Nuclear Research and Consultancy Group, NRG, … / 19 Uncertainty evaluations and validations D. Rochman and A.J. Koning Nuclear Research and Consultancy Group, NRG, Petten, The Netherlands

1 / 19

Uncertainty evaluations and validations

D. Rochman and A.J. Koning

Nuclear Research and Consultancy Group,

NRG, Petten, The Netherlands

May 22, 2008

Contents

2 / 19

① Motivations for a change:=⇒ a roadmap to ban covariance files

② Concept:=⇒ Monte Carlo from nuclear data to large-scale systems

③ Where can we apply it ?=⇒ (needed tools & knowledge)

④ How does it work ?

⑤ Examples with Pb isotopes:=⇒ keff benchmarks and reactors

⑥ Examples on global scale:=⇒ keff benchmarks, fusion shielding, reactivity swing

⑦ Pros, Cons and Conclusions

Introduction: Motivations for a change

3 / 19

Usual procedures in uncertainty propagation imply

☞ rigid format,

☞ need for fixed libraries of cross section values,

☞ need for processing, sensitivity and perturbation codes,

☞ group scheme,

☞ simplification of covariance matrix (restricted correlation),

☞ necessity of linearizing inherently nonlinear relationships,

☞ and so on. . .

➸ Most of these routines were developed decades ago when the supportfor nuclear data and nuclear reactor physics research was sufficient to

allow them to be produced !

Introduction: Motivations for a change

3 / 19

Usual procedures in uncertainty propagation imply

☞ rigid format,

☞ need for fixed libraries of cross section values,

☞ need for processing, sensitivity and perturbation codes,

☞ group scheme,

☞ simplification of covariance matrix (restricted correlation),

☞ necessity of linearizing inherently nonlinear relationships,

☞ and so on. . .

➸ Most of these routines were developed decades ago when the supportfor nuclear data and nuclear reactor physics research was sufficient to

allow them to be produced !

Introduction: Motivations for a change

3 / 19

Usual procedures in uncertainty propagation imply

☞ rigid format,

☞ need for fixed libraries of cross section values,

☞ need for processing, sensitivity and perturbation codes,

☞ group scheme,

☞ simplification of covariance matrix (restricted correlation),

☞ necessity of linearizing inherently nonlinear relationships,

☞ and so on. . .

➸ Most of these routines were developed decades ago when the supportfor nuclear data and nuclear reactor physics research was sufficient to

allow them to be produced !

Introduction: Motivations for a change

3 / 19

Usual procedures in uncertainty propagation imply

☞ rigid format,

☞ need for fixed libraries of cross section values,

☞ need for processing, sensitivity and perturbation codes,

☞ group scheme,

☞ simplification of covariance matrix (restricted correlation),

☞ necessity of linearizing inherently nonlinear relationships,

☞ and so on. . .

➸ Most of these routines were developed decades ago when the supportfor nuclear data and nuclear reactor physics research was sufficient to

allow them to be produced !

Introduction: Motivations for a change

3 / 19

Usual procedures in uncertainty propagation imply

☞ rigid format,

☞ need for fixed libraries of cross section values,

☞ need for processing, sensitivity and perturbation codes,

☞ group scheme,

☞ simplification of covariance matrix (restricted correlation),

☞ necessity of linearizing inherently nonlinear relationships,

☞ and so on. . .

➸ Most of these routines were developed decades ago when the supportfor nuclear data and nuclear reactor physics research was sufficient to

allow them to be produced !

Introduction: Motivations for a change

3 / 19

Usual procedures in uncertainty propagation imply

☞ rigid format,

☞ need for fixed libraries of cross section values,

☞ need for processing, sensitivity and perturbation codes,

☞ group scheme,

☞ simplification of covariance matrix (restricted correlation),

☞ necessity of linearizing inherently nonlinear relationships,

☞ and so on. . .

➸ Most of these routines were developed decades ago when the supportfor nuclear data and nuclear reactor physics research was sufficient to

allow them to be produced !

Introduction: Motivations for a change

3 / 19

Usual procedures in uncertainty propagation imply

☞ rigid format,

☞ need for fixed libraries of cross section values,

☞ need for processing, sensitivity and perturbation codes,

☞ group scheme,

☞ simplification of covariance matrix (restricted correlation),

☞ necessity of linearizing inherently nonlinear relationships,

☞ and so on. . .

