You are on page 1of 17

The current issue and full text archive of this journal is available on Emerald Insight at:

https://www.emerald.com/insight/1355-2511.htm

JQME
28,1 Implementing an integrated
maintenance management system
for monitoring production lines:
180 a case study for biscuit industry
Received 3 June 2020 Mokhtar Ali Amrani
Revised 8 August 2020
Accepted 30 September 2020 Industrial Engineering, Taiz University, Taiz, Yemen;
Al-Janad University for Science and Technology, Taiz, Yemen and
Center for Graduate Studies (CGS), Taiz University, Taiz, Yemen
Mansour Alhomdi and Badiea Aswaidy M.
Taiz University, Taiz, Yemen
Atef M. Ghaleb
Industrial Engineering, King Saud University, Riyadh, Saudi Arabia
Mohyeddine Al-Qubati
Department of Physics, Taiz University, Taiz, Yemen and
Al-Janad University for Science and Technology, Taiz, Yemen, and
Mutahar Shameeri
Industrial Engineering, Taiz University, Taiz, Yemen
Abstract
Purpose – This study provides a unique integrated diagnosis system to investigate the causes of low
productivity, profitability, machinery health conditions and wear severity of medium-size biscuit industry
assets in Taiz, Yemen.
Design/methodology/approach – The evaluation is based on an integrating of the overall equipment
effectiveness (OEE) and oil-based maintenance (OBM) approaches. The data are collected using the company’s
operational records, interviews and observations, while the used lubricating oil samples are also collected from
production lines’ machineries. Scanning electron microscope (SEM) is used to study the wear debris particle
features and wear mechanism. Different other analysis tools such as fishbone, 5 whys and Pareto charts are
also used to investigate the root causes and plausible recovery solutions of machinery failures.
Findings – This study demonstrated that a large proportion of machinery failures and production loss are of
management concerns. Also, this study inferred that the analysis of wear debris is unique and informative for
determining machinery wear severity and useful life. Finally, the current conditions of production lines are
clarified and suggestions to use a mixed preventive/predictive maintenance management approach are also
elucidated.
Originality/value – This work implemented an integrated OEE/OBM diagnostic maintenance system to
investigate the root causes of low productivity and machine failures in real production lines and suggested
robust decisions on the maintenance duties.
Keywords Predictive maintenance, OEE, Wear debris, Machine failure, Oil-based maintenance, Proactive
maintenance
Paper type Research paper

1. Introduction
The maintenance management is an essential process that involves decision-making at
different stages of operations in industrial systems, especially for dynamic and complicated
Journal of Quality in Maintenance
Engineering
Vol. 28 No. 1, 2022
pp. 180-196 The authors thank Yemen Lubricant Manufacturing Co. Ltd., Taiz, for allowing them to conduct the oil
© Emerald Publishing Limited analysis measurements. The authors also thank Prof. Ahmed Al-Fatish, Department of Chemical
1355-2511
DOI 10.1108/JQME-06-2020-0049 Engineering, King Saud University, for doing SEM analysis.
systems. Of note, a poor decision leads to a direct loss of resources (materials, machines, time Integrated
and money). Maintenance expenses can reach up to 40% of the operational costs and could maintenance
uprise more if not planned well, affecting the productivity and the useful life of
manufacturing systems (Kumar and Soni, 2015; Eti et al., 2006; Kent et al., 2018). As a
management
result, maintenance management systems have shown an exponential development during system
the last few decades (Garg and Deshmukh, 2006).
Recently, there are numerous diagnostic/prognostic maintenance strategies have been
employed; such as total productive maintenance (TPM), reliability-centered maintenance 181
(RCM), condition-based maintenance (CBM) and computerized maintenance management
system (CMMS) (Jain et al., 2014; Ahuja and Khamba, 2008; Prabhakar and Raj, 2014; Sakib
and Wuest, 2018; Randall, 2011; Wakiru et al., 2019). Furthermore, advanced maintenance
management strategies involve all levels of employees, management and manufacturers in
the maintenance activities and employ online monitoring systems (Jain et al., 2014; Ahuja and
Khamba, 2008). However, information-based online monitoring approaches properly fit with
the critical and most expensive systems. Also, these approaches should be initially
considered at the design stage, are extremely hard to implant for a system in use and identify
only limited parameters in narrow ranges (Randall, 2011; Sakib and Wuest, 2018). For that,
off-line monitoring approach is still dominating in the maintenance systems (Wakiru
et al., 2019).
Overall equipment effectiveness (OEE) is an off-line metric tool and a TPM pillar that has
been commonly implemented in industries to diagnose the failure of equipment/system and
thus provide a support for the right decisions (Jain et al., 2014; Nurprihatin et al., 2019; Kumar
and Soni, 2015; Cheah et al., 2020). The data acquired by OEE are crucial for monitoring the
overall performance of equipment; it shows up many hidden costs such as the costs of
downtime, slow production and defect products. In fact, the manufacturing process is
subjected to blind control if the initial OEE analysis is neglected. In the other hand, OEE has
many drawbacks; it considers an equal weight for all types of losses regardless of their
importance, neglects the bottleneck stoppages in production lines, depends substantially on
the quality of collecting data and has different benchmarks and definitions (Da Costa and de
Lima, 2002; Aminuddin et al., 2016; Cheah et al., 2020). Furthermore, OEE analysis is limited to
study one equipment only and requires modifications to study production lines, plants and
large factories (Braglia et al., 2009), even if the OEE-based evaluation is not sufficient
(Wudhikarn, 2016; Piran et al., 2020). In addition, OEE analysis provides no information about
the dynamic behavior of machinery, its wear severity and the remaining useful life.
Given the previous barriers, OEE-based maintenance decision-making is not a
straightforward process. OEE-based developed maintenance approaches such as the TPM,
lean manufacturing and six sigma require a precise and planned collecting data procedure,
investing money to employ and engaging all levels of management and employees in the
process (Aminuddin et al., 2016; Da Costa and de Lima, 2002; Braglia et al., 2009; Cheah et al.,
2020; Raouf, 1994). As a result, it is unfair to rely on the OEE to make maintenance decisions
in the absence of the aforementioned requirements. Therefore, it is a fascinating trend of
advanced monitoring technologies to implement an appropriate mix of measuring tools that
fit with a conditioned maintenance job (Wudhikarn, 2013). The probable interactions and
relationships between the chosen approaches have to be addressed clearly (Wudhikarn, 2016;
Piran et al., 2020).
Mechanical analysis of dynamic machines is probably an ideal choice to support the
maintenance management decisions. Of course, the machine that has a transmission system,
such as gears, requires focused maintenance since the transmission system contributes to
about 60% of total maintenance costs (Cao et al., 2019). Of note, technical approaches such as
5 whys technique, Pareto chart and failure mode and effect analysis (FMEA) are commonly in
use to identify the root causes of failure (Cheah et al., 2020). However, these tools rely mainly
JQME on technician interviews, direct observations and historical data, which are imprecise data
28,1 collection tools. Instead, a systematic analysis of the machinery’s lubricating oil could
investigate the root causes of failure efficiently. Therefore, an integration of OEE with the oil-
based maintenance (OBM) could offer compelling and comprehensive analysis.
OBM is an efficient predictive maintenance tool that is broadly implemented to
continuously identify failures earlier by monitoring the lubricating oil properties and
impurities (Sakib and Wuest, 2018). It behaves like blood, urine or stool tests for identifying
182 human health conditions. Similarly, used oil properties such as viscosity, moisture, acid
number, density and debris particles could be determined to identify the health condition of
the machines. Nowadays, OBM employs different evaluation approaches, from conventional
statistics to artificial intelligence, in order to support and make the right decisions (Wakiru
et al., 2019). It is one of the most cost-effective off-line tools for maintenance purposes that
offers a systematic analysis of machinery failures, wear rates and useful life (Kumar and Soni,
2015). For example, wear debris analysis investigates whether the machine is in normal,
abnormal or catastrophic operations and therefore helps in a robust decision-making to
implement overhaul maintenance or replacement.
For that, OEE/OBM integrated system is an ideal choice to identify failures of machines
that have a transmission system and large movable parts. It is an optimum choice in cases
where the quality of collecting data is inferior, and the management is reluctant to develop so
that a development toward TPM, lean manufacturing or six sigma is unviable. To the best of
the author’s knowledge, no study has reported the root causes of equipment failures in the
production line using OEE/OBM integrated maintenance system. The OEE/OBM-based
analysis could provide secured solutions to the technical and managerial maintenance
problems and make a plain impact on equipment effectiveness and productivity as well. This
works to precisely diagnose the maintenance problems, investigate the root causes of failure
and suggest reliable and sustainable solutions that warrant high profitability and
performance of biscuit production lines using OEE/OBM integrated system.

