Coming to terms with your parents (hulls)...

What does the historical record say about using parent hulls in acquisition programs?

Sep 23, 2024

Introduction

Few concepts seem as popular in the United States as mandating that a new acquisition program be based on a proven parent hull. It seems a simple concept: reduce risk and time by re-using hull lines, structures, machinery, and overall integration from a proven design. Given the documented difficulty and frustration around design maturity in federal acquisition programs, requiring a previous design to be used as a starting point would seem to be an obvious step in the right direction. Yet the recent record of adopting parent hulls is mixed at best - both the FFG-62 program and the USCG Heavy Icebreaker have run into extensive difficulties despite being “parent hull” programs. Others point to the success of the USCG smaller cutters - the 110s, 87s, and new Sentinel-class cutter to show that parent hull mandates work. So what is the true story?

Carefully analyzing parent hull programs over the last five decades points to a more nuanced view. Parent hull approaches can be useful when employed skillfully but are not a universal recipe for success. There are successful examples for large naval vessels, and the USCG small-cutter success story isn’t anywhere near as simple as parent hull advocates make it out to be. Here, I will mainly focus on the concept of a parent hull, not that of a modular base design (e.g., MEKO), which is a fascinating extension but deserving of its own post.

There are three main reasons to consider using a parent hull approach in acquisition:

An organization lacks access to the knowledge structure necessary to create a design for the given mission. DAMEN is notable for offering a range of options to help organizations (largely governments) in this situation, from adaptable designs to construction help.
An existing design can be adapted for a new role, saving significant engineering work, and potentially gaining efficiencies in production and maintenance1
The acquisition process (not just the naval architecture step) required to create clean-sheet new designs is too slow and/or cumbersome, and it is desired to justify “short-circuiting” this process by using a parent hull to accelerate the start of a material solution to the design problem.

Notably, reasons #2 and #3 are not about reducing risk but largely about accelerating schedule or saving labor. When discussing parent hull approaches, it is often assumed they are a lower-risk option. This is fundamentally untrue. For reasons #2 and #3, programs may pay a penalty with increased downstream design integration risk in return for achieving faster upstage times.

What the historical record says

For large Navy ships, two papers published almost three decades apart provide insight into the actual performance of parent hullforms. The first is Covich and Hammes “Repeat Ship Design Facts and Myths”2 from 1983, which is nicely balanced by the second, Keane and Tibbits's more recent “The Fallacy of Using a Parent Design: The Design is Mature.3” from 2013. Covich and Hammes review several auxiliary and combatant designs, noting that roughly half the designs from 1970-1983 were “repeats”, often, but not always, sharing similar hulls. Keane and Tibbits focus on more recent designs, highlighting the problems that arise from holding one or more dimensions to match that of a parent hull.

A series of designs based on the USS Spraunce, DD 963 hullform, span both papers and give a good idea of the benefits and challenges of the parent approach. The base DD 963 design was modified twice - once into the DDG 993 Kidd class (originally ordered by Iran), which took an anti-submarine destroyer with a significant growth margin and made it into a general-purpose guided missile destroyer. The design was further, and far more extensively, modified to become the CG-47/CG-52 class cruisers. Covich and Hammes note that moving from the DD 963 to the DDG 993, a change made when the DD 963 design was still in active production, saved significant engineering labor and was an overall success. However, the more complex DDG 993 still took longer to design and build than the parent design. The CG-47/52 is the flip side of this story; the required modifications consumed all the margin remaining in the DD 963 design, and were still hungry for more. The result was a crash program to try to reduce the weight of the design in production (particularly so the VLS system would fit in CG-52 onward). This program was called TOTS, for Take Off Tons Sensibly. Unfortunately, the ocean had different ideas about whether the reductions were sensible, leading to in-service structural failures. Follow-on programs to reinforce the ship (with further weight growth) were required after damage in service was discovered. The CG-52s struggled through life with minimal service life growth margins and extremely challenging drydocking characteristics. The class ended its life with great difficulty in achieving the last ~5 years of desired service, owing in part to hull structure issues. The highly-effective AEGIS system would have been better served to go to sea on a new hullform, not the DD 963 parent. It’s worth noting that this effectively happened with the DDG-51, a design that has supported significant growth and is still in active production in 2024. Furthermore, the first DDG-51 was laid down only four years after CG-52.

