Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com
ScienceDirect
ScienceDirect
Procedia
Available Manufacturing
online 00 (2018) 115–121
atatwww.sciencedirect.com
Available
Procedia online www.sciencedirect.com
Manufacturing 00 (2018) 115–121 www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia
ScienceDirect
ScienceDirect 
Procedia Manufacturing 22 (2018) 107–113
Procedia Manufacturing 00 (2017) 000–000
www.elsevier.com/locate/procedia
11th International Conference Interdisciplinarity in Engineering, INTER-ENG 2017, 5-6 October
11th International Conference Interdisciplinarity in Engineering,
2017, Tirgu-Mures, Romania INTER-ENG 2017, 5-6 October
2017, Tirgu-Mures, Romania
Determination of Johnson-Cook material model parameters by an
Determination of Johnson-Cook
Manufacturing Engineering Society International material
Conference model parameters
2017, MESIC by June
2017, 28-30 an
optimization approach 2017, Vigo using the fireworks
(Pontevedra), Spain algorithm
optimization approach using the fireworks algorithm
a, a
Costing modelsNikolaos
for E. Karkalos
capacity a, *, Angelos P.in
optimization P

P
P0F

P0F
Markopoulos
Nikolaos E. Karkalos *, Angelos P. Markopoulosa 4.0: Trade-offIndustry P

between usedUniversity
capacity of Athens,and
Heroonoperational efficiency
a
National Technical University of Athens, Heroon Polytechniou 9, Athens 15780, Greece
P P

a
National Technical
P P Polytechniou 9, Athens 15780, Greece

Abstract
A. Santanaa, P. Afonsoa,*, A. Zaninb, R. Wernkeb
Abstract a
University of Minho, 4800-058 Guimarães, Portugal
The accurate description of materials behavior b is essential
Unochapecó, when investigating
89809-000 Chapecó, SC, the application of these materials in the industrial
Brazil
The accurate description of materials behavior
practice. For this purpose, experimental data are is essential
fitted to when investigating
suitable constitutivethelaws
application of thesemethods.
using various materialsIninthis
thepaper,
industrial
the
practice. For this
determination purpose, Cook
of Johnson experimental
material data
model areparameters
fitted to suitable constitutive
for AISI laws using
316L stainless various
steel is methods.
performed usingInanthis paper, the
optimization
determination
approach. More of specifically,
Johnson Cook the material
Fireworks model parameters
algorithm for AISIto 316L
is employed stainless
determine steel is performed
the appropriate materialusing an optimization
parameters for AISI
Abstract
approach.
316L steel.More
The specifically, the aFireworks
results indicate sufficientalgorithm
performance is employed to determine
of the optimization the appropriate
approach, which canmaterial parameters
be further for AISI
developed and
316L steel. The results indicate a sufficient
applied for other types of materials as well. performance of the optimization approach, which can be further developed and
Under
© 2018 the
applied for concept
Theother of materials
"Industry
typesPublished
Authors. 4.0", B.V.
as well.
by Elsevier production processes will be pushed to be increasingly interconnected,
© 2018 The under
information
Peer-review Authors. Published
basedresponsibility
on by
of Elsevier
a real time basis
the B.V. necessarily,
and,
scientific committee of much
the 11thmore efficient.Conference
International In this context, capacity optimization
Interdisciplinarity in
© 2018 The Authors.
Peer-review under Published by
responsibility Elsevier
ofof
thecapacityB.V. committee of the 11th International Conference Interdisciplinarity in
scientific
Engineering.
goes beyond the traditional aim maximization, contributing also for organization’s profitability
Peer-review under responsibility of the scientific committee of the 11th International Conference Interdisciplinarity in Engineering. and value.
Engineering.
Indeed, lean management and continuous improvement approaches suggest capacity optimization instead of
Keywords: stress-strain
maximization. The relationship; Johnson-Cook
study of capacity material model;
optimization andartificial
costingintelligence;
models isFireworks algorithm;
an important optimization;
research topicAISI 316L
that deserves
Keywords:
stainless stress-strain relationship; Johnson-Cook material model; artificial intelligence; Fireworks algorithm; optimization; AISI 316L
steel.
contributions
stainless steel. from both the practical and theoretical perspectives. This paper presents and discusses a mathematical
model for capacity management based on different costing models (ABC and TDABC). A generic model has been
developed
1. Introductionand it was used to analyze idle capacity and to design strategies towards the maximization of organization’s
value. The
1. Introduction trade-off capacity maximization vs operational efficiency is highlighted and it is shown that capacity
optimization
Nowadays,might hide operational
the development inefficiency.
of new alloy materials and generally, the need for using new materials in high end
© 2017 The Authors.
Nowadays,such Published
the as
development by Elsevier
of new B.V.
alloy
applications, aerospace industry, leadsmaterials and generally,
to a consequent the need
development for using
regarding thenew materialsstudies
theoretical in high
of end
the
Peer-review
applications, under
suchresponsibility
as aerospace of the scientific
industry, committee
leads to of the Manufacturing
a consequent development Engineering
regardingSociety
the Internationalstudies
theoretical Conference
ofnew
the
behavior
2017.
of these materials. Thus, considerable work is conducted in order to determine the material behavior of
behavior
materials ofusing theseexisting,
materials.modified
Thus, considerable work is conducted
or new constitutive material in order toAfter
models. determine the material
the parameters of behavior
the modelsof new
are
materials using existing, modified or new constitutive material models.
Keywords: Cost Models; ABC; TDABC; Capacity Management; Idle Capacity; Operational Efficiency
After the parameters of the models are