2. Methodologies and experiments


2.1 Data collection method
The data were collected from a medium-size food company in Taiz city, Yemen. The Biscuit
section consists of three production lines; they are Sandwich, Wafer and Marie biscuit lines,
which are then denoted in the analysis as Line1, Line2 and Line3, respectively. These
production lines were first purchased from Imaforni SpA, Italy, and put into service for the
first time in 1993. Figure 1 displays an illustrative diagram for the production lines and assets
under study and the methodologies used in the present work. Operational data were precisely
recorded through Jan.–Dec. 2017 for the three lines, while a complete data of Line 2 were
collected during the first six months of 2018 for its six machines as shown in Figure 1. Table 1
summarizes the cumulative required operational data from the three production lines for one
year. The data involve planned, actual and waste products and time.
The equations (Eq. 1–4) govern the measurements of OEE (Nakajima, 1988), while
equations (Eq. 5 and Eq. 6) represent the mean time between failure (MTBF) and mean time to
repair (MTTR).
OEE ¼ Availability 3 Performance 3 Quality (1)
Planned time  Breakdown time
Availability ¼ (2)
Planned time
Actual production time
Performance ¼ (3)
Planned production time
Integrated
maintenance
management
system

183

Figure 1.
An illustration
diagram of the major
production lines and
equipment and the
proposed
methodologies that
were considered in the
present study

Parameter Line 1 Line 2 Line 2 Table 1.


The required data that
Planned time (h) 8640.00 8,640 8,640 are collected from the
Breakdown time (h) 1337.04 2592.29 1546.23 three production line as
Actual time (h) 7302.96 6047.71 7093.77 per the official
No. of planned products 792,000 1,566,000 756,000 operational records of
No. of good products 522,692 834,346 478,010 food company in Taiz
No. of defect products 106,709 97,484 31,089 city, Yemen, during
No. of actual products 629,401 931,830 509,099 Jan.–Dec. 2017

Good products
Quality ¼ (4)
actual products

The initial scanning of OEE data has led us to implement the OEE tool for Line 2, the most
risky line. Of course, the sequence of production processes of Line 2 (Sandwich line) started
in the mixing machine and ended at the Packaging Machines 1 and 2, as illustrated in
Figure 1. Basically, the operational data of Line 2 were collected for six months (Jan.–June
2018) and used to evaluate the OEE of its six machines. In detail, information of planned,
actual and good products, the time required to execute each product and breakdown time
have been collected for each equipment on Line 2 and organized in Table 2. Of note, the
planned time is considered the total time of operation based on 24 h daily and seven days
weekly for the time period of study. The MTBF and MTTR have also been estimated using
Eq. (5) and Eq.(6), respectively. Different tables, graphs, Pareto charts, fishbone diagrams
and 5 why questions were used to measure and analyze data.
Actual Oper: Time ðhrs:Þ
MTBF ¼ (5)
No: of Breakdown
MTBF
MTTR ¼  MTBF (6)
Availability
JQME Mixer Forming Filling Packaging Packaging
28,1 Parameter machine machine Oven machine machine 1 machine 2

