Hydrogen Embrittlement of Low Alloy Steel forPressure Vessels Based on a Large SizeSpecimen

著者 和田 洋流号 52学位授与番号 3863URL http://hdl.handle.net/10097/37579

~~

,,,~, ~'!~

'~.',;~ ~.*;;

,~ ; **;:j~*

~~

J~(;

~~ ~ ~~i*+-

~~4*+･･1~~ J~~)~~~~!+~~~~~:

~~4,+･････-･"*~^~C~:~

~~~~~~

~1) t･-i J~5 ~

~D FEI ~~l~ ~i:

t~~ (IC:~~) l~~~l 9~~9J~ 1 2 ~

~~:4~~~~!;~U~;~ 4 ~t~~&~~ I T_:~ ~:~t~~~~;~C~~~~~t*Ill~~~~~~~4 (t~1b~"'-'~~~~~) ~) )~ ~~~7;~~:~:~

Hydrogen Embrittlement of Low Alloy Steel for Pressure

Vessels Based on a Large Size Specimen (~~;P7P**'i~~~)f iC J~ ~ ~E j~~~~~" ~! fi A f~:~~~i~~~IO)71~~;~~~:4~ }C f~~:~ ~ ~F~f~)

~Ci~: Bilal Dogan (GKSS R*se***h C**t**)

~ F~~ ~: ,-._.'-'･-*~ J*･ ~:J RRl~ ~~F'i ~( ~l 1 . Introduction

Many old vintage reactors still operate in an aggressive thermal hydrogen gas environment for more than

40years. The failure of the large thick walled vessel means devastating disaster and such a failure must be definitely

avoided. Decision as to the replacement or repair is required when defect size approaches to the critical one.

Hydrogen in a reactor wall absorbed during high temperature/pressure operation may have the potential to cause

critical or sub-critical crack growth duririg shut down. Now it has been found that among the possible failure modes,

hydrogen embrittlment cracking is the principal concern for the reliability of reactors.

Conventionally; Iaboratory crack growih experiments have been conducted using small fracture mechanics

specimens that are thermally hydrogen charged at high pressure/temperature autoclave or catholically charged then

tested in moist air for longer period of time. Then, fracture mechanics experiments have been compromised due to H

10ss from both bulk specimen and region of the erack tip. Since it is desirable to conduct hydrogen embrittlment

experiments for times between 10 and 100hours, which condition are relevant to reactor startup and shutdown

cycles and ultimately at temperatures up to 150'C in order to establish minimum pressurization temperature, it is

necessal~r to employ a method to retain a near-constant H concentration in a compact tension specimen during

loading. Therefore, an alternative procedure must be developed and employed to reliably mininize hydrogen loss

during long'term hydrogen embrittlment crack growth test.

The objectives of this study are 1) to clarify how absorbed hydrogen in steel influences the long term, time

dependent crack growth behavior by realizing hydrogen charging large size specilnen, 2) to establish the critical

telnperature regime where internal hydrogen embrittlement (IHE) of old vintage steel (=temper embrittled steel) is

eliminated, 3) to interpret the internal hydrogen embrittlment phenolnenon of large size specimen by numerical

analysis in consideration of the effect of both hydrogen dissipation and ambient temperature and 4) to interpret the

characteristics of hydrogen environment embrittlment (HEE) problem encountere.d in H2 st,orage system by

- 45 -

applying the numerical analysis method proposed in this study.

E 35 E

- 30 ce

<1_ 25 J: ~

o) 20 c G)

2; 15 ~

o* 10 o)

~(o 5 ,Q * n O ~

. r IJ2 -1 K~:~o'ooc5 MPam s

r 3.5T-CT J

i,.i~,~~] 1 .OT-CT $ fixcMOD

o 2 4 6 8 10 12 14 16/xIOOks

O I OO 200 300 400 Time t

Fig. I Crack growth test results

lh 4

E~3 '~

~

~2

a3c~_

C,1

1

o

'~r 5flT *-[ ~ -3 5T CT

j : - Jo; ' ':i : Hydregen charge :

:;i i r:1::~'~'5:~~]~~i_: 420Qcxl2MPa ~";

~: ~ :1i i No stress i

{ ;:: :j 1 OT CT

: _:_, _i____L.__]_._

'0= : ,;o,=:

'~ i~~jra o '~;

=//>=== = = = ==== = = '=" 1 OOO 1 OO 10

O

Exposure Time in Air RT, tlks

Fig.2 Decrease in hydrogen concentrations with time

ll

2. Effect of specimen size on crack growth under hydrogen environment

High'temperature high-pressure hydrogen charging of a large 3.5T-C(T) specimen using 2.25Cr- IMo, a

new-generation steel, was realized and it was compared with a conventional 1.0TC(T) specimen to examine the

influence of specimen size on the crack growth behavior over a prolonged period of time. It was found that while

there was virtually no difference between the large-size specimen and small-size specimen in terms of the lower

bound threshold for hydrogen assisted cracking (: J~iH), the g 'owth period of cracking greatly depended on specimen

size (Fig. 1). The hydrogen concentration distribution inside the 3.51~C(T) specimen was examined and it was found

that the hydrogen concentration inside the large-size specimen showed an inverse U-shape distribution, where

hydrogen concentration decreased near the surface. Furthermore, the decrease in the rate of hydrogen concentration

in the central part that was nearly one-tenth of that of the 1.0TLC(T) specimen when they were left in the

atmosphere at ambient temperature (Fig.2). The crack growth experiment on 3.5T-C(T) specimen was interrupted,

and the hydrogen concentration around the extended crack was examined by freeze cutting the specimen (Fig.3). A

high hydrogen concentration was determined around a crack tip (Fig.4). It was also found that this highly

concentrated area migrates as a crack progresses, and such data that are highly valuable were obtained.

FP HA

Determination of Hydrogen by

the inert gas fusion thermal conductivity method (JIS Z2614)

Fig.3 Freeze cutting procedure of crack growth specimen for hydrogen analysis

HAc: Hydregen ~l ~ Assisted crack

-A- Line I ~o- Line A :d I -c- ~ine 2 ~c- ~ine B :/ :

-A- Line C

c AA

:: 2 ~~~~~~'~~c,v;~~¥~ o o

O 20 40 60 80 O 20 40 60 80 Oistance from Dist~nce from

