Python flying at 40,000 feet - part 1

Presentation by Marko Vasiljevski given at PyData London 2016

Transcript

4. Python flying at 40,000 ft Marko Vasiljevski & Raffaele Rainone
5. Instead of ‘Click here to add title’ Flight data monitoring – what? Flight data analyst in 1 minute Engine health monitoring Problem statement Raff on the stage - solution (hopefully!)
6. Flight Data Monitoring (FDM) Airlines worldwide obliged to do FDM Spotting deviations from safe operation Non-punitive Confidential
7. Quick Access Recorder Flight Data Recorder Data Acquisition Unit
8. How?! Take-off, fly, land Copy data to servers Data cleansing (“numpy masked arrays”) Data analysed by software Events detected
14. Engine performance monitoring N1 N2 Fuel Flow Exhaust Gas Temperature
15. Engine performance monitoring Fuel Flow Exhaust Gas Temperature
16. Engine performance monitoring
17. End of take-off flow
18. End of take-off flow
19. End of take-off flow
20. End of take-off flow
21. End of take-off flow
22. End of take-off flow
23. End of take-off flow
24. End of take-off – a problem If not detected properly – problems with engine health Difficult to do deterministically, i.e. with classical signal processing What if a parameter fails (happens not so rarely)
25. End of take-off – a solution Time at vertical navigation mode selected (VNAV) Drop/rise in fuel flow (FF) and/or exhaust gas temperature (EGT) Search for changes around that point in time Gather the knowledge Calculate probability for EOT for unseen flight
26. End of take-off – a solution Time at vertical navigation mode selected (VNAV) Drop/rise in fuel flow (FF) and/or exhaust gas temperature (EGT) Search for changes around that point in time Gather the knowledge Calculate probability for EOT for unseen flight
27. Hkjk;l