V ol. 18 No.

2 CH INES E JO U RN A L O F AER ON A U TICS M ay 2005

Study on the Machining Distortion of Thin walled Part

Caused by Redistribution of Residual Stress

WANG Zhao jun, CHEN Wu yi, ZHANG Yi du, CH EN Zhi tong , L IU Qiang
( School of M echanical Engineer ing and A utomati on , Beij i ng Univer si ty of A eronauti cs
and A st ronaut ics, Beij ing 100083 , Chi na)

Abstract: In or der to reduce the w eight of airplane and increase its mechanical behaviors, more and
more larg e integrated parts ar e applied in modern aviat ion industr y. When machining t hin walled aer o
plane parts, more than 90% of the mater ials would be removed, resulting in sev ere distortion of the
parts due to t he weakened rigidity and the release of residual str ess. T his might also lead to str ess co n
centration and damag e of the parts. T he effect of mater ial removal from residually stressed billet is simu
lated using F EA softw ar e M SC. Marc and the causations of disto rtion is analyzed. T o verify the finite el
ement simulation, a high speed milling test on aluminum alloy 7050T 7351 is carried out . T he results
show that the simulation result is co nsistent with the ex perimental one. It is concluded that the release of
r esidual stress is t he main cause of machining distort ion.
Key words: finite element method; simulat ion; residual stress; machining distort ion
残余应力重分布引起的薄壁零件加工变形研究. 王兆 峻, 陈五一, 张以都, 陈志 同, 刘强. 中国 航空
学报( 英文版) , 2005, 17( 2) : 175- 179.
摘 要: 现代航空工业中为 减轻 飞机重 量, 提高 飞机的 各项 机械性 能, 整体 构件 越来 越多 地被 使
用。加工大型整体 薄壁构件时, 有 90% 以上的材料 被切削 加工去 除, 由 于材料 去除后零 件刚度 的
降低以及 应力的释放, 造成过大的加工变形。本文用 M SC. M arc 有限元 软件仿真了 铝合金预 拉伸
板材料去 除对于加工变形的影响, 并分析 了加工 变形的 成因。为验 证有限 元结果 的正确 性, 在 高
速数控铣床上加工了同样的试件。结果表明仿真结果与实验结果 一致, 残余应力的 释放与重 分布
是薄壁零件加工变形的主要原因。
关键词: 有限元法; 仿真; 残余应力; 加工变形
文章编号: 1000 9361( 2005) 02 0175 05 中图分类号: V214. 4 + 1 文献标识码: A

M odern aviation is developing t owards high

speed and heavy load. A lot of thin w alled aero
plane components, machined f rom solid aluminum
billet s rat her t han assembled from separate parts,
have been increasing ly applied in modern aviat ion
indust ry t o reduce the assembly t ime/ costs and t o Fig . 1 T he deformed shape of thin w alled part
improve t he performance of the planes. T hese part s after machining
are thin in t hickness, com plex in st ruct ure and
( 1) F ix an aluminum blank on the w orkt able.
tight in machining tolerance. Due t o t he w eakened
( 2) Perform a 2 D machining on the t op sur
rigidit y and the release of residual st ress aft er more
f ace w ith small DOC and reach t he size of the 2 D
than 90% of t he mat erials are removed, severe dis
profile direct ly, i . e. , leaving no allow ance for fin
t ortion is of ten observed on the part s, F ig 1.
ishing.
Presently, t he procedure to machine a large
( 3) M achine the blank layer by layer t ill ar
scale integrat ed part in w orkshops is oft en arranged
riving at t he f inal shape.
as f ollows:
R eceived dat e: 2004 10 10; R evision received dat e: 2005 01 24
Foundation it em: N at ional Defen ce Fundamental Research

© 1994-2010 China Academic Journal Electronic Publishing House. Open access under CC BY-NC-ND license. http://www.cnki.net
 176 WA NG Zhao jun, CHEN Wu yi, ZHAN G Y i du, CHEN Zhi tong, L IU Q iang CJA

