Before we start with simulation, we need to collect data.
First the difference between attibute data and variable data.
OK / NOK 21,7° C
0 / 1 19 sec
Good / False 2m 66cm
Minimum sample 100 Minimum sample 30
From this analogy, one can see that variable data provides more information than
attribute data. As input for our simulation projects we prefer to receive variable
Variation is our enemy, but a need for simulation accuracy. Still we should aim to
keep the variation for measurement as small as possible.
Known as the 5Ms and 1P, these sources of variation are common to all processes,
whether manufacturing or business-based.
Machines – The equipment used in the transformation of inputs into outputs, e.g.,
a computer that turns various sources of information into an organized report.
Materials – The elements transformed from inputs to outputs, e.g., the paints and
canvas used by a painter.
Methods – The procedures that transform inputs into outputs, e.g., a standard procedure
for billing and collections.
Measurement – The tools that monitor a process’s performance, e.g., a doctor’s reading
of a patient’s blood pressure.
Mother Nature – The environmental elements that influence the ability to meet a customer
need or requirement, e.g., regulation of temperature and humidity in a paper warehouse.
People – The staffing that influences the ability to meet customer needs and requirements,
e.g., the number of operators per shift
Variation in a process may be caused by any one of these factors or by the interaction
of two or more, such as machines and materials.
Populatie vs Sample
A population is the entire group of objects from which information is needed for
a statistical study. A sample is the group of objects selected from the population
from which data is gathered for a statistical study.
We can consider the relationship between the population and the sample in two ways:
Physically, the sample is a subset of the population.
Statistically, a properly selected sample should mirror the population.
Sample size should be small enough to be affordable, and large enough to accurate.
GOOD MEASUREMENT LEADS TO….
Accuracy – how close the indicated value is to the true value
Stability – ability to produce the same results over time
Linearity – difference in accuracy across the range of measures
Repeatability – variability in values when everything else is held constant
Reproducibility – variability in values with different people measuring elements
under the same conditions
THINGS TO CHECK…
- Is your data validated?
- Is the method of measurement OK?
- Is the data un-interpretable?
- Only include data intended to use.
VALUABLE INFORMATION FOR SIMULATION?
· Products, demand
· Timing : set-up, proces, MTBF, MTTR
· Controls en priority
· Rework?? Hidden workflow??
You can only use input which is available and accurate.
Good to know is that a complexity level of 20 % already returns a model accuracy
of 80 %.
Did you know that…
CAD files, JT files, JPEG’s, etc can be imported for more realistic and personalized