manufacturing for
ket accelerate the speed of products replacement, causing an reduction up to 60% energy, 70% materials, 50% cost, 80% air
exponential increase in the production of waste, and saturating pollution and is considered to be an important part of circular
landfills (Bulmus et al., 2014). Remanufacture adds value to waste economy.
streams by returning them to working order rather than reducing Remanufacturing process involves complete disassembly of the
them to their raw material value only (Karamouzian et al., 2011). cores, followed by cleaning, reconditioning or recovery processes
Further, re-integrating the used products into the production pro- like repair, rework, refurbish and replacement and reassembly and
cess reduces disposal (and hence landfill), processing, and virgin inspection (Lee et al., 2010). The prices of the remanufactured
material extraction (Li et al., 2013; Gehin et al., 2008; Tian et al., items are generally 30e40% of that of new ones (Mukherjee and
2014). Mondal, 2009). Thus, remanufacturing not only provides eco-
Remanufacture returns used products to an “as good as new” nomic benefits for customers, but also utilizes minimum resource
functional state with warranty to match, thus releasing residual and reduce waste disposal (Sundin and Bras, 2005). Despite the
value in these components (Guide et al., 1999). Remanufacturing tremendous advantages offerred by remanufacturing, customers's
perception of remanufactured products are very low as it is often
believed that the quality of remanufactured goods is inferior to
* Corresponding author. that of new ones and this in turn affects their willingness to pay
E-mail addresses: [email protected], [email protected] (Z. Jiang).

http://dx.doi.org/10.1016/j.jclepro.2015.11.037
0959-6526/© 2015 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037
2 Z. Jiang et al. / Journal of Cleaner Production xxx (2016) 1e9

for these products (Hazen et al., 2012). Since quality is a broad perform remanufacturing operations, the parameters of these op-
term embracing availability and reliability aspects, which have erations and the order in which they will be executed (Kin et al.,
caused many issues in remanufacturing, especially in reliability of 2014).
remanufactured products (Du and Cao, 2014). Thus, overcoming In addition to the reliability, the cost in remanufacturing is
challenge of reliability is one of the major concerns for prolifer- another criterion that should be taken into consideration for pro-
ation of remanufacturing in the market. The reliability of rema- cess planning. High initial investment to build up the infrastructure
nufacturing requires consideration of several factors like quality and equipment for carrying out remanufacturing is required when
issues for returned products or components, and operating pro- starting a remanufacturing business. Remanufacturers can go for
cesses such as disassembly and reassembly. These factors deeply this only if they can make profits from the remanufacturing activ-
affect to some the extent to the reliability of remanufactured ities (Sabharwal and Garg, 2013). Thus, another key challenge of
products (Ortegon et al., 2013). While not necessarily a technical remanufacturing process planning is to optimize a process plan by
issue, the success of a remanufacturing business is yet very guaranteeing the reliability of remanufacturing whist keeping the
dependent on the acquisition of high quality used products or cost at minimum.
cores in order to satisfy the reliability for remanufactured prod- With respects to the goal of improving reliability and reducing
ucts (Jiang et al., 2011). Thus, developing a product recovery sys- process cost, this paper presents a novel optimization method for
tem and estimating influence of the reverse logistics network to remanufacturing process planning integrating these two objectives
the reliability of remanufacturing is an initial step in remanu- for the first time to the best knowledge of the authors. The method
facturing process planning. Hatefi and Jolai (2014) proposed a has two principal characteristics: (1) establishing the reliability and
reliable model for reverse logistics network based on robust cost model of remanufacturing process planning by taking the
optimization method to protect the network against uncertainty. faults of returned products into consideration; and (2) genetic al-
Vahdani et al. (2012) developed a bi-objective mathematical gorithm (GA) is used to solve the optimization problem of rema-
programming formulation to construct a model for a reliable fa- nufacturing process planning. Compared to other optimization
cility network design. Lin and Yeh (2010) focused on the optimal method, GA mimics the process of natural evolution to optimize
carrier selection problem based on network reliability criterion, multi-dimensional nonlinear problems, and has been reported by
and a GA-based algorithm is developed to determine the optimal many researchers. For instance, Ismail et al. (2008) devised and
carrier selection. Once the acquisition of returned items in the employed two modified genetic algorithms to provide the best
reverse logistics network is completed, the performance of the approximate process planning solution. Fan et al. (2012) estab-
remanufacturing process including disassembly and assembly is lished a multi-objective optimization function and applied GA for
another challenge for remanufacturing reliability. Zussman and process routing decision. Wang et al. (2004) developed an opti-
Zhou (1999) proposed a mathematically sound disassembly Petri mization method for process routing by using GA to determine the
net that can be modified to deal with the reliability of resources operation sequence with consideration of the sequence constraints.
performing the operations. Song and Wang (2014) analyzed the Thus, GA is a well-recognized multi objective optimization method
disassembly process of LCD and simulated its reliability. Suzuki and it is also feasible to be used to select and order the alternative
et al. (2003) introduced an effective new design for quality tool operations in remanufacturing process planning. The most impor-
that visualizes assembly fault potential based on product reli- tant aspect in this optimization method is that it considers the
ability. Jiang et al. (2010) defined the reliability of assembly entire reliability and cost problem from the faults of returned
connection and proposed a model to assign the assembly reli- products. The incorporation of faults impacts into process planning
ability to the parts and sub-systems. Although significant amount certainly will provide more efficient solutions for remanufacturing.
of research has been conducted in reliability related to the quality
of used cores, disassembly and assembly, not much effort is re- 2. The proposed framework of remanufacturing process
ported reliability on the impact of remanufacturing process plan planning
optimization to reliability.
Remanufacturing process planning is a vital step in remanu- The goal of the proposed strategy for remanufacturing process
facturing which provide a pivotal link between used products and planning is to identify optimal remanufacturing process plan in
remanufactured products. It is focused on specifying the machines, dynamic environments. In seeking an optimal solution, it is pro-
operations, operation sequences, tools required to recovery the posed to employ a multi-criteria optimization method to achieve
used products as good as new ones (Yin et al., 2014). Due to its the goal of improving reliability and reducing cost of remanu-
significant important of remanufacturing process planning, a great facturing process. The proposed optimization method is divided
deal of research related to it has been undertaken. Kernbaum et al. into three sequential phrases: (1) characteristic framework of
(2009) presented a mixed integer programming approach to opti- remanufacturing process planning; (2) evaluation criterion for
mize remanufacturing process plan in terms of cleaning, injecting, remanufacturing process planning; (3) optimization model of
disassembly, repairing and reassembly. Parkinson and Thompson remanufacturing process planning based on GA.
(2004) put out a systematic approach for the planning and execu-
tion of product remanufacturing based upon the failure mode and 2.1. Characteristic framework of remanufacturing process planning
effect analysis method. Li et al. (2011) adopted a Graphical Evalu-
ation and Review Technique method for remanufacturing process In remanufacturing process planning, the returned (cores) are
plan decisions, which considered the uncertainty in the quality of first disassembled into individual parts the quality of which are
incoming used components. Li et al. (2013) established remanu- subsequently inspected and evaluated. Since the parts have
facturing process model based on a fuzzy Petri net in view of the different quality levels which are infected by various damages
uncertainties in remanufacturing process routing and process time. (wear, corrosion, fatigue, etc.) at different level (slight, medium and
Undoubtedly, these studies provided a very useful reference for serious) and varying location. In responding to the type and level of
remanufacturing process planning. However, there are relatively damage of the parts that needs recovery, a variety of operations,
few studies considering the reliability of remanufacturing due to e.g., material addition (welding, thermal spraying, etc.), material
the uncertainty of returned used products. High reliability of removal (machining, laser cutting, etc.), and surface treating (heat
remanufacturing is a vital target for the process planning to treating, anodizing, etc.) are employed. For instance, if a used lathe

