'Big data' trend analysis?
Thread Starter
Join Date: Jan 2008
Location: US/EU
Posts: 694
Likes: 0
Received 0 Likes
on
0 Posts
'Big data' trend analysis?
Years ago, a client I worked with was in the business of performing data extraction and analysis of manufacturer warranty repairs and maintenance records for auto makers, appliance makers and such. The theory was that they were able to crunch this data, which involved text as well as numeric data, and tease out trends that could result in recalls well before millions of units were in the field.
Not knowing anything about the aircraft industry and what systems are in place to record and track aircraft performance, maintenance and incident data, I was just curious if there are any similar big data type analysis systems in place, either by aircraft or component manufacturers or the government to help spot trends and nip problems in the bud. Couldn't this kind of analysis also be applied to human issues like the recent complaints by some about a decline in piloting skills and an inability to manually fly aircraft?
Not knowing anything about the aircraft industry and what systems are in place to record and track aircraft performance, maintenance and incident data, I was just curious if there are any similar big data type analysis systems in place, either by aircraft or component manufacturers or the government to help spot trends and nip problems in the bud. Couldn't this kind of analysis also be applied to human issues like the recent complaints by some about a decline in piloting skills and an inability to manually fly aircraft?
Join Date: Mar 2013
Location: Siliconia
Age: 63
Posts: 44
Likes: 0
Received 0 Likes
on
0 Posts
Mark,
Maintenance data analysis has been going on since in the late 1980's my employer started using the automatically downloaded digital maintenance data in combination with manually downloaded data recorder files in the analysis of (what we concluded to be) ice crystal effects.
Real-time controller data files (from sub-assembly and flight test programmes) are occasionally cut down to parameters of interest to validate completely new controller algorithms and optimise associated data variations.
There's clearly a great deal of money associated in after market services here is a spread of reading material: -
Health and Usage Monitoring Systems conference
Airbus AIRMAN
Boeing AHM
General Electric Company - Integrated Vehicle Health Management
Honeywell - Trend Monitoring
Pratt & Whitney - Engine Management Programs
Rolls-Royce - Engine Health Management
The provision of operational data recorder files is a problematic one, requiring legal protections; data leakage in view of flight crew rights, across the commercial (let alone public) domain is simply not tolerable. See
The AAIB investigation report for G-YMMM (section 1.19) describes some data mining, and states The provision of QAR data is a sensitive one and this development meant that airlines were able to provide a summary of results, rather than the QAR data itself.
1 n 0
Maintenance data analysis has been going on since in the late 1980's my employer started using the automatically downloaded digital maintenance data in combination with manually downloaded data recorder files in the analysis of (what we concluded to be) ice crystal effects.
Real-time controller data files (from sub-assembly and flight test programmes) are occasionally cut down to parameters of interest to validate completely new controller algorithms and optimise associated data variations.
There's clearly a great deal of money associated in after market services here is a spread of reading material: -
Health and Usage Monitoring Systems conference
Airbus AIRMAN
Boeing AHM
General Electric Company - Integrated Vehicle Health Management
Honeywell - Trend Monitoring
Pratt & Whitney - Engine Management Programs
Rolls-Royce - Engine Health Management
The provision of operational data recorder files is a problematic one, requiring legal protections; data leakage in view of flight crew rights, across the commercial (let alone public) domain is simply not tolerable. See
The AAIB investigation report for G-YMMM (section 1.19) describes some data mining, and states The provision of QAR data is a sensitive one and this development meant that airlines were able to provide a summary of results, rather than the QAR data itself.
1 n 0
Join Date: Feb 2006
Location: Hanging off the end of a thread
Posts: 32,754
Received 2,738 Likes
on
1,166 Posts
Prior to all the electronic systems, in the RAF on the VC10 they used to have a paper trend log book for each Aircraft where the Flight Engineer (remember those) would record all the engine parameters at set periods during the flights, so you had an idea if one was going amiss before it did.
Join Date: Feb 2006
Location: Uk
Age: 59
Posts: 76
Likes: 0
Received 0 Likes
on
0 Posts
There are also monitoring groups in engineering that look at trends in all systems. The manufacturers also monitor and interact with all operators . So in theory any trends should be detected and acted upon .
