The FAA contracted with Lockheed Martin to create a brand new computer system for enroute controllers called En Route Automation Modernization (ERAM).
ERAM is slated to replace the aging HOST computer systems that run the current enroute radar displays, and is part of the FAA’s grand NextGen scheme, an overhaul of the country’s air traffic control system; a large part of which hasn’t even yet been clearly defined.
However, in spite of the fact that the FAA insists the ERAM program is on time and on budget, there have actually been numerous delays and lots of problems.
My facility (Minneapolis Center, ZMP) was last scheduled to go Initial Operational Capability (IOC) sometime last November, but that date was pushed back because of problems and a moratorium on testing over the holidays. (IOC is the first live use of ERAM on air traffic at the respective facilities and is a significant date with respect to the contractor deadlines.)
As of today we’re now slated to go IOC on February 5.
Controllers have already received training on the new ERAM system, and needless to say not too many are thrilled about the prospect of working real air traffic with it.
That’s mostly because they know the system doesn’t work very well.
From the beginning of our training we were told that every time they fix bugs in the software, they create new bugs. And there have been lots of bugs fixed, only to be replaced by other bugs.
We keep testing the same functionality, over and over because because they have no confidence that an update hasn’t broken what they already had working.
So controllers are more than a little concerned about the project especially since they’re the ones left holding the bag if and when the system crashes.
However it doesn’t seem like anyone else at the FAA or Lockheed Martin is that concerned about all the problems.
(To reassure us we’ve had managers tell us, “Remember when they first put in the HOST computer system; it crashed all the time too!”, like that’s supposed to make controllers feel better about using ERAM…)
This weekend Salt Lake Center (ZLC) is doing an “Ops Run” that is supposed to last for eight days that started at 5 AM this morning CST.
However, tonight there are already problems only 16 or some hours into the test, and ZMP controllers were told to verify any flight plan coming from Denver Center (which abuts Salt Lake Center) with the pilots because they’ve discovered that the flight plan information coming from ZLC’s ERAM computer system may not be accurate.
The flight plan information is the route of flight for each aircraft, that is supposed to be passed via computer from one facility to the next. Controllers are able to separate airplanes safely because they know where airplanes are going next.
However, if the flight plan information is erroneous, then they don’t know where aircraft are going next, which causes obvious problems. If an aircraft takes an unexpected turn there can be serious consequences.
So the FAA already knows that ERAM at ZLC isn’t working properly some 16 hours into an eight day test; possibly not processing flight plan information properly, but they’re going to continue using it, and essentially telling controllers downstream to “watch out”.
This isn’t the first test for ERAM with live traffic at ZLC, and it certainly won’t be the last. But it seems pretty clear that ERAM still isn’t fit for use with live traffic even yet.
But the FAA is going to roll the dice with it anyway. As the project slips further and further behind schedule, it makes one wonder what problems ERAM would have that the FAA would consider to be serious enough to suspend the testing.
The fact is that originally only two facilities were supposed to be “key sites” for the project too (Salt Lake Center and Seattle Center):
The FAA chose two sites to introduce ERAM, instead of the usual single key site. This is because the system requires facilities to access each other’s flight data, says Dan Watts, the FAA’s ERAM program director. By bringing up two facilities, the FAA also can see how flights are handed off between them.
The plan is to achieve initial operating capability (IOC) at Salt Lake City and Seattle next month. ERAM will be evaluated and tested for about three or four months, and when the results are analyzed, “we will decide whether we are ready to go ahead and turn [ERAM] on nationwide,” Watts says.
The idea was that they would test ERAM at the two key sites and make sure everything was working properly before deploying it elsewhere. No other facilities were supposed to be going IOC until the key sites had met important testing benchmarks.
Of course that was the plan about a year ago, when ERAM was closer to being on schedule (and on budget) than it is now.
But with all the problems and resulting delays, the deadlines weren’t being met, so instead of waiting for the key sites to get all the bugs worked out and testing ERAM over the course of several months, they decided to just expand the project anyway.
With a wave of a magic wand, Minneapolis Center (ZMP) became a “supplemental” key site, which means that they don’t have to wait for the two key sites to successfully complete their operational evaluations before expanding the deployment of ERAM.
So despite the fact that there are lots of problems with ERAM, they’ve changed the original plan and are putting it into use at more facilities anyway. In spite of all the problems that have delayed the program significantly, now they’re actually accelerating the deployment of ERAM!
Since the FAA keeps saying the ERAM project is on time and on budget, when it clearly isn’t, it’s pretty obvious they’re willing to make bald-faced lies about the project. They know ERAM still isn’t ready for use with live traffic, but are willing to test it with live traffic anyway.
And keep in mind that when the FAA says, “Safety was never compromised“, what they’re really saying is that safety likely was compromised but no one died.
Repeat after me (and keep your fingers crossed, kids), “Safety was never compromised…”