The question of the effect of trim upon the speed of vessels is one that has been much debated by those that go down to the sea in ships for hundreds—perhaps thousands—of years. It is altogether probable that Noah's sons thought that the old man kept the ark at a bad trim for speed. Certainly in the days of sailing ships the question of trim was regarded as a very live one, but change of trim for sailing vessels affects other qualities besides speed; for instance, maneuvering power. Naval architects of the present generation, however, regard moderate variations of trim as having little effect upon resistance for vessels of ordinary types and to them it will be somewhat surprising to find in No. 133 of the PROCEEDINGS OF THE NAVAL INSTITUTE (March, 1910) a paper entitled, "Wasted Horsepower and Economical Trim" in which there is an apparently serious attempt to demonstrate that a comparatively small variation from a supposed "Economical Trim" will result in "enormous loss of wasted horsepower."
As this article ignores well-known reasons which have influenced naval architects in reaching their present frame of mind as to the effect of trim, I propose to discuss briefly some of these reasons and then consider the arguments in the "Wasted Horsepower" article.
There are two factors involved—namely, the resistance of the ship and the efficiency of propulsion, the latter being the ratio between the net horsepower which would be absorbed in towing the ship—called the effective horsepower—and the indicated horsepower. This ratio varies in practice from ship to ship roughly between the limits .5 and .6. It is affected by engine friction, propeller efficiency, wake factor and thrust deduction. The three last mentioned factors are liable to vary with variation of trim and hence the efficiency of propulsion is liable to vary with trim as well as the resistance.
Let us consider first the resistance. There is, of course, only one way to demonstrate with rigid accuracy the effect of change of trim of a full-sized ship upon its resistance. That is to tow the ship at various trims—all conditions except trim remaining unchanged—and determine curves of resistance at the various trims.
Such experiments were made by Mr. William Froude in 1871 when as one of an Admiralty committee he conducted the famous Greyhound experiments. The Greyhound was a vessel 172 feet 6 inches long between perpendiculars, of 33 feet 2 inches extreme breadth and 1161 tons normal displacement.
She was fitted with dynamometer gear and towed by H. M. S. Active from the end of a boom fitted to keep the Greyhound clear of the wake of the Active. For present purposes it is not necessary to consider the results of all the experiments which covered other questions than that of change of trim. Fig. 1 shows four experimental curves of resistance for the Greyhound at normal displacement and various trims, and Fig. 2 three curves for three other conditions when the displacement was light.
In considering Figs. 1 and 2 two things should be borne in mind. In the first place the curves from the nature of the case cannot be exact. Froude himself says in this connection: "In proceeding to compare the results shown by the several curves of resistance, it must be observed that the minutiae of the features they present cannot be insisted upon as absolutely exact, because minor discrepancies in result were in some cases noticeable when an experiment was repeated with unchanged conditions."
In the second place the range of trim variation covered was relatively enormous. A range of 6 feet for a vessel 172 feet 6 inches long is, as regards angle, equivalent to a range of 15.65 feet for a 450-foot vessel such as the Connecticut.
While the maximum speed was only about 12 knots this speed of 12 knots for a vessel 172 feet 6 inches long corresponds to a little over 19 knots for a 450-foot vessel.
Considering Figs. 1 and 2 it is seen that for the Greyhound, broadly speaking, at low speeds trim by the stern was disadvantageous and at the upper speeds it was advantageous.
Froude's own conclusion in this connection is as follows: "As dependent on differences of trim, the resistance does not change largely; indeed, at speeds between 8 and 10 knots it scarcely changes appreciably even under the maximum differences of trim. In proportion as the ship is down by the head the resistance is, on the whole, increased at the higher speeds and diminished at the lower, and this character of difference is maintained at both 'normal' and 'light' displacements."
With the comparatively small changes of resistance of the Greyhound for such a large range of trim it would seem obvious that for this vessel the variation of resistance with the variation of trim possible in practice would be secondary.
Of course, it would not be safe to draw general conclusions from trials of a single vessel. But as a result of Froude's work it was generally accepted that the resistance of a full-sized vessel in any condition could be estimated with reasonable accuracy from results of trials of a model of the same.