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037
 Z. Jiang et al. / Journal of Cleaner Production xxx (2016) 1e9 3

guide is in slight wear condition, chrome plating may be used for Table 1
the recovery whist moderate wear condition, machining Quantitative values assigned to attributes.

operations may be employed. The relationship between faults and Qualitative measure of attributes Assigned value of aij
its corresponding recovering process operations is depicted in Exceptionally low 1
Fig. 1. Extremely low 2
In fact, the faults of the cores strongly affects the type, sequence Very low 3
and time of remanufacturing processes, resulting in various Below average 4
Average 5
remanufacturing process plans, which in turn have great influences
Above average 6
on reliability and cost. Reliability of remanufacturing process is the High 7
ability to accomplish recovery mission of returned products in the Very high 8
specified processing time and condition (Zhang and Liang, 2013). Extremely high 9
Exceptionally high 10
The processing time and condition are mainly determined by the
degradation failure rate of machine (lathe, tools, etc.). However, the
machining resources decision are subject to process planning,
which cannot be specified until operation sequencing for recovery
the returned cores is performed (Huang et al., 2012). In addition, Rk ðtÞ ¼ elk ðtÞ$t (1)
the cost for each process plan is different due to the alternative
operators and machines which affected by the returned cores in Where lk(t) is the failure rate for processing resources in time t.
different failure attribution. For the faults of returned cores can range from minor scratches
to extensive damage, the quality deviations of returned compo-
nents may result in the increasing depletion of machines and tools.
2.2. Evaluation criterion of remanufacturing process planning Thus, lk(t) can be formulated as follows:

