thermodynamics of friction
TRANSCRIPT
Thermodynamics of friction
Noa Mitsui
1
Contents
1. Introductions ・tectonics & earthquakes ・theory of earthquakes ・relation between earthquakes and friction
2. Purpose: thermodynamic modeling of rock friction
3. Research ・introduction: rock experiments ・thermodynamic modeling
4. Summary
2
Tectonics & earthquakes in the world
• Plates move relatively at several cm/yr
• Japan locates on some boundaries
• Many earthquakes occur around Japan
Plate movements and their boundaries
Earthquake epicenter (M5.5<) 3 M= -3.2 + 0.67log10E (E: energy)
Tectonics & earthquakes in the world
• Plates move relatively at several cm/yr
• Japan locates on some boundaries
• Many earthquakes occur around Japan
Plate movements and their boundaries
Earthquake epicenter (M5.5<) 4 M= -3.2 + 0.67log10E (E: energy)
Tectonics & earthquakes in the world
• Plates move relatively at several cm/yr
• Japan locates on some boundaries
• Many earthquakes occur around Japan
Plate movements and their boundaries
Earthquake epicenter (M5.5<) 5 M= -3.2 + 0.67log10E (E: energy)
Difficulty in understanding of earthquake
(Source area= ruptured area estimated by detected seismic waves)
earthquakes distribution (seismicity) along plate boundaries
Eurasian Plate
Pacific Plate
Philippines sea Plate
N. American Plate
yr
yr
Tokyo Fukushima
Nagoya
6
• People want to know “WHEN & WHERE” earthquakes will happen
Difficulty of the forecast
Incompleteness of theory
Difficulty in understanding of earthquake
(Source area= ruptured area estimated by detected seismic waves)
earthquakes distribution (seismicity) along plate boundaries
Eurasian Plate
Pacific Plate
Philippines sea Plate
N. American Plate
yr
yr
Tokyo Fukushima
Nagoya
7
• People want to know “WHEN & WHERE” earthquakes will happen
Difficulty of the forecast
Incompleteness of theory
Difficulty in understanding of earthquake
(Source area= ruptured area estimated by detected seismic waves)
earthquakes distribution (seismicity) along plate boundaries
Eurasian Plate
Pacific Plate
Philippines sea Plate
N. American Plate
yr
yr
Tokyo Fukushima
Nagoya
8
2011
Cross section along the plate movement
(1964-2007 & 2011, M5<, USGS)
• People want to know “WHEN & WHERE” earthquakes will happen
Difficulty of the forecast
Incompleteness of theory
Incomplete theory = elasticity
It cannot explain small earthquakes (=seismicity)
occurring in the plate ↓
realistic deformation theory is necessary (= motive)
• current bases of seismology: All strain released by slip at contact surface plate boundary
||
Schematic diagram of deformation at plate boundary
trench
strain increase
tsunami
rebound 9
All strain released
Slip only at boundary
Only in theory
Main theme of today’s talk
☆ How we can obtain deformation theory of medium? (bases of the earthquakes occurrences) fracture and deformation in the earth’s crust
Research object: friction law in rock experiments ・earthquakes are compared to frictional slips along the tectonic plate boundary ・easiness to understand: rock experiment > earthquakes I will focus on the rate- and state-dependent friction law (Dieterich, 1979)
10
Sand particles in frictional layer between rocks
• Even in the small frictional layer, sand particles are deformed, crushed, compacted.
• These phenomena probably causes the complexity of friction
• The mechanism of friction can be related to that of earthquakes
11 (Mair & Marone,1999)
Before dislocation After dislocation
1mm
Sand particles in frictional layer between rocks
• Even in the small frictional layer, sand particles are deformed, crushed, compacted.
• These phenomena probably causes the complexity of friction
• The mechanism of friction can be related to that of earthquakes
12 (Mair & Marone,1999)
Before dislocation After dislocation
1mm
2011
• wide distribution along the plate boundary
• Magnitude: M9.0-M2(detectable limit)
Wideness of earthquakes
The mechanism can be related to them
2011
Cross section along the plate movement
Magnitude-frequency diagram
Gutenberg-Richter law (Power law)
freq
uenc
y
Magnitude (1924-2007, M5<, JMA)
(1964-2007 & 2011, M5<, USGS)
13
← Another model of G-R law called as “cumulative Gompertz distribution” can have the same entropy formula as that in heavy ion physics (Biro et al, 2015)
Main theme of today’s talk
☆ How we can obtain deformation theory of medium? (It causes the earthquakes occurrences) fracture and deformation in the earth’s crust
Research object: friction law in rock experiments ・earthquakes are compared to frictional slips along the tectonic plate boundary ・easiness to understand: rock experiment > earthquakes I will focus on the rate- and state-dependent friction law (Dieterich, 1979)
14
Contents
1. Introductions ・tectonics & earthquakes ・theory of earthquakes ・relation between earthquakes and friction
2. Purpose: thermodynamic modeling of rock friction
3. Research ・introduction: rock experiments ・thermodynamic modeling
4. Summary
15
Outline of rock experiments
