Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
A High Performance, Continuously Variable Engine Intake Manifold
2011-01-0420 Published 04/12/2011
Adam Vaughan and George J. Delagrammatikas Cooper Union Copyright © 2011 SAE International doi:10.4271/2011-01-0420
ABSTRACT Manifold tuning has long been considered a critical facet of engine design and performance optimization. This paper details the design, analysis and preliminary testing of a continuously variable, carbon fiber intake manifold for a restricted 2003 Suzuki GSXR-600® engine. The device achieves a large dynamic runner length range of 216-325 mm through the use of a half-tube, sliding shell design that differs substantially from traditional variable intake approaches. A combination of Ricardo WAVE® and 2D/3D Ansys Fluent® simulations were used to aid in the design of the intake along with a custom software routine to optimize restrictor geometry through fully automated CFD simulations. The sliding mechanism was actuated via a cable linkage system and powered by a small servo motor. This motor was controlled by a Microchip dsPIC® microcontroller that was embedded in a custom power distribution PCB for the 2009 Cooper Union Formula SAE® entry. The controller communicates with the engine's MicroSquirt® ECU over CAN to read instantaneous engine speed and commands the servo based on an empirically tuned look-up table. Initial testing of the intake showed the expected torque and power variation, maintaining maintaining over 95% of the peak engine torque for an additional 60% of the usable engine speed range in addition to a peak power improvement of 5% relative to a baseline static intake configuration. A peak power improvement of over 22% was also achieved relative to the 2008 FSAE® intake. The variable intake system adds less than 1% to overall powertrain weight and is able to actuate the full dynamic range in less than 1.0 s. Additional gains are expected through optimized cam timing coupled with refinements to the initial engine calibration.
INTRODUCTION As is characteristic in any rotary power-plant (electric, internal combustion, etc.), there exist various operating points where torque output reaches a local or global maximum [1]. Because torque affects the acceleratory performance of a vehicle, optimizing it is a critical factor for any vehicle designer. For engines, torque is strongly affected by Volumetric Efficiency Efficiency (VE), which is a measure of the actual air inducted versus the swept volume of the piston [2]. VE, in turn, is closely coupled to the resonance conditions that develop in the engine's manifolds, the timing of the induction and exhaust processes, fluid flow losses and the average speed of fluid flow [1]. Many empirical and computational studies have shown a clear relationship between the geometry of manifold ducts and volumes to the resonance peaks favorable to high VE performance (see Figure 1) [1]. As a result, an engine manifold designer is able to “tune” an engine to achieve higher VE at certain engine speeds by carefully selecting appropriate geometry. However, drivability also remains a concern, and because highly “tuned” engines typically exhibit narrow operating bands of peak performance, it is the goal of this paper to explore dynamically varying manifold geometry in an effort to produce an engine that is “tuned” over a greater range of engine speeds. By taking this approach, a performance engine can be made more drivable (through a more constant, or “flatter” torque curve) without a corresponding increase in engine displacement to compensate for low torque regimes.
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
• Engine displacement must not exceed 610 cc. • Any component capable of throttling the engine must be upstream of the flow restriction. • All intake components must be contained within an envelope defined by the main roll hoop and rear tires.
Figur e 1. Generalized intake geometry for a restricted FSAE® engine
A “flatter” torque curve is especially important in the displacement limited Formula SAE® competition, which is the target application for the design to be prese nted. It is worth mentioning that modern Variable Valve Timing (VVT) systems can also be utilized to produce a more drivable vehicle, with greater torque gains as well as benefits in reduced exhaust emissions and improved fuel economy [3] [4]. However, a variable intake is more easily adapted to the small, fixed cam timing motorcycle engines commonly used in FSAE®. It is also simpler i n overall construction. Numerous models have been developed to help predict VE curves as a function of runner length and other manifold geometry [2]. These range from a basic mass-s pring model (Helmholtz) [2], to 1D pressure wave simulations (such as Ricardo WAVE®) [5], to even transient 3D CFD codes that strive to model the complete fluid state (such as Ansys Fluent® and Ricardo VECTIS®) [6] [7] [8]. Each scheme has its own advantages and disadvantages. The Helmholtz model, for instance, is a simple equation that roughly predicts the VE peak but does not yield information about other operating points [2]. While the Helmholtz model is convenient for calculations, it can show a sizable error at high speed engines [9]. Ricardo WAVE® discretizes duct lengths and plenums as 1D flow elements to in order to rapidly compute a fairly accurate VE curve, but it struggles to capture effects produced by turbulence (such as from the restrictor [5] [7] [8]). Ansys Fluent® and Ricardo VECTIS® are each capable of providing a more complete model compared to Ricardo WAVE® (as they include empirically verified turbulence models), but they do so at great computational expense [7] [8]. As mentioned earlier, this paper will only focus on the FSAE® class of vehicles and will, therefore, be compliant with the rules that govern the 2010 competition. The key requirements for the competition are: • There must be a 20 mm diameter flow restriction on all air entering the engine.