Planned time (h) 4,320 4,320 4,320 4,320 4,320 4,320


Breakdown time 464.55 503.91 607.65 544.33 1570.38 1398.08
(h)
Actual op. time 3855.45 3816.08 3712.35 3775.66 2749.62 2921.91
184 (h)
No. of 649 750 717 782 1,025 1,160
breakdowns
Planned 1.5 18,000 18,000 18,000 9,000 9,000
products per h
Table 2. Actual products 1.1215 14275.75 13997.1 13449.04 6534.11 6523.1
Manipulated raw data per h
collected from the Line Good products 1.09 14054.3 13503.51 13110.72 6065.24 6101.274
2 as per the official per h
operational records Defect products 0.0315 221.432 493.6 338.325 468.87 421.82
during Jan.–June. 2018 per h

2.2 Oil-based condition experiments


The base oil of high-grade ISO VG 68 was purchased from a local agent and used for the
analysis along with used oil samples. Besides, used oil samples were collected from
packaging machines 1 and 2 after 3,000 working hours and then coded as S1 and S2,
respectively. Interestingly, in spite of six major equipment in Line 2 as demonstrated in
Figure 1, only packaging machines 1 and 2 constitute transmission systems and thus the
analysis of OBM will be assigned to them only.
The samples of wear debris were magnetically extracted using a simple method. In brief, a
permanent magnet was put in a 0.5 L of oil sample, S1, diluted with ethanol and then shaken
to enhance the attraction of wear debris particles. After that, the magnet was carefully
separated from the oil sample, washed several times with distilled water and ethanol and
dried in the oven at 120 8C for 6 h. The same procedure was also conducted for the S2 sample.
In the final step, debris particles samples were packed and then characterized using scanning
electron microscope (SEM) model (JCM 6,000Plus, JEOL, Japan), to identify the particle
features (sizes and morphologies) and surface information of wear debris. The mass of debris
particles in the oil samples was calculated by measuring the mass of magnet before and after
the experiments and the difference was counted.
Water content (moisture) measurements of oil samples were obtained by weighing a
predetermined mass of oil samples (20 g ± 0.2) that was heated at 150 8C on an electronic
magnetic hot and stirrer device for 1 h to allow water/moisture in the oil to evaporate.
Subsequently, the masses of oil samples before and after the dehydration experiment were
recorded using a digital balance (Ohaus Scout Pro SP401) and the difference was considered.
This simple measuring technique assumes that water is the only evaporated element at this
temperature limit (150 8C). Separately, density measurements were conducted by measuring
the mass of base and used oil samples of the same volume, and the difference was recorded
and then converted to density units.
For viscosity measurement, an ASTM D445–06 standard test of kinematic viscosity
(ASTM D445–06, 2018) was conducted using the Seta Vis Kinematic Viscometer (PMT
THOMSON) of a glass capillary. The measurements of viscosity were first conducted at 40 8C,
while the viscosity index calculation follows an ASTM D2270 standard at 40 and 100 8C
(ASTM D2270–10, 2016). The procedure could be summarized in the following steps; all
measurement tools were carefully cleaned and dried by passing a warm air bath. Integrated
Subsequently, a pipette was filled with 10 mL of oil sample and then charged to a U-Tube, maintenance
while the temperature of the bath was fitted at 40 8C using a thermometer (accuracy
of ± 0.02 8C). During the flow of oil in the tube, a stopwatch was calibrated to measure the
management
efflux time between two preset lines on the tube. In the final step, the time was recorded, system
converted to the required units and multiplied by a device factor as recommended. The
procedure was repeated for S2 once, and the whole procedure was repeated at 100 8C and the
results were recorded accordingly. 185

3. Results and discussion


The analysis will be divided into two main compartments. The first subsection reviews the
analysis of OEE findings and different evaluation tools such as MTBF, MTTR, Pareto chart
and fishbone diagram. Furthermore, the second subsection discusses the evaluations of OBM
measurements such as viscosity, moisture, density and wear debris.

3.1 OEE analyses


Table 3 demonstrates the outcomes of the OEE analysis from Line 1, Line 2 and Line 3.
Critically, the results denoted that 15.5, 30 and 17.9% of the planned time transferred to
wasted time plausibly due to breakdowns and other unplanned stops at Line 1, Line 2 and
Line 3, respectively. The aforementioned breakdown analysis should be taken seriously,
owing to its reflected costs on the company budget. Of course, large losses in resources (time,
materials, machines, energy and products) will definitely lead to additional costs
(Wudhikarn, 2012).
The aforementioned analysis could also be represented by OEE metric analysis.
Accordingly, the obtained results of OEE in Table 3 for Line 1, Line 2 and Line 3 are alarming
at 56, 38 and 52%, respectively. These findings are much below the world standard of good
performance (60–85%) (Nakajima, 1988), which undoubtedly demonstrated that the company
is employing only run-to-failure as a maintenance strategy without implementing any
maintenance management system. What is more, the OEE initial evidence showed that Line 2
is the most stressed line with only 38%, which is a disastrous condition, as per OEE world-
class standards. It exhibited the lowest availability, performance and quality as compared to
the other lines, specifying a high level of failures and wear-out machines. Besides, the
performance of Line 2 is also critical at only 60%, revealing a large bottleneck that requires
more investigation. By implementing the 5 whys technique, the technicians have attributed it
to the use of only three heating zones instead of four heating zones of the drying system, as
recommended by the manufacturer. In fact, a machine on a series of a production line that has
the lowest performance will cause a bottleneck for the whole system and then a reduction of