( 4) Unf ast en t he part from t he w orkt able. 1 1 Material properties

At t his st age, unacceptably large distort ions T he mat erial of t he part is aero aluminum al
of ten occurres. T he highly distort ed part may no loy 7050T 7351. T he alloy is assumed t o be an elas
long er be able t o serve its designated f unctionalit y t ic plast ic material. Its physical specif icat ions are
or may require signif icant reworking to render it s listed in T able 1.
funct ion. T hese problems result in high scrap rates Table 1 Physical parameters of aluminum
and increase manuf acturing costs. alloy 7075T7351 [ 6]
It is assumed that t he release and t he redist ri Propert ies V alue Remark
bution of t he residual stresses in the billet are t he Y oung s modulus 71 GPa
major reasons of t he distort ions. T he residual st ress Poission s rat io 0. 33
comes mainly from t he rolling process of the billet s Densit y 2800 kg/ m 3
and is also influenced, t houg h less signif icant ly, by Thermal conduct ivit y 155 W/ m ! 25 !

cut ting f orce, fastening force, cut t ing t em pera Specif ic heat 960 J/ kg ! 100 !

t ure, etc. Many methods to eliminat e residual 1. 2 Geometry of the part

st ress have been adopted. Af ter a rolled aluminum T he blank of the part is a pre stretched alu
blank is st ret ched along rolling direct ion to produce m inum plate. Its dimension is 300 mm ∀ 104 m m
2 5% permanent distort ion, it s residual st ress ∀ 24 mm. A t hin w alled st ruct ure is desig ned sim
magnitude may drop 90% . Anot her effect ive ilar to a t ypical aero plane component , see F ig 2.
[ 1]
met hod is uphill quenching . Nevert heless, it
seems impossible t o elim inat e t he residual st ress
complet ely. Machining distort ion may occur as
long as t he residual stress ex ist s.
T he residual stress and its measurement w ere
ext ensively invest igat ed during 1970 s t o 1990
s[ 2 4] . T he dist ribut ion of the residual st resses par
t icularly in aluminum billet s w as st udied by Wang,
[ 5] Fig. 2 Designed specimen
et al . How ever, its effect s on m achining dist or
t ion and t he t echnique to reduce the dist ortion have Due t o its symmetry, t he 1/ 2 model is used,
not yet been f ully explored. T he paper present ed and t he model is divided into 49896 elements and
here det ailes a research concerning t he effect of ma 55600 nodes. F ig 3 is the F EA model of t he block
t erial removal on the dist ortion of the thin w alled in simulat ion.
part s vi a FEM sim ulat ion and experim ent al w ork.

1 F EM Simulation of M achining Distortion

During the CNC machining process, m any
factors, such as residual stress, cutt ing force, cut
t ing temperature, fastening f orce, f ix ture conf igu
rat ion et c. are not t aken into considerat ion, when
coding the CNC machining program. T o reveal t he
real eff ect s of t hese f actors on the m achining pro
cess, a f inite element simulation can be adopt ed.
F ig. 3 F EA M odel
T he FEM simulat ion of t he cut ting process has
g reat value t o understand t he cut t ing process and t o 1. 3 Boundary conditions
reduce t he number of ex periments. T he f inite element met hod is a pow erful tool