Fatigue Crack tip* r / mm Patigue Crack tip, r/ mm

(a) Interrupted (b) Interrupted at t~:238ks at t=238ks

Fig.4 Hydrogen concentration distribution along extended crack

- 46 -

~

{.} }

~

j~~ ~

~

{Ii {; :'*; ',,{ I

~*?.

*~'

'~~; .*~ +'",i~ {; *,'~ ~ ,;:': ,

,*i; ~*i{ i

~

~.' ~

~

~

,~';"'~~ ~

i ~ " I ~

"~'~ *{i '*i*;;'

,:,*

;

i"~{

'~' I }

I

;"'=' ,;*~

='*{Ii il

ji;

.* I,

ljj 'f";'j

~

,･~~~,;i

:',;{ ';,:i

**, {{-

*';~ ,**{-

*~!

;~~;;

~~*,

~･;

<*;*;~

~;;

3. Hydrogen Embrittlement test of aging pre8sure vessel steels using

a large thick specimen

2.25Cr-lMo, an old-generation steel, whose susceptibility

to temper embrittlement was increased on purpose, was selected,

and a comparison was made between this old-generation steel and

the new'generation steel in terms of susceptibility to hydrogen

elnbrittlement and elucidated the difference in fracture toughness.

On the basis of this comparison, it was found that hydrogen

absor~Ption affects critical flaw growth that occurs as cracks develop,

and confirmed there was an interaction between temper

enrbrittlement and hydrogen embrittlement. Furthermore, it was

discovered that the influence of hydrogen absor~Ption ceases to play a

role in crack development and a degradation of fracture toughness in

the temperature region between 86 oC- 1500C.

,:~* E

ce Q* 2 ~ ~<:~~

70

60

50

40

30

20

io

~

calculation

O 3.5T-CT O O)11'11 - - - - Literature data O 1.0T-CT

*1 Mo steel

<;Ys=500MPa

.O 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0

H dro en Concentration C / nm y g , (1'23) p*

Fig.5 Relationship between l~iH and hydrogen concentration at l/2 thicknesses.

~ ~ ~a' 16* ,~ ~

~~ 120

~ IS

coao LL

~ ,h

~ 40 ~;

,,, $,, ~

~5 O

;-----/:/lv/ : : t 1/ ~ t a' t L/ I e I E '(( C~ack t "" '

~, ~ Ineubation (// propa9ation t <1 '

l ~;: :

//// ~~ ~ef*S2~{~11~*'Af$~1l~. tl~'n 5Af$~ll --- Analyiicgl jine *

-*-* Predleted line No Cracking e: E erimental data!B~88.9mm

lco

~ 20

80

40

o

10* Io'* Io' Io* Io' Io' Io'/ks

4. Numerical analysis on hydrogen difftision and coneentration io~' Io' Io' ~o' Ic / h 1 o"

around the crack tip Time, t Fig.6 Trend of the annihilation of hydrogen

A physical model based on the hydrogen distribution inside embrittlment shown by the j~i~i increase with time the steel obtained was constructed, and the hydrogen diffusion using

this physical model was analyzed. High-precision analysis method for the local stress gradient term that was the

driving force of hydrogen difftision using a hydrogen diffusion formula was originally proposed, and the behavior of

hydrogen diffusion and concentration around a crack tip in consideration of the inverse U-shaped initial hydrogen

distribution was ascertained.

(1) The experimental results obtained were verified by the numerical analyses under the conditions in which stress

increased at various loading rates. The analysis demonstrated that even under a short･term rising load condition

(dK/dt = O.005 MPa 1:nl/2/s), hydrogen is accumulated around a crack tip at a concentration equivalent to that

observed under a long"term rising load condition (dK/dt = 0.0005 MPa mi/2/s). Thus, it is considered that even in a

small fracture'mechanics specimen, a value of J~:H that is equivalent to that obtained with a large

fracture-mechanics specimen can be obtained, because the effect of hydrogen dissipation from the specimen is small

in such a short period of test time.

(2) It was demonstrated that the threshold stress intensity factor d~i:H that is necessary for crack initiation increases

when the hydrogen coilcentration decreases within the thickness of steel (Fig.5). However, for a stress intensity

factor of more than around 55 MPa Inll2, hydrogen diffasion and accumulation characteristics become insignificant,

because the stress gradient that is a driving force of hydrogen becomes insignificant. On the basis of the above, it

was obtained basic information for use in predicting the duration of crack growth for specimens with various

thicknesses in which hydrogen is charged at the beginning of a test (Fig.6).

- 47 ~

(3) It is shown by the analysis that with the increase in temperature, the

maximum hydrogen concentration attained around the crack tip

decreases. This is considered to be related fundamentally to the

attenuation of stress gradient ~t elastic ' plastic boundary due to the

yield strength reduction at higher temperature. The analysis

demonstrates that the faster diffusion and concentration occurs in a

shorter time with less saturated hydrogen concentration. This would

indicate that lower bound I~iH can be obtained under the short-ternl

rising load condition even at elevated temperature condition. Therefore,

the observed elimination of hydrogen crack growth in the temperature

regime between 86 and 150~C, which tests were conducted under

short-term test condition, is considered to represent the threshold for

hydrogen cracking where hydrogen concentration would not occur at

extended period of test time any more.

(4) Based on the obtained results shown above, an algorithm to predict the ~

~

crack growth life of steels with various thicknesses was constructed (Fig.7) ~ and the example problem was introduced (Fig.8). ~

~S ~

o

5. Effect of hydrogen gas pressure on the mechanical properties of low auoy

steel for hydrogen pressure vessels

S.TE~P1. ~ DeterSline,SSeei.th.icS;ness. :B. Determine K and dK/dt Fiaw shepe* ~aw en'entatipn~~K. so!Su cycle~ ~dxld~ * ' - '

