Applied Energy 275 (2020) 115328

Contents lists available at ScienceDirect

Applied Energy
journal homepage: www.elsevier.com/locate/apenergy

Optimal design of electric vehicle battery recycling network – From the T

perspective of electric vehicle manufacturers
⁎
Lei Wanga,b, , Xiang Wanga,c, Wenxian Yangd
a
State Key Laboratory of Power Transmission Equipment & System Security and New Technologies, Chongqing University, Chongqing, China
b
College of Automation, Chongqing University, Chongqing, China
c
College of Mechanical Engineering, Chongqing University, Chongqing, China
d
School of Engineering, Newcastle University, Newcastle upon Tyne NE1 7RU, UK

H I GH L IG H T S

• ATherecycling network of Lithium-Ion batteries considering CO emission is proposed.

2

• The factors affecting the cost and location optimization are analyzed.
• carbon tax can influence the optimal network configuration.

A R T I C LE I N FO A B S T R A C T

Keywords: Driven by the global campaign against climate change, the market of electric vehicles has boomed across the
Electric vehicle world in recent years. Since Lithium-Ion batteries are commonly used to power electric vehicles, a huge amount
Battery recycling of batteries will soon reach their end-of-life; how to recycle them to reduce environmental pollution and pro-
Carbon emission moting the sustainable development of the electric vehicle market has become an urgent challenge today.
Network design
Implementation of the secondary reuse of used electric vehicle batteries is a valuable recycling strategy.
However, there is a lack of research investigating electric vehicle batteries recycling network design at the
enterprise level, which impedes the sustainable development of electric vehicles. Driven by this, this paper
developed a model considering carbon emission for simulating the recycling of electric vehicle batteries. The
proposed model takes into account three potential battery handling strategies (recycling, remanufacturing, and
disposal) to processing used vehicle battery cells of different quality levels at different centers. A real case study
from a Chinese electric vehicle manufacturer is presented, wherein a 5.7% decrease in total cost and a 21.8%
reduction in CO2 emission can be achieved. Moreover, the results of the scenario analysis show that transpor-
tation costs, carbon tax, and the number of used batteries, which can change both the configuration of the
network, have been identified as three major factors affecting the optimal design of recycling networks. In
addition, developing more economical recycling technology for electric vehicle manufacturers to further reduce
the total cost of the recycling process is the main direction. In all, this research will provide foreign researchers
with a perspective on Chinese companies in terms of electric vehicle battery recycling at the enterprise level, and
promote economically and environmentally sustainable development in the electric vehicle battery industry.

1. Introduction 5 million, which is almost 4 times the amount in 2015[2].Accom-

panying the large-scale adoption of electric vehicles, EV batteries
In recent years, the growing concern with environment and climate (EVBs), which are a key component of EV, will be utilized in the next
has promoted the fast development and deployment of Electric Vehicles few years. At present, there are four main EVB technologies in use: lead-
(EVs). According to the International Energy Agency, the EV market is acid, nickel-metal hydride, lithium-ion (Li-ion), and nickel-nickel
expected to expand to about 18,000,000 vehicles by the year 2020[1]. chloride [3]. Among them, lithium-ion batteries (LIBs) have unques-
And particularly in China, the accumulated sales of EVs are projected at tionable advantages in terms of raw material manufacturing, cost,

⁎
Corresponding author at: State Key Laboratory of Power Transmission Equipment & System Security and New Technologies, Chongqing University, Chongqing,
China.
E-mail address: [email protected] (L. Wang).

https://doi.org/10.1016/j.apenergy.2020.115328
Received 22 February 2020; Received in revised form 20 May 2020; Accepted 2 June 2020
0306-2619/ © 2020 Published by Elsevier Ltd.
L. Wang, et al. Applied Energy 275 (2020) 115328