It is worth briefly commenting on the difficulties that arise from the FSAE® constraints before delving into greater detail in the subsequent sections. The 20 mm flow restriction presents what is probably the greatest modeling challenge. Not only does it cause the engine to always operate at part load conditions, but it also necessitates a modeling approach that can handle compressible, transient flow as well as turbulence. Thus, it is important to incorporate elements of 3D simulations to effectively model this element. The FSAE® displacement limitation constrains “off the shelf ” engine choices to those of the small engine motorcycle/ATV market. Finally, the envelope limitation presents a handful of system packaging issues, particularly with the long restrictor.
DESIGN The general design methodology taken for this study begins first with restrictor optimization, followed by a discussion of plenum sizing and finally the selection of variable runner length geometry through 1D and 3D CFD codes and packaging limitations. Although not presented, an embedded intake servo control system was developed for this project using a Microchip® dsPIC30F4011/30I microcontroller. This system utilized a Controller Area Network (CAN) interface to query the MicroSquirt® Engine Control Unit (ECU) for instantaneous engine speed and to subsequently command an optimal runner length based off an empirically determined look-up table.
ENGINE SELECTION Previous Cooper Union FSAE® entries have used the 2003 Suzuki GSXR-600® (specifications given in Table 1) engine primarily because its stock configuration provides more lowend torque compared to other engines within its displacement class. Of course, adding a restrictor and packaging an alternate set of manifolds will alter this characteristic [1]. The engine has proven itself a reliable workhorse, and given the team's prior experience with this engine as well as the availability of spare parts within the lab, it was decided to continue using this power plant. The engine is in stock configuration with the exception of the ECU, intake manifold and an increased compression ratio provided by a Wiesco® piston upgrade.
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
Table 1. A list of general specifications for the 2003 Suzuki GSXR-600® engine
shift the VE resonance peak lower. Directly conflicting these results, Hamilton et al. empirically found that a large 6 L plenum increased the peak torque output by 17% [12]. However, it is worth noting that the runner length used for their study was 150 mm, which is significantly shorter than the runner lengths detailed in this and other studies. Because the ultimate goal of this project was to study the effects of a variable runner length intake and given the highly non-linear results found by previous studies with large plenums, the final plenum volume was sized primarily off the space required for runner actuation (3.1 L). It is left to future computational and empirical studies to further optimize this variable.
RESTRICTOR
1D CFD OPTIMIZATION
With one of the key handicaps of FSAE® requirements being the choked flow induced by the 20 mm diameter restriction, extensive studies were performed to optimize the restrictor element for maximum mass flow. A custom software routine was developed in C# to perform an exhaustive Design of Experiments (DoE) search quantifying the relative contributions of various geometric variables. The details of this study are described in [10], however the final selected geometry is presented in Table 2.
As runner length is a variable that can produce sizable VE effects [1], a method of varying length was incorporated into the intake design. To initially explore the feasibility of such a setup, preliminary calculations were first performed using a simple Helmholtz model [2]. The model results showed that in order to cover a substantial range of VE peaks (8000-12000 rpm), the runner length must more than double.
