FIGURE 4. A view of the louvered strip mixing device showing the two elements in series.

the air passing through any two adjacent louvered areas was deflected through equal and opposing angles from its normal path. For example, the air passing through one area was deflected left by the louvers and in the adjacent area the deflection of the air was to the right. The air streams issuing through adjacent louvered strips moved in opposing directions, thus creating a shearing action between the streams at their interface. Both the forces generated by the shearing action and the resistance set up by opposing flow caused the formation of a highly turbulent field downstream from the mixing device.

Two louvered strip mixers of this construction were placed in series for this study. The first mixer provided shearing action in the horizontal plane. The second, rotated 90° in respect to the first mixer and located downstream, provided shearing action in the vertical plane (see fig. 4).

3.2. Louver-Baffle

The louver-baffle mixing device, as shown in figure 5, is a modification of one developed by Wile [4]. It was modified to fit into a round duct rather than a square or rectangular duct. For this mixer, the circular area was divided into six vertical strips each 4 in wide (left, fig. 5). One-half of the area of each strip was completely covered by a fixed metal baffle, thus making up a pattern of staggered baffles and openings over the large circular area. Behind the opening in each strip a set of louvers was attached to direct the flow of air from one side of the duct to the other. By blocking half of the total cross section of the duct, the set of

FIGURE 5. A view of the louver-baffle mixing device showing the two elements in series.

baffles approximately doubled the average air velocity in the plane of each mixing element. The baffles created in their wake, low-pressure regions into which the deflected air could flow. A pattern of turbulence was set up by the action of the adjacent layers of air moving in opposing directions. Two of the louver-baffle elements were used in series. Through the first element the air stream was deflected vertically and in the second element a horizontal deflection was achieved by rotating the element 90° from the first.

In each test, temperature values were recorded for each of the 24 thermocouple positions at the upstream and downstream measuring stations. A statistical analysis was made of the temperature values of the two stations. Five sets of temperature measurements were made for each test, and standard deviations of the upstream and downstream distribution were made for each set. An average upstream standard deviation and an average downstream standard deviation were obtained from the five sets. A value of effectiveness, which would give an indication of mixing capability, was calculated by subtracting from unity the ratio of downstream to upstream standard deviations. The mathematical relationships used for determining both the standard deviation and the effectiveness are defined by eqs (1) and (2).

unity the ratio of the average downstream to upstream ranges. This relationship is shown in eq (3):

Range DS

Effectiveness

=

(Range)

(1.

1.0

Range us

× 100%

(3)

where:

Range Ds=

= average (max min) downstream

This latter method of determining effectiveness, based on the range, is much simpler to calculate but could be more influenced by a single temperature value. If, for instance, either the maximum or minimum value were in error or differed considerably from the average value, then the effectiveness based upon the range would reflect this one value to a greater extent than would the method based upon standard deviation.

Since the air flow was in the turbulent region for these tests, it was of interest to determine to what extent the air was mixed during the simple process of flowing along the duct with no mixing device in the duct. Earlier tests [1] with no mixing device in the duct showed that mixing caused by the inherent turbulence and duct configuration was less than 7 percent when calculated by the range method of determining effectiveness. When calculated by the standard deviation method, a higher value of 17 percent was obtained for the natural mixing because this method takes into account all of the observed values. Some mixing, which is not indicated by the range method, does occur at the interface between streams of different temperature air.

4.1. Louvered Strip

Table 1 is a summary compilation of all of the tests conducted on the louvered strip mixing device and gives details of the test conditions and resulting performance at these conditions. In all illustrations which relate effec

fourth of the stream at a different temperature than the remaining three-fourths is shown in the center of figure 7. From the results shown by the three bar graphs in figure 7, it is evident that the lowest effectiveness occurred for a pattern composed of 50 percent warm air and 50 percent cool air with the interface along the diameter of the duct. The next most difficult pattern of nonuniformity to mix was that of three-fourths of the stream at a different temperature than the remaining one-fourth. The easiest pattern of nonuniformity to mix was that of a concentric distribution. With the concentric pattern, the 99.0 percent effectiveness of the mixer approached very nearly the limitation of the temperature-measuring system.

