Scribd
Upload
63 views

A Finite Element Method Based Process Model For Prediction of Spring Back During Forming of PFBR Vessels

1) A finite element method (FEM) based process model was developed to predict springback during forming of multiple curvature plates for vessels. The model was validated against experiments and showed 94% accuracy in predicting plate profiles. 2) The process model used shell elements and simulated loading explicitly and unloading implicitly. It will help predict final shapes for a given die/punch/plate setup and correct for springback. 3) An FEM model was developed and validated experimentally for forming 316L stainless steel plates with curvatures required for the main vessel of a prototype fast breeder reactor. The model accurately predicted plate profiles and can be used to design the plate forming process.

Uploaded by

sreeshpsomarajan
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
63 views

A Finite Element Method Based Process Model For Prediction of Spring Back During Forming of PFBR Vessels

1) A finite element method (FEM) based process model was developed to predict springback during forming of multiple curvature plates for vessels. The model was validated against experiments and showed 94% accuracy in predicting plate profiles. 2) The process model used shell elements and simulated loading explicitly and unloading implicitly. It will help predict final shapes for a given die/punch/plate setup and correct for springback. 3) An FEM model was developed and validated experimentally for forming 316L stainless steel plates with curvatures required for the main vessel of a prototype fast breeder reactor. The model accurately predicted plate profiles and can be used to design the plate forming process.

Uploaded by

sreeshpsomarajan
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

SCIENCE-2

A Finite Element Method based Process Model for


Prediction of Springback during forming of PFBR Vessels
n EXECUTIVE SUMMARY
A finite element method (FEM) based process model had been developed for estimation of springback during forming of
multiple curvature plates. The results obtained from the simulation were validated with plate forming experiments and
the correlation between the simulated results and experiments is excellent. The accuracy level of the validated process
model was up to 94% in predicting the profiles of the formed plates. The process model was prepared with reduced
integration shell elements incorporating dynamic explicit FEM for the loading phase and static implicit FEM for the
unloading phase during plate forming. The FEM based process model would be very helpful to predict the final shape of the
product for a given die, punch and plate configuration and subsequently use this information to correct the die and punch
shapes for springback to manufacture the sectors of large vessels to the desired accuracy.

n OUTLINE
The fabrication of main vessel, safety vessel, inner vessel and transition
pieces of 500 MWe Prototype Fast Breeder Reactor (PFBR) involves
assembling and welding of multiple curvature preformed plates
(hereafter called as petals in this report) to form the final shape. These
petals need to be fabricated with stringent dimensional tolerances in
order to fabricate the vessels within the desired level of tolerance
conforming to the operational and functional requirements.
Springback is a major factor controlling the dimensional accuracy of
the petals. Springback occurs due to recovery of the elastic strains after
removal of the punch at the end of the forming operation and the tools
need to be designed with appropriate springback allowances for near-
net shape forming of petals. Springback is an intrinsic characteristic of
the deformation process and it can only be minimized by proper
process design, but cannot be completely avoided. FEM based process
models for plate forming operation would help in predicting the
springback as well as facilitate in designing dies for near-net-shape
forming. Such a model using FEM was developed for simulating Fig. 1 : Plate profiles along transverse centrelines of the
plate
multiple curvature plate forming operations and the results were
experimentally validated.
The optimum FEM formulation for plate forming was identified as
sequential dynamic explicit formulation with contact modeling for the
pressing phase and static implicit formulation for the unloading phase
for springback calculation in the Lagrangian framework. The results
from the explicit analysis were imported into an implicit FEM code and
a static analysis was performed to determine the springback and the
residual stress state in the plate after springback. The displacement
obtained at the end of the step is the springback and the stresses and
strains give the residual stress-strain state in the material after
springback.
To validate the FEM process model, double curvature forming
experiments were carried out on 300 mm square and 15 mm thick
316L(N) stainless steel plates. The profiles of the formed plates were
measured using a CNC ZEISS Universal Measuring Machine. Figures 1
and 2 show the centre-line and diagonal sectional profiles from both
experimental and FEM results superimposed on the desired (punch)
surface profile. From the figures 1 and 2 it is evident that the FEM Fig. 2 : Plate profile along diagonal of the plate
simulations of the plate forming trials compare well with the
experiments, the differences between the simulated and the
experimental profiles of the plates being within 6-10%. The deviation of
the simulated profiles from experimental ones is more towards the edge
of the plate than that at the centre-line. The radius of curvature (ROC)
values at various sections were obtained for both experimental and
FEM data. The variation of percentage springback, which was defined
as the percentage deviation of final ROC at a section from the
corresponding punch ROC, is shown in figure 3. Figure 3 clearly
depicts that springback increases towards the central sections of the
plate. From the FEM results it was found that the equivalent plastic
strain is maximum at the centre due to maximum deformation
occurring at the centre during forming. The differential distribution of
strains that occurs during plate forming leads to differential
springback in the plate and this in turn leads to a product shape
different from the die and punch shapes. So the springback corrections
on the dies and punches have to be imparted to obtain the desired
product shape. It was also found from the experiments that springback
is inversely proportional to the tool stiffness. Fig. 3 : Variation of springback across formed plate

