1.

4 Manufacturing Shop Floor- Heart of Manufacturing

Due to the Industry 4.0 initiative, industrial manufacturing has witnessed histrionic upheaval
of technology, but few features remain unchanged. The manufacturing shop floor is still
considered as the heart of any manufacturing operation, and it is a place where equipment,
processes, and people merge collectively to add value to material and manufacture products
for trade (Demchenko, Ngo and Membrey, 2013; Data Magnum- The Big Data Connection,
2015). In simple words, the manufacturing shop floor is an area where production is carried
out manually by workers, semi-automatic by workers and machines, or automatic by
machines. This area includes inventory, warehouse, and equipment. It is common in
manufacturing to contemplate the shop floor operations as the least important factor to be
addressed by the manufacturers (Demchenko, Ngo and Membrey, 2013; Data Magnum- The
Big Data Connection, 2015; Russwurm, 2014). Asymptomatic issues that can be measured
using KPIs such as idle equipment, inventory inaccuracies, emergency breakdown, worker
inefficiency, or stations/equipment/machine waiting for attention can go unnoticed until a
severe customer-affecting event arises. Overlooking shop floor operations is a grave insight
(Deloitte, 2015). Problems that may initially appear insignificant, such as unexplained
variations in the cycle time for a manufacturing operation, can swiftly escalate and become a
significant issue (or result in a substantial setback)—affecting the overall production cycle,
customer order fulfillment time, profit, and growth across many internal departments (Chong,
Ng, and Goh, 2016).
This research is mainly focused on the manufacturing shop floor KPIs because it reveals the
loopholes in production or process that can severely impact manufacturers through efficient
monitoring. It can also reduce the occurrence of errors in the output or process by enabling
proactive rather than reactive decision-making, finding deviation from standard procedures
and processes, and increasing staff productivity by continuously monitoring their
performances and enabling the smart scheduling of staff and resources operations. Thus,
assisting the manufacturing industries in bringing out usable, tangible, and superior quality
products as per the customer specifications from their shop floor seamlessly.
1.4.1 Identifying the Challenges Faced by Manufacturers While Selecting KPIs
According to Kucukaltan et al., in manufacturing industries, performance measurements are
based on numerous performance indicators. The information on determining the right
performance indicators (i.e., KPIs) and identifying the interrelationships between several
performance indicators have been lacking. For manufacturers, selecting and implementing
KPIs to remain competitive in the current business environment is severe. These challenges
include:
What are the appropriate KPIs?;
How many KPIs are needed?;
How to prioritize the KPIs?;
How can these KPIs help to improve performance?;
Finding relevant KPIs to measure manufacturing performance is challenging and time-
consuming. It considers all current and future business objectives and targets the weak spots
in the production that needs attention. There are no KPIs mentioned in the literature that can
give manufacturers a complete picture of the shop floor performance. Each KPI presents a
partial insight into the performance and hence not sufficient to serve as a source for decision-
making. For this reason, the selection of a set of KPIs that can capture the complete picture of
the shop floor performance is a necessary but challenging task for the manufacturers.
When selecting KPIs and using them efficiently, everyone involved in the manufacturing
facility must know:
How the KPIs affect their manufacturing operation (work or responsibilities)?;
Do the KPIs help them to improve their manufacturing operations?;
What kind of changes shall be incorporated from the information received by KPIs?
Moreover, the selection of KPIs must be in-line with the approach and business objectives
that the manufacturers are aiming to accomplish. In a complex business environment, where
the objectives need to be transformed over time, the KPIs selected should also be altered. A
lack of actionable KPIs can cause delays in identifying weak spots and severely impact
performance improvement. In order to improve overall performance, manufacturers should
carefully answer the following questions while identifying and selecting KPIs for monitoring
their shop floor operations:
Is there a link between KPIs and business objectives?;
How many KPIs should be used?;
How often should the KPI be monitored? KPI accountability?
