The nearly worthless DOT&E annual report has been released and one of the very few sets of actual data is for the Ford EMALS, AAG, and elevator systems. Let’s take a look at the data.
Note: Reference is made in the quotes to ‘ISE’ which is an abbreviation for Independent Steaming Event which is part of a series of at sea, post-delivery tests and trials. It’s basically one of a sequence of test periods at sea. The number indicates which test/event it is in the sequence. So, for example, ISE 12 would be the twelfth at sea test period.
EMALS
DOT&E reports this data for the Ford EMALS,
During the 8,157 catapult launches conducted through ISE 18, EMALS achieved a reliability of 272 mean cycles between operational mission failures (MCBOMF), where a cycle is the launch of one aircraft. This reliability is well below the requirement of 4,166 MCBOMF. [1, p.144]
Compare this to the preceding year’s report,
During the 3,975 catapult launches conducted post PSA through ISE 11, EMALS demonstrated an achieved reliability of 181 mean cycles between operational mission failure (MCBOMF), where a cycle is the launch of one [2, p.137]
What is not clear is whether the phrase ‘through ISE 18’ refers to the total since the beginning of the program up through ISE 18 or whether it refers to the incremental change since the last DOT&E report (2020) which was ISE 11. If it refers to the cumulative lifetime total then the results are skewed by the initial higher failure rate. If it refers to the incremental number since the previous report then we can get a good feel for the current performance.
My interpretation is that the numbers are cumulative. This means we can, therefore, take the difference between the current (2021) figure of 8,157 launches and the 2020 figure of 3,975 launches, do some simple arithmetic and get the performance for the last year (ISE 13 through ISE 18) which, presumably, should show substantial improvement over the initial performance and give us a good look at the current status.
Simple arithmetic gives us the actual number of failures and allows us to assemble the following table of data.[3] Calculated data is indicated as such. Remaining data is from the DOT&E report.
Launches |
Failures |
Actual MCBOMF |
Required MCBOMF |
8,157 (through ISE 18) |
30 (calculated) |
272 |
4,166 |
3,975 (through ISE 11) |
22 (calculated) |
181 |
4,166 |
4,182 (calculated - 2021) |
8 (calculated) |
522 (calculated) |
4,166 |
We see, then, that the performance for last year (2021) showed substantial improvement from the initial failure rate of 181 to 522, however, it is still woefully short of the required 4,166.
Okay, it’s still very bad but if it keeps getting better we’ll be okay, right?
Unfortunately, no. As with any developmental effort, progress and improvement is rapid in the initial stages and then slows quickly as the easy improvements are made and progress becomes increasingly more difficult. Even if progress were to continue in a linear fashion at the current rate – say, improving from 181 to 522 each year – it would take just over 12 years to get from a failure rate of 522 to 4,166 … but, as I just explained, that’s not how it works. Barring a developmental breakthrough, I just don’t see EMALS ever being able to meet the failure rate (MCBOMF) or even come close.
Advanced Arresting Gear
For the AAG system, the current DOT&E report gives,
During 8,157 recoveries, AAG achieved a reliability of 41 MCBOMF, where a cycle is the recovery of a single aircraft. This reliability estimate falls well below the requirement of 16,500 MCBOMF. [1, p.144]
And, from the previous year’s DOT&E report,
Through the first 3,975 recoveries, AAG demonstrated an achieved reliability of 48 MCBOMF, where a cycle is the recovery of a single aircraft. This reliability estimate falls
well below the requirement of 16,500 MCBOMF. [2, p.137]
Treating the data as was done for EMALS,
Traps |
Failures |
Actual MCBOMF |
Required MCBOMF |
8,157 (through ISE 18) |
199 (calculated) |
41 |
16,500 |
3,975 (through ISE 11) |
83 (calculated) |
48 |
16,500 |
4,182 (calculated - 2021) |
116 (calculated) |
36 (calculated) |
16,500 |
The data shows that we can only recover around 40 aircraft before experiencing an operational failure. So, if we were to launch a full air wing strike, only half the aircraft could recover!
There are two things that are absolutely stunning about this data:
1. The failure rate has barely changed from the initial data. For practical purposes, there has been no improvement from the start until now! In fact, the previous year’s failure rate is actually slightly worse than the initial rates!
2. The difference between the actual MCBOMF of around 40 and the requirement of 16,500 is staggering. No amount of further development is going to eliminate that gap and, as with EMALS, the easy improvements have already taken place and progress from now on will be slower. Yikes!
Elevators
For the Advanced Weapons Elevators (AWE), we have only one set of data. From the DOT&E report,
Therefore, only preliminary reliability estimates are available to compare to the requirement of 932 hours between operational mission failure. Through the first 14,842 elevator cycles, 68 operational mission failures were reported. [1, p.145]
Again, simple arithmetic gives us the following:
cycles = 14,842
failures = 68
failure rate = 218 versus that target of 932
While this tells us that the performance is still well short of the requirement, it is at least numerically conceivable that the gap can be closed with further development. On the other hand, with the amount of effort and man-hours that have already gone into this, it is hard to imagine that there is much more improvement left to be had.
Summary/Conclusion
The basic function of a carrier is to arm planes (weapon elevators), launch planes (EMALS), and recover planes (AAG). The Ford is currently unable to perform those basic functions at anywhere near the required level of reliability which means the Ford is unable to perform its basic function.
How do we get from where we are to the requirement? Disturbingly, for the reasons discussed, it likely can’t be done. We’re well into the diminishing returns portion of the development curve. In other words, we’re entering the flatter portion of the development curve and improvements will be fewer and less impactful.
So, what’s the alternative? What can be done? Well, we could rip out the systems and install something proven like the Nimitz systems but that’s almost prohibitively expensive and difficult since the Fords are not designed for those systems. Instead, I’m going to predict that the Navy will do what they always do when they can’t meet a specification … change the requirement! I predict the Navy will accept the current performance and simply change the specification. And just like that, they’ve met the spec!
As bad as all this is, the really scary part is that we’re in the process of building three more identical carriers with, presumably, the same problems. We are building an entire class of carriers that can’t perform their basic function.
The only slight hope is that the engineers have learned from the Ford and significantly modified the equipment for the subsequent carriers so that they will be able to perform their basic function. Unfortunately, this is yet another example of concurrent construction. We’re building multiple ships before the first one is complete and working properly so the EMALS, AAG, and AWE for the subsequent carriers were ordered long ago, before there was time for the development feedback loop to function and, therefore, the equipment is almost certainly identical, for all practical purposes. Same equipment … same problems.
Of course, given DOT&E’s new practice of concealing almost all data, it’s likely that we’ll never see any data for the other carriers so we’ll have only the Navy’s official statements to inform us … and those will be glowing, without a doubt, as the Navy’s official pronouncements about the Ford have been. In fact, I’m quite surprised that DOT&E exposed even this data.
On the plus side, China has committed to EMALS for their carriers so they’ll likely have the same problems!
_________________________________
[1]Director, Operational Test & Evaluation, 2021 Annual Report, Jan 2022
[2]Director, Operational Test & Evaluation, 2020 Annual Report, Jan 2021
[3] Launches / failure rate = number of failures
8,157 / 272 = 30 fails
3,975 / 181 = 22 fails
Taking the difference between those two data sets
4,182 / 522 = 8 fails