Impact of Smart Incoming Inspection

System on the Production, in a Medical

Device Manufacturing MSME

Puneeth S. Kannaraya, G. H. Shreya, Manish Arora,

and Amaresh Chakrabarti

1 Introduction

Quality-related issues and challenges are critical to be addressed during manufac-

turing [1]. Quality control in manufacturing plays a significant role in finding defec-
tive parts and processes and assuring quality outcome throughout the intended life
cycle of the product. It is a vital part in the Quality Management System (QMS)
which dictates the requirements of quality assurance as a regulation and standard
[2]. Hence, to integrate the quality system, organizations need to follow the stan-
dards recommended by regulatory agencies and government bodies. The standards
are so important that if not followed, it might affect the life of the consumer (e.g.
medical devices, automotive parts, aerospace parts).
Quality assurance is defined in ISO 9000–2000 as “part of quality management
focused on providing confidence that quality requirements are fulfilled” [3]. The
term quality control refers to the activities or tools that are used to provide this
assurance. Quality-related in-process inspection/verification is an essential part of
quality control in manufacturing. Inspection in manufacturing includes measuring,
examining, testing or gauging one or more characteristics of a product or process
and comparing the results with specified requirements to determine whether the
requirements are met for each characteristic [4]. To perform inspection activities,
accurate measuring instruments are needed. Normally, this instrument is purchased;
however, it may be necessary to design and build it in cooperation with process
design [5]. Using human for 100% inspection requires considerable manpower that
significantly increases costs and the time for inspection [6]. Even though the recent
advancements in manufacturing systems have been characterized by precision of

P. S. Kannaraya (B) · G. H. Shreya · M. Arora · A. Chakrabarti

Centre for Product Design and Manufacturing, Indian Institute of Science, Bengaluru 560012,
India
e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 157
A. Chakrabarti et al. (eds.), Industry 4.0 and Advanced Manufacturing, Lecture Notes
in Mechanical Engineering, https://doi.org/10.1007/978-981-19-0561-2_15
158 P. S. Kannaraya et al.

work through automation [7]. It is very difficult to automate any manufacturing

system due to budget constraints, space constrains or lack of skilled labour [2].
The issues and challenges are mostly encountered in the MSMEs. The Indian
MSMEs are the main contributors of the Indian economy, especially in medical
devices, 65% of the indigenous manufacturers are categorized under MSME in India.
There are 750–800 domestic medical device manufacturers in India, with an average
investment of $2.3–2.7 million and an average turnover of $6.2–6.9 million and
contribute 30% or (USD 1.1 billion) to the Indian medical devices market [8]. Even
with this huge contribution the MSMEs struggle to compete with the global market.
There are various reasons for the enterprises like lack of manpower, lack of resources,
time, regulatory challenges, cost etc.
This research started in a focus to identify the challenges faced during the imple-
mentation of QMS for the certification of medical devices in Indian MSMEs. During
understanding of the quality control activities of a local orthotic footwear manu-
facturer, it was found the inspection of the raw materials was not being performed.
This was mainly due to the unaffordability of the state-of-the-art inspection setup.
Therefore, a bespoke smart incoming inspection system was developed which uses
off-the-shelf cameras and image processing techniques to capture image, analyse
the defects and show the defects on the UI. The testing of the developed system for
various light condition and for different colours of the raw materials was established
earlier [9]. Brewer et al. [10] rightly said that reduction of waste, production cycle
time, other non-value-added time or inventory is the purpose of a new technology,
not the introduction of the technology itself.
The performance of the system was analysed using the six sigma DMAIC method-
ology. Statistical analysis of the final inspection data before and after introducing
the smart incoming inspection system showed a significant change in the rejection
rates. The impact on production was indicated by improvement in the FPY.