Remanufacturing process planning has different consideration lk ðtÞ ¼ l0 ðtÞ þ aðm0  ms Þ2 (2)
for different objects. For a used core, a process plan provides
detailed information about the selection of remanufacturing op- Where l0(t) is the initial failure rate, m0 is the evaluated quality of
erations and resources to guarantee the reliability of remanufac- the returned cores, ms is the desire quality of the returned cores for
tured parts. For a remanufacturer, a process plan is a method to the remanufacturer, and a is the correction factor.
return a used product to an as-good-as condition at a price that is In order to calculate the reliability of remanufacturing process,
significantly less than a new product. Thus, the criterion of max the quality evaluation of returned cores may be conducted firstly.
reliability and minimum production cost is adopted for remanu- The quality of returned components is affected by the faults
facturing process planning. F ¼ {F1,F2,/,Fi,/,Fn}. The type of faults can be classified as the inside
defect, dimension, surface roughness, geometric tolerance and
non-working surface damage and so on (Jones et al., 1997; Zhou
2.2.1. Reliability model et al., 2012). These faults can present different quality levels by
For the remanufactured products, reliability is a key concern of the types, degree and location of damage. The influence of the each
customers. The remanufactured products must be of high reliability faulty on the reliability is presented by a method of weight. To
and quality to meet the same specifications of new products in the obtain the weight of each fault, entropy weight method is used. The
market. main steps of the entropy weight method include the formation
As the reliability model of remanufacturing process is a series and standardization of the evaluation matrix, and the calculation of
system, the processing resources (machine tools, cutting tools, etc) the entropy and the entropy weight.
of the remanufacturing process system are sequential and the
failure of each process is independent (Tavakkoli-Moghaddam a. The formation of the evaluation matrix. The number of faults is
et al., 2008). The failure rate of procedure obeys index distribu- n, the number of marking criteria is m, and then a matrix
tion (He and Shen, 2014). Thus the reliability for the procedure k A ¼ (aij)mn, aij is the influence value of the ith type of fault on the
can be calculated as follows: marking criteria j to the quality of returned cores. The degree of

Fig. 1. The relationship framework of different cores with different faults and the corresponding remanufacturing process operations.

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037
4 Z. Jiang et al. / Journal of Cleaner Production xxx (2016) 1e9

damage can be represented quantitatively by a number between in the process plan, the expense of machine tools and cutting tools
1 and 10 as shown in Table 1. Among these values, the bigger it is a key consideration in this work.
is, the more influence it is on the quality. If the influence value is
10, the quality of returned core is extremely low, and may not be (1) The total cost of machine (TMC) in remanufacturing process
suitable for remanufacturing. can be written as follows:
b. The standardization of the evaluation matrix. The
matrixA ¼ (aij)mn, is normalized to form a new matrix
X
N
P ¼ (pij)mn via the Equation (3). TMC ¼ MCIi  t 0 (11)

 i¼1
max aij  aij
rij ¼

 (3)
max aij  min aij Where MCIi is the machine cost for processing a core per unit time;
and t 0 is the machine time for machine to finish the ith process.

c. The calculation of the entropy. For marking element j, the en- (2) The total cost of tool (TTC) in remanufacturing process can be
tropy value for ith type of fault is defined as follows: expressed as follows:

X
n
hj ¼ k fij ln fij (4) X
N
j¼1 TTC ¼ TCIi  t 0 (12)
i¼1
P
Where k ¼ 1/lnn, fij ¼ rij = nj¼1 rij , and when fij ¼ 0, fijlnfij ¼ 0.
Where TCIi is the tool cost for processing a core per unit time.
d. The calculation of the entropy and the entropy weight. The However, the processing time t 0 for the ith process is largely
weight for marking criterion j can be calculated as follows: depended on the quality of returned components. The different
quality between returned products and the ideal recovery com-
!
1  hj X
m ponents determine the processing time. Let the rated time for
uj ¼ P 0  uj  1; uj ¼ 1 (5) machine and tool in ith process be ti, and the relationship between
m m i¼1 hj i¼1 processing time and returned products satisfies the following
The weight of fault type uj, can be calculated based on the en- function:
tropy method, the comprehensive evaluation for the ith type of fault
and the jth marking criterion are calculated by Equations. (6) and (7) t 0 ¼ ti þ hðm0  ms Þ (13)
respectively:
Where h is a certain coefficient which can be achieved by the
X
m
experimental method. m0 is the evaluated quality of the returned
gi ¼ aij uj (6)
core, and ms is the ideal quality of the remanufactured core.
i¼1
Thus, the total cost of the remanufacturing process can be
calculated as follows.
X
n
gj ¼ aij ui (7)
j¼1 X
N
C¼ ðMCIi þ TCIi Þ$½ti þ hðm0  ms Þ (14)
The weight of the faults comprehensive impact on the quality of
i¼1
returned core is calculated as follows:

X
n X
m
m0 ¼ gi ui $ gj uj (8)
i¼1 j¼1

Similarly, the desired quality ms of returned cores can be

calculated. The reliability of the procedure k can be calculated by
Equations. (1)e(8), and is shown as follows:
 h  i
Rk ðtÞ ¼ exp  l0 ðtÞ þ E ak ðm0  ms Þ2 $t (9)

Then, the reliability of the remanufacturing process can be

calculated as follows:

Y
N
Rk ðtÞ ¼ Rk ðtÞ (10)
k¼1

Where N is the total number of remanufacturing process operation.

2.2.2. Cost model

The total cost of remanufacturing process planning contains the
expense of machine tools and cutting tools, the cost for machine
tools and cutting tools transformed, labor cost and so on. In order to
illuminate the influence from the faults of returned products to cost Fig. 2. The optimization steps of remanufacturing process planning.

