AAE 4510-11 Experimental Projects

Modular UAV Engine Air Particle Separator

Final Report

D. Fritz 1, C. Lapworth 2, A. Sarachene 3, A. Stankovic 4, T. Wolford 5

The Ohio State University, Columbus, Ohio, 43210

The widespread usage of Unmanned Aerial Vehicles (UAVs) has taken traditional air breathing engines into flight environments that experience frequent dust storms and other hazardous weather conditions, resulting in Foreign Object Damage (FOD) to the engine and other critical internal components. FOD repair costs necessitate a more refined form of particulate separation for UAV air intake systems without drastically impeding the airflow into the engine. This research focuses on the implementation of a Vortex Tube Separator to accomplish separation. The scope of implementing a Vortex Tube Separator (VTS) involves the impact of helix pitch angle on particle separation efficiency, and the magnitude of the pressure gradient across the engine inlet. Through use of low order SolidWorks® flow simulation, high fidelity pressure models in ANSYS® Fluent®, and environmental engineering testing methodology MIL-STD-810G, this report details the relationship of helix pitch angle and pressure gradient in order to find an acceptable balance between the two. Through experimentation and multiple computational models, it was determined that the optimal pitch angle range that meet a  60% separation efficiency and a pressure drop of less than 25%, is within the 40°- 50°.

Nomenclature

Dh = Hydraulic Diameter (m)

𝜀 = Energy Dissipation Rate (m2/s3)

I = Turbulent Intensity

k = Turbulent Kinetic Energy (m2/s2)

madded = Mass Added (g)

mbag after = Mass of Bag After (g)

mbag before = Mass of Bag Before (g)

mseparated = Mass Separated (g)

η = Separation Efficiency

∇P = Pressure Gradient

Re = Reynolds number

u = velocity (m/s)

u’ = velocity perturbation (m/s)

1. Introduction

Nations around the world deploy Unmanned Aerial Vehicles (UAVs) in support of a wide variety of

mission objectives, whether they are humanitarian in disaster aid, or militaristic in support reconnaissance and similar missions. Widespread usage of UAVs inevitably causes exposure to various atmospheric conditions and environmental hazards. Over time these hazards can cause serious damage to static and rotating components as well as degradation to the internal engine. Projected costs indicate Foreign Object Damage (FOD) causes around $4 billion annually[1].  This necessitates a more refined form of particulate separation for UAV air intake systems without drastically impeding the airflow into the engine. The method that best allows this balance is a Vortex Tube Separator (VTS).

Several forms of Engine Air Particle Separators (EAPS) exist for the sole purpose of increasing particulate separation efficiency for an air breathing engine. The team will design and fabricate an EAPS utilizing centrifugal force to separate a substantial percentage of particulates from the airflow (80%) into an MQ-1 Predator Rotax 914F engine. Maximum efficiency will be determined via testing multiple configurations of helix angles ranging from 20° to 60° to achieve less than a 25% drop in static pressure. This particle separation efficiency and pressure drop values were determined through an initial investigation into VTS studies. These studies are evaluated and summarized in the literature review below.

1. Literature review

• Nicholas Bojdo and Antonio Filippone, Comparative Study of Helicopter Engine Particle Separators, May-June 2014

1. This article takes a comparative analytic approach to the 3 available methods of particle separators for a helicopter's engine intake; Vortex Tube Separators (VTS, which this project is focused on), Inlet Barrier Filters (IBF) , and Integrated Inlet Particle Separators. The writers’ study focused on the pros and cons of each device while also using computational methods and analytical theory to gather and generate performance data for each device. Their studies found that VTS achieved the lowest pressure drop of the devices tested, while still maintaining a reasonably high separation efficiency (where the IBF was found to be the most effective in this category).  For the group’s research purposes this article provides important theory-based information concerning particulate ingestion, pressure drop, and separation efficiency for the VTS.

• L. A. C. Klujszo et al, Design of Stationary Guide Vane Swirl Air Conditioner, April-June 1999]

-The purpose behind this document is to determine an efficient design of a stationary vortex vane that will act as a particle separator. This document is likely going to be the centerpiece of our focus due to its in-depth analysis of how pitch angles affect both the separation of particles from the flow as well as the pressure drop at the engine intake. Due to how similar the paper’s focus is to the project, the previous team used this document as a base to design their own testing apparatus.