1. Introduction
* Corresponding author. Tel.: +30-210-772-3897.
* E-mail
Corresponding
address:author. Tel.: +30-210-772-3897.
nkark@mail.ntua.gr
The cost
E-mail of idle
address: capacity is a fundamental information for companies and their management of extreme importance
nkark@mail.ntua.gr
in modern©production
2351-9789 systems.
2018 The Authors. In general,
Published it isB.V.
by Elsevier defined as unused capacity or production potential and can be measured
in several©under
2351-9789
Peer-review ways: tons of production,
2018responsibility
The Authors. Published
of available
by Elsevier
the scientific B.V.hours
committee of manufacturing,
of the 11th etc. The
International Conference management
Interdisciplinarity of the idle capacity
in Engineering.
Peer-review underTel.:
* Paulo Afonso. responsibility
+351 253 of the761;
510 scientific committee
fax: +351 253 604of741
the 11th International Conference Interdisciplinarity in Engineering.
E-mail address: psafonso@dps.uminho.pt

2351-9789 © 2017 The Authors. Published by Elsevier B.V.

Peer-review under responsibility of the scientific committee of the Manufacturing Engineering Society International Conference 2017.
2351-9789 © 2018 The Authors. Published by Elsevier B.V.
Peer-review under responsibility of the scientific committee of the 11th International Conference Interdisciplinarity in Engineering.
10.1016/j.promfg.2018.03.017
108 Nikolaos E. Karkalos et al. / Procedia Manufacturing 22 (2018) 107–113
116 N.E.Karkalos and A.P. Markopoulos / Procedia Manufacturing 00 (2018) 115–121

