Professional Documents
Culture Documents
Outline
1 Fundamentals
Motivation of Simulation
Conclusion
WZL/Fraunhofer IPT
Seite 2
WZL/Fraunhofer IPT
Seite 3
System grinding
Input-Parameters
WZL/Fraunhofer IPT
chip thickness
WZL/Fraunhofer IPT
Seite 5
Characteristics of grinding
cutting edges possess different geometries mainly highly negative chip angle varying distance of the cutting edges and thus different chip thicknesses varying distance of the cutting edges from the rotation axes tool consists of three components (grain, bonding, pore) tool can be dressed in the machine
workpiece grinding wheel rake angle
in-feed
chips
pore
Seite 6
WZL/Fraunhofer IPT
Definition of grinding
Grain is fixed in the tool and moves on a kinematic curve Machining with geometrically undefined cutting edges
cutting edge chip
workpiece
Surface characteristics and deformation in the rim zone depend on the grinding direction
force-bound
track-bound
room-bound
energy-bound
WZL/Fraunhofer IPT
Seite 7
Fn,S Ft,S
bonding
phase of elastic deformation is followed by a phase of elastic and plastic deformation (II)
workpiece
vc
bulging T I elastic deformation II elastic and plastic deformation III elastic and plastic deformation and chip formation grit (cutting edge) chip
the chip formation begins, when the cutting depth reaches T (III)
source: Knig
WZL/Fraunhofer IPT Seite 8
Fn,S
grit path bonding
Ft,S vc
primary particles
WZL/Fraunhofer IPT
Seite 9
vw ae hcu ,max, SG 2 d eq vc d eq
Fc, 1/Rz, rS = f (hcu)
vw ae
v s
d eq
fas
WZL/Fraunhofer IPT
Process
set-values char. values cutting mechanisms
thermal mechanical
Result
technology workpiece
form-/ shapeaccuracy surface roughness influence on the rim zone
workpiece
geometry material
grinding wheel
profile specification
wear mechanisms
cutting speed mechanical thermal tribo-chemical
grinding wheel
wear clogging
dressing tool
type specification
efficiency disturbances
vibrations temperature output unit costs
cooling system
type system
Source: Knig, Klocke
WZL/Fraunhofer IPT
Seite 12
cylindrical grinding
externalinternal-
surface grinding
surfacerotary-
circumferential-
crosswise grinding
traversegrinding
crosswise grinding
traversegrinding
Seite 13
WZL/Fraunhofer IPT
WZL/Fraunhofer IPT
4 5
S8 S7 S6
3
S5 S4
2
S3 S2
1
S1
Grains engage to different degrees and are thus strained in differing amounts.
LS78 LS67
LS56
LS45 LS34
LS23
LS12
LS, kin
source: K. Steffens
WZL/Fraunhofer IPT
LS, stat
the kinematic cutting edge distance arises from the grinding parameters the static cutting edge distance can be measured
Seite 15
Abrasives
symbol abrasive Knoop hardness [N/mm2] thermal stability [ C] thermal conductivity [W/(m C)] fields of application
corundum
1950 to 2200
2000
medium-tough to hard materials below 60 HRC (Rm< 500 N/mm2) such as: steel malleable cast iron surface grinding of cemented carbides, cast-iron, ceramics, non-ferrous metals dressing hardened steel low alloy steel HSS precision grinding carbon saturated steel glas, stone, ceramics tungsten carbide, cermets
Seite 16
3000
1300
55
4700
1370
200-700
diamond
8000
900
600 - 2100
WZL/Fraunhofer IPT
Bond types
symbol bond type characteristics
brittle and thus sensitive to impact load high modulus of elasticity high thermal stability, low resistance to thermal shock chemical resistance against oil and water high impact and shock resistance as well as resistance to side pressure application in cut-off and roughing wheels high elasticity in fine-grinding wheels enables high surface qualities high wear resistance difficult to dress high thermal conductivity
vitrified bond
resin bond
M G
WZL/Fraunhofer IPT
Seite 17
Outline
1 Fundamentals
Motivation of Simulation
Conclusion
WZL/Fraunhofer IPT
Seite 18
Motivation
Product
innovative reliable cost-effective
Components
high quality high precision long product life
Manufacturing process
economic reproducible flexible Continuously increasing demands of the market lead to increasing requirements for manufacturing processes.