2 Literature

Quality means freedom from deficiencies and features of products which meet
customer needs and thereby provide customer satisfaction [1]. Freedom from defi-
ciencies is freedom from errors that require doing work over again (rework).
Quality-related research has been an interesting topic among researchers for decades.
Various researchers have worked on the rejection rates using different quality
control methods. Most notable is by using six sigma methodology using DMAIC
(Define-Measure-Analyse-Improve-Control) approach.
Gijo and Scaria [11] discuss the implementation of six sigma methodology in
reducing rejection and rework in a honing process in an automobile part manufac-
turing company, where six sigma implementation in a manufacturing process for
improving quality resulted in reduction of rejection and rework, thus improving the
first pass yield from 88 to 100%. At one of the Indian small-scale unit, to improve
rejection/rework rate in manufacturing products by pressure die casting process,
Impact of Smart Incoming Inspection System … 159

quality was improved from 3.1σ to 3.7σ by reducing the rejection rate from 15.50%
to 4.47% which is 71.2% improvement [12]. By making modification in the design
of tool, viz., jig, the process sigma level is improved from 1.5 to 4.15, while the
reworking rate in the manufacturing of track shoe was decreased to almost zero;
this resulted into no delay in delivery and improved customer satisfaction [13]. To
reduce the rejection rates, some jigs and fixtures were designed which in turn reduced
the setup time and adjustments for a pump casing manufacturing using the DMAIC
methodology [14]. Also, the smart manufacturing (referred to as Industry 4.0 also)
has the major contribution on QMS, using smart factories, cyber-physical systems
(CPS), Internet of Things (IoT), etc. [15].

3 Methodology

To understand the quality-related issues associated with QMS, a local medical device
manufacturing micro-enterprise was selected. The company produces a low-risk
orthotic footwear used to cure medical conditions like calcaneal spurs, corns, heel
cracks, heel pain, etc. by providing customized footwear with medial arch for flat feet,
cushioned heel for heel pain, scooped insoles or offloading for ulcers, crow shoes and
ankle shoes for specific foot disorders, outsole modifications for differential height,
soft packed insole for corns and so on.
The footwear consists of different materials in both upper and sole parts. The
bottom sole part is divided into three parts, namely, bottom sole, midsole and insole.
The bottom sole is made of 10-mm-thick rubber material. The midsole is a 10 mm
Ethylene–Vinyl Acetate (EVA) foam material. The insole is also a 2 mm EVA foam
material. The insole is the part which directly encounters the human foot. Therefore,
the raw material of the insole is considered for the inline inspection just before it
gets assembled with the midsole.
To Define the problem—during the production process, the rejections are made if
the defects are encountered (see Fig. 1). These rejected parts are then scrapped and
considered as waste. Figure 2 shows the percentage of the insole rejections, where
close to 64% are because of the rejections of insoles with defect. This is the Measure

Fig. 1 Defects encountered

160 P. S. Kannaraya et al.

Fig. 2 % Rejections post insole cutting

Fig. 3 Cause-and-effect diagram for the rejections of the insole

of the problem. To Analyse the problem, a cause-and-effect diagram is constructed

(see Fig. 3) for the rejection of insoles with defects.

4 Experimentation

To avoid the wastage of resources like material, time and manpower, a smart incoming
inspection system is proposed to introduce in the incoming stage of the raw materials
to the production line (see Fig. 4). This incoming quality control stage will not only
serve as a quality assurance but also a necessary inclusion to comply with the quality
management system of the organization.
The design of the smart inspection system, identification of the defects and testing
for various light settings were performed successfully. It has been proved the defects
of at least 1 mm of size are identified within a range of 3100 lx to 3250 lx with both
surface and bottom light settings [9].
Impact of Smart Incoming Inspection System … 161

Fig. 4 Production line with the smart incoming inspection included in the insole cutting line