determined, the characteristics of these new materials and the applicability of them in various manufacturing
processes can be investigated using numerical simulations such as Finite Elements (FEM) simulations.
The determination of suitable constitutive model parameters according to experimental data can be conducted
using methods such as regression analysis [1]. One promising category of methods is pertinent to Soft Computing,
which comprises methods such as Artificial Neural Networks (ANN). In the relevant literature, several works exist
which employ soft computing or even optimization methods for the prediction of stress-strain relationship. For
example, Gupta et al. [2] studied the flow stress of 316 stainless steel during dynamic strain aging regime with the
use of an ANN model. In their model, data from tests covering a temperature range of 350-650oC and strain rate
range of 10-4 and 10-2 s-1 were employed. Kong et al. [3] used an ANN model combined with a phenomenological
model (Estrin-Mecking model) and found out that this approach was superior to other constitutive models, even at
conditions outside the range of experimental data which were used to develop their model.
Furthermore, various works also exist, where an optimization method is employed for the determination of
material behavior based on experimental data. For example, Dimatteo et al. [4] used genetic algorithms (GA) for the
prediction of Mean Flow stress during rolling process. At first, they obtained results using models available in the
literature and then, using GA, they derived optimum values for the existing models and compared the results. It was
found that optimized models outperformed the original models and the results were consistent with industrial
measurements. Özel and Karpat [5] also employed an optimization approach in order to derive the Johnson-Cook
material parameters for 4 different materials, namely AISI 1045 and 4340 steels, AA 6082-T6 aluminum alloy and
Ti6Al4V titanium alloy. For this purpose they employed a swarm intelligence method (cooperative Particle Swarm
Optimization) and found out that this method outperformed simpler Particle Swarm Optimization variants. Lin and
Yang [6] were able to fit their experimental data for Ti6Al4V alloy at 927oC to viscoplastic constitutive equations
using a Genetic Algorithm multi-objective approach and noted the excellent capabilities of this approach to provide
material parameters for several material behavior models. Huang et al. [7] employed a cluster global optimization
algorithm to obtain material properties for two types of steel and they managed to obtain high levels of accuracy for
their predicted stress-strain curves.
Finally, as material modeling is also very important for numerical simulations of manufacturing processes, some
researchers employ their derived model to actual manufacturing processes simulation or even attempt to predict
suitable material model parameters by matching the simulation results to the experimental ones. Chen et al. [8]
derived a Johnson-Cook material model for Ti-6Al-4V and compared simulation results concerning cutting forces
and machining chip morphology to experimental ones. Some representative works regarding the determination of
Johnson-Cook parameters from machining experiments can be found in works such as [9, 10].
The purpose of this paper is the determination of parameters of Johnson-Cook constitutive model for 316L
stainless steel, using an optimization approach. This is achieved through two major steps; the first step involves the
collection of stress-strain curve data and the second involves the fitting of these data to Johnson-Cook material
model by using for this reason the Fireworks algorithm, a relatively new and promising stochastic optimization
method.

2. Fireworks algorithm

2.1. General aspects

In the current work, the Fireworks optimization algorithm (FWA) is employed for the determination of
constitutive material model parameters. This method belongs to the Swarm Intelligence (SI) category of
optimization methods, was introduced in 2010 by Tan and Zhu [11] and has found many applications in various
fields of science. This method was inspired from actual fireworks displays and contains elements from the
observation of tracks of sparks which were produced by the fireworks. Thus, the "sparks" from an initial "explosion"
of fireworks fill a region which is considered as the search space for the algorithm and until the optimum is detected,
"explosions" from specific regions occur, with a view to approach the optimum point.
 Nikolaos E. Karkalos et al. / Procedia Manufacturing 22 (2018) 107–113 109
N.E Karkalos and A.P. Markopoulos / Procedia Manufacturing 00 (2018) 115–121 117