-   Chengming Song, Binbin Pei, Mengting Jiang, Bo Wang, Delong Xu, Yanxin Chen, Numerical analysis of forces exerted on particles in cyclone separators, Powder Technology, Vol 294, June 2016, Pages 437-448, ISSN 0032-5910,

• Computational Fluid Dynamics (CFD) and experimental results were utilized to track motion of different sized particles in an airflow subjected to a centrifugal force field (Note: results found for cyclone separator, not vortex tube separator). Particle motion is largely governed by the drag force (pushing inwards) and the centrifugal force (pushing outwards) of the airflow. Larger particles are pushed to the outside wall, smaller particles aren’t pushed as hard and stay close to the center. Very fine particle motion is determined entirely by the flow of the gas.

II. Experiment Description

1. Overview

This project focused on creating multiple designs of an EAPS with varying helix pitch angle for each design. Each design was created and initially subjected to low order testing through the use of SolidWorks® flow simulation. Higher order testing was then conducted in ANSYS® Fluent® to confirm that the pressure drop for each model is within the acceptable range below 25%. Maintaining a reasonable balance between efficiency and pressure drop was the main focus throughout computational analysis. It is important to emphasize that helix pitch angle is directly proportional to both efficiency and pressure drop; an increase in helix pitch angle means an increase in separator efficiency, but also an increase in pressure drop across the separator.  Once computational modeling was completed, the most effective designs meeting the project requirements were fabricated utilizing the Ohio State Rapid Prototyping lab and used in experimentation to validate the computational models.

 These models were utilized throughout both facets of experimental testing, pressure gradient testing, and particle separation efficiency testing, with the acrylic box test apparatus . In order to remove the risk of damaging sensitive instrumentation at the Aerospace Research Center (ARC), these tests were conducted in two testing facilities. Pressure gradient testing was conducted in the ARC, whereas particle separation testing took place in Bolz Hall laboratory room 103. Once the necessary data was logged from testing, the necessary data consolidation had taken place and the results of the experimentation had been summarized within this report.

2. Computational Modeling and Operating Conditions

Low order models were initially created in SolidWorks® flow simulation to achieve a fundamental understanding of how the particles would behave in the experimental setup. A simple tube was modelled with a separator placed at the midsection and a volumetric flow was introduced into the system through a pressure opening. This model showed the onset of rotational flow about the separator, however the pressure would equalize after a certain amount of time had passed. This information led to the experimental design having the trailing edge of the separator placed at a fixed location closer to the engine inlet in order to maintain pressure gradient. While the SolidWorks® model showed a reasonable approximation of particulate behavior in the flow it did not have the fidelity required for proper pressure gradients or separation due to centripetal forces. This led to the need for more complex models.

Higher order analysis was conducted utilizing the ANSYS® Fluent® Fluid Flow. This CFD package utilizes Reynolds averaged Navier-Stokes equations (RANS) for mass, momentum, and energy conservation in order to account for turbulence within each model. Each model consisted of fluid workpiece dimensionalization about the SolidWorks® rendered models of each respective separator. Each separator was positioned 0.1143 meters from the defined outlet of the fluid workpiece to remain consistent with separator positioning within the testing chamber during experimentation. The radial cushion of each separator was set at 0.00254 meters in order to mitigate the possibility of reversed flow near the wall of the fluid. Fluent® automesh meshing method, accounting for proximity and curvature, produced a cell count on the order of e+06 for each separator and was deemed sufficient given the cylindrical geometry of the fluid region.

 Pressure-based RANS were utilized within the computational model given incompressible flow within the flight regime of heavy particulate ingestion for experimentation and Rotax engine operation. A realizable k-𝜀 turbulence model was selected for the transport equations of turbulent kinetic energy and dissipation rate. Enhanced wall treatment conditions were employed accounting for pressure gradient effects in order to realistically implement turbulent kinetic energy readings at the wall. The boundary conditions of the fluid regions were defined via the establishment of a pressure outlet and velocity inlet respectively. Pressure outlet values were determined directly from readings obtained during experimental pressure testing.  Velocity readings at the inlet were computed from pressure readings at the first two ports detailed in the experimental testing apparatus before the onset of turbulent flow. Values of k and 𝜀 were then computed via equations 1-3 shown below.