Parameter OEE standard [a] Line 1 Line 2 Line 3


Table 3.
Planned time (h) 8,640 8,640 8,640 8,640 OEE evaluation
Breakdown time (%) 10 15.50 30.00 17.90 findings of the three
Defect products (%) 1.0 26.94 10.50 06.10 production lines of food
company in Taiz city,
Availability (%) 90 84.50 70.00 82.00 Yemen. The
Performance (%) 95 79.50 60.00 67.00 corresponding
Quality (%) 99 83.00 90.00 94.00 benchmarks are also
OEE (%) 85 56.00 38.00 52.00 listed in the second
Note(s): [a](Nakajima, 1988) column
JQME production rate. To this, Line 2 could be selected for further OEE deep analysis of its
28,1 individual assets.
Not only the low availability and performance are explaining the low productivity of
production lines, a large number of products were classified as damaged products of the
production lines as presented in Table 3. In numbers, 16, 10.5 and 6.1% of products are
defective products in Line 1, Line 2 and Line 3, respectively, as compared to only 1.0% world
benchmarks (Al-Nozili et al., 2014). Unluckily, absolutely excessive results are obtained from
186 Line 1 with more than one-sixth damaged products in the process, revealing senseless
management since this large loss is so expensive waste. Of course, the large imperfect
products could be attributed to the nature of the product (Biscuits), the old design of
production lines and bad management system. Seriously, the company should hire
consultants along with overhaul maintenance to enhance the quality measure in the Line 1
and relatively other production lines. Once this problem is treated, the OEE measures of
Line 1 will jump to a safe range. Likewise, the results of Line 3 revealed low performance at
67%, indicating a severe bottleneck in the line that needs to be improved. However, the line
has reasonable quality at 94% as compared to the other lines. Thus, employing a strict
maintenance firm and offering sufficient maintenance budget could increase its OEE
readings readily.
Table 4 records the calculated values of breakdown time, defect products, MTBF, MTTR
and OEE of the six assets of Line 2. That is to clarify the unclear faults as per the
aforementioned analysis of Table 3. Obviously, the results conveyed that among the six
equipment of Line 2, the packaging machines 1 and 2 are the most stressed machines with
almost all examined parameters. Besides, a Pareto chart of cumulative breakdowns of
equipment of Line 2 is plotted in Figure 2. The analysis implies that this line is in the worn-out
stage. In numbers, packaging machines 1 and 2 constitute 58.33% of all breakdowns, while
the remaining other four equipment are within the world standard range (10%). Similarly, the
damaged products in the packaging machine 1 and 2 represent 7.2 and 6.5% of total products,
respectively, as compared to the 1% world benchmark. Unhappily, unbelievable quantitative
findings showed that every 2.68 h and 2.52 h, there is a single failure in the packaging

OEE [a] Mixer Forming Filling Packaging Packaging


Parameter standard machine machine Oven machine machine 1 machine 2

Planned time 8,640 4,320 4,320 4,320 4,320 4,320 4,320


(h)
No. of – 649.0 750.0 717.0 782.0 1,025 1,160
Breakdowns
(year)
Breakdown 10 09.13 09.90 11.94 10.67 30.86 27.47
time (%)
Imperfect 01 02.81 01.55 03.53 02.52 07.18 06.47
products (%)
Availability 90 89.00 88.00 86.00 87.00 64.00 68.00
(%)
Table 4. Performance 95 75.00 79.00 78.00 75.00 73.00 73.00
OEE evaluation (%)
findings of six major
equipment in Line 2. Quality (%) 99 97.00 98.00 96.00 97.00 93.00 94.00
The corresponding OEE (%) 85 65.00 68.00 64.00 63.00 43.00 46.00
benchmarks are also MTBF (h) – 05.94 05.09 05.18 04.82 02.68 02.52
listed in the second MTTR (h) – 00.72 00.67 00.85 00.69 01.53 01.20
column Note(s): [a](Nakajima, 1988)
Integrated
maintenance
management
system

187

Figure 2.
Pareto chart of the
accumulated
breakdown time of the
machines of Line 2

