In Situ Calibration Algorithm to Optimize Energy Consumption in an

Automotive Stamping Factory Process

Ivan Peinado-Asensi

1,2 a

, N. Montes

1 b

and E. Garc

ıa

2 c

Mathematics, Physics and Technological Sciences Department, University CEU Cardenal Herrera,

C/ San Bartolom

e 55, 46115, Alfara del Patriarca, Valencia, Spain

Ford Spain, Pol

ıgono Industrial Ford S/N, 46440, Almussafes, Valencia, Spain

Keywords:

Predictive Maintenance, Machine Health Monitoring, Energy Consumption, IIoT, Sustanibale Development,

In Situ Calibration.

Abstract:

The world’s large factories in all sectors consume a great deal of resources, either raw materials or energy,

to develop their products. Saving resources can have a positive impact on the sustainable development of the

planet. Automotive manufacturers are a clear example of how to save by investing resources in improving

technologies and optimizing processes. This article focuses on one of the most common processes in the

automotive sector: the stamping process. For the optimization of this process, previous simulations are usually

carried out in order to deﬁne the optimal parameters and which should only be applied for a correct operation.

The real circumstances of the plant show there is a large discrepancy between the parameters obtained by

simulation and the real process because of the difference in material properties, lubrication, press operation,

etc. The solution is that the operators must adjust the parameters a posteriori and the only criterion to follow

is obtaining the right quality of the part. In many cases, the parameters are well above the ideal. This article

presents some algorithms used in order to perform an in situ calibration of the stamping presses to ﬁnd the press

parameters that, guaranteeing the quality of the part, allow to adjust the energy consumption to the minimum.

At the end of this article the experimental results from this in-situ calibration process and the energy savings

are shown.

1 INTRODUCTION

The stamping process consists of applying a force on

a sheet of reduced thickness (approximately 1mm)

with moulds designed according to the geometry of

the part to be manufactured. The main mechanical

characteristic that must be taken into account to ob-

tain a correct shaping of the part is the material elas-

tic limit. If the parts are of large dimensions, therefore

so must be the dies used. That is, the larger the die,

the greater the weight to be moved through the press,

being able to weigh up to 30 tons both the upper and

lower die. Displacing these large moulds and properly

shaping materials such as steel or aluminum requires

the use of industrial presses with high workforce ca-

pacity, that is, for example when forming larger parts

such as a side, larger presses will be needed as com-

https://orcid.org/0000-0002-5603-5910

https://orcid.org/0000-0002-0661-3479

https://orcid.org/0000-0002-4210-9835

pared to manufacturing small parts such as door rein-

forcements.

Regarding the manufactured parts, the fundamen-

tal priority in the plant is to avoid passing defective

parts to the following assembly processes, since a de-

fective part in the subsequent processes increases the

cost of the losses. For this, comprehensive quality

controls are carried out at the end of the line where

we can verify whether the product is suitable or not.

In the event of a change in the quality of the prod-

uct, adjustments are made to the process to verify that

with the parameters entered the parts are within the

quality margins in order to continue producing nor-

mally. This is a work that requires great knowledge

since in the stamping process more than 40 different

variables are involved, some of them are more impor-

tant than others but just a small modiﬁcation in one

of them can cause quality defects, whether they are

wrinkles, breaks, stretches, etc. And these variables

are both due to the type of materials and the sheet

thickness, the surface roughness, the amount of lu-

Peinado-Asensi, I., Montes, N. and García, E.

In Situ Calibration Algorithm to Optimize Energy Consumption in an Automotive Stamping Factory Process.

DOI: 10.5220/0011272600003271

In Proceedings of the 19th International Conference on Informatics in Control, Automation and Robotics (ICINCO 2022), pages 169-176

ISBN: 978-989-758-585-2; ISSN: 2184-2809

169

bricant used, and the anisotropy and elasticity among

others, as well as due to the equipment where wear in

the die, mismatch in parameters such as the regulation

or the stamping pressure of the material among others

may happen.

