Future processes will embed more technology and may become smarter, but they’ll remain processes. This trivial reminder is a hint that use cases for Six Sigma will still exist. Capabilities and deviations must be mastered, drifts must be controlled and there will be needs to pinpoint the few influent parameters in an even greater ocean of many.
With more automation and the potential faculty to mass-manufacture unique products, the traditional - and sometimes convenient - excuse of lack of repetition in high mix low volume production, hence lack of statistical significance should vanish.
Smart automated processes should produce more data and faster to which smart objects will add theirs, creating a permanent data flood. New parameters will be measured by embedded or affixed sensors, like temperature, moisture, acceleration, orientation, pressure, brightness, etc.
That’s where Six Sigma could need to evolve as this data flood will enter the Big Data world.
Analysis of such mass of data will not be done with usual Six Sigma statistical techniques but with scenario correlation analysis. The new statistical management and control models will be multi-criteria base, both because the techniques allow it and the new processes require it.
Those techniques are said to give predictive and self-learning abilities to machines, smart material and all objects associated; eg. molds, tools, test benches, ovens, etc.
Design of experiment, originally designed to reduce the number of required experiments to the bare minimum in order to isolate the few influent parameters, could go obsolete as High Performance Computing takes over. The latest uses low cost resources (massively paralleled multicore processors) to calculate at high speed large amounts of data, thus allowing exhaustive explorations of several scenarios.
Among the three TLS (ToC, Lean & Six Sigma) approaches, Six Sigma is by its nature the one most likely to be passed over to the machines. Future machines and equipments will take over, individually or in collaboration, statistical controls, real time analysis, self corrections, and so on.
If I’m right, Black Belt and Master Black Belts expertise can be passed over to machines, maybe even to smart objects. The need for experts will be reduced to those engineering the embedded Six Sigma intelligence.
In this future nevertheless, problems will still happen and the way to understand the causes and solve the issues will not change fundamentally. New tools will show up, but good old DMAIC, PDCA, Pareto or fishbone diagrams will still be used.
Despite possible transfers to smart devices, problem solving will still require intuition and the kind of reasoning (say gut feeling) machines should not have so soon. Therefore, until artificial intelligence demonstrates human-equivalent abilities, some (human) earthlings will still be required in smart factories.
| < Précédent | Suivant > |
|---|