WZL/Fraunhofer IPT Seite 19
Motivation
For controlling and improving the grinding process a high degree of technological knowledge is required. A high potential to increase the process knowledge and the process optimization is based on modeling and simulation the grinding process.
WZL/Fraunhofer IPT Seite 20
Goals
increase of the process comprehension increase of the process knowledge
Today
without process simulation
4 12
FEM-Calculation
specification sheet product concept design and choice of material lay-out manufacturing aspects
76 weeks
prediction of the process stability prediction of the component characteristics Reduction of planning- and development steps Cost reduction
WZL/Fraunhofer IPT
16
8
design manufacturing planning manufacturing work piece-test
Seite 21
24
manufacturing tests
12
Goals
increase of the process comprehension increase of the process knowledge prediction of the process stability prediction of the component characteristics Reduction of planning- and development steps Cost reduction
WZL/Fraunhofer IPT
Future
4 12
14
FEM
lay-out
54 weeks
design
12
simulation
manufacturing planning
12
manufacturing
work piece-test
Seite 22
Goals
increase of the process comprehension increase of the process knowledge prediction of the process stability prediction of the component characteristics Reduction of planning- and development steps Cost reduction
WZL/Fraunhofer IPT
Requirements
High result quality
Seite 23
Number of articles
500 400 300 200 100 0 1970- 1975- 1980- 1985- 1990- 1995- 20001974 1979 1984 1989 1994 1999 2004
WZL/Fraunhofer IPT
Seite 24
Outline
1 Fundamentals
Motivation of Simulation
Conclusion
WZL/Fraunhofer IPT
Seite 25
Perception
WZL/Fraunhofer IPT
Seite 26
Modeling
Sensorium
IT-based converting
WZL/Fraunhofer IPT
Seite 27
Basic of reasons
The particular Reality
Inductiv
Deductiv
Inductive reasoning
Deductive reasoning
Derivation of reason from theoretical derived models
WZL/Fraunhofer IPT
Seite 28
Definitions
Model
A model is an abstract system that corresponds to a real system and is used for expensive and/or impossible - investigations, - calculations and - explanations- or demonstration purposes. It delivers general information about elements, structure and behavior
Simulation
A simulation is a replication of a dynamic process in a model.
WZL/Fraunhofer IPT Seite 29
microscopic
heuristic and empirical models are limited and difficult to transfer from one process to another
physical
Finite Element models are complex to apply and the necessary material properties are often not known Molecular dynamics are very fundamental fundamental models can be regression models with physical background kinematics models can be used for applicable simulations
x, x, x
heuristic
empirical
x x x xx x x xx x x x
Source: CIRP Keynote Paper 2006, Brinksmeier et al. WZL/Fraunhofer IPT Seite 30
Molecular Dynamics (MD) kinematics Finite Element analysis (FEA) fundamental regression artificial neural nets rule based
Source: CIRP Keynote Paper 2006, Brinksmeier et al.
WZL/Fraunhofer IPT Seite 32
mx+bx-cx = CU0sin(t)
x, x, x
x xx xx xx x x x xx
CPU needed
knowledge needed
highly
amount of data
maintenance + development
Outline
1 Fundamentals
Motivation of Simulation
Conclusion
WZL/Fraunhofer IPT
Seite 33
Advantage
Large 3D atomistic modeling and simulation Good combination with other model types
Disadvantage
Enormous CPU-power and long calculation time needed Unknown suitable material potential functions and parameters Less further development
WZL/Fraunhofer IPT Seite 34
Limitation: Long calculation time Unknown suitable material specific potential functions
Source:CIRP Keynote paper 2006, Brinksmeier et al WZL/Fraunhofer IPT
Groove scratching with 2 grits (side/top view - 360 000 time steps, 144 ps)
Seite 35
Kinematics
Based on the grinding wheel and workpiece geometry and the process kinematics, significantly different kinematic models following varied approaches have been developed. There are two basic kinematic approaches, and another two grinding models, kinematicgeometrical models and kinematic-empirical models.
