rapid control prototyping for the production of ... · rapid control prototyping source: d. abel...
TRANSCRIPT
WERK Kassel
Rapid Control Prototyping for the Production of Functionally Graded Materials with Tailored Microstructural Properties utilizing COMSOL Multiphysics J. Clobes, M. Alsmann, H.-J. Watermeier, H.-H. Becker, K. Steinhoff
WERK Kassel H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 2
Overview
1 Introduction
2 Extendible Control Architecture for Thermo-Mechanical Processes
3 Inductive Heating with COMSOL
4 Optimization of Process Specific Heating Strategies
5 Rapid Control Prototyping
6 Conclusion and Outlook
WERK Kassel
Tailored Microstructural Components
Wear-optimized tailored components
Load-optimized tailored components
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 3
WERK Kassel
Process Challenges
Heating Strategy
Radiative and
convective losses
Intra-billet heat flux
Source: Me*orm – University of Kassel
Transfer Forming Process Cooling Strategy
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 4
WERK Kassel
FE Modeling
FEM Simulation
Maxwell equations:
Induced current density:
Coupled thermo-electrical system (intra-billet heat flux):
Simplifications:
• 2D axissymmetric model
• Inclination of coils neglected
Geometry Coupled Differential Equations
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 5
WERK Kassel
Model Verification
Specifications
power (Pind) = 18 kW
frequency (f ) = 8 kHz
time (tind) = 12 s
Time [s]
Tem
pera
ture
[°C
]
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 6
WERK Kassel
Matlab and Comsol
Comsol Model Boundary Conditions
and Parameters
Simulation Output
FEM Based Control Architecture
Parameters
Control Parameters
Metal Forming Process
Mapping
SPS Control Architecture
Sensors and Actuators
FEM Verification
1. fem1.sol = asseminit(femmodel,
'init',femmodel.sol);
2. fem.sol=femtime
(fem,'init',fem1.sol, ...
,[…]);
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 7
WERK Kassel
Numerical Optimization of the Heating Strategy Steepest gradient descent Genetic algorithm
Advantages and disadvantages: - Suited for complex, nonlinear problems - Stochastic search - Usually slower
Advantages and disadvantages: - Fast convergence - Requires gradient - Start-value dependent
Initial Population
Random Variation
Evaluation of Fitness
Apply Selection
Evaluate Cost Function
Determine Gradient (dk)
Update Ik[m]: Ik=Ik-1+αkdk
Cost function
Steepest Gradient Descent Genetic Algorithm
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 8
WERK Kassel
above AC3
Strong Core-Surface Gradient Heating Strategy
r
z Time [s]
Tem
pera
ture
[°C
]
Radius [mm]
Tem
pera
ture
[°C
]
Heating Strategy Temperature Profile
0s soaking 2s soaking 6s soaking
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 9
WERK Kassel
Rapid Control Prototyping
Source: D. Abel and A. Bollig, Rapid Control Prototyping, Springer 2006
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 10
WERK Kassel
Fuzzy Logic Controller
Fuzzy Logic Control Online Adaption
"↓$%$& =[█■0&0&0&0&0&0&0@0&0&0&0&0&0&0@0&0&0&0&0&0&0@0&0&0&0&0&0&0@0&0&0&0&0&0&0@0&0&0&0&0&0&0@0&0&0&0&0&0&0 ]
-
.-
No pirior process information
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 11
WERK Kassel
Simulative Evaluation of Cycle Time Impact
20ms cycle time
50ms cycle time
250ms cycle time
Time [s]
Time [s]
Time [s]
Tem
pera
ture
[°C
] Te
mpe
ratu
re [°
C]
Tem
pera
ture
[°C
]
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 12
WERK Kassel
Application Results
Fuzzy Controller development utilizing
Comsol
Fuzzy Controller optimization utilizing Comsol
Direct translation into C-code with no alterations
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 13
WERK Kassel
Conclusion and Outlook
Conclusion - Control of highly nonlinear system - Complex optimization algorithms - Implementation of complex, self adjusting control algorithms
Opportunities: - Highly flexible production system for high precision mass customization - Stable, tailored product quality properties (dimensions, microstructural distribution, etc.) - Adaptable to all thermo-mechanical processes
Future Challenges: - Multi-form flanged shafts - Complex geometries - Mapping to sheet metal forming
Source: Institute of Mechanics –Chair of Numerical Mechanics
H – Dipl.-Ing. J. Clobes – Tel.: +49561 49059541 – Stand: 25.10.2011 – 14
WERK Kassel
Thank you for your attention!