To all this we must add the day-to-day work of

an automotive plant, where it is essential to have the

ability to be ﬂexible and dynamic to adapt to the de-

mand required at all times. In the stamping plant

at Ford Spain facilities we have a large number of

lines of different capacities, which allows us to have

a great adaptation to the demand required. There are

a large number of car chassis parts manufactured in

the factory for the ﬁve models that are currently being

made, these may undergo changes in the manufactur-

ing standards, either by the type of material, geometry

of the part, line change or other factors. Therefore,

try-out tests are carried out for a new adjustment and

to ensure that the quality of the product during mass

manufacture is correct.

The tests carried out in the try-out procedure for

the new adjustment consist of modifying the manu-

facturing parameters until reaching the most optimal

point of resources use, both equipment and material.

For example, adjustments are made to the working

parameters of the press in the calibration to control

the pressure with which the part is made, parameters

of the pressure made in the press are also adjusted to

control the amount of material drawn into the die in

order to ensure no cracks occur due to excess stroking

or wrinkles due to the lack of it. Other types of mate-

rial can even be used instead of the one initially pro-

posed, taking into account a different mechanical be-

haviour with which the work parameters must be ad-

justed again.

As for the characteristics of the presses we have in

the plant, we can classify them in two groups, the me-

chanical and hydraulic presses. Most of the presses

in the plant belong to the ﬁrst group that includes,

on the one hand, the cutting presses, which work at

high speeds and perform the cutting of the coil in the

blank parts that will later be used in the stamping pro-

cess. We have also the stamping presses, character-

ized by being the largest ones, with which different

operations are carried out, such as deep-drawing, cut-

ting, drilling, bending and spring-back. These presses

work at high levels of pressure due to the size of the

moulds used for the aforementioned operations, espe-

cially those of deep-drawing. Within the mechanical

presses we have two types, single-action and double-

action. The latter are the ones that have been used for

a long time for forming stamping parts, characterized

by having two eccentric transmission systems in the

press head. And on the other hand the single-action

presses, which are more efﬁcient, incorporating an in-

telligent hydraulic cushion at the bottom of the press

including only one eccentric transmission system on

the head.

Following the current trends in predictive mainte-

nance, we intend to implement at plant level a mon-

itoring system of the presses we have in the factory

to ﬁnd out their working status and be able to an-

ticipate possible faults. It is known that the imple-

mentation of this type of industrial projects requires

a great economic investment, but in our case, follow-

ing the philosophy proposed by the Miniterms (Gar-

cia and Montes, 2019), we intend to take advantage

of the maximum of available sensors and taking into

account the information that can be extracted from

these develop new solutions to monitor the health of

the equipment. This is a great advantage we have in

the stamping plant, since most presses come equipped

with a lot of sensors thus having at our disposal a lot

of information at no extra cost.

The sensorization of the presses with strain gauges

has been used for years to deﬁne pathologies of the

equipment from different points of view, including di-

agnosing failures in the stamping process of both the

equipment and the manufactured product (Koh et al.,

1996). This can be done by applying different tech-

niques to obtain information from measured data such

as wavelets (Jin and Shi, 1999), relying on experi-

ments (Jin and Shi, 2000) and even applying machine

learning techniques by using neural networks (Bassi-

uny et al., 2007). Going a step beyond the detection of

pathologies, process control systems have also been

developed based on the graph obtained from the ton-

nage of each cycle (Zhou et al., 2015) or by ﬁnding

variations in the lubrication of the process and wear

of the die (Voss et al., 2017). Following the trends

of internet of things in industry (IIoT) we now have a

lot of available data in real time to model the process,

as it has been done in this ﬁeld by (Niemietz et al.,

2020). Hence, in this paper we show the ﬁrst insight

we have obtained from the process and the advantages

we are taking from the application developed for solv-

ing detected issues and optimizing the process from

the point of view of energy consumption.

Optimizing the process to achieve energy savings

is vitally important due to two major factors. One of

them aims to achieve a sustainable development of the

planet, reducing pollution and saving on available re-

sources. And the second is the economic factor, since

during the last year the price on the electricity bill at

the factory has doubled the price and it is predicted

that this upward trend will continue, thus these costs

indirectly will affect the proﬁt per car produced at the

factory.