2.2. Description of the algorithm

In this subsection, the basic details about the FWA will be presented. However, several variants of this algorithm
have been proposed since its invention, either by the inventors themselves or by other researchers, something that
will be discussed in subsection 2.3.
The basic idea, on which the algorithm was established, was that after a firework is set off, a "cloud" of sparks
will be dispersed within a distance from the firework [12]. This observation bears a resemblance to a search in the
space around a point, which is essential function of an optimization algorithm. The explosion of fireworks is taking
place at several locations in the search area (n explosions). After the sparks are created and reach the final location,
the corresponding candidate solutions are evaluated using a fitness function and the process continues until the
termination criteria are met.
Concerning the efficiency of explosions, it was observed that a "good explosion" involves a sufficient number of
sparks, almost uniformly distributed around the location of the explosion. "Good fireworks" are the ones located in
an area which is probably close to the optimum point, so a larger number of sparks are produced from these
locations in order to reach the optimum point. On the contrary, fewer sparks should be produced from a "bad
firework" and the distance of sparks from the explosion point should be larger in order to move away from the area
near the "bad firework" and thus obtain better values of fitness functions [12].
Another important parameter for the fireworks algorithm is the amplitude of explosions (A). This parameter has a
smaller value for "good fireworks", as it is related to the distance at which the sparks will land. As it is anticipated,
this parameter is adjusted accordingly in the case of a "bad firework". Finally, the diversity of the location of the
sparks is further ensured by the generation of a number of sparks from random locations determined according to a
Gaussian function.
At the end of the evaluation of the sparks by the use of the fitness function, the best candidate solution and n-1
other solutions randomly chosen from the total sparks produced according to the selection probability value are
selected and the algorithm continues until termination criteria are met.

2.3. Variants of the algorithm

It is common for optimization algorithms to be modified in order to produce favorable results in a more efficient
way (e.g. reduced computational cost) or provide higher possibilities of reaching the global optimum. Regarding the
FWA, several modified versions exist in the relevant literature [13-16] and will be described afterwards.
The cooperative framework for fireworks algorithm addresses the problems of reduced exploitation ability of
"bad fireworks" and limited efficiency of Gaussian mutation operator [13]. Furthermore, the opposition-based
adaptive fireworks algorithm aims to enhance the performance of adaptive fireworks algorithm using opposition-
based learning, a technique employed to aid several other optimization techniques [14]. The adaptive mutation
dynamic search fireworks algorithm [15] is also an effective variant of fireworks algorithm which was shown to
outperform the simpler adaptive and dynamic search FWA. Finally, the GPU-based parallel fireworks algorithm
[16] constitutes an extremely fast implementation of the FWA which produces also very satisfactory solutions as
well.

3. Methodology

3.1. Description of the proposed approach

As it was aforementioned, the determination of Johnson-Cook material constitutive model parameters for 316L
stainless steel [17] will be conducted by fitting of experimental data using Fireworks algorithm optimization
method. The Johnson-Cook model can describe the flow stress for a variety of materials and is formulated as
follows:
110
118 Nikolaos
N.E.Karkalos E.Markopoulos
and A.P. Karkalos et al. / Procedia
/ Procedia Manufacturing
Manufacturing 00 22 (2018)
(2018) 107–113
115–121

( )
σ = [ Α + Β ( ε p ) ][1 + C ln ε*p ][1 − T * ( )
n m
] (1)

ε p
where ε p is the equivalent plastic strain, ε *p is the normalized strain rate, defined as , which is the ratio of the
ε0
T − TR
plastic strain-rate ε p and the reference strain rate ε0 , Τ* the normalized temperature, defined as with T R
TM − TR
representing the reference temperature and T M the melting temperature. Constants A, B, C, n and m are the model
parameters which represent initial yield stress at the reference strain rate and reference temperature, hardening
modulus, strain rate dependency coefficient, work-hardening exponent and thermal softening component,
respectively. So, the first term of Eq. 1 represents the influence of plastic strain (strain hardening), the second term
represents the influence of strain rate (viscous behavior) and the third term represents the influence of temperature
change (thermal softening).
At first, data concerning stress-strain curves of these materials at various temperatures and strain rate values are
collected and stored. Then, the optimization algorithm is employed to fit these data to Johnson-Cook material model
by determining the appropriate parameters, namely A, B, C, n and m (see eq. 1). This is achieved by defining the
objective function as follows:

n
min ∑ ( yi − f ( xi )) 2 (2)
i =1

where, y i represents the experimental data vector (stress values) and f(x) represents the material model results
vector (Johnson-Cook).This function represents the sum of root squared errors between experimental data and data
predicted by Johnson-Cook model. It is evident that, when the minimization of this function is achieved, the
appropriate parameters that enable Johnson-Cook model to model the experimental data adequately, will be
determined. As Fireworks algorithm is a stochastic optimization algorithm, the range within the search for optimum
values for each input parameter will be conducted is required to be defined. For that reason, the ranges of values for
each parameter are defined, in accordance to values for similar materials in the literature.