To address these challenges, the manufacturers face no framework or methodology to
systematically, methodically, and/or scientifically select KPIs for a manufacturing facility. A
KPI guidelines approach is developed to guide the manufacturers to analyze and choose
appropriate KPIs systematically. Furthermore, prioritizing the KPIs is also discussed within
the scope of this research using SMART criteria.
1.4.2 Manufacturing Shop Floor-Related Problems
Visibility—from the many tools and techniques that are used as a part of the Industry 4.0
initiative, manufacturers have several options for managing and monitoring their shop floor.
Because there may be siloes of data from non-integrated systems, overall operation visibility
might be hard to achieve and 26 challenging to organize these fragmented reports together
(Kaifei He, 2016). Lacking real-time integration to a single version of manufacturing status,
identifying errors or inefficiencies may be impossible (Donald N. Heirman, 2002). Every
interruption in production means the waste of resources and failure to fulfill customers'
demands on time. Manufacturers might have a nebulous sense that the production throughput
KPI is not "where it should have been" disruptions "are hard to control" and "something"
needs to be fixed. But what? Maybe machines in a few stations are failing to keep up with the
production takt time? Perhaps communication gaps between different areas of the shop floor
are delaying orders? Maybe labors are performing inefficiently? Maybe shop floor managers
have limited knowledge? Maybe they do not measure the right performance indicators needed
to improve the shop floor efficiency.
With the lack of shop floor data visibility, manufacturers are stuck with reactive rather than
proactive decision-making. These types of problems in manufacturing can be expected and
essential characteristics of complex manufacturing. A shop floor with complete visibility
typically will potentially reduce these types of problems.
Automation—for any manufacturing industry, it is essential to recruit and retain a skilled
workforce. However, on the shop floor, spotting idle personnel is expected. Often the
production staff is made to wait for the final product confirmation or authorization from the
customer on the design amendments (Krumeich et al., 2014). For some manufacturers,
understanding labor costs and inadequacies is a complex challenge. Automation stands as the
solution to many of these issues but it is crucial to experience which jobs to automate. Most
manufacturers lack insights into the shop floor, directly impacting the overall production rate
and time invested in fulfilling customers' orders. Automation can also help manufacturers to
achieve high product quality and guarantee compliance with corporate obligations. For
instance, product quality KPI checks can be controlled and automated using sensor
technology superior to human inspection methods (Niggemann et al., 2015). Maintaining
high quality helps to improve brand image and increases customer satisfaction.
Digitalization—for manufacturers, is seen as the application of Information Communication
Technology (ICT) to enhance manufacturing processes' productivity and efficiency.
Digitalization can be defined as "embracing or escalating the use of computer technology by
an organization or industry" (Merkel et al., 2017). Regrettably, the acceptance of digital
technologies in the manufacturing industries is slow-paced, with most manufacturers
adopting a 'wait and see' method. Digitalizing industrial processes should be viewed as a
continuing process, not just as a one-stop solution, as manufacturers incessantly evaluate
'what is worth digitalizing' (Prades et al., 2013). In the current industrial revolution, digital
technologies, such as RFID scanners, telemetry, machine sensors, and GPS tracking, are
evolving at a rapid pace. To survive and thrive, manufacturers have to adapt to and utilize
these new technologies. Shop floor digitalization can be applied across manufacturing
operations such as material handling, scheduling, inventory, planning, logistics, labor, and
time tracking (Georgoudakis et al., 2006).
Out-of-date shop floor equipment, poorly combined with inadequate software solutions,
cannot offer the much-needed dexterity and visibility required by the shop floor to meet
current market demands. The shop floor is where most manufacturing profits are made, and
customer satisfaction is retained (Williams, 1994). If the shop floor is functioning poorly, the
entire organization suffers. It is the reason why manufacturers should consider shop floor
processes as their top priority. Digitalizing shop floor processes certainly makes everyone
involved in manufacturing jobs easier. Digitalization makes the complete manufacturing
process completely secured, automated, continually monitored and managed as a global
entity. That, in turn, helps achieve reduced downtime, improved efficiencies, increased return
on investment, and greater resource utilization (Hankel and Rexroth, 2015).