Advantage
Low starting effort High convergence of modeling results to real grinding processes
Disadvantage
High knowledge needed Long calculation time Only trends not absolute values of results
WZL/Fraunhofer IPT Seite 36
Types of models
Octree-model
Steric partitionated model Segmentation of the space into eight smaller subspaces Homogeneous areas can be described by big spaces with less memory From a certain depth of the tree of the octree the memory requirements will be very big
Dexel-model
Illustration of the model with pins By combination of several blocks in different axes it is possible to display complex bodies Reduction of the memory requirements, accuracy and velocity compared to the octree-model
modelfaults
Y X
WZL/Fraunhofer IPT
Y X
Seite 38
Kw
z y
Assumptions:
1. workpiece is fixed 2. grinding wheel moves around the workpiece
Ks
xs
workpiece
grinding wheel
WZL/Fraunhofer IPT Seite 39
P
rw g rs zs V a xw
zw
yw b
Kw
z y
rw
= Ts, w
rs
Ks
xs
transformation matrix
cosb cosg
sina sinb cosg - cosa sing sina sinb sing +cosa cosg sina cosb 0
cosa sinb cosg + sina sing cosa sinb sing - sina cosg cosa cosb 0
x y z 1
Seite 40
Ts, w =
D V 0T 1
WZL/Fraunhofer IPT
nothing or just workpiece workpiece behind grinding wheel modify front part of the dexel delete dexel create dexel modify back part of the dexel Grinding wheel behind workpiece just grinding wheel
Grinding wheel
2 3 4 5 6 Workpiece 7 8
WZL/Fraunhofer IPT
Seite 41
Simulation-step
Dexel-models Diffusion calculation Metal removal volume Force-model
CL-file
Kinematics of the machine Movement of the axis
Actual movement
Rough discretisation
Target movement
Dynamic displacement
Machinemodel
WZL/Fraunhofer IPT
Seite 42
Movements of axis
WZL/Fraunhofer IPT
Diffussion calculation
Fi [N]
Advantage
Little experience needed High transferability to other processes
Disadvantage
High computational effort and high CPU-power needed Limited by current measuring techniques within the process High effort for experiments and data analysis
WZL/Fraunhofer IPT Seite 45
Boundaries
Two-dimensional model Linear moving heat source
Contact length lg Table speed vw Workpiece Fixed at the table lg Adiabatic surface
Heat source qw
Temperature-independent thermal material properties The surface of the solid is adiabatic Bottom surface is set to 20 C Maximum temperature of the coolant lubricant is tB = 120 C due to the boiling point of emulsion
Workpiece
Table speed vw
0 = +20 C
WZL/Fraunhofer IPT Seite 46
426C 10C
Material 100Cr6 (HRC 62) Grinding wheel B181 LHV 160 Grinding parameters vw = 12 m/min Qw = 40 mm/mms vs = 160 m/s Coolant lubricant Emulsion (5%)
Austenizing temperature was not reached during the simulation of different grinding processes. Therefore, it is assumed that no phase transformation will take place.
WZL/Fraunhofer IPT
Seite 47
WZL/Fraunhofer IPT
Seite 48
Fundamental
Fundamental analytical approaches (FA) develop predictive models that are deductively derived from basic physical interrelationships. Based on the knowledge of a process and the selection of appropriate physical quantities, physical models can be developed using mathematical formulations. FA can provide a good way to illustrate insights of the grinding process.