3.2. Material description and properties

AISI 316L stainless steel is a variety of the regular 316 grade stainless steel which exhibits low-carbon content
(no more than 0.03% compared to 0.08% for AISI 316 steel). The general composition for AISI 316L stainless steel
is presented in Table 1. This steel is one of the most important austenitic stainless steels is preferred for medical
applications such as various medical devices and is generally more expensive than many stainless steel types.

Table 1. Typical composition of AISI 316L stainless steel.

Cr (%) Ni (%) Mo (%) C (%) Mn (%) Si (%) P (%) S (%) N (%) Fe (%)
16-18 10-14 2-3 <0.03 2 0.75 0.045 0.03 0.1 Bal

In general, stainless steels are highly resistant to corrosion due to the low carbon content in conjunction with the
high chromium content. In particular, the low carbon content of 316L steel renders it favorable for welding
applications as the occurrence of grain boundary carbide precipitation can be avoided. Furthermore, the slightly
higher Mo and Ni content render it more suitable for marine environments, as it is resistant to corrosion in
aggressive chloride environment (such as NaCl) and it also suitable for areas with low temperatures as it can retain
its toughness at low temperature. Moreover these two elements are responsible for the austenitic (FCC) structure of
AISI 316L steel.
 Nikolaos E. Karkalos et al. / Procedia Manufacturing 22 (2018) 107–113 111
N.E Karkalos and A.P. Markopoulos / Procedia Manufacturing 00 (2018) 115–121 119

Regarding the constitutive material model parameters for AISI 316L, there exist several values in the relevant
literature, which were derived from different tests. In Table 2, some of these values for Johnson-Cook model
parameters are summarized. In this work, the experimental data used for the derivation of material parameters are
obtained from [17], which includes experimental curves for strain rate values of 1000, 3000 and 5000 s-1 and
temperatures of 25, 200, 400 and 800oC.

Table 2. Johnson-Cook model parameters from various researchers [18].

A B C N m 𝜀𝜀̇0
305 1161 0.010 0.610 0.517 1
305 441 0.057 0.100 1.041 1
301 1472 0.090 0.807 0.623 10-3
280 1750 0.100 0.800 0.850 2*10-2
514 514 0.042 0.508 0.533 10-3

4. Results and discussion

4.1. Optimization algorithm parameters

Before presenting the results concerning the determination of material model parameters, it is considered
important to briefly present some details on the numerical parameters of the FWA employed in the present work. In
order to obtain the best possible performance using the FWA, a short investigation of the optimum values of FWA
was conducted. Furthermore, it is important to note that the definition of lower and upper limit values for each of the
five constants should be chosen with special attention. These limits are defined according to common values used in
the literature and are subjected to constraints in order to have sense e.g. constants A and B should not be negative
and C, n, m are usually derived in the range 0-1.5. Finally, the FWA is run 20 times for each case for a maximum of
20000 generations in order to ensure that the random choice of initial candidate solutions does not affect
significantly the determination of the optimal solution.

