creation-of-a-novel-tool-for-the-design-and-evaluation-of-uas-propellers

Paper ID #35078 Creation of a Novel Tool for the Design and Evaluation of UAS Propellers Mr Brett Dekker Bennett, Baylor University M.S.M.E Student at Baylor University B.S.M.E May 2019, Baylor University Dr Kenneth W Van Treuren, Baylor University Ken Van Treuren is an Associate Professor in the Department of Engineering at Baylor University He received his B S in Aeronautical Engineering from the USAF Academy in Colorado Springs, Colorado and his M S in Engineering from Princeton University in Princeton, New Jersey After serving as USAF pilot in KC-135 and KC-10 aircraft, he completed his DPhil in Engineering Sciences at the University of Oxford, United Kingdom and returned to the USAF Academy to teach heat transfer and propulsion systems At Baylor University, he teaches courses in laboratory techniques, fluid mechanics, energy systems, and propulsion systems, as well as freshman engineering Research interests include renewable energy to include small wind turbine aerodynamics and experimental convective heat transfer as applied to HVAC and gas turbine systems c American Society for Engineering Education, 2021 Paper #35078 Creation of a Novel Tool for the Design and Evaluation of UAS Propellers Brett Bennett, Kenneth Van Treuren: Advisor Mechanical Engineering Department Baylor University Abstract Recent times have seen a tremendous increase in the development and deployment of Unmanned Aircraft Systems (UASs) for both military and commercial applications As the utilization of UASs continues to increase, the design of improved propellers for these UASs is an important area for research The power requirements of UASs are highly influenced by how efficiently electrical power can be turned into propulsive power Employing a more efficient propeller allows a UAS to either operate longer using the same power source or to use a smaller power source to improve the performance A major source of inefficiency in propellers is the induced drag created by the vortices from the tip of the propeller One potential method to reduce these vortices and increase efficiency is unloading the tip of the propeller by specifying the distribution of the lift coefficient (𝐶𝐿 ) A common method for designing and evaluating propellers is the Blade Element Momentum Theory (BEMT) Currently, the QMIL and QPROP programs, based on BEMT, are the most frequently employed programs for designing and creating propellers However, QMIL and QPROP have several limitations that make them difficult to use They use DOS commands which can be complicated to run on newer operating systems These programs are also limited in their ability to accurately model 𝐶𝐿 and the drag coefficient (𝐶𝐷 ) as a function of angle of attack (𝛼) and to create a propeller with a prescribed 𝐶𝐿 distribution Due to the limitations presented by QMIL and QPROP, an Excel spreadsheet, called CLPROP, was created using BEMT to enable the design and evaluation of propellers with a prescribed 𝐶𝐿 distribution CLPROP allows for the rapid comparison of potential propeller designs This enables the determination of optimal configurations for testing and further assessment CLPROP’s accuracy was validated by testing two different propeller designs utilizing the ClarkY airfoil It produced very similar results to QPROP CLPROP when compared to experimental results underpredicted RPM, torque, and power required, however, when correcting for the torque contribution caused by the weight of the blade, CLPROP’s predictions for the experimental torque and power required were very similar Introduction Recent times have seen a tremendous increase in the development and usage of UASs for both military and commercial applications With this rapid increase in UAS utilization, there is an opportunity for UAS design improvement One such area for improvement is the design of propellers Propellers must become more efficient More efficient propellers result in UASs Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education using less power to operate, leading either to an increased range or time airborne An increased range allows UASs to carry out operations over larger areas which could impact their use in logistics services such as package delivery Therefore, it is important to develop new propeller designs to improve their performance A major source of inefficiency in propeller