        (1)

                (2)

                (3)

3. Experimental Test Apparatus

Experimental testing was split into 2 distinct test methodologies. These methods are the pressure distribution testing, and the particle separation testing. Each test used the same acrylic box apparatus with the exception of modified inlet tubes; The designated particle inlet tube includes an opening on the top that allows  particulates of the required size to be added into the flow in a uniform fashion, while the other inlet has holes carefully drilled along the span of the pipe to allow for pressure rakes to read the airflow inside the pipe. The experimental apparatus has been created by the previous project team for use in testing multiple separator designs. Schematics and photos of the apparatus are found in Figures 1 and 2 below.

Figure 1: Test Apparatus Schematic

Figure 2: Test Apparatus

  The test station consists of an acrylic tube fitted to a 12 horsepower shop vacuum, simulating a scaled down version of the MQ-1 Predator Rotax 914F engine, while also generating airflow through the apparatus. The separator filtered the flow while the separated particles were transferred into an acrylic box where the tube is mounted. The team measured the efficiency of the EAPS through use of modified particle mass flow equations, and multiple test runs to establish a good basis for particle separation efficiency. Pressure drop was assessed via static pressure sensors attached along the length of the test section, where multiple rakes were positioned before and after the separator to assess overall pressure loss.

The experimental testing consisted of fitting the particle separator in the model inlet and activating the vacuum to simulate the Predator engine airflow. Measurements of particle separation and pressure probe readings of flow were taken both before and after passage through the EAPS to test design efficiency. Testing methodology is described in fuller detail within the Test Matrices section of this report.

4. Fabrication

Continuing on from AAE 4510, 5 separators were modeled and fabricated using Ohio State’s Rapid Prototyping Laboratory in the East wing of Scott Lab by the first 2 weeks of December of 2016. While the fabrication of the separators occurred on schedule, there was a noticeable flaw within the details of the modeling. The separators were modeled to the inlet pipe’s outer diameter rather than the inner diameter. Causing the separators to be 0.635cm  larger than required. Following this setback, the flawed separators had been subjected to be hand sanded in attempt to fit within the inlet orifice. This attempt was unsuccessful.

Due to the nature of the separators’ 3-D printed material base, the edges of the models would fringe and become unruly on the outer edges. The frayed edges would accumulate additional particulates if subjected to testing, as well trip the airflow into turbulence across the separator itself. This attempt to salvage the now defunct separators called for new model design. Design of the new models focused around fabrication to the correct sizing of the test apparatus inlet, while also slightly modifying the original design with a filleted leading edge to ensure smoother flow transitioning across the separator. The 60° filleted separator model shown in figure 3 below.

Figure 3: 60° Filleted Separator Model

5. Scope of tests

        The primary focus of this report is to investigate the impact of helix pitch angle on particle separation efficiency and the magnitude of the pressure gradient across the engine inlet. Once applied to the correct Reynolds number of the full scale scenario, the data can then be applied for use at the AFRL with the MQ-9 Reaper drones.

 While experimental testing was still in its planning phase, the computational focus of this research sought to simulate particle and separation testing for each separator design. Once the initial models were created and run through Solidworks® flow simulator, a baseline data set was established as an initial benchmark. This data set provided enough data to begin more complex simulations within ANSYS® Fluent®.

1. Extent of Computational Approach

The computational scope was primarily focused on simulating real world scenario of a VTS on the Rotax 914F engine. Reliant on boundary conditions provided by the experimental testing, pressure gradient testing was conducted to verify the experimental setup. The span of this study focuses only on verifying the relationship between helix pitch angle and pressure loss. Alternative computational research for implementing the EAPS is discussed further in the future work section of this report.

2. Extent of Experimental Approach 

The experimental scope had been primarily focused on capturing the real world scenario of a VTS placed within the Rotax 914F engine. Simulating the engine intake with the 12 gallon shop vac gave results based on this smaller scale test, that had been conducted with the purpose of validating the computational modeling. The span of this study focuses only on validating the relationship between helix pitch angle and separation efficiency. Alternative experimental research foci for implementing this EAPS is discussed further in the future work section of this report.

