Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com

ScienceDirect
ScienceDirect
Procedia
Available Computer
online Science 00 (2019) 000–000
at www.sciencedirect.com
Procedia Computer Science 00 (2019) 000–000 www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia
ScienceDirect
Procedia Computer Science 180 (2021) 1024–1033

International Conference on Industry 4.0 and Smart Manufacturing

International Conference on Industry 4.0 and Smart Manufacturing
Statistical Process Control of assembly lines in a manufacturing
Statistical Process Control of assembly lines in a manufacturing
plant: Process Capability assessment
plant: Process Capability assessment
Eleonora Bottaniaa, Roberto Montanariaa, Andrea Volpiaa, Letizia Tebaldiaa, Giulio Di Mariaaa
Eleonora Bottani , Roberto Montanari , Andrea Volpi , Letizia Tebaldi , Giulio Di Maria
a
Department of Engineering and Architecture – University of Parma, Parco Area delle Scienze 181/A, 43124, Parma (Italy)
a
Department of Engineering and Architecture – University of Parma, Parco Area delle Scienze 181/A, 43124, Parma (Italy)

Abstract
Abstract
Among the main strategies adopted by companies for enhancing their competitive advantage as well as for improving the internal
Among
efficiencytheismain strategies
the quality adopted by Several
management. companies forcan
tools enhancing their when
be involved competitive
dealingadvantage
with this as wellone
issue; as for
of improving
these is thethe internal
Statistical
efficiency is the quality
Process Control, management.
which includes Several tools
the employment of can be involved
statistical methods when
anddealing
metricswith this issue;
to monitor and one of these
control is the quality.
a process’ Statistical
In
Process Control,
this paper, indeed,which includes metrics
two statistical the employment of statistical
are involved methods
for assessing and metrics
the process to monitor
capability and
of a filler controlproduced
machine a process’byquality. In
an Italian
this paper,operating
company indeed, two statistical
in the metricsSpecifically,
food context. are involvedtwo forprocesses
assessingare theinspected:
process capability
the slewingof aring-pinion
filler machine produced
backlash and by
thean Italian
handling
company operating
clamps height in the
check, foodshowing
both context. excellent
Specifically, two processes
performances are having
after inspected: the slewing
carried out the ring-pinion
control andbacklash and the
provided handling
appropriate
clamps height
adjustments. check,areboth
Results also showing
comparedexcellent
with thoseperformances
obtained fromafter having
the Six Sigmacarried
theory,outanother
the control and provided
tool involved appropriate
for quality controls
adjustments.
which is in lineResults are also compared
with principles with those obtained
of lean manufacturing. fromfor
Moreover, thethe
Sixsecond
Sigmaprocess,
theory, another
a software toolwas
involved for quality
implemented controls
for speeding
which is in lineand
up operations withachieving
principlesbenefits
of lean manufacturing.
in terms of time. Moreover, for the second
The reliability of theseprocess,
analysisa software was implemented
is confirmed for speeding
by the application of the
up operations
ANOVA GageandR&R achieving
tool, whichbenefits in terms
allowed of time.
to assess The reliability
the precision of these analysis
of the measurement system is involved.
confirmed by the application of the
ANOVA Gage R&R tool, which allowed to assess the precision of the measurement system involved.
© 2021 The Authors. Published by ELSEVIER B.V.
©© 2021
2021 The
The Authors.
Authors. Published
Published by by ELSEVIER
Elsevier B.V.B.V.
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
This is an
Peer-review open access
under article under
responsibility of the scientific
CC BY-NC-NDcommitteelicense (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee ofof
thethe International
International Conference
Conference on on Industry
Industry 4.0 4.0
andand Smart
Smart Manufacturing
Peer-review
Manufacturing under responsibility of the scientific committee of the International Conference on Industry 4.0 and Smart
Manufacturing
Keywords: Automation; Logistics; Industry 4.0; Food Industry.
Keywords: Automation; Logistics; Industry 4.0; Food Industry.