➸ Most of these routines were developed decades ago when the supportfor nuclear data and nuclear reactor physics research was sufficient to

allow them to be produced !

After all, not a new idea

4 / 19

Here is the mantra in which we believe and motivates us:

“Researchers should cease trying to be clever in devising refinements toold methods that were developed when computational resources were

limited.

Instead, their creative instincts should be redirected to unleashing the fullpotential of computers for brute force analysis”

D. Smith, Santa Fe 2004

=⇒ Most straightforward way: Global Monte Carlo Approach !

After all, not a new idea

4 / 19

Here is the mantra in which we believe and motivates us:

“Researchers should cease trying to be clever in devising refinements toold methods that were developed when computational resources were

limited.Instead, their creative instincts should be redirected to unleashing the full

potential of computers for brute force analysis”

D. Smith, Santa Fe 2004

=⇒ Most straightforward way: Global Monte Carlo Approach !

We are proposing a conceptual revolution

5 / 19

(integral, differential)

Experiments(basic quantities)

Evaluation 1(uncertainties)

Evaluation 2

Formatting

Checking

(MCNP, Perturbation)

Benchmarks(NJOY)

Processing Library

�

�

� � � �

� �

� �

(basic quantities + uncertainties)

Evaluation

We are proposing a conceptual revolution

5 / 19

(integral, differential)

Experiments(basic quantities)

Evaluation 1(uncertainties)

Evaluation 2

Formatting

Checking

(MCNP, Perturbation)

Benchmarks(NJOY)

Processing Library

�

�

� � � �

� �

� �

�

�

(basic quantities + uncertainties)

Evaluation

We are proposing a conceptual revolution

5 / 19

(integral, differential)

Experiments(basic quantities)

Evaluation 1(uncertainties)

Evaluation 2

Formatting

Checking

(MCNP, Perturbation)

Benchmarks(NJOY)

Processing Library

�

�

� � � �

� �

� �

�

�

(basic quantities + uncertainties)

Evaluation

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)=⇒ Deterministic codes (APOLLO, WIMS...)=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)=⇒ Deterministic codes (APOLLO, WIMS...)=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)

=⇒ Deterministic codes (APOLLO, WIMS...)=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)=⇒ Deterministic codes (APOLLO, WIMS...)

=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)=⇒ Deterministic codes (APOLLO, WIMS...)=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)=⇒ Deterministic codes (APOLLO, WIMS...)=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)=⇒ Deterministic codes (APOLLO, WIMS...)=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)=⇒ Deterministic codes (APOLLO, WIMS...)=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Fields of application & Needs

6 / 19

Where can it be applied ?

✑ Everywhere where there are nuclear data (not hardcoded)

=⇒ Monte Carlo codes (MCNP, Tripoli...)=⇒ Deterministic codes (APOLLO, WIMS...)=⇒ Quantities: criticality, flux ( + all from SG-26), shielding and fusion (withEAF files)

In the following, we will restrict ourselves to elements between 19F and 209Bi (forthe time being):

✑ Stable Nuclear reaction code: TALYS

✑ Monte Carlo transport code: MCNP

✑ Tabulated resonance parameters

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 1

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 1

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 1(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 1(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 2

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 2

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 2(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 2(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 3

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 3

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 3(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 3(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 4

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 4

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 4(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 4(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 5

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 5

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 5(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 5(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 6

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 6

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 6(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 6(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 7

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 7

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 7(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 7(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 8

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 8

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 8(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 8(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 9

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 9

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 9(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 9(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 10

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 10

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 10(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 10(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

n = 15

θn = 61 degEn = 5.2 MeV

(n,xn)

Energy (MeV)

d2σ/d

E/d

θ(b

/eV

/sr)

6543210

10−6

10−7

10−8

10−9

10−10

n = 15

En = 500 keV

(n,el)

Angle (deg)

dσ/d

θ(b

/sr)

150100500

2.4

2.0

1.6

1.2

0.8

0.4

n = 15(n,γ)

Incident Energy (eV)

Cro

ssse

ctio

n(b

)

1000800600400200

10−1

10−2

10−3

10−4

n = 15(n,2n)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

201510

2.0

1.5

1.0

0.5

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 20

keff = 1.01451± 703 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 25

keff = 1.01403± 696 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 30

keff = 1.01395± 678 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 40

keff = 1.01337± 712 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 60

keff = 1.01429± 676 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 100

keff = 1.01358± 693 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 200

keff = 1.01306± 766 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 300

keff = 1.01304± 743 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

10

8

6

4

2

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 400

keff = 1.01355± 769 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

25

20

15

10

5

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 600

keff = 1.01372± 745 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

25

20

15

10

5

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv MCNPTALYS

n = 800

keff = 1.01363± 744 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

25

20

15

10

5

0

Hands on “1000 ×(Talys + ENDF + NJOY + MCNP) calculations”