As a matter of fact, it was in connection with his classical Greyhound experiments that the elder Froude first demonstrated the value and reliability of model experiments.
Now in a model basin it is a very simple matter to test the effect of change of trim upon resistance. The matter has been investigated thus at the U. S. model basin during years past for models representing practically every modern vessel in the U. S. Navy and also for a large number of experimental models. In no single instance has there been found any great change of resistance for the usual range of trim variation—say 2 or 3 feet for a 450-foot vessel. The majority of the models of our earlier ships show slightly less resistance when trimmed a little by the stern than when upon an even keel but the difference is of no practical importance. With the forms that have been adopted for our later ships the variations of resistance of the models with changes of trim practicable for the ships are usually even smaller than for the older ships.
When the Kansas on trial showed such poor results trimmed two or three inches by the head a great many people jumped to the conclusion that it was because of the trim by the head. During the design of the class the effect of change of trim had been investigated as usual with the usual result that it was found to have no material effect. After the Kansas trial additional very careful experiments were made with a model of the Connecticut (the same as the Kansas). The differences were hardly large enough to measure with accuracy. Fig. 3 shows the E. H. P. curve for the vessel on an even keel and percentage variations of power for various trims as determined from model experiments. The effect of trim change is expressed in the figure by curves of percentage variation because it would be impossible to make a clear diagram of the size of the figure with curves of E. H. P. for each trim.
As shown in Fig. 3 the experiments indicate that the Connecticut class offer slightly more resistance when trimmed by the stern for speeds below 16 knots. Above 16 knots trim by the stern is slightly advantageous but the benefit is not as much as i per cent for practicable variations of trim.
So far, then, as the resistance is affected by change of trim it may be stated positively:
1. That in the case of the Greyhound where this specific question was carefully investigated by trials of a full-sized vessel the effect upon resistance of variations of trim such as are found in practice was of a minor nature.
2. That numerous experiments at the U. S. Model Basin extending over years have shown that so far as can be determined from model experiment the effect upon resistance of vessels such as compose the U. S. Navy of variations of trim such as are found in practice is of a minor nature for all types.
It may be added that experience of other model basins along this line is in full accord with that of the Washington establishment.
The evidence brought forward in the "Wasted Horsepower" paper is, however, from trials of ships when self propelled, not from trials to determine resistance only. It is certain that change of trim may be expected to have some effect upon efficiency of propulsion but if is difficult to conceive from the nature of the phenomena concerned how the effect could be great. However, we are not now concerned with how things happen but with what does happen.
Before considering some of the results of trials quoted in the paper on "Wasted Horsepower" it may be remarked that the determination of accurate curves of speed and power on trial is not such a very easy matter. Indicators should be carefully calibrated and standardized, and handled by people skilled in taking diagrams, if power results which can be relied upon as practically correct are to be obtained. Hence it is never safe, from a few speed and power curves, to draw conclusions which depend upon the minute accuracy of the curves.
Furthermore, there are a large number of factors which affect the speed and power of a ship and it is obviously wrong to ascribe a variation in power for a given speed to one factor which is varied unless we know that the others are not varied or that the variation has no effect.
A good deal of space is given in the paper on "Wasted Horsepower" to results of six sets of trials of the Kearsarge, curves of I. H. P. and speed being given for five of them. These five curves upon examination are seen to fall into two groups—a lower group of three curves and an upper group of two curves. The five trials varied as regards displacement, time out of dock, and trim. It will be found that for the upper group the times out of dock are stated to be 16 weeks and 23 weeks respectively. For the lower group these times were one week, two weeks and six weeks. There are various abnormal features about some of the curves which render it doubtful if they are of a high order of accuracy, but assuming them to be substantially correct it is impossible to draw reliable conclusions as to the effect of change of trim when ignoring the fact that some trials were made with the ship much longer out of dock than others and presumably more or less foul. However, passing this by let us examine the demonstration of "Wasted Horsepower" for the Kearsarge. We find on page 210:
When the Kearsarge is at her "Economical Trim"—between 20 and 28 inches by the stern—the following facts obtained from Plate VIII should be noted:
1. Fewer revolutions are required to maintain the same speed.
2. Consequently less horsepower is needed.
3. And therefore from 5 to 6 per cent less coal is burned than when trimmed 12 inches by the stern.
In preparing Plate VIII referred to above the results of three trials only were used, namely when the Kearsarge was 12" by the stern, 20" by the stern and 36" by the stern. The first conclusion quoted above agrees with Plate VIII. The second conclusion "Consequently less horsepower is needed" cannot be derived from Plate VIII, which deals with revolutions, speed and trim only. No doubt this conclusion ought to be true but a reference to the original curves of speed and horsepower given in the paper shows that they are not in accord with this conclusion.