ICINCO 2022 - 19th International Conference on Informatics in Control, Automation and Robotics

170

For the optimization of this process, previous sim-

ulations are usually carried out in order to deﬁne the

optimal parameters which should only be applied for

a correct operation. The real circumstances of the

plant show there is a large discrepancy between the

parameters obtained by simulation and the real pro-

cess either because of the difference in material prop-

erties, lubrication, press operation, etc. The solution

is that the operators must adjust the parameters a pos-

teriori and the only criterion to follow is obtaining the

right quality of the part. In many cases, the param-

eters are well above the ideal. Being able to have

knowledge of a process and applygin in situ solu-

tions is a trend that is increasingly used to develop

technological solutions, as can be seen in (Grasso and

Colosimo, 2017) and (Buchli et al., 2018). Hence, in

this article presents some algorithms used to be able to

perform an in situ calibration of the stamping presses

in order to ﬁnd out the press parameters that, guar-

anteeing the quality of the part, allow to adjust the

energy consumption to the minimum.

The paper is structured as follows: in the next sec-

tion we will explain previous considerations to take

into account which are our purpose with the proposed

methodology and in section 3 the following method-

ology to optimize the process will be shown. In the

fourth section we will show a real case of how we

have detected an abnormal function of the process and

the modiﬁcations carried out. Finally, the conclusions

and future works are proposed in section 5.

2 BACKGROUND

Together with the consequences that happen in an au-

tomotive plant as explained in the previous point, to

this must be added the consumption of resources that

is generated, due to raw materials, labour, and en-

ergy consumption. Therefore, for proper sustainable

development and to be able to manufacture with the

least possible impact on the environment, the aim is

to optimize the use of available resources even by

reusing materials that are classiﬁed as scrap due to

the process.

What is sought in this research is to reduce the en-

ergy consumption of the stamping plant from the data

extracted from the process in order to know how the

presses are working with the different sets of dies and

to look for an optimal point of the working parameters

which ensures minimum energy consumption without

affecting the quality of the manufactured material.

2.1 Previous Considerations

Throughout the stamping process, as mentioned

above, the presses that carry out the Deep-Drawing

process are the ones that can apply the highest pres-

sure. Therefore, these are the presses with the largest

electric motors and on which we will focus to develop

our tool.

In stamping, two phases are required to get a cor-

rect conﬁguration of the work parameters, in a ﬁrst

phase the simulation of the process is carried out with

programs speciﬁcally designed for stamping. From

which the working parameters of the press are deﬁned

by taking into account the design of the part and char-

acteristics of the material. But due to the variations

between the simulation and the real world, a second

phase is required in which try-out tests are carried

out in the plant by performing an adjustment of the

equipment in order for the product to meet the quality

requirements, in which the ﬁnal process adjustments

can vary with respect to those deﬁned by the design.

The main properties of the material to take into ac-

count in the forming are the deformation the sheet un-

dergoes in the different areas along its entire surface.

Due to the non-linear plastic deformation character-

istic of the deep drawing and the non-homogeneous

material ﬂow during the process, it is very difﬁcult to

control its efﬁciency.

E =

(1)

In the forming process of a steel sheet there are

several properties of the materials that are fundamen-

tal to take into account to achieve a correct deep draw-

ing without causing defects. One of the possible de-

fects is wrinkles which can be caused by several fac-

tors in the process, but mainly because the elastic limit

of the material has not been exceeded and therefore it

tries to return to its initial position.

To do this, a certain tension must be applied that

allows exceeding the elastic limit of the material but

without causing breaks or stretches that do not pro-

vide structural stability. Therefore, the result of the

tension against deformation can be seen in the red

zone in ﬁgure 1 where the material is in the hardening

zone.