7 / 19

Γn

acompound

atarget

Γγ

av

rv

=⇒ uncertainty due to nuclear data ' 740 pcm

Statistical uncertainty ' 68 pcm

MCNPTALYS

n = 800

keff = 1.01363± 744 pcm

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.011.000.99

25

20

15

10

5

0

Examples with Pb isotopes

8 / 19

❅ Full evaluations for 204−208Pb (see NIM A589 (2008) 85 for evaluation)

❅ Parameter uncertainty assessment and Monte Carlo maximum likelihoodestimate with “accept-reject” method,

❅ 5000 random ENDF files

❅ Applied on keff and βeff for thermal and fast criticality benchmarks (LCT-10and HMF-64) and to ADS and LFR

204Pb(n,inl)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

20151050

3

2

1

0

208Pb(n,γ)

Incident Energy (keV)

Cro

ssse

ctio

n(b

)

90807065

10−1

10−2

10−3

10−4

10−5

Examples with Pb isotopes

8 / 19

❅ Full evaluations for 204−208Pb (see NIM A589 (2008) 85 for evaluation)

❅ Parameter uncertainty assessment and Monte Carlo maximum likelihoodestimate with “accept-reject” method,

❅ 5000 random ENDF files

❅ Applied on keff and βeff for thermal and fast criticality benchmarks (LCT-10and HMF-64) and to ADS and LFR

204Pb(n,inl)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

20151050

3

2

1

0

208Pb(n,γ)

Incident Energy (keV)

Cro

ssse

ctio

n(b

)

90807065

10−1

10−2

10−3

10−4

10−5

Examples with Pb isotopes

8 / 19

❅ Full evaluations for 204−208Pb (see NIM A589 (2008) 85 for evaluation)

❅ Parameter uncertainty assessment and Monte Carlo maximum likelihoodestimate with “accept-reject” method,

❅ 5000 random ENDF files

❅ Applied on keff and βeff for thermal and fast criticality benchmarks (LCT-10and HMF-64) and to ADS and LFR

204Pb(n,inl)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

20151050

3

2

1

0

208Pb(n,γ)

Incident Energy (keV)

Cro

ssse

ctio

n(b

)

90807065

10−1

10−2

10−3

10−4

10−5

Examples with Pb isotopes

8 / 19

❅ Full evaluations for 204−208Pb (see NIM A589 (2008) 85 for evaluation)

❅ Parameter uncertainty assessment and Monte Carlo maximum likelihoodestimate with “accept-reject” method,

❅ 5000 random ENDF files

❅ Applied on keff and βeff for thermal and fast criticality benchmarks (LCT-10and HMF-64) and to ADS and LFR

204Pb(n,inl)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

20151050

3

2

1

0

208Pb(n,γ)

Incident Energy (keV)

Cro

ssse

ctio

n(b

)

90807065

10−1

10−2

10−3

10−4

10−5

Examples with Pb isotopes

8 / 19

❅ Full evaluations for 204−208Pb (see NIM A589 (2008) 85 for evaluation)

❅ Parameter uncertainty assessment and Monte Carlo maximum likelihoodestimate with “accept-reject” method,

❅ 5000 random ENDF files

❅ Applied on keff and βeff for thermal and fast criticality benchmarks (LCT-10and HMF-64) and to ADS and LFR

204Pb(n,inl)

Incident Energy (MeV)

Cro

ssse

ctio

n(b

)

20151050

3

2

1

0

208Pb(n,γ)

Incident Energy (keV)

Cro

ssse

ctio

n(b

)

90807065

10−1

10−2

10−3

10−4

10−5

Examples with Pb isotopes

9 / 19

LCT10-1Thermal criticality benchmark

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.00

40

30

20

10

0

Lead Fast Reactor

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.00

40

30

20

10

0

keff = 1.01028±(60 pcm and 212 pcm) keff = 1.00894±(60 pcm and 240 pcm)

< keff >

hmf64-1

204−208Pb

Sample number

Update

dk

eff

120010008006004002000

1.019

1.017

1.015

1.013

1.011

< σ >

Sample number

Updat

edunce

rtai

nty

(pcm

)