© 1994-2010 China Academic Journal Electronic Publishing House. Open access under CC BY-NC-ND license. http://www.cnki.net
 M ay 2005 St udy on t he M achining D istortion of T hin w alled Part Caused by R edist ribution of R esidual St ress
to assess potent ial distort ions caused by the ma
chining process. It provides a good approx imate so
lut ion to continuum problems. T he FEA sof tw are
M SC. Marc is used t o analyze t he dist ort ions for it
can define the elements to be inact ive. T his tech
nique can simulat e the met al removal act ion in cut
t ing procedure.
Once the part is fixed on t he worktable, it can
not move t o the t able but can move away from it.
M ost FEA soft w are, including MSC. M arc, does
not provide single side const raint . In t his paper
cont act analysis has been used for t he solut ion of
single side constraint in order t o simulat e t he real
fixt ure more aut hent ically. T he w orkt able is con
sidered as a rig id body in t he F EA model, so t he
part can not move tow ards it.
1. 4 Simulation Results
In the simulat ion, t he part mat erial is re
moved layer by layer. Each layer is 3mm in thick
ness ex cept t he last layer that is 2mm thick. T he
removed layer is set to be inact ive to simulate t he
milling process. In order to st udy t he deformat ion
of t he bot tom point # A ∃ is chosen, Fig 4. Fig 5
shows the deformation of point # A ∃ af ter each lay
Fig. 4 Point selected to measure the defor mation
F ig. 5 Bottom deformation of t he specimen
© 1994-2010 China Academic Journal Electronic Publishing House. Open access under CC BY-NC-ND license. http://www.cnki.net
177
er is removed. Fig 6 is t he def ormat ion nephogram
when t he removed lay er is 9 mm in thickness.
Fig . 6 Distortion w hen 9 mm layer is removed
It can be seen from Fig 5 t hat the deformat ion
increases st ep by step unt il the fourt h layer is
reached. T hen it drops dow n f rom t he fourt h t o
the sixt h layer. During the removal of the last
three layers the def orm at ion varies very lit tle. It is
anticipated that w it h more and more met al being
removed, t he distort ion w ill increase.
T he reason w hy t he distort ion decreases when
half of the material has been cut aw ay m ay be ex
plained as follow s.
Wit h the removal of the residually st ressed
mat erial, t he rem aining st resses in the workpiece
redistribut e t o reach new equilibrium . For pre
stret ched aluminum billet s, the surf ace is more
likely t o ex hibit t ensile stress[ 7] . T he inner part of
the piece is t herefore st ressed compressively . T hat
is, t he stress changes from t he t op surf ace t o t he
bott om in a tensile com pressive t ensile manner.
T he FEA result s using M SC. M arc show t hat when
three layers of t otally 9 mm material have been re
moved, t he stress distribut ion alt ers dramat ically,
see Fig 7.
Although t he overall st resses are st ill bal
anced, t he concent rat ion of the tensile stress in t he
upper part of the specimen w ill cause both ends of
the specimen t o deflect upwards, F ig 6. As t he
machining progresses, t he t ensile st ress wit hin t he
low er part increases g radually. Af ter the f if th layer
has been cut aw ay, the low er part t ensile st ress of
the specimen becomes st rong enough t o balance t he
 178 WA NG Zhao jun, CHEN Wu yi, ZHAN G Y i du, CHEN Zhi tong, L IU Q iang CJA

T he cutt ing speed is 18 000 r/ min, t he depth

of cut is 0 5 mm, and the f eed rat e is 4000 mm/
min. T he dist ort ion of the bot tom of t he part is
measured in a coordinate measuring machine.
T able 2 show s bot h t he simulated and mea
sured values of dist ort ion. It can be seen from
T able 2 that t he m easured dat a increase in t he f or
mer four st eps, then decrease from t he fift h st ep.
T he deformat ion does not change drastically in t he
last three steps. T he measurement and the F EA
simulation of distort ion receive very sim ilar change
F ig. 7 Residual stress after removing the third layer
able t rend. Since t he simulation result is derived
upper part one, as show n in Fig 8, and the up under t he assumpt ion of neg lect ing cut t ing heat and
w ard deflect ion reduces as a result. When the ma tool v ibration, t here is relative error especially in
chining on the six th t o t he sevent h layer have been the last step.
performed, most residual st resses, no mat ter t en Table 2 The distortion of the bottom
sile or com pressive, are released and the distort ion R emoved t hickness/ mm 3 6 9 12
of t he specimen decreases t o a very low level, Simulat ed result s/ mm 0. 049 0. 122 0. 165 0. 193
M easured dat a/ mm 0. 068 0. 139 0. 192 0. 218
F ig 5.
R emoved t hickness/ mm 15 18 21 23
Simulat ed result s/ mm 0. 170 0. 049 0. 048 0. 034
M easured dat a/ mm 0. 180 0. 088 0. 081 0. 069