Thus referring to the original curves of speed and power we have for 12 knots speed:
Trim of Trial. Revolutions. I. H. P.
12" by stern 78 3450
20" by stern 74 3430
36" by stern 77 4000
In the above I have used 74 revolutions for 12-knot speed 20" by the stern as given in the paper in Plate I instead of 75.2 as given on page 210, the latter being evidently in error. None of the revolutions for the various speeds at 20" trim, as given in Table IV, page 210, agree with the revolutions for the same Speeds and same trim as given in Table II, p. 201, although since they refer to the same trial they should presumably be identical.
It is seen that at 12 knots speed in spite of the difference in revolutions there is a difference of but 6/10 of one per cent in horsepower between the 12" trim and the 20" trim. Most people would conclude that this would result in about 6/10 of one per cent of difference of coal consumption instead of the 5 to 6 per cent deduced in the paper. At 13 knots, though the revolutions for the 20" trim are still well below those for the 12" trim, the power is actually greater. At 12 knots for the 36" trim the revolutions are 1 less than for the 12" trim but the power is some 16 per cent greater.
Evidently the dictum that fewer revolutions need less horsepower does not apply to the data of the Kearsarge trials put forward, and as a matter of fact it would seem that the only positive conclusion that can be drawn from the data given for the various Kearsarge trials is that some of it is materially in error. Hence it does not appear necessary to take up some errors of method in the deductions from the Kearsarge data in addition to those already pointed out.
Let us consider now the Connecticut class results from which there are drawn in the "Wasted Horsepower" article some very sweeping and positive conclusions.
The paper gives the powers for four vessels of the class for 18 knots on contractors trials—plotted on trim. In Table II below will be found these powers for all six vessels of the class; the powers given being the main engine I. H. P.:
TABLE II—SHOWING DISPLACEMENT AND TRIM ON STANDARDIZATION TRIAL AND I. H. P. AT 18 KNOTS FOR CONNECTICUT AND VIRGINIA CLASSES.
Name of Ship. I.H.P. at 18 knots. Trim. Disp. In Tons.
Georgia 14,815 Even Keel 14,963
Nebraska 15,000 2" by Stern 14,940
New Jersey 16,060 64" by Stern 14,883
Rhode Island 14,450 Even Keel 14,940
Virginia 15,800 3" by Stern 14,990
Connecticut 15,500 24" by Stern 16,375
Kansas 18,800 2 7/8" by Head 16,065
Louisiana 15,720 15" by Stern 16,000
Minnesota 14,980 31 9/16" by Stern 16,016
New Hampshire 15,800 254" by Stern 16,023
Vermont 16,460 8 1/8" by Stern 16,007
There will also be found in Table lithe same data for the five vessels of the Virginia class. In Fig. 4 these horsepowers are shown plotted upon the trim. The spots for the Minnesota, Louisiana, Vermont and Kansas are joined by a curve as in Plate 9 of the "Wasted Horsepower" article.
If we accept the conclusion that the differences of horsepower shown in Fig. 4 are all due entirely to the differences in trial trim it follows that in the case of the Connecticut class trim by the stern is advantageous and in the case of the Virginia class it is very markedly disadvantageous.
As a matter of fact, there are entirely too many variables involved in these trials apart from the unavoidable inaccuracy of all trial data to warrant any general conclusion being drawn from the variation of trim alone.
For instance, the curve connecting the Kansas, Vermont, Louisiana and Minnesota results is given its characteristics almost entirely by being passed through the spot for the Kansas.