120010008006004002000

1400

1300

1200

1100

1000

900

800

Examples with Pb isotopes

9 / 19

LCT10-1Thermal criticality benchmark

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.00

40

30

20

10

0

Lead Fast Reactor

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.00

40

30

20

10

0

keff = 1.01028±(60 pcm and 212 pcm) keff = 1.00894±(60 pcm and 240 pcm)

< keff >

hmf64-1

204−208Pb

Sample number

Update

dk

eff

120010008006004002000

1.019

1.017

1.015

1.013

1.011

< σ >

Sample number

Updat

edunce

rtai

nty

(pcm

)

120010008006004002000

1400

1300

1200

1100

1000

900

800

Examples with Pb isotopes

9 / 19

LCT10-1Thermal criticality benchmark

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.00

40

30

20

10

0

Lead Fast Reactor

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.00

40

30

20

10

0

keff = 1.01028±(60 pcm and 212 pcm) keff = 1.00894±(60 pcm and 240 pcm)

< keff >

hmf64-1

204−208Pb

Sample number

Update

dk

eff

120010008006004002000

1.019

1.017

1.015

1.013

1.011

< σ >

Sample number

Updat

edunce

rtai

nty

(pcm

)

120010008006004002000

1400

1300

1200

1100

1000

900

800

Examples with Pb isotopes

10 / 19

HMF64-1Fast criticality benchmark

keff value

Num

ber

ofco

unts

/bin

s

1.051.041.021.010.99

50

40

30

20

10

0

Accelerator Driven System

keff value

Num

ber

ofco

unts

/bin

s

0.980.970.96

50

40

30

20

10

0

Examples with Pb isotopes

10 / 19

HMF64-1Fast criticality benchmark

keff value

Num

ber

ofco

unts

/bin

s

1.051.041.021.010.99

50

40

30

20

10

0

Accelerator Driven System

keff value

Num

ber

ofco

unts

/bin

s

0.980.970.96

50

40

30

20

10

0

Better fit with the “Extreme Value Theory”, or EVT:F(z) = e−z−e−z with z = X−µ

σMean µ′ = µ+ γσ

Standard Deviation σ′ = σ π√6

Examples with Pb isotopes

10 / 19

HMF64-1Fast criticality benchmark

keff value

Num

ber

ofco

unts

/bin

s

1.051.041.021.010.99

50

40

30

20

10

0

Accelerator Driven System

keff value

Num

ber

ofco

unts

/bin

s

0.980.970.96

50

40

30

20

10

0

Better fit with the “Extreme Value Theory”, or EVT:F(z) = e−z−e−z with z = X−µ

σMean µ′ = µ+ γσ

Standard Deviation σ′ = σ π√6

HMF-64.1 ADSkeff 1.00848 0.96648

µ′=1.01394 µ′=0.96785σk ×105 855 291

σ′=1097 σ′=345

Why not a Normal distribution for keff ?

11 / 19

(1) The central limit theorem does not apply

(2) Cross sections do not follow a Gaussian distribution:

En = 3MeV

206Pb(n,γ)

Cross Section (mb)

Cou

nts

/chan

nel

6420

50

40

30

20

10

0

Cou

nts

/Chan

nel

20

10

0

Cross section (b)

3.0 2.6 2.2 1.8 1.4

Ener

gy(M

eV)4.0

3.5

3.0

2.5

206Pb(n,inl)

Any safety related issue regarding high keff tail ?

Why not a Normal distribution for keff ?

11 / 19

(1) The central limit theorem does not apply(2) Cross sections do not follow a Gaussian distribution:

En = 3MeV

206Pb(n,γ)

Cross Section (mb)

Cou

nts

/chan

nel

6420

50

40

30

20

10

0

Cou

nts

/Chan

nel

20

10

0

Cross section (b)

3.0 2.6 2.2 1.8 1.4

Ener

gy(M

eV)4.0

3.5

3.0

2.5

206Pb(n,inl)

Any safety related issue regarding high keff tail ?

Global calculations: from 19F to 208Pb

12 / 19

“In general, this paper will or will not be a breakthrough in methodology if the[practicality and robustness] can or can not be demonstrated.”,

ANE Referee, May 2008

Okay, let’s go from academic solutions to mass production !