Combining the conclusions presented in both papers, a few points stand out from the historical record:

The use of parent designs makes the most sense when a recent design, still in production, can achieve new missions with limited modifications.
Extensive passage of time, new requirements, or significant mission requirement differences almost always make a parent design less effective than a clean sheet design.
Using an inappropriate parent often results in designs with lower weight and center of gravity growth margins. Such margins are critical to keeping ships in service and effective over 30-40 year service lives.
There is little historical evidence for significantly faster overall program lengths when using parent or repeat designs, except for low production or one-off very close repeats.

At the other end of the size spectrum, the USCG experience with parent hulls for small cutters is often given as an example of a successful parent hull approach. Careful analysis indicates, however, that the USCG is deeply involved in extensive modifications to these designs before production. Furthermore, using parent hulls did not prevent problems from emerging in service.

The first USCG example is the 110-foot Island-Class patrol boat. In a review of the propulsion aspects of the program, McCarthey4 notes that political pressure to ramp up drug interdictions necessitated a more rapid acquisition program than the USCG’s prior approach of designing cutters in-house. The resulting cutter was based on a Vosper Thornycroft design, with a strict parent hull requirement (the contract was re-awarded after the original winner was disqualified for proposing to reduce the number of engine cylinders within the same engine model to better match the USCG’s needs). The 110s had a rough entry into service, including:

Structural reinforcements for the bow region, owing to higher loads in USCG service (different concept of operations/environment from parent). A second and third flight of this design were built with a modified bow structure.
Forced to use a de-rated engine as there was no freedom to re-design the propeller or shafting given the parent hull requirements. The engine manufacturer had improved the engine in the ~10 years since the parent hull was developed, but this higher power could not be put through the parent shafting and propeller. Low-speed operations and towing light vessels were challenging with the propulsion setup as the minimum two-shaft speed was over 10 knots, as derating the engines did not lower their cut-in powers4 .
In-service propeller cavitation was so bad that holes were drilled between the face and back surfaces of the propeller to equalize pressure4 (e.g. destroying thrust). The final 110 was heavier than the parent craft, this forced the parent propeller into an off-design condition, yet it could not be re-designed owing to the strict parent requirement.
A far-off optimal hydrodynamic performance. The 110s were finally fixed with a stern flap, new propellers, and other modifications 15 years after the first vessels entered service5.

The more recent 154-foot fast response cutters (FRCs) are also held up as an example of a straightforward application of a parent hull. But again, the true record of the program tells a different story. For this class, the USCG selected a modern hullform from DAMEN, DAMEN designs adaptable vessels as part of their business plan. While keeping the successful hull shape, the USCG extensively adapted the design for their own needs6:

Fixed pitch propellers replaced the controllable pitch propellers on the parent.
A new structural design using higher-strength steel to meet ABS HSNC requirements (vs. LR on the parent) and to free up weight for other modifications needed (e.g. an additional growth margin through design modification).
Three new transverse watertight bulkheads were fitted to move from a one-compartment damage standard on the parent craft to the USCG’s requirement of two compartments.
Extensive layout and equipment changes, including reduced firefighting capability, upgraded machinery and foundations, the addition of hull-mounted weaponry, and various equipment switch-outs.

In documenting the process Bollinger used for the parent hull design on the FRC, Fanguy6 repeats many of the same principles from Covich and Hammes. The DAMEN design was selected as a parent as it was recent, designed to modern standards, designed to be adaptable (e.g., had margins), and was close to the USCG mission. A more relaxed interpretation of the parent hull requirement compared to the 110s, and careful weight control (a lesson learned from the 110 problems?) ensured that the resulting design was within 1% of the displacement of the parent hull. In turn, this ensured that the parent’s strong performance was retained.