4
SCIENCE-2

n ADDITIONAL INFORMATION ON VESSEL TOLERANCES


The petals need to be fabricated within tight radial dimensional tolerances in order to fabricate the vessel within the
desired overall tolerance limits. The amount of residual cold work is also a crucial issue in the nuclear vessels and it is
specified to be within a predetermined level to ensure that the fabricated vessels see the intended life. For large vessels like
the main, safety and inner vessels the cold work limit has to be adhered to strictly during the forming stage because heat
treatment after forming as well as fabrication would be very difficult due to large size of the petals. While on assembly
during welding, the areas of localized deformation will relax and will cause variation in dimensions of the assembled
pieces. Since it is very difficult to quantify the level of cold work in the zones of localized deformation, the control of the
dimensions in the assembled pieces is very difficult.

n GENERAL EXPLANATION RELATED TO THE DESCRIPTION


The main vessel (MV) of PFBR is the most critical component in the reactor
assembly and forms a part of the core supporting path. It carries ~1150 tons
of liquid sodium along with the core and reactor internals (~920 tons). The
vessel, made up of type AISI 316L(N) stainless steel, has cylindrical top and
paraboloidal bottom with multiple radii of curvature. The vessel will be
fabricated by sequentially joining together a number of preformed plates
(called as petals in this report) such that the final desired shape of the vessel
is obtained. Figure 4 gives the schematic of the main vessel for PFBR and a Petal
typical petal. The paraboloidal portion of the vessel consists of 40 such petals
of different dimensions welded along three layers above a circular crown
plate. The radius of curvature (ROC) of the petals varies from 1500 mm to
19350 mm and thickness of the petals varies from 30 mm to 40 mm. The ratio
of radius of curvature to plate thickness (R/t) varies from 50 to 645 with R/t
in the range of 100 to 200 for most of the petals and the ratio of deflection to
the projected length in a plane (y/L) varies from 0.0123 to 0.099 with most of
the petals having the ratio as 0.053. The petals will be formed into the Petal
individual shapes and dimensions by cold forming in a hydraulic press. Near-
net shape forming of petals would help in keeping the cold work levels within
Main Vessel
the acceptable tolerance range. Estimation of springback in the petals is an Axis
essential step in design of the plate forming setup for near-net shape forming
of petals. Since the individual petal shape is complex, finite element method
(FEM) would be an ideal tool for proper design of the plate forming process
and the setup by predicting the springback and the cold work level in the
Fig. 4 : PFBR Main Vessel & a Typical Petal
formed plate.

n BRIEF DESCRIPTION OF THEORETICAL BACKGROUND


Complete 3-D finite element analysis (FEA) is necessary for the petal forming process because the geometries do not
conform to any idealizations like plane strain or axisymmetric analyses. The plate forming process includes both
membrane and bending effects in the workpiece and so a 3D shell formulation was used to account for both these effects.
The dynamic explicit FEM is a conditionally stable scheme with stable time increments of the order of 10-7s for 316L(N)
stainless steel and so use of reduced integration elements would be economical without losing accuracy of solution. The
inevitably high surface to volume ratio in plate forming would result in large number of contact conditions in the FEM
model necessitating increased execution time and convergence difficulties for a fully implicit FEM model. Geometric
nonlinearity needs to be incorporated to tackle the springback, which may occasionally be quite large.

n ACHIEVEMENT
An FEM based process model for multiple curvature plate forming was developed and the model was experimentally
validated. This process model would be very useful for predicting the final shape of the product for given die, punch and
plate configurations. Subsequently this information can be used to correct the die and punch shapes for springback to
manufacture the sectors of large vessels to the desired dimensional accuracy and cold work level.

n PUBLICATIONS ARISING OUT OF THIS STUDY AND RELATED WORK


1. Utpal Borah, Shreedhar Venugopal, R. Nagarajan, P. V. Sivaprasad and S. Venugopal. Trans. IIM, 60 (2007) 511.
2. Utpal Borah, Shreedhar Venugopal, R. Nagarajan, P. V. Sivaprasad, S. Venugopal and Baldev Raj, Intl. Journal of
Mechanical Sciences (Accepted for publication)
3. Utpal Borah, Shreedhar Venugopal, R. Nagarajan, P. V. Sivaprasad, V. Maduraimuthu, S. Venugopal and
S.L. Mannan. Proc. “Inter. Conf. on Pressure Vessels and Piping, OPE–Chennai” Feb 2006, Chennai.

Further inquiries:
Shri Utpal Borah, Materials Technology Division
Metallurgy and Materials Division, IGCAR, e-mail: [email protected]

You might also like