® Default TALYS calculation + Resonance parameters (RP)

® Default TALYS uncertainties + RP uncertainties derived from the Atlas

® 100 to 2000 ENDF files per isotope from 19F to 208Pb (100 isotopes)

® 150 criticality-safety benchmarks calculated (> 30 000 calculations) fromS. van der Mark’s list

® All Oktavian shielding benchmarks (neutrons and gammas)

® Reactivity swing for a LWR using an “Inert Matrix Fuel” (Pu and Mo),Westinghouse 3 loops type reactor

Global calculations: from 19F to 208Pb

12 / 19

“In general, this paper will or will not be a breakthrough in methodology if the[practicality and robustness] can or can not be demonstrated.”,

ANE Referee, May 2008

Okay, let’s go from academic solutions to mass production !

® Default TALYS calculation + Resonance parameters (RP)

® Default TALYS uncertainties + RP uncertainties derived from the Atlas

® 100 to 2000 ENDF files per isotope from 19F to 208Pb (100 isotopes)

® 150 criticality-safety benchmarks calculated (> 30 000 calculations) fromS. van der Mark’s list

® All Oktavian shielding benchmarks (neutrons and gammas)

® Reactivity swing for a LWR using an “Inert Matrix Fuel” (Pu and Mo),Westinghouse 3 loops type reactor

Global calculations: from 19F to 208Pb

12 / 19

“In general, this paper will or will not be a breakthrough in methodology if the[practicality and robustness] can or can not be demonstrated.”,

ANE Referee, May 2008

Okay, let’s go from academic solutions to mass production !

® Default TALYS calculation + Resonance parameters (RP)

® Default TALYS uncertainties + RP uncertainties derived from the Atlas

® 100 to 2000 ENDF files per isotope from 19F to 208Pb (100 isotopes)

® 150 criticality-safety benchmarks calculated (> 30 000 calculations) fromS. van der Mark’s list

® All Oktavian shielding benchmarks (neutrons and gammas)

® Reactivity swing for a LWR using an “Inert Matrix Fuel” (Pu and Mo),Westinghouse 3 loops type reactor

Global calculations: from 19F to 208Pb

12 / 19

“In general, this paper will or will not be a breakthrough in methodology if the[practicality and robustness] can or can not be demonstrated.”,

ANE Referee, May 2008

Okay, let’s go from academic solutions to mass production !

® Default TALYS calculation + Resonance parameters (RP)

® Default TALYS uncertainties + RP uncertainties derived from the Atlas

® 100 to 2000 ENDF files per isotope from 19F to 208Pb (100 isotopes)

® 150 criticality-safety benchmarks calculated (> 30 000 calculations) fromS. van der Mark’s list

® All Oktavian shielding benchmarks (neutrons and gammas)

® Reactivity swing for a LWR using an “Inert Matrix Fuel” (Pu and Mo),Westinghouse 3 loops type reactor

Global calculations: from 19F to 208Pb

12 / 19

“In general, this paper will or will not be a breakthrough in methodology if the[practicality and robustness] can or can not be demonstrated.”,

ANE Referee, May 2008

Okay, let’s go from academic solutions to mass production !

® Default TALYS calculation + Resonance parameters (RP)

® Default TALYS uncertainties + RP uncertainties derived from the Atlas

® 100 to 2000 ENDF files per isotope from 19F to 208Pb (100 isotopes)

® 150 criticality-safety benchmarks calculated (> 30 000 calculations) fromS. van der Mark’s list

® All Oktavian shielding benchmarks (neutrons and gammas)

® Reactivity swing for a LWR using an “Inert Matrix Fuel” (Pu and Mo),Westinghouse 3 loops type reactor

Global calculations: from 19F to 208Pb

12 / 19

“In general, this paper will or will not be a breakthrough in methodology if the[practicality and robustness] can or can not be demonstrated.”,

ANE Referee, May 2008

Okay, let’s go from academic solutions to mass production !

® Default TALYS calculation + Resonance parameters (RP)

® Default TALYS uncertainties + RP uncertainties derived from the Atlas

® 100 to 2000 ENDF files per isotope from 19F to 208Pb (100 isotopes)

® 150 criticality-safety benchmarks calculated (> 30 000 calculations) fromS. van der Mark’s list

® All Oktavian shielding benchmarks (neutrons and gammas)

® Reactivity swing for a LWR using an “Inert Matrix Fuel” (Pu and Mo),Westinghouse 3 loops type reactor

Global calculations: from 19F to 208Pb

12 / 19

“In general, this paper will or will not be a breakthrough in methodology if the[practicality and robustness] can or can not be demonstrated.”,

ANE Referee, May 2008

Okay, let’s go from academic solutions to mass production !