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037
 Z. Jiang et al. / Journal of Cleaner Production xxx (2016) 1e9 5

2.3. The optimization model for remanufacturing process planning

based on GA

The proposed remanufacturing process planning method is

divided into three steps, as shown in Fig. 2: failure feature identi-
fication in which the fault set is identified by analyzing the returned
components, operation generation where feasible operations is
identified, and process plan generation based on GA, in which
multi-objectives (reliability and cost) is used to search for the
process plan for the maximization fitness function value.

Step 1: Identification of faults of the used component.

Major faults in the components are identified to obtain the fault

set (f1,f2,f3,/,fn) which is mapped with remanufacturing operations
based on the type, degree and location of damages.

Step 2: Determination of the alternative remanufacturing

operations.

According to the previous definition of the fault, a part can be

decommissioned due to a set of faults. If these attributions can be
recovered by sequence of operation, the part remanufacturing
would be fully completed. Based on the characteristics of ith failure
attribution, a set of alternative remanufacturing operation set
(opi1,opi2,/,opim) can be determined.

Step 3: Generate a process plan using GA.

Owing to the faults of used components and the various choices

for process, remanufacturing process plan is a NP-hard (Non-
deterministic polynomial hard) problem with the objective to
improve the reliability of remanufactured products and reduce the
process cost. A process plan will be generated based on GA that Fig. 3. Algorithm flow of initial population.
utilizes a function with multiple objectives.
A representation scheme based on the faults of cores, with
2.3.3. Fitness function
corresponding genetic operations and fitness evaluation functions
The fitness function is used to select the best individuals. The
is described in the following subsections.
remanufacturing process planning optimization problem carries
out two objective functions. The objective optimization of rema-
2.3.1. Encoding nufacturing process planning is to maximize reliability R(x) and
The generation of remanufacturing process and the search minimize cost C(x). Assuming that, at the kth generation, across the
ability for GA is deeply influenced by solution encoding. In GA, entire population of solution points, find the minimum cost Cppmin
chromosomes is very important for the representation of the and maximum reliability Rppmax, and the reliability and cost for
attribution of remanufacturing process plan. In this paper, the chromosome j are R(x)j and C(x)j. A selection function for chro-
attribution of remanufacturing process includes the faults of a used mosome is formulated as follows:
component, the operations selected to recover faults and selection
of machines and tools for each operation, thus the gene encoding is f1 ðxÞj ¼ Rppmax  RðxÞj ; RðxÞj < Rppmax (15)
shown in Table 2.
f2 ðxÞj ¼ CðxÞj  Cppmin ; CðxÞj > Cppmin (16)
2.3.2. Initial population
The fitness value for chromosome j can be calculated as follows:
After encoding, the initial population is set using random solu-
tions to search for any solution in a search space. However, the FðxÞ ¼ l1 f1 ðxÞj þ l2 f2 ðxÞj (17)
process plan cannot be generated randomly due to some con-
straints, e.g., the degree of the faults, the operation sequence. When
Where l1,l22(0,1), l1þl2 ¼ 1.
generating the candidate process plan, the constraints should be
The fitness function can be formulated as follows:
met, as shown in Fig. 3.
gðxÞ ¼ minFðxÞ (18)
Table 2
The gene encode.

Fault number Operation number Machine number Tool number Gene

2.3.4. Genetic operators
1 Crack 1 Cold welding M1 1110 The design of appropriate genetic operators, including the
1 Crack 2 Grinding M2 T2 1222 reproduction, crossover, and mutation, plays a crucial role in the
2 Wear 3 Grinding M2 T1 2321
successful implementation of GA (Sardinas et al., 2006).

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037
6 Z. Jiang et al. / Journal of Cleaner Production xxx (2016) 1e9

Table 3 To illustrate the proposed method, the remanufacturing process

Two viable chromosomes of remanufacturing process. planning for a used C6132 lathe bed is presented.
2,3,2,1 2,5,3,0 2,6,2,0 1,1,1,0 1,2,2,3 3,7,2,1 3,8,4,1 3,9,2,2
Chromosome one
2,3,2,2 2,4,3,0 3,10,2,3 3,11,1,0 3,9,2,3 1,1,1,0 1,2,2,4 0,0,0,0
3.1. Faults inspection of lathe bed
Chromosome two
According to real circumstances, the repairable components of
C6132 lathe bed include guide, saddle and spindle. The lathe bed is
Table 4 fault product because of some defects in guide, saddle and spindle.
An offspring child chromosome. The location, degree and type of these defects make it possible to
have different remanufacturing operations. Thus, inspection is the
2,3,2,1 2,5,3,0 3,10,2,3 3,11,1,0 3,9,2,3 1,1,1,0 3,8,4,1 3,9,2,2
first step to assess the quality of the returned cores. The inspection
results of guide, saddle and spindle are shown in Table 5.