To obtain stamped product information, we are

able to obtain the deformation (ε) of a stamped sheet

to ensure that we will have no quality defects with the

ﬁnal conﬁguration. The process consists of printing

a mesh in the plate by electrolytic marking. After-

wards, the stamped part is taken to the laboratory and

with a high-resolution camera, with 20 million pixels,

photos are taken of the different areas to be analyzed,

the example shown in ﬁgure 5 is the central beam of

In Situ Calibration Algorithm to Optimize Energy Consumption in an Automotive Stamping Factory Process

171

Figure 1: Stress–Strain Curve.

the measured part. Using the AutoGrid® software we

can obtain an output of the FLD of the area analyzed

with the strain of each node measured, showing the

maximum strain ε

and ε

per node in the diagram

with the bounds of the diagram previously deﬁned for

the speciﬁc material.

Figure 2: FLD of a Real Stamped Part.

According to this research development, once the

working parameters are known in real time and the

stamping process is understood in depth, the process

will be optimized from the point of view of press

work, where the least amount of energy resources is

used for the shaping of the parts, we verify that the

consumption of the electric motor of the press and

the measured tonnage are directly related in a propor-

tional manner. In the result section, we have veriﬁed

this through real measurements of the process.

2.2 Previous Work

Real-time data monitoring is essential to know the sta-

tus of the machine. By applying IIoT (Industrial In-

ternet of Things) techniques we are able to receive

a large amount of process information, the pressure

made by the presses, pressures of the hydraulic and

pneumatic groups, press movement activation time,

speeds and positions of the different moving ele-

ments among others. Of these, a real-time application

(Peinado-Asensi. et al., 2021) was developed based

on the information obtained from the tonnage sensors

so that we have been able to ﬁnd out the quality of

the press throughout the cycle in each and every press

stroke during production.

In order to obtain information on the relevant pro-

cess, an alarm sending system has been deﬁned so that

for each of the different parts of the car produced, we

can store the press cycles that fail including detailed

information on which press position said anomalous

event has occurred. The reason why it has been made

for each part is because depending on the geometry

of the die the pattern of the gravity centre may vary

considerably. We also have a parallel system used to

control the press at all times, which warns us about

how much it is deviating from normal functioning to-

wards the limits of the press. We currently have these

margins deﬁned as shown in the picture below.

0,00

500,00

1000,00

1500,00

2000,00

2500,00

-2500 -1500 -500 500 1500 2500

Slide length

LIMIT FORCE X AXIS

Fx max Warning Safe

Figure 3: Health Monitoring Limits.

To validate that the tool worked as expected and

the data received from the Gravity Centre were cor-

rect, a test was carried out with 4 columns mounted at

the same height at the base of the press, then we un-

balanced them by adding 1 mm thick sheets between

the different columns and after shifting the gravity

centre towards each of the sides, we found out that the

results obtained coincided with what was expected,

being able to determine that our system works as ex-

pected.

The value of the sensor has been taken for each

press position to obtain the gravity centre and the

equation shown below has been applied,

∑

, i = 1, . . . , 4, j = 0, . . . , 359. (2)

Where T is the value of the tonnage sensor and

D the distance from the rod foot to the centre of the

press slide.

ICINCO 2022 - 19th International Conference on Informatics in Control, Automation and Robotics

172

Figure 4: Gravity Center Graph.

3 IN SITU CALIBRATION

ALGORITHM PROPOSAL

We have 3 critical variables that directly affect the en-

ergy transmitted by the ﬂywheel to the eccentric trans-

mission system and this energy of the ﬂywheel is cal-

culated as shown

E =

· J · ω

, (3)

where J is the inertia transmitted by the ﬂywheel

and ω is the angular speed at which the ﬂywheel ro-

tates. In this case the inertia is ﬁxed since it depends

directly on the geometry of said ﬂywheel where the

mass and the radius are always the same, therefore the

greater the energy requirement, the greater the speed

of rotation transmitted by the electric motor of the

press by means of a pulley transmission system.

One of the parameters evaluated as critical with

respect to the energy required for the press movement

is the die weight, which depending on the part pro-

duced can be higher or lower, between 5 and 30 tons

approximately. The second parameter is the compen-

sation pressure, which consists of two large volume

pneumatic cylinders in which the pressure is adjusted

depending on the die weight in order to help the press

raise the slide and die at the end of the cycle. And

third, the adjustment of the press slide, the lower the

height of the slide, the greater the pressure in the die

to exert a travel of the upper die and therefore greater

energy will be needed to be transmitted from the iner-

tia ﬂywheel.