® Default TALYS calculation + Resonance parameters (RP)

® Default TALYS uncertainties + RP uncertainties derived from the Atlas

® 100 to 2000 ENDF files per isotope from 19F to 208Pb (100 isotopes)

® 150 criticality-safety benchmarks calculated (> 30 000 calculations) fromS. van der Mark’s list

® All Oktavian shielding benchmarks (neutrons and gammas)

® Reactivity swing for a LWR using an “Inert Matrix Fuel” (Pu and Mo),Westinghouse 3 loops type reactor

Examples of keff benchmarks for 19F

13 / 19

keff = 1.01735 ± 939 pcmhmf7-34 (19F)

keff value

Num

ber

ofco

unts

/bin

s

1.051.041.031.021.011.000.990.98

40

30

20

10

0

keff = 1.00270 ± 1562 pcm

hmf5-1 (92−100Mo)

keff value

Num

ber

ofco

unts

/bin

s

1.061.041.021.000.980.960.94

30

20

10

0

keff = 1.01295 ± 144 pcmlst1 (19F)

keff value

Num

ber

ofco

unts

/bin

s

1.0201.0151.0101.005

80

70

60

50

40

30

20

10

0

keff = 1.00383 ± 470 pcmhmf7-32 (19F)

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.000.99

80706050403020100

keff = 1.03929 ± 1062 pcm

hst39-1 (19F)

keff value

Num

ber

ofco

unts

/bin

s

1.081.071.051.041.021.01

40

30

20

10

0

keff = 1.01347± 733 pcmhmf7-33 (19F)

keff value

Num

ber

ofco

unts

/bin

s

1.051.041.031.021.011.000.99

70

60

50

40

30

20

10

0

Examples of keff benchmarks for 27Al and 92−100Mo

14 / 19

keff = 1.01469 ± 155 pcmhmt8d (27Al)

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.00

25

20

15

10

5

0

keff = 1.02529 ± 161 pcmhmt8s (27Al)

keff value

Num

ber

ofco

unts

/bin

s

1.041.031.021.01

20

10

0

keff = 1.00432 ± 198 pcmhmt22 (27Al)

keff value

Num

ber

ofco

unts

/bin

s

1.0151.0101.0051.0000.995

25

20

15

10

5

0

keff = 1.01846± 68 pcmhmi1 (27Al)

keff value

Num

ber

ofco

unts

/bin

s

1.0251.0201.015

25

20

15

10

5

0

keff = 1.00469 ± 853 pcmhmf5-2 (92−100Mo)

keff value

Num

ber

ofco

unts

/bin

s

1.041.021.010.990.98

30

20

10

0

keff = 1.00186 ± 755 pcmhmf5-5 (92−100Mo)

keff value

Num

ber

ofco

unts

/bin

s

1.031.021.011.000.990.980.97

30

20

10

0

Examples of keff benchmarks for 59Co and 90−96Zr

15 / 19

keff = 0.99650± 158 pcmmmf11-2 (90−96Zr)

keff value

Num

ber

ofco

unts

/bin

s

1.0051.0000.9950.9900.985

25

20

15

10

5

0

keff = 0.96338 ± 3777 pcmhci5-5 (90−96Zr)

keff value

Num

ber

ofco

unts

/bin

s

1.051.000.950.90

10

5

0

keff = 1.01044 ± 326 pcmict3-133 (90−96Zr)

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.000.99

20

10

0

keff = 1.00571 ± 322 pcmict3-132 (90−96Zr)

keff value

Num

ber

ofco

unts

/bin

s

1.021.011.000.99

20

10

0

keff = 0.99977 ± 84 pcmlct51-4 (59Co)

keff value

Num

ber

ofco

unts

/bin

s

1.0051.0000.995

60

50

40

30

20

10

0

keff = 1.00026 ± 89 pcmlct51-8 (59Co)

keff value

Num

ber

ofco

unts

/bin

s

1.0051.0000.995

60

50

40

30

20

10

0

Examples of keff benchmarks for 54−58Fe

16 / 19

keff = 0.99129 ± 59 pcmimf10-1 (54−58Fe)

keff value

Num

ber

ofco

unts

/bin

s

0.9960.9930.9900.987

30

20

10

0

keff = 1.01140 ± 123 pcm

pmf12 (54−58Fe)