6. Testing Method (Calibration and Error Mitigation)

1. Pressure Loss 

Once the testing apparatus had been properly assembled, the separator to be tested was inserted into the designated pressure inlet tube. Each separator was placed 1.125 diameters (11.43 centimeters) away from the outlet to ensure the pressure rake reading location immediately after each separator would be the same for all separators. 8 pressure rakes in total were secured along the inlet pipe: 1 at the inlet of the pipe, 3 across the separators, 3 immediately after the separator, and finally one positioned before the vacuum inlet. Once the separator was positioned properly, the vacuum as run and the pressure transducer recorded data. Each separator collected 2 runs of around 1 minute per run.

2. Particle Separation 

A designated particle injection inlet tube was then used for the separators. Again, each separator was positioned at the same trailing edge location of 1.125 diameters and the test apparatus was secured and sealed with additional aluminum tape and duct tape. The testing method for the particle testing is as follows:  3 runs minimum were completed for each of the 5 separator models. Roughly 50 grams of Red China clay, as per military standards[2], were measured out into individual distribution containers, and the shop vac bag was weighed prior to the test.

A large-mouthed funnel was placed into the tube drilled orifice and the vacuum was run. Once flow had been generated inside the simulated engine inlet, the particulates were introduced through a flour sifter positioned above the large mouthed funnel in order to produce a uniform particulate flow rate. The run would finish when all 50 grams of particulate had been introduced into the flow. Particle separation was then measured from the weight of the bag after the run had been completed. Computing the difference between the bag’s final and initial weights along with the amount of particulates added gives the amount of particulates separated during that run. The separation efficiency equation is shown below.

        (4)

3. Error Mitigation Experimentation 

After 3 consecutive runs were completed, the inlet tube was removed for cleaning to ensure consistent testing data across all separators. From the relatively low power of the shop vac, buildup of the Red China clay had accumulated immediately after the separator. This averaged out to ~2.9 grams of particulates per 3 runs. This average buildup leads to error mitigation within both test methodologies. In such a small scale experimental setup, the 12 gallon shop vacuum would suffice to simulate the Rotax 914F engine. The overall power of the vacuum was underestimated, however, when the additional particulate buildup was discovered by the team. In order to minimize this build up, a can of compressed air was utilized to encourage the particulates into the flow at the tube inlet.

        Regarding other areas of error mitigation within testing, introducing a refined particle induction method to ensure a smooth and uniform rate of particle flow was the main concern with the experimental team lead. The use of the flour sifter method provided a more manageable mass flow rate for particulates entering the flow, as well as the purchase of a higher accuracy mass scale for bag measurements. Compared to the scale that was initially used, the new compact digital bench top scale measured weight within 0.01 grams of accuracy, ensuring a high fidelity result set compared to the original 0.1 gram accuracy with the scale prior. Each method described was in attempt to keep experimental error to a minimum while also producing higher precision results.

4. Error Mitigation Computational

 Error mitigation within this computational model consisted of model refinement techniques and solution parameter calibration. A realizable k-𝜀 turbulence model employs constraints on Reynolds stresses that sufficiently mirror physical restrictions on turbulent rotating flows. Solution initialization methods eschewed hybrid initialization in favor of standard initialization utilizing pressure and velocity conditions obtained experimentally at the inlet and outlet. Residual error tolerance was set to the order of e-03 with convergence primarily governed by the continuity equation of error. An acceptable mass flow rate flux of an order of e-07 between the inlet and outlet was also employed to ensure error mitigation.

7. Test Matrices

Pressure gradient testing relies on the data gathered by the pressure taps located along the wall of the inlet tube. Correlating this raw data into useful results by means of a test matrix is shown in table 1 below.

Table 1: Pressure Gradient Test Matrix

        This test matrix addresses the measured pressure and pressure differences from ambient in units of inches of H2O and pounds per square inch (psia). The main focus of this test matrix is the last 2 columns to the far right--the pressure drops from each test run from ambient temperature. Particle separation testing had focused on the weights of the particulates added into airflow, and the shop vac bag weight before and after each test. Determining the difference in these weight resulted in amount of particles separated in each run of the selected separator. Table 2 below shows the text matrix used for all particulate testing.

