-
Thermal fatigue degradation effects occurred at austenitic T
connections:
• cyclic feeding (Civaux, FR)
• valve leakage (GKN, DE)
Potential consequences?
• surface stresses
• crack initiation
• stresses in wall
• crack propagation
stratification
Type 2
convectioncold
coldflow induced turbulences (eddy)
Type 1
main pipeflow
Type 3
Type 4
cold
cold
hot
flowleakage
orflow
hot mixing area
cold
Figure 1: Turbulent mixing effects in piping system T
connections
-
1. Collation of field experience
5. Evaluation and development of road map
4. Verification tests3. Analysis
Integrity evaluation and testing
2. Thermal load determination
Figure 2: Work packages flow chart
-
Figure 3: THERFAT consortium
Name Country Organisation Wilke, U. Germany E.ON Faidy, C.
France EDF Le Duff, J. A. France FANP-F Braillard, O. France CEA
Cueto-Felgueroso, C. Spain Tecnatom Varfolomeyev, I. Germany FHG
Solin, J. Finland VTT Schippers, M. Germany FANP-D Stumpfrock, L.
Germany MPA Nilsson, K.-F. Netherlands JRC Vehkanen, S. Finland FNS
Seichter, J. Germany SPG Abbas, T. United Kingdom CINAR Figedy, S.
Slovakia VUJE Carmena, P. Spain ENDESA Cizelj, L. Slovenia JSI
-
Temperature loads
Dead-end configuration
Small leakageTurbulent temperature fluctuation in branch
Civaux event„swinging streak“
∆T range
T T∆T < 2 x 150 K∆T = 2 x 150 = 300 K
t t
Figure 4: Field experience on high cyclic turbulent temperature
mixing
cold
hot
T with an inclined dead-end configuration
-
Figure 5: SPG, glass models test matrix
Dimensions of T Objective Parameters Remarks Status
50 x 50 (90°-T) Flow Visualisation Various Flow Directions and
Mass Flows
50 x 50 (45°-T) Flow Visualisation Various Flow Directions and
Mass Flows
70 x 24 (90°-T) Flow Visualisation Various Flow Directions and
Mass Flows
Tests finished
100 x 100 (90°-T) Flow Visualisation Flow Direction A Mass Flows
see table below
Tests at Room Temperature Variation of Fluid Density
A B Tests
finished
50 x 50 (90°-T) Electric
Conductivity Measurement
Main Flow in kg/s: 2 and 4
Leak Flow in kg/s:
0.03, 0.06 and 0.12
Tests at Room Temperature Variation of Fluid Density
Tests finished
Main Mass Flow in kg/s Leak Mass Flow in kg/s
DN 100 x 100 (di = 100) 20 10
0.015 0.03
-
Figure 6: SPG, glass model, electrical conductivity
measurement
2555
105155
34540
0
12312911
5 4 8 6 7
10
LeckstromH
aupt
stro
mleak flow
mai
n flo
w
-
T and flow orientation Main mass flow
s/gk in m&
Leak mass flow
s/gk in m&
Temperature difference (hot – cold
water)
K in T∆
Circumferential measurement
position
Status
DN 50 x 50 (di = 48)
3,9 1,95
0.015 0.03 0.06 0.12 0.23
90 45 6 ... 12 o’clock
Tests finished
DN 80 x 20 (di = 78 x 20)
5,5 2,75
0.015 0.03 0.06 0.12
90 45 6 ... 12 o’clock
Tests finished
A B
Steel models (pipe wall thickness 1 mm) Test matrix
Figure 7: SPG, steel models, test matrix
-
Figure 8: CEA, Fatherino II experiment, test rig
The THERFAT mock-up
Fatherino facility overview
Locations for sensors- flux meter
- mix sensor fluid + wall- fluid thermocouples
Disk with rulerto check the
angular position of the mixing branch