A framework for thinking about parent hull or mod-repeat designs

The historical record is clear - parent hulls aren’t a magic solution to reduce programmatic risk. Normally, differences in operations, requirements, and standards mean there isn’t a stock design for a given government requirement. It is tempting to modify a parent in this situation, but the magic is in determining if a good parent exists. Forcing close adherence to a sub-optimal parent is not a long-term solution; ultimately, the physics of the ocean cannot be fooled. If the design doesn’t meet the requirements, problems such as structural damage, cavitation, and machinery failures will arise. Growth margin and through-life performance will likey decrease. There are historical examples of all of these occurring in parent hull designs. The parent hull, or mod-repeat approach, is not a magic wand but one tool amongst many for naval architects to consider.

So when does a parent hull make sense? I would argue that there is only one set of conditions where it is an all-around winner: for a set of requirements that can be met by small (or carefully controlled and checked) modifications to an existing in-production or recently-produced design that has growth margin to spare. The “flight” modifications of the DDG-51 fall in this category. The USCG’s FRCs do too - the modifications here were more extensive, but the weight control on the program was excellent, and the resulting design seems to be working well in service. Furthermore, for the DDG-51, the Navy’s decision to stop after Flight III, as the growth potential of the hull was exhausted, shows a logical endpoint for the parent hull. In contrast, the CG-47/52 and LSD-41 stories (see Keane and Tibbitts) show the pitfalls of not stopping in time and trying to squeeze one more iteration out of a now-poorly-fitting parent.

It’s clear from recent and historical examples that accelerating a program by re-using a more distant parent raises both the design integration risk and the potential for in-service problems. This does not mean this approach won’t be pursued when the need is urgent. However, one needs to be upfront with the goals and tradeoffs so people understand what is likely to happen downstream. It’s rare that a distant parent will ever be both a faster and a lower-risk approach. It sounds like the design team on the Polar Security Cutter is extensively re-working their base design. While many people are piling on to these growing delays, I highly doubt time spent getting the design right is ever wasted. Neither the knowledge nor the product structure for a modern USCG polar icebreaker exist in this country, and they take time and effort to develop. As the saying goes, if you want it bad, you get it bad.

So, if the risk answer is not a parent hull, how do you control programmatic risk? Leading designers and shipbuilders have long looked to minimize risk by re-using specific features (e.g. engine room design, architecture of a specific system, component-level re-use) between designs. The idea that new ship designs should be limited to trying out at most one or two new features is also a widely repeated saying in the field that I can trace back at least as far as the 1960s. Programs like DDG-1000, with a new hullform, radars, VLS, gun system, propulsion architecture, superstructure material etc. are inherently extremely high risk. Yet many defense system engineering acquisition approaches struggle to determine an overall risk that comes from taking many “small” risks at once. (If you have a red/yellow/green risk matrix for ship design where 4-5 yellows don’t equal a red, your risk approach and the reality of the problem are not in alignment…). These risks are much harder to control than the well-known and bounded risks of developing new hull geometry, structure, and layout with proven components.

When there isn’t a good parent, limiting subsystem development risk and providing a hullform with weight and volume margin for integration (this is common in most modern design methods - e.g. Andrew’s Design Building Block7, Keane’s inside-out design8 and even Levander’s System-Based design I would argue) is the best approach. This will almost always be a lower-risk approach than forcing a poorly fitting parent on the design problem. If one looks at the DDG-X briefings, many aspects of this approach are apparent: building off a proven combat system and propulsion architecture with a low-risk but larger hullform for Flight I. Controlling your risk appetite is best when the parent approach doesn’t fit. So use parents sparingly, and don’t be fooled into thinking they always mean lower risk - when process speed is the objective, a parent hull approach often increases integration risk to buy that speed.

a2NavArc’s Substack

Discussion about this post