operation is induced drag resulting from the vortex flow generated at the tip of a propeller This produces an increase in power required for the propeller to operate One method of potentially improving a propeller’s efficiency is by altering the propeller’s lift characteristics to minimize the loading at the tip of the propeller, thus, lowering the induced drag Currently, the widely used program for propeller design is Massachusetts Institute of Technology’s (MIT) QMIL and the primary program utilized for evaluation of propeller design’s is MIT’s QPROP QMIL and QPROP are limited in their ability to create a blade with desired lift characteristics Most of a blade’s thrust is produced near the tip and inaccurate performance predictions can occur QMIL and QPROP model the relationship between coefficient of lift (𝐶𝐿 ) and angle of attack (𝛼) as linear In many cases, especially when 𝐶𝐿 is near zero as is the case when the propeller tip is being unloaded, the relationship between 𝐶𝐿 and 𝛼 is not entirely linear for many airfoils which causes inaccuracies Additionally, QMIL and QPROP can be both difficult and time-consuming to use due to their reliance on DOS commands Therefore, a Microsoft Excel based program called CLPROP was created utilizing Blade Element Momentum Theory (BEMT) It allows the design of propellers with a userdesignated 𝐶𝐿 distribution, resulting in propellers being quickly created and compared for rapid determination of optimal designs CLPROP designed propellers results were compared to QPROP results for the same propellers and tested experimentally to validate the accuracy of CLPROP’s performance predictions Theoretical Background BEMT is a commonly utilized method to calculate the performance of a given propeller It combines momentum theory and blade-element theory (BET) into a more complete method for evaluating a propeller’s performance The momentum theory evaluates a propeller’s performance by calculating the change in momentum of the stream-tube as it passes through the propeller disk This determines the momentum imparted on the flow by the propeller which causes the flow to accelerate as shown in Figure The thrust (𝑇) produced by the blade is determined as a function of the fluid’s density (𝜌), the propeller disk area (𝐴), the freestream velocity (𝑉∞ ), and the induced axial velocity (𝑣) as shown in Eq (1) [1] 𝑇 = 𝜌𝐴(𝑉∞ + 𝑣)2𝑣 Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education (1) Figure 1: Control volume analysis for the momentum theory [1] Often, when evaluating a propeller, it is divided into small radial subsections to characterize the changes in performance of a propeller over the course of the blade (see Fig 2) The momentum theory can also be used to estimate the elemental thrust (𝑑𝑇) and elemental torque (𝑑𝑄) produced by each radial subsection as functions of 𝜌, 𝑉∞ , the axial induction factor (𝑎), the angular induction factor (𝑎′), the radial distance from the center of the hub (𝑟), the length of the subsection (𝑑𝑟), and the rotational speed of the propeller (Ω) using Eq (2) and (3) respectively [2] 𝑑𝑇 = 𝜌𝑉∞2 4𝑎(1 − 𝑎)𝜋𝑟 𝑑𝑟 (2) 𝑑𝑄 = 4𝑎′ (1 − 𝑎)𝜌𝑉∞ 𝜋𝑟 Ω𝑑𝑟 (3) The momentum theory does not account for the geometry of the blade or the aerodynamic properties of the airfoil [1] BET evaluates a propeller’s performance by dividing the blade into small radial subsections and determining the performance of each subsection based on the performance of the airfoil at the 𝛼 of each subsection as shown in Fig 𝑑𝑇 and 𝑑𝑄 produced by each subsection are calculated as functions of 𝜌, 𝑉∞ , Ω, 𝑟, the chord length (𝑐), 𝑑𝑟, 𝐶𝐿 , the coefficient of drag (𝐶𝐷 ), the geometric pitch angle (𝛽), and 𝛼 using Eq (4) and (5) [3] 𝑑𝑇 = 0.5𝜌(𝑉∞2 + (Ω𝑟)2 )𝑐 𝑑𝑟(𝐶𝐿 cos(𝛽 − 𝛼) − 𝐶𝐷 sin(𝛽 − 𝛼)) (4) 𝑑𝑄 = 0.5𝜌(𝑉∞2 + (Ω𝑟)2 )𝑐 𝑟 𝑑𝑟(𝐶𝐿 sin(𝛽 − 𝛼) + 𝐶𝐷 cos(𝛽 − 𝛼)) (5) Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education Figure 2: The components of BET [3] However, BET does not account for the induced velocities generated by the propeller blades which can substantially alter the lift and drag performance of each subsection [3] Therefore, these two theories are combined into BEMT BEMT combines the application of induced losses determined by the momentum theory with the determination of the performance of individual subsections in the BET By calculating the performance of individual subsections of a propeller, because of the induced performance of the