1.5 Need for KPIs in Manufacturing Shop Floor
The significant reasons for deploying KPIs in manufacturing shop floor are (Bhanot, Rao and
Deshmukh, 2015, 2017; Ghazilla et al., 2015; Star et al., 2016; Zackrisson et al., 2017;
Zailani et al., 2017):
Monitoring the health: KPIs acts as an essential tool for manufacturing industries to track and
monitor their operational and strategic performances. The bottleneck equipment, process,
station, or production line can be acutely observed to increase performance by reducing
downtime. The major challenge arises when the KPIs selected for monitoring are
inappropriate and incomprehensible by the manufacturers leading to further degradation in
performance. The number of KPIs used for monitoring and its accountability plays a
significant part in sustaining improved industrial health. The selected KPIs should
incorporate each department of production, namely: product, process, and resource, making
sure that overall improvement in performance is achieved.
Measuring progress over time: KPIs such as cycle time, stop time, machine idle time, etc.,
need to be measured in real-time to improve quick decision-making capabilities. Strategic
KPIs, for instance, employee performance, production line monthly performance, gross
margin, and revenues, need to be measured periodically since they are result indicators used
to track the industry's progress (weekly, monthly, quarterly, or annually) towards achieving
its strategic objectives. With appropriate selection of KPIs that can be measurable in the real-
time, periodical, and on-demand can support manufacturers to monitor their performance
both in short- and long-term perspectives.
Making modifications and tackling opportunities: to survive in the current complex
environment, manufacturing industries make frequent alterations to their operational and
strategic objectives. KPIs monitors the performances based on new objectives and support
manufacturers to make modifications to the set objectives. These modifications can be
achieved by selecting the right set of KPIs, which has a balanced set of leading and lagging
indicators. New opportunities, such as increasing the production rate, throughput rate, etc.,
can be realized if the manufacturers succeed in achieving their objectives. Therefore,
manufacturing industries depend on the results generated by KPIs to make modifications and
tackle opportunities.
Analyzing patterns in performance over time: manufacturing industries tend to measure a
small number of commonly known KPIs frequently over time, such as production ratio,
quality, OEE, etc. Measuring KPIs over time (quarterly, annually, or a more extended period)
tends to generate a pattern. These patterns tend to support manufacturers in countless ways.
For example, patterns can help them be aware of the lowest quarter of production rate and use
that time to carry out maintenance activities or other productive initiatives. It can help to
know the shifts' performances over a certain period, which shift underperforms or over-
performs. Patterns act as a forecaster for manufacturers to understand what can be expected
in the next production phase.

Categorization of KPIs Based on Various Research Publications

Kang et al. understanding the complexity of selecting the right KPIs for improving manufacturing
system operations, suggested categorizing the KPIs, mainly into primary and comprehensive
KPIs; and supporting metrics. Basic KPIs were used to represent a single aspect of system
performance, and comprehensive KPI was used to describe the overall performance. The
supporting metrics were directly monitored elements. The categorization was based on a
hierarchical structure with several categories at several levels, as shown in figure 8. The
supporting elements level included time and quantity related to production, maintenance, and
quality-related to the machine, order, and worker. These supporting elements help compute basic
KPIs and desired mathematical operations over primary KPIs to raise comprehensive KPIs. Kang
et al. failed to discuss critical parameters: the audience, timing, and production methodology
essential for KPI selection (Kang et al., 2015).

Li et al. worked on one of the limitations of the ISO 22400 standard (i.e., standard applicability).
The KPIs mentioned in ISO 22400 standard are applicable for discrete production, making them
technically unsuitable for the process industry. Li et al. proposed a framework for classifying
KPIs in the process industry. The framework follows a similar hierarchy mentioned in ISO 22400
and can be evaluated in three levels: equipment KPIs, measurement elements, and process KPIs.
