Percentile values in Fill analysis results

Percentiles are the set of numbers from 0 to 100 that divide a set of ranked data into 100 class intervals with each interval containing 1/100 of the observations.

A particular percentile, say the 5th percentile, is a cut point with 5% of the observations below it and the remaining 95% of the observations above it. Thus for an X percentile, at least X% of the observations are below this value.

Percentile results are available in the analysis log for a Fill analysis.

For example, in the Filling phase results summary for the part:

• Bulk temperature-95th percentile, suggests that at least 95% of the part volume is below or at this temperature.
• Bulk temperature-5th percentile, suggests that at least 5% of the part volume is below or at this temperature.

Using this result-95th and 5th Percentiles

The maximum and minimum values of variables are well defined. However, the 95th and 5th percentile values are more useful in the development of molding windows and for comparing results. The 95th and 5th percentile values are the relative position (weighted by volume) in an ordering of the values of a variable. In other words, 95% of the part or runner volume has a value of the variable below the 95th percentile value. Similarly, only 5% of the part or runner volume has a value of the variable below the 5th percentile value.

The 95th and 5th percentile values reflect the overall range of variation of the variables for the majority of the part or runners. The absolute maximum and minimum values may sometimes be misleading. For example, a few small, thin ribs on an otherwise uniform and relatively thicker part will result in a very low minimum temperature at the end of the filling stage. Using the 5th percentile value, the colder area on the rib can be excluded. This provides a more realistic lower bound for the temperature on the part. Similarly, the 95th percentile value can exclude the effect of excessive viscous heating occurring only near the gate region, and provide a better upper bound of the temperature on the part. Since the 95th and 5th percentile values provide a better idea of the variation of the particular variable under consideration, these values should be used to evaluate design considerations and to compare results between different designs.