Statistical Process Capability Calculations for a Group of Rockwell Hardness Readings
A collection of values reflecting Rockwell C scale readings from a heat treated surface, are shown below. The specifications for this process call for 55.0 +/- 2.5. The main goal was to determine the percent in tolerance from the sample as well as its process capability, Cp, and Cpk values.
Grouped data analysis is important because it is a pictorial way of relaying data sets. This helps with communication between operators and the engineering team. Not everyone can interpret the
data, but a picture relating the normal distribution and current distribution of the manufactured products is a common language amongst all.
​
The pictorial chart shown below represents how the data set is all spread out, without any trend, which indicates some sort of corruption within the manufacturing process. The data points go outside of the normal curve, which is apparent when looking at the red normal curve compared to the blue frequency bars. This process incapability is further proven when the Cp and Cpk values were calculated. A good process normal has a Cp value around 1.3, but the ones calculated here are well below 1 meaning significantly defected parts are being produced.
Overall, these analysis techniques help assess the problem sooner than later and keep the quality within the production line high.