subsection, a more complete picture of propeller performance is obtained This allows for the optimization of propeller performance throughout the entire blade Currently, MIT’s QMIL and QPROP programs are widely used for the design and evaluation of propellers QMIL designs propellers to operate at a designed thrust or power for specified operating conditions QPROP evaluates the performance of a propeller at specified operating conditions QMIL and QPROP both have shortcomings despite their widespread deployment Both programs require the utilization of DOS commands which can be difficult as the usage of DOS commands is not commonplace in modern computing Both programs include a single airfoil type over the blade and this airfoil’s characteristics are modeled as a linear relationship between 𝛼 and 𝐶𝐿 and a parabolic relationship between 𝐶𝐿 and the coefficient of drag (𝐶𝐷 ) While these models are typically accurate at an 𝛼 that is not near stall or low angle of attacks for most airfoils, they are not always precise especially when designing propellers to operate at near stall conditions In addition, QMIL does not allow the specification of a constant chord or to designate a chord distribution Also, while it allows a limited 𝐶𝐿 distribution to be specified, it often results in inconsistencies in the 𝛽 or chord distribution This leads to sudden changes in geometry over the blade with the potential for unexpected aerodynamic consequences QPROP is cumbersome to setup and evaluate the performance of a propeller, requiring the manual input of chord length, radial position, and 𝛽 for each substation of the propeller being evaluated Due to these difficulties, a new propeller design and evaluation program was created to allow for more design options, quicker design and evaluation, improved ease of use, and customization Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education CLPROP A Microsoft Excel spreadsheet, CLPROP, was designed based on BEMT to provide a propeller design program that is easier to operate, more robust, and easily edited In CLPROP, the user inputs a desired 𝐶𝐿 distribution relative to the radial stations that are being evaluated for the propeller design CLPROP then models 𝛼 and 𝐶𝐷 as 6th-order polynomial fits of 𝐶𝐿 This provides a much more robust model than used by QMIL and QPROP The user also inputs several design parameters into cells including 𝑉∞ , the number of blades (𝐵), atmospheric conditions, the airfoil conditions at the maximum allowed 𝛼, the desired blade tip, and the design thrust The user then enters the value of the desired chord length (𝑐) for the propeller To create a propeller, 𝛽 needs to be determined at each radial station This is accomplished by calculating 𝐶𝐿 as a function of the helix angle (𝜙), Prandtl loss factor (𝐹), 𝐵, 𝑟, the tip radius (𝑅), the local wake advance ratio (𝜆𝑊 ), the radial speed ratio (𝜆𝑟 ), the local blade solidarity (𝜎′), 𝑉∞ , Ω, and 𝑐 using Eqs (6-10) The relationship between 𝐶𝐿 and 𝜙 is obtained by relating 𝑑𝑇 and 𝑑𝑄 obtained from the momentum theory with 𝑑𝑇 and 𝑑𝑄 obtained from BET and then solving for 𝐶𝐿 as a function of 𝜙 [2], [4] 𝜎′ = 𝐵𝑐 2𝜋𝑟 (6) 𝜆𝑟 = Ω𝑟 𝑉∞ (7) 𝑟 tan 𝜙 𝑅 (8) 𝜆𝑊 = 𝐹= 𝐵 𝑟 ) ( )(1− )( 𝑅 𝜆𝑊 ) cos −1 (𝑒 𝜋 cos 𝜙 − 𝜆𝑟 sin 𝜙 ) 𝐶𝐿 = −4𝐹 sin 𝜙 ( ′ 𝜎 (sin 𝜙 + 𝜆𝑟 cos 𝜙) (9) (10) Next, the Excel GRG Nonlinear solver is utilized to determine 𝜙 of each subsection and the minimum RPM at which the design parameters can be satisfied The program accomplishes this by solving for 𝜙 required at each subsection to equate the calculated 𝐶𝐿 and the desired 𝐶𝐿 while iterating the value of the RPM until the minimum value at which all design parameters can be satisfied is obtained Once 𝜙 is known at each point, 𝛽 can then be determined as a function of 𝛼 and 𝜙 using Eq (11) 𝛽 =𝛼+𝜙 (11) When CLPROP’s results converge, in addition to providing the 𝛽 angles necessary to manufacture the propeller, it estimates the expected torque to be produced by the propeller CLPROP also provides the expected propulsive power produced by the propeller and the shaft power required to spin the propeller as well as the propeller’s expected propulsive efficiency as Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education Figure 3: Outputs of CLPROP shown in Figure By rapidly determining these outputs, CLPROP allows for potential designs to be quickly accepted or rejected This prevents manufacturing resources from being used