machine 1 and 2, respectively, while the meantime spent on repairing each failure is also the
highest among other equipment. In other words, the failure frequency of packaging machine
1 and 2 is in the order of several times daily while the time spent to repair each failure takes
more than 1 h on average. Gravely, the aforementioned outcomes of Line 2 revealed that
∼30% of the planned time is classified as wasted time. Indeed, this evidence indicates the
basest maintenance management at all times. Of course, the aforementioned analysis draws
attention to the packaging machines 1 and 2, reflecting their risky conditions. In this regard,
we could attribute that to their large movable parts and transmission systems as compared to
the other equipment in the same production line. The aforementioned unusual outcomes are
questionable and the investigation of the root causes of large failures is highly recommended.
Technical data on the packaging machines 1 and 2 has to be identified.
The aforementioned analysis could also be strengthened by the results of availability,
performance, quality and OEE of all assets of Line 2 as detailed in Table 4, which
demonstrated that all equipment are within the world standard range with the exception to
the packaging machine 1 and 2. Likewise, the findings of OEE of all six assets indicated that
packaging machines 1 and 2 are the most promoted machines to failure with 43 and 46%,
respectively, while the other equipment behave relatively better in the range 63–68%.
Unexpectedly, equipment’s performance (%) evaluation exhibited values in a narrow range
(73–79%), expressing that the causes of delay in this line could be assigned to the whole
system and not a single asset.
All in all, the aforementioned analysis inferred that Line 1 and Line 3 operate at abnormal
conditions and the company should take immediate action to implement efficient
maintenance management and invest sufficient budget for maintenance or to purchase
new systems with online monitoring tools, if budget allows. In this regard, the whole system
of Line 2 is in a catastrophic condition and the company unintentionally loses a large amount
of money and reputation. Accordingly, our recommendation goes to immediately replace it.
However, the aforementioned decisions were made based on OEE analysis only, which
suggests too expensive decisions for the management to invest. Of course, OEE findings
JQME could cast doubts on the quality of data since the analysis showed a careless management at
28,1 different stages of manufacturing processes. Thus to provide a robust decision, the
packaging machines 1 and 2 need another monitoring approach that investigates the
mechanical dynamic system of these machines with a goal to predict the root causes of
frequent failures and their effects. For that reason, the next section converges to study the
lubricating oil properties and impurities of these two machines.
188 3.2 Oil-based conditions maintenance
Condition-based maintenance is an effective tool that belongs to the predictive maintenance
family. It relies on the monitoring of operating conditions of machines (Sakib and Wuest,
2018). Recently, the analysis of used lubricating oil either from the gearboxes or from the
engine crankcase has been confirmed to afford valuable information regarding the machine
health (Wakiru et al., 2019). Herein, the oil analysis comprises the kinematic viscosity, density,
wear debris and water/moisture contents in the oil, which is the most off-line technology
commonly in use to investigate the conditions and lifetime of machines. Of note, the analysis
constitutes only packaging machines 1 and 2, which are denoted in the analysis as S1 and S2,
respectively, while S0 is donated to the base oil. That is because the other machines do not
have an oil system.
3.2.1 Viscosity of lubricant oil. Table 5 records the kinematic viscosity (centistokes) and
viscosity index (VI) findings, both at 40 and 100 8C. Basically, the allowable reduction of used
oil viscosity should be within ±10% as a good signal for good machinery (Al-Nozili et al.,
2014). In fact, oil viscosity is the most crucial property of a lubricant and its reduction reflects
the amount of oil impurities, oil physiochemical reactions and the machinery wear rate. Thus,
the analysis of oil viscosity is essential for failure analysis. In this regard, the results of
Table 5 showed that the kinematic viscosity reductions of S1 and S2 at 40 8C are 28.42 and
21.36%, respectively, while the corresponding kinematic viscosity of S1 and S2 oil samples at
100 8C is also reduced by 24.7 and 16%, respectively. The high reduction of kinematic
viscosity could be assigned to the low-viscous contaminants such as water, water vapor, wear
products and pollution carried out by air to the gearbox. In addition, evidence of VI findings
revealed a reduction as compared to the new oil sample (S0). All in all, the measurements of
kinematic viscosity and VI inferred that the equipment related to the S1 oil sample is more
intense than the machine of S2, which is in line with the previous findings of the OEE
diagnostic metric system.
3.2.2 Water content and density measurements. Water in the lubricating oil is harmful even
at low concentrations. It increases the wear rates several folds and enhances the oxidation
and additives reactions (Soni and Patel, 2017). Thus, lubricating oil should be sealed tightly
with the surrounding environment. Table 6 depicts the findings of density and water content
measurements of the oil samples revealing an abrupt reduction as compared to the new oil
sample. Accordingly, the density outcomes showed that oil sample S1 has higher density
reduction (2.12%) as compared to the oil sample S2 (1.27%), plausibly due to low-density
impurities in the oil. Similarly, the amount of water content in g/g of oil was measured to be
high for S1 (1.46%) as compared to (0.966%) for S2. The major source of emulsified water in
the oil could be traced to the slack sealing, which allows humid air to enter to the oil system.
As a benchmark for a good machine, the highest allowed limit of water in the lubricating oil
should be below 0.2% by weight (Smiechowski and Lvovich, 2002). Honestly, the results of
water contents in the oil are alarming and investigation to elucidate the reasons why a large
amount of water was emulsified to the oil is required. The results of density reduction and
water content are in agreement with the previous evidence of viscosity and OEE.
3.2.3 Wear debris analysis. Engineering surfaces are fabricated to be rough enough to
permit energy transformation between surfaces. As the relative speed between them is
progressing, a third material is generated due to friction and it is called debris particles. In
Sample Code K. Vis. @ 40 8C K. Vis. @ 100 8C Viscosity index (VI) Vis. Change (%) @ 40 8C Vis. Change (%) @ 100 8C

S0 237.28 21.18 106.00 – –


S1 169.845 15.95 96.00 28.42 % 24.7 %
S2 186.6 17.80 104.00 21.36 % 15.96 %
Integrated
maintenance

189
system
management

changes of S0, S1 and

and 100 8C
Table 5.

S2 oil samples at 40
index and viscosity
viscosities, viscosity
Measurement
outcomes of kinematic
JQME fact, gears contact is a special type of surface contact that is considered a periodical contact
28,1 surface (Hong et al., 2018). In this regard, when a periodical load is applied to a mechanical
system, a cyclic surface fatigue wear takes place owing to the repeated and concentrated
stresses on the gear meshing contact teeth (Hong et al., 2018; Dupuis, 2010; Dan, 2013). Of
course, such cyclic loads cause gear pitting and perhaps spalling (Dupuis, 2010). Basically,
the main function of lubricating oil is to make a thin layer that prevents surface-to-surface
direct contact. In fact, the operation conditions such as speed, hardness, temperature, load
190 and oil properties lead to the formation of different sizes and morphologies of wear debris
particles (Hong et al., 2018; Dupuis, 2010; Dan, 2013; Raadnui, 2005). However, the presence of
solid and hard contaminant particles in the lubricating oil is of research debates (Zmitrowicz,
2005; De Oliveira et al., 2019; Xiang et al., 2014). Some arguments claimed that fine debris
particles could be entrapped into the interfacial layer between the parent surfaces and may
cause degradation to the gear surfaces (Zmitrowicz, 2005), while other studies claimed that
the presence of debris particles is fundamental to the formation of a protective tribolayer that
prevents metal-to-metal contact and readily reduces the wear rate (De Oliveira et al., 2019).
The measurements of debris particle weights have shown that the pure masses of ferrous
debris particles from S1 and S2 magnet samples were 1.27 and 1.24 g, respectively. These
values show that packaging machines 1 and 2 have large concentrations, indicating severe
wear. These outcomes are consistent with those of viscosity, density, water content and OEE.
Plate 1 presents SEM micrographs of wear debris particles of S1 and S2 oil samples.
The images of debris particles of the S1 sample displayed a wide range of particle sizes and
morphologies. Accordingly, different particles of different shapes identical to spheres,
rods, fibers and needles are clearly observed in the figure. Besides, the variety of particle
sizes and shapes represent a fertile environment for identifying the wear nature and
mechanism. Close observation of the images implies that particles greater than 100 mm
constitute a large fraction while flake-like particles are the dominant morphology.
Similarly, the images of debris particles of S2 samples are also shown in Plate 1. Evidently,
the results exhibited smaller particles and narrower distribution than of the S1 sample,
while the majority of particle sizes are below 40 mm but their shapes eventually vary. Of
course, the shape of particles could be classified into three key categories; sphere-like,
flake-like and fiber-like particles, with almost equal volume for the three identified shapes.
Besides, the analysis of particle size–shape interaction of S2 showed that sphere-like and
fiber-like particles almost all are in the range of 1–20 mm. However, the flake-like particles
are slightly larger in dimensions in the range of 20–50 mm and a few particles are above
this range.
Figure 3 displays image acquisition results and particle size distribution diagram of S1 and
S2 oil samples to evaluate wear state and severity. The histograms provide sufficient
information to identify the condition of the machines under study. In fact, the analysis was
carried out for more than 100 particles for each sample. It is observed that, for the S1 sample,
there are 35% of the particles set in the size range above 100 mm, 53% of the particles placed in
the size range 20–100 mm and 12% are in the range 1–20 mm. Unlike the S1, the analysis of the
size distribution of the S2 sample exhibited that only 6% of the particles have sizes range 100–