1. Introduction
1. Introduction
In an increasingly competitive market, one of the main strategies adopted by companies for gaining advantages is
In an increasingly
to monitor and achieve competitive market, and
quality in product one processes
of the main [1].strategies adopted by companies for gaining advantages is
to monitor and achieve quality in product and processes [1].
The basic definition of quality refers to one or more desirable characteristics that a product, a service or a process
The possess
should basic definition of quality
[2] in order refers
to satisfy to one
implicit andorexplicit
more desirable
customer’s characteristics
requirements. that a product,
Indeed, a service
the fulcrum or acustomer
is the process
should possess [2] in order to satisfy implicit and explicit customer’s requirements. Indeed, the fulcrum is
satisfaction: when a consumer receives quality products, in return his loyalty increases, the company’s position on the the customer
satisfaction: when a consumer
market is maintained receives quality
or even improved, products,
liability risks areinreduced,
return his
theloyalty
brand increases,
gets good the company’s
reputation and,position on the
consequently,
market is maintained or even improved, liability risks are reduced, the brand gets good reputation and, consequently,
benefits arise. It follows that, within the industrial context, an appropriate quality control is essential at the process
benefits arise. Itcompliances
stage to ensure follows that,ofwithin theitem.
the final industrial context, an appropriate quality control is essential at the process
stage to ensure compliances of the final item.
1877-0509 © 2021 The Authors. Published by ELSEVIER B.V.
1877-0509 © 2021
This is an open Thearticle
access Authors. Published
under by ELSEVIER
the CC BY-NC-ND B.V.(https://creativecommons.org/licenses/by-nc-nd/4.0)
license
This is an open
Peer-review access
under article under
responsibility CC BY-NC-ND
of the scientific license
committee (https://creativecommons.org/licenses/by-nc-nd/4.0)
of the International Conference on Industry 4.0 and Smart Manufacturing
Peer-review under responsibility of the scientific committee of the International Conference on Industry 4.0 and Smart Manufacturing

1877-0509 © 2021 The Authors. Published by Elsevier B.V.

This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0)
Peer-review under responsibility of the scientific committee of the International Conference on Industry 4.0 and Smart Manufacturing
10.1016/j.procs.2021.01.353
 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033 1025
2 Author name / Procedia Computer Science 00 (2019) 000–000

Among the main tools having this purpose, it is worth mentioning the Statistical Process Control (SPC), which
monitors and controls quality by tracking production metrics, and the famous Six Sigma method, which uses five key
steps (Define, Measure, Analyze, Improve and Control – DMAIC or Define, Measure, Analyze, Design and Verify -
DMADV) to ensure that products meet customers’ requirements and have zero defects [3]; specifically, the term Six
Sigma refers to a statistically derived performance target of operating with only 3.4 defects per million opportunities
(DPMO) [4], which represents a 99.99966% process yield meaning that 99.99966% of output products do not have
any defects, and consequently requires significant efforts [5]. For instance, it is worth mentioning studies by [6] as far
as a general quality assurance through process improvement, or by [7] who carried out an exhaustive bibliometric
analysis concerning the topic of Six Sigma.
The concept of quality is a keyword and a pillar of a particular philosophy of production developed at the beginning
of the Third Industrial Revolution [8], namely the well-known lean manufacturing (LM). It can be seen as a business
strategy implementable through different principles, tools and techniques, which is focused on the elimination of any
type of waste (called in this perspective Muda), aiming at supplying the exact quantity of products required by
customers, in the right time, with an extremely high level of quality and low cost, thanks to its being smart [9]. From
the joint union of the LM thinking and one of the abovementioned tools for keeping quality under control, the Lean
Six Sigma (LSS) was born, namely a hybrid initiative that combines the Six Sigma’s structured problem solving using
statistical tools with lean operation’s emphasis on flow improvement [10]. Literature in this field is extremely copious
and various; for a complete overview and comprehension of the state-of-art, [11] and [12] are recalled. Specifically,
in this last paper, its adoption in European organizations is treated.
On the bases of these brief theoretical premises, the aim of this paper is to present two case studies carried out on a
company based in the north of Italy, designing and producing food machinery; specifically, an SPC was implemented
for a filler machine.
The novelty of the presented paper is threefold: on one side it presents a real and practical implementation of the
theoretical process control methodology; second, it provides readers and quality inspectors with real and tangible
results and then, on the basis of these outcomes, it precisely describes the design and development of a software tool
for improving process capability indices.
The remainder of the paper is as follows: section 2 illustrates material and methods involved, followed by section
3, which deals with the first case study including the measuring system evaluation and the capability process
calculation. The second case study, instead, is proposed in section 4, whose structure recalls that of case study 1.
Section 5, finally, presents conclusions and future research directions in the light of the obtained results.