The raw materials were investigated using the system and labelling was done on
the defects as highlighting. This helped the worker at the cutting station to notice the
defective sheets. The barcode was used as the identification method which helped in
registering the number of defects identified in each sheet during the smart inspection.
The cutting operation was then performed by altering the stacking procedure of the
sheets, which prevented the defective areas of the sheet being included in the cut-out
insole pieces. The batch of cut sheets into insole was then moved to the concealing
area where it was then glued with the midsole using a special adhesive. Prior to this
stage is the inspection of cut insole and most of the rejection takes place here.
The program was designed to capture the defects in two forms. One is capturing
the image of the defective area and storing it in local as well as cloud database
and another to store the coordinate values of the defects. The latter will also have
the information of the sheet, viz, dimensions, colour and unique ID (Barcode). The
inspection was carried out for a batch of black and brown colour EVA sheets which
was issued by the inventory for production. These sheets are examined for any defects
but the data on rejection at the concealing process was collected for 4 weeks before
the experiment was conducted.
162 P. S. Kannaraya et al.

To establish statistical significance for the two test groups, the experiment was
based on rejecting the null hypothesis. (H0: σ1 2 = σ2 2 ). The hypothesis testing is by
accepting alternate hypothesis (Ha: σ1 2 = σ2 2 ).

5 Results

To test the hypothesis, the data of rejection rate at the pre-concealing stage over the
span of 5 weeks were collected N = 15 with M = 1.609 and SD = 1.282. Similarly,
the data was collected at this stage for another 2 weeks, but for those raw materials
which had undergone smart incoming inspection N = 7 with M = 0.421 and SD =
0.630. The introduction of incoming inspection enabled the production supervisor
to record the defective areas in the raw material sheets. The recorded data were then
statistically analysed to check for the significance and reject the null hypothesis (H0 :
σ1 2 = σ2 2 ). An F-test (see Table 1) was performed to test the homogeneity of the
variances of two populations with α = 0.05 as criterion for significance F (6,14) =
4.14, p = 0.045. This test is a one-tailed test as to reject null hypothesis and shows
the two-sample means are not equal, i.e. alternative hypothesis (Ha: σ1 2 = σ2 2 ). The
independent sample t-test (see Table 2) was conducted to show the significance since
the sample sizes were not the same, t (20) = 2.911, p = 0.0086. As the significance
value α is greater than the value obtained by t-test, the null hypothesis can be rejected.
There is enough evidence to establish the means of two samples are different and to
claim the alternate hypothesis. Thus, it is evident the rejection rates after introducing
incoming inspection are significant than the rejection rate without the inspection.
Figure 5 shows the rejection rates (RRo ) without inspection and rejection rates
post inspection (RRi ) and the standard error mean of 0.33 and 0.23, respectively.
The Improve of the DMAIC is addressed by analysing the results after the statis-
tical testing. The inspection of raw materials not only reduced the rejection rate
but also improved the First Pass Yield (FPY) from an average of 97.65–99.52%
contributing up to increase in 1.88%. As the production initiated post incoming

Table 1 F-test two-sample

Rejection rate Rejection rate (RRi)
for variances of rejection
(RRo)
rates with and without the
smart incoming inspection Mean 1.609 0.421
Variance σ1 2 = 1.645 σ2 2 = 0.397
Observations 15 7
df 14 6
F 4.14
P(F < = f) one-tail 0.045
F Critical one-tail 3.956
Impact of Smart Incoming Inspection System … 163

Table 2 t-test: two-sample

Rejection rate Rejection rate (RRi)
assuming unequal variances
(RRo)
of rejection rates due to the
unequal sample sizes Mean 1.609 0.421
Variance σ1 = 1.645
2 σ2 2 = 0.397
Observations 15 7
df 20
t Stat 2.9108
P(T < = t) one-tail 0.0043
t Critical one-tail 1.7247
P(T < = t) two-tail 0.0086
t Critical two-tail 2.0859

Fig. 5 Average rejection

rates without and with
inspection

Fig. 6 Average weekly First 100%

Pass Yield (FPY) tracked 99%
Production and Rejection (%)

over the weeks of the study.

