Abridgment of paper on concluding study of ventilation of turbo alternators multiple path radial system

In 1924 two papers were presented before the A. I.E. E. on Turbo Alternator Ventilation. In one of the papers tests on two models for two methods of ventilation were described, and data from the tests were given. The other paper contained a mathematical treatment for one system of ventilation, which was based upon the data obtained from the tests. It was recognized that the tests were not sufficiently accurate to evaluate the loss coefficients, nor was it possible to obtain data on the distribution of volumes for the intake vents. For that system, (see Fig. 1), it was found that the influence of rotation upon total volumes and their distribution could be neglected; consequently the investigation could be continued on stationary models. Those tests, the methods of determining the losses and the equations derived therefrom are given in this paper. Since the effect of rotation could be neglected, the test could be reduced to a model which represented only one axial row of vent ducts. On this model the stator vent ducts were imitated by square brass tubes with a plaster of paris restriction cast in one end to imitate the vent duct restriction so that the pressure drop obtained for either direction of flow was approximately the same as in a stator vent. A long steel channel or duct represented the section of the air-gap. Some of these brass tubes, representing intake vents, lead from a large sheet metal box to the gap channel. Other tubes, representing discharge vents, lead from the gap channel to the atmosphere. The gap channels could be inter-changed readily to represent various sizes of air-gaps. Any desired number of intake and discharge tubes could be used so as to represent any desired layout of the machine. The volumes in individual intake tubes were measured by reading on a manometer the drop in pressure from the intake chamber to a point a little ways down the tube. For the discharge tubes small impact tubes were employed. Each tube was carefully calibrated before making any other tests; a thermal volume meter was used for calibrating an orifice, and the orifice was used for the lube calibrations. Various difficulties arose, and the means of overcoming them are given. Tests were made with a group of intake vents, and the losses were separated into (1) those in the tubes; (2) those accompanying a right angle turn; (3) those due to a stream of one velocity impinging upon a stream of another velocity; (4) those due to sudden increase in velocity; (5) those due to surface frictions. Similar tests on the discharge side showed that the losses there could be separated into (1) those in the tubes; (2) those accompanying a right angle turn; (3) those due to sudden decrease in velocity; (4) those due to surface friction. The losses were put, for the most part, into comparatively simple expressions. They were then combined in order to obtain final solutions. There was no difficulty on the intake side in obtaining a differential equation which admitted of ready integration. On the discharge side, however, it was necessary to use approximations. With the aid of the final equations the total volumes and their distribution can be calculated for a given pressure drop. One difficulty is that the equations include trigonometric and hyperbolic functions of quantities involving the distance between the point where the gap velocity is maximum to the points where the gap velocity is zero. The latter point, called the “balance point” is not known, and a simultaneous solution of the transcendental equations needed for its determination, is impossible. Suggestions for a direct simple approximate solution are given, which may be followed by trial and error methods. In most of the applications, only one or two trials were required. The equations were checked for accuracy by comparison with tests made on the tubes, on the turbo model of 1922 and 1923, and with those on an actual machine. The agreement in total volumes was very close, and is considered to be quite good for distribution also.