Measurement elements, the middle level of the framework contains the data that can be directly
collected and monitored throughout the production process, as shown in figure 9. Equipment and
process KPIs are calculated based on the measurement elements. This framework provides useful
ideas for decision-makers and manufacturing engineers to describe and measure appropriate KPIs
for process industry process assessment. Li et al. also provided the description and formula for
the KPIs but failed to discuss few parameters: the audience, timing, and production methodology
essential for KPI selection (Li et al., 2009).

Hester et al. proposed a method for KPI assessment in manufacturing organizations. The
proposed method is profoundly dependent on the stakeholder contribution at varying levels
during the complete course of the KPI assessment process. The aim is to improve KPI assessment
methods by introducing a mathematical foundation centered on value-focused thinking. The steps
involved in the assessment are shown in figure 10, divided into three main activities: preparatory
manufacturing activity, stakeholder values, and preference elicitation activity, and manufacturing
company KPI characterization and analysis activity. Ranking criteria used for assessment are
based on KPI characterization.
The parameters which are considered for KPI characterization are: verified, standardized,
relevant, predictive, quantifiable, inexpensive, actionable, inexpensive, accurate, documented,
independent, traceable, understandable, and timely. This proposed method lets the stakeholders
assess the organization’s KPIs to determine its performance compared to predetermined KPI
thresholds. A case study was conducted within a chemical manufacturing company to test its
validity. With stakeholders playing a pivotal role in the KPI assessment method, any failure in
understanding the manufacturing process and KPIs can result in the incompetence of the
developed method (Hester et al., 2017).

Carlucci et al. addressed one of the significant challenges faced by manufacturing industries, i.e.,
selecting the most meaningful KPIs for monitoring manufacturing operations. The KPI selection
was described as an MCDM problem involving several factors and associated interdependencies.
Using the ANP method, a decision-based model was proposed to help manufacturers in KPI
selection. This model was based on the contemplation that KPIs can be selected and evaluated
based on a set of criteria. These set of criteria are relevance, reliability, understand-ability,
representational quality, comparability, and consistency. A set of relevant questions is prepared
for the decision-makers founded on criteria to assess the manufacturing company's performance.
While selecting KPIs, it was mentioned that decision-makers do not rigorously consider the
dependency of criteria and interdependencies among the chosen indicators, compromising the
indicators' quality.
The developed model provides a more viable approach to deal with this problem consisting of
two clusters, namely criteria and performance indicators, as shown in figure 11. Carlucci et al.
proposed a model to help decision-makers select KPIs' best set for their manufacturing
operations. For this model to work competently, decision-makers must know the manufacturing
process and all the relevant KPIs available in the literature to monitor the performance. It is
because the proposed model only provides that KPIs selected by decision-makers are appropriate
or vice-versa. Decision-makers do not always have to compete for knowledge of the KPIs
available in the literature, so this does not seem to work efficiently (Carlucci, 2010).

Lindberg et al. state that benchmarking KPIs against other KPIs from comparable equipment,
process, and industry can identify and select appropriate KPIs to monitor industrial performance.
A case study was carried out on a heat and power plant; a set of KPIs was suggested to the
managers after benchmarking them with the literature on similar industries. The list of suggested
KPIs was divided into energy, raw-material, equipment, operation, control performance,
maintenance, planning, inventory, and buffer utilization KPIs. It was seen that the boiler
efficiency was considerably improved by deploying benchmarked KPIs in their heat and power
plant. With the surge in the literature related to KPIs, using the benchmarking method for KPI
selection would result in using outdated KPIs that have been replaced with more efficient ones or
new KPIs that have been generated recently. Also, for KPI selection, the manufacturers need to
know its description, audience, timing, production methodology, etc. Without a complete
understanding of the KPI, it would merely remain a number (Lindberg et al., 2015).
kibira et al. present a procedure for evaluating, monitoring, and improving KPIs for sustainable
manufacturing industries. The purpose of this procedure was to address KPIs' inconsistent
meanings, lack of compelling selection, and evaluation methods for environmental KPIs in
manufacturing processes. The procedure was developed based on ASTM International standard
guidelines. The development stages include identifying appropriate KPIs from existing literature,
defining new KPIs if needed, selecting KPIs based on set criteria, and assigning weights. The
weights are used to reduce deteriorating effects that result from emphasizing similar or
interrelationship KPIs.