extended cycle life, nickel metal and high specific capacity, so they 6.62% compared to manufacturing batteries from raw materials; Hao
have become the most widely used EVB in the market [4].Due to the et al [26] analyzed electric vehicle (including EVB) recycling tech-
booming EV market, the global Compound Annual Growth Rate of Li- nology in China. The analysis results show that about 10% of lifecycle
thium-ion batteries (the batteries most used as power sources in EVs) greenhouse gas emissions can be reduced by recycling. Ciez and Whi-
will grow by 36% from 2015 to 2020 [5].Along with the rapid growth tacre [27] compared the carbon emission situation of three recycling
of the EV market and the application of LIBs, a huge amount of batteries technologies (i.e. pyrometallurgical recycling, hydrometallurgical re-
will soon reach their end-of-life. According to the China Automotive cycling, and direct cathode recycling), and found that direct cathode
Technology and Research Centre, 120–170 thousand tons of used EVBs recycling has the potential to reduce emissions and is economically
will need to be processed by 2020 [6]. It is well known that all materials competitive.
used for making EVBs are extremely hazardous to both the environment In addition to the technologies adopted for disposing of used EVBs,
and human health, and are able to permeate into the ground soil and the design of recycling networks also has a significant impact on costs
thus water supplies when they are directly placed into landfills. For this or profits. A recycling network usually consists of: collection center,
reason, how to properly deal with so many used EVBs has become an disassembly center, material recycling center, and waste disposal
urgent challenge today. center. Because transportation between these centers causes costs and
Currently, information is abundantly available concerning the carbon emissions, there is no doubt that profits can be increased also by
Chevrolet Volt battery. The cost of manufacturing a new Chevrolet Volt optimizing the design of the recycling network. To reach this purpose,
battery is estimated to be $10,000 [7]. According to a report from the Kannan et al. and Subulan et al. [28,29] proposed a closed-loop supply
Argonne National Laboratory Center for Transportation, the breakdown chain network to reduce the total cost of battery recycling and to re-
of the manufacturing cost of an EV battery shows that material, labor, cover valuable material from used EVBs; Li et al [30] also proposed a
and overhead account for 80%, 10%, and 10%, respectively [8]. In similar network for remanufacturing LIBs, and found that integrating
order to lower battery costs, one potential way is to remanufacture from remanufacture into LIB supply chains can help to increase profits; Gu
used batteries, since recycling materials from used EVBs is critical to et al [31] formulated optimal pricing strategies for manufacturers and
saving on material costs. In the meantime, the increasing use of Lithium remanufacturers, and studied the relationship between the rate of re-
for EVBs is leading to the rapid depletion of Lithium, which implies an turn, the rate of sorting, and the rate of recovery to optimize total profit
urgency to reduce the consumption of raw materials by recycling in different periods.
[9,10]. In addition, EV manufacturers need to take greater responsi- In short, the recycling of car batteries can lead to more sustainable
bility in battery recycling. For example, The Ministry of Industry and EVB production, which in turn supports the mass adoption of electric
Information Technology of China (MIIT) promulgated the Provisional vehicles. However, there is not much literature relating to EV battery
Regulations on Traceability Management of Recycling and Utilization of recycling network design, and there is a lack of research investigating
Power Batteries for New Energy Vehicles in July 2018, which clearly states EVB recycling at the enterprise level, which impedes the implementa-
that EV manufacturers must provide battery recycling service and tion of the secondary reuse of used EVBs. In fact, it is crucial that dif-
promise to recycle in accordance with requirements, placing particular ferent handling strategies be adopted for different quality levels of used
emphasis on the traceability management of batteries. In response to batteries for recycling network design; these strategies include, for ex-
government regulations, it is mandatory for EV manufacturers to ac- ample, recycling raw materials, remanufacturing battery cells, and
tively get involved in building recycling systems to reuse certain com- landfill disposal. Although [28,29] have designed a closed-loop supply
ponents, thereby yielding savings in raw materials, manufacturing chain network for used batteries, unfortunately, their research rarely
costs, and energy consumption, leading to a reduction in environmental considers the differences between these used battery handling strate-
impact. Such a policy is very helpful to the conservation of raw mate- gies, and their main consideration is to recycle and reuse raw materials
rials, the lowering of manufacturing costs, and the reduction of energy from used batteries. Meanwhile, the design and optimization of re-
consumption and environmental pollution risks. cycling networks also has a significant impact on the reduction of
Recycled lithium-ion batteries from electric vehicles could provide a carbon emissions, but to date, carbon tax, as an effective energy saving
valuable secondary source of materials [11]. It is also gratifying that, and emission reduction mechanism, has not been considered in the
currently, more and more EV manufacturers have considered recycling context of EVB recycling networks.
used EV batteries. Nissan, Volkswagen, and BMW require their EV The purpose of this research is to fill this knowledge gap by de-
customers to return used batteries to licensed points or local authority veloping a mathematical model to optimize recycling networks for
battery collection schemes [12,13]. These returned used batteries are EVBs. Unlike other models [20,28,29,30,31], the model proposed in
expected to be used as home energy storage instead of other energy this work considers three potential strategies to handle used batteries,
storage equipment [14,15,16], considering the current price of lithium- including recycling, remanufacturing, and disposal (more information
ion batteries. In addition, Chevrolet has established an energy storage is given in Section 2.1). Used battery cells are classified according to
station using used EV batteries at the General Motors plant in Michigan quality, and are assigned to different processing centers in the recycling
[17]. In Europe, Tesla has begun recycling in cooperation with Umicore network. Moreover, to make the model more suitable for the develop-
[18]. And the recycling of EVBs has been studied from various per- ment of the global low-carbon economy, carbon tax is also considered,
spectives. For example, Georgimaschler et al [19] reviewed the re- to explore how the design of the EVB recycling network would be af-
cycling techniques for used batteries manufactured before 2012; Tang fected by the impact of carbon tax. The proposed recycling model is
et al [20] investigated the social-economic-environmental impacts of justified by a real case study of Chang'an Automobile (Group) Co., Ltd.
recycling used EVBs under reward-penalty mechanisms; Liu et al [21] (China). This research will bring guidance and decision-making to re-
evaluated the latest technologies for used battery recycling and utili- levant companies at the enterprise level, that is, it supports EV manu-
zation; Qiao et al [22] compared Cradle-to-Gate greenhouse gas emis- facturers in reducing environmental burdens and total costs in the re-
sions of internal combustion engine vehicles and battery electric ve- cycling network. Most importantly, this research will bring huge energy
hicles, and found that the production of EVBs causes an approximate and environmental benefits to the electric vehicle industry, while also
20% increase in greenhouse gas emissions; Hyung Kim et al [23] esti- reducing the consumption of fossil fuels, and promote economically and
mated the Cradle-to-Gate greenhouse gas emissions of an EV battery, environmentally sustainable development in the EVB industry. By
and found it accounted for 45% of total emissions. Similar conclusions conducting this research, the following three fundamental questions
are also drawn in the research reported in [24]. To identify an effective will be answered:
way to reduce carbon emissions, Xiong et al [25] found that EV battery RQ1: How to design a recycling network model to reduce costs and
recycling and remanufacturing can reduce greenhouse gas emissions by carbon dioxide emissions?

2
L. Wang, et al. Applied Energy 275 (2020) 115328

L3 cells

Remanufacturing Plant

Material
EVB Collection Center

L2 cells Waste

Recycling Plant Waste disposal center

Recycling &Remanufacturing Center

Fig. 1. EVB recycling network.