on ineffective designs Also, the widespread utilization of Microsoft Excel allows for operators to alter the program for other propeller design scenarios or to calculate additional parameters if desired This gives CLPROP an added element of adaptability that is not present in other design software such as QMIL and QPROP Experimental Methods To determine the accuracy of the results produced by CLPROP, two different 2-bladed propeller designs were created One design featured a constant 𝐶𝐿 distribution typical of most propeller designs, and the other a modified Prandtl bell-shaped lift distribution as described by Sanchez [5] The designs had a Clark Y airfoil because of its frequent commercial utilization and wellknown aerodynamic properties The 𝐶𝐿 , 𝐶𝐷 , and 𝛼 data for the Clark Y was obtained using XFOIL A maximum chord length of 1.5 𝑖𝑛 was selected so that the propeller would be thick enough to allow sufficient detail when manufactured Previous attempts at using smaller chords resulted in propellers that were unable to be produced or had easily damaged trailing edges The propeller design had an Oval tip profile as described by Liller [6] By utilizing a tapered tip profile, such as the Oval 2, the induced drag produced by the blade is reduced, thereby the torque 𝑓𝑡 produced by the blade is lower 𝑉∞ was selected to be 44 𝑠 as was a reference thrust of 0.5 𝑙𝑏𝑓 A propeller diameter of 𝑖𝑛 reduces the effect of wind tunnel blockage and eliminates the need for data corrections The U.S Standard Atmosphere at sea level determined the theoretical atmospheric conditions After the design criteria were selected, CLPROP designed the propellers The designed propellers were then analyzed using QPROP by varying the RPM at the selected design conditions until the design thrust of 0.5 𝑙𝑏𝑓 was achieved This allows for the comparison of CLPROP’s performance relative to both QPROP and experimental data Next, A modified MATLAB code originally written by Liller [6] created Solidworks macros for each blade The macros then rendered the individual blade and attached it to a dovetail used for inserting the blade into a center hub This process results in the Solidworks design shown in Figure Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education Figure 4: Clark Y Prandtl blade in Solidworks Each blade was manufactured using two different rapid prototype printers to confirm that the manufacturing process and material does not have an effect on the experimental results Blades were produced from acrylonitrile butadiene styrene (ABS) on an Objet 30 Printer and from polylactic acid (PLA) with a CraftBot 3D printer The ABS blades have a smoother surface finish than the PLA blades due to a higher resolution However, the PLA blades are lighter, weighing 0.01875 𝑙𝑏 each compared to the ABS blades which weigh 0.025 𝑙𝑏 each The resulting blades are shown in Figure Figure 5: Prandtl (top) and baseline (bottom) blades produced from PLA (left) and ABS (right) Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education Experimental testing was then conducted with the Baylor Research Innovative Collaborative (BRIC) Aerothermodynamics Lab low-speed subsonic wind tunnel The test stand, pictured in Fig 6, featured a motor supplied by the United States Air Force Academy mounted to an Interface MRT-0.2Nm Miniature Overload Protected Flange Style Reaction Torque Transducer and a Transducer Techniques LSP-1 load cell for measuring the thrust produced An Omega HHT20-ROS remote optical sensor (ROS) is also attached to the test stand to measure the propeller RPM Figure 6: Propeller Test Stand The torque transducer and load cell produced analog voltage outputs which were converted to digital signals using a National Instruments (NI) data acquisition (DAQ) system utilizing NI9205 voltage input modules in a NI 9178 DAQ chassis A LabVIEW Virtual Instrument (VI) developed by Sanchez [5] was used to record the measurements of the torque transducer and load cell and to convert the voltage signals into torque and thrust measurements utilizing known calibration constants RPM was manually recorded from a Monarch Instruments ACT-3X-1-1-10-0-0 tachometer that displayed a digital readout of the ROS’s measurement Propeller speed was controlled using a Futaba T4EXA remote-control capable of both course and fine adjustment of its throttle settings An RPM sweep was performed using an Advanced Precision Composites (APC) 8x6 propeller that had previously been tested to confirm the accuracy of the calibration, then the designed propellers were tested Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education Results and Discussion The predicted results from CLPROP and from QPROP are shown in Table Table 1: Comparison of CLPROP and QPROP analysis at 0.5 𝑙𝑏𝑓 Baseline RPM Torque (𝑙𝑏𝑓 ∗ 𝑖𝑛) Power Required (𝑊) CLPROP 4560 0.752 40.6 QPROP 4519 0.767 41.0 % Diff -0.90 1.98 1.07 Prandtl CLPROP QPROP % Diff RPM 6493 6492 -0.02 Torque (𝑙𝑏𝑓 ∗ 𝑖𝑛) 0.554 0.579 4.41 Power Required (𝑊) 42.6 44.5 4.39 CLPROP predicted very similar results to QPROP for the baseline propeller While the RPM prediction for the Prandtl is near identical from the two, CLPROP predicts that less torque and, therefore, less power is required to produce the desired thrust However, CLPROP uses a different model for predicting 𝐶𝐿 and 𝐶𝐷 so it is expected that differences can occur, especially when considering a propeller with a non-constant 𝐶𝐿 distribution such as the Prandtl Overall, the similarity of the CLPROP and QPROP results support the accuracy of CLPROP The results of the RPM sweeps for thrust and torque are shown in Figure Figure 7: a) RPM vs thrust; b) RPM vs torque The experimental tests resulted in consistent data with thrust and torque behaving as expected and increasing consistently as RPM increases The PLA and ABS blades displayed similar performance suggesting that the material was not a factor Since, both materials yielded similar results, either is acceptable for further testing Due to lower cost and ease of access to printers, PLA will be used for the propeller tests The experimental results for the PLA propellers are displayed and compared to the CLPROP propellers in Table Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education 10 Table 2: Comparison of CLPROP and experimental results at 0.5 𝑙𝑏𝑓 Baseline RPM Torque (𝑙𝑏𝑓 ∗ 𝑖𝑛) Power Required (𝑊) CLPROP 4560 0.752 40.6 Experimental 5257 0.833 51.8 % Diff 15.29 10.77 27.70 Prandtl CLPROP Experimental % Diff RPM 6493 6744 3.87 Torque (𝑙𝑏𝑓 ∗ 𝑖𝑛) 0.554 0.607 9.53 Power Required (𝑊) 42.6 48.5 13.76 The Prandtl experimental propeller produced less torque and required less power than the baseline design despite rotating at a higher RPM This indicates that the Prandtl design is producing less drag than the baseline design which will improve efficiency The performance of the Prandtl propeller displays the effect of unloading the blade tip and suggests that this should be the basis for future propeller designs CLPROP underpredicted the RPM for the baseline propeller However, it only slightly underpredicted the design point RPM for the Prandtl propeller This suggests that the modeling of induced losses near the tip of the blades is important and should be studied further The main difference between the baseline and the Prandtl designs is the unloading of the blade tip to reduce the tip losses The small diameter of these propellers likely makes them more sensitive to small perturbations in aerodynamic performance than larger propellers CLPROP also underpredicted the torque at the design point by around 10 percent for both designs This could also be due to aerodynamic properties as the Prandtl prediction is more accurate However, another potential factor is that CLPROP does not account for the effect of the weight of the blades If the chord length distribution of each blade is assumed to be proportional to the mass distribution, then the center of mass of the blade (𝑐𝑚 ) can be estimated This is utilized to predict the torque created by the mass of the blades as a function of 𝐵, 𝑐𝑚 , and the weight of each blade (𝑊) using Eq 12 𝑄𝑚𝑎𝑠𝑠 = 𝐵𝑐𝑚 𝑊 (12) For the PLA propellers, this results in a predicted increase in the CLPROP torque of 0.075 𝑙𝑏𝑓 ∗ 𝑖𝑛 This correction brings CLPROP’s performance more in line with the experimental results as shown in Table Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education 11 Table 3: Comparison of CLPROP corrected for mass and experimental results at 0.5 𝑙𝑏𝑓 Baseline RPM Corrected Torque (𝑙𝑏𝑓 ∗ 𝑖𝑛) Corrected Power (𝑊) CLPROP 4560 0.827 44.6 Experimental 5257 0.833 51.8 % Diff 15.29 0.72 16.12 Prandtl CLPROP Experimental % Diff RPM 6493 6744 3.87 Corrected Torque (𝑙𝑏𝑓 ∗ 𝑖𝑛) 0.629 0.607 -3.52 Corrected Power (𝑊) 48.4 48.5 0.21 The corrected torque predictions for CLPROP are similar to the experimental results