4.2. Results for the case of AISI 316L stainless steel

After the test cases are completed, the results concerning AISI 316L steel are presented. The set of parameters
which gave the minimum value of fitness function was: (A, B, n, C, m) = (310.8, 881.38, 0.178, 0.19, 1.25). In
comparison to the data presented in the Table 2, it can be observed that parameters A,B and n lie within the range of
values predicted by other researchers, whereas values for parameters C and m are slightly higher than the values in
the literature. As can be seen in Fig.1 (a) to (d), the predicted curves are generally close to the experimental points,
indicating a sufficient level of fit.
In general, the predicted curves for 25, 200 and 400oC are more close to the experimental points, whereas the
predicted curves at 800oC exhibit lower level of prediction accuracy, underestimating the actual stress values and
thus predicting higher degree of thermal softening. Moreover, although the effect of strain rate is predicted
sufficiently for many cases, the highest deviations occur for the lowest value of strain rate and especially at cases
with 400oC and 800oC. This can be attributed to the fact that, the curves for 1000 s-1 strain rate exhibit higher change
of curvature than the curves at higher strain rate and so they are more difficult to be predicted.
In order to further analyze the predicted results, RMSE (Root Mean Square Error) and MAPE (Mean Absolute
Percentage Error) values were computed for each of the 12 different curves in respect to the experimental ones. As
can be seen in Table 3, RMSE values vary from 10.542 to 91.112 whereas MAPE values vary from 0.762% to 18%.
These values are in accordance with the previous observations from the comparison between experimental and
predicted curves; higher RMSE and MAPE values are computed mainly for cases with low strain rate value or high
temperature value.
112
120 N.E.KarkalosNikolaos
and A.P.E.Markopoulos
Karkalos et /al. / Procedia
Procedia Manufacturing
Manufacturing 22 (2018)
00 (2018) 107–113
115–121

a b

c d

Fig.1. Comparison of predicted and experimental curves: (a) at 25oC; (b) at 200oC; (c) at 400oC ; (d) at 800oC.

These larger errors for the curves at 800oC could be attributed to the fact that the parameters of Johnson-Cook
model are not correlated with each other and thus, for example the effect of strain rate is the same at every
temperature while the deviations of the predicted curves from the experimental data, especially at 800oC, indicate
that it cannot be exactly the same. Also, more experimental curves within the range 400-800oC are needed in order to
establish a better correlation and derive a more realistic value for parameters C and m.

Table 3. RMSE and MAPE values for all predicted curves.

Cases RMSE (min-max) RMSE average MAPE (min-max) MAPE average
o
25 C 10.542-64.625 39.510 0.752%-4.595% 2.710%
200oC 40.624-74.947 59.140 3.716%-6.421% 4.734%
o
400 C 24.813-60.071 43.510 1.544%-6.616% 3.813%
800oC 49.068-91.112 69.148 5.371%-17.332% 10.189%

In conclusion, the results prove that a decent level of correlation between experimental and predicted data was
achieved but also several improvements should be carried out in order to determine material behavior using the
proposed approach.
 Nikolaos E. Karkalos et al. / Procedia Manufacturing 22 (2018) 107–113 113
N.E Karkalos and A.P. Markopoulos / Procedia Manufacturing 00 (2018) 115–121 121

5. Conclusions

In the present work, a promising optimization algorithm, the Fireworks algorithm, was employed for the
determination of constitutive material model parameters for AISI 316L stainless steel. Experimental data for
temperature values in the range of 25-800oC and strain rates in the range 1000-5000 s-1 were used for the fitting
process. The constitutive model used was Johnson-Cook model and the optimization method was employed in order
to fit the experimental data to this model by deriving the suitable model parameters. From this preliminary work
several conclusions were drawn:
• The determination of material model constants is able to be sufficiently conducted using the proposed approach
by minimizing the difference between predicted and experimental results. However, optimization algorithm
parameters and bounds for optimization variables should be carefully chosen in order to avoid premature
convergence to a solution far from the optimal point.
• From the obtained results it was found that an acceptable level of fit was attained for the experimental data range.
Nevertheless, lower level of prediction accuracy was observed for cases at 800oC and some cases at 5000 s-1
strain rate where the flow stress was underestimated. Similarly, larger deviations from the experimental results
occurred for curves with higher curvature.
• RMSE and MAPE values confirm the observations regarding the experimental and predicted curves; cases at
800oC and cases at lower strain rate generally exhibited higher levels of error.
• Based on the results of the preliminary study, it is proposed that, for the amelioration of predicted results several
modifications should be carried out. For example, a comparison of the performance of newer versions of the
FWA could be conducted and results from a wider range of temperatures and strain rates or even results from
machining simulations could be incorporated to the model in order to increase the level of correlation attained
and predict material model parameters more close to the actual ones.