RQ2: How do different factors affect the optimal location of facil- and
ities in the recycling network? • Class 3 (L3) – capacity is lower than 60% of initial capacity.
RQ3: How to further reduce costs to promote recycling of used
batteries based on the designed recycling network model? Based on the test results, different downstream processing strategies
Accordingly, the rest of the paper is organized as follows. A math- will be adopted for dealing with different qualities of cells, i.e.
ematical model is developed and discussed in Section 2. A case study is
conducted to validate the effectiveness of the proposed model in • the L1-class cells will be transported to the remanufacturing plant
Section 3. Scenario analyses are performed in Section 4. The paper is and reassembled into new EVBs;
finally ended in Section 5 with several important conclusions and a plan • the L2-class cells will be transported to the recycling plant, where
for future research. some of the cells will be transported again to the remanufacturing
plant and reused directly as batteries in other applications. The
2. Model of recycling network other L2-class cells in the recycling plant will be dissected, and only
the reusable materials in them will be transported to the re-
2.1. Designing recycling network manufacturing plant to make new battery cells;
• the L3-class cells and the waste from the remanufacturing plant and
This section introduces a model for EVB recycling networks con- the recycling plant will be transported directly to the waste disposal
sidering carbon emissions. This model includes: collection centers, re- center.
cycling and remanufacturing center, and waste disposal centers, as seen
in Fig. 1. The proposed model will be able to determine the number of 2.2. Main modeling assumptions
each type of center constructed in the recycling network, and the
number of EVB batteries that should be transported from one center to In order to simplify the model development, the following as-
another. The objective of the model is to minimize the total cost in an sumptions are introduced:
EVB recycling logistics network, which includes transportation costs,
fixed costs, acquisition cost, cost for processing activities and environ- (1). (NiX Co y Mn1 − xy)O2 (NMC) is the most in-demand technology for
mental costs associated with carbon emissions. The environmental costs EVBs, mainly due to its low self-heating rate and high energy
are measured by monitoring greenhouse gas emissions, such as the CO2 density [32,33]. Therefore, all EVBs are assumed to be the same
generated by the transportation and disposal processes. The amount of type.
carbon emissions is measured throughout the entire EVB recycling (2). All centers have disposal capacity constraints;
cycle, from the collection center to the reuse and waste disposal facil- (3). The possible locations of centers are known in advance;
ities. As illustrated in Fig. 1, the used EVBs stored at the collection (4). Cells in EVBs will be processed with multiple potential strategies
center will be disassembled first; then their cells will be tested and (remanufacturing, recycling, and disposal) based on the cells’
categorized based on quality. Usually, the cells in a used battery can be quality; where the number of cells processed (corresponding to the
roughly classified into the following 3 classes based on their capacity: different strategies) follows Normal Distribution [17];
(5). The geographical location of each facility, and the distance be-
• Class 1 (L1) – capacity is higher than 80% of initial capacity; tween each facility, are determined by each facility’s latitude and
• Class 2 (L2) – capacity is between 60% and 80% of initial capacity; longitude, and accordingly straight-line distance is considered

3
L. Wang, et al. Applied Energy 275 (2020) 115328

between centers; formulated in Eq. (6).

(6). Based on the IPCC (Intergovernmental Panel on Climate Change)
Guidelines for National Greenhouse Gas Inventories, only the CO2 fPC = ∑ pcci ∙QBi + ∑ ∑ pcrj ∙QCRij + ∑ ∑ pcmj ∙QCMij +
i i j i j
emissions from the EVB recycling network are considered [34];
∑ ∑ pcwi ∙QCWik
i k (6)
2.3. Proposed model
where
Based on the above assumptions, a mixed integer programming
problem model is formulated to optimize the locations of the EVB re- QCRij - quantity of cells flowed between collection center i and re-
cycling network, minimizing the total cost of the entire recycling net- cycling plant j
work as shown in Eq. (1)–(3). QCMij - quantity of cells flowed between collection center i and re-
min F = Foperation + FCO2 emission (1) manufacturing plant j
QCWik - quantity of cells flowed between collection center i and
Where waste disposal center k
Foperation = fAC + fFC + fPC + fTC pcci - processing cost of collection center i
(2)
pcrj - processing cost of recycling plant
FCO2 emission = fFEC + fTEC (3) pcmj - processing cost of remanufacturing plant
pcwk - processing cost of waste disposal center k
As indicated in (1), the recycling cost comprises two parts, i.e. op-
eration costs and CO2 emission costs. Since all used batteries have to be The transported products mainly include battery cells and waste
recycled, based on [16,17], EV battery recycling and remanufacturing from recycling and remanufacturing processes. The transportation cost
can reduce greenhouse gas emissions compared to manufacturing bat- from one node to another is formulated in Eq. (7).
teries from raw materials. In addition, the operation of the EVB re-
cycling network will generate corresponding costs and carbon emis- fTC = ∑i ∑j Dij ∙tcc∙QCRij + ∑i ∑j Dij ∙tcc∙QCMij
sions, but these can be reduced by optimizing the design of the + ∑i ∑k Dik ∙tcc∙QCWik + ∑j ∑k Djk ∙tcw∙QWjk
recycling network.
The former is the sum of acquisition cost fAC , fixed cost fFC , pro- (7)
cessing cost fPC and transportation cost fTC ; the latter is the sum of the where
costs incurred in the processing and transportation processes, which are
denoted by fFEC and fTEC , respectively. tcc - transportation cost of one cell
tcw - transportation cost per unit waste
2.3.1. Operation costs
The acquisition cost means that in order to promote the recycling of
2.3.2. CO2 emission costs
used batteries, subsidies could be offered to consumers. The acquisition
The respective costs of CO2 emissions from processing used battery
cost of collection centers from the consumer market is calculated by Eq.
cells in the collection center, the recycling and remanufacturing plant,
(4).
and the waste treatment center can be estimated by Eq. (8).
fAC = ∑ ∑ QBi ∙ru ∙αu fFEC = ∑i pct ∙feci ∙QBi + ∑i ∑j pct ∙fer j ∙QCRij
i u (4)

where + ∑i ∑j pct ∙femj ∙QCMij

+ ∑i ∑k pct ∙fewk ∙QCWik (8)

QBi - quantity of used batteries returned to collection center i
ru - acquisition cost for used EVBs at different quality levels (L1/L2/ where
L3)
αu - percentage of batteries at different quality levels (L1/L2/L3) feci - unit CO2 emissions in collection center i,
ferj - unit CO2 emissions in recycling plant,
For the recycling and remanufacturing centers in the recycling femj - unit CO2 emissions in remanufacturing plant,
network, the total fixed cost is described by Eq. (5). fewk - unit CO2 emissions in waste disposal center k,
pct - carbon tax of unit carbon emission.
fFC = ∑ xi FCi + ∑ xj FCj + ∑ xk FCk
i j k (5)
Meanwhile, the carbon emission cost of the transportation process,
where including transporting battery cells and waste, is shown in Eq. (9).