keff value

Num

ber

ofco

unts

/bin

s

1.0201.0151.0101.005

20

15

10

5

0

keff = 1.01066± 475 pcmhmt13-625 (54−58Fe)

keff value

Num

ber

ofco

unts

/bin

s

1.031.021.011.00

20

15

10

5

0

keff = 1.00848± 137 pcmhmt13-15 (54−58Fe)

keff value

Num

ber

ofco

unts

/bin

s

1.0151.0101.005

30

20

10

0

keff = 1.04501 ± 1852 pcm

hmi1-1 (54−58Fe)

keff value

Num

ber

ofco

unts

/bin

s

1.101.081.061.041.021.000.98

20

10

0

keff = 0.99813± 82 pcmlmt1 (54−58Fe)

keff value

Num

ber

ofco

unts

/bin

s

1.0051.0000.995

25

20

15

10

5

0

Examples of shielding benchmarks and reactivity swing

17 / 19

ExperimentCalculation

Oktavian Cr Neutron Leakage

Emission Neutron Energy (MeV)

Lea

kage

curr

ent/

leth

argy

per

sourc

epar

ticl

e

10521

1.0

0.5

0.2

0.1

1σ Mo data uncertainty

12 % Pu, 88 % Mo (volume)Westinghouse 3 loops type reactor

LWR Inert Matrix Fuel

δT ' 35 efpd

Reactivity Swing

Time (efpd)

keff

2000150010005000

1.4

1.3

1.2

1.1

1.0

0.9

0.8

(Blind Taly calculations)

E Also applied to Mn, Co, Al, Cu Oktavian benchmarksE and industrial PWR reactor for life-time extension (uncertainty on the reactorpressure vessel damage)

Examples of shielding benchmarks and reactivity swing

17 / 19

ExperimentCalculation

Oktavian Cr Neutron Leakage

Emission Neutron Energy (MeV)

Lea

kage

curr

ent/

leth

argy

per

sourc

epar

ticl

e

10521

1.0

0.5

0.2

0.1

1σ Mo data uncertainty

12 % Pu, 88 % Mo (volume)Westinghouse 3 loops type reactor

LWR Inert Matrix Fuel

δT ' 35 efpd

Reactivity Swing

Time (efpd)

keff

2000150010005000

1.4

1.3

1.2

1.1

1.0

0.9

0.8

(Blind Taly calculations)

E Also applied to Mn, Co, Al, Cu Oktavian benchmarks

E and industrial PWR reactor for life-time extension (uncertainty on the reactorpressure vessel damage)

Examples of shielding benchmarks and reactivity swing

17 / 19

ExperimentCalculation

Oktavian Cr Neutron Leakage

Emission Neutron Energy (MeV)

Lea

kage

curr

ent/

leth

argy

per

sourc

epar

ticl

e

10521

1.0

0.5

0.2

0.1

1σ Mo data uncertainty

12 % Pu, 88 % Mo (volume)Westinghouse 3 loops type reactor

LWR Inert Matrix Fuel

δT ' 35 efpd

Reactivity Swing

Time (efpd)

keff

2000150010005000

1.4

1.3

1.2

1.1

1.0

0.9

0.8

(Blind Taly calculations)

E Also applied to Mn, Co, Al, Cu Oktavian benchmarksE and industrial PWR reactor for life-time extension (uncertainty on the reactorpressure vessel damage)

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA

§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Pros and Cons

18 / 19

© + No MF 32-35 (no 2 Gb files) but every possible cross correlation included

© + No approximation but true probability distribution

© + Only essential info for an evaluation is stored

© + No perturbation code necessary, only “essential” codes

© + Feedback to model parameters

© + (Random) EAF libraries

© + QA§ - Needs discipline to reproduce

§ - Memory and computer time

§ - Complexity for full reactor core calculation unknown

§ - Role of data centers would change

Conclusions and future improvements

19 / 19

X New methodology to propagate nuclear data uncertainty to integralquantities (keff benchmarks, shielding benchmarks, reactivity swing,neutron flux for commercial reactor) via Monte Carlo

X Proof of principle with high quality Pb evaluations

X Mass production tested on more than 150 benchmarksX “Paper to present the principle”: under revision for ANE� Needs for better sampling in the resonance region

� Needs for a better “accept-reject” mechanism

� What if nuclear modeling does not match the accuracy of themeasurements ? (how to sample ?)

