99% Reliable 100% of the Time: How an airline meets amazing reliability metrics under…
An airline operational environment is one of the toughest in the world for maintenance and reliability. What makes it tough is the level of complexity involved and in the level of reliability that must be consistently maintained. The cards are definitely stacked against running an efficient and reliable airline.
The equipment is highly mobile
The airline does not have parts or mechanics everywhere the equipment might go
Operations are affected by numerous events other than maintenance
Operating components to failure is not an option
Maintenance procedures are Federally mandated and controlled
Mechanics have strict licensing requirements
So, how do you keep the aircraft operating without interruption? Redundancy, Predictive Maintenance, Preemptive Maintenance, Scheduled Downtime, constant attention to condition, tracking everything – even things that you might not consider to be important, believing what the statistics tell you unless there is specific evidence to the contrary, and don’t assume ‘cause and effect’ without specific, verifiable proof thereof.
Aircraft are designed with built-in safeguards. Backups on the backups on the backups. Let’s look at DC electrical power on a Boeing 767 for example. First, there are starter generators on both engines, each capable of supplying all the power the aircraft needs. There is also an Auxiliary Power Unit in the tail with an identical starter generator on it. If all three of those are lost, in the left wheel well there is an electric motor and a hydraulic pump connected together on the same shaft If you put hydraulic pressure to the pump, you get DC power – or you can put DC power to it and get hydraulic pressure out of the unit. Okay, so what if we lose that also?
A bit further back behind the wheel well, there is a RAT – a Ram Air Turbine. It is normally retracted flush with the fuselage and hidden from view, but can be extended into the air stream when it is needed. It has a unit similar to the one in the wheel well, but with a propeller mounted on it, so when it is extended into the air stream, the propeller turns which spins the generator / pump. If you happen to lose that also, well, all that leaves is the three onboard batteries.
Redundancy. It helps with reliability.
Predictive Maintenance, Preemptive Maintenance, and Scheduled Downtime
If you don’t plan your maintenance, your equipment will plan it for you. Trust me on this, your equipment is not very good at planning. In order to plan your maintenance, you need to be able to not only predict when things are going to fail, but how.
What his takes is constant attention to condition, tracking everything – even things that you might not consider to be important, believing what the statistics tell you unless there is specific evidence to the contrary, and never assuming ‘cause and effect’ without specific, verifiable proof.
So – what do you track and how? On an aircraft, we track almost everything. I can tell you with a high degree of accuracy when and how an item is going to fail, and in some cases, even where it is going to fail. Okay – so what about things that are ‘unpredictable’? What about things that don’t have a discernable pattern? What about something like – tires and brakes?
I can tell you about how many landings I get before a tire wears to limit. I can tell you which brand of tire is going to last longest. I can tell you how often tires are found with low pressure and how many times they are found flat. I can tell you how many times tires have foreign object damage. I also track this by calendar time, by cycles, and by location. So, can I accurately predict when an aircraft is going to have a flat? Nope. Not accurately. Then why track it?
Let’s see what I do know from this information. I know that:
I will have higher incidents of low tire pressure in the Northeast in the winter.
I will have a dramatic increase in warped brake disks in the Northeast in winter.
I will have a dramatic increase in foreign object damaged tires in the Northeast in the spring and summer.
I will have a marked increase in tire wear in the south in summer.
I will have a marked increase in brake problems in the south in summer.
While this information doesn’t give me solid numbers to predict failure on, they do tell me where I need to position tires, wheels, and brakes to ensure that my aircraft aren’t stranded on the ground waiting for parts.
So – now you’re asking – what causes these spikes in usage? Yes, I know the answer to that also. Tires give the appearance of being low when they are cold. The northeast is cold in winter, and pilots tend to write up more tires for appearing low. Pilots are not allowed to check tire pressure, so all they can do is write them up.
The increase in warped brake disks in the northeast in winter is also simple. By the time the pilot slows the aircraft enough to pull onto a taxiway, the brake rotors are hot. Quite hot, actually. Taxi the aircraft through snow or a puddle of cold water that splashes up on the brake rotors, and things happen. Splash ice cold water on hot metal – and it warps.
The increases in tire damage in the northeast in spring and summer? Simple. That’s when construction happens in the northeast. Lots of construction activity, meaning lots of construction materials lying around and potholes and such.
Once you have the data and understand it you can take some action to improve your reliability. For example, you can intentionally throw things out of phase. Let’s say I have a pump that typically fails between 5,025 and 5,050 hours of operation. Because of redundancy, I have two identical pumps on the aircraft, and the aircraft can easily operate with just one pump.
Let’s also assume that I start with two new pumps, each with zero time. At 2500 hours, I would schedule one of the pumps to be replaced with a new, zero time pump. The pump I replaced would go back into stock as a ‘time continued’ item and would be used as a spare. The reasoning is that I would normally change the pumps at 5,000 hours, well before the predicted failure time. By replacing one pump at half-life, I have now given my self a safety net, and can run the pump that I didn’t replace to 5,020 hours, knowing that the second – backup – pump will only be at half life, so if the higher time pump fails I have a reliable backup. When I replace the higher time pump with a zero time pump, I still have the same situation – one pump nearing it’s life limit and one at half life. The chances of both pumps failing at the same time in this situation are slim.
If both pumps were at the same ‘phase’ I would change the pumps at 5000 hours as my safety net would not be there. By keeping the pumps out of phase, I gain at least an additional 20 hours of operation on every pump without increasing the risk of a dual pump failure.
“R.A.I.” the Reliability.aiTMChatbot
You can ask "R.A.I." anything about maintenance, reliability, and asset management.