FCj , FCj , and FCj - fixed cost of collection center i, recycling and fTEC = ∑i ∑j pct ∙Dij ∙te ∙QCRij ∙ω + ∑i ∑j pct ∙Dij ∙te ∙QCMij ∙ω
remanufacturing center j and waste disposal center k, respectively + ∑i ∑k pct ∙Djk ∙te ∙QCWik ∙ω
x i - binary decision variable equal to 1 if collection center i is open
and 0 otherwise + ∑j ∑k pct ∙Djk ∙te ∙QWjk ∙ω (9)
xj - binary decision variable equal to 1 if recycling and re-
manufacturing center j is open and 0 otherwise where
xk - binary decision variable equal to 1 if waste disposal center k is
ω - number of goods per container,
open and 0 otherwise.
te - CO2 emissions of shipping one container of products per kilo-
The processing cost of each center is mainly composed of three meter.
parts, i.e. the cost of detection and disassembly in the collection center,
the production cost in the recycling and remanufacturing centers, and 2.3.3. Constraints
the processing cost in the waste disposal center. This can be fPC To solve the model, the following constraints are considered.

4
L. Wang, et al. Applied Energy 275 (2020) 115328

2.3.3.1. Mass balance. 3.2. Input data

bc∙ ∑ QBi= ∑ ∑ QCRij + ∑ ∑ QCMij + ∑ ∑ QCWik ∀

The current numbers of used EVBs handled in different centers are
i i j i j i k
summarized in Table 1. Due to lacking information, the quantities of
i ∈ I, j ∈ J, k ∈ K (10) the cells at different levels are obtained based on assumption 4. The
specifications of an EVB are given in Table 2.
∑ ∑ QWjk = ∑ ∑ QCRij ∙wb ∙φ + ∑ ∑ QCMij ∙wb ∙φ ∀ The other costs needed by the calculation are presented in Table 3.
j k i j i j
Different recycling technologies result in different carbon emissions
i ∈ I, j ∈ J, k ∈ K (11) [24]. In this study, hydrometallurgical recycling is considered. The
where corresponding emission data are shown in Table 4. Moreover, based on
the results of [38], for transportation by trucks (Load: 9.0 ton/truck),
wb - the weight per cell (kg), the normal fuel consumption is about 0.33 l/km, which implies a
bc - number of cells included in an EVB, carbon emission density of 0.00249 kg CO2/(ton·km), based on the
φ - The mass of cells will change after being processed in the re- carbon emission density of diesel, 0.06805 kg CO2/l [39].
cycling center, which will directly affect the cost of transport and
the carbon emissions costs. Therefore, the mass ratio of the out- 3.3. Optimization of the studied case
flowed waste from recycling and remanufacturing center to the
disposal center is φ [35]. In order to verify the effectiveness of the proposed model, the cur-
rent Chang'an automobile battery recycling network is firstly opti-
Constraints (10–11) represent the mass equilibrium of the recycling mized. A genetic algorithm (GA) is used to solve the model for the case
and remanufacturing centers and waste disposal centers, respectively. study. For the GA, the crossover probability and mutation probability
are 0.6 and 0.05, and the number of iterations is 100. The optimized
2.3.3.2. Capacity constraints. The collection capacity of the collection recycling network is shown in Fig. 3, and the cost savings are shown in
center, the recycling capacity of the recycling plant, and the Fig. 4.
remanufacturing capacity of the remanufacturing plant are limited. As seen from Fig. 3 and Fig. 4, the optimized network reduced the
Constraints (12–14) indicate the capacity restriction constraints of the number of collection centers and waste disposal centers. Correspond-
collection center, the recycling and remanufacturing centers, and the ingly, the transportation routes between centers have also been
waste disposal centers, represented by Ii , Jj , and Kk , respectively. changed, which can result in a reduction of the total costs of the current
Chang'an EVB recycling network. Moreover, carbon emissions can be
∑ QBi ⩽ ∑ Ii xi ∀i∈I effectively reduced by 20.11%.
i i (12) To meet the requirements of the booming market, Chang'an is
planning to build one more recycling and remanufacturing center in the
∑ ∑ QCRij + ∑ ∑ QCMij ⩽ ∑ Jj xj ∀ i ∈ I, j ∈ J
current recycling network. According to city planning, there are 3 al-
i j i j j (13)
ternative addresses available to build the new recycling and re-
∑i ∑k wb ∙QCWik manufacturing center. They are also marked in Fig. 5, where blue solids
+ ∑ ∑ QWjk ⩽ ∑ Kk xk ∀ i ∈ I, j ∈ J, k represent existing centers, and red solids represents potential locations
bc j k k
for the new centers. In addition, it is worth noting that J1, J2 and J4 are
∈K (14) all in Yubei District, while J3 is in Jiangbei District in the adminis-
trative division of Chongqing. In order to identify the best choice
2.3.3.3. Decision variables constraints among these options, thereby achieving the lowest cost and carbon
emissions, the model developed in Section 2 is applied to optimize the
QBi , QCRij , QCMij , QWjk ⩾ 0 (15) network.
x i , x j , xk ∈ {0, 1} The proposed model is adopted to achieve the following goals: (1)
(16)
determine the location of the new recycling and remanufacturing
Constraints (15–16) are related to the binary and non-negativity center; (2) optimize the transport routes of the recycling network. In
restrictions on the corresponding decision variables. order to minimize the total cost of the recycling network, some col-
lection centers may be closed; and (3) comprehensively understand the
3. Case study impact of key factors on the location-allocation and operating cost of
the EVB recycling network.
3.1. Description of current network The new network design to optimize the location of the new re-
cycling and remanufacturing center for EVB recycling is shown in
Chang'an Automobile (Group) Co., Ltd. (Chang'an) is a manu- Fig. 6. The black and red dashed lines represent existing and proposed
facturer of EVs and therefore is responsible for recycling used power transport routes, respectively, and the closed centers have been marked.
batteries. Chang'an has established an EVB recycling network, in- Moreover, a comparison of the costs of existing and proposed networks
cluding: eight collection centers, which are responsible for providing is shown in Fig. 7. The corresponding costs per unit of battery recycling
services for detecting and dismantling used batteries; one recycling and are shown in Fig. 8.
remanufacturing center; and two waste disposal centers. These are From Fig. 6, it can be seen that J3 was selected as the best address of
shown in Fig. 2, in which green, blue, and black dashed lines represent the new recycling and remanufacturing center. In addition, it was found
the material flows of L1 & L2 cells, L3 cells, and waste produced in the that, compared to the previous recycling network, the optimized net-
recycling and remanufactured process, respectively. I, J, and K are used work is divided into two parts. The operation of the first part is centered
to indicate the collection centers, recycling and remanufacturing cen- on J2, while the operation of the second part is centered on J3.
ters, and waste disposal centers, respectively. Because EV battery re- Moreover, as the operation of the new center will inevitably increase
cycling and reuse mainly include the recycling of raw materials from fixed costs, the optimization results recommended that existing col-
used batteries and the remanufacturing of used batteries, the operation lection centers I4 and I7 be closed to reduce the costs of the whole
of the recycling network is organized according to the location of the recycling network.
recycling and remanufacturing center. Compared to the costs of the previous network, as shown in Fig. 7,