Table 2. Ranges of variables used for DoE restrictor study, with final selected values
PLENUM SIZING Plenum volume is known to affect engine VE [1]. In unrestricted engine simulations, Winterbone and Pearson found that small volumes were favorable at low engine speeds and that plenum effects were marginal at high speeds. They went on to state that plenum sizing is mostly a concern of idle speed control. For restricted engines, Blair has recommended that plenum volume be as large as possible [11], and up to 6 L has been successfully used by other FSAE® teams without throttle response issues [12]. McKee et al. stated that 3.5 L was the optimal volume for a 600 cc engine from their simulations [11]. Through simulations of conical spline FSAE® intakes, Claywell et al. found that large plenum volumes (>3.5 L) beneficially affected VE, but the results were highly dependent on the way plenum geometry was adjusted to vary the volume [7] [8]. In general, they found that the gains from increased plenum volume showed diminishing returns (<1% of VE) after increasing beyond 5 L. Additionally, they were able to show that adjusting the shape of a large plenum volume could
In order to further quantify the expected gains of varying various runner lengths, a Ricardo WAVE® model was developed for the 2003 Suzuki GSXR-600® engine. A rubber mold of the engine's intake port was made, digitized and incorporated into the engine model to improve its accuracy. The rubber mold geometry was also used to aid in injector placement later on in the design. A contour plot of the resulting torque at Wide Open Throttle (WOT) for various runner lengths across different engine speeds is presented in Figure 2. This chart clearly shows how varying runner length can shift the engine's torque peak, allowing the engine to produce more torque over a greater range of engine speeds. To more clearly show the simulated gains, a plot of the percent difference relative to the torque output at a runner length of 320 mm is shown in Figure 3. While the Ricardo WAVE® model requires further calibration using dynamometer results (data which prior to this project were unavailable for the restricted Suzuki GSXR-600®), its general trends were found to agree with those of the Helmholtz model. Note that the simulated torque gains of 5% shown here were significantly less than the 10-15% found in the McKee et al. empirical study with similar runner length variation but different geometry than used here [11]. From the results of this Ricardo WAVE® study and those of [7], [8] and [11], a desired length variation of 205-330 mm was selected. This maintained high torque output for a substantial portion of the torque band (7500-10,000 rpm) and did not unnecessarily increase runner length to be optimal for engine speeds that did not produce adequate power. The difficulties of packaging this length
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
variation will be discussed later, but the shortest packageable runner (without modifying the engine head) is roughly indicated in Figure 2.
• Dynamic range is further reduced because the housing tube must devote space for a fuel injector, a bend for restrictor packaging, and a mounting flange. • While not simulated, the relatively harsh bend (necessary to increase dynamic range) would most likely induce recirculation within the head port.
Figure 2. Ricardo WAVE® simulated contours of torque (N·m) for various runner lengths at WOT Figure 4. Section view of telescoping runner design
After iteratively optimizing the design around these considerations, the final geometry could vary (at most) from 240-300mm, which significantly fell short of the desired range of 205-330 mm. Nevertheless, this geometry was rushed to fabrication in order to have a design ready for the 2009 FSAE® competition.
Figure 3. Ricardo WAVE® simulated % difference in torque at WOT from 320mm runner length baseline
METHODS OF RUNNER VARIATION In order to package the desired runner length variation of 205-320 mm, two approaches were pursued. The first was a telescoping design (Figure 4), wherein a movable tube (red) would slide into and out of larger fixed tube (light grey) much like a trombone (see [10]). While relatively simple, this design proved less than ideal because of the following considerations: • The dynamic range is geometrically limited, as the sliding tube cannot actuate to a length longer than the tube within which it is housed.
For the 2010 FSAE® season, the partially complete telescoping design was abandoned to pursue a new design that showed the promise of remedying the dynamic range issues while building on the fabrication experience developed during the 2009 FSAE® season. This new design (Figure 5) operated under the principle that only the large area change at the end of the tube into the plenum served as a wave reflecting resonance device, and that the geometry after the end of the tube did not contribute appreciably to the resonance characteristics. To accomplish length variation, only half the tube (red) was actuated to increase or decrease the runner's centerline length, while the other half of the tube remained stationary and built into the plenum wall. By taking this approach: • The full dynamic range could be realized. • Harsh bends within the intake could be avoided. • The linkage system could provided by a simple cable mechanism.
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
STEADY-STATE 3D CFD OPTIMIZATION
Figure 5. Overview of half-tube concept
As an exhaustive, transient simulation of manifold dynamics was not possible within the time constraints of this project, a steady-state solution was used to help identify areas for improvement. This was done in an effort to minimize minor flow losses which have a much more profound effect on VE at high engine speeds [1]. It is understood that steady-state solutions do not fully capture the complex manifold dynamics that develop within the manifold (see [7] and [8] for a more exhaustive study).