� Needs to develop best central-value evaluations (non-fissile and fissile) ?

Conclusions and future improvements

19 / 19

X New methodology to propagate nuclear data uncertainty to integralquantities (keff benchmarks, shielding benchmarks, reactivity swing,neutron flux for commercial reactor) via Monte Carlo

X Proof of principle with high quality Pb evaluations

X Mass production tested on more than 150 benchmarksX “Paper to present the principle”: under revision for ANE� Needs for better sampling in the resonance region

� Needs for a better “accept-reject” mechanism

� What if nuclear modeling does not match the accuracy of themeasurements ? (how to sample ?)

� Needs to develop best central-value evaluations (non-fissile and fissile) ?

Conclusions and future improvements

19 / 19

X New methodology to propagate nuclear data uncertainty to integralquantities (keff benchmarks, shielding benchmarks, reactivity swing,neutron flux for commercial reactor) via Monte Carlo

X Proof of principle with high quality Pb evaluations

X Mass production tested on more than 150 benchmarks

X “Paper to present the principle”: under revision for ANE� Needs for better sampling in the resonance region

� Needs for a better “accept-reject” mechanism

� What if nuclear modeling does not match the accuracy of themeasurements ? (how to sample ?)

� Needs to develop best central-value evaluations (non-fissile and fissile) ?

Conclusions and future improvements

19 / 19

X New methodology to propagate nuclear data uncertainty to integralquantities (keff benchmarks, shielding benchmarks, reactivity swing,neutron flux for commercial reactor) via Monte Carlo

X Proof of principle with high quality Pb evaluations

X Mass production tested on more than 150 benchmarksX “Paper to present the principle”: under revision for ANE

� Needs for better sampling in the resonance region

� Needs for a better “accept-reject” mechanism

� What if nuclear modeling does not match the accuracy of themeasurements ? (how to sample ?)

� Needs to develop best central-value evaluations (non-fissile and fissile) ?

Conclusions and future improvements

19 / 19

X New methodology to propagate nuclear data uncertainty to integralquantities (keff benchmarks, shielding benchmarks, reactivity swing,neutron flux for commercial reactor) via Monte Carlo

X Proof of principle with high quality Pb evaluations

X Mass production tested on more than 150 benchmarksX “Paper to present the principle”: under revision for ANE� Needs for better sampling in the resonance region

� Needs for a better “accept-reject” mechanism

� What if nuclear modeling does not match the accuracy of themeasurements ? (how to sample ?)

� Needs to develop best central-value evaluations (non-fissile and fissile) ?

Conclusions and future improvements

19 / 19

X New methodology to propagate nuclear data uncertainty to integralquantities (keff benchmarks, shielding benchmarks, reactivity swing,neutron flux for commercial reactor) via Monte Carlo

X Proof of principle with high quality Pb evaluations

X Mass production tested on more than 150 benchmarksX “Paper to present the principle”: under revision for ANE� Needs for better sampling in the resonance region

� Needs for a better “accept-reject” mechanism

� What if nuclear modeling does not match the accuracy of themeasurements ? (how to sample ?)

� Needs to develop best central-value evaluations (non-fissile and fissile) ?

Conclusions and future improvements

19 / 19

X New methodology to propagate nuclear data uncertainty to integralquantities (keff benchmarks, shielding benchmarks, reactivity swing,neutron flux for commercial reactor) via Monte Carlo

X Proof of principle with high quality Pb evaluations

X Mass production tested on more than 150 benchmarksX “Paper to present the principle”: under revision for ANE� Needs for better sampling in the resonance region

� Needs for a better “accept-reject” mechanism

� What if nuclear modeling does not match the accuracy of themeasurements ? (how to sample ?)

� Needs to develop best central-value evaluations (non-fissile and fissile) ?

Conclusions and future improvements

19 / 19

X New methodology to propagate nuclear data uncertainty to integralquantities (keff benchmarks, shielding benchmarks, reactivity swing,neutron flux for commercial reactor) via Monte Carlo

X Proof of principle with high quality Pb evaluations

X Mass production tested on more than 150 benchmarksX “Paper to present the principle”: under revision for ANE� Needs for better sampling in the resonance region

� Needs for a better “accept-reject” mechanism

� What if nuclear modeling does not match the accuracy of themeasurements ? (how to sample ?)

� Needs to develop best central-value evaluations (non-fissile and fissile) ?