Barrus, Boston, 1891, p. 18, and in 'Experimental Engineering,' Carpenter, p. 386. 1 VI. The Wire-drawing Method. This was first proposed by Professor Peabody. Mr. Barrus and others have devised modified forms of the apparatus. It depends on the principle that slightly moist steam is superheated by wire drawing. Fig. 5 shows Mr. Barrus's arrangement of the apparatus. The steam passes from a chamber A to a chamber B through a very small aperture (inch in diameter). The full steam pressure is in A, and the pressure in B differs little from atmospheric pressure. Ther mometers t, t, give the temperatures of the chambers. Let t, to be the temperatures of the steam before and after wire drawing, and t, the temperature of saturated steam corresponding to the pressure in B. t-t3 is the amount of superheating of the steam in B. Let h1, L, be the liquid heat and latent heat of steam at t,; h3, L3, corresponding quantities for steam at t3. 1 See A Universal Steam Calorimeter. By G. H. Barrus, Amer. Soc. Mech. Engineers, 1890. Or taking the specific heat of water constant : No weighing is required, and temperatures only have to be observed. The apparatus must be in operation for twenty to thirty minutes before observations are commenced. Then observations may be begun and continued for as long a period as is desired. It is not a mere small sample of steam, therefore, which is observed, but a considerable quantity taken from the steam pipe at a uniform rate. The observations are as simple as possible, and the apparatus is used with the greatest facility. Presently some tests will be given tending to show that the instrument is very trustworthy. Suppose the steam is initially quite dry. Then x=1 in the formula above. h1+L1 =x3+L3+0·48(12-13) From the small excess of pressure over atmospheric pressure tz is about 214°. Then t2=0·604t,+84.5. This may be termed the normal temperature in the chamber B for dry steam. In proportion as the temperature falls below this the steam is initially moist. If it contains initially more than a certain amount of moisture, the temperature in the chamber falls to 214° or below it, and no calculation of the dryness is possible. If t2=220° and t3=214° are taken as the limiting values for which the instrument can safely be used, there being then 6° of superheating in B, then give the minimum initial dryness or maximum amount of initial moisture in steam for which the apparatus is suitable. VII. Combined Separating and Wire-drawing Method. To extend the usefulness of the wire-drawing method Mr. Barrus added a separator (C, fig. 5). The steam first passes through the separator, leaving most of its moisture, and the remainder is measured by a wire-drawing calorimeter. When thus arranged the use of the instrument is much more complicated, as the amount of steam flowing through and the amount of moisture separated must be weighed. A condenser must therefore be used, at any rate occasionally, to determine the amount of steam flowing in a given time. Assuming the orifice to remain unchanged, a formula can be found for the weight of steam discharged per minute at any given pressure. VIII. Superheating by Addition of Heat.-A method has been proposed by Mr. W. R. Cummins, in which a vessel of constant volume is used, completely surrounded by a steam jacket. Steam is at first blown through both jacket and vessel till there is a fair sample of the steam to be tested in both. Next the jacket and vessel are cut off from the steam pipe and from each other. The steam in the jacket is then heated by compression (by a steam compressing pump) or by heating. The simultaneous rise of pressure and temperature in the vessel containing the sample of steam are observed. While moisture is evaporating the relation of pressure and temperature will follow Regnault's law. From the moment when superheating begins the pressure will increase much less rapidly with the temperature. Let t1 be the temperature of the sample of steam initially, t, the temperature at which superheating is observed to begin, v, v, the corresponding specific volumes of dry saturated steam, x the initial dryness of the steam, V the volume of the vessel, and w the weight in pounds of steam admitted. Then where the specific volumes vv, can be calculated from the observed temperatures tot. The method is quite accurate in principle, but it would involve considerable skill to carry it out accurately, and it does not appear to have been actually tried. IX. Chemical Methods.-These suppose that some soluble salt is added to the boiler water so as to form a solution of known strength, and that of all the feed supplied to the boiler part is evaporated and flows away as steam, part as priming water. The former removes no salt; the latter is of the saltness of the water then in the boiler. Hence, either the decrease of saltness of the boiler water or the amount of salt present in the steam can be used in determining the amount of priming water in the steam. Most commonly table salt (NaCl) has been used because the amount in solution can be easily determined by the nitrate of silver method. In other cases sulphate of soda has been used, the sulphuric acid being determined by precipitating with chloride of barium. It is convenient to describe three variations in the way in which the chemical method is applied. (a) Decrease of Saltness of the Boiler Water measured.