Despite these advantages, the design time for the half-tube approach was greater due to packaging issues with the long restrictor and the long (220 mm) sliding surface along which the half-tubes actuate. These two features together were difficult to fit around the engine block, the frame and the FSAE® surface envelope constraint while trying to minimize the number of bends and number of fabricated parts.
The key areas of interest were:
Figure 6 shows a detailed section view of the new design concept. The cross-sectional area of the fixed runner was carefully designed to have at most 1.5% variation relative to the intake port area as it transitioned from the engine's oval geometry to a circular shape. Blair had found that such a transition was beneficial to VE [6]. Additionally, the variable runner geometry includes a 25% reduction in inlet crosssectional area for the fully extended runner position which was sized using unpublished, empirical cross-sectional area studies done by Purdue FSAE®. Reducing runner diameter is discussed extensively in the literature [1] to improve low engine speed VE, although the particulars of the approach described here are unique in that only the end condition area is reduced. The reduction in area was primarily done to improve the flow characteristics of the half-tube design, by providing a more gradual transition within the packaging constraints detailed above. That said, future studies should further o ptimize this transition.
• Whether there were any localized regions of Mach >0.5 outside the restrictor.
• The flow field leading into the half-tube bellmouth. • How the flow distribution varied as the runners were actuated between the various runner positions. • If recirculation developed in the diffuser region after the restrictor terminated at the top of the plenum.
The following discussion will display contour plots along various planar slices through the combined fluid volume of the plenum and runners. These slices are necessary to visualize a 3D flow field on a 2D piece of paper. Figure 7 should be used as a reference as to which cross-sectional slice is being discussed.
Figure 7. Contours of Mach number for various rendering planes under 6900 Pa pressure drop (regular ellipse geometry)
Figure 6. Detailed section view of half-tube concept
The pressure-based 3D solver was used with the realizable κε turbulence model for these studies [7]. The imposed pressure drop was 6900 Pa, which was chosen such that the flow was partially choked. Claywell et al. had determined that the restrictor is never fully choked, even at an engine speed of 14,000 rpm, and thus a partially choked condition
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
was felt to be more reflective of the typical operating regime in which the intake would be operated [8]. Again, a full transient simulation is necessary to truly capture the manifold dynamics. With regard to Mach number, Winterbone and Pearson discussed empirical data that intake com ponents should be sized such that the Mach number does not exceed 0.5, otherwise a precipitous drop of in VE would ensue [1]. By this criterion, the FSAE® mandated restrictor is clearly the biggest shortcoming of the designed intake (Figure 7). The first problem region identified through steady-state 3D CFD simulation was an area of recirculation within the original runner design (Figure 8). This issue was not easily remedied, as the original runner geometry had been designed to package the intake within the FSAE® surface envelope and other components of the 2009 vehicle. Nevertheless, the mass flow gains (>8%) from removing this region of recirculation were enough to justify a review of the original packaging. After extensive iterative optimization of the original centerline curve, an arrangement was found that packaged the intake well around the previously discussed criteria, with only a marginal (<10 mm) increase in fixed runner length.
Figure 9. Top view of velocity contours along mid plenum plane
The simple radius proved to be a localized point of turbulence generation, and is included in Figure 9 as a baseline for comparison for later intake geometries. The reader should note that the maximum mass flow is appreciably lower in the simple radius design primarily due to recirculation in the runners (see Figure 8) which was remedied in the later designs. In order to explore the mass flow variation through the entire dynamic range of runner lengths, an additional steady-state study was performed to identify fluid flow losses as the intake was actuated from its shortest to longest position. The results of the study are presented in Table 3. These simulations were done using the regular ellipse geometry (except where elongated ellipse geometry is noted). Table 3. Steady-state mass flow variation for various runner positions
Figure 8. Velocity vectors along the mid-runner plane for original and final runner geometry
Blair identified the optimal bellmouth geometry to be a short, wide ellipse (the parameters of which are best described within [6]). He came to this conclusion using transient, 3D simulations of simple radii, airfoils and ellipse geometries with differing lengths and widths. In the int erest of comparing the relative contribution of the bellmouth geometry on peak mass flow for the designed intake, a series of studies were performed using the largest ellipse geometry that could package well (Figure 9).