5
L. Wang, et al. Applied Energy 275 (2020) 115328

I1
K1

I2 J1

I8
I3
The flow of L1&L2 cells
I4 The flow of L3 cells
The flow of wastes
I7

I5 I Collection Center
K2 J Recycling & Remanufacturing Center

I6 K Waste Disposal Center

Fig. 2. The current recycling network of EVBs in Chang'an.

Table 1 Table 4
The flows cells at different centers. Unit emission parameters.
Collection No. of Recycling and remanufacturing centers To Waste Parameters Values
center used disposal
EVBs for To To centers CO2 emissions in collection centers (CO2 kg/cell) 2.43 [24]
Chang'an RemanufacturingPlant RecyclingPlant (L3 cells) CO2 emissions in recycling plants (CO2 kg/cell) 8.55 [40,41]
(L1 cells) (L2 cells) CO2 emissions in remanufacturing plants (CO2 kg/cell) 31.03 [21,42]
CO2 emissions in waste disposal centers (CO2 kg/cell) 0.35 [40,42]
I1 420 62 260 98
I2 351 54 217 80
I3 684 107 421 157
cost represented the biggest share in the total cost; even though it was
I4 539 82 333 124
I5 574 83 360 132 reduced by 7.3%, it showed the most obvious impact. These data prove
I6 637 97 391 149 that the developed model does work in optimizing the design of the
I7 314 48 184 82 recycling network.
I8 720 117 421 164 From Fig. 8, the cost per unit battery is 3977.4 USD in the recycling
network; this is much lower than the cost of manufacturing a new
battery ($10,000) [7,8]. Meanwhile, the battery cells produced by the
Table 2
recycling and remanufacturing process can be reassembled in new
EVB parameters.
batteries, thereby further reducing the cost of raw materials purchased
Weight (kg) the weight per cell (kg) Cells within an EVB Capacity (kWh) by enterprises in the process of producing new batteries. In addition,
the cost of a unit battery is based on the current recycling amount of
365 3.5 96 52.56
used batteries and cost parameters. As more and more used batteries are
recycled, and recycling technology develops, the value will change.
Table 3 Therefore, the optimization in this paper mainly involves transportation
Cost parameters. costs and carbon emissions generated during transportation.

Parameter classification Value

Fixed cost ($/center) [250,000–650,000 29] 4. Scenarios analysis

Processing cost in collection center ($/cell) 8.3 [36]
Processing cost in recycling plant($/cell) 20.5 [36] According to the cost equations in the objective function, many
Processing cost in remanufacturing plant($/cell) 62.3 [36]
factors are found to affect the design and the operating cost of the
Processing cost in waste disposal center($/cell) 0.08 [36]
Unit transportation cost ($/cell⋅km) 0.08 [27] network, such as the number of used batteries, carbon tax, unit trans-
Unit transportation cost ($/waste⋅km) 0.012 [27] portation cost and unit processing cost. We expect that as transporta-
Acquisition cost for L1/L2/L3 batteries ($/battery) (850, 220, 0) [23] tion costs and carbon taxes change, in order to reduce carbon emissions,
The price of carbon tax ($/ton) 44 [37] EV manufacturers will significantly differ from the baseline situation in
the location and allocation of facilities in the recycling network. In
addition, changes in processing costs will not affect the design of re-
the new network can achieve a reduction of the total cost by 5.7%. The
cycling network locations. In order to further understand the impacts,
transportation cost showed the biggest drop (reduced by 22.9%), fol-
five scenarios are considered in the research:
lowed by the carbon emission cost (reduced by 21.8%). The processing

6
L. Wang, et al. Applied Energy 275 (2020) 115328

The flow of L1&L2 cells

The flow of L3 cells I1
K1
The flow of wastes

I Collection Center
J Recycling & Remanufacturing Center
I2 J1
K Waste Disposal Center
I8
I K Closed Center

I3

I4

I7
I5

K2

I6

Fig. 3. The recycling network after optimization.

cost saving (%) I1

70.00 I2
60.60
60.00
J2 J1

50.00
I3 K1

J4
40.00

I4
30.00

19.27 20.11 I8
20.00

K2
10.00
9.12 J3
5.43
I5
0.00
0.00
Acquisition Fixed Processing Transportation Carbon emission
Cost cost cost cost cost
Total I6