 Table 2: Particle Separation Test Matrix

III. Results and Discussion

1. Description of Results

        Following completion of computational models and experimental testing, the results of both facets of research were compiled and consolidated. Regarding pressure gradient testing, both computational and experimental testing show similar results for the pressure drop across each separator. Tables 3 and 4 below showcase each set of results.

Table 3: Pressure Gradient Computational Results

Table 4: Pressure Gradient Experimental Results

        Initially the experimental results revealed a larger pressure drop between the pressure rakes. This was due to pressure loss from the skin friction of the inlet pipe. The distance between the first two pressure rakes was large enough to show a decrease of nearly 15 pascals. Correction factors for this pressure drop are discussed in the analysis section of this report. Particle separation testing results are shown in Table 5 below.

Table 5: Particle Separation Experimental Results

        Compiling the above data shows a jump in both particle separation and pressure drop for the 20° separator. This jump is analyzed in the result analysis section of this report. The compiled graph is shown in figure 4, as well as the average pressure drop vs. separator distance in figure 5 below.

Figure 4: Separation Efficiency and Pressure drop vs. Helix Pitch Angle

Figure 5: Pressure Drop vs. Distance from Separator (Average of 2 Runs)

Figures 5-9 show the wall static pressure readings for each separator completed in ANSYS® Fluent®.

Figure 6: Wall Static Pressure Readings (20° Separator)

Figure 7: Wall Static Pressure Readings (30° Separator)

Figure 8: Wall Static Pressure Readings (40° Separator)

Figure 9: Wall Static Pressure Readings (50° Separator)

Figure 10: Wall Static Pressure Readings (60° Separator)

2. Result Analysis

        As illustrated by Figure 4, computational and experimental trends of fluctuations in pressure drop are consistent across the testing range of separator helix angles. Initial assumptions of a  linear coupling decay between separation efficiency and pressure drop held for the range of 30°- 60° however, as illustrated in Figure 4, an increase in pressure drop and separation efficiency for the 20° separator indicates nonlinearity across all range of separator helix pitch angles. While this general trend holds for separator ranges of 30° - 60° , additional experimentation is required to determine the nature of nonlinear effects below the determined 30° threshold.

        Static pressure reading trends at the wall of each respective separator, as determined via experimentation in Figure 5, adhere to the same static pressure trends of the wall obtained via the computational model in Figures 6-10. Due to the limitations of finite discrete locations of wall pressure reading in experimentations, noticeable pressure drop and pressure gradient effects at the location of the separator within the tube are not directly illustrated within the experimental data. Static pressure drops of greater severity exist within computational models due to the ability of the CFD package to export all static pressure readings along the curvature of each separator blade. Thus accounting for realistic pressure drops inherent to swirling turbulence in the flow.

3. Application

        Upon verifying the computational model employed across each of the separators, the team utilized the established Fluent® fluid flow model to determine static pressure drops across a newly fabricated separator of a helix angle within the range of optimal separation efficiency and pressure drop as determined via the results of Figure 4. Additional simulations were run at flight conditions of takeoff and landing for the  Rotax 914F engine indicated in Table 6  below with new values of k and 𝜀 determined via equations 1-3 above. Figures 10-11 below show the wall static pressure readings for a 45° separator computed in ANSYS® Fluent®.

Table 6: Rotax 914F Performance Data

Figure 11: Wall Static Pressure Readings of a 45° Separator for Rotax Landing Conditions

Figure 12: Wall Static Pressure Readings of a 45° Separator for Rotax Takeoff Conditions

4. Comparison to Theory

        The theory proposed by the group can be supported by both experimental and computational results for separator range from 30º to 60º in helix pitch. This initial assumption of a linear relationship between the pitch angle and separation efficiency as well as pitch angle and pressure drop holds as an approximation for this range. Reviewing the total data set of results, there is an overall indication that the pressure loss and separation efficiency relate to the variation in helix angle nonlinearly.

        For 80% separation, this was unattainable both computationally and experimentally, given the helix pitch angle domain. This efficiency was too high to accomplish in conjunction with a 25% pressure drop. For the 25% pressure drop, this was both computationally and experimentally attainable for helix angles approximately less than 50º.