Reproduction is used to guarantee the selection of optimal indi-

vidual. In this operator, “elitism” strategy is applied by copping the 3.2. Determination of alternative operations
best individual of the population (the one with highest fitness
value) to the next generation and then the “roulette wheel” method From the defect inspection for the lathe bed in Table 5, crack,
is applied to the remain strings. Crossover contains reproduction wear and corrosion are the main faults to be considered. Based on
procedure from parent's chromosome, which is essential in GA. the remanufacturing requirements, the alternative operations to
Crossover is applied, at a given probability, to create new genera- recovery the faults are shown in Table 6. The available machine and
tions resulting from two parent strings exchanged randomly be- available tool information related to the operators are given in
tween the two crossover points. For example, two strings are Tables 7 and 8. Based on international field study, the process in-
selected in Table 3. There are two crossover sites: x ¼ 3, y ¼ 7. An formation for lathe bed in this remanufacturing enterprise is shown
offspring child chromosome is generated based on crossover in Table 9.
operation, as shown in Table 4. The mutation operator makes
random changes to one or more elements of the string. Mutation 3.3. Quality evaluation for the lathe bed
operator randomly modifies two or more elements to obtain a new
string with a value between 0.001 and 0.1. If the string is feasible, it The alternative recovery operation will be selected based on the
will appear in future generations. reliability and cost model by using genetic algorithm. However, the
Based on the above method, the optimal remanufacturing pro- reliability and cost of remanufacturing process are influenced by
cess plan will be obtained. The process planning optimization the quality of used components and the machine that is used to
method in this paper is obtained based on the reliability and cost perform the recovery operations. The quality of lathe bed and the
functions. The precious data about the reliability and cost of process desire quality of returned lathe bed are evaluated by entropy
impacted by the quality of used cores can be obtained by the weight method, and the scores are in Tables 10 and 11 which can be
function which can guide remanufacturer in making better obtained from Table 1 based on the evaluation of experts.
decisions. The number of quality factor is 3, and the number of marking
criteria is 3, i.e., damage, degree, location, and then the evaluation
3. Case study matrix A ¼ (aij)33 can be obtained. The evaluation matrix
A ¼ (aij)33 is normalized to obtain a matrix P ¼ (pij)33. The results
This section describes the reliability and cost optimization of of weight calculation for faults and marking criteria are shown in
remanufacturing process planning for a lathe bed based on GA. The Table 12.
lathe bed, as a typical electromechanical product, is of great po- For the lathe bed remanufacturing, the correction factor
tential for remanufacturing. A remanufactured lathe may cost only a ¼ 0.001, h ¼ 0.015. Then, the quality for lathe bed, m0, and the
40%e60% of that of a new lathe whist offering better machining desired quality, ms, can be calculated by Equation. (7), m0 ¼ 14.38,
accuracy and production efficient (Du et al., 2013; Zhou et al., 2014). ms ¼ 7.24.

Table 5
The inspection results of lathe bed.

Components Guide Saddle Spindle

Faults' Type Inside defect Surface roughness Non-working surface damage

Damage Crack Wear Corrosion

Location On the surface On the surface Outside the surface
Inspect result 0.05 mm 1.2 mm 0.2 mm depth
Degree Slight Medium Medium

Table 6
The alternative operators of lathe bed.

Faults Op1 Op2 Op3 Process

Crack Cold welding 1 Grinding 2 1e2

Wear Grinding 3 Electrodeposited chromium 4 3e4
Laser cladding 5 Fine grinding 6 3-5-6
Corrosion Milling 7 Thermal spraying 8 Milling 9 7-8-9
Grinding 10 Cold welding 11 10-11-9

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037
 Z. Jiang et al. / Journal of Cleaner Production xxx (2016) 1e9 7

Table 7 Table 10
Available machine information. The score for returned lathe bed.

Number Machine type MCI Type of faults Damage Degree Location

M1 Electrospark welding machine 10 Inside defect 9 4 2

M2 Guideway grinder 40 Surface roughness 4 8 1
M3 Chrome-plating equipment 35 Non-working surface damage 3 5 3
M4 CNC vertical miller 25
M5 Vertical miller 20
M6 Semiconductor laser cladding equipment 40
M7 Plasma spraying machine 50 Table 11
M8 Arc thermal spraying equipment 30 The score for desire returned lathe bed.

Type of faults Damage Degree Location

Inside defect 3 2 2
Table 8 Surface roughness 3 5 1
Available tool information. Non-working surface damage 2 3 3
Number Tool type TCI

T1 Welding tool 8
T2 Grinding wheel 20 h i
T3 Guideway grinder sharpener 15 FðxÞ ¼ min 0:65f1 ðxÞj þ 0:35f2 ðxÞj (19)
T4 Milling tool 1 10
T5 Milling tool 2 20
T6 Milling tool 3 10