3 Conclusions
Wit h t he removal of the f irst few layers on t he
specimen, distort ion increases ow ing t o uneven
concent rat ion of the stress.
T he residual stress in t he blank varies severely
accom panied by t he remove of mat erial. T his re

Fig. 8 Residual stress after removing the fifth layer

2 Experiment Verification
In order t o verify the simulat ion results, t he
designed specimen part is machined on a F idia CNC
machining cent er, see Fig 9.
sult s in t he machining deformat ion of thin w alled
part . T he release of the residual st ress in the blank
is the main cause of machining dist ortion especially
for t he t hin w alled part .
Aft er half of t he mat erial is cut aw ay, t he
stress distribut ion becomes bet ter balanced and t he
dist ortion reduces gradually. T here is a crit ical val
ue of material removal amount. When the amount
of t he mat erial removed ex ceeds the critical value
the machining deformat ion drives to st eady.

References
[ 1] Sun J, K e Y L, K ang X M . Study on technology strat egies for
guarantee machining precision of large scale int egrat ed aircraf t
Fig. 9 M achined specimen parts [ A ] . Progress of M achining Technology [ C ] . Beijing:

© 1994-2010 China Academic Journal Electronic Publishing House. Open access under CC BY-NC-ND license. http://www.cnki.net
 M ay 2005 St udy on t he M achining D istortion of T hin w alled Part Caused by R edist ribution of R esidual St ress
A viat ion Indust ry Press, 2002. 695- 701.
[ 2] Iw ata K , Osakada K, Terasaka Y. Process modeling of or
t hogonal cut ting by the rigid plast ic f inite element met hod[ J] .
T rans of t he A SM E, J of Engineering f or Indust ry, 1984,
106: 132- 138.
[ 3] Lin Z C, Lai W L, Lin H Y, e t al . Residual st ress w it h dif
f erent t ool flan k wear lengths in t he ult ra precision machining
of N i P alloys [ J] . Journal of M at erials Progress Technology,
1997, 65 : 116- 126.
[ 4] Kalhori V . M odeling and simulation of mechanical cutt ing
[ D ] . S weden: Lulea U n iversit y, 2001.
[ 5] 王秋成, 柯映林, 章巧芳. 7075 铝合金板材残余应力深度 梯
度的评估[ J] . 航空学报, 2003, 24( 4) : 336- 338.
W ang Q C, K e Y L, Zhang Q F. Evaluation of residual st ress
dept h prof iling in 7075 aluminum alloy palt es[ J] . Act a A ero
naut ica et A stronaut ica Sinica, 2003, 24( 4) : 336- 338. ( in
Chinese)
[ 6] 吴学仁. 飞机结构金属材料力学性能 手册[ M ] . 北京: 航 空
工业出版社, 1996. 448- 453.
W u X R. Handbook of m echanical properties of aircraf t struc
t ural metals [ M ] . Beijing: A viat ion Indust ry Press, 1996.
© 1994-2010 China Academic Journal Electronic Publishing House. Open access under CC BY-NC-ND license. http://www.cnki.net
179
448- 453. ( in Chinese)
[ 7] Sasahara H, Obikaw a T, Shirakashi T . FEM analysis of cut
t ing sequence eff ect on mechan ical charact erist ic in mach ined
layer[ J] . Journal of M at erials Processing T echnology, 1996,
62: 448- 453.
Biographies:
WANG Zhao jun Born in 1976 in
Shandong, he receiv ed his Ph. D. degree
in machine manufacture and automation
from Beijing U niversity of Aeronautics
and Astronautics ( BU AA ) . M ajor in
new technolog y of mater ial processing,
cutting simulat ion and machine desig n.
T el: 010 82317754, 13121647009, E mail: zhaojunbuaa @
163. com
CHEN Wu yi Born in 1951 in Shanx i, doctor and profes
sor , he is dev oted to t he research o f par allel machine tool,
cutt ing principle and machine desig n in BU AA. T el/ F ax:
010 82317754, E mail: w ychen@ buaa. edu. cn