98%
Weeks 1–5 are without smart
inspection system while 97%
weeks 6 and 7 were after 96%
introduction of smart
95%
inspection system
94%
week 1 week 2 week 3 week 4 week 5 week 6 week 7
Observations (No of weeks)

inspection, the rejections on week 6 and week 7 are comparatively low which
contributed on the overall FPY (see Fig. 6).
The Control is shown using a control chart (Fig. 7) with mean of μ = 1.231 and
an upper control limit (UCL) of μ + 3σ = 4.943. The number of rejections which
was closer to the UCL can be seen to be under control and equalizes to zero as the
number of days go by with the inclusion of smart incoming inspection system to
164 P. S. Kannaraya et al.

6
UCL 3σ

No. of Rejections
5
4
3
2
MEAN
1
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22

No. of Days

Fig. 7 Control chart for the number of rejections shown across the number of days. Number of
rejections reduced from day 16 once the smart incoming inspection was introduced. (the lower
control line is assumed 0, since rejections cannot go negative)

the production line. Since the mean is very close to zero the lower control line is
assumed to be zero as the number of defects cannot become negative.

6 Conclusions

The quality-related challenges in a medical device manufacturing MSME were the

adoption of state-of-the-art inspection system. A bespoke smart incoming inspection
system was proposed and developed. It was tested for its performance in live produc-
tion line of partnering MSME. The six sigma DMAIC methodology was adopted for
the quality control. The incoming raw materials were tested, and inspection data was
analysed using statistical process control method for the rejection rates before and
after introducing incoming inspection stage and smart incoming inspection system in
the production line. The analysis resulted in more than threefold decrease in rejections
and improvement in FPY by 1.88%.

References

1. Juran JM, Godfrey AB (eds) (1999) Juran’s quality handbook, 5th edn. McGraw Hill
2. Kang CW, Ramzan MB, Sarkar B, Imran M (2018) Effect of inspection performance in smart
manufacturing system based on human quality control system. Int J Adv Manuf Technol
94(9):4351–4364
3. Quality Glossary. American Society for Quality (ASQ). (Accessed 02 Jan 2020)
4. ISO 2859-1:1999, Sampling procedures for inspection by attributes — Part 1: Sampling
schemes indexed by acceptance quality limit (AQL) for lot-by-lot inspection; 3.1.1
5. Besterfield DH (2004) Quality control. Pearson Education, India
6. Kosmopoulos D, Varvarigou T (2001) Automated inspection of gaps on the automobile
production line through stereo vision and specular reflection. Comput Ind 46(1):49–63
7. Cheng K, Bateman RJ (2008) e-Manufacturing: characteristics, applications and potentials.
Prog Nat Sci 18(11):1
8. Recommendations of the Task Force on the Medical Devices Sector in India, Department of
Pharmaceuticals, GOI, 2015 (Accessed 08 Mar 2020)
Impact of Smart Incoming Inspection System … 165

9. Atul KSP, GH S, Pradipta B, Manish A, Amaresh C (2019) Designing an affordable system

for early defect detection using image processing. Adv Transdiscip Eng 121–128
10. Brewer A, Sloan N, Landers TL (1999) Intelligent tracking in manufacturing. J Intell Manuf
10(3):245–250
11. Gijo EV, Scaria J (2010) Reducing rejection and rework by application of Six Sigma
methodology in manufacturing process. Int J Six Sigma Compet Adv 6(1/2):77
12. Malek J, Desai D (2015) International journal for quality research. Int J Quality Res 9(4):577–
604
13. Gholap PC, Desai TN (2012) Reduction of rework the Six Sigma way: case study of an Indian
small scale industry. Int J Six Sigma Compet Adv 7(1):92–116
14. Yousaf F, Butt SI, Ahmad R, Six sigma implementation to reduce rejection rate of Pump Casings
at local Manufacturing Company
15. Foidl H, Felderer M (2015) Research challenges of industry 4.0 for quality management. In:
International conference on enterprise resource planning systems, pp 121–137. Springer, Cham