TECH TOPICS

Dave Register Bartlesville, Oklahoma regdave@aol.com Planform Optimization for Hand Launch and Discus Launch Gliders, Continued

icking up from last month, well continue our review of planform optimization for Hand Launch and Discus Launch Gliders (HLG, DLG). The driver for this study is to get a better feel for the appropriate aspect ratio (AR: wing span/average chord) for a good DLG design. Previous work in this column reported a value of ~ 10 as pretty good for a 2 channel HLG (javelin launch). In that case we dealt with a total glider weight of ~ 8 oz. using the SD7080 airfoil. DLG designs require a bit more beef in the wings and fuselage to survive the launch stresses which can reach upwards of 80Gs. They also require somewhat larger vertical stabilizers for yaw correction of the spin and yawroll coupling at the moment of release. Additionally, were seeing more control surface functionality (flaps, ailerons). All of these factors add to the total weight of the sailplane. Coupling this with another trend - the move towards thinner, lower lift airfoils using laminar flow and weve got several interesting twists to the earlier problem. Last month we developed the basic input data for the project. Using Prof. Mark Drelas X-Foil code, we calculated Cl and Cd data for the SA7035 and AG12 airfoils as a function of Reynolds number (Re - Epplers notation). The SA7035 represents the general class of airfoils used a few years ago for HLG while the AG12 represents the newer class of airfoils used for DLG. The Cd vs Cl curves are repeated here as Figures 1 and 2. As noted, the drag bucket is wider and the Cl is higher for the SA7035. But the AG12 is definitely better behaved in the lower Re range
Page 12

which is important for the cruise and thermal stage of flight. At high Re and low Cl (launch conditions) the Cd values for both sections are comparable.

Since were going to perform a polar analysis (sink rate vs horizontal velocity over the flight profile of the sailplane), we need to make some consistent assumptions about weights, moment arms, fuselage area and empennage. That data is summarized in Table 1 for the HLG case and Table 2 for the DLG case. Three differences should be noted:

R/C Soaring Digest

for this volume coefficient may use S rather than S/2 so be sure you know which is being used by the particular author. The numbers are taken from a series of HLG designs Ive built over the past several years. That basic design was converted to DLG. Some trial and error (Umm, that really means wreckage!) resulted in the extra structures that I needed to add to keep things in one piece. You may be able to build lighter and stronger so well guesstimate from this work what the likely trends could be as a function of weight. In Table 2 weve done the SA7035 DLG only for a 2 channel ship. Since the AG12 works well in this example, weve added estimated weights for 4 servo and 6 servo cases to see how wing loading affects things with this airfoil. Comparing polars for the simple HLG case, weve plotted the results (Figures 3 and 4) for AR of 6, 8, 10 and 12. Although this may seem an extreme range, some craft have been made at the 7 end while a few have been flown at 11 or slightly higher. So were bracketing the wing planform values actually used at the field. For HLG, good minimum sink is very important. But its also important to achieve that at a relatively low speed so you can linger longer in a light thermal. Additionally, youd like to be as efficient as possible when youre cruising around so a good Lift to Drag (L/D) ratio is also desired. In the earlier work with the SD7080, the UIUC database was used for Re dependence. Since this data ended at 60k, and the higher aspect ratios require values down to ~ 35k, a linear extrapolation of that data was made to the lower range. As can be seen in Figure 1, below 50k a linear extrapolation of Cd with Re is not appropriate. Consequently, the poorer performance of the SA7035 at AR = 10 and 12 is likely due to the high drag at low Re exhibited by this section. For the AG12, an aspect ratio of around 10 appears to be a reasonably good optimum. However, anywhere in the 8 to 10 range is probably indistinguishable in practice. So I would conclude that using the older large
Page 13

1)

The wing area density is higher for the DLG. This represents the stronger construction needed for this case, The vertical (rudder) volume coefficient (VVC) is larger for the DLG. This represents the subrudder contribution needed to maintain good initial launch trajectory, and The fuselage weight is higher for the DLG. This is the added structure needed to withstand

high G launches as well as the piezoelectric gyroscopic stabilizer circuit (lingua franca: piezo or gyro) used for launch control. Please note that my definition of VVC is: VVC = (Av * Dm) / (Aw * S/2) where Av is the vertical stabilizer area, Dm is the moment arm ( chord wing to chord stabilizer), Aw is the wing area and S/2 is the total span (30 inches in this case). Some definitions

2)

3)