From these data, it is clear that the outer two cylinders do not receive as much flow at long runner lengths, as they are offset from the centerline of the restrictor. This may have been remedied through a longer diffuser section after the r estrictor, but packaging and fabrication constraints did not afford enough space to accomplish this. Also shown is that the elongated ellipse geometry reduced the cylinder imbalance from 2.8% to 2.3% for the 90 mm runner length.
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
In addition to mass flow studies, the inlet to the runner for the final geometry was carefully inspected using a 3D vector view of the flow field. It was found to have a smooth transition into the runner, with no significant points of recirculation or other anomalies. This suggests that the method of runner length actuation itself does not present appreciable flow losses while still allowing the full dynamic range of runner lengths.
MECHANICAL DESIGN Hobby servo motors, linear actuators and solenoids were screened for their potential use in actuating the variable runners. It was found that solenoids and linear actuators were unnecessarily heavy for this application and required a linkage system to cover the full dynamic range. High-end hobby ser vos did not present these issues, and thus an actuator scheme was implemented around a high-torque Futaba® S5801 sail winch. This servo provided 6 full rotations (which represents significantly more than the typical servo rotation of 180 degrees) and it came stock with a winch drum for winding up cables. This drum proved to be of an appropriate size for actuating the full dynamic range of the variable runners. To efficiently package the system, a mount was incorporated into the fuel rail for the servo and cable. One of advantages to the half-tube intake concept was that a simple, low-tolerance cable linkage setup could slide the runners along the bottom plenum surface. Additionally, by using a cable linkage, the servo could be placed external to the intake without a complicated sealing method. The reasons this was advantageous were two-fold: For one, the servo was perceived to be the most frequently serviced component and easy access to it was desirable. The second reason was that it minimized the quantity of small parts residing within the plenum. In the event of an intake failure, such small components could easily make their way into the cylinder and cause engine damage. The basic actuation mechanism is shown in Figure 10. It consists of a hardened stainless steel shaft and two high-misalignment linear bearings that guide the variable runners. To seal the variable runners, o-ring stock was installed in channels along the bottom of the variable runners (see Figure 6). With the application of lubricant, having the o-ring stock ride along the surface of the bottom plenum offered a simple solution to sealing the runners as they actuated. Additionally, this approach is able to take advantage of the negative dynamic pressure developed within the runner from the mean flow into the cylinders. To seal the rear portion of the variable runner, flaps of rubber were incorporated and glued onto the back side of the fixed runner (see Figure 6). The final selected geometry is shown in Figure 11. Note that only a single fixed runner is displayed to simplify the diagram.
Figure 10. The sliding mechanism
Figure 11. Final design overview
The material of choice for this low production run, weight sensitive application, with thin walls and complex curves is Carbon Fiber Reinforced Polymer (CFRP). While the particulars of CFRP design and fabrication are outside the scope of this paper, a quasi-isotropic layup schedule was used to aid in FEA simulation and because of the expected uniform pressure loading. It was necessary to use a coring material (Divinycell® H100 Semi-rigid PVC foam) to enhance the strength of the predominantly flat plenum surfaces. A geometrically stronger curved plenum design where the variable runners rotated was considered, but the fabrication of the design was deemed too complex. A detail view of the fabricated intake actuation system is shown in Figure 12. After assembly, timed trials of the actuation system showed that the intake was able to transition its full dynamic range (shortest to longest runner length) in under 1.0 s. The static friction resistance to sliding with lubrication was measured to be 8 N. Figure 13 shows the fabricated intake manifold on the engine prior to the installation of the top plenum half.
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
PRELIMINARY RESULTS The fabricated intake was installed on a recently rebuilt 2003 Suzuki GSXR-600® (specifications given in Table 1). The engine had completed a break-in and calibration procedure using the 2008 intake. For the new intake, the engine was recalibrated for maximum torque at WOT with the runner length actuated to the shortest runner position.