2. Material and Methods

The metrics investigated are the process capability indices Cp and Cpk, which measure the ability of a process to
meet engineering limits [13]. The first index evaluates the performance of the process related to the production
specifications, and it is obtained by applying the following formula (1):

𝑈𝑈𝑈𝑈𝑈𝑈−𝐿𝐿𝐿𝐿𝐿𝐿
𝐶𝐶𝑝𝑝 = (1)
6𝜎𝜎

Where USL is the Upper Specification Limit of the quality characteristic, LSL the Lower, and 𝜎𝜎 is the process
standard deviation. In case its value is greater than 1.00, the process is capable. Conversely, Cpk takes into account the
process location, namely whether a process deviates from half of its range of specifications. It is computed as follows
(formula 2):

𝑈𝑈𝑈𝑈𝑈𝑈− 𝜇𝜇 𝜇𝜇−𝐿𝐿𝐿𝐿𝐿𝐿
𝐶𝐶𝑝𝑝𝑝𝑝 = ( ; ) (2)
3𝜎𝜎 3𝜎𝜎

Where USL and LSL were already defined, and 𝜇𝜇 is the mean of the process output.
These indices are both common when dealing with this kind of measures; see for instance [14] or [15], recent studies
found in literature.
1026 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033
Author name / Procedia Computer Science 00 (2019) 000–000 3

The values obtained are then compared with those deriving from the Six Sigma theory; more in details table 1
reports the correlations between achieved sigma level, goods conformity percentage, Cpk index, PPM (part per million)
of defective goods, time wasted for bad production in a month [16].

Table 1. Correlations between sigma level, Cpk index, PPM, time wasted
σ Conformity % Cpk PPM Time wasted / 720 h

±1 68.26 0.33 317400 228.5 h

±2 95.46 0.67 45500 32.8 h
±3 99.73 1 2700 1.94 h
±4 99.994 1.33 63 2.74 min
±5 99.99994 1.67 0.57 1.49 min
±6 99.9999998 2 0.002 0.005 s

Before collecting data and start with measurement, a crucial issue is clearly to possess the appropriate instruments,
as well as correct methods and trained operators. For validating these instruments, several tools can be involved; one
of the most spread, adopted in this case, is the ANOVA gage repeatability and reproducibility (AGRR, also known as
ANOVA Gage R&R), which allows to assess the precision of a measurement system. The peculiarity of this method
is that of considering the variation due to the instrument as composed by two different variances, respectively inherent
to the repeatability and the reproducibility. The step for applying the AGRR are the following: (1) determine an
experimental design (e.g. the number of operators, number of parts, number of replicates) according to rule of thumb,
budget and availability; (2) measure the parts for each treatment; (3) conduct the ANOVA using the observations; (4)
estimate the variance components for each factor and interaction; (5) calculate various performance metrics using the
estimates (e.g. Cp or Cpk); (6) judge the adequacy precision for the measurement system according to criteria; (7)
perform subsequent actions such as improvements of the system according to the results [17]. According to the
Automotive Industry Action Group (AIAG) [18], in case Gage R&R < 10% the measurement system is acceptable; if
10% < Gage R&R < 30% the system is conditionally acceptable; not acceptable for remaining values. In this study,
Minitab software was implemented for carrying out the ANOVA analysis. Several studies in literature implement this
method for validating instruments; see for instance recent works by [19] or [20].
The first case study is related to a quality check which is carried out during the assembly operations of a mechanical
machine for the food industry. More in details, this check consists of the measurement of the slewing ring-pinion
backlash; it verifies the correct backlash between the teeth of the pinion and the teeth of the slewing ring. The pinion
is a toothed wheel connected to the main motor which moves the carousel through the engagement with the teeth of
the slewing ring as illustrated in figure 1. A certain backlash between the teeth, according to the design specifications,
is important to avoid friction and overheating problems or unexpected forces.
The second analyzed process, instead, refers to another kind of check which is performed during the assembly
operations of a mechanical system for handling plastic bottles. Specifically, the clamps check is performed on all the
handling clamps connected to the rotation support (namely star), for every clamp the precise height “C” with respect
to a fixed reference point is measured as shown in figure 2. The star is made up of several clamps used for sorting
bottles and transferring them from one point to another on the machine. In one machine there are several stars rotating
and interacting with each other; given the interaction that there must be between the stars and therefore between their
clamps, it will be necessary that they are positioned correctly so that the bottle transfer can be carried out correctly.