Mass of 20 ml oil Water content Density


Sample Before After (g/g of oil) Density reduction
Table 6. code heating heating (g/20 ml oil) % (kg/m3) %
The measurements of
density and water S0 18.87 18.83 0.04 0.21 943.5 –
content of oil samples S1 18.47 18.20 0.27 1.462 923.5 2.12
S0, S1 and S2 S2 18.63 18.44 0.18 0.966 931.5 1.27
Integrated
maintenance
management
system

191

Plate 1.
Low- and high-
magnification SEM
micrographs of wear
debris particles
collected from oil
samples of packaging
machine1 (S1) and
packaging machine 2
(S2) in the Sandwich
production line

Figure 3.
A histogram displays
the wear debris particle
sizes distribution as
measured from SEM
images of S1 and S2
samples. The
histogram represents
the size range in mm on
the x-axis and the
number of particles on
the y-axis. S1 and S2
denote the debris
samples of packaging
machines 1 and 2,
respectively

200 mm, 71% of the particle sizes were found in the range 20–100 mm and 23% are below 20 mm.
Of course, a critical analysis goes toward the particles that have sizes greater than 100 mm (Dan,
2013). In numbers, 35% of S1 particles have sizes above 100 mm as compared to 6% of the S2
particles. The aforementioned findings loudly imply that the conditions of machine 1 are too
JQME critical than Machine 2. What is more, the presence of ∼71% of particles in the size range of
28,1 20–100 mm suggests the abnormal operation of machine 2 (Hong et al., 2018).
Based on the aforementioned analysis, the plausible wear mechanism could be elucidated
in these few points: combined rolling fatigue and severe sliding (flake-like particles with
blocky and flat surfaces) is the major wear mechanism of the transmission system of
packaging machine 1 (S1). Basically, this mechanism is dominated for gears where high loads
and slow motion could offer a good environment to large pitting (Dan, 2013). Besides, the
192 occurrence of large flakes with parallel grooves and striations is another warning of the
severe sliding (Dan, 2013; Raadnui, 2005). More to the aforementioned point, the
concentration of sphere-like particles increases rapidly as the service life decreases (Dan,
2013). The presence of fiber-like straight and curved particles of aspect ratio in the range 5–15
is also a mark of the severe cutting wear mechanism (Raadnui, 2005). Consequently, a similar
mechanism of cutting, fatigue and sliding wear could be assigned for machine 2 (S2), but with
low intensity than that of machine 1 (S1), in spite of their similarities in the design and the
operating conditions.
Plate 2 shows high-magnification images of debris particles of S1 and S2 oil samples. A
magnified surface investigation of the large flake-like particles of S1 indicated that their
surfaces exhibit high roughness. The outer layer of the surface looks like a large number of
fibers aggregated together to form this irregular layer. This phenomenon could be explained
in which large and relatively smooth surface particles were removed from the gear teeth at an
early stage and act as a rolling bearing and protective tribology medium to reduce the friction
rate and improve the tribology characteristics of lubricating oil. Basically, the difference in
hardness between the debris particle and the gear surface when it comes into contact could
explain the fatigue, grooves and irregularity on the surface of debris particles (De Oliveira
et al., 2019). In fact, it was claimed that ferric debris in oil habits the formation of an oxide
layer, which functions as a protective tribolayer (Xiang et al., 2014). A similar behavior could
also be observed from the high-magnification SEM image of S2 (Plate 2). The image tells that
the morphology of flake-like particles has surface features quite similar to those of S1, in spite
of their size difference. Mostly, the circulation of lubricating oil containing debris particles
could offer a secondary contact between gear teeth and debris particles, which in turn cause
the existence of new debris particles but with small features. This process is clearly observed
in the S1 sample (Plate 2), where small spheres and fibers are seen in bulk around the large
particle.
It is worth mentioning that packaging machines 1 and 2 were located in parallel at the end
of the Sandwich production line (see Figure 1) to prevent bottlenecks and enhance the
production rate. The aforementioned analysis of lubricant oil has inferred that the gearbox of