Figure 12. Detail view of fabricated half-tube concept
After calibration, the engine was allowed to slowly sweep through a speed range of 7000 to 12000 rpm (at WOT) while torque was logged. This was repeated twice for 10 different runner lengths. The two runs at each runner length were averaged and all of the results are plotted in Figure 14. Note that the torque data for each sweep were collected after the engine's internal gearbox using a Mustang® MDT-70 eddy current dynamometer and corrected using SAE® J1349 compensation. Pump gasoline (93 AKI) was used for all runs and an Innovate Motorsports® LC-1 wideband Air-Fuel Ratio (AFR) sensor was used for closed-loop fueling control.
Figure 14. Contours of measured torque (N·m) at WOT
Although a direct measurement of VE would have more accurately separated intake performance from combustion and friction considerations, it was not available for these tests. For these preliminary trials this was deemed acceptable provided fueling was adjusted under closed-loop AFR control. Because engine torque is proportional to VE when all other variables are constant [2], these data do provide a rough indication of the VE variation. Figure 15 presents the percent difference in torque from a 325 mm runner length baseline (similar to Figure 3). Figure 13. Fabricated intake on the engine prior to installation of the top plenum half
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
through a continuously variable system. In addition to being variable, these results also demonstrate a peak power output improvement of greater than 22% versus the 2008 intake design.
Figure 15. % difference in measured torque at WOT from 325 mm runner length baseline
The results show the expected shift in the resonance peak of the engine. Compared to the simulations presented earlier, the minimum torque is significantly higher in value while the peak torque follows a more complex distribution. Figure 16 shows an envelope of all the torque values presented in Figure 14. The torque curve when the runner length is equal to 289 mm is also shown as a baseline for comparison between fixed and variable operation. This baseline runner length was selected on the basis that it has a relatively long, flat torque curve that would be desirable for the driver on a road course if the intake was operated in a fixed runner length configuration.
Figure 17. Power curves for selected runner lengths at WOT
Transient operation of the variable intake under servo control is presented in Figure 18 for a constant engine speed of 9500 rpm. Since the system is able to respond to a less than 1.0 s step input, these torque data support the assumption that closed-loop fuel adjustment is able to compensate for VE variation. The reader should note that the torque data shown here are higher in value than those presented earlier. This discrepancy is due to improvements that were made to runner sealing after the initial dataset was acquired (Figure 14). The gradual decay in torque output is due to inadequate cooling within the test cell. Heat was also identified as a source of a servo failure, and it will necessitate either relocating the servo away from the engine or designing a new servo system with greater temperature tolerance.
Figure 16. Envelope of all torque values at WOT compared to 289 mm performance
By shifting the runner position, the intake is able to improve the torque output of the engine by 2-5% relative to the 289mm baseline. Figure 17 shows how the power peak transitions for selected runner lengths. This plot shows that the typical variable intake approach of toggling between two runner lengths misses a sizable portion of the gains to be had
Figure 18. Transient runner length transitions at constant engine speed (9500 rpm) and WOT
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
CONCLUSIONS & RECOMMENDATIONS The design, analysis and successful testing of a continuously variable, carbon fiber engine intake manifold has been described in this work. The intake can vary its runner length from 216 to 325 mm. Preliminary testing of the intake showed the expected torque and power variation, with a peak power improvement of over 22% relative to the 2008 design and a 5% torque output gain compared to a 289 mm runner length baseline. This variable intake system increases the powertrain weight by less than 1% and can transition through its full dynamic range in less than 1.0 s. The intake was designed through published data, a Ricardo WAVE® model and steady-state 2D/3D Ansys Fluent® CFD code. Optimization of restrictor geometry was accomplished through a custom software routine that was developed to automatically mesh and execute cases in Ansys Fluent® [10]. Final intake geometry was determined through a series of packaging and fabrication compromises detailed in this work. The variable runner mechanism was actuated through a servo motor and a cable linkage which was controlled by a Microchip dsPIC® microcontroller. In the future, restrictor DoE software should be expanded into a more general automated design optimization package to aid in more than just restrictor design. An extension has already been coded to interface to MATLAB® for such routines. It is believed that by limiting trials to good performers from the current steady-state results and systematically limiting the selection process with a gradient-based optimization scheme, the simulations can be accelerated enough to allow a full transient design optimization. Additionally, the preliminary performance data should be incorporated into the Ricardo WAVE® model to better calibrate it and allow for better computer aided engine optimization. It is hoped that future studies can build on the hardware and software developed for this project to further improve and characterize the half-tube variable intake concept. The current engine calibration should be refined (with optimized cam timing) and coupled with the empirically determined runner length versus engine speed look-up table code embodied in the Microchip dsPIC® microcontroller. Controller code should be expanded to incorporate traction control and driving style considerations. Finally, a direct measurement of VE should be used for future studies.