By analyzing the relations between the number of KPIs and their supporting matrices, it can be
probable to compute a similarity score to select appropriate KPIs. For instance, if several KPIs
are calculated using the same supporting matrices, it can be possible to compute a similarity score
to help in effective KPI selection. Using repositories that store all the relevant manufacturing
processes, KPIs were used to provide pre-defined KPIs for initial selection. The procedure is
based on both quantitative methods and human judgment. Humans are the think tank experts
aware of all the key activities critical for successful business performance. Figure 12 highlights
the keys steps involved in the designed procedure (Kibira et al., 2018).

Ding et al. develop a data-driven scheme to predict and diagnose KPIs for static and dynamic
processes. It aims to apply this scheme to solve the KPI prediction and diagnosis problems faced
in the steel manufacturing industry's hot strip mill process. This paper discusses the two most
popular data-driven methods to achieve fault diagnosis tasks, i.e., Partial Least Squares (PLS) and
Principal Component Analysis (PCA) for complex industrial processes. Despite having wide
applications in the industry, these schemes have been reported to have certain limitations. Hence,
the paper builds upon these schemes to give two algorithms for KPI computation and prediction
for static and dynamic processes. To implement KPI prediction in the hot strip mill process, a
four-step procedure is defined in the paper, which includes defining the KPI, variable selection,
off-line training, and finally, on-line prediction and diagnosis.
KPI is to be defined according to specific process characteristics, quality requirements, and
economic benefits. Variables that are responsible for the given KPIs are selected. This
implementation shows that the dynamic approach's prediction performance is better than that of
the static approach. There is a smooth variation between the actual and predicted values of the
process variable. The dynamic process also gives more useful results for fault-detection.
However, Ding et al. suggest that selecting a static or dynamic method should depend on the
underlying process's nature, memory cost, computational cost, design complexity, and the
required robustness. This paper provides an insight into which methods should be utilized for KPI
prediction and diagnosis in complex processes where a sophisticated modeling procedure is not
realizable. Therefore, model-based techniques cannot be utilized.
Tatsiana et al. identify process-based KPIs that can control the pre-processing stage in a
construction project. This is done by conducting a pilot-study and validating its results through
interviews of experts in this industry. It is first identified that even though the pre-project stage is
vital to a project’s success, it does not always perform well in the construction industry leading to
time delays and budget overruns. Therefore, the project performance must be governable in the
early stages of the construction process. An extensive literature review conducted by the authors
highlights that most of the KPIs in the construction industry focus on end-products rather than
providing measures to control the process while it is in progress, meaning there is a lack of
process-oriented KPIs, which their paper aims to develop.
A pilot study was conducted in which experts were asked to review a list of main sub-processes
in the pre-project stage, and as a result, eight process-based KPIs were developed. These were
again validated by experts mainly from large construction organizations as they experienced
dealing with large projects' early stages. This method is advantageous because large projects can
be challenging to control, involve many

stakeholders with differing interests that are tough to align. Bad decisions taken in the early
stages of the process can have massive consequences for the later stages. After considering the
suggestions and comments of these experts, the KPIs were reduced from eight to five.
These are initial problem definition, management of client requirements, and alignment of
stakeholders’ requirements, design solution, and stakeholder involvement. The experts pointed
out that it is necessary to align clients' needs on different levels before defining and managing
client requirements. Additionally, the alignment of interests of different stakeholders involved by
defining their objectives and influence on the project was also deemed essential. Keeping the
stakeholders involved continuously in the project regarding the decisions that are to be made is
also one of the critical indicators of performance.