It has been well established since the trial of the Kansas that the excessive horsepower she showed on trial was due to the type of propeller with which she was fitted. Even in the "Wasted Horsepower" paper we find on page 213 the following statement: "The propellers of the Kansas have been changed since the above data were taken and less power is now needed in order to make 18 knots; the horsepower necessary to make this speed more nearly approaches a mean of her sisters."
This being the case to still assume that the excessive horsepower shown by the Kansas was due to her trim alone and to base sweeping conclusions upon this assumption as is done in the paper upon "Wasted Horsepower" appears rather remarkable.
It may be remarked in passing that on page 213 of the "Wasted Horsepower" paper a statement is made in comparing the Minnesota and Kansas that "on the Kansas 4950 horsepower were apparently wasted." Table 5 on the same page, however, shows that the Kansas took at 18 knots 3950 horsepower more than the Minnesota instead of 4950.
It does not appear necessary to pursue further the discussion of the evidence and arguments brought forward in the "Wasted Horsepower" paper. The author's main conclusion as regards trim appears to be that trim by the stern is necessarily beneficial. The results of the Virginia class shown in Fig. 4 are alone sufficient to refute this conclusion if, following the method of the "Wasted Horsepower" paper, we are to assume the trial data exact and that all the differences shown are due to variation of trim.
Following this method it is very easy to reach all sorts of absurd conclusions. For instance, in Table III below will be found the average indicated horsepower to make 18 knots for the four most recent classes of completed battleships with reciprocating engines. In the case of the Connecticut class the Kansas is not used on account of her abnormal results which were due to abnormal propellers. The average displacements of the classes of vessels are also given.
It is seen that the Delaware of 33 3 per cent more displacement than the Virginia class made 18 knots on trial with nearly 20 per cent less power. It does not follow, however, that increase of displacement involves of itself less resistance. The decreased power of the Delaware is not because of her increased displacement but in spite of it. Again considering the South Carolina class and the Connecticut class, the South Carolina class has the same dimensions except beam and practically the same displacement. The beam is about 4 1/2 per cent greater and the power for 18 knots about 17 per cent less.
It would not be safe, however, to conclude that the greater the beam the less the power. In some cases it is possible to decrease the resistance for a given displacement, length, and draught, by increasing the beam and modifying the model, while in other cases this process would result in a material increase of resistance.
The conclusion that trim by the stern is necessarily beneficial to speed is inconsistent with a number of results of practical experience. For instance, in Fig. 5 will be found curves of I. H. P. on trial of the three colliers of the Mars class. These were sister vessels tried on the same course by the same Board under practically the same conditions. It is seen that up to 12 knots, which was the contract speed, the Mars which was materially by the head took less horsepower than the Hector which was slightly by the head and the Vulcan which was slightly by the stern. The differences shown, allowing for unavoidable errors of 'observation, are not large. It may be remarked, however, that experiments showed that the model of these colliers towed somewhat more easily when trimmed by the head.
During the recent trials of the Michigan on three courses of different depths some special trials were made to investigate the question of the effect of change of trim upon speed. On June 2, 1910, the Michigan was standardized on the Delaware Breakwater course. She was trimmed 16" by the stern.
On June 3, 1910, two series of three runs each were made, one with the ship 3"by the head, the other 42" by the stern. The displacement for all conditions was within a few tons of 16,000, and thus practically the same. Observers, etc., on all trials were the same and as they had shortly before run several standardizations of the Michigan on other courses, and as all trials were run within two days, it is reasonable to suppose that observation errors on these trials were reduced to a minimum and that the only factor materially changed was the trim of the vessel.
The Trial Board plotted and considered the results and stated: "The results may be summarized as follows: For a speed of 18 2 knots the following H. P. was developed:
Condition (a) (1' 04" by stern) 15,430
Condition (b) (0' 03 1/8" by head) 15,650
Condition (c) (3' 06" by stern) 16,725
For 16,000 H. P. the following speed was obtained:
Condition (a) (1' 04" by stern) 18.62 knots.
Condition (b) (0' 03 1/8" by head) 18.57 knots.
Condition (c) (3' 06" by stern) 18.36 knots.
