OEE and downtime analysis on a six-machine CNC cell, with a kaizen plan

Brenner's valve-body cell has six horizontal machining centres. Each part takes an ideal 4.2 minutes, and the cell runs two shifts of 7.5 hours, five days a week. A tractor maker has offered a contract for 18% more parts from next quarter. The plant manager would rather find the capacity in the existing cell than buy a seventh machine at €420,000.

Nobody knows the real OEE. The shift reports say the cell is "busy", but the monthly output suggests it is closer to 55%. The machines log state changes (running, idle, alarm, changeover, waiting for material, planned maintenance) and the quality station logs scrap and rework. Nobody has ever put the two together.

The Spartanburg plant runs an identical cell and wants to use whatever method you build.

Your job is to: - generate a realistic four-week machine and quality log from the distributions in task 1, since real plant data is confidential; - calculate availability, performance, quality and OEE per machine and per shift; - build a Pareto of losses using the six big losses; - run a root-cause analysis on the top two; - write a kaizen plan, including a changeover reduction study, with a target and an owner for every action.

You do not need a factory. The work is done in Python with pandas, or in a spreadsheet if you prefer. Done looks like an OEE calculation that one command re-runs on new data, a Pareto that the shift leaders agree with, root causes for the top two losses, and a kaizen plan showing how the cell reaches 70% OEE and whether that covers the 18% extra demand.