3.4. Generation of the process plan based on GA 3.4.4. Genetic operators

The genetic algorithm operator of the chromosome reproduc-
3.4.1. Encoding tion, crossover, and mutation is used to generate optimal rema-
Based on the above information, the gene can be coded as nufacturing process plan. The parameters for GA are set as follows:
(1,1,1,1), which represents that the fault of crack (1) is recovered by iteration number is 200, initial population size N is 100, and
cold welding (1). The selected machine and tool related to the cold crossover and mutation probabilities are 0.85 and 0.10, respectively.
welding are electrospark welding machine (1) and welding tool (1). After finishing the optimization steps, the optimal operation
sequence can be obtained as shown in Table 13, and the informa-
tion for this process plan is shown in Table 14. It is worth to noting
3.4.2. Initial population
that the process plan selected for the lathe bed is incredibly close to
The lathe bed contains three components, so the operations are
the optimal solution and appropriate for the given remanufacturing
divided into three groups which are (1,2) (3,4) or (3,5,6) (7,8,9) or
factory. The used lathe before remanufacturing is shown in Fig. 5.
(10,11,9). For the each group, the operations sequence cannot be
The guide (Part model: C6132D, Size: 2300 mm  490 mm) is
changed, for example, the operation 2 must follow the operation 1
remanufactured through cold welding and grinding. Comparison
in group (1,2). This relationship of procedures in remanufacturing
chart between before and after repair for the lathe guide is shown
process constrains the initial population generated randomly.
in Fig. 4. Saddle (Part model:C6132A1, Size:645 mm  615 mm) is
remanufactured using grinding, laser cladding and fine grinding to
3.4.3. Fitness function recover wear. Spindle is remanufactured using the milling, grinding
Fitness function is an evaluation for the represented solution in and cold welding. The used lathe is restored into a new remanu-
remanufacturing process planning optimization, which has been factured lathe with the similar even better performance, as shown
calculated in Eq. (17). However, the weight distribution for reli- in Fig. 6.
ability and cost is variable from different remanufacturing enter- To reveal the performance of the optimization model, the lathe
prises constrained by their capability in production and technology. beds remanufactured by two different process plans are compared
For the case study the main focus for enterprise is to improve the with that of new ones as shown in Table 15. Scheme 1 is the used
reliability. Thus, the weight of reliability (f1) is 0.65, and the weight C6132 lathe bed remanufactured by the process plan which
of cost (f2) is 0.35, the fitness function is calculated as bellow: generated with the optimization model. Scheme 2 is the used

Table 9
The alternative machines, cutting tools, the initial failure rate and process time of each process stage for lathe bed remanufacturing.

Fault Operation l0  103 M&T Time/min Fault Operation l0  103 M&T Time/min

1 Crack 1 Cold welding 4.0 M1T1 9.4 3 Corrosion 7 Milling 3.6 M4T4 13.2
2 Grinding 3.0 M2T2 5.8 M4T5 10.8
M2T3 7.2 M5T5 14.3
2 Wear 3 Grinding 2.5 M2T2 10.4 M5T6 17.2
M2T3 9.8 8 Thermal spraying 2.8 M7 15.9
4 Eletrodeposited chromium 5.0 M3 4.6 M8 12.8
5 Laser cladding 7.0 M6 3.9 9 Milling 6.0 M4T4 15.5
M7 5.8 M4T5 13.6
6 Fine Grinding 3.8 M2T2 11.7 M5T5 22.3
M2T3 9.5 M5T6 16.8
M4T4 6.8 10 Grinding 4.5 M2T2 14.7
M4T5 7.6 M2T3 12.8
11 Cold welding 3.4 M1T1 8.9

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037
8 Z. Jiang et al. / Journal of Cleaner Production xxx (2016) 1e9

Table 12
The weight of quality factor and marking criteria.

Weight of faults Weight of marking criteria

Type
Inside defect Surface roughness Non-working surface damage Damage Degree Location

Returned lathe bed 0.23 0.23 0.43 0.27 0.48 0.25

Desire returned lathe bed 0.39 0.34 0.27 0.40 0.32 0.27

Table 13
The optimal operation sequence.

1,1,1,1 1,2,2,2 3,9,5,6 3,10,2,3 3,11,1,1 2,3,2,3 2,5,6,0 2,6,4,4

Table 14
The information of the optimal remanufacturing process plan.

Order Component Damage Operation Machine Tool

1 Guide Crack 1 Cold welding M1 T1

2 Crack 2 Grinding M2 T2
3 Spindle Corrosion 9 Milling M5 T6
4 Corrosion 10 Grinding M2 T3
5 Corrosion 11 Cold welding M1 T1
6 Saddle Wear 3 Grinding M2 T3
7 Wear 5 Laser cladding M6
8 Wear 6 Fine Grinding M4 T4

C6132 lathe bed remanufactured by traditional process plan

without considering the reliability. Fig. 5. The used Lathe before remanufacturing.
Obviously, the optimization model has demonstrated significant
improvement in terms of reliability and cost of the used C6132 lathe
bed remanufacturing. The proposed method can be used to obtain
the optimal process route for remanufacturing.

