1、OR-05-14-2 (RP-1010) Inlet Distortion Effects on the Air Performance of Axial Flow Fans John A. Murphy, PhD Member ASHRAE ABSTRACT The air performance efects expected due to inlet velocity distortion for axial fans is predicted based on basic airfoil data. This expectation is compared with the test
2、data gathered in Rp-1010. Sound data for various inlet conjgurations are also presented to help understand the test data. INTRODUCTION ASHRAE research project 1010 was conducted to deter- mine the effects of the inlet connection on the air and sound performance of axial flow fans. This project invol
3、ved the test- ing of three axial fans, all of the same diameter but with hubs of different size. Each fan was tested at four different blade angles. A corner, consisting of two perpendicular walls, was placed near the fan to produce some inlet flow distortion. The comer was placed at ten different l
4、ocations with respect to the fan inlet. The complete results of the project can be examined in the final report, and an excellent summary of these results is given by Dr. William Swims paper (Swim 2005). The purpose of this paper is to discuss why inlet flow distortion can be expected to affect the
5、fan performance and just how these effects might be manifested. BACKGROUND Although some knowledge of the details of axial fan design is certainly desirable, a complete discussion of the subject is beyond the scope of this paper. Understanding the basic velocity triangle, however, is crucial to the
6、discussion of inlet distortion effects. The aerodynamic angle of attack of any blade section is determined by the relationship between the rotational velocity and the axial velocity. Figure 1 shows the velocity triangles for the hub, pitch, and tip sections of a typi- cal axial fan. The blade twist
7、is selected so that the aerody- namic angle of attack is approximately constant at all radii. As the flow is decreased, i.e., by closing the throttle during a performance test, only the axial velocity decreases and the angle of attack increases. Thus is the classical slope of the performance curve a
8、nd the existence of a stall (if the blade setting angle is high enough). It is always assumed that the velocity at any given radius is independent of angular position. For simplicity it is assumed that the axial velocity is indepen- dent of the radial location. This assumption is not valid for many
9、designs, but the effects considered in this paper and the conclusions reached will still be accurate. INLET DISTORTION When the fan inlet is situated so that the flow is non- uniform at the fan inlet, i.e., the fan inlet is placed in a corner, the axial flow will no longer be independent of angular
10、posi- tion. The rotational component of the velocity triangles is unchanged. This means that the angle of attack will vary as any individual blade rotates, and at any particular time some blades will be at higher angles than the mean (averaged around one complete rotation), while others will be at l
11、ower angles. To evaluate what effect, if any, this will have on the fan perfor- mance, we need to look at some basic airfoil data. NACA conducted a series of cascade airfoil tests, and the results (for our purposes) may be summarized in Figure 2. This figure shows the turning angle (which is directl
12、y proportional to pressure rise) and 1iWdrag (LD) ratio plotted versus the angle of attack for a typical airfoil. Notice that peak L/D (which corresponds to peak efficiency) occurs at an angle of attack close to peak turning. Also notice that the turning versus the angle of attack curve has very lit
13、tle curvature except near the peak. The WD versus the angle of attack curve shows signif- icant curvature at most angles of attack. John A. Murphy is vice president at JOGRAM, Inc., New Philadelphia, Ohio. 1 O00 02005 ASHRAE. ro tallo n axial IO tatio n Ixla1 Figure 1 Velocity triangleS.for the hub,
14、 pitch, and tip sections of a typical axial fan. 60 50 40 20 10 O O 5 10 15 angle of anack Figure 2 Results of the NACA cascade airfoil tests. 20 25 ASHRAE Transactions: Symposia 1 O01 7.00 6.00 5.00 4.00 I- n. 3.00 2.00 1 .o0 0.00 O 5000 10000 i 5000 20000 25000 30000 35000 40000 CFM Figure 3 Press
15、ure vs. volume curves for the 36-1 7 fan at 25 degrees blade angle and all wall positions. For simplicity, assume an inlet velocity pattern that has a single maximum and minimum value as a function of angle, and further assume that the deviation is approximately anti- symmetric (i.e., the portion of