Plate 2.
High-magnification
SEM images for
surface information of
large debris particles
related to packaging
machine 1 (S1) and
packaging
machine 2 (S2)
packaging machine 1 is in catastrophic situations as compared to the gearbox of packaging Integrated
machine 2, which is also in an abnormal condition. Therefore, both packaging machines maintenance
require intensive efforts to investigate the causes of excessive wear and support a robust
maintenance decision that is able to prevent machine failures. Indeed, the root causes of oil
management
deterioration could be summarized in the following technician and managerial points: system
(1) time-based maintenance duties such as the time of replacing oil and filters are not strictly
followed; researchers’ observations have found used oil outside the sealing indicating a
plausible leakage, while 5 whys technique has found that the technicians often go for adding 193
more new lubricant oil whenever the oil level decreases in the gearbox without identifying the
causes of oil degradation during operation. In fact, the gradual decrease in oil level is another
warning signal of machine failure. It means that the oil is leaking or evaporating into the
surrounding, which is another warning sign to the wear-out machine; if the oil not sealed
properly, it is plausibly for air to enter to the oil case while recent studies have inferred that
humid air in the oil increases the wear rate promptly. The wear mechanism could be
attributed to the presence of water vapor in air, which habits oxidation of metallic parts
(Blades et al., 2020); (2) the useful lifetime of the gearboxes is almost over (the packaging
machines were fixed in 1993); (3) by referring to the manual, we found that the company uses
a substandard low-viscous oil of grade ISO VG 68 as compared to ISO VG 150 grade that
should be used as per the manufacturer’s specification. Accordingly, low oil viscosity
prompted to reduce the film thickness of the oil between the mating parts. All the
aforementioned root causes promote the increase in wear debris size and concentration.
Finally, from a maintenance view of point, the analysis of debris particles concentration,
size distribution and shape could come to the conclusion that the transmission system of
packaging machine 1 undergoes a disastrous failure mode and immediate overhaul
maintenance or replacement has to promptly be decided based on the evidence furnished
here. Furthermore, the transmission system packaging machine 2 is also under critical
conditions and requires to implement a mixed preventive/predictive maintenance strategy
periodically to prevent machine failure.
Figure 4 illustrates the root causes of failures and studies their influences on the
productivity of Line 2. The diagram classifies the root causes to four main partitions: (1)
The raised problems due to the materials and machines could be summarized in the
communication with suppliers, bad inventory management, old equipment and the nature

Figure 4.
Fishbone diagram
illustrating the major
causes–effects of
failures in Line 2.
Causes written with a
brown color could be
elucidated from OBM
analysis
JQME of products, which require more care during transmission and packaging; (2) technicians
28,1 and employee play a crucial role in the whole processes. The low level of experience,
training, income, incentives and rewards in the company are the most problems of
technicians; (3) the maintenance work employs run-to-failure strategy with a low-level of
scheduled-based maintenance. No workshop facilities while the tools of maintenance and
measurements are not completed. The long period of equipment in service explains the
high-frequency failure and damaged products; (4) the management is family-based, unable
194 to invest sufficient money for maintenance purposes and also reluctant to use advanced
systems for management as well as maintenance. These causes affect the productivity and
profitability of the production lines.

4. Conclusions
An OEE/OBM integrated diagnostic and prognostic approach was successfully implemented
in three production lines of a biscuit factory and the outcomes are conforming. This
integration approach provides robust information on the current conditions of each line. The
results of OEE metric inferred that Line 2 is the most risky one as compared to other lines. A
magnified analysis of the six major machines on the Line 2 using the OEE approach has
conveyed that the packaging machines 1 and 2 are in catastrophic conditions, showing the
lowest values in terms of failure frequency, deficient products, and bottleneck. In addition, the
OEE analysis finds that 43 and 46% for the packaging machines 1 and 2, respectively,
constitute ∼58% of breakdown losses of the production line while other equipment in the line
are in good condition. The unusual losses obtained by OEE analysis were authenticated
using mechanical system that can investigate the health conditions of these machines on the
production line. The analysis of used lubricating oil of packaging machines 1 and 2 has
expressed a high reduction in oil viscosity and density and large water content much more
than the allowable limits for good lubricants. A comprehensive analysis of wear debris
particles comes with interesting findings as a proactive monitoring tool. The quantitative
results of particle size distribution have elucidated that 36 and 6% particles are above
100 mm, which indicated catastrophic conditions of the corresponding machines as OEE
analysis also suggests. The wear mechanism investigation has found that combined rolling
fatigue, severe sliding and cutting wear are the plausible wear mechanism of Packaging
machines 1 and 2, indicating overloads and abnormal operating. Different analysis
techniques were implemented such as 5 whys, fishbone, Pareto chart, direct observations by
the research team and interviews with technicians, which make final conclusions and
recommendations to the production system. Finally, this study could be a robust feasibility
study to the factory management to immediately replace Line 2 with new lines that are
supported by online monitoring technologies. Also, suggestions for Line 1 and Line 3 were to
take immediate action to implement efficient maintenance management and invest sufficient
budget for maintenance.

