ATSF 3463 Rebuild Project - Trains Magazine

Herewith $0.02 worth (actually nearer $2.00) of thoughts about the proposed restoration of ATSF 3463, based on the project website, from a UK fan.

The good news. It’s terrific that something is being done to rescue ATSF 4-6-4 3463 from rusting away.  It is the sole surviving US 4-6-4 from the late 1930s, the pinnacle of US design development, Baldwin’s latest and best high speed 4-6-4. The cosmetic restoration is done. Marvellous! Since not a single NYC Hudson survived from this period, nor any of the MILW Hudsons, (which have as good a claim to the World Hall of Fame as any steam locomotive I know of), if there is to be a streamlined shroud, I suggest there should be three, a Blue Goose from the ATSF Historical Society, a Dreyfuss from the NYCHS and an F7 one from the MILWHS. Only about 0.01% of the population would know that that 5450 and 103 thus created weren’t the real Alco thing, and less than half of those would care, so with any reasonable rounding up there would be 100% satisfaction. Four locos for the price of one. Can’t wait. I recognise that this would be a departure from historic practice, in which fans of a particular railroad restore in their own locality. Whether interest in steam locomotives in the US nowadays transcends the fan base of old companies I do not know, but in essence what the 3463 group is proposing is a non-partisan effort. Not sure this concept will fly in the US, but if it’s the only way to enjoy the sight of a latter day US Hudson working, I’m all for it.

The people behind it. Lots of good folks with relevant skills, it seems.

The financial plan. Details of sponsors are given, but it’s not clear they have the multimillions needed . There is a link to the University of Minnesota, which may get State or federal aid to study the proposed fuel, torrified biomass.

The environmental case.  The basic pitch is that by burning torrified biomass, a bit like coke, you could create the world’s first ‘carbon neutral’ steam locomotive. 3463 is to be used as a test bed before a ‘proper’ 21^st Century locomotive is designed.The oil burning 3460s were designed for easy conversion back to coal firing, good news. However, ‘carbon neutral’ is perhaps a bit ambitious - you sure need a lot of carbon to produce 363 tons of steel, build facilities to maintain it, produce and deliver its fuel and keep the army needed to operate and maintain it alive.

Further, as has been pointed out here, the idea that farmed wood could become a ubiquitous fuel of the future, including in a locomotive boiler, is a complete environmental and economic nonsense. Most stuff on the internet about torrified biomass is written by enthusiasts. This seems a reasonably balanced view:

From this, it seems that the idea is in fact to make better use of waste from wood processes, rather than farming trees to produce coke. How much waste wood there is to burn, and where it’s best to burn it are unanswered questions. Not in the boiler of a reciprocating steam locomotive, I would have thought.  (In my view farming biomass for fuel is scandalous. Supplies of e.g. phosphate fertiliser are not limitless, and phosphate is essential for life. I feel it in my bones. But try telling the Senators who have got ethanol production in their States that). So, maybe there is a Federal or State gravy train that can be tapped into, and by sequestering someone else’s money, 3463 might steam again, as we would all like to see. Surely you want this project to succeed rather than someone else’s, as Machiavelli might have said?

The Marketing Case. US passenger rail hangs on by a thread. There is no non-electrified railroad that that has or plans 125mph speeds, so the market for a 125mph locomotive is not clear. Few US railroads want 80 year old steam locomotives anywhere near their tracks. No community would want a fire throwing behemoth polluting their atmosphere on a regular basis. Not a strong case.

The economic case.  As has been pointed out here, the reciprocating steam locomotive died for very good economic reasons, many not related to its thermal efficiency and only by spending a large fortune can they be kept running. This is true the world over, and even given free, carbon free fuel, nothing would change in my view if that fuel could be burnt elsewhere. Surely there are always better things to do with energy sources? You can boil water with about 80% efficiency, but even the best reciprocating engines waste 80% of the heat in the steam, and precious little of this can be usefully recycled, thus maximum 16% efficiency referred to cylinder output, less at the drawbar. What about a future in which oil or natural gas is phenomenally expensive? A lot of other things will have changed by then, but I’m going to bet rail transportation will not be going back to solid fuel reciprocating steam. I will not be putting my pension fund towards supporting the proposition.

The improved efficiency case. It is implied that modernisation would lead to a significant improvement in efficiency. I believe this is optimistic thinking too. Looking at the specifics on the website:

