Manufacturers publish the envelopes that their aircraft or components can operate in. This can be in simple terms, such as extending the flaps up to 180kts doesn’t require a maintenance inspection, or in slightly more complex terms, such as the engines can operate up to 120% torque, so long as they do it for less than 20 seconds, and the propeller is spinning at less than 90% speed. The important thing to note for the purposes of this discussion though, is that what to monitor, and the thresholds to put on it, are very well defined for the maintenance side. For the safety management side, there is often no such rigid guidance.
Determining thresholds for safety events can be done in a few ways, basing thresholds:
- Relative to SOPs,
- Relative to the spread of previous data for that measure
- Relative to other operators with similar fleets and missions
Basing event triggers on SOPs gives a general idea of where the thresholds will be. FDM systems allow users to discriminate on severity of an exceedance by using different levels. Now, there are different schools of thought about which level should be based on the SOP. For this example, let’s assume that the SOP states “Below 10,000ft, the airspeed shall be below 250kts”. This is quite a common SOP, and in some countries and airspace, a law too. In this case, a common approach is to set a level 1 threshold at the 250kt SOP, and to set a level 2 threshold slightly higher, for example 265kts. The level 3 threshold is the one which will be of most significance to most FDM programmes; a point at which some kind of action, for example crew contact, needs to happen should be determined. In short, a level 3 event should require some kind of action or in-depth Flight Data Analysis.
Take speeding while driving for example, many of us have been guilty of it at some point in our lives. But what’s ‘beyond an acceptable margin’? This is the question to ask yourself when picking a level 3 threshold. Is 5mph over the limit too much? 10mph? 15mph? You have to determine a point at which you will need to carry out some kind of action, whatever that may be, but also balance that with only triggering a quantity of events that will support your safety management system, rather than flood it with too much work to manage, which could result in you missing something important. Let’s say you decide that 5mph over the limit is too much, would you want to receive a notification every time I drive 5mph over? Well, yes, and no. Sometimes the important thing isn’t the absolute amount by which I exceeded the limit, but the percentage by which I exceeded it. If I drive at 20mph in a 15mph school zone, that could be considered to be worse than driving at 70mph on a 65mph limit freeway. You could decide to set your level 2 and level 3 thresholds as percentages of that particular SOP, for example 110% and 120% respectively, and this is a perfectly valid way to set thresholds.
Sometimes you’re creating an event for which there is no SOP. For these events you can either use experience of flying the aircraft to set the thresholds, but sometimes this may create too many events when you realise the aircraft is not being flown how you expected. If your FDM programme has already accumulated a large enough bank of data, you can analyse the spreads for measures taken on previous flights to set the event thresholds.
This gives you an advantage over starting a brand new FDM programme, because you can get a rough idea of how many triggered events your Safety Management System will be dealing with, depending on the thresholds you pick. In L3Harris’ Flight Data Connect platform, the system will approximate how many of each event level would have triggered in the previous 3 months with your prospective new thresholds.
What if your operation is not performing in line with other operators of the same type? Setting thresholds at intervals based only on your own previous data could easily hide operational traits. This is where data sharing really comes into its own.
One advantage of outsourcing in this case is the access to extra knowledge and experience that you might otherwise not have available. An outsourced provider will be able to help in many ways, not least suggesting which algorithms to use to monitor for your SOPs, as well as what are ‘normal’ thresholds for certain occurrences, and possible exceptions. Access to large amounts of data for numerous operators makes this advice more reliable; looking at measurements across millions of flights, the standard error will be smaller than a sample size of a few thousand. Being part of a data sharing programme has many advantages, but one of the main positives is being able to connect your data into it and compare the data spread and event rates for other aircraft of the same type, across many more flights than your operation alone could fly.
In reality, none of the methods alone set up the strongest FDM programme; a mixture of the three methods will provide the best outcome. Some events may benefit from a more SOP-based approach, and some may benefit from being more influenced by the distributions of previously flown flights. Events and thresholds are not set in stone, and should be where necessary to improve your FDM programme and SMS, particularly where SOPs change or a safety campaign looks to address certain issues.
FDM providers or consultants can provide additional insight when setting event thresholds, and membership of a data sharing programme will aid further. L3Harris offer an outsourced FDM/FOQA solution which links into the IATA Flight Data Exchange programmes to access over 16 million flights for comparative data records. Flight Data Analysts can provide additional insight when setting event thresholds, based on templates derived from industry best practice and out experience. For event monitoring, thresholds, or anything related to flight data monitoring, L3Harris Flight Data Services are here to help.