Content uploaded by John Masse
Author content
All content in this area was uploaded by John Masse on Jun 12, 2019
Content may be subject to copyright.
Public
Ecalibration Methodology Example
Enhanced Calibration for parameters is a useful
tool when the simulation is too time consuming
to perform calibration by numerical optimization.
It reinforces the human intelligence in the
expertise, knowledge and explanation of the role
and influence or cumulate influence of
parameters on simulation outputs.
It gives us instantly the Ecalibration forecast
parameters to use for your model.
Task
Run parameter
variation around
nominal value P0
Example:
Np+1 simulation = 3
Ecalibration allows to perform calibration of numerical models
'parameters against reference data (experiment,
simulation,…), in a very short timeframe.
From uncertainties intervals of parameters, Ecalibration
provides new forecast parameters 'values within a confident
range and the cumulated effect of parameters on the output.
(Statistical Characterization)
Objective is to prevent of time consuming numerical simulation
by decreasing the error between the simulation and the
reference with the criteria defined by:
1
2
0
+∞ 𝜀2𝑑𝑡, with 𝜀 = 𝑦𝑒𝑥𝑝𝑒 − 𝑦𝑠𝑖𝑚𝑢
Contact
http://www.ecalibration.fr/
Ecalibration at appedge.com
User Benefit
oNo additional simulation using CFD software
oNo additional programing is required and none modification of the
model/simulator too.
oFast analysis of parameter influence and weight on simulation outputs
oStatistical characterization of the parameters with to respect to the outputs.
oFast exploration of numerous designs of physical and mathematical models
oLink with many CFD software ( Converge, …)
oPhenomenological analysis and numerical conditioning (Solver parameters).
o….
Fields Cover by Ecalibration
oAnalytic sensitivity and statistical analysis ( Sort parameters by influence and correlation)
oSpray model calibration (https://api.convergecfd.com/wp-content/uploads/Yohan-
Blacodon_Gasoline-Spray-Model-Calibration-Under-Diesel-Engine-Like-Conditions.pdf)
oCombustion model calibration (ECFM, ECFM3Z,…)
oTransient design
oInjector map generation
o…
Task
Analyze parameters ‘ influences on
the simulation output and assess
inter-parameter correlation.
By Example: The parameter
khact_c_tcav are similar to
discharge_coeff at 20.06 %
khact_c_tcav explains 40.98 % of Spray_Penet95
discharge_coeff explains 4.57 % of Spray_Penet95
Forecast Ecalibration
Task
Find new parameters values and relative
variation from nominal one.
Parameters P0 Pecalibration
Variation
Trust on the
variation
value
khact_c_tcav
0.01
0.014029
40.29%
99.99%
discharge_coeff
0.81
0.80954
-
0.0564%
91.2%
Replay of simulation
Task
Replay of simulation with
Ecalibration parameters
Error reduction
=
982.2%
Calibration Process
Pre-treatment RMS khact_c_tcav & discharge_coeff
on Spray-Penet95