First, I got it wrong. It's not 165.6 g/L at 1atm and 20k, it's 162.4 g/L at 1 atm and 18.5 K. 18.5 K is the triple point at 1 atm, where deuterium can be both solid, liquid and gaseous. If I knew what I was doing I could then calculate the density at many pressures. But I can't.
As I think I said in the TOS Rom BOP thread, there is evidence that deuterium can be stored considerably more densely than the "standard" 162.4 g/L. It's called "ultradense deuterium" and is highly speculative at the moment but promises densities of 100 to 150 tonnes per liter. That is not a type-o.
The circumstances for this density are, for the moment, not particularly conducive to antimatter storage: it's supposedly found in the pores of certain metals under odd circumstances. And the researcher who claims to have detected it has had his findings questioned. None the less, the theoretical foundations are apparently sound. So it is a possibility. Thus we can make the assumption that up to this rather large limit, we can store as much antideuterium as we need in as small a container as necessary.
So, let's calculate M/AM mass requirements from 9.6wf for 12 hours and this will give a lower limit for our needs. For Ent-D 9.6wf needs 4.58e18 watts (2.4e15 W/coch * 1909c). Twelve hours of this is 213e21 joules, at 93% power conversion, equivalent to 2.36 tonnes M/AM, or 1.18 tonnes of antideuterium. Within a 3000kL volume, that's 394 kg/kL.
You will note the total amount of fuel-mass here is nearly twice what I calculated earlier for three years at warp six! So, again the numbers don't match.
Page 69 gives a normal deuterium tankage of 62,500 kL. Again assuming a density of 162.4 g/L --a questionable assumption since they specifically state a temperature of 13.8 K-- that's 5.28 tonnes of deuterium. I figure that's a lower limit for deuterium storage. If you had the same amount of antideuterium, you'd need 1760 g/L... I enter this data point for thoroughness as I do not expect this tto be accurate. However, I think I proved in the TOS Rom BOP thread that the bussard collectors could gather enough fuel to supplement the deuterium tanks for impulse power while in warp flight if you can fuse hydrogen into deuterium... And maybe even if you can't, but I'll have to double check that. So maybe this isn't a completely useless data point. Of course, that much fuel would run the engines at warp 6 for 31yrs, 10mo.
So, I think a valid wild-ass-guess for antideuterium density is between 500 and 1000 kg/kL. That's less then 1% of the above ultradense limit and more than what's needed for the 12h @ 9.6wf requirement.
I have no estimate for what 75% of its power output would be. Do you? ... I could never think of a way of estimating it, or even limiting the range.
However, I just found on page 64 of TNG:TM that the cross sectional area of the reaction has a radius between 2.1 & 9.3 cm. Thus the "reaction chamber" does not have a volume but an area. Further, page 59 says the magnetic strictures between the injectors and the dilithium crystal "add between 200 & 300 m/sec velocity." Put together, this gives us a volume between 0.554 & 16.3 kL per second of both matter and antimatter. But we don't know at what densities so we can't convert that to mass or energy usage...that I can see.
However, given that 9.6wf requires 4.58e18 watts and 93% conversion, that's 54.7 kg and 3.36 kg/kL in the injectors. Not particularly dense, which you'd expect. But, once again that only tells us what's needed for the warp drive.
However, page 138 suggests the shields can draw a momentary peak of 3.31 TW (7 generators at 473 GW each) but a more normal maximum and continuous load of 2.688 GW. But even the larger number does not add to fuel consumption much: 39.6 milligrams per second of M/AM. And page 88 seems to place an upper bound to deflector power projection at 2.53 TW (3 projectors with 80% of peak momentary output of 675 GW each). Again, not a huge hit on fuel reserves: 30.2 milligrams per second (remember the assumed 93% energy to power conversion)....And this is supposed to represent maximum output of these systems. So I don't think the manual can be correct.
And I agree with your assessment that the manual implies that warp speed limitations are in the materials of the nacelles, not the fuel supply.
I will look for your in-universe post.