Fig. 1: Rotating structure. Fig. 2. Clamp height measurement.

 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033 1027
4 Author name / Procedia Computer Science 00 (2019) 000–000

3. Case Study 1: slewing ring-pinion backlash check

The correct value of backlash, the minimum and the maximum value are reported in a control sheet, which reports
also a very brief description of the operations required to perform the check: the slewing ring tooth must be positioned
in perfect tangency with the pinion’s tooth. This can be achieved in two different steps. The first is to visually center
the pinion tooth with the slewing ring so that the symmetry line of the pinion tooth overlaps with the radius of the
slewing ring (they are perfectly aligned centered). Then, the slewing ring must be manually adjusted so that one side
of the pinion tooth is in contact with the tooth of the slewing ring, as shown in figure 3.

Fig. 3. Contact between the pinion tooth and the tooth of the slewing ring.

At this point, the backlash between the teeth is manually measured by means of a thickness gage having a resolution
of 0.05 mm. After the first measurement, according to the abovementioned procedure, the pinion must be rotated 360
degrees so that the backlash between the same tooth of the pinion is checked again and measured with respect to
different tooth of the slewing ring. The procedure is repeated five times.

3.1. Measuring system validation

Minitab software, used to perform ANOVA analysis of the measurement values and compute the Gage R&R,
returned the outcomes reported in table 2.

Table 2. Results from the AGRR analysis.

%Contribution Study Var %Study Var
Source VarComp (of VarComp) StdDev (SD) (6 × SD) (%SV)

Total Gage R&R 0.0011806 51.52 0.0343592 0.206155 71.77

Repeatability 0.0005556 24.24 0.0235702 0.141421 49.24
Reproducibility 0.0006250 27.27 0.0250000 0.150000 52.22
Operators 0.0000648 2.83 0.0080508 0.048305 16.82
Part-To-Part 0.0011111 48.48 0.0333333 0.200000 69.63
Total Variation 0.0022917 100.00 0.0478714 0.287228 100.00

It should be noticed that the Total Gage R&R value (%Study Var) is 71.77%, so the variance of the results obtained
according to the described measurement procedure is 71.77% caused by the low repeatability and reproducibility of
the measurement process. This is a very poor result, as acceptable values of %Study Var must be less than 30%
according to the AIAG standard in order to have a reliable measuring system.
As a consequence, the calculation of the process capability indexes cannot be performed until the measuring system
is not improved. Thanks to the observation of two operators performing the measurement, some differences in actions’
execution were noticed; thus, as a first step, the measuring method has been standardized as much as possible,
performing more training on the operators, explaining them how the measurement should be performed without relying
on the spontaneity of the operators. A SOP (Standard Operation Procedure) consisting of clear and simple images was
designed, as reported in figure 4.
1028 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033
Author name / Procedia Computer Science 00 (2019) 000–000 5

Fig. 4. SOP steps.