Gonzalez et al. aim to establish a useful set of KPIs to measure maintenance services'
performance using an MCDM methodology. After a comprehensive literature review of different
MCDM methods, the authors propose a methodology based on the original ELECTRE I method,
which uses a preference model that uses the concept of concordance and discordance to carry out
paired comparisons between the alternatives. This method has evolved to allow the ranking of
alternatives from best to worst with complementary analyses. This ranking method is deemed
suitable for selecting KPIs in the maintenance framework as it has been successfully applied to
other ranking problems. Moreover, it is easy to use, and the logic behind it is rational, and the
computation process systematic and well-organized. It is also the most used method for ranking
alternatives. The first step in the proposed methodology is defining the maintenance management
objectives. Then a set of criteria to evaluate the KPIs are specified to which appropriate weights
are assigned. Based on the defined objectives, KPIs are pre-selected to be assessed as competing
alternatives. Then those alternatives are evaluated against the selected criteria by the decision-
makers. The higher the response level on the Likert scale, the more critical it is for KPI (Gonzalez
et al., 2017).
Effendi et al. emphases on the problems of KPI development in an aerospace manufacturing
enterprise. The KPIs were targeted to provide the enterprise with the best PM tools to remain
cost-competitive and effectively implement lean manufacturing. Through techniques such as
internal benchmarking, intensive literature review, and best industrial practices, a set of KPIs was
developed and measured both subjectively and objectively. To analyze and investigate the
communication issues between different manufacturing levels (ISA-95 architecture), focus group
discussions and semi-structured interviews were used. To overcome these issues, a standard
report was generated. The enterprise was suggested to recruit personnel at every manufacturing
level to report and review their manufacturing process to their subordinates’ regularly. This paper
highlighted several problems related to KPI developments, such as lack of ownership of
responsibility, misinterpretation of company objectives, no normalization of reporting, and
difficulty measuring employee performance in the given aerospace manufacturing enterprise.
Nevertheless, the paper could not resolve these problems apart from lack of ownership of
responsibility (Effendi et al., 2008).
Rødseth et al. focused on the challenges faced by silo departments working independently in a
number of disciplines within manufacturing industries resulting in a sub-optimized effect on
production. It is due to poor communication between different departments and disciplines
working together. The disciplines focussed were maintenance management and Manufacturing,
Planning & Control (MPC). The purpose of this paper was to identify the relevant KPIs through a
literature review for these disciplines. Accordingly, throughput time was marked as an essential
KPI for MPC; OEE for maintenance management and MPC disciplines used integrated planning
concepts. This paper also highlights the relationship between several leading and lagging
indicators. However, identifying two indicators to optimize production performance by
improving communication between two disciplines was not realistic since manufacturing
industries are placed where many disciplines have to work simultaneously to achieve
organizational objectives. Moreover, two KPIs are insufficient to solve the silo problem in
manufacturing industries (Rødseth, Strandhagen, and Schjølberg, 2015).
May et al. addressed the challenges faced by current manufacturing industries in supporting the
development of energy-related KPIs (e-KPIs). The challenges included a lack of applicable KPIs
to compare energy-use profiles of equipment, processes and benchmarking them alongside
competitors’ energy performance measures. Aiming to address these challenges, this paper's key
motive was to outline a method to assist manufacturing industries in e-KPIs development. The
paper presents an e-KPIs development method containing the following sequential steps:
definition of the reference manufacturing system; identification of manufacturing resource power
requirement; investigation of the causes related to energy inefficiencies through exploring
manufacturing states of the given resource; connecting energy requirements with appropriate time
drivers; constructing a hierarchical structure of power resource energy consumption; e-KPIs
development, design, and management.
The developed e-KPIs allows the manufacturers to interpret cause-effect relationships and help
them in making enhanced operative decisions. This method also identifies strong and weak areas
for energy efficiency enhancements related to MPC. The granularity of the accessible data is one
of the drawbacks of the developed method. In the current industrial practices, the energy
monitoring is done at level-2, a finer granularity at level-0, and 1 of the ISA-95 functional
hierarchy model is needed (May et al., 2015).