~~~~

STE=P 2' EVai~ate ~x a~c time. at a= ry~st. ,~f '~--::1;:; '~t 1~:"rl~

/v c**s~

~ (~1i/(/Pfl"~s'tipa ~L e' I~ S (l <

f -~'1Q~nj$ + c'lldii'f '

e't ~o '~o

Tim~, f 'b ~'c~"

ST~P' 3. :;' ~~t. i~at~ ; {:cr'aick't is9'row'th ~tei;:::d' ~dj; '; ~" by wr ~'sr'ture d~t~ ~::. ~1~~ s'f"~' tlttt at RT

~: c~f'*s'ss~ro~:);t~~'~x'l$'4'~#~ co't st"~ il RT

'~''~"(t*oe;7~;k""' 't""~ ~ ~~ "I'

'~'ll~'~"~

As~='~~P~o~' ~~fQn~feet'o~esf' ~#P tee~~idered ~:~~

Prediciton of amount of crack growth Aa

Fig.7 Flow chart for crack growth prediction

i

-8*2s4~*~ 280 = cr~ck arF~~te~

-B*se$m~ pT'die'd t **3~i~e..~ -~*2s~~* '+ 240 B-2s4~lm -so*2..e~'s~~"~ E~pe'~T1'~te~ ~~/B*o.9 -* J

200 ~ F~' c'~0D #. ' leo : ; ,

1 20 , = crack ariested

80 ct~e~ sff*st~d t.,-.***477h t.,*~;**3s.5 t.~~?39oh B=88.ernrn 40 ~*,*~~4sh AslB*o~

B*25Amm o ~*!~*o~

10 Ioo looo loooo

Time , tl ~

Fig.8 A problem of exangPle The physical model proposed in this study was applied to the subjects in the hydrogen-gas environment,

the resulting values was compared with the hydrogen charged condition. It was demonstrated that hydrogen

concentration around a crack tip occurs at a conspicuously accelerated rate in the case of hydrogen gas environment

compared to hydrogen charged environment.