4. Conclusion

This work presented a dual-objective remanufacturing process

optimization method to maximize the reliability and minimize
process cost for the first time. The critical optimization objective of
reliability enhancement in the model is evaluated by the failure rate
of process and capability decay of machines and tools which are
affected by the quality of returned products. The classical optimi-
zation objective in remanufacturing process such as cost is also
integrated in the model. For this multi-objective optimization
problem, genetic algorithm is used to solve the model. The case
study with remanufacturing process planning optimization for the
lathe bed is presented to evaluate the feasibility of the model. The
results showed that it is more efficient for remanufacturing to
improve the reliability and reduce process cost by integrating the Fig. 6. The remanufactured new lathe.

Fig. 4. Comparison chart between before and after repair for the lathe guide.

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037
 Z. Jiang et al. / Journal of Cleaner Production xxx (2016) 1e9 9

Table 15
Comparison of the remanufactured lathe with the standard of new lathe.

Items Accuracy value New C6132 lathe bed Remanufactured C6132 lathe bed by scheme 1 Remanufactured C6132 lathe bed by scheme 2

Roundness (mm) 0.0085 0.0082 0.0080

Flatness (mm) 0.012 0.010 0.008
Pitch error 0.030 0.028 0.026
Surface hardness 55 52 50
Surface parallelism 0.030 0.026 0.025
Repeatability of positioning X axis 0.015 0.012 0.010
from feed (mm) Y axis 0.020 0.018 0.015
Cost 20,000 6500 8000