After the introduction of the SOP and the training of the operators, a second measurement session was carried out
and 30 values were obtained; specifically, these 30 values come from the operations repeated three times by two
operators for five different components, according to the same procedure previously described. After data analysis,
the following results were obtained (table 3):

Table 3. Results from the AGRR analysis after the SOP and training of operators.

%Contribution Study Var %Study Var

Source VarComp (of VarComp) StdDev (SD) (6 × SD) (%SV)

Total Gage R&R 0.0003065 21.69 0.0175075 0.105045 46.58

Repeatability 0.0003065 21.69 0.0175075 0.105045 46.58
Reproducibility 0.0000000 0.00 0.0000000 0.000000 0.00
Operators 0.0000000 0.00 0.0000000 0.000000 0.00
Part-To-Part 0.0011063 78.31 0.0332614 0.199569 88.49
Total Variation 0.0014128 100.00 0.0375877 0.225526 100.00

The introduction of SOP methodology in the measurement system has reduced the %Study Var from 71.77% to
46.58%, which is still too high for a reliable measurement system. In order to further lower the value, the main actions
that could somehow bring variability have been specifically addressed.
As mentioned above, in order to check the backlash between the pinion tooth and the slewing ring, the pinion tooth
must be first perfectly aligned with a radius of the slewing ring. This operation is performed by the operators without
the aid of any tool or equipment, relying only on his visual skills and experience. This visual inspection which guide
the alignment can lead to a great variability in the measurement.
An alignment jig could improve the process by satisfying the need for positioning and by increasing the accuracy
of the centering of the two teeth; in fact, it could indicate the correct position of the tooth in relation to the slewing
ring. Since this specific tool was not available on the market, the tool was designed using 3D modeling software and
then 3D printed thanks to additive-manufacturing techniques; the model is shown in figure 5.

Fig. 5. 3D Model of the alignment tool.

 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033 1029
6 Author name / Procedia Computer Science 00 (2019) 000–000

After the introduction of the jig, the SOP was updated with its assembly and positioning instructions; once done,
the tooth positioning becomes easier as the jig provide fixed reference points for tooth alignment.
A third measurement campaign was performed with 30 measures and excellent results were obtained as reported
by Minitab in table 4.

Table 4. Results from the AGRR analysis after jig introduction.

%Contribution Study Var %Study Var

Source VarComp (of VarComp) StdDev (SD) (6 × SD) (%SV)

Total Gage R&R 0.0003065 2.42 0.017508 0.105045 15.56

Repeatability 0.0003065 2.42 0.017508 0.105045 15.56
Reproducibility 0.0000000 0.00 0.000000 0.000000 0.00
Operators 0.0000000 0.00 0.000000 0.000000 0.00
Part-To-Part 0.0123563 97.58 0.111159 0.666954 98.78
Total Variation 0.0126628 100.00 0.112529 0.675176 100.00

The %Study Var has dropped to 15.56%, significantly below the threshold of 30% typical of a reliable measuring
process. The introduction of the mask has significantly reduced the probability of error caused by the operators. The
measuring system can be considered reliable and therefore repeatable and reproducible, allowing the computation of
the process capability analysis.

3.2. Capability Process Calculation

Data collected during the validation process of the measurement system has been plotted in figure 6 and used to
compute the capacity indexes of the whole process by means of an Excel spreadsheet.

Figure 6. Capability process calculation (case study 1).

The computation of the indexes gives a result of Cp=1.51 and Cpk=1.45. These values fit perfectly the target imposed
by the company, more precisely they exceed the expectations giving a Sigma level of 4.3. This is consistent with
plotted data of figure 6, which clearly shows that the process is centered with respect to the project specifications (LSL
- USL, the red lines of the graph). Thus, the control process can be declared under statistical control and no further
improvements nor optimizations have been identified since the new SOP has not caused any slowdown in the process
of assembly check (backlash measurement).