6. Conclueions

In this study, effect of specimen size on crack growth under hydrogen environment was primarily

discussed in order to clarify how absorbed hydrogen in steel influences the long term, time dependent crack growth

behavior by realizing hydrogen charging the large size specimen. High-temperature high-pressure hydrogen

charging of a large 3.5T-C(T) specimen using 2.25Cr-1Mo steel was realized and it was compared with a

conventional 1.0TLC(T) specimen. The large size specimen revealed the prolonged crack extension with a higher

hydrogen concentration remained for a longer time. While it was demonstrated that in a small fracture-mechanics

specimen, a value of ~iH that is equivalent to that obtained with a large fracture'mechanics specimen can be

obtained under a short term test condition. The numerical analysis was carried out in consideration of the effect of

applied stress intensity ~), rate of stress intensity (dl~dt), rate of hydrogen dissipation from specimen and ambient

temperature. The analysis results well explain those experimental observations of large size specimen. Based on

those, an algorithm to predict the crack growth life in thick'sectioned pressure vessel was constructed. Finally, this

physical model was applied in the hydrogen environment embrittlement problem.

- 48 ~

II

* ,**

* ~ tj li i' li i{{

,, ~ "i; ~

~

~.

~

I

~

論文審査結果の要旨

犀力容器用低合金鋼の水素脆化挙動は,実用的に重要であるにも関むらず,試験片からの水素散逸な どにより,未だ解明されていないのが現状である.著者は,水素散逸が軽微となる大型試験片を用いた

実験方法の提案と系統的な実験を行い,さらに水素拡散数値解析を行って,両者の績果を組み合わせる ことにより,圧力容器用低合金鋼の水素脆化挙動に及ぼす試験片寸法,水素環境条件および力学的効果 を明らかにしている.本論文はこれらの成果についてまとめたものであり,全編6章よりなる.

第1章は緒論であり,本研究の背景および目的と意義について述べている.L

第2章では,新材相当の2.25Cr・1Mo鋼を用い,大型3.5TC(T)試験片に高温高圧環境で水素添加.し

て,水素脆化試験を行い,従来の1.OTC(T〉試験片と比較することに.より,大型試験片を用いることに

より,適切な水素脆化試験参行えることを示している.さらに,大型試験片内部の水素濃度分布状態を 極低温処理して調帯,き裂先端力)ら少し離れた位置に高濃度り水素が凝集していることも初めて明らか 1;している、本結果は,大型試験片1こよる水素脆化試験の妥当性を示し,き裂朱端近傍の実際の水素濃 度分布を初めて明らかにしているもので工学的に重要な成果である、

第3章では,経年劣化を模擬した焼き戻し脆化2.25Cr4Mo鋼(脆化材)を用いて,第2章で明らか にした新材との水素脆化敏感性の比較を行っている.その結果,一脆化材に見られるき裂の急速成長破壊 に水素吸蔵の影響が認められることを見出し,焼戻脆化と水素脆化の相互作用現象解明に新たな知見を 与えている.また,・86～150℃付近以上の温度域において,き裂成長と破壊靱性の低下に及ぼす水素吸

蔵の影響が消失することを明うかにしている.本結果は経年劣化材の水素脆化敏感性を明らかにするも ので,工学的に有用な成果である.

第4章では,第2章で得られた鋼材内部の水素分布状態を基にして物理モデルを構築し,このモデル に基づいた水素拡散数値解析を行っている.本章では,水素拡散方程式において,水素駆動力となる局

所応力項ρ高精度表示を行い,本物理モデルに即した初期水素濃摩分布におけるき裂先端近傍の水素拡. 散凝集挙動を明らかにし,第2章で得られた実験結果の検証と大型試験片の長時間水素脆化割れによる

き裂成長寿命を予測するア々ゴリズムを構築している.本結果は,的確な物理モデルの構築と数値解析 により水素の濃度分布とき裂成長寿命を予測できることを示すもので,水素脆化の問題に対する計算科 学の工学的応用の意義を示す有用な成果である, 第5章では,試弾片の外部から水素ガスが浸入する水素ガス脆化の問題に本研究で確立した脆化予測

方法を適用し,第2,4章で得られた水素添加条件下での結集と比較して水素ガス脆化は,より加速的 に進行することを示している.

第6章は結論である.

以上要するに拳論文は・大型試験片による水素脆化試験法の提示とそ抑に整斉くき裂成長挙動と寿命 予則法を水素拡散解析と組み合わせて確立したものである.尊こで得られた成果は,水素脆化に関わる プラント等,機器構造物の安全性,信頼陸の向上に有益なものであり,ナノメカニクスおよび機械工学一 の発展に寄与するところが少なくない.

よって,本論文は博士(工学)の学位論文として合格と認める.

一49一