Therefore, the control algorithm proposed to ad-

just the working parameters of the press and ensure

that the energy consumed is minimal is as follows.

We can see there are two variables that can be

modiﬁed to optimize the power consumption, the ad-

justment of the slide and the pressure of the compen-

sation cylinders.

Die Weight

Regulation

Counterbalance

Pressure

Tonnage

Energy

Flywheel

Car Part

Output

Figure 5: In-loop Control Diagram.

3.1 Slide Adjustment

On the one hand we have the slide adjustment, this

parameter is very sensitive and critical in terms of

part quality. Since modifying tenths of a millime-

tre, both upwards and downwards, signiﬁcant quality

defects, wrinkles or breaks may appear, respectively,

thus turning the product into scrap. This parameter

will be modiﬁed by try-out tests in the event that the

press pressure is detected outside the designed work

value. A real case of this setting will be shown in the

next point.

Being a parameter directly related to the force ex-

erted by the press, with the measurement of the ton-

nage sensors we can obtain information on whether

the adjustment is adequate or programmed at a value

that can damage the mechanical elements of the press

by performing a force greater than necessary.

The way to proceed will be as follows, starting

from a tonnage detected as anomalous in the sense of

exceeding the design working conditions. Tests will

be carried out by modifying the adjustment in half a

millimeter each time. Then a stamping cycle is car-

ried out verifying, on the one hand that the quality

of the produced part has not been affected and on the

other hand, the resulting tonnage value, in case of be-

ing able to reduce the tonnage, this process will be

repeated until reaching a value of the slide adjustment

which does not cause wrinkle or springback defects in

the material.

3.2 Counterbalance Pressure

On the other hand, we have the compensation, so for

this variable the energy consumption in a cycle will

be measured for different pressure values. With the

values obtained, we can proceed as explained below.

As explained above, this is an auxiliary system

that exclusively affects the operation of the equip-

ment, which will cause two cases of mismatch. On the

one hand, the manufacturer provides us with a list of

compensation pressure values depending on the upper

die weight. On several occasions this value has been

In Situ Calibration Algorithm to Optimize Energy Consumption in an Automotive Stamping Factory Process

173

modiﬁed with readjustments after a fault, for exam-

ple, or external factors that cannot be controlled. In

some cases, depending on the consumption, we will

need to optimize the value required by the manufac-

turer, which may not be the most optimal one. The

following methodology is therefore proposed for ad-

justing the compensation value. To achieve the min-

imum energy, the gradient descent method is applied

for n iterations

n+1

= x

− α∇ f (x

), (4)

for a known function f (x) that will give us the energy

value with regard to a pressure value, but the problem

is that from our measurements we cannot deﬁne an

exact function for the behaviour of the measured val-

ues. To do this, we are going to obtain an approximate

function p(x) by using the method of approximation

by polynomials of degree n,

p(x) = a

+ a

x + a

+ ··· + a

, (5)

where having n points we will need n + 1 measure-

ments, being

p(x) ' f (x). (6)

At this point we must obtain the values of a

, . . . , a

by solving the system of n equations with n unknowns

that we would have left, being n the number of total

measurements made. Once the function is obtained,

the gradient descent method is applied to obtain the

lowest energy value with respect to the compensation,

obtaining the optimal value for the process and adjust-

ing it online according to our control system shown in

ﬁgure 5.

4 EXPERIMENTAL VALIDATION

For the measurement of the energy consumed with

the different working parameters, the Fluke 438-II

Power Quality and Motor Analyzer equipment has

been used, connected to the output of the electric mo-

tor drive as can be seen in ﬁgure 6, being able to mea-

sure the energy consumed directly by the electric mo-

tor.