4. Case Study 2: handling clamps check

As already stated in section 2, as far as case study 2 it is firstly necessary to determinate the different heights.
All the measures are then grouped together, the difference between the highest and lowest clamp is computed and
checked against the limit value of 0.2 mm defined in the control sheet. This value allows to have a precise and reliable
interaction between the clamp that yields and the one that receives during the bottle transfer.
1030 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033
Author name / Procedia Computer Science 00 (2019) 000–000 7

4.1. Data collection

In order to calculate the difference between the highest and lowest clamp, the height of each individual clamp is
measured with respect to a reference point, this operation is performed by means of a laser scanner interfaced with a
computer using laser triangulation on the clamp surface. Since the reference point is fixed, the values found can be
compared and thus the difference easily computed between the clamps.
During measurement with the laser, the star is rotated 360 degrees for three times, so that the measurement of each
clamp is averaged over three different values, the standard deviation for each clamp is calculated and the highest and
lowest clamp with the respective value is displayed. Figure 7 shows the measurements on a graph.

0,2
0,18
Height Difference (mm)

0,16
0,14
0,12
0,1
0,08
0,06
0,04
0,02
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Clamp

Fig. 7. Clamp heights.

In case the difference between the highest and lowest clamps is more than 0.2 mm, the clamps must be manually
adjusted by means of the report composed of diagram and raw data. The operator identifies, completely relying on his
own experience and skills, the clamps needing intervention for height adjustment. After the clamps have been
modified, the previous step of measure and height calculation shall be repeated in order to check whether the values
after the intervention are within the project specifications reported in the control sheet.

4.2. Measuring system validation

Following the same procedure as the previous check, the height difference between clamps of five different rotating
stars was measured by repeating the measurements three times with two different operators. The analysis, always
performed with Minitab software, shows a %StudyVar of 27%, which is not very far from the upper limit of 30% that
allows a positive evaluation of the measurement process. Since the value of the %StudyVar is lower than 30%, the
measurement system has been considered as reliable and repeatable at the first stage. After the validation of the
measuring system, the next step is the calculation of the capacity indexes in order to evaluate assembly process
performance.

4.3. Capability Process calculation

After gathering all required data, performance indexes have been computed, finding a Cp=1.45 and Cpk=0.30 (figure
8).

Fig. 8. Capability process calculation (case study 2).

 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033 1031
8 Author name / Procedia Computer Science 00 (2019) 000–000

The observation of the indexes shows that the process is not centered; indeed, although the value of the Cp is
satisfying and higher than the target of 1.33, the value of the Cpk is very low, leading the system to a sigma level of
0.9, very far from the target Sigma Level of 4. In order to improve the Cpk several simulations have been carried out
using an Excel spreadsheet; the results show that having a maximum difference in clamps height in a range between
0.09 mm and 0.11 mm brings to a very high Cpk, centers the process and achieves a Sigma Level of 6 with Cp=4.23
and Cpk=4, as reported in figure 9.

Fig. 9. Capability process calculation after many simulations in order to centre the process (case study 2).

In order to achieve this result, the difference in height between the highest and lowest clamps is reduced to 0.10
mm with a tolerance of ± 0.01 mm. To do this, a deeper study of the clamp’s adjustment process was required in order
to reengineer it. It was then decided to use washers with a thickness of 0.1 mm, instead of 0.2 mm as previously used,
to get a finer height adjustment of the clamp. This improvement brought very good results; in fact, the difference
between highest and lowest clamp was reduced in a range from 0.09 mm to 0.12 mm as shown in figure 10.
Conversely, the time needed to perform the height adjustment raised from 45 minutes (current scenario) to 1 hour
and 15 minutes for each star (reengineered process). Thus, although the capability indexes have increased, also
machine cycle time has increased, and optimization of the process is clearly needed.

0,12

0,1
Height Difference (mm)

0,08

0,06

0,04

0,02

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Clamp

Fig. 10. Clamp heights (with reduced difference).