2. Heywood, J. B., Internal Combustion Engine Fundamentals, McGraw-Hill, New York, ISBN 0-07-028637X, 1988. 3. Brüstle, C. and Schwarzenthal, D., “VarioCam Plus - A Highlight of the Porsche 911 Turbo Engine,” SAE Technical Paper 2001-01-0245, 2001, doi:10.4271/2001-01-0245. 4. Kreuter, P., Heuser, P., Reinicke-Murmann, J., Erz, R. et al., “Variable Valve Actuation - Switchable and Continuously Variable Valve Lifts,” SAE Technical Paper 2003-01-0026, 2003, doi:10.4271/2003-01-0026. 5. Cauchi, J., Farrugia, M., and Balzan, N., “Engine Simulation of a Restricted FSAE Engine, Focusing on Restrictor Modelling,” SAE Technical Paper 2006-01-3651, 2006, doi:10.4271/2006-01-3651. 6. Blair, G. P., Cahoon, W., “Special Investigation: Design of an Intake Bellmouth,” Race Engine Technology: 34-41, September 2006. 7. Claywell, M., Horkheimer, D., and Stockburger, G., “Investigation of Intake Concepts for a Formula SAE FourCylinder Engine Using 1D/3D (Ricardo WAVE-VECTIS) Coupled Modeling Techniques,” SAE Technical Paper 2006-01-3652, 2006, doi:10.4271/2006-01-3652. 8. Claywell, M. and Horkheimer, D., “Improvement of Intake Restrictor Performance for a Formula SAE Race Car through 1D and Coupled 1D/3D Analysis Methods,” SAE Technical Paper 2006-01-3654, 2006, doi: 10.4271/2006-01-3654. 9. Zimmerman, S., Cordon, D., Anderson, M., and Beyerlein, S., “Development and Validation of an Impedance Transform Model for High Speed Engines,” SAE Technical Paper 2005-01-3803, 2005, doi:10.4271/2005-01-3803. 10. Vaughan, A. and Delagrammatikas, G.J., “Variable Runner Length Intake Manifold Design: An Interim Progress Report,” SAE Technical Paper 2010-01-1112, 2010, doi: 10.4271/2010-01-1112. 11. McKee, R.H., McCullough, G., Cunningham, G., Taylor, J.O. et al., “Experimental Optimisation of Manifold and Camshaft Geometries for a Restricted 600cc Four-Cylinder Four-Stroke Engine,” SAE Technical Paper 2006-32-0070, 2006, doi:10.4271/2006-32-0070. 12. Hamilton, L. J., Cowart, J. S., Lee, J. E., and Amorosso, R. E., “The Effects of Intake Plenum Volume on the Performance of a Small Naturally Aspirated Restricted Engine,” ASME Paper ICEF2009-14036, 2009.
REFERENCES 1. Winterbone, D. E., and Pearson, R. J., Design Techniques for Engine Manifolds, Professional Engineering Publishing, London, ISBN 186058179X, 1999.
ACKNOWLEDGMENTS Fabrication of the parts described in this work would not have been possible without the CNC expertise of Sinisa Janjusevic. James Abbott's guidance was instrumental with the various acoustical aspects inherent to this project. Numerous Formula
Downloaded from SAE International by SRM Univ, Friday, April 25, 2014 03:50:31 AM
SAE® team members have also assisted with this project: Muneeb Hai, Greg Shikhman, Rob Smith, Kim Meehan and Tim Fedullo.
The Engineering Meetings Board has approved this paper for publication. It has successfully completed SAE's peer review process under the supervision of the session organizer. This process requires a minimum of three (3) reviews by industry experts. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of SAE. ISSN 0148-7191
Positions and opinions advanced in this paper are those of the author(s) and not necessarily those of SAE. The author is solely responsible for the content of the paper. SAE Customer Service: Tel: 877-606-7323 (inside USA and Canada) Tel: 724-776-4970 (outside USA) Fax: 724-776-0790 Email:
[email protected] SAE Web Address: http://www.sae.org Printed in USA