Fig. 4. The cost savings after optimization. I7

Collection Center Recycling & Remanufacturing Center
Scenario 1 (S1) represents the reference scenario, with all inputs at
Waste Disposal Center Potential Recycling & Remanufacturing Center
the current level;
Scenario 2 (S2) considers an increase of used EVB by 30%, in which Fig. 5. The distribution of facilities.
the increased number of used batteries does not exceed the pro-
cessing capacity of the current opened center;
From the results of S2, as shown in Fig. 9, the optimal location of the
Scenario 3 (S3) considers changes to carbon tax, assuming an in-
recycling and remanufacturing center has been changed, from J3 to J4.
crease of carbon tax by 50%;
The closed centers and the transport routes were also different from S1.
Scenario 4 (S4) considers changes to transportation cost, assuming
Increased amount of used batteries will lead to increased costs in the
an increase of fuel price by 50%;
recycling network; therefore, several existing centers (I2, I4, I5-I7, and
Scenario 5 (S5) studies the impact of processing cost, assuming that
K2) will be closed in the optimized recycling network to reduce costs. In
advances in technology have reduced unit processing costs by 25%.
the optimized network, both the new and existing recycling and re-
manufacturing centers J4 and J1 were the core of the operation of the
The optimal designs of the network under different scenarios are
EVB recycling network. Compared with S1, the recycling network was
illustrated in Fig. 9; the total cost and cost breakdown are shown in
more centralized in S2.
Fig. 10; and the number of batteries handled in each center is shown in
According to S3, changes in carbon tax can also clearly affect the
Table 5. The results clearly show that the effects of processing cost and
optimal design of the recycling network. Fig. 9 shows that J3 will be
carbon tax on location optimization were mild, and that the effects of
selected for the new center. Since carbon emissions occur mainly during
the number of used batteries and fuel price can change the optimal
transportation of goods between the centers, and during processing in
location of the EVB recycling network.

7
L. Wang, et al. Applied Energy 275 (2020) 115328

J2
I1
K1
I Collection Center

J Recycling and Remanufacturing Center

J4

Waste Disposal Center I2 J1

K

The existing material flow

I8
The proposed material flow I3

I K Closed Center
I4
J3
I7
I5

K2

I6

Fig. 6. The recycling network after determining location of new opened center.

Fig. 7. The various cost comparison of existing and proposed.

Fig. 8. Cost breakdown: per unit of battery recycling.

8
L. Wang, et al. Applied Energy 275 (2020) 115328

J2 J2
I1 I1
K1 K1

J4 J4

I2 J1 I2 J1

I8 I8

I3 I3

I4 I4
J3 J3
I7 I7

I5 I5

K2 K2

I6 I6
S1&S5 S2

S3 S4

J2
I1 J2
K1 I1
K1

J4
J4
I2 J1
I2 J1

I8
I8
I3
I3

I4
J3 I4
J3
I7
I7
I5
I5
K2
K2
I6
I6

Fig. 9. The recycling network of S1-S5.

MUSD S1 S2 S3 S4 S5
25

20

15

10

0
Acquisition Fixed Processing Transportation CO2 emission
TOTAL
cost cost cost cost cost

Fig. 10. The costs breakdown of S1-S5.

9
L. Wang, et al. Applied Energy 275 (2020) 115328

Table 5 improving recycling technology.