·         Conversion to GPCS firebox. GPCS has not proved a viable option so far, but with a big stretch of the imagination, this could deal with the number 2 loss of efficiency in a steam locomotive at high steam rates, the loss of unburned fuel. These losses depend on a) the quality of coal b) the specific rate of evaporation (lbs steam/sqft grate/hr) and c) the degradation of coal in the mechanical stoker screw. For good quality hand fired coal, unburned losses are pretty low at less than 600 lbs/sqft/hr. So GPCS only ‘works’ above this rate. Now US steam passenger locomotives were generally not steamed at more than 600lbs/sqft/hr. This was sufficient to produce 70000lbs/hr steam to the cylinders on a feedwater heated 100 sqft grate type, the best part of 5000 cylinder horsepower with steam age superheat and exhausts. In fact, no more than 4000HP was needed to time schedules. I do not know, but suspect that the reason for the low specific evaporation rates was to minimise boiler maintenance costs. Conclusion if so: higher working steam rates at which GPCS might provide a significant benefit would require a complete redesign of the boiler. Further, and more importantly, the mechanistic reason for the loss of unburned fuel is not well understood. With coal, it is related to the production of fine char, and the smaller the coal feed particles, especially from stoker feed, the higher the losses. My suspicion is that the two key factors are the surface area of the coal and the rate at which that surface is burning. The faster the burn and the higher the surface area, the more char is produced per unit of heat produced. However, this applies to burning coal! (a different mechanism applies to unburned losses in an oil fired boiler). Will torrified biomass produce the same char flow as coal? Could it be worse? Does torrified biomass produce char at all? Does anyone know? The design of steam locomotives requires that very large amounts of heat are produced in a very small space, under high draught, and this provides some unique challenges, which need to be understood for burning torrified biomass before anyone starts worrying about GPCS.

·         Application of modern boiler water treatment system. Can’t argue with this, but this is about maintenance, not efficiency

·         Increased superheat temperature. The effect of improvements such as this on engine efficiency can now be estimated accurately by a computational fluid dynamics (CFD) package. We know from the ATSF test report on 3461 that the inlet steam could reach 750-800^oF, but was more often 650-700^oF. This is because superheat increases with specific evaporation rate, and as noted above, this was relatively low in the US. How much efficiency could be gained by going to 750-800^oF at normal rates? About 3%. 

·         Reduce pressure drop in steam circuit. There are two drops to consider, between the regulator and the steam chest, and from the steam chest to the cylinders. Looking at the ATSF data, boiler pressures on test were around 290 psi, and the indicator cards show that steam chest pressures were generally about 260 psi at high steam rates, say a 30psi loss. There was a constant loss of ca 12psi between the superheater inlet and the valve chest, but pressure drop between the throttle and superheater inlet increased from 4 to 18psi as steam rate increased, indicating a restriction in flow through the throttle. If this total pressure drop could be reduced by 20psi, what would be the gain in efficiency from being able to use shorter cut offs this would allow? CFD says about 2%. On the second question, the ATSF data show that both on their 4-6-4, but more particularly on their 4-8-4, maximum cylinder pressure was some way below steam chest pressure, it’s difficult to get 30000+lbs steam/hr into a single cylinder! For this reason, lead on US locomotives was quite long, 0.3125” on the 4-6-4, 0.25” on the 4-8-4. So, pressure held up better in the 4-6-4, a good thing for efficiency but the longer lead is marginally worse for efficiency. How does it all work out? Well, it’s very finely balanced and CFD says it doesn’t matter much, so steam age engineer’s gut reaction that low initial cylinder pressure was a bad thing is not really correct. 

·         Improve adjust valve settings. The 3460s had relatively short steam lap (1.125”) and this means that steam flow at a given cut off is less than it might be, so  to achieve a target power, cut off has to be lengthened,  reducing efficiency. If the valve events of the J3 were adopted (1.625” lead) CFD says efficiency at speed would improve by about 2%, and if the device used on the ATSF 4-8-4s to increase their steam lap to 2.125”, the benefit could be as much as 3%. What happens if you increase valve size? Well, steam flows faster, so you can work in more efficient shorter cut offs, but CFD says there’s little efficiency gain because the larger valves allow steam to escape faster at the end of the expansion; smaller valves mean the steam escapes more slowly, and it does more useful work in the ‘toe’ of the indicator diagram.

·         Improve backpressure. This only works if there’s a serious backpressure problem. There wasn’t on the 3460 class at the steam rates at which they were normally worked, but backpressure was about 10psi at their highest rate of working. What happens if you halve this? A gain of about 4% in efficiency.

·         Summary of engine efficiency improvements.  Below are the actual estimates from CFD of the effect of the above changes working at constant steam rate, based on an illustrative CFD calculation for 44% cut off at 64 mph which reproduces pretty faithfully what the ATSF actually found for 3461.  ‘As is’ condition has 260psi steam chest pressure, 13” valves, 1.125” steam lap, 4 *3.5” nozzles, and 695^oF inlet. ‘Full package’ includes all the improvement options. The combined benefit package adds up to an efficiency improvement of but 11%! How come Chapelon was able to claim so much more? The answer is on the final row. If you take the ‘full package’ at the same operating condition as ‘as is’ you see there is a spectacular 25% increase in power.  But steam has gone up by 20%. Only the increase in superheat and reduction in backpressure directly improve efficiency; the other approaches require reduction in cut off.  In his Compounds, the higher steam flow options helped, because restricted steam flow into the HP cylinders can limit power (this does not apply to simples working at speed which are boiler limited). 

·         Streamlining. This does offer significant advantages at speeds of 70+ mph. Not any old streamlining however. Looking at how modern traction has evolved, I would think the MILW design might have had some useful aerodynamic purpose.