-
TC1 (272 µm)TC2 (1375 µm)TC3 (2575 µm)
THERFAT mock-up: 50 mm diameter (7,11 mm) 304L
x
z
0
zy
40°40°
2 locations by section
∅ int = 54 mm
∅ ext = 73 mmSection
CEA flux meter
Fatherino II instrumentation
Figure 9: CEA, Fatherino II, test configuration
-
Figure 10: SPG, steel model, turbulent-temperature load
spectrum
THERFATExample: turbulent-temperature load spectrum in
branch
Measured temperature ranges – rain-flow evaluation
Vertical T 50 x 50 (wall thickness = 1 mm)
Measuring position 6 o‘clock
Secondary pipe, fluid temperature
Mai
n P
ipe;
T1
Secondary pipe; T2
Measured temperature ranges – rain-flow evaluation
Vertical T 50 x 50 (wall thickness = 1 mm)
Measuring position 6 o‘clock
Secondary pipe, outside wall temperature
0
10
20
30
40
50
60
70
80
90
100
1 10 100 1000 10000
Number of Cycles with Range greater than TR
TR-T
empe
ratu
re R
ange
(% o
f ∆T)
0
10
20
30
40
50
60
70
80
90
100
1 10 100 1000 10000
Number of Cycles with Range greater than TR
TR-T
empe
ratu
re R
ange
(% o
f ∆T)
-
THERFAT – WP 2.2 Deliverable D8Thermo-hydraulic tests on steel
models (50 x 50 and 80 x 20)
• Steady flow in main pipe - one leg locked (closed valve) but
leakage
• Temperature difference ∆T (main flow – leakage) up to 90 K
• Temperature measurement outside and inside the wall (thickness
1 mm)
Results • Temperature alterations, load spectra (percentage of
∆T)
• Mean heat-transfer coefficients found by inverse temperature
calculation
• Report BLP-SB/27-04
T and flow orientation Temp. alterations Heat-transfer
coefficient
DN 50 x 50 (di = 48)
Dead leg: > 90 %
Main flow: ≤ 70 %
Dead leg: ≤ 4000 W/m²K (A) ≤ 7000 W/m²K (B) Main flow: ≤ 6000
W/m²K (A)
≤ 10000 W/m²K (B)
DN 80 x 20 (di = 78 x 20)
Dead leg: negligible
Main flow: ≤ 70 %
Dead leg: no relevant information
Main flow: ≤ 10000 W/m²K
A B
Figure 11: SPG, steel model, test results
-
Figure 12: SPG, glass model, test results
THERFAT – WP 2.2 Deliverable D8Thermo-hydraulic tests with glass
models (50 x 50 and 100 x 100)
• Steady flow in main pipe - one leg locked (closed valve) but
leakage
• Temperature difference ∆T simulated by different specific
fluid densities
• Electrical conductivity measurement
Results: • “Temperature” alterations (percentage of ∆T)
• Report BLP –SB/50-04
T and flow orientation “Temp.” alterations
DN 50 x 50
Dead leg: ≤ 80 %
DN 100 x 100
Dead leg: ≤ 40 %
-
Figure 13: SPG, CFD benchmark analysis experiment/CFD
analysis
THERFAT – WP 2.3 Deliverable D10/D11CFD benchmark calculation by
Technical University of Dresden (TUD)
Density versus time in the leakage pipe at 6-o’clock
position
Distance 0.5 d Distance d Distance 2 d Distance 3 d
Summary
- Qualitative agreement with test results (large peaks at low
frequency between small amplitudes)
- Time period covered by calculation: 12 s (decay time
forstart-up effects in tests about 100 s)
- Relation costs/benefit too large
Main flow: 2.1 kg/s
Secondary flow: 0.03 kg/s
-
Figure 14: Benchmark of CFD analysis SPG, FANP-D
SPG FANP-D
CFD – Benchmark calculation in WP 2.3
TU Dresden
DN 50:50