References

[1] A. Dorogoy, D. Rittel, Determination of the Johnson–Cook material parameters using the SCS specimen, Exp. Mech. 49 (2009) 881-885.
[2] A.K. Gupta, S.K. Singh, S. Reddy, G. Hariharan, Prediction of flow stress in dynamic strain aging regime of austenitic stainless steel 316
using artificial neural network, Mater. Des. 35 (2012) 589-595.
[3] L.X. Kong, P.D. Hodgson, D.C. Collinson, Extrapolative prediction of the hot strength of austenitic steels with a combined constitutive and
ANN model, J. Mater. Process. Technol. 102 (1-3) (2000) 84-89.
[4] A. Dimatteo, M. Vannucci, V. Colla, Prediction of mean flow stress during hot strip rolling using genetic algorithms. ISIJ International 54
(1) (2014) 171-178.
[5] T. Özel, Y. Karpat, Mater. Identification of constitutive material model parameters for high-strain rate metal cutting conditions using
evolutionary computational algorithms. Manuf. Process. 22 (5) (2007) 659-667.
[6] J. Lin, J. Yang, GA-based multiple objective optimisation for determining viscoplastic constitutive equations for superplastic alloys. Int. J.
Plasticity 15 (1999) 1181-1196.
[7] Z. Huang, L. Gao, Y. Weng, F. Wang, J. Mater. Determination of the Johnson-Cook Constitutive Model Parameters of Materials by Cluster
Global Optimization Algorithm. Eng. Perform. 25 (2016) 4099-4107.
[8] G. Chen, C. Ren, W. Yu, X. Yang, L. Zhang, Application of genetic algorithms for optimizing the Johnson–Cook constitutive model
parameters when simulating the titanium alloy Ti-6Al-4V machining process. P. I. Mech. Eng. B. J. Eng. Man. 226 (8) (2012) 1287-1297.
[9] A. Shrot, M. Bäker, Determination of Johnson–Cook parameters from machining simulations. Comput. Mater. Sci. 52 (1) (2012) 298-304.
[10] M. Bäker, A new method to determine material parameters from machining simulations using inverse identification. Procedia CIRP 31
(2015) 399-404.
[11] Y. Tan, Y. Zhu, in: Y.Tan, Y.Shi, K.C. Tan (Eds.), Fireworks algorithm for optimization. Advances in Swarm Intelligence, Springer Verlag,
Berlin Heidelberg, 2010, pp. 355-364.
[12] Y. Tan, Fireworks Algorithm, A Novel Swarm Intelligence Optimization Method, Springer Verlag, Berlin Heidelberg, 2015.
[13] S. Zheng, J. Li, A. Janecek, Y. Tan, A cooperative framework for fireworks algorithm. IEEE ACM T. Comput. Bi. 14 (1) (2017) 27-41.
[14] C. Gong, Opposition-Based Adaptive Fireworks Algorithm, Algorithms 9(3) (2016) 43
[15] X.-G. Li, S.-F. Han, L. Zhao, C.-Q. Gong, X.-J. Liu, Adaptive Mutation Dynamic Search Fireworks Algorithm. Algorithms 10(2) (2017) 48
[16] K. Ding, S. Zheng, Y. Tan, GECCO '13, A gpu-based parallel fireworks algorithm for optimization.Proceedings of the 15th annual
conference on Genetic and evolutionary computation (2013) 9-16.
[17] W.-S. Lee, T.-H. Chen, C.-F. Lin, W.-Z. Luo, Dynamic mechanical response of biomedical 316L stainless steel as function of strain rate and
temperature. Bioinorg. Chem. Appl. (2011):173782.
[18] D. Umbrello, R. M’Saoubi, J.C. Outeiro, Int. J. Mach. Tool Manu. 47 (3-4) (2007) 462-470.