December 2002

plane airfoils should tend towards the low end of this range to avoid poor Re performance. The more recently designed laminar flow sections can still perform well at the upper end of this aspect ratio range. Whats interesting to note is that the polars all become broader as an inverse function of Re at higher speeds. Although not shown here, this is the opposite of the tendencies for this same type of calculation for open class ships. For the HLG/DLG wingspan, a high aspect ratio apparently produces poor performance in the intermediate speed range due to Re scaling. For open class ships it appears you are above a critical Re value for these problems and higher aspect ratios are generally favored over the entire speed range. If we look at an expanded speed range for the AG12, we find that the curves indeed cross over to the expected tendency at speeds in excess of 65 ft/ sec. This is shown in Figure 5 only for AR 6 and 12 to keep the graph from being too cluttered. I hope this emphasizes the need for careful consideration of Re effects at the scales encountered with HLG and DLG designs. This factor is not nearly as critical for open class ships. Turning our attention to the DLG weight class, for the two channel configuration, the polar is shown in Figures 6 for the AG12. Although the increase in weight is only a few ounces, it tends to shift the optimum performance to a slightly lower aspect ratio. This trend favors both Re and wing loading and suggests that a value of ~ 9 may be a good choice. For the DLG case we can now add two more parameters for evaluation - the estimated launch height and the estimated flight time at minimum sink. These numbers are derived from a launch height calculation which is coupled to the polar program and uses the calculated drag coefficients derived from the polar evaluation. To arrive at these numbers, an estimated release velocity and launch angle is assumed to calculate the launch height. The flight time is then the sum of the launch time and the value for the height divided by the minimum sink rate.
Page 14

For this study, the launch angle is 70 degrees and the release velocity is 65 ft/sec. The results will roughly scale with initial launch velocity. The launch height for a given velocity is not much changed for angles > 60 degrees. So these are conservative, but practical numbers for sport flying. It is notable that the launch heights are all comparable for most cases but the flight time varies somewhat more

widely. This is mostly in response to the change in minimum sink. Comparing these values with the minimum sink and max L/D terms we also find an optimization somewhere between 8-10 for the aspect ratio. In Figures 7 and 8, the polars for the AG12 for a 4 servo and 6 servo estimated ship are also shown. There are subtle differences as the weight is increased. The minimum sink, max L/
R/C Soaring Digest

advantage by increasing effective camber for light air conditions and using reflex for launch and moving around in the wind. But for the simple pleasures of sport flying, there is no disadvantage (and maybe even an advantage!) to keeping it real simple. The case of flaperons may be a very nice compromise for contest conditions (4 servos). As a reality check on the overall results, an AR of 8 converts to about 450 sq.in. wing area while an AR of 10 gives 360 sq.in. In prowling the web for present day designs, I find that most ships bracket the 350 sq.in. to 420 sq.in. range. One could claim Q.E.D. and leave it at that. But why should one celebrate about predicting the obvious? Well, if these calculations had said that an AR of 16 is great for a 4 channel DLG, they wouldnt have much credibility. Since the results do seem to match with reality, we can continue to use this as a tool to simulate performance for conditions we havent yet tried (different airfoils, details of wing layout, etc.). Ultimately, the proof is at the flying field, but studies like this can provide some confidence in the expected performance of untested configurations. I also like knowing that for just messing around, a properly designed two channel ship with a real light sport type radio (Hitec III, for instance) is probably going to give me just about as much performance as a top end multi-function sailplane. Now I wonder what would happen if I change the base airfoil to increase camber for the higher wing loading? Do I get back to a somewhat higher aspect ratio? Sounds like another study is developing here. Lets look at that next time. D and flight time values optimize around an aspect ratio of 8 at the higher weight. So we can conclude that a heavier plane probably favors a somewhat lower aspect ratio. However going much below 8 does not seem to offer any significant advantage - unless you build somewhat heavier than I do. Although the changes appear subtle, a final comparison of a subset of this data is shown in Figure 9. This plot
December 2002

compares the polar and associated key flight parameters, for the 2 channel and 6 channel DLG. Here we see that the 2 channel DLG in straight flight will generally outperform the 6 channel estimation for minimum sink, total flight time and efficiency as seen in the max L/D value. In the heat of battle (contests) the added functionality of the multifunction ship very likely has the

Until then, hope youve had a happy holiday. Remember that flying something is a whole lot better than flying nothing so have fun and dont sweat the details too much! (Figures 7-9 are on the next page! Ed.)

Page 15

SCHEDULE OF SPECIAL EVENTS

February 1-2, 2003 Southwest Classic Phoenix, AZ March 15-16, 2003 The Classic Mid-Winter Southern California Torrey Pines Vintage Sailplane Regatta http://www.agcsc.org May 15-18, 2003 Midwest Slope Challenge Wilson Lake, KS www.alltel.net/~mwsc May 24-25, 2003 So. California PSS Festival Cajon Summit, CA Brian Laird, Slope_Scale@compuserve.com <ourworld.compuserve.com/homepages/ slope_scale> October 10-11, 2003 Texas National Tournament (TNT) Dallas, TX www.SLNT.org

For Sale - Personal Dodgson Camano 100 with E205 Eismann Jet JM Glasscraft Songbird 78 JM Glasscraft Songbird 100 Sailplanes Intermatal RacerCS Sailplanes Intermatal Axle $120.00 $175.00 $75.00 $75.00 $150.00 $150.00

Craig Christensen, 3261 45th Street, Webster, MN 55088; 952-652-5483 after 4:00 PM.

