In this country, cooling radiators for aeronautical engines have been mainly adaptations of those developed for the cooling of automobile engines where air speeds are very much slower and where Weight and particularly head resistance are of minor importance.
The rational design of radiators to meet the requirements of maximum cooling capacity with minim head resistance and minimum weight under the velocity, pressure, and temperature conditions met with in flight requires considerable fundamental information regarding the laws of air flow and of heat dissipation in the honeycomb air-tube and vertical flat water-tube types of radiator construction used to-day. At the start of the investigation practically none of this information was available although recently received reports of the British and French commissions have added much that is of value
A radiator when mounted on any particular airplane will have a definite figure of merit depending upon the type and speed of the plane, the location of radiate; the temperatures of the air and water, the barometric pressure, and the rate of circulation of the cooling water. The figure of merit as used herein is defined as the ratio of the power (i. e., in the form of heat) abstracted from the cooling water to the power absorbed in carrying the radiator. Part of this power absorbed in carrying the radiator is used in overcoming head resistance and part in supporting its weight. If the ratio of lift to drift for the plane is known, or assumed, these two factors can be combined into a single figure for equivalent head resistance and therefore of power absorbed in carrying the radiator.
The problem suggested above is too complex to permit of a complete general solution, hence it was necessary to lay out a program which could be completed in a reasonable time and which would give results of most immediate importance.
The questions of more immediate importance appear to be some what as follows: What are the effects on the figure of merit and its component factors of (a) changes in air velocity though the radiator, (b) changes in head velocity, (c) changes in ratio of depth to diameter of radiator cells or air passages with changing diameter and width of air passages, (d) addition or omission of secondary cooling surface i. e., metallic surface not in direct contact with cooled water), (e) addition of devices or changes of design to increase turbulence of the air stream, (f) changes in velocity of cooling water, (g) changes in turbulence of water streams as affected by design of water passages.
These questions could be best answered by the study of a sufficient number of- special models designed and constructed specifically for the purpose of observing each factor separately but the practical difficulty of constructing such a series of modefs in point of time made it advisable to limit the initial experiments at least to specimen radiators of types already in use with such modifications of dimensions as could be devised and built by the various makers, thus extending as far as possible the experimental range.
In pursuance of this program, two experimental wind channels were designed and constructed to accommodate specimen radiators with an exposed frontal area of 8 by 8 inches of typical air cells or passages, the necessary water connections not being included in this area.
The first wind channel shown diagrammatically in figure 1 is constructed inside a steel chamber in which the air pressure can be reduced to a fraction of an atmosphere. Air velocities in this channel at atmospheric pressure up to about 70 m. p. h. can be obtained without the specimen in place, or from 50 to 65 m. p. h. through the air cells of some of the types of specimen radiators when placed in the channel. Provision is made for measurement of velocities, pressure differences, and temperature differences in both the air and water streams; for metering the air and the water and for close control over the temperatures and rates of flow of both water and air.
The velocities obtainable in this channel, while reasonably adequate for calorimetric measurements were not deemed sufficient for a study of head resistance at very high speeds. Therefore, the second channel was constructed in which velocities of about 120 m. p. h. are attained in the open channel and up to 90 m. p. h. with certain radiators in place. In this channel, measurements of air velocity and resistance to air flow are made, but it has been impracticable to duplicate the calorimetric equipment which is elaborate and complicated, requiring numerous instruments which could not be purchased or built without much delay.
In order to correlate the figure of merit as determined in these experiments with the performance of the radiator in an open air stream measurements of the head resistance and tube velocities in a wind tunnel of large size, at least 4 feet in diameter, are necessary’. The necessary equipment for these measurements will be available at the Bureau of Standards by February 1.
A series of some 45 or more specimen radiators have been secured representing practically all the radiator designs now in use on airplanes or automobiles together with a number of modified designs to cover more -fully the desired experimental range of cell dimensions. Measurements for determining the figure of merit of a number of these specimens have already been completed, observations having been made in both of the 8-inch wind channels. The immediate program calls for the detailed study of about 20 of the radiators in hand, so selected as to give the necessary data for a preliminary report on the several questions outlined above. Until the observations on these specimens are completed, none of the results can be put in final form, although the following tentative conclusions may be drawn from the results already obtained:
(1) The amount of heat transferred is nearly proportional to the velocity through tubes when the ratio of the length of the air cells to their diameters is greater than about 8. This relation holds for velocities up to about 60 m. p. h. through the tubes.
(2) The pressure difference necessary to force air through the tubes is in nearly all cases proportional to Va where V is the velocity through the tubes and varies from 1.5 to 2.5, depending on the air cells construction.
(3) Various features added for the purpose of producing turbulent flow in the air passages are by no means equally efficient. Some types introduce resistance that is not compensated for by a corresponding increased cooling capacity.
(4) When properly designed, indirect cooling surface, i. e., metal not in direct contact with the water, appears from the present data to be good practice.
(5) Increasing the water velocity in different types of radiator construction in general results in increasing the rate of heat transfer, but by no means in like proportion in all radiators.