UNLOCKING THE SECRETS Patterns in Data. Techniques Plotting Log-log, semi-log, scaling Statistics Frequency domain (fft) xmgrace xy plotting tool Derivatives.

UNLOCKING THE SECRETS

Patterns in Data

Techniques

Plotting Log-log, semi-log, scaling

StatisticsFrequency domain (fft)

xmgrace xy plotting tool

Derivatives and integrals – hare/lynx

Dimensionless numbers Reynolds number, prandtl, etc

Why and When to Use

Particularly important for emergent behavior projects (emergent math)

Can be used for experiments and for calculation analysis

Discriminator for top projects Appears to be an improvement vs Improves the result with a 98% confidence level

Statistics

Averages or MeansProbabilities and Random NumbersRegression and CorrelationANOVA or t-tests (determine whether

samples are from one population or two)Bayesian – (advanced, hard!)

Pressure Record from Explosive Test

Second curve is impulse -- integration of data Shows ‘energy’

Frequency analysis is also useful (60 hz, tube diameter)

Predator-PreyHare/Lynx

Integration

Derivatives

Birth Rates, Death rates

Frequency Analysis

Period

Dimensionless Numbers

Concept is to pick combination of terms that eliminates units Reynolds number Prandtl number Mach number

Used in experiments to scale down size

Additional Thoughts

Comparison to experimental data is important Looking at only experimental data or calculational

results yields little insight – The great leaps in learning occurs when you compare the calculational results to experimental data

Verification and Validation is critical for a model Running a model for a known simple case to validate

it is a critical step