quality evaluation of returned products to establish a remanu- Jiang, Z.L., Xuanyuan, S.S., Li, Z.Q., Meng, X.X., 2010. Optimization of product as-
sembly relations based on connection reliability. Key Eng. Mater. 431, 289e292.
facturing process model.
Jones, G., Willett, P., Glen, R.C., Taylor, R., 1997. Development and validation of a
The work presented herein not only provides a tool to optimize genetic algorithm for flexible docking. J. Mol. Biol. 267 (3), 727e748.
the remanufacturing process planning, but it also provides decision Karamouzian, A., Teimoury, E., Modarres, M., 2011. A model for admission control of
support for high reliability remanufacturing. It may be used for the returned products in a remanufacturing facility using queuing theory. Int. J. Adv.
Manuf. Technol. 54 (1e4), 403e412.
selection of remanufacturing process plan that is reliable and Kernbaum, S., Heyer, S., Chiotellis, S., Seliger, G., 2009. Process planning for IT-
economical. In addition, future study may be focused on a general equipment remanufacturing. CIRP J. Manuf. Sci. Technol. 2 (1), 13e20.
framework model integrated the environmental issue in remanu- Kin, S.T.M., Ong, S.K., Nee, A.Y.C., 2014. Remanufacturing process planning. Procedia
CIRP 15, 189e194.
facturing process planning optimization. The research work in this Lee, H.B., Cho, N.W., Hong, Y.S., 2010. A hierarchical end-of-life decision model for
area is expected to be one of the key areas of remanufacturing determining the economic levels of remanufacturing and disassembly under
research in the near future. environmental regulations. J. Clean. Prod. 18, 1276e1283.
Li, C.B., Tang, Y., Li, C.C., Li, L.L., 2013. A modeling approach to analyze variability of
remanufacturing process routing. Autom. Sci. Eng. IEEE Trans. 10 (1), 86e98.
Acknowledgments Lin, Y.K., Yeh, C.T., 2010. Optimal carrier selection based on network reliability
criterion for stochastic logistics networks. Int. J. Prod. Econ. 128 (2), 510e517.
Li, C.B., Tang, Y., Li, C.C., 2011. A GERT-based analytical method for remanufacturing
The work described in this paper was supported by the National process routing. In: Automation Science and Engineering (CASE), IEEE Confer-
Natural Science Foundation of China (51205295, 71471143, ence on. IEEE, pp. 462e467.
Mukherjee, K., Mondal, S., 2009. Analysis of issues relating to remanufacturing
51405075), Wuhan Youth Chenguang Program of Science and technologyea case of an Indian company. Technol. Anal. Strateg. Manag. 21 (5),
Technology(2014070404010214), A1202 supported by Science 639e652.
Foundation of Wuhan University of Science and Technology, and Ortegon, K., Nies, L.F., Sutherland, J.W., 2013. Preparing for end of service life of
wind turbines. J. Clean. Prod. 39, 191e199.
Funds for International Cooperation and Exchange of the National Parkinson, H.J., Thompson, G., 2004. Systematic approach to the planning and
Natural Science Foundation of China (51561125002). These finan- execution of product remanufacture. Proc. Inst. Mech. Eng. Part E J. Process
cial contributions are gratefully acknowledged. Mech. Eng. 218 (1), 1e13.
Sabharwal, S., Garg, S., 2013. Determining cost effectiveness index of remanu-
facturing: a graph theoretic approach. Int. J. Prod. Econ. 144 (2), 521e532.
References Sardinas, R.Q., Santana, M.R., Brindis, E.A., 2006. Genetic algorithm-based multi-
objective optimization of cutting parameters in turning processes. Eng. Appl.
Artif. Intell. 19 (2), 127e133.
Bulmus, S.C., Zhu, S.X., Teunter, R., 2014. Competition for cores in remanufacturing.
Song, S., Wang, Y., 2014. Research on design and reliability of LCD easy to disas-
Eur. J. Oper. Res. 233 (1), 105e113. semble structure. Mach. Des. Manuf. 7, 252e255.
Du, Y.B., Cao, H.J., Chen, X., Wang, B., 2013. Reuse-oriented redesign method of used
Suzuki, T., Ohashi, T., Asano, M., Arai, T., 2003. Assembly reliability evaluation
products based on axiomatic design theory and QFD. J. Clean. Prod. 39, 79e86. method (AREM). CIRP Ann.-Manuf. Technol. 52 (1), 9e12.
Du, Y.B., Cao, H.J., 2014. Overview and perspectives on reliability of machinery Sundin, E., Bras, B., 2005. Making functional sales environmentally and economi-
equipment remanufacturing. Comput. Integr. Manuf. Syst. 20 (11), 2643e2651.
cally beneficial through product remanufacturing. J. Clean. Prod. 13 (9),
Fan, S., Wang, J., Li, S., 2012. Decision and optimization of process routing based on 913e925.
genetic algorithm. Manuf. Technol. Mach. Tool 3, 031.
Tavakkoli-Moghaddam, R., Sassani, F., Safari, J., 2008. Reliability optimization of
Gehin, A., Zwolinski, P., Brissaud, D., 2008. A tool to implement sustainable end-of series-parallel systems with a choice of redundancy strategies using a genetic
-life strategies in the product development phase. J. Clean. Prod. 16, 566e576. algorithm. Reliab. Eng. Syst. Saf. 93 (4), 550e556.
Guide, V.D.R., Jayaraman, V., Srivastava, R., 1999. Production planning and control Tian, G., Chu, J., Hu, H., Li, H., 2014. Technology innovation system and its integrated
for remanufacturing: a state-of-the-art survey. Robot. Comput.-Integr. Manuf. structure for automotive components remanufacturing industry development
15 (3), 221e230.
in China. J. Clean. Prod. 85, 419e432.
Hazen, B.T., Overstreet, R.E., Jones-Farmer, L.A., Field, H.S., 2012. The role of ambi- Vahdani, B., Tavakkoli-Moghaddam, R., Modarres, M., Baboli, A., 2012. Reliable
guity tolerance in consumer perception of remanufactured products. Int. J. Prod. design of a forward/reverse logistics network under uncertainty: a robust-M/M/
Econ. 135 (2), 781e790. c queuing model. Transp. Res. Part E Logist. Transp. Rev. 48 (6), 1152e1168.
Hatefi, S.M., Jolai, F., 2014. Robust and reliable forwardereverse logistics network Wang, Z., Wang, N., Chen, Y., 2004. Optimization of process routing based on the
design under demand uncertainty and facility disruptions. Appl. Math. Model.
genetic algorithm. J. Tsinghua Univ. 44 (7), 987e992.
38 (9), 2630e2647. Yin, R., Cao, H., Li, H., Sutherland, J.W., 2014. A process planning method for reduced
He, Y., Shen, Z., 2014. Reliability analysis modeling of manufacturing systems based
carbon emissions. Int. J. Comput. Integr. Manuf. 27 (12), 1175e1186.
on process quality data. J. Beijing Univ. Aeronaut. Astronaut. 40 (8), 1027e1032. Zhang, J., Liang, G.Q., 2013. Reliability analysis of remanufacturing production sys-
Huang, W., Hu, Y., Cai, L., 2012. An effective hybrid graph and genetic algorithm tem under part quality uncertainty. Ind. Eng. J. 16 (2), 117e121.
approach to process planning optimization for prismatic parts. Int. J. Adv.
Zhou, J., Huang, P., Zhu, Y., Deng, J., 2012. A quality evaluation model of reuse parts
Manuf. Technol. 62 (9e12), 1219e1233. and its management system development for end-of-life wheel loaders.
Ismail, N., Musharavati, F., Hamouda, A.M.S., Ramli, A.R., 2008. Manufacturing
J. Clean. Prod. 2012 (35), 239e249.
process planning optimisation in reconfigurable multiple parts flow lines. Zhou, F., Jiang, Z.G., Zhang, H., Wang, Y., 2014. A Case-based Reasoning Method for
J. Achiev. Mater. Manuf. Eng. 37 (2), 671e677. Remanufacturing Process Planning. Discrete Dynamics in Nature and Society.
Jiang, Z.G., Zhang, H., Sutherland, J.W., 2011. Development of multi-criteria decision Zussman, E., Zhou, M., 1999. A methodology for modeling and adaptive planning of
making model for remanufacturing technology portfolio selection. J. Clean. disassembly processes. Robot. Autom. IEEE Trans. 15 (1), 190e194.
Prod. 19 (17), 1939e1945.

Please cite this article in press as: Jiang, Z., et al., Reliability and cost optimization for remanufacturing process planning, Journal of Cleaner
Production (2016), http://dx.doi.org/10.1016/j.jclepro.2015.11.037