·         The rest talked about. Sundry improvements to reliability, no game changers, and as with any design changes it might prove they are more difficult than it seems, or more trouble than they are worth.

·         Conclusion. Even if you spend millions of dollars in addition to the basic restoration costs to improve power and efficiency, you’re not going to get anything that’s dramatically better than 3463 as was, certainly nothing that would change the underlying economics of steam.

Improved boiler output? As noted above the oil burning 3460s were not tested at very high specific evaporation rates. There are many reasons why this might be, some related to the use of oil as a fuel, but what is clear is that solid fuel grates, whilst also not steamed much above 600lbs/sqft/hr in daily service, could be steamed at 1000+lbs/sqft/hr for show off stunts (Niagara and T1 tests, Chapelon 4-8-0, Mallard to mention some). Now these efforts were generally made with very high calorific value coal, 13500-15000Bthu/lb. Torrified biomass is about 11000Bthu/lb, and it may well be that the ‘stunt’ limit of a boiler fired with this material would be less. I am going to suppose that the ‘stunt’ limit for 3463 would be about 85000lbs/hr with torrified biomass.

Could 3463 ever reach 130 mph, even with the above upgrades? The short answer is no. The only 125+mph rated track is in the northeast corridor. A non starter. The only track with 120mph steam pedigree in the US is from Crestline west on the PRR- another non-starter - and after Caledonia near Milwaukee. So let’s suppose that nice Mr Hunter Harrison would cough up the $$$$ to rebuild the MILW from Rondout to Milwaukee. Let’s not stretch Mr Harrison’s generosity too much, and only ask him to rebuild/resignal/superelevate the line to 100mph standards as far as Sturtevant, 125 mph beyond there- back to the 1940s, for beyond Sturtevant, a MILW Hudson averaged 120mph for 5 miles I believe. Now reaching 130mph is about a) mechanical robustness, and a reciprocating steam locomotive is a pretty daft contraption to try to achieve high speed with, but I’d back 3463 for a one off attempt and b) HP. The Santa Fe rated the 3460s at 4350HP, but as noted above, this was at pretty modest steam rates for the size of the boiler. If you thrashed an upgraded 3463 to a boiler limit of 85000lbs/hr, you could get about 6000IHP at 130 mph. Completely unknown territory from a mechanical perspective, but the boiler at least would be up for this. So we run 3463 out to Sturtevant with the Afternoon Hiawatha consist, cruising at 100mph. This requires only 3000-3800IHP, completely in its comfort zone. (I would have the MILW shroud on). The good old days return, and there’s not a dry eye at the trackside. At Sturtevant we go for it, giving it a full 6000IHP, the change of grade kicks in, and what do we get at the foot of the 0.67%? About 125mph. What do need to get 130mph? Over 7000IHP. This is just simple mechanics. 130mph is a fantasy. The only option I can suggest is that one sweet-talks that nice Mr Rose into using the Santa Fe west from Kingman. Just by releasing the brakes at the top of the downgrade, 3463 and the Afternoon Hiawatha would be doing about 130mph by Topock simply from the effect of gravity. There is however a bit of a curve after Yucca, so it might be best to apply a bit of steam to get to 130mph before there. In fact, a mere 2000HP would get you well into the 140s. Whether you could then stop is a different question.

Overall conclusion. The good news is someone cares enough to get the restoration process started, and they have a good set of skills. They don’t appear to have the funds needed for the project. They are selling the idea on an environmental platform, possibly to get Government funding, and to tap the wider world’s pockets there is a claim, spurious in my view, that 3463 could reach 130mph. There is no doubting my mind that some improvements to the efficiency and power of 3463 could be made, but it is exceedingly unlikely to change the economics of the reciprocating steam locomotive. I can’t imagine if torrified biomass did become a viable fuel that you would want to burn it in a locomotive boiler. In any case, work needs to be done to find out how useful this is in a locomotive boiler. 3463 modified to burn coal is a place to start, though there are easier, less costly ones. Send a few tons of the stuff over here, for example. These do not seem sound starting points if the project is to be funded by disinterested or profit orientated third parties.

Personally, I’d be more than happy just to see 3463 running as was. If I may be undiplomatic, it is scandalous you guys let it get into such a state in the first place. If anyone really cared about US steam history it would have been up and working a long time ago. Sure, that might have cost a few million dollars, but as anyone knows the only thing you need to look after a steam locomotive is a small fortune, and to get that, you have to start with a big one, and have abundant free time and relevant skills. Plenty of folk over here in the UK have come to terms with that, and that’s why you’re never more than 50 miles from a working steam line in England and Wales. Dreams cost time and money.

If the people on the project have the finance, and the business, engineering and people skills to make this happen, they deserve every bit of support.  If they want to tinker with 3463 to get some improvements in efficiency and reliability, and put a solid fuel grate in, if that’s their entry price, I’m ok with that too. But the project will only succeed if it’s based on rational premises, or there is someone or some group with a large, irrational cheque book. That's why Tornado exists.  Time to get you cheque books out, steam fans, I think.