5. Error Discussion

        The issues with error in the pressure testing measurements were fairly simple to determine. Two tests were performed on five of the six separators to get a bearing on the error in experiment rather than analyzing the device. The average error deviation using this data was within the resolution of the instrument (±0.0005 psi) with a maximum error of 0.002 psi occurring on the 50°  separator. Error between the two 50° separator runs are significantly higher than for the 20°- 40° , and this is likely due to a sealant that was used to stick the pitot tubes to the inlet. This sealant was not used in later tests.

        Error in the experimental particle testing is more difficult to substantiate. Multiple tests were performed because the random error was anticipated to be high. The mean absolute deviation from the five separators was highest with the 20° seperator (7.18%) and lowest with the 30° seperator (1.31%). The uncertainty was high in the 20° and 60° separators, therefore the trends at the two ends of our plot on Figure 4 may not hold up to scrutiny.

        Implementation of a granular phase into the Fluent® fluid flow multiphase model consisted of establishing 2 distinct Eulerian phases with slip velocity mixture. Properties of the kaolinite dust were imported into material properties for the fluid type of the second mixture phase [2]. The multiphase volume fraction parameter was established given the ingestion port and tube inlet geometry in accordance with particle density specifications. Calculations implementing the multiphase model resulted in divergence in the volume fraction residual equation for the second mixture phase. Due to constraints in the project timeline, implementation of a functioning multiphase model was not achievable for the purposes of obtaining separation efficiency data from the computational model.

6. Relation to hypothesis

The main objective of this research involves assessment of the theory that centrifugal forces in a stationary VTS yields measurable separation of particulates from airflows, with minimal static pressure drop. Data obtained from computational analysis undertaken in SolidWorks® with created CAD models will yield direct static pressure gradients across each model of varying helix angles. This data, obtained via low order approximations of turbulent flows, will provide insight regarding the capabilities of a VTS and allow flow streamline visualization to indicate relative particulate separation present for each pitch angle variation.

Data obtained from computational analysis undertaken in Fluent® will directly allow for measurements of particulate separation efficiency with the definition of a granular phase in the multiphase flow. Additionally, higher order turbulence models utilized in this computational analysis will allow for static pressure gradient data acquisition for the pressure drop. The project goal of determining the feasibility of VTS will be met via experimental testing of each fabricated model for the data acquisition of static pressure readings and measurements of degrees of particulate separation present. This data can then be utilized to refine model characteristics of the high order computational analysis such that newly obtained numerical data mirror the physical parameters of particulate separation in a laboratory setting. This analysis can then be utilized in accordance with an established testing matrix indicating real-world parameters of flight conditions to determine whether or not sufficient particulate separation can be achieved while minimizing static pressure drop on the basis of VTS for engine intakes of unmanned aerial vehicles.

IV. Summary and Conclusion

1. Summary of findings

        When comparing experimental and computational results, the pressure gradient analysis trends are accurate to within approximately 15 Pa. Experimental data collected is indicative of a coupling relationship between pressure loss and separation efficiency. The variation in helix pitch angle across the domain of study (20º- 60º) demonstrates efficiencies spanning from approximately 48% to 65% with pressure drop percentages spanning from approximately 16% to 32%.

2. Assessment concerning hypothesis

The original hypothesis that was presented as the research went forward was that 80% of the particulates would be separated with at most a 25% static pressure loss. From both experimental and computational findings, this cannot be achieved. By approaching the 80% separation that was initially hypothesised the static pressure loss would approach values of 50% or higher, drastically reducing the power the engine would produce. The pressure gradient from a model that achieves this 80% separation efficiency would be too great to justify any practical use in the professional field.  From this information, the initial hypothesis for separation efficiency must be reduced such that experimental and computational findings will fit the success criteria. Aside from the overshoot in separation efficiency, the initial hypothesis of a pressure drop below 25% across the separator is feasible within the 20°- 60° helix pitch range. Both computational and experimental results confirm the hypothesis with a maximum pressure drop of 32.62% for the 60° model, the lowest readings of 13.89% drop for the 30° model.

The team’s initial hypothesis was based on initial investigations into VTS studies evaluated and summarized in the literature review. Information gathered from these studies provided a baseline goal for the team as they sought to replicate the hypothesis established. From the results of both experimental and computational endeavors the original statement was revised to ensure a  substantial percentage of 55% of the particulates from the airflow into an MQ-1 Predator Rotax engine while ensuring less than a 25% drop in static pressure.