4.1 In Situ Slide Regulation Calibration

Among the available lines in the stamping shop, there

are two twin lines to manufacture big car body parts

Figure 6: Energy Measurement.

with 2500 tons of capacity in the deep drawing op-

eration, where the gravity centre system is currently

monitoring data. During the period the app has been

working, no signs of malfunction or imbalance have

been found in the slide. There is an alarm system that

warns us about the number of strokes and which press

position the imbalance is taking place. There are two

ways of reading the information of the gravity centre,

a weekly report with the productions status of each die

and a real time dashboard where it has been counted

the number of times our security limits are exceeded.

So we can monitor the working status of the machine

all the time.

Until now, all the monitored parameters men-

tioned before have been correct without having any

abnormal behaviour, but there is still other important

limit to check: the working load limit of the press.

The press manufacturer recommended not to exceed

at least 80% of the limit force, as those machines are

designed to work with 2500 tons maximum, but the

advised number should be 2000 tons to ensure a cor-

rect working status.

Our alarm system detected a part that was be-

ing manufactured with 2300 T as can be seen in ﬁg-

ure 7, an unusual load that had not been previously

recorded. We checked it and found out the follow-

ing events: the part produced was the roof of the Ford

Transit van in its long version shown in 8, when man-

ufacturing this part one of the largest forming dies

available in the plant was used.

It was shocking at ﬁrst because a simple geometry

car part should not cause problems unless something

strange happened, but everything seemed normal. So

ICINCO 2022 - 19th International Conference on Informatics in Control, Automation and Robotics

174

Figure 7: Gravity Center Graph.

after making some research to ﬁgure out what would

have caused this event, we realized that the die set

was changed from one line to another, that is, new

parameters were input in the stamping line and clearly

some error would have caused this.

Figure 8: Car Body Part - Roof.

Comparing working parameters from one line to

another, before the tonnage data monitoring the load

was around 1500 tons. So try-out jobs were organized

to modify parameters to optimize the process. Table

1 shows the measured values before and after the in

situ adjustment.

Table 1: Working parameters comparation.

PRE POST

Tonnage (T) 2292 1618

Counterbalance (bar) 7.4 7.4

Slide reg. (mm) 1371.5 1373.5

Energy (kWh) 416 359

Where you can see that the tonnage has been re-

duced by about 700 tons by modifying the adjustment

only 2mm and the energy consumption in one hour of

work has been reduced by 57 kWh.

4.2 In Situ Counterbalance Adjustment

The test for compensation was performed on a smaller

press as compared to the anomaly presented in the

previous case. In this case we used a double-action

mechanical press with a maximum capacity of 1500

tons, where smaller parts are manufactured as com-

pared to those used in the presses described in the

previous point. Here we have the peculiarity that for

each type of die, because of the weight difference be-

tween them, different energy values will be obtained.

Therefore we should repeat the process for each of the

different car chassis parts manufactured.

The values shown in table 2 are the measurements

for a press cycle showing the consumption for each of

the pressures from 3 to 6 bars during several strokes.

After measuring several strokes for a ﬁxed pressure,

the average Wh conpsumtion per hour was obtained.

Table 2: Working parameters comparation.

Measure Pressure (bar) Energy (Wh)

1 3 45

2 3.5 44.25

3 4 44.5

4 4.5 46.75

5 5 44.25

6 5.5 45.5

7 6 48.75

The function that deﬁnes the distribution of points

should have a shape like the one shown below and it

will be the one we need to look for, an approximate

function p(x).

25,00

30,00

35,00

40,00

45,00

50,00

55,00

60,00

3 3,5 4 4,5 5 5,5 6 6,5 7

Energy (Wh)

Counterbalance (bar)

Figure 9: Couterbalance - Energy Graph.

To obtain the function that describes the behaviour

of the consumption with respect to the pressure, we

will apply the method of approximation by polyno-

mials deﬁned in equation 5, therefore we will have

the following expression:

p(x) = a

+ a

x + a

+ a

(7)

There is a system of 6 equations with 6 unknowns

to solve in order to have the values of the parame-

ters a

, . . . , a

and obtain the approximate function

that will give us the energy consumption for a given

compensation. The values obtained for this die are as

follows;

In Situ Calibration Algorithm to Optimize Energy Consumption in an Automotive Stamping Factory Process

175

= −2.0585 · 10

, a

= 2.9929 · 10

= −1.7837 · 10

, a

67063

= −971, a

1331

, a

= −

Once the function is obtained, the gradient descent

method is applied to obtain the lowest energy value

with respect to the compensation, obtaining the opti-

mal value for the process and adjusting it online ac-

cording to our control system shown in ﬁgure 5.