validating the potential accuracy of the aerodynamic calculations of the spreadsheet The power required is a function of the rotational speed and the torque, so its accuracy is dependent on the accuracy of the predictions for RPM and torque Therefore, even after the correction of the torque, the power required is underpredicted due to the difference between the predicted RPM and the experimental RPM However, the power prediction for the Prandtl blade is nearly identical to the experimental value after it was corrected for weight This again suggests that the modeling of tip losses could be improved Conclusions In summary, CLPROP underestimated the experimental RPM, torque, and power required to generate the design thrust of 0.5 𝑙𝑏𝑓 Improvements in the modeling of the effects of tip losses and blade weight on propeller performance could improve the accuracy of CLPROP’s predictions However, despite underestimating propeller parameters, CLPROP performed very well compared to the industry-standard of QPROP, indicating the potential for CLPROP to be utilized in the field of propeller design The addition of propeller noise prediction capabilities to CLPROP is an avenue for improvement especially relating to the design of propellers for UASs Currently, the ability to gain insight into a propeller’s noise profile is very limited prior to actual testing so implementing noise prediction to CLPROP would improve the ability to determine how to optimize various aspects of the propeller design to reduce noise Future studies should also look at the usage of flow visualization techniques such as computational fluid dynamics (CFD) or smoke visualization to further characterize the effects of the tip losses on aerodynamic performance to obtain insight for improving CLPROP’s modeling accuracy References [1] L M Nicolai and G E Carichner, Fundamentals of Aircraft and Airship Design, vol Volume 1-Aircraft Design, vols Reston, Virginia: American Institute of Aeronautics and Astronautics, Inc., 2010 [2] J F Manwell, J G McGowan, and A L Rogers, Wind Energy Explained: Theory, Design and Application, 2nd Edition Chichester, U.K: Wiley, 2010 [3] J Roskam and C.-T E Lan, Airplane Aerodynamics and Performance Lawrence, Kansas: Design, Analysis and Research Corporation, 1997 Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education 12 [4] M Drela, “QPROP Formulation.” MIT Aero & Astro, Jun 2006, Accessed: Sep 14, 2020 [Online] Available: http://web.mit.edu/drela/Public/web/qprop/qprop_theory.pdf [5] R D Sanchez, “Aerodynamic and Aeroacoustic Design of Small Unmanned Aircraft System Propellers at Low Reynolds Numbers,” Master’s Thesis, Baylor University, Waco, Texas, 2020 [6] W R Liller III, “The Design of Small Propellers Operating at Low Reynolds Numbers and Associated Experimental Evaluation,” Master’s Thesis, Baylor University, Waco, Texas, 2015 BRETT D BENNETT Mr Bennett is a master’s candidate in the Department of Mechanical Engineering at Baylor University His research interests are in the fields of aerodynamics, thermofluids, design optimization, and programming with particular interest in the development of a program for the design and optimization of small propellers for UAS utilization He also currently serves as the student manager of Baylor University’s Additive Manufacturing Lab KENNETH W VAN TREUREN Dr Van Treuren is a Professor in the Department of Mechanical Engineering at Baylor University and serves as the Associate Dean in the School of Engineering and Computer Science He received his B S in Aeronautical Engineering from the USAF Academy in Colorado Springs, Colorado and his M S in Engineering from Princeton University in Princeton, New Jersey After serving as USAF pilot in KC-135 and KC-10 aircraft, he completed his DPhil in Engineering Sciences at the University of Oxford, United Kingdom and returned to the USAF Academy to teach heat transfer and propulsion systems At Baylor University, he teaches courses in laboratory techniques, fluid mechanics, energy systems, and propulsion systems, as well as freshman engineering Research interests include renewable energy to include small wind turbine aerodynamics, small propeller design, and experimental convective heat transfer as applied to HVAC and gas turbine systems Proceedings of the 2021 ASEE Gulf-Southwest Annual Conference Baylor University, Waco, TX Copyright © 2021, American Society for Engineering Education