Advantage
Low computer power needed A good way to illustrate insights of the grinding process
Disadvantage
Limited to the ranges of the different parameters Depending on the accuracy of input parameters High effort for experiments and data analysis
WZL/Fraunhofer IPT Seite 49
Regression
Regression analysis (RA) is the generic term for any mathematical statistical method that aims to find a functional interrelation between dependent random variables (output parameters and the machining results) and one or more independent random variables (input parameters).
Advantage Low computer power needed Vast field of possible application Big potentials Disadvantage Depending on the effort and number of experiments Limit application for a developd model Extensive coefficient calculations
WZL/Fraunhofer IPT Seite 50
The mathematical part of modeling is enhanced by using new and more complex polynomials obtaining a higher quality of simulation.
Empirical models were compared to newly developed modeling techniques such as artificial neural networks and fuzzy set theory.
N kin =
h=
F '=
c gw
1 q
e1
e1 ae 2
e1 ae 2
c gw
c gw cwp
1 q
e1
e1 1 2 d eq e1 1 2
N St = n z
1 q
e1
ae e2
d eq e 1 3 d eq
z max =
Rt =
e e e e1 e2 e3 e4 e5 Aae 6 vw 7 deq 8 z A1 ae vw vc d eq e
fundamental roughness
c gw cwp
A + c gw cwp
Rz =
1 e2 a e q e 1 2 e1 a e q
e1
1 d eq vc e3
e3
V 'we4
Seite 52
WZL/Fraunhofer IPT
Advantage
Be able to handle so-called soft input parameters without a numerical value Easily adaptable to different problems Be able to handle slightly incomplete data bases
Disadvantage
Relies heavily on the quality and number of data sets for testing Amount of data
WZL/Fraunhofer IPT
Seite 54
Source: CIRP Keynote paper 2006, Brinksmeier et al. WZL/Fraunhofer IPT Seite 55
Advantage Takes the knowledge into account which is based on human experience High simulation quality Easy to combine with other model approaches Disadvantage Sophisticated knowledge base needed Difficult to transfer to other grinding processes as the rule sets are unique for every grinding process
WZL/Fraunhofer IPT
Seite 56
Outline
1 Fundamentals
Motivation of Simulation
Conclusion
WZL/Fraunhofer IPT
Seite 58
Conclusion
In grinding several types of models are used:
x x x xx x x xx x x x
physical (kinematics, MD, fundamental, FEA) empirical (regression, ANN) heuristic (rule based)
They describe important process parameters (wheel topography, workpiece roughness, forces, temperatures, etc.). Every model has advantages and disadvantage when simulating the grinding process. A model can not take into account both the macroscopic and the microscopic view.
WZL/Fraunhofer IPT
Seite 59
Temperature gradients:
106 C/s / 10 3 C/mm
Forming speeds:
up to 107 1/s
vS
vW
WZL/Fraunhofer IPT Seite 60
Good description of special problems little development input necessary for easy problems
Exact formulation of the context often impossible high development input necessary
WZL/Fraunhofer IPT
Seite 61
Backup
Kinematics
Two Basic kinematic approaches describe surface grinding and/or cylindrical grinding processes using two-dimensional model grains, by calculating the roughness, or the topological cross section, of a workpiece and the chip thickness. Kinematic-geometrical grinding models foucs on the geometric penetration of workpiece and grinding wheel. kinematic-empirical models has similar output parameters to above models, but different way of calculating the process characteristics and results.
Applications in grinding : Modeling of the grinding wheel topography Modeling of the surface profile and the surface roughness of workpiece
Source:CIRP Keynote paper 2006, Brinksmeier et al
WZL/Fraunhofer IPT Seite 64
Coordinate transformation
WZL/Fraunhofer IPT
Seite 65
Multistage grinding
high material removal rates and good surface finish are exclusive roughing multi-step machining with different set-values finishing spark-out the dressing strategy must be adapted different grinding wheel specifications can be used for roughing and finishing
Qw
Ra,z
=0
30% 0%
70%
Vw tc
15% 60% 25%
Seite 66
WZL/Fraunhofer IPT