It is seen then that the Michigan, when tried with care and skill above the ordinary and under conditions where practically the only variation was in trim, ran a little better trimmed 16" by the stern than trimmed 3/8" by the head, the difference being, however, not material. When trimmed 42" by the stern she required over moo H. P. more to make her contract speed-18% knots—than when nearly on an even keel. Model experiments with the Michigan did not indicate material change of resistance for the variations of trim concerned and it would seem that in this case, even allowing for data inaccuracy, change of trim by the stern may have affected the efficiency of propulsion prejudicially. The water on the course is only from 20 to 25 fathoms deep and in deep water the results might be slightly different but personally I hold the somewhat heretical opinion that for smooth water work deep immersion of the screws is apt to be prejudicial to efficiency. However, that may be there is no question that there is no appreciable "Wasted Horsepower" when the Michigan is trimmed three inches by the head, the trim which is alleged to have caused a "waste" of nearly 4000 H. P. in the case of the Kansas. It does not appear necessary to multiply instances.
As regards trim, I believe the following statements fully warranted.
1. So far as can be inferred from trials of the Greyhound and large numbers of trials in model basins, ordinary variations of trim of actual vessels cause variations of resistance which are entirely secondary in their nature.
2. There is no evidence which will stand analysis to show that actual ships show material variations of horsepower for small variations of trim or that change of trim by the stern is necessarily beneficial.
3. "Enormous loss of wasted horsepower" due to trim was deduced by erroneous methods from inaccurate data.
In concluding this rather discursive paper I should like to make a few remarks taking as a text the quotation below from page 199 of the paper of "Wasted Horsepower and Economical Trim"—"Ships in commission and displacing 1000 or 1500 tons more than when running their contract trials often equal or exceed the contract speed. And no reasons are usually given; but certainly there are causes for such astonishing results! Why is it that these increased displacements, in such a large number of cases, make such small differences in speed on full-power runs?"
As our vessels do not carry official explainers of "astonishing results" it is not surprising that "no reasons are usually given" but as a matter of fact these results are not astonishing and the reason is very simple. Other things being equal, when a ship develops more power she will go faster and if a vessel in commission develops more power than on her contract trial and is in equally good condition for speed the contract speed will necessarily be exceeded if the displacement is the same as on the contract trial. Moreover, many ships can reach a given speed at a materially increased displacement for a comparatively small increase of power. The reason is simply that the resistance does not increase in proportion to the increased displacement. Hence if the power of the contract trial is exceeded the contract speed, in spite of the increased displacement is likely to be equalled or exceeded.
Some years ago it was an unusual thing in the service for ships in commission to develop as much power as on contract trials. Naturally they did not develop as much speed. It was convenient to ascribe the falling off of speed to the increase of displacement in commission over trial displacement and to say very little about the falling off in power. I believe this had a great deal to do with the undoubted prevalence in the service at that time of erroneous ideas as to the reduction of speed due to increased displacement.
During the last few years it has become much more common for ships in commission to develop as much power as on contract trials and one hears much less about the baneful effect upon speed of increased displacement in commission.
The effect of variation of displacement upon resistance is a thing which has been investigated for years at the Experimental Model Basin, the established practice being as a routine matter to test all models of actual ships io per cent light and 10 per cent heavy as compared with the designed normal displacement. In addition special investigations of other displacements are often made.
Broadly speaking, the effect of variation of displacement is materially greater in the case of small fast vessels than in the case of large vessels of moderate speed. The situation may be fairly well typified by Fig. 6 which shows for the Connecticut at 18 knots and the Dahlgren, at 30 knots the effect upon speed of variation of displacement, assuming that the indicated horsepower would change in the same proportion as the effective horsepower which, of course, is substantially the case. The Connecticut for a 10 per cent variation of displacement should show about a 2 per cent variation of speed only. That is to say, if for a given power she would make 18 knots at 16,000 tons, she should make 17.64 knots for the same power at 17,600 tons, and a little under 18.36 knots for the same power at 14,400 tons.
The Dahlgren at 30 knots would fall off 1.2 knots for an increase of io per cent of displacement, power remaining unchanged, and gain a little less than a knot for a decrease of io per cent of displacement. I do not believe that in the case of any of our battleships would the variation of displacement in service affect the speed for a given power as much as the variation in foulness of the bottom or the variation of condition of the sea from smooth to moderately rough.