It was also of interest to determine how the magnitude of an initial temperature difference might affect the value of effectiveness. Tests using the concentric temperature pattern were made by varying the temperature difference between the warm and cool portions of the air stream. Observations of temperature differ

ences between portions of the unmixed stream of 4.0, 6.5, and 15.5 deg F revealed no change in the percentage effectiveness of the mixer. In studying all other parameters, a temperature difference in the unmixed air stream of approximately 3 to 4 deg F was used for all mix

ers.

Using two mixing elements in series, it was desirable to know at what position the elements should be located with respect to each other for best mixing. Tests were made to determine the optimum placement by varying the location of the second element with respect to the first element. Figure 8 shows a curve of effectiveness as the distance between the mixing elements was changed. As a result of space limitations, the overall distance from the inlet of the mixing device to the downstream measuring station was confined to 4.75 duct diameters. Because of the space limitation, the distance from the second mixing element to the downstream measuring station varied as the distance between mixing elements varied. Under the conditions of the tests, the effectiveness increased gradually from 92.5 to 96.1 percent as the spacing was increased from 0.5 to 1.5 duct diameters; it remained constant at spacings greater than 1.5 duct diameters.

Figure 9 presents the effectiveness as a function of overall distance between the first mixer element and the downstream measuring station with the spacing between mixing elements held constant at 2.0 duct diameters. Presented in the figure is the performance of the mixer for two temperature patterns, the concentric distribution and the quadrant distribution. As can be seen, mixing effectiveness for the concentric pattern was consistently the higher of the two.

Measurements were made over a selected range of air flow corresponding to average air stream velocities ranging from 150 to 600 fpm. Figure 10 illustrates the negligible effect upon the effectiveness of mixing as the flow or mean velocity was varied over the above range.

Measurements were also made to compare the effectiveness of the mixer at two different louver angle positions. The results illustrated in figures 7 to 10 inclusive are for a position of 60° from the normal path of flow along the axis of the duct. Using the quadrant temperature distribution pattern, the effectiveness decreased from 96 to 93 percent for the respective settings of 60° and 45° from the normal path of flow. For the comparative tests, the flow rate was 1400 cfm, the spacing between mixing elements was 2.0 duct diameters, and the downstream measuring station was 4.75 duct diameters from the first mixing element.

The loss of static head due to the presence of a mixing device in the duct system was of interest because of the energy required and size of equipment needed to move the air through such devices. The pressure loss from the louvered strip mixer was relatively low, as is shown in the plot of figure 11. When the loss in static pressure head is expressed in terms of equivalent multiples of velocity head, the static pressure drop is equivalent to approximately

100

seven velocity heads. For the other two mixers, the louver-baffle and the concentric louver, the equivalent velocity head multiples are 38 and 5, respectively.

4.2. Louver-Baffle

A summary of the test conditions and performance of the louver-baffle mixer is given in table 2.

The same three temperature patterns as previously described, having quadrant and annular distributions, were used in evaluating the performance of the louver-baffle mixing device. Shown in figure 12 are bar graphs which compare the performance for the three temperature patterns. By comparing figures 6 and 12, it is seen that the effectiveness of this mixer is not as much affected by temperature distribution as was that of the louvered strip mixer.

From the curve of figure 13, it is seen that there was little change in effectiveness beyond the 3 deg F level of temperature difference. It is possible that the computed values of effectiveness for temperature differences of less than 3 deg F at the mixer inlet (0.6 and 1.7 deg F) could have been influenced by the instrument error, since the ratio of the instrument error to the temperature gradient increases as the temperature gradient gets smaller. At the downstream measuring station, an error of 0.01 deg F for an initial temperature difference of 0.5 deg F between quadrants would cause an error of 2 percent in effectiveness. To reduce the influence of instrument error upon the observed effectiveness, all other tests were conducted at a temperature difference of 3 deg F or higher.

Tests were conducted to determine at what distance the two elements should be placed with respect to each other and at what overall length the mixer would be highly effective (97% or higher). Figures 14 and 15, respectively, show the effect when the distance between elements is varied, and the effectiveness