Effect of Primary Intake Runner Tapers and Bellmouths on ...engine.osu.edu/ASJP/JP/J72.pdf · with a one-cylinder diesel engine, found that the engine would not start unless the compression

400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-0790 Web: www.sae.org

SAE TECHNICALPAPER SERIES 2007-01-0382

Effect of Primary Intake Runner Tapers andBellmouths on the Performance of a

Single Cylinder Engine

V. Mariucci and A. SelametThe Ohio State University

K. D. MiazgowiczFord Motor Company

Reprinted From: Modeling of SI and Diesel Engines, 2007(SP-2079)

2007 World CongressDetroit, MichiganApril 16-19, 2007

Author:Gilligan-SID:12324-GUID:50206991-164.107.10.90

By mandate of the Engineering Meetings Board, this paper has been approved for SAE publication uponcompletion of a peer review process by a minimum of three (3) industry experts under the supervision ofthe session organizer.

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, ortransmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise,without the prior written permission of SAE.

For permission and licensing requests contact:

SAE Permissions400 Commonwealth DriveWarrendale, PA 15096-0001-USAEmail: [email protected]: 724-776-3036Tel: 724-772-4028

For multiple print copies contact:

SAE Customer ServiceTel: 877-606-7323 (inside USA and Canada)Tel: 724-776-4970 (outside USA)Fax: 724-776-0790Email: [email protected]

ISSN 0148-7191Copyright © 2007 SAE InternationalPositions 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. A process is available by which discussionswill be printed with the paper if it is published in SAE Transactions.

Persons wishing to submit papers to be considered for presentation or publication by SAE should send themanuscript or a 300 word abstract of a proposed manuscript to: Secretary, Engineering Meetings Board, SAE.

Printed in USA


ABSTRACT

The present experimental study investigates systematically the effects of primary intake runner configurations on a firing single cylinder research engine. Twelve different intake configurations were fabricated to investigate runners with tapers and bellmouths. For each configuration, the length from the base of the test configuration to the start of the inlet radius and pressure measurement locations were retained in an effort to isolate the effect of runner geometry only. Each configuration is presented against a baseline case with constant cross-sectional area and a ratio of inlet radius, Ri, to internal diameter, D, of 0.35. For the seven tapered runners, the length of the taper varied from 25% to 100% of the overall length of the test piece, and the taper area ratios (TAR) varied from 1.5 - 3; all tapers retained the inlet radius of the baseline. The four bellmouth runners had a constant cross-sectional area, and varying inlet radii from Ri/D = 0.05 to 1.0. The time-averaged quantities such as volumetric efficiency and brake power, and time-resolved intake pressures are presented for each configuration. For the bellmouth runners, Ri/D > 0.20 was found to be most beneficial to volumetric efficiency, and a TAR larger than 1.5 was detrimental to the intake tuning for higher-speed tuning peaks.

INTRODUCTION

Whether the motivating factor is more power or better fuel economy, several techniques have been employed throughout the years to improve the IC engine. One of the most well-known and used methods of enhancing an engine’s performance is by intake tuning. Due to the unsteady nature of the gas exchange process in the IC engine, pressure fluctuations are generated when the engine takes in air. During the intake stroke of a single cylinder engine with intake pipe leading from atmosphere to the cylinder, for example, as

the piston draws in fresh charge, an expansion wave propagates toward the inlet of the intake pipe. When this expansion wave reaches the inlet of the pipe, it is partially reflected as a compression wave. If this compression wave arrives back at the intake valve while it is still open, it can aid the breathing process of the engine, thus increasing the volumetric efficiency ( v).This charging effect will happen near certain engine speeds, governed by the intake geometry, giving rise to “tuning peaks.”

Numerous early works have investigated the effect of varying primary intake runner geometry on intake tuning. In 1924, Matthews and Gardiner, working with a one-cylinder diesel engine, found that the engine would not start unless the compression pressure was increased. They found that a long suction pipe attached to the inlet port gave them the pressure increase they needed, and proceeded to test several lengths of pipe up to 177.8 cm in addition to the case where no pipe was present from 500 to 1800 RPM; they noticed an increase in compression pressure of 17% with the 62 in. pipe compared to no inlet pipe. The study was purely experimental and did not attempt to explain the reason for the increase in pressure.

Morse et al. (1938) developed a simple method of intake tuning prediction, assuming linear, acoustic behavior in the intake duct. This method suggests that the quasi-standing wave (QSW) developed in the intake pipe when the intake valves are closed dictates intake tuning. It was found that when the ratio q of pipe frequency ( = c/4 ) to engine frequency ( = RPM/120) was equal to 3, 4, or 5, beneficial resonance would occur since the pressure fluctuations at the valves will be large, leading to

qc

N30

, (1)

1

2007-01-0382

Effect of Primary Intake Runner Tapers and Bellmouths on the Performance of a Single Cylinder Engine

V. Mariucci and A. Selamet The Ohio State University

K. D. Miazgowicz Ford Motor Company

Copyright © 2007 SAE International


where N [RPM] is the engine speed of beneficial resonance, c [m/s] is the speed of sound, [m] is the length of the intake pipe, and q = 3, 4, or 5. They verified their method by testing several lengths of intake pipe of 6.5 cm diameter at a constant engine speed of 1220 RPM.

Kastner (1945) studied the intake tuning effect of various lengths and diameters of primary runners on two separate single-cylinder engines. Standard air consumption, intake port and in-cylinder pressures were analyzed. One engine, a Rolls-Royce, was motored between 1000 and 2000 RPM with several lengths of intake pipes of diameters 3.5, 5.3, and 7.3 cm. The other engine, a J.A.P. motorcycle engine, was fired from 1200 to 4000 RPM with various runner lengths of diameters 3.2, 3.5, and 5.3 cm. Kastner found that, for an engine running over a wide range of speeds, the intake pipe “natural period” should be between 160-180 crank angle degrees (CAD) at the engine speed of desired tuning, where natural period, p, is defined as

pVN

p24

[deg] , (2)

where N [RPM] is the engine speed, [cm] is the effective intake length, and Vp [cm/s] is the velocity of propagation of the waves. Note that when p = 180 CAD, the expression is equal to Eq. (1) with q = 4. Kastner also noted that Vp increased as the intake diameter increased.

Downing (1958) performed a study on the effects of primary runner length and diameter using one cylinder of a motored six-cylinder engine. He tested runner lengths of 13.3, 27.9, 38.1, 54.6, and 82.6 cm for diameters of 4.0 and 5.0 cm from 1500 to 6000 RPM for the purposes of designing a complete fuel-injection system for a racing engine, including intake manifold. He found that, for both intake diameters, as the length increased, the tuning peak moved toward lower speeds. Also, for same-length runners, the tuning peak of the 5