4.4. Process Optimization

The process optimization was carried out in order to find a quicker solution to speed up the clamp control. The time
increase gives excellent capability index values but is not acceptable.
Since the height adjustment is a completely manual operation, a tool capable of identifying the clamps needing
intervention and indicating the number of washers to add could improve the efficiency of the whole process.
A visual software was implemented on Excel using Visual Basic for Application language (whose interface is
shown in figure 11). A specific and tailored user interface was used to make the program more intuitive, to guide the
operator in entering data and to facilitate the interpretation of the results.
The software created has the following specifications:
1032 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033
Author name / Procedia Computer Science 00 (2019) 000–000 9

1. Acquisition of the height of each clamp from the laser scanner;

2. Calculation of the difference between highest and lowest clamp;
3. In case the difference is not within the specified range, it identifies the best solution (in terms of fewer clamps
to be modified) in order to get the desired difference between the clamps’ height;
4. Show the number of washers to add in each clamp;
5. Acquisition of the height of each clamp from the laser scanner after adjustment and final check;
6. In case the obtained result is not acceptable, the process is repeated from step 2.

Fig. 11. Developed software interface.

The use of the software helps the operator’s activity, as he only has to follow the procedure:
1. enter the required specifications;
2. look for the clamps to be modified;
3. adjust them accordingly.
This allows a great saving of time in the execution of the control; in fact, the whole procedure in performed in about
20 minutes, leading to a time saving of 73%. The software has been tested on various stars of different machines
achieving excellent results in every scenario. Thanks to the implementation of the software, it has been possible to
save a significant time both for the capability project and for the optimization of the machine cycle, because the process
has been speeded up and the human factor has been eliminated.

5. Conclusions and Future Developments

The process capability project introduced in the two case studies reported excellent results.
As far as it concerns the backlash control, after the first measurement according to the current procedure, the process
showed a very high value of Gage R&R of 71.77%, significantly above the threshold of 30% typical of a reliable
measuring process. Thus, several corrective actions have been made in order to improve it, such as the standardization
of the measurement process and the introduction of an operating instruction (SOP). Such interventions didn’t improve
enough the value of the Gage R&R and consequently a tool for pinion tooth centering was created and introduced in
the SOP. Eventually, the Gage R&R value has been reduced to 15.56%, which means that the measuring system is
reproducible and repeatable. Subsequently, the process capability indexes were calculated and the process was
centered with a sigma level of 4.3. This is a very good result, satisfying the current target set by the company, although
the final goal will be to try to improve the process to reach a sigma level equal to 6.
For the star clamp comparison check, on the other hand, the current procedure has been promptly validated as the
Gage R&R value was acceptable (27%). Thus, Cp and Cpk indexes were calculated and evaluated as not acceptable to
achieve a target sigma level of 4. After various simulations and tests, the feasible solution was to adjust the clamps’
height using washers with a thickness of 0.10 mm, keeping the maximum difference between them within 0.12 mm.
In this case, a sigma level equal to 6 has been achieved, although the reengineered process brought to a substantial
 Eleonora Bottani et al. / Procedia Computer Science 180 (2021) 1024–1033 1033
10 Author name / Procedia Computer Science 00 (2019) 000–000

increase in time in the execution of the checks. In order to speed up this phase, a dedicated software tool has been
implemented to reduce wasted time according to Lean methodology.
Below, table 5 summarizes the benefits in terms of time:

Table 5. Benefits resulting from the implementation of the software.

Process Check time [min] Time difference [%]

Current 45 -
Enhanced (manually managed) 75 +66.7 (than the original state)
-55.5 (than the original state)
Enhanced (software driven) 20
-73.3 (then the previous state)

Future developments will encompass different CTQ (Critical to Quality) parameters impacting the overall quality;
such parameters will be selected not only in pre-assembly department controls but also in other departments of the
company. The automation of data acquisition is another future step; this will lead to automatic population of Excel
spreadsheet for real-time computation of process capability indexes.