Gonçalves et al. stated that several KPIs present in the literature could be applicable for
measuring maintenance performance. Still, manufacturers must choose an appropriate set of KPIs
that impact their maintenance-related activities. Creating a suitable set of maintenance KPIs is
subjected to manufacturers’ maintenance objectives and is considerably associated with definite
business processes, systems, strategies, and contexts. The maintenance team was responsible for
selecting the KPIs, which can help them realize their business goals. Cesar et al. offered a new
method to choose necessary maintenance KPIs using a methodology based on the ELECTRE II
(ELimination Et Choice Translating REality), an MCDM method. The proposed ELECTRE I
established methodology supports manufacturers’ resolve their ranking issues by assessing
numerous conflicting options through MCDM. Since the manufacturers are generally
inexperienced with the mathematical formulation required for MCDM methods, a software tool
was designed utilizing Visual Basic for Applications (VBA).
All the needed mathematical formulation was encoded using VBA to offer computational
analysis for decision making. A case study was carried out with two maintenance experts in
service quality at the airports to validate the methodology. Using the methodology, the experts
ranked the alternatives from good to bad and steadily selected KPIs with higher interest to airport
quality service teams. The case study outcome proved that the methodology is an efficient tool to
assist maintenance teams in relevant KPI selection based on maintenance strategies and
objectives (Gonçalves, Dias, and Machado, 2015).
Alsyouf states that the “fix when it breaks”-concept, which describes the corrective maintenance
method, mostly moved towards the preventative maintenance concept after the end of the second
world war due to advancements in industrial production systems with planned maintenance
actions aiming to prevent machine breakdown from occurring (Alsyouf, 2007). Nevertheless,
with the globalization of industries and more generous than before worldwide competition,
planned maintenance methods used for replacement or repair of machine parts did not stand out
to be the best solution to solve the given problem.
Therefore, in order to make the most of industrial assets and curtail unwanted stops for running
preventative maintenance procedures, the condition-based maintenance method alongside
maintenance-related KPIs was established (i.e., Mean Time To Repair (MTTR) and Mean Time
Between Failure (MTBF)). The maintenance-related KPIs were used to closely monitor machine
behavior using simple visual check-ups to evaluate real-time data using computer-aided
technology continuously. The critical parameters were carefully observed, and with any small
variation in their values compared to a set of predetermined values, condition-based maintenance
actions were initiated.
Pereira et al. used a predictive maintenance fuzzy logic technique for on-line monitoring of
induction motors using related KPIs (Pereira and Augusto, 2016). The authors’ main concern was
using maintenance-related KPIs to analyze the induction motor's vibration spectrum resulting
from a mechanical failure such as misalignment, unbalanced disk, bearing faults, and mechanical
clearance during motor operation. This vibration spectrum was investigated using the fuzzy logic
technique. In this technique, the motor operating conditions are defined in fuzzy linguistic
variables; these variables are then exploited to present the result using monitoring programs such
as Fast Fourier Transforms (FFT). To prove the proposed control method's effectiveness, an
experimental test was performed at Dynamic Systems Laboratory in the Federal University of
São João del-Rei (UFSJ). They were able to identify and diagnose faults in the motor casing
using maintenance-related KPIs (i.e., MTTR and MTBF).
Bastos et al. present an architecture aimed to collect the data produced during the industrial
maintenance schedules and predict future fiascos based on data exploration. Rapid Miner, a
product of software innovation lab Limited Liability Company (LLC), is used to analyze
maintenance data, apply various prediction algorithms to gather data, and then compare their
accurateness in the detection of predictions and patterns using applicable maintenance KPIs.
Rapid Miner is incorporated with a real-time system that gathers data utilizing automatic agents.
These data results are presented in KPIs to the maintenance team for decision-making (Bastos,
Lopes, and Pires, 2014).