References
Ahuja, I.S. and Khamba, J.S. (2008), “Total productive maintenance: literature review and directions”,
International Journal of Quality and Reliability Management, Vol. 25 No. 7, pp. 709-756.
Al-Nozili, M.S., Abeed, F.A. and Ahmed, M.M. (2014), “Evaluation of the lubricating oil’s optimal drain
interval for petrol public transport means at the local operating conditions”, The International
Journal of Emerging Technology and Advanced Engineering, Vol. 4 No. 12, pp. 79-87.
Aminuddin, N.A.B., Garza-Reyes, J.A., Kumar, V., Antony, J. and Rocha-Lona, L. (2016), “An analysis
of managerial factors affecting the implementation and use of overall equipment effectiveness”,
International Journal of Production Research, Vol. 54 No. 15, pp. 1-18.
ASTM D2270 – 10 (2016), Standard Practice for Calculating Viscosity Index from Kinematic Viscosity Integrated
at 40 and 100 8C, West Conshohocken, PA.
maintenance
ASTM D445 – 06 (2018), Standard Test Method for Kinematic Viscosity of Transparent and Opaque
Liquids (And Calculation of Dynamic Viscosity), West Conshohocken, PA.
management
Blades, L., Hills, D., Nowell, D., Evans, K.E. and Smith, C. (2020), “An exploration of debris types and
system
their influence on wear rates in fretting”, Wear, Vols 450–451, pp. 203-252.
Braglia, M., Frosolini, M. and Zammori, F. (2009), “Overall equipment effectiveness of a manufacturing 195
line (OEEML) an integrated approach to assess systems performance”, Journal of
Manufacturing Technology Management, Vol. 20 No. 1, pp. 8-29.
Cao, W., Zhang, H., Wang, N., Wang, H.W. and Peng, Z.X. (2019), “The gearbox wears state
monitoring and evaluation based on on-line wear debris features”, Wear, Vols 426–427,
pp. 1719-1728.
Cheah, C.K., Prakash, J. and Ong, K.S. (2020), “Overall equipment effectiveness: a review and
development of an integrated improvement framework”, International Journal of Productivity
and Quality Management, Vol. 30 No. 1, doi: 10.1504/IJPQM.2020.107240.
Da Costa, S.E.G. and de Lima, E.P. (2002), Uses and Misuses of the Overall Equipment for Production
Management, Cambridge, August 18–20, s.n.
Dan, R.M. (2013), Multi-view and Three-Dimensional (3D) Images in Wear Debris Analysis (WDA), PhD
thesis, University of Manchester.
De Oliveira, M.M., Costa, H.L., Silva, W.M. and De Mello, J.D.B. (2019), “Effect of iron oxide debris on
the reciprocating sliding wear of tool steels”, Wear, Vols 426–427, pp. 1065-1075.
Dupuis, R. (2010), “Application of oil debris monitoring for wind turbine gearbox prognostics and
health management”, Annual Conference of the Prognostics and Health Management Society,
pp. 6-10, Portland, Oregon, USA.
Eti, M.C., Ogaji, S.O.T. and Probert, S.D. (2006), “Reducing the cost of preventive maintenance (PM)
through adopting a proactive reliability-focused culture”, Applied Energy, Vol. 83 No. 11,
pp. 1235-1248.
Garg, A. and Deshmukh, S.G. (2006), “Maintenance management: literature review and directions”,
Journal of Quality in Maintenance Engineering, Vol. 12 No. 3, pp. 205-238.
Hong, W., Cai, W., Wang, S. and Tomovic, M.M. (2018), “Mechanical wear debris feature, detection,
and diagnosis: a review”, Chinese Journal of Aeronautics, Vol. 31 No. 5, pp. 867-882.
Jain, A., Bhatti, R. and Singh, H. (2014), “Total productive maintenance (TPM) implementation
practice: a literature review and directions”, International Journal of Lean Six Sigma, Vol. 5
No. 3, pp. 293-323.
Kent, M.D., Costello, O., Phelan, S. and Petrov, K. (2018), “Cost oriented maintenance management
systems for manufacturing processes”, IFAC-Papers OnLine, Vol. 51 No. 30, pp. 48-53.
Kumar, J. and Soni, V.K., (2015), “An exploratory study of OEE implementation in Indian
manufacturing companies”, Journal of The Institution of Engineers (India): Series C, Vol. 96
No. 2, pp. 205-214.
Nakajima, S. (1988), Introduction to TPM: Total Productive Maintenance, 11th ed., Productivity Press,
New York.
Nurprihatin, F., Angely, M. and Tannad, H. (2019), “Total productive maintenance policy to increase
effectiveness and maintenance performance using overall equipment effectiveness”, Journal of
Applied Research on Industrial Engineering, Vol. 6 No. 3, pp. 184-199.
Piran, F.A.S., De Paris, A., Lacerda, D.P., Camargo, L.F.R., Serrano, R. and Cassel, R.A. (2020), “Overall
equipment effectiveness: required but not enough-an analysis integrating overall equipment
effect and data envelopment analysis”, Global Journal of Flexible Systems Management, Vol. 21,
pp. 191-206.
JQME Prabhakar, D. and Raj, J. (2014), “CBM, TPM, RCM and A-RCM - a qualitative comparison of
maintenance management strategies”, International Journal of Management and Business
28,1 Studies, Vol. 4 No. 3, pp. 49-56.
Raadnui, S. (2005), “Wear particle analysis – utilization of quantitative computer image analysis: a
review”, Tribology International, Vol. 38, pp. 871-878.
Randall, R.B. (2011), Vibration-Based Condition Monitoring: Industrial, Aerospace and Automotive
Applications, 1st ed., The Atrium, Southern Gate, Chichester, West Sussex.
196
Raouf, A. (1994), “Improving capital productivity through maintenance”, International Journal of
Operations and Production Management, Vol. 14 No. 7, pp. 44-52.
Sakib, N. and Wuest, T. (2018), “Challenges and opportunities of condition-based predictive
maintenance: a review”, Procedia CIRP, Vol. 78, pp. 267-272.
Smiechowski, M.F. and Lvovich, V.F. (2002), “Electrochemical monitoring of water/surfactant
interactions in industrial lubricants”, Journal of Electroanalytical Chemistry, Vol. 534,
pp. 171-180.
Soni, A. and Patel, V.K. (2017), “Review and study of effect of water contaminants in lubrication in ball
bearing through vibration analyses”, Materials Today: Proceedings, Vol. 4, pp. 2717-2722, .
Wakiru, J.M., Pintelon, L., Muchiri, P.N. and Chemweno, P.K. (2019), “A review on lubricant condition
monitoring information analysis for maintenance decision support”, Mechanical Systems and
Signal Processing, Vol. 118, pp. 108-132.
Wudhikarn, R. (2012), “Improving overall equipment cost loss adding cost of quality”, International
Journal of Production Research, Vol. 50 No. 12, pp. 3434-3449, doi: 10.1080/00207543.2011.
587841.
Wudhikarn, R. (2013), “A framework for integrating overall equipment effectiveness with analytic
network process method”, Internal Journal of Innovation, Management and Technology, Vol. 4
No. 3, pp. 351-355.
Wudhikarn, R. (2016), “Implementation of the overall equipment cost loss (OECL) methodology for
comparison with overall equipment effectiveness (OEE)”, Journal of Quality in Maintenance
Engineering, Vol. 22 No. 1, pp. 81-93.
Xiang, L., Gao, C., Wang, Y., Pan, Z. and Hu, D. (2014), “Tribological and properties of magnetite
nanoflakes as additives in oil lubricants”, Particuology, Vol. 17, pp. 136-144.
Zmitrowicz, A. (2005), “Wear debris: a review of properties and constitutive models”, Journal of
Theoretical and Applied Mechanics, Vol. 43 No. 1, pp. 3-35.

Corresponding author
Mokhtar Ali Amrani can be contacted at: mokh_ali2007@yahoo.com

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

You might also like