Page 16

R/C Soaring Digest

THE NATURAL SIDE OF THERMAL SOARING

by Lee Murray Appleton, Wisconsin lmurray@athenet.net The degree of stability of the lifted air determines whether cloudiness will be predominately stratiform (stretched horizontally) or cumuliform (stacked vertically). In chapter 5 of his book4 , Dennis Pagen talks about the weather fronts, wind direction and lift. A comparison is made to cooler sea breezes that displace and run under warmer air. On a larger scale, the situation of cold fronts converging with warm, moist air often involves thunderstorms, a condition not recommended for full size soaring. On the day I mentioned, a cold front arrived from the northwest, and with it, a distinct line of dark clouds with increasing elevation into the distance. The cold air was coming in from Canada with a North West direction. The ground level warm breeze was out of the West. You can see the wind directions on the left side of the Soarcast Plot in Figure 2. The angled lines show wind direction (going straight up indicating wind out of the north). A line extending from the left edge of the plot to the right is a graph of the wind velocity.

Time Deg F Dew Point 8:45 59.0 57.2 9:45 60.8 51.8 10:45 64.4 46.4 11:45 64.4 44.6 12:45 66.2 42.8 13:45 66.2 39.2 14:45 66.2 39.2 15:45 64.4 35.6 16:45 62.6 39.2 18:45 59.0 42.8

%RH 0.94 0.72 0.52 0.49 0.43 0.37 0.37 0.34 0.42 0.55

Event Scattered Clouds Scattered Clouds Scattered Clouds Scattered Clouds Scattered Clouds Mostly Cloudy Mostly Cloudy Mostly Cloudy Overcast Overcast

Temperature history for September 21st, 2002 in Appleton, WI.

n a recent column, I wrote about a frontal system coming through during a local thermal duration contest and how it affected the flight times. The lift was neutral to certain point then became good as the ground level air temperatures reached a trigger level allowing the height of the mixed boundary layer to go much higher1 . During a recent fun flying experience, a point was reached when the lift changed from neutral to wide-areas and long in duration2 . The Soarcast* indicated that it should have been a good day. However, flight times had been about 3-5 minutes until about 3 PM when they were 20+ minutes for several of us. Was it just some fluke in the stability of the atmosphere not
December 2002

predicted by lapse rate and ground level air temperature or did it have something to do with the approaching cold front? This is the topic of this article.

The confluence, or the frontal line, where the air was mixing was advancing slowly to the east as the afternoon The book Aviation Weather describes went on. According to one theory, the what can happen with frontal weather colder air was slipping underneath the when masses of air with different warmer air as shown in Figure 1. temperatures mix3 . Models in the air were gaining altitude as the air they were in was generally Weather occurring with a front depends on being forced up. As we flew the sky (1) the amount of moisture available, (2) became darker, we packed up and left. the degree of stability of the air that is A few hours later it was raining. forced upward, (3) the slope of the front, (4) the speed of frontal movement, and (5) Weather front theory has to get a little the upper wind flow. more complex to explain some limited areas where the air was not going up. I Sufficient moisture must be available for found two possible explanations for clouds to form, or there will be no clouds.
Page 17

this, either one or both of which may have existed. Explanation 1: The ground was now warmer than the air filling in under the rising warmer air. This gave the possibility of convective cooling and sinking air mixed in with the general rising air. Explanation 2: There was turbulence along the interface of the two air masses that would cause areas where the model would be sinking. Weather information was downloaded from the Internet: The temperature, dew point, wind, RH and conditions history was recovered from: http://www.wunderground.com/ history/airport/KATW/2002/9/21/ DailyHistory.html KATW is the abbreviation for Appletons Outagamie County Airport just a few miles from the flying field. The conventional theory for why we had good conditions so late in the day would rely on the temperature history for the day and the lapse rate. The lapse rate data from Green Bay (about 25 miles away) suggests that the temperature exceeded the trigger temperature by noon. The good thermals didnt happen when expected, perhaps because of some special condition imposed by the weather font. Or perhaps the very local conditions had too much evaporative cooling to participate in the convective cooling (thermal processes) until more instability took place. The scattered clouds in the weather history earlier in the day suggest this may have been the case. The exercise of looking at weather information should be useful to you in being able to analyze your local weather conditions. I believe the more in-tune an RC or full size sailplane pilot is to his local atmospheric conditions, the better thermal flyer he can be. Im closing with a photo of a line of cumulus clouds seen in late October in SW Florida shown on page 17. This is the result of converging coastal front of cooler coastal breezes running under
Page 18

Wind Speed, mph/Temperature, F

warmer interior air with enough moisture to form clouds. Why there isnt more full size thermal soaring in Southern Florida? Perhaps the air is too unstable. I bet some readers have some answers. 1 2 RCSD: Jan 99 pg10, May 99 pg 24, Sep 99 pg 22, Apr 00 pg 22 and Jul 00 pg 14. A method developed by Pearson predicts when thermals will lift sailplanes to specific altitudes when a temperature is reached.

Soarcast Version 1.0.4 was provided as a free download from the web page of the Soaring Society of America. Newer versions are currently available. US Depts. of Transportation and Commerce, Aviation Weather for Pilots and Flight Operations Personnel Reprinted by ASA Publications, 1975 ed. Pagen, Dennis, Understanding the Sky, p. 185, Dennis Pagen Publisher.

ZIKA

R/C Soaring Digest