Lecture 14 Multipass Evaluation Scheme

1
Lecture 14 Multi-pass Evaluation Scheme
Engr 691, Section 10University of
MississippiFall 2003
2
Multi-pass Evaluation

• Multi-pass evaluation flow chart

Start
Determine initial vector map S0n n1,2,???N
Process and evaluate PIV recordings to obtain
vector map Sn
S0nSn
NO
NO
Convergence criteria fulfilled?
Too many iterations?
YES
YES
Stop
3
Multi-pass Evaluation

• Multi-pass evaluation schemes
• - Window shift techniques
• - DWS CWS for correlation-base interrogation
  to reduce errors
• - CDI for both correlation-based interrogation
  and image pattern tracking
• to reduce bias error
• - Image correction techniques
• - For both correlation-based interrogation and
  image pattern tracking to
• reduce random error resulting from particle
  image pattern distortion
• - Multiple interrogation grid
• - Iteration started with lower grid density to
  reduce computation time
• - Multiple interrogation window
• - Iteration started with big window for high
  reliability
• - Iteration ended with small window for high
  spatial resolution
• - Multiple search radius
• - Search radius reduced in latter iteration
  cycles to avoid big errors

4
Multi-pass Evaluation

• Convergence criteria of multi-pass evaluation
• - Convergence factor
• - Maximal iteration number
• In case of convergence factor vacillation
  above the convergence criteria
• a number limit is set to stop the iteration.

5
Multi-pass Evaluation

• Methods helpful for iteration convergence
• - Detect erroneous vectors replace them with
  interpolation
• - Smooth the vector map
• - Loosen the iteration procedure

Loose factor ? selected between 0 and 1
6
Synthetic PIV image

• Particle image simulation

7
Synthetic PIV image

• Single pixel random noise simulation

- Determined by given mean (Im) and standard
deviation (Ir)
Histogram (PDF)
Image sample
8
Synthetic PIV image

• Low frequency noise simulation

- Determined by given amplitude (A) and
wavelength (?x, ?y)
9
Synthetic PIV image

• Gray value distribution in synthetic images

- Root-sum-square (RSS) of Ip, Isp and Ilf
10
Synthetic PIV image

• Simulated flows

Rotation potential
Pipe flows
11
Synthetic PIV image

• Brownian motion simulation

Sb
- Determined by given intensity
12
Homework

• Practice with EDPIV software
• Sample DOUBLE-EXPOSED RECORDING
• Sample SINGLE-EXPOSED RECORDING PAIR
• Familiar with particle image simulation tools
• Prepare at least 1 topic for discussion in next
  class