5 CONCLUSIONS AND FUTURE

WORKS

This article shows two in situ calibration techniques

of stamping process parameters that allow to reduce

the energy consumption of the manufacturing pro-

cess. In-situ calibration methods are commonly used

to adjust sensors or processes, which cannot be ad-

justed from the factory since the discrepancy between

the ideal and actual situation is signiﬁcantly differ-

ent. This is the case of the stamping process where

the values assumed in the simulation process and the

reality of the process require the operators to make

adjustments on the process and where it is common

to ﬁnd them oversized. The calibration algorithms

proposed in this article make it possible to perform

this calibration and achieve signiﬁcant energy sav-

ings, as demonstrated in this paper. As future works,

we will intend to generalize these in-situ calibration

techniques to the rest of the presses and also other

processes of the factory having the same problem.

ACKNOWLEDGEMENTS

This study was supported by the Universidad CEU

Cardenal Herrera, Ford Spain S.L. and Fundaci

para el Desarrollo y la Innovaci

on (FDI), Spain,

which the authors gratefully acknowledge.

REFERENCES

Bassiuny, A., Li, X., and Du, R. (2007). Fault diagno-

sis of stamping process based on empirical mode de-

composition and learning vector quantization. Inter-

national Journal of Machine Tools and Manufacture,

47(15):2298–2306.

Buchli, J., Giftthaler, M., Kumar, N., Lussi, M., Sandy,

T., D

orﬂer, K., and Hack, N. (2018). Digital in situ

fabrication-challenges and opportunities for robotic in

situ fabrication in architecture, construction, and be-

yond. Cement and Concrete Research, 112:66–75.

Garcia, E. and Montes, N. (2019). Mini-term, a novel

paradigm for fault detection. IFAC-PapersOnLine,

52(13):165–170.

Grasso, M. and Colosimo, B. M. (2017). Process defects

and in situ monitoring methods in metal powder bed

fusion: a review. Measurement Science and Technol-

ogy, 28(4):044005.

Jin, J. and Shi, J. (1999). Feature-preserving data compres-

sion of stamping tonnage information using wavelets.

Technometrics, 41(4):327–339.

Jin, J. and Shi, J. (2000). Diagnostic feature extraction from

stamping tonnage signals based on design of experi-

ments. J. Manuf. Sci. Eng., 122(2):360–369.

Koh, C. K., Shi, J., Black, J. M., and Ni, J. (1996). Ton-

nage signature attribute analysis for stamping pro-

cess. Transactions-North American Manufacturing

Research Institution of SME, pages 193–198.

Niemietz, P., Pennekamp, J., Kunze, I., Trauth, D., Wehrle,

K., and Bergs, T. (2020). Stamping process modelling

in an internet of production. Procedia Manufacturing,

49:61–68.

Peinado-Asensi., I., Montes., N., and Garc

ıa., E. (2021).

Towards real time predictive system for mechanical

stamping presses to assure correct slide parallelism.

In Proceedings of the 18th International Conference

on Informatics in Control, Automation and Robotics -

ICINCO,, pages 396–402. INSTICC, SciTePress.

Voss, B., Pereira, M., Rolfe, B., and Doolan, M. (2017). Us-

ing stamping punch force variation for the identiﬁca-

tion of changes in lubrication and wear mechanism. In

Journal of Physics: Conference Series, volume 896,

page 012028. IOP Publishing.

Zhou, C., Liu, K., Zhang, X., Zhang, W., and Shi, J. (2015).

An automatic process monitoring method using recur-

rence plot in progressive stamping processes. IEEE

Transactions on Automation Science and Engineer-

ing, 13(2):1102–1111.

ICINCO 2022 - 19th International Conference on Informatics in Control, Automation and Robotics

176