Achieving 85% or more for the OEE measurable requires not only equipment uptime and availability, but it must be consistently producing good parts.
So how do we go about defining what is a measure of consistency?
Should you have done the course on Statistical Process Control (SPC), this is where all those good things you learnt start to come into play. The feature we need to understand here is Process Capability.
In everything we do in life, no matter how hard we try, there is always variation, even if it is small, it is still variation. In manufacturing, we need to control that variation in order to maintain a stable process.
A stable process is a predictable process - using statistics we can predict how capable and stable this process is on an ongoing basis given no “special causes” that disrupt the capability.
If you are designing a new process or adding additional equipment to an existing process, you need to conduct a capability analysis. Nearly all the big problems I have assisted in resolving quality concerns on were due to the lack of APQP run@rate capability trials. When purchasing equipment it is imperative that this measureable is declared in the tender process to your prospective suppliers. They must demonstrate the capability of the equipment as part of the equipment buy-off procedure. Believe me, if you don’t ask for it, you won’t get it! If you ask after the equipment has been commissioned, most good machine builders will offer, but this will then come with a time delay and most times with additional cost. Hence the inclusion in APQP (Advanced Product Quality Planning).
Process capability is measured using the Cp/Cpk and or Pp/Ppk index.
So again with your data collection sheets, collect the measurement results needed to calculate the Potential Process Capability Ppk. The accepted minimum standard for a Ppk analysis is a Ppk of 1.67 or better. However should this characteristic under study be rated a Critical or Significant characteristic, then the customer within his rights, may request a Ppk of 2.0.
Without fully teaching SPC, here is the crux:
Voice of the Customer (VOC) / Voice of the Process = Cp, Pp
Standard Deviation | P(%) | Ppm | Cp |
± 3,75 S | 99,982 | 178 | 1,25 |
± 4 S | 99,994 | 64 | 1,33 |
± 4,5 S | 99,9994 | 6 | 1,50 |
± 5 S | 99,99994 | 0,6 | 1,67 |
± 6 S | 99,9999998 | 0,002 | 2.0 |
Remember this table...
A Cp of 1.33 shows a probability of producing 64 defects per 1 million parts produced.
A Cp of 2 is basically Zero defect territory.
Therefore another quantifiable target for a Robust Process is Cp=>1.33 and/or Pp=>1.67
So now we have a capable process, but is it accurate?
Remember that Cp/Pp determines capability without measuring to target. Cpk/Ppk measure capability to target value (xbar).
Next up - maintaining a robust process….
Written by:
Matthew Woodford (ht+a Trainer & Consultant)
Comentarios