A pump jet for the future submarine? Not so fast (or slow)
14 Feb 2018|

A couple of months ago Andrew Davies offered some comments on the pump-jet propulsion system that’s proposed as a key feature of the Shortfin Barracuda. As he observed, there seems a clear case for propellers being more efficient at low speeds—which is very important for conventional submarines. Comments from Naval Group’s director of the future submarine program seemingly confirmed this.

Andrew raised two important questions that might help resolve the apparent contradiction between Defence’s confidence in the pump-jet solution and seemingly credible evidence that its low-speed inefficiency makes a pump jet an odd choice for a diesel-electric submarine.

The first was that the cross-over point—where a pump jet becomes more efficient than a propeller—was unknown. If the pump jet was actually more efficient at plausible higher transit speeds, on balance that may offer a benefit over the entire mission profile, even if the pump jet is marginally worse at very low patrol speeds. The second was the effect of the ‘hotel load’. Since the square law for drag means that the energy required for propulsion becomes extremely small at very low speeds, underwater endurance might be more affected by the power required to keep the lights, sonars and air conditioning running than by the power necessary to move forward.

Both of these arguments really require some better understanding of the likely efficiency curves in order to be answered with any confidence, as Andrew notes. But it’s quite possible that sufficient evidence does lie in the public domain to rule some possibilities in or out.

The first question requires the establishment of a plausible cross-over point. Here, all available evidence points firmly against the idea that the cross-over could lie anywhere near the plausible transit speeds of submarines (around 10 knots). For surface vessels, open propellers are adopted essentially universally by all ships that work at speeds up to around 30 knots, including relatively fast ferries. Only extremely fast vessels (35 knots plus) tend to have waterjets. Australian companies specialising in such very-fast vessels have spent considerable effort investigating the prospect of using jets in the 20–30 knot range and discovered that the efficiency penalty of jets in this range is substantial, up to a factor of two.

And at low speeds the penalty can be very high indeed. Other research shows that the propulsive performance of waterjets falls off towards zero as craft velocity decreases. Marine enthusiasts may observe that ‘waterjets’, which eject water on or above the surface, are not quite the same as the completely submerged pump jets often found on the aft end of submarines and torpedoes. But the more I dived into the literature on fully submerged pump jets, the less advocacy I found for efficiency in any speed range. In fact, it seemed that pump-jet efficiency had to be traded-off to eliminate cavitation at higher speeds.

The method by which this occurs is simple enough to warrant mentioning. The role of the duct in a pump jet is to slow the water down before it reaches the spinning rotor that adds energy to the flow. This slowing of the water stream also increases its pressure, and the elevated pressure helps reduce cavitation. Problems arise when you try to slow down water that’s not going very fast in the first place, which is what happens at low speeds. The result is predictable: some of the water stops, or actually starts moving in the other direction, spinning around in eddies and vortices in a phenomenon called ‘flow separation’. The slower you go, the more that occurs, and the worse the efficiency becomes. The more the duct acts to raise the pressure to eliminate cavitation, the more slowing occurs, and so the sooner flow separation begins, further worsening efficiency.

Consequently, it seems unlikely in the extreme that the acoustic advantage offered by pump jets at higher speed can be decoupled from an efficiency penalty at lower speeds. It also seems highly unlikely that pump jets are more efficient than propellers at submarine transit speeds of around 10 knots.

The second question requires some quantitative assumptions about the curve and hotel load, as well as a more generalised model to be built out of the algebra described in Andrew’s ‘geeky annex’.

A good reference for plausible curves is a 2008 BMT and Rolls-Royce study of a pump-jet concept for a very quiet anti-submarine warfare surface ship. While the precise jet design may not be identical, this jet was at least fully submerged, unlike most surface ship jets. The study replicates the same efficiency curve one would expect from theory, with efficiency rising to  its peak at around 30 knots, and declining towards zero at very low speeds.

With a hotel load of 100kW (consistent with Andrew’s assumption that it might equal propulsion power at patrol speeds), it appears that in general the impact of the low efficiency of a pump jet will still be very significant in most of the important speed ranges for a submarine, particularly around 3–7 knots. Altering a range of assumptions doesn’t generally change that conclusion, as can be seen in the full discussion here. Switching back to a propeller could realistically result in improvements in dived range and endurance of 60% or more at these lower speeds. This could amount to a couple of days dived endurance, or hundreds of miles additional dived range, and credibly five or ten times that difference if air-independent propulsion was included in the submarine.

Unless the French have achieved a quantum leap in pump-jet technologies in the past few years and none of the previous physics or experimental results apply, it seems that the choice of a pump jet is curious indeed. Exactly what kind of advantage would justify accepting such a penalty in terms of dived range, dived endurance, indiscretion ratio and overall range is quite hard to imagine when building a ‘regionally superior’ submarine. Defence has made crucial errors of judgement with grave long-term consequences in acquisition projects before. We would do well to make sure that the same doesn’t happen with the future submarine.