3. Suggestions for Future work

The primary focus of this research was to study the linear relationship between particle separation efficiency and pressure drop across the separator by adjusting the helix pitch angle of each separator model. While the results of this research provided useful information relating toward the subject, this field of study is not a single problem, single solution situation. It is a multi faceted problem with multiple areas of research that could be examined in the future. During particulate testing the 20° separator showed results that did not follow the theory of  the separator physics, resulting in the 3 additional runs for this model to verify if any errors in testing or recording were present. Particle separation efficiency had a higher value for the 20° separator than or the 30° model.

After the 30° model each other separator follows the expected trend in both pressure drop and separation efficiency. It was determined that there exists a threshold pitch angle at which the separator is unable to accelerate the particulates enough to separate out into the collection box, leading to non uniform separation. Future research could go into examining this non-uniform separation threshold and at which pitch angle is the cutoff for said threshold.

Other research could delve into improving the experimental setup. The particle injection method used throughout experimental testing was not without flaw. Multiple funnels were used in order to avoid any dust clumping up in the inlet, as well as using a flour sifter to attempt to introduce the flow in a steady rate.The particulates were funneled through the sifter that was agitated to emulate a constant 104 beat per minute rate for particulate introduction. One facet of research could be to develop a particulate injection method that can reduce or even eliminate clumping of the dust, as well as investigate other methods of introducing the particles in a uniform flow.

Adjusting the geometry and position of the separator within the engine inlet could also be another facet of research that could be looked into. Initially the team had conceived of separator models with tapered pitch angles; The separators tapered by 5° pitch angles at ⅓ increments of the separator length such as 20°- 25°- 30° taper. Position of the separator along the inlet could also have a factor on engine performance. Whether or not the particulates could be separated properly if they encountered the separator earlier or later in the flow, or if the pressure drop would reduce from the fact that a separator placed closer to the inlet would give the flow time to reset as it enters the engine inlet. Any of these ideas could provide valuable information to this multi faceted research.  

V. Appendices

1. Final Project Budget

        In accordance with the Air Force Research Laboratory (AFRL) , the research proposal was approved by the project sponsor Dr. James Gilland, as the team received over $6750.00 for the entire project budget.  Table 6 below shows the final project budget over the 8 month research progress. Over $600 of the budget was allocated towards the 3D printing of the separator models. After the flaw was realized in the initial model set, the 2nd filleted model set was sent out for production immediately after design. Purchase Request 1 and 2 consisted of  items necessary for primarily the particle separation portion of experimental testing. Objects consisted of supplementary shop vac filter bags, respirators, gloves smocks, and aluminum tape for any leaks found in the test apparatus. The key items in these purchase request consisted of the compact digital bench top scale, an additional shop vac hose, a flour sifter, and a new wide mouthed funnel. The final two  45° models were sent out to print post report, and will be manufactured for to the Ohio State Senior Engineering Capstone Showcase on 4/25/2017 where the team will present their findings to their peers. 

Table 7: Final Budget

2. Final Project Schedule

1. Milestones + Gantt Chart

Table 9. Gantt Chart

Major research milestones for this project include the first batch of separators that were manufactured on 20 January 2017, completion of pressure gradient testing on 25 January 2017, completion of particulate testing on 29 March 2017, and final report consolidation on 17 April 2017.

3. Acknowledgements

The team would like to give special thanks to Professor Clifford Whitfield for giving the research excellent direction and motivation. Shop supervisor  Kevin Wolf  at Scott Lab’s Rapid Prototyping Laboratory was quick, professional, and very helpful when consulting the fabrication of the separators themselves, special gratitude goes to his involvement in this research. Lastly, the staff at the ARC, Jake Allenstein and Robin Prenter were of great help when it came to pressure gradient testing as well as conceptual ideas for CFD models. This research would not have proven as presented without the help of these four individuals.

4. References

[1] B. Bachtel Foreign Object Damage Analysis [website] Available:

http://www.boeing.com/commercial/aeromagazine/aero_01/textonly/s01txt.html 

[2] U.S. Department of Defense, “MIL-STD-810G,” 2008.