[TEC] Unit of measurement fix for merchant ship slowdown

[nematode]

PROBLEM
Many of Silent Hunter 4's merchant ships, and some of its warships, slow down to a few knots when they turn, and accelerate back to speed very slowly. This problem appears whenever the ships turn at a waypoint, or turn during "evade mode", which is triggered when a player sub is detected.


INTRODUCTION
This article documents a unit of measurement bugfix to eng_power, which is a game measure of a ship's horsepower found in ships' .sim files.

This bug is UNRELATED to any "lost leader" or any other AI problem. This isn't a "rivet counter" issue, or a how-to for a realism mod. The UBIdevs gave us eng_power values for the game that they collected from historical references for the real life ships' sea trial horsepower. But apparently without the UBIdevs realizing it, one of the sources that they used put them on the wrong track with a badly flawed unit of measurement. I'll explain all that in a minute.

By posting my research here, I'm making it freely available to anyone who wants to incorporate it into SH4. If anyone uses my work and cares to credit me, that'd be nice, if not, I won't lose sleep over it.

All of my findings came from a non player modified version of SH4. Fresh install, completely stock, US CDROM plus v1.2 patch.

I would like to acknowledge Timetraveller and Nvdrifter for their work creating SH3 Mini-Tweaker and resource files, without which this article would not have been possible.


DISCOVERIES
While troubleshooting an SH4 custom mission, I noticed that the Kuma light cruiser would slow down to about 2 knots when it turned at a waypoint or started evading after it spotted my sub. Firing up Mini-Tweaker, I found that eng_power in NCL_Kuma.sim was 70. No wonder Kuma was having problems. SH4 was simulating trying to accelerate a large ship with the equivalent of a motorcycle engine.

The similar Naka cruiser in SH4 doesn't have those problems, and it has 90000 eng_power in NCL_Naka.sim. Opening my copy of "Warships of the Imperial Japanese Navy, 1869-1945" by Jentschura et al, I found that the real life Kuma and Naka were both rated at 90,000 shaft horsepower.

After editing 90000 eng_power into NCL_Kuma.sim and testing the fixed Kuma in an SH4 mission, it performs properly: no extreme slowdown on turning, with acceleration out of the turn that's consistent with other warships.

I next started checking other SH4 warships. There are several mistakes, but most warships checked out fine, and their eng_power values match real life sea trial horsepower values found in historical references. I've listed corrections for the mistakes that I came across in the table below.

That's when it dawned on me that one cause of the SH4 merchant ship slowdown might be eng_power problems. I started digging, and found that yep, all but one of the Japanese merchant ships in the game are underpowered, and some of them are extremely underpowered. After more digging, I spotted a trend...

It appears that the UBIdevs used Charles Hocking's "Dictionary of Disasters at Sea During the Age of Steam" (London: Lloyd's Register of Shipping, 1969) as a major source of information on the Japanese merchant ships. This book has been digitized and put online in PDF format, available at several websites including http://perso.orange.fr/cdasm.56/dico.htm

Hocking's book documented a ship's horsepower with a unit of measure called nominal horsepower. Basically, nhp was a system dating from the 19th century to measure steam engine horsepower by formulating together the engine's dimensions. For reasons that aren't entirely clear to me, Lloyd's Register, the merchant shipping information company whose databases Hocking used to write his book, continued to list nhp in its publications well into the 20th century.

While nhp was a useful way for Lloyd's Register to maintain "rule of thumb" style values for ships' horsepower, it's an extremely inaccurate way to measure it. The problem with nhp was recognized over a century ago by mechanical engineers. A typical criticism of nhp can be found in "The Mechanical Engineering of Steam Power Plants", by Frederick Hutton, published in 1908, which has been by digitized by Google (search for the title to find the online book if you care to). Hutton wrote, "Nominal Horsepower is an old term now properly disused, which was based on an untenable assumption that all engines of a given diameter and stroke (or cylinder volume) were of the same horsepower, whatever the mean pressure on the piston or the speed of its traverse."

Engineers have long since come up with improved methods of measuring sea trial horsepower, by installing a meter somewhere on the engine or drive train. These are known as indicated horsepower (ihp), brake horsepower (bhp), or shaft horsepower (shp), depending on the type of engine and testing method.

These methods weren't perfect, but they were light years more accurate than the nhp formula. When compared to ihp, bhp, or shp, nominal horsepower underestimates the sea trial horsepower of ship engines by anywhere from 400% to 1500%.

I've been able to locate ihp, bhp, or shp sea trial horsepower values for most, but not all - I had to make several "best guesses" - of the Japanese merchant ships, and they're listed in the table below. I've included notes to document how I came up with my numbers, as well as showing where I believe UBIdevs got their nhp numbers.

By replacing the incorrect eng_power values in SH4 with my numbers, the game's merchant ships and merchant convoys will no longer come to a screeching near-halt. Or rather, if they do, it's because of AI problems, whether they are "lost leader" or bad pathing/collision avoidance stuff. Hopefully that'll be attended to by the UBIdevs. Data follows in next post......