The following text may have been generated by Optical Character Recognition, with varying degrees of accuracy. Reader beware!

TAE MaRINE REVIEW | Factors of oe in Marine Engineering - BY PROF. J OHN OLIVER ARNOLD, OF SHEFFIELD UNIVERSITY, On the vital importance of the sub- ject dealt with in this paper it is obvi- ously unnecessary to dilate, and in deal- ing with it the author will endeavor to sustain the somewhat difficult, delicate, but, unfortunately, necessary part of act- ing as a sort of connecting link between engineers and metallurgists. He fully expects some adverse criticism from the ranks of the first-named, since it will be his duty to deny the truth of dangerous dogmas which have to a considerable ex- tent, in connection with engineering cal- culations, held the field since 1871. In discussion upon matters such as those dealt with in this paper, engineers have' occasionally insisted on the great - importance of basing all calculations on the elastic limits of the materials used in construction; but, such contributors to debate have been somewhat akin to finger-posts, which point out the way but never travel by it. Broadly speaking, engineers always base their calculations on the maximum stress (often absurdly called the "breaking strain")+ capable of being endured by the metal employed. The reasons. for this procedure are: (1) The maximum stress is a value 'capable of accurate and rapid measure- 'ment on the static machines used in or- dinary works practice. (2) It is more or less an accepted 'canon in engineering faith that the maxi- mum stress is really an indirect measure ment of the elastic limit in the ratio of about 2:1. There is no doubt that in the 'majority of cases the rule just specified 'is approximately true, but it is also equally 'true that in a relatively small, but, never- theless, extremely important number of instances, this widely accepted datum is hopelessly inaccurate. So far as the author has been able to ascertain, factors of safety in marine and other engineering practice are calculated as follows: Firstly--In' a rough-and-ready fashion by taking care to subject the material of construction only to a working stress _varying according to circumstances from about one-quarter to, say, one-tenth of the maximum stress of being endured by. the material. Secondly.--In scientific engineering a more elaborate course of procedure ob- tains. Let 1 = the maximum stress of the material and m = the factor of safety, 1 then -- = working load. The factor n is n *Read at Institution of Naval Architects. +The "breaking strain" is the total elonga- tion in inches, and yet one often sees report sheets expressing this linear dimension in tons per square inch. speaking generally, ° obtained as a product of several sub- factors, which may be designated as a, b, c and d. Sub-factor a is the ratio of the elastic limit to the maximum stress, 'and is almost untiversally, in connection with mild structural steel, assumed to equal 114 to 21%4. Sub-factor b is a vari- able, dependent on the character of the stress to which the material is subjected, and, broadly speaking, is based on the classical work published by Wohler in 1871, which indicated that metals sub- jected to rapidly varying stresses, will ul- timately break under a load within the elastic limit. : TABLE 1.* SHOWING DEVIATION OF FACTORS OF SAFETY FROM THE COMPONENT SUB-FACTORS, ~ stated. bodied therein, were considerably under- In connection with the main points at issue, namely, the properties of more or less mild steels connected with structural materials of engineering, it is necessary to quote only a fraction of the data just referred to, namely, that concerned with steels ranging in carbon from about 0.1 per cent to 0.4 per cent. The steels investigated were submitted to tests in two widely differing micro-chemi- cal conditions, produced by two widely diverse rates of cooling. These two con- ditions may be specified by the terms, "normalized" and "annealed," or, to be strictly accurate, "annealed to death." The tabulated results of these particular steels are now reproduced from the pro- Showing Derivation of Factors of Safety from the Component Sub-Factors. Type of purpose for which the a steel is to be used. (Elasticity). Boilers ee ree ter ten cate he, 2 Double-acting connecting rod....... 1.5--2 Single-acting connecting rod......... 1.5--2 Shaft cartyine propeller = -0acn..e. Ab? Steel "cast wheeler esse. ey ee 2 Steel: CAShiNOS Or re es ee 2 Oils tempered steel... iin a) Nickel steel? ee ona pean wea 15 *The figures in this table were kindly compiled for 'he author by Nature of sub- pe involved------- d n=: b CRapidity Unknown ax pXex d= (Wohler Phe- .. of Contin- Factor of nomenon). Load). gencies). Safety. 1 2u%--3 4.5--6 3 2 15 13.5--18 2 eZ 15 9--12 3 1 1S 6.75--9 dj 1 4 8 1 1 2 4 1 ii is) 225 J 1 15 2.29 Mr. J. W. Kershaw, M. Sc., of the Engineering Department of Sheffield University. Taking three typical cases, we have: (1) A steady load, as in bridges and boilers, in which b == 1, (2) Loads varying between zero. ad a maximum as in single-acting connect- ing rods, where b = 2. (3) Alternating loads, where stress varies from tension to compression, as in a double-acting connecting rod, where b = 3: Sub-factor c is a variable dependent: upon the rapidity with which the stress is applied. A sudden load is supposed to stress the material to twice the extent of one gradually applied, therefore, c = 1 for a steady load, 2 for suddenly applied load, and, say, 3 for load including im- pact. Sub-factor d is used as a margin for unknown conditions and contingen- cies, and is a variable ranging from 114 to 3. Table I embodies concrete exam- ples of the application of the foregoing sub-factors. : In 1895 the author read before the Insti- tution of Civil Engineers a paper on "The Influence of Carbon on Iron," which engineers have apparently considered of small practical value. After an interval of a dozen years the author again ven- tures to call the attention of marine engineers in particular to the remarkable statical test data published in that paper, and also to show that some of the suffi- ciently disconcerting mechanical facts em- _ ceedings of the Institution of Civil Engi- neers in Tables II and III of the present paper. The requisite thermal definitions are as follows: The normalized steels were heated to a bright red or orange, say, 950° C., and were spontaneously cooled in air. The annealed samples were main- tained at a similar temperature for about 72 hours, and were then allowed to cool to milk warm during an addi- tional 100 hours. TABLE II. NORMALIZED STEELS. per cent, dd ad ¢ 5 Gon n wW Be ae Og : S Eu w as 6 wo. vo ° ¢ a So. 2S } A a " & ees ee or ee wn = on 1 q ~~ a "a we 3.0 8 Bo a pe 88" oe n © i) = & ee 1 0.08 12:19 21.39 46.6 74.8 1y% 0.21 17.08 25.39 42.1 67.8 2 0.38 17.95 29.94 34.5 56.3 TABLE III. ANNEALED STEELS. S eg ae = ° On, oO oO _o Ho Ag oO te f=} nan nn '6 oO wn +H ® a a 1 u ag oO ae gh ee ke os oO q 32 B. . = oo ag hea bo oe g Oo 90 He 488. oe N oO aa] = q fo 1 0.08 8.82 18.34 52:2 76.7 1y% 0.21 9.02 21.25 42.3 65.7 2 0.38 9.55 25.02 35.0 50.6 In connection with factors of safety the important columns in the forego-