The number of handled batteries of S1-S5 at different centers.
Scenarios S1 S2 S3 S4 S5 In fact, different types of batteries from different manufacturers can
use the same recycling network for secondary use of used batteries. It is
Collection centers I1 945 1911 420 1191 945 worth noting that the results will also vary from case to case due to the
I2 486 × 723 867 486
different actual conditions of different companies and regions. For ex-
I3 628 1747 1058 1045 628
I4 × × × × ×
ample, adopt different recycling processes or areas with very high
I5 8666 × 946 × 666 carbon tax. Therefore, relevant companies should refer to the scenarios
I6 637 × × × 637 analysis in this article and combine their environment to analyze, in
I7 × × × × × order to accurately determine the number of facilities and transporta-
I8 877 1852 1092 1136 877
tion routes in the recycling network.
Recycling and remanufacturing J1 1253 2012 1325 1824 1253
centers J2 × × × 1352 × In the next step, the research reported above can be further dee-
J3 22004 × 1928 × 2004 pened by considering more factors to achieve broader insights into the
J4 × 2249 × × × recycling and reusing of used electric vehicle batteries. For example, the
Waste disposal centers K1 337 1249 375 1063 337
model could consider the impact of different transportation methods
K2 635 × 611 × 635
and the types of used batteries on the total cost and the corresponding
carbon emissions of the recycling network, thereby making the model
each of them, identifying the shortest route for transportation through more accurately simulate the operation of an actual battery recycling
optimization became more important when the carbon tax was in- network. In addition, besides carbon emissions, other environmental
creased by 50%. In order to offset the fixed costs due to the opening of pollutants (e.g. PM2.5 or other gaseous pollutants) generated during
new recycling and remanufacturing centers J3, while minimizing the the electric vehicle battery recycling process could be also considered in
carbon emissions generated, the existing collection centers I4, I6 and I7 the model. That would enable the model to provide a more reliable
were closed in the optimized network, which was different from the prediction of the impact of the recycling of used EVBs on the en-
other scenarios. vironment.
In S4, the optimal location of the recycling and remanufacturing
center has been changed from J3 to J2. Meanwhile, as the increased CRediT authorship contribution statement
cost of fuel prices mainly affects the transportation routes, the closed
collection centers and waste disposal centers as well as the transport Lei Wang: Conceptualization, Methodology, Software, Resources,
routes were also different from S1. According to Fig. 10, increasing fuel Writing - review & editing, Conceptualization, Methodology, Software,
price can clearly increase the transport cost. But its impact on carbon Resources, Writing - review & editing. Xiang Wang: Investigation,
tax was not obvious, due to the optimized routes. Validation, Formal analysis, Writing - original draft. Wenxian Yang:
Supervision.
5. Conclusion and future work
Declaration of Competing Interest
Due to the urgent need for the recycling of used electrical vehicle
batteries, this work developed a model which considers three used The authors declare that they have no known competing financial
battery potential handling strategies (recycling, remanufacturing, and interests or personal relationships that could have appeared to influ-
disposal). This includes the waste battery disposal strategies of most ence the work reported in this paper.
domestic and foreign companies, and is generally representative.
Different quality levels for the used battery cells are considered in Acknowledgments
model, and are used to assign processing jobs to different centers within
the network for the design optimization of battery recycling networks. The authors gratefully acknowledge the support provided by the
This research can provide planning and decision-making methodology National Natural Science Foundation of China (NO. 51875058);
for electric vehicle manufacturers for recycling and secondary reuse at Chongqing Basic Science and Frontier Technology Research Special
the enterprise level. To reduce the costs and carbon dioxide emissions (NO. CSTC2018jcyjAX0414; Chongqing Municipal Education
of the entire network, the developed model has been applied to a real Commission Science and Technology Research Project (NO.
case study, based on the network of Chang'an electric vehicle manu- KJQN20180118) and Central University Frontier Discipline Special
facturer. And the results obtained are summarized as follows: Project (NO. 2019CDQYZDH025). We are also grateful to the anon-
ymous reviewers for their valuable comments and constructive criti-
(i) The proposed model can effectively optimize the design of re- cism.
cycling networks. Compared to the current layout of the network,
the total cost can be reduced by 5.7%, and CO2 emissions can be References
reduced by 21.8% after optimization. The model can also be used
to select optimal locations when additional centers are planned. [1] International Energy Agency. Global EV Outlook 2017: Two Million and Counting.
(ii) Transportation costs, carbon tax, and the number of used electric IEA 2017, Paris. 10.1787/9789264278882-en.
[2] General Office of the State Council. Energy Saving and New Energy Vehicle Industry
vehicle batteries have been identified as three major factors af- Planning from 2012 to 2020. http://www.gov.cn/zwgk/2012-07/09/content_
fecting the optimal design of the recycling network. Scenario 2179032.htm.
analyses reveal that variations in the key factors can change both [3] Andwari AM, Pesiridis A, Rajoo S, Martinezbotas R, Esfahanian V. A review of
battery electric vehicle technology and readiness levels. Renew Sust Energ Rev
the configuration of the network and the total cost. From the 2017;78:414–30. https://doi.org/10.1016/j.rser.2017.03.138.
perspective of total cost, processing cost has the most obvious [4] Yun L, Linh D, Shui L, Peng X, Garg A, Le MLP, et al. Metallurgical and mechanical
impact; meanwhile, carbon tax has the least impact. methods for recycling of lithium-ion battery pack for electric vehicles. Resour
Conserv Recy 2018;136:198–208. https://doi.org/10.1016/j.resconrec.2018.04.
(iii) Although processing cost accounts for the largest proportion of the 025.
total cost, changes in processing cost will not affect network op- [5] Chung D, Elgqvist E, Santhanagopalan S. Automotive lithium-ion cell manu-
timization. Therefore, for electric vehicle manufacturers, it is facturing: Regional cost structures and supply chain considerations (No. NREL/
TP–6A20-66086). NREL National Renewable Energy Laboratory, Golden, CO
possible to further reduce the total cost of the recycling network by
(United States) 2016. https://doi.org/10.2172/1247459.

