To get the best results out of predictive maintenance, predictive results should be immediately and automatically transferred to an airline’ or MRO’s inventory management system, notes Micheál Armstrong, CEO of Armac Systems, an inventory software provider that is owned by SR Technics. Typically, this transfer would be done by sending predictions to the maintenance management or ERP system, which would then transmit them to inventory tools.
That is not happening now, Armstrong says, chiefly because, apart from engines, predictive maintenance is so young, used by only a few airlines and for a few parts. Armstrong argues that providers of predictive tools must incentivize more early adopters to accelerate use of their systems.
Don’t predictive tools pay users in avoided AOGS and delays? Yes, eventually. But Armstrong says there are a lot of trials and errors in the early stages of adoption, and these deter some airlines from using predictive maintenance.
This should change over time, and then tying predictions into inventory systems will be a natural next step. Armstrong divides part demand forecasting into four types: planned, preventative, probabilistic and predictive. “The future opportunity will be to feed demand signals from predictive maintenance to supply chains for fulfillment . . . we have commenced integration but we are still in the prototyping phase.”
Inventory optimization itself plans on three time scales, strategically, tactically and for actual fulfillment. Predictive maintenance would trigger the last segment, planning for fulfillment. But even here there are challenges. “For example, how certain is the requirement?” Armstrong asks. False positives associated with No Fault Founds can falsely increase demands for part replacements. And there are tradeoffs. “How do you make the best supply-chain response to a possible requirement in the next 10 days versus a certain requirement in five days with one part on the shelf?”
Inventory optimization can be used for rotables, repairables, consumables and expendables. It can include different decisions, for example exchange versus buy or repair versus not repair. When prediction, integration and inventory responses are done correctly, Armstrong says major gains will be possible: improved utilization of stocks, reduced logistic costs for urgent AOG shipments, reduced AOGs, better technical dispatch, and, in Europe, reduced flight compensation.