Goundar et al. work is centered on the real-time monitoring of three-phase AC industrial
induction motors using KPIs (Goundar et al., 2015). The KPIs monitored here are the vibration
and temperature of induction motors. Any failure resulting from lubrication, motor ventilation,
motor load, electrical consideration, and alignments directly impacts motor temperature and
vibrations. The data set collected while monitoring these motors predicts the motor’s bearing
breakdown by comparing the healthy and faulty motor working operations' variances. To measure
vibration and temperature KPIs, necessary sensors such as accelerometer and thermistor are
incorporated. Furthermore, to obtain the sensor data, communicate with various other devices,
and store the sensor data, the author uses Waspmote IDE pro v1.2 board. The data collected from
the sensors is further scrutinized and converted from the time domain to the frequency domain so
that FFT techniques can be applied to solve the motor bearing problem. This technique
transforms the data into frequency plots that can distinguish the frequency spectrum from a faulty
to a new bearing.
Susto et al. explain how flexible, adaptive predictive maintenance based planning decision
support system employing regularized regression and machine learning approaches are used to
reduce maintenance linked cost and downtime KPIs in a semiconductor manufacturing plant for
Ion Beam etching process (Susto et al., 2014). To reduce the maintenance cost and downtime,
Susto et al. made use of the newly processed data accessible from the semiconductor process
equipment is to enhance remaining valuable life estimates, thus decreasing unpredicted breaks
and entire equipment lifetime. The above predictive maintenance-based system was validated on
a real industrial dataset associated with semiconductor manufacturing plant experimentally out-
performing preventative maintenance and run-to-failure maintenance.
Munir et al. emphasized the importance of improving the existing KPIs assessment methodology.
The main objective was to develop a better assessment methodology for KPIs to improve the
manufacturing plant process. By combining qualitative methods such as questionnaires and
assessment matrices adapted from EFQM, this methodology was developed. By using these
qualitative methods, it was seen that KPIs to be considered as an essential element for planning
manufacturing strategies were quality, cost, delivery, inventory and flexibility; and the top five
KPIs to be used where the return on investment, conformance to specifications, profitability,
overhead cost, and customer satisfaction. Based on the importance of specific processes, it was
possible to prioritize the KPIs linked to those processes using the developed methodology.
However, the applicability of using the resultant five KPIs was limited because the nature of
operations within different manufacturing industries tends to alter, and the tactical and
operational goals. Merely suggesting the top five KPIs will not improve the process utilization for
every manufacturing industry. There is still a need for a generalized KPI assessment methodology
that can apply to all manufacturing industries.
Stricker et al. understanding the problems arising due to increasingly distributed production,
growing variant diversity, and disruptions due to shorter product life cycles on the production
performance and quality of the manufacturing industries, recommended an approach for selecting
KPIs. Since KPIs are commonly used tools for detecting any changes in production system
performance to organize suitable countermeasures. It underlined that the need for choosing an
appropriate number of KPIs is significant. The manufacturers should not be overburdened or
given insufficient KPIs to monitor the production performance overstraining decision-makers'
cognitive capabilities. By knowing the critical areas of performance improvement, a set of KPIs
were selected based on employing a mathematical linear programming approach (based on
integer linear programming). This approach focused on bridging the gap between information
content and simplicity. The appropriate KPIs needed for this selection approach are assumed to
be linked to the manufacturers' business objectives. Nevertheless, from a manufacturer’s
perspective, they require an approach that can make them understand, analyze, and link the KPIs
to the business objectives before moving towards the selection procedure.
Bongsug et al. recognized that developing KPIs for monitoring critical business performance is
challenging, and the set of practical guidelines available in the literature are limited. A list of
essential KPIs and a practical approach to performance measurement are presented in this paper
to address this challenge. The list of essential KPIs was developed based on incorporating a rich
industrial experience and Supply Chain Operations Reference (SCOR) model and hierarchically
clustered (i.e., primary and secondary). Furthermore, these KPIs were grouped into five
categories: planning, sourcing, production, and delivering, as shown in figure 13, to cover the
whole supply chain management structure. As a limitation, the KPIs presented were lagging in
nature and specific to managers.