• Static & dynamic friction are tested with various condition:
rock type, temperature, humidity, normal stress with/without sand at the interface
16
Load (velocity controlled)
stress stress
Frictional interface
rock
(Dieterich, 1981)
Servocontrolled to constant
Static friction increases with time
• Static friction ∝ log (time)
• Friction decreases to dynamic one with displacement
(Marone, 1998) 17
Feature1/3: Stable dynamic friction ∝ log(slip velocity)
• Rock samples loaded with constant velocity
• Friction weakens or strengthen with velocity depending on the tested condition
ex. Interface sand without sand: weaken with sand: strengthen
(Marone, 1998)
Dyna
mic
fric
tion
Slip velocity
Velocity weakening
18
logV dependency
Feature2/3: frictional jump ∝ log(velocity step)
(Dieterich, 1981)
• Loading velocity step causes instantaneous jump of dynamic friction
• Following the jump, the friction transitions to the stable one
• Both jump and relaxation depend on ln(velocity step)
velocity (µm/s)
2.5
19
∆F<0
Velocity weakening
Velocity weakening case
25
fric
tion
Feature2/3: frictional jump ∝ log(velocity step)
(Dieterich, 1981)
• Loading velocity step causes instantaneous jump of dynamic friction
• Following the jump, the friction transitions to the stable one
• Both jump and relaxation depend on ln(velocity step)
Dc
velocity V1
20
Velocity weakening case
V2
fric
tion
bln(V2/V1) aln(V2/V1)
Feature3/3: symmetric relaxation with constant displacement
(Dieterich, 1981)
• Both case of velocity increase and decrease has frictional jump and relaxation with constant displacement (Dc)
vel. increase: jump up
vel. decrease: jump down
fric
tion
Dc
velocity (µm/s)
2.5 2.5 25
21
Velocity weakening case
Dc
Empirical equation of the friction law
μ: shear stress t: time V: slip velocity μ0:reference friction for V0 Dc: characteristic slip distance θ: state variable a,b: positive parameters
Empirical equations of rate- and state-dependent friction law
(Dieterich, 1979)
• Neither of them can reproduce both (static & dynamic) friction. ↓ • We will reproduce the experimental result with thermodynamics.
Simulated frictional change with displacement
V2=10µm/s V1=1µm/s V1=1µm/s
22
aln(V2/V1) bln(V2/V1)
−=
LV
LV
dtd θθθ lnor
(Ruina, 1983)
friction = dissipative event -> thermodynamic issue
: Static friction Dynamic friction
① ②
Previous thermodynamic model (Mitsui and Van, 2014)
Rock sample
external force
Start point
permanent displacement
Rock sample
xm
23
damping force
( )
amics thermodynof law 2nd0)( :balanceentropy
22:entropy
:energy total
:motion ofequation
22
≥−=
−−=
==
−=
rkrVFUST
krmVEUS
VFxFE
FFxm
d
ee
de
↑ internal energy
↑ elastic modulus
Result of linear model (Mitsui and Van, 2014)
fric
tion
velocity (µm/s) 2.5 2.5 25
(Dieterich, 1981)
Improvement of thermodynamic model • Important feature ・introduced elements 1. instantaneous jump 1. partial steady state aln(V2/V1)
2. following relaxation bln(V2/V1) 3. symmetric relaxation with disp. 2. logarithmic increment Dc
3. factor of slip reduction in elastic displacement
24
③
② ①
rVr
rxddr
dtdr
αα −=
−→
)(
( ) 22
22
2200
22
krVVmEU
krmVEU
−−−=
↓
−−=
( ) )/ln( 00 VVVV µµ ∆→−∆
(Van, Mitsui, Hatano, 2015, arXiv:1501.04608v1)
based on thermal activation theory (chemical reaction: Eyring, 1936; rock friction: Nakatani, 2001)
Improved constitutive equations
25
( )rlVVlrlVVlVr
rlVVl
κακ
κµ
2021
2021
1201
)/ln(1)/ln(
)/ln(
−+−
=
−=∆↓
(Van, Mitsui, Hatano, 2015, arXiv:1501.04608v1)
( ) 0/ln)(
0)( :balanceentropy
000n
n
≥−
−∆+=
↓
≥−=
rVrrVVV
AUST
rrVAUST
ακµµ
σ
κµσ
1 2 3
1. partial steady state 2. logarithmic increment 3. factor of slip reduction
Improvements
constitutive equations:
Vr
dxdr
dtdr
=→
cf. Ruina law (1983)
Results
Data comparison
26
( )rlVVlrlVVlVr
rlVVl
κακ
κµ
2021
2021
1201
)/ln(1)/ln(
)/ln(
−+−
=
−=∆
mLllblalµ
κα
20/102.0,015.0
8.0,1
221
121
=======
==
(Van, Mitsui, Hatano, 2015, arXiv:1501.04608v1)
Dieterich law
Ruina law (α=0, κ=1)
(α=8, κ=0.85)
α→larger: asymmetric relaxation
Static friction
• log(t)-like increase in static friction ⊿μ is reproduced • Both dynamic & static friction are reproduced with
one set of constitutive equations
27
(Marone, 1998)
002.0=κ
074.0=κ
Data ●: ■: 074.0
002.0==
κκ
(Van, Mitsui, Hatano, 2015, arXiv:1501.04608v1)
Summary • First motive: to understand mechanism of earthquakes ↓ • Current motive: deformation theory of medium under stress
object: static & dynamic friction of rocks (quasi-thermostatic condition) Results: 1. friction at reference velocity can be explained by using partial steady-state condition of internal energy 2. logV dependencies of friction are reproduced by the logarithmic increments of entropy production 3. symmetric relaxation can be explained by introducing a factor of slip reduction in elastic evolution 4. Above three elements reproduce log-t increase of static friction Advantage: simple model can be applied to geophysics with upscaling Future plan: physical bases of assumptions in this model, etc. 28