Big data in Aircraft Maintenance

Big data, a term we hear around us more and more nowadays. But what is big data, what underlying principles are being impacted by it, how can we utilize big data and its underlying principles in the aircraft maintenance industry and ultimately how can we benefit from it? Questions to which in my coming posts I will try to shed some light on.

Big data, or in other words, information of extreme size, diversity and complexity (this is a definition used by Gartner INC, one of the world's leading institutes on IT) revolves around the capability of organizations (or any other institute for that matter) to:

  1. Timely access and retain data;

  2. Analyse and interpret data;

  3. Base decision making on data

Timely access and retain data

Your organization produces data (e.g. company ERP system), your employees produce data (e.g. email or social media), your suppliers produce data (e.g. their ERP system), your suppliers employees produce data, your customers produce data, the employees of your customers produce data, the world produces data (..the internet..).

Data is being produced 24 hours a day, 7 days a week throughout your company's value chain. The challenge here is to be able to access this data (information), access it in a timely manner (it does not make sense to have access to data of last year only two years afterwards) and have the computing power to retain it.

Analyse and interpret data

Having access and being able to retain all this information has no value whatsoever unless you have the organizational capability to analyse and interpret this data. Having this capability starts with having adequate talent and tools in your organization to perform such analyses. In addition an organization needs to have the governance structure to allow proper data interpretation.

Aspects such as data quality and available funds to support the infrastructure come into play here. A note for adequate talent has to be made here as well; talent does not revolve around having an army of “data scientists” in a staff organization providing endless reports to management. Talent revolves around having people in key-places in your organization that are able to analyse and interpret data once it is provided to them.

This is where we will notice the shift in years to come, as collecting and providing data to management becomes more and more automated and emphasis starts to move to interpretation (e.g: What does this particular graph tell me?).

Three items need to be distinguished:

  1. Predictive analyses; use available information to identify expected future trends or outcomes

  2. Behavioural analyses; use available information to create models to drive e.g cost reductions, product changes, innovation, customer satisfaction or quality improvements

  3. Real-time analyses; use available information to drive decision making in the here-and-now

None of the above matter if we don’t use the analyses and interpretation of information available to us to drive (strategic) decision making and exploit it to our benefit.

This is best illustrated by an example I have seen being used on different occasions:

Whiskey and ice is bad for you liver, Vodka and ice causes short term memory loss, Pepsi and ice damage your teeth. Hence ice is bad for you and henceforth banned from your drinks.

Big data and its three principles in Aircraft Maintenance

So what can we do with these three principles of big data in our aircraft maintenance environment and how can we put it to good use? 

So what can we do with this phenomenon of big data and its three principles in Aircraft Maintenance? Before I go into the answer of this question, let's first of all get rid of the obvious answer:

Sure, we can gain operational efficiency and reduce cost if airlines and maintenance companies would use data more effectively but this is a narrow minded, “stuck-in-the-here-and-now” view of things. It completely misses the fact that big data in Aircraft Maintenance has the potential to fundamentally change the industry as a whole! In order to see the fundamental change we need to look beyond the borders of our own airline or maintenance company, and take a helicopter view on the industry.

Imagine that, for whatever reason (most probably for reasons of efficiency), most of the industry’s data is collected real-time into a platform (timely access & capacity to retain data). Now add the capacity to analyse and interpret all this data and you have a fundamental change taking shape. Let’s illustrate this based on an example:

An Example 

Flight EX-1234 from New York JFK to Amsterdam-Schiphol, during flight the on-board maintenance systems indicate that one of the flight control computers is failing and needs to be replaced. Already during he flight this is transmitted to the platform at which point it knows that a flight control computer is required, with an engineer to replace it, on Amsterdam airport for flight EX-1234 on its Estimated Time of Arrival. Now suppose that the platform would have access to the airline’s parts inventory to determine if a spare Flight Control Computer is available, where it is located and the estimated transit time to Amsterdam to get this part there. Now let’s take it one step further, and assume it also has access to the parts inventory of all other part suppliers, airlines and MROs. This enables the platform to check global availability of the flight control computer, the price and how soon it can be delivered in Amsterdam. In short, it would be able to completely automate the buying decision for airlines based on lead-time to delivery, quality and costs.

The latter example is based on a unscheduled maintenance situation, just imagine the impact on all those scheduled purchases that take place in the industry (safe to say that it would be around 70% of total purchases done in the industry). Now let’s take a look at the engineer that we need in order to replace the flight control computer. The same principle applies here, it would be able to look into the staff availability within the airline’s rostering and determine if a suitable engineer can be made available on Amsterdam at the ETA of flight EX-1234 and what it would cost to get him there (if not already on the airport). Additionally, it would be able to source with other maintenance providers to see if they have the capacity to provide a proper licensed engineer to replace the flight control computer and list costs, quality and performance. Thus, also in this case, the planning cycle is fully automated. Again, this is an unscheduled scenario, just imagine the possibilities for scheduled maintenance. Hopefully by now you realise that we have eliminated the human factor all the way up until the moment the actual flight control computer needs to be physically replaced on the aircraft.

Conclusion

Elimination of most human involvement until the moment of actual maintenance being done, enables numerous efficiency gains and process streamlining: production of spare parts being based on actual need rather than commercial drive, full accuracy in aircraft record keeping, provisioning of maintenance capabilities based on actual need rather than competition, optimal use of tools & equipment on airports, proper aircraft lifecycle planning to reduce our environmental impact, and many additional positive effects you might identify yourself. Obviously all these improvements result in significantly lower maintenance costs and improved operational efficiency. 

However, foremost as an industry, it would allow us to eliminate a large portion of human error (either fatal or not-fatal) by reducing the actual human involvement in all maintenance related processes.

Previous
Previous

Part II : Big data in Aircraft Maintenance

Next
Next

Continued Airworthiness 2030