References

[1] Andres-Jimenez, Andres, Medina-Merodio, Jose-Amelio, Fernandez-Sanz Luis, Martinez-Herraiz, Jose-Javier and Ruiz-Pardo, Estefania.
(2020). “An intelligent framework for the evaluation of compliance with the requirements of ISO 9001:2015”. Sustainability 12: 5471.
[2] Montgomery, Douglas C. (2009) “Introduction to Statistical Quality Control”. Jhon Wiley & Sons, Inc. Sixth edition.
[3] Graphic Products Staff. “Quality Control in Manufacturing”. Available at https://www.graphicproducts.com/articles/quality-control-in-
manufacturing/ [Accessed on July, 2020].
[4] Pande, P.S., Neuman, R.P., and Cavanagh, R.R. (2000). “The Six Sigma way”. McGraw-Hill.
[5] Linderman, Kevin, Schroeder, Roger G., Zaheer, Srilata, and Choo, Adrian S. (2003). “Six Sigma: a goal-theoretic perspective”. Journal of
Operations Management, 21: 193-203.
[6] Sharma, G.V.S.S., Rao, P.S., and Babu, B.S. (2016). “Quality assurance through process improvement – A concise review”. Engineering
Journal, 20 (5): 103-114.
[7] Ninerola, A., Sánchez-Rebull, M.-V., and -Lara, A.-B. (2019). “Six Sigma literature: a bibliometric analysis”. Total Quality Management and
Business Excellence.
[8] Sposito Valamede, Luana, and Santos Akkari, Alessandra Cristina. (2020). “Lean 4.0: a new holistic approach for the integration of lean
manufacturing tools and digital technologies”. International Journal of Mathematical, Engineering and management Sciences, 5 (5): 851-868.
[9] Tayaksi, Cansu, Sagnak, Muhittin, and Kazancoglu, Yihit. (2020). “A new holistic conceptual framework for leanness assessment”.
International Journal of Mathematical, Engineering and management Sciences, 5 (4): 567-590.
[10] Bonome Message Costa, Luana, Godinho Filho, Moacir, Fredendall, Lawrence D., Miller Devos Ganga, Gilberto. (2021). “Lean six sigma in
the food industry: Construct development and measurement validation”. International Journal of Production Economics, 231: 107843.
[11] Sordan, J.E., Oprime, P.C., Pimenta, M.L., Chiabert, P., and Lombardi, F. (2020). “Lean Six Sigma in manufacturing process: a bibliometric
study and research agenda”. TQM Journal, 32 (3): 381-399.
[12] Panayiotou, N.A., and Stergiou, K.E. (2020). “A systematic literature review of lean six sigma adoption in European organizations”.
International Journal of Lean Six Sigma.
[13] Arzak, M.E., Wazeer, A., Essam, K., Saied Ayman, A., and Abd-Eltwab, A. (2020). “Process capability analysis in filling operation – A case
study”. International Journal of Scientific and Technology Research, 9 (3): 6650-6655.
[14] Majumder, H., and Maity, K. (2020). “Performance analysis in WEDM of titanium grade 6 through process capability index”. World Journal
of Engineering, 17 (1): 144-151.
[15] Siraj, I., and Bharti, P.S. (2020). “Process capability analysis of a 3D printing process”. Journal of Interdisciplinary Mathematics, 23 (1): 175-
189.
[16] Oakland, J. (2003). "Statistical Process Control”. Oxford: Butterworth-Heinemann, Fifth edition.
[17] Park, Sejoon, and Ha, Chunghun. (2020). “Determination of optimal experimental design for ANOVA gage R&R using stochastic
programming”. Measurement, 156: 107612.
[18] Automotive Industry Action Group – AIAG. (2017). Measurement System Analysis. Fourth Edition: Automotive Industry Action Group.
[19] Sharma, Mithun, Sanhi, Sanjeev P., and Sharma, Shilpi. (2016). “Validating a destructive measurement system using Gage R&R – A case
study”. Engineering Management in Production and Services, 11 (4).
[20] Doshi, Jighar A., and Desai, Darshak, A. (2017). “Measurement system analysis for continuous quality improvement in automobile SMEs:
multiple case study”. Total Quality Management & Business Excellence, 30 (5-6): 626-640.