16、 the inlet face seeing velocities below the average is approximately the same size as the portion seeing velocities above average). As an individual blade travels through a complete rotation, the angle of attack will be greater than the average value when the velocity is below average and will be le
17、ss then the average when the velocity is above average. Since the turning versus the angle of attack curve has little curvature, the pressure rise will be essen- tially unaffected by the velocity distortion unless the critical angle of attack is exceeded. Some additional effect will occur due to the
18、 performance of the straightening vanes, but this effect should also be small. Because of the increased curva- ture of the L/D curve, the efficiency effect should be slightly more pronounced. It has been this authors experience that velocity distortions of +20% are not unusual with fans located in c
19、orners. Figure 3 shows the pressure versus volume curves for the 36-17 fan at 25 degrees blade angle and all wall positions. Figure 4 shows the efficiency versus volume curve. Examina- tion of these curves reveals that the only performance effect is a slight movement of the peak pressure point (towa
20、rd higher flows, as would be expected). The shape and position of both curves is not significantly influenced by wall position except for the “worst”case (OD, 0.5 D). The performance loss for this t5D. 50 -t-D, ZD D, 1.5D -x-D, D +SD, 15D +.5D, D -o, 1.5D O, .5D wall position can be explained by rec
21、ognizing that the flow into the fan must pass through the space between the inlet bell and the wall in front of the fan and that all the flow must come from one side and the top. This flow must turn 90” to enter the fan. The measured loss equates to essentially one velocity head for this flow. Figur
22、es 5 through 9 show the 1/24th octave band data for the 36-17 fan at 25 degrees blade angle plotted versus wall position. Only four wall positions are shown (5D, 5D; 0.5D, D; O, D; O, 0.5D). Figures 10 through 12 show the same data for the 18 degree blade angle at free flow, point 3, and top. These
23、data show that at the closest wall position, the turbulence level in the flow at the rotor is higher than for any other wall posi- tion. An interesting anomaly shows up in the 18 degree data at the top of the curve. Notice the high sound levels below 600 Hz. This may be due to this performance point
24、 being taken a little farther into the incipient stall region than the top at 25 degrees. It is very interesting, and unexpected to this author, that the change in efficiency due to inlet distortion was essentially zero until the wall was placed unreasonably close to the inlet. The sound data also i
25、ndicate that there is no significant change in the character of the flow. Perhaps the inlet bell on this fan does a surprisingly good job of directing the flow into the fan inlet. Or perhaps the fact that the walls did not go clear to the ceiling, thus allowing significant flow over the top ofthe wa
26、lls, can explain the results. 1002 ASHRAE Transactions: Symposia 80.0 70.0 60.0 50.0 E Y 40.0 w 30.0 20.0 10.0 o .o o 5000 1 0000 i5000 20000 25000 30000 35000 40000 CFM Figure 4 Eficiency vs. volume curves for the 36-1 7 fun ut 25 degrees blade angle und all wallpostons. 90 80 70 60 50 d P m 40 30
27、20 10 Figure 5 1/24th octave bund datu for the 36-1 7 fun ut 25“ blade angle at freeflow. ASHRAE Transactions: Symposia 1003 90 80 70 60 50 iI n a 40 30 20 O O 20 40 60 80 1 O0 120 140 160 180 20 o Band Number -+-5D,5D +O 5D,D 0.D -*sO,O 5D 1 Figure 6 1/24th octave band data for the 36-1 7 fan at 25
28、“ blade angle atpoint 2. 90 80 70 60 50 iI P a 40 30 20 10 O O 20 40 60 80 1 O0 120 140 1 eo 180 200 Bend Number 1&5D 5D +O 5D,D 0,D 4- 0.0 5D 1 ._ Figure 7 1124th octave band data for the 36-1 7 fan at 25“ blade angle atpoint 3. 1004 ASHRAE Transactions: Symposia O 60 80 t O0 i 20 i 40 160 180 200
29、20 40 +5D,Sd -.-O 5D,D O D L _ . _. - .-. . Figure 8 1/24th octave band data for the 36-1 7 fan at 25“ blade angle at point 4. 90 80 70 60 30 20 10 0 O 20 40 60 80 100 120 140 160 180 200 Band Number -5D.5D +O 5D,D O D &O O 5D Figure 9 1124th octave band data for the 36-1 7 fan at 25“ blade angle to
30、p. ASHRAE Transactions: Symposia 1005 90 80 70 60 50 iI a m 40 30 20 10 O O 20 40 60 80 1 O0 120 140 160 180 20 o Band Number &5D,5D +O 5D.D O,D -+o,o 50 I Figure 10 1124th the octave band data for the 36-1 7 fan at 18“ blade angle freeflow. 90 80 70 60 50 iI m a 40 30 20 10 O O 60 80 1 O0 120 i 40 160 i80 20 o 20 40 Bnad Number -5-5D.D OD i: a 40 30 20 io O O 40 60 80 100 120 140 I60 180 200 20 Bend Number -5D.5D +O 5D,D OD -x- 0,OdD Figure 12 1/24th octave band data for the 36-1 7 fun at 25“ blade angle top. ASH RAE Transactions: Symposia 1007
