Reducing Costs - Optimizing Value - Suggest Using Alternate Methods
Degraded systems cost directly related to "quality escape" reduces revenues and profits; identifying degradation early allows for scheduled interruptions minimizing disruptions to production.
Non-redundant critical part failure induce safety incidents.
Summary if you look at a "cloud" - in linear world you only see a square or a rectangle, all interesting features are removed.
Due to simplifying assumptions many systems break unexpectedly disrupting production introducing additional costs in re-planning, scheduling, personnel, tools and parts acquisition.
Alternate methods increase visibility into root cause methods, in turn, increasing Reliability, Safety, in turn decreasing revenue generation disruptions.
Non linear metrics enable better visibility into system dynamics and degradation patterns.
For instance in the aviation domain:
Jet Engine Turbine Data (from NASA PHM Data Set)
Non Linear similarity metrics reveal root-cause systemic patterns, that are invisible to std. Correlation.
MTBF - Mean Time Between Failure of systems, subsystems and parts
RUL metrics enables Logistics inventory bins - how many of which part needs to be in warehouse, across the world to support a specific product design.
Statistical closed form mathematical models, like Weiibull and many models, fit the data to the model. In my professional experience, using Gaussian (or any other traditional model) to smooth out experimental data rarely conforms to reality based experimental findings. The engineering community sometimes forgets to validate their models against reality.
Data Driven RUL - Remaining Useful Life provide insight into actual End-of-Life dynamics.
Inventory costs dominate overall systems costs across a design lifetime. Parts degrade at different rates dependent on stresses degrading system over time. Systems can fail if one critical part fails.
Comments
Post a Comment