-Salt is introduced with the feed till the boiler water contains 1 or 1 per cent. of salt. Then, during a test in which the amount of feed is measured, a sample of the boiler water is drawn off at the beginning and end of the test. Care must be taken that the level of the water in the boiler (as shown by the gauge glass) is exactly the same when the two samples are taken. Let W be the weight of water in the boiler, and 8, 8, the percentage of salt in the two samples. The amount of salt removed from the boiler during the test is W(s,-82)/100. Let w be the amount of feed supplied, and a the dryness fraction of the steam. Then xw lb. of pure steam are 1 Trans. North-east Coast Inst. of Engineers and Shipbuilders, 1893. generated, and (1-x) w lb. of priming water carried over. saltness of the priming water is (8, +82)/2. The mean Hence The method deals with the whole amount of steam produced, and not with mere samples of it. On the other hand, the disadvantages of the method are: (1) It measures only priming water mechanically projected into and carried by the steam, and not moistness produced in other ways. (2) Great exactness of analysis is necessary, the difference of saltness being a small quantity. (3) The feed, being continuously supplied at one part of the boiler, tends to produce, in spite of the circulation, a nonuniform distribution of saltness throughout the mass of water in the boiler. If the saltness is not uniform, the difference of saltness of the samples gives a quite erroneous result. If there is salt in the feed water, it can be allowed for. This method was first used at Mulhouse in 1859; much later it was revived, and is often called Brauer's method. (B) Percentage of Salt in the Steam determined.-At intervals during a test, samples of the boiler water are taken, and at the same time some steam is condensed in a small surface condenser. Equal quantities of the water samples are mixed together, and also equal quantities of the condensed steam. Let s be the percentage of salt in the averaged sample of boiler water, and 8, that in the averaged sample of condensed steam. If is the dryness fraction of the steam, each pound contains x lb. of steam and 1- lb. of water of the saltness of the boiler water. (1 − x) 81 = 82 82 The disadvantages of this method are (1) that very small portions of the steam are tested, and, as the mechanical action producing priming is probably irregular, it is very uncertain whether they fairly represent the average condition of the steam; (2) the accurate determination of 8, can only be made if the boiler water is uniformly salt; (3) the amount of salt in the steam is very small compared with that in the boiler water, so that the accurate determination of 8, is difficult. (7) Escher's Method. A quite different method of considerable interest has been proposed by Mr. R. Escher, of Zurich. Mr. Escher considered the case of a boiler fed with impure water. Of the salts in solution some are thrown down as incrustation, others remain in solution. In the boiler water during working there is a steady concentration of these soluble salts, which tends towards a fixed limit. At this limit the priming removes as much salt in the form of concentrated boiler water as the feed brings in. Lets be the saltness of the feed water in pounds per pound of water, and k, the saltness at any given time of the boiler water. Then each 1 Civilingenieur, 1879, p. 51. pound of steam removes (1-x) k lb. of salt, while each pound of feed introduces s lb. of salt. In steady working when the limit is reached Let k/s c, the ratio of concentration of the boiler water. Then, for example, if the boiler water contains twenty times as much salt as the feed, x=1=0.95, or the steam contains 5 per cent. of priming water. The method assumes that the feed contains a definite percentage of easily soluble salt, and that the trial has continued till the maximum concentration is reached. But if the boiler is freshly filled it may be a long time before the limiting condition is reached. Let W be the quantity of water in the boiler in pounds, and F the quantity of feed per unit of time. Let k be now the saltness of the boiler water at t from the beginning of the test. In a short interval, dt, the quantity of salt Fsdt enters the boiler with the feed, and the quantity F(1-x) kdt is carried away in the priming water. The increase of salt in the boiler is therefore The constant of integration is obtained by putting k=s when t=0. Let also 1+x t=n. α Then From this equation, c being ascertained by analysis of the boiler water, x can be determined at any period of the test. As the equation is not easily solved x may be approximated to thus. If c is the concentration after t hours' working, a first approximation to x is x=1-c. Put this in the expression above for c; a value c, will be obtained which would be the concentration in the time t, if c were the limiting concentration. Then assume |