10
L. Wang, et al. Applied Energy 275 (2020) 115328

[6] National Energy Administration. Whether the proper disposal of EV batteries can be lithium-ion batteries from electric vehicles. Waste Manage 2020;102:579–86.
achieved 2016. http://www.nea.gov.cn/2016-06/06/c_135415477.htm. https://doi.org/10.1016/j.wasman.2019.11.013.
[7] Sam Abuelsamid. General Motors builds first Volt battery pack on production line. [26] Hao H, Qiao Q, Liu Z, Zhao F. Impact of recycling on energy consumption and
http://green.autoblog.com/2010/01/07/general-motors-builds-first-volt-battery- greenhouse gas emissions from electric vehicle production: The China 2025 case.
pack-on-production-line/. Resour Conserv Recy 2017;122:114–25. https://doi.org/10.1016/j.resconrec.2017.
[8] Gaines L, Cuenca R. Costs of lithium-ion batteries for vehicles. United States. 10. 02.005.
2172/761281. [27] Ciez RE, Whitacre J. Examining different recycling processes for lithium-ion bat-
[9] Neubauer J, Pesaran A. The ability of battery second use strategies to impact plug-in teries. Nat Sustain 2019;2:148–56. https://doi.org/10.1038/s41893-019-0222-5.
electric vehicle prices and serve utility energy storage applications. J Power Sources [28] Kannan G, Sasikumar P, Devika K. A genetic algorithm approach for solving a closed
2011;196(23):10351–8. https://doi.org/10.1016/j.jpowsour.2011.06.053. loop supply chain model: A case of battery recycling. Appl Math Model
[10] Harper G, Sommerville R, Kendrick E, Driscoll L, Slater PR, Stolkin R, et al. 2010;34(3):655–70. https://doi.org/10.1016/j.apm.2009.06.021.
Recycling lithium-ion batteries from electric vehicles. Nature 2019;575:75–86. [29] Subulan K, Baykasoglu A, Ozsoydan FB, Tasan AS, Selim H. A case-oriented ap-
https://doi.org/10.1038/s41586-019-1682-5. proach to a lead/acid battery closed-loop supply chain network design under risk
[11] Vikstrom H, Davidsson S, Hook M. Lithium availability and future production and uncertainty. J Manuf Syst 2015;37:340–61. https://doi.org/10.1016/j.jmsy.
outlooks. Appl Energy 2013;110:252–66. https://doi.org/10.1016/j.apenergy. 2014.07.013.
2013.04.005. [30] Li L, Dababneh F, Zhao J. Cost-effective supply chain for electric vehicle battery
[12] Nissan USA. Recycling Your Old Car Battery. https://www.nissan.co.uk/recycle- remanufacturing. Appl Energy 2018;226:277–86. https://doi.org/10.1016/j.
your-old-car-battery.html. apenergy.2018.05.115.
[13] Volkswagen. Battery recycling. UK, 2016. https://www.volkswagen.co.uk/need- [31] Gu X, Ieromonachou P, Zhou L, Tseng M. Developing pricing strategy to optimise
help/environment/recycling. total profits in an electric vehicle battery closed loop supply chain. J Clean Prod
[14] Ayre J. Nissan reuses EV batteries for home energy storage XStorage. https:// 2018;203:376–85. https://doi.org/10.1016/j.jclepro.2018.08.209.
cleantechnica.com/2016/05/15/nissan-recycles-ev-batteries-home-energy- [32] Hannan MA, Hoque MM, Hussain A, Yusof Y, Ker PJ. State-of-the-Art and Energy
storage/. Management System of Lithium-Ion Batteries in Electric Vehicle Applications:
[15] Dalton A, 2016. BMW will repurpose I3 batteries for home energy storage. https:// Issues and Recommendations. IEEE Access 2018;6:19362–78. https://doi.org/10.
www.engadget.com/2016-06-21-bmw-will-repurpose-i3-batteries-for-home- 1109/ACCESS.2018.2817655.
energy-storage.html. [33] Gandoman FH, Jaguemont J, Goutam S, Gopalakrishnan R, Firouz Y, Kalogiannis T,
[16] Song Z, Feng S, Zhang L, Hu Z, Hu Z, Yao R. Economy analysis of second-life battery et al. Concept of reliability and safety assessment of lithium-ion batteries in electric
in wind power systems considering battery degradation in dynamic processes: Real vehicles: Basics, progress, and challenges. Appl Energy 2019;251:113343https://
case scenarios. Appl Energy 2019:251. https://doi.org/10.1016/j.apenergy.2019. doi.org/10.1016/j.apenergy.2019.113343.
113411. [34] Hasanov P, Jaber MY, Tahirov N. Four-level closed loop supply chain with re-
[17] Voelcker J. Reusing electric-car batteries: great idea, lots of practical challenges. manufacturing. Appl Math Model 2019;66:141–55. https://doi.org/10.1016/j.apm.
Green Car Report 2016. https://www.greencarreports.com/news/1103363_ 2018.08.036.
reusing-electric-car-batteries-great-idea-lots-of-practical-challenges. [35] Diekmann J, Hanisch C, Frobose L, Schalicke G, Loellhoeffel T, Folster A, et al.
[18] Kelty K. Tesla's closed loop battery recycling program. Tesla 2011. Ecological recycling of lithium-ion batteries from electric vehicles with focus on
[19] Georgimaschler T, Friedrich B, Weyhe R, Heegn H, Rutz M. Development of a re- mechanical processes. J Electrochem Soc 2017;164:6184–91. https://doi.org/10.
cycling process for Li-ion batteries. J Power Sources 2012;207:173–82. https://doi. 1149/2.0271701jes.
org/10.1016/j.jpowsour.2012.01.152. [36] Standridge CR, Corneal L. Remanufacturing, repurposing, and recycling of post-
[20] Tang Y, Zhang Q, Li Y, Li H, Pan X, Mclellan B. The social-economic-environmental vehicle-application lithium-ion batteries; 2014. No. CA-MNTRC-14-1137. https://
impacts of recycling retired EV batteries under reward-penalty mechanism. Appl rosap.ntl.bts.gov/view/dot/27425.
Energy 2019;251:113313https://doi.org/10.1016/j.apenergy.2019.113313. [37] Tian X, Dai H, Geng Y, Huang Z, Masui Z, Fujita T. The effects of carbon reduction
[21] Liu C, Lin J, Cao H, Zhang Y, Sun Z. Recycling of spent lithium-ion batteries in view on sectoral competitiveness in China: A case of Shanghai. Appl Energy
of lithium recovery: A critical review. J Clean Prod 2019;228:801–13. https://doi. 2017;197:270–8. https://doi.org/10.1016/j.apenergy.2017.04.026.
org/10.1016/j.jclepro.2019.04.304. [38] Nieuwenhuis PA, Beresford AK, Choi AK. Shipping or local production? CO2 impact
[22] Qiao Q, Zhao F, Liu Z, Jiang S, Hao H. Cradle-to-gate greenhouse gas emissions of of a strategic decision: an automotive industry case study. Int J Prod Econ
battery electric and internal combustion engine vehicles in China. Appl Energy 2012;140:138–48. https://doi.org/10.1016/j.ijpe.2012.01.034.
2017;204:1399–411. https://doi.org/10.1016/j.apenergy.2017.05.041. [39] Korea LCI Database Information Network. http://www.edp.or.kr/lci_db.asp.
[23] Kim HC, Wallington TJ, Arsenault R, Bae C, Ahn S, Lee J. Cradle-to-Gate emissions [40] ANL, 2018a. Closing the loop on battery recycling. https://www.anl.gov/article/
from a commercial electric vehicle Li-Ion Battery: A comparative analysis. Environ closing-the-loop-on-battery-recycling-0.
Sci Technol 2016;50(14):7715–22. https://doi.org/10.1021/acs.est.6b00830. [41] Dai Q, Spangenberger JS, Ahmed S, Gaines L, Kelly JC, Wang M. EverBatt: A Closed-
[24] Dunn JB, Gaines L, Sullivan J, Wang M. Impact of recycling on cradle-to-gate en- loop Battery Recycling Cost and. Environmental Impacts Model. Argonne National
ergy consumption and greenhouse gas emissions of automotive Lithium-Ion bat- Laboratory, Chicago 2019. https://doi.org/10.2172/1530874.
teries. Environ Sci Technol 2012;46(22):12704–10. https://doi.org/10.1021/ [42] Dunn JB, James C, Gaines L, Gallagher K, Dai Q, Kelly JC. Material and energy flows
es302420z. in the production of cathode and anode materials for lithium ion batteries. Acta
[25] Xiong S, Ji J, Ma X. Environmental and economic evaluation of remanufacturing Chem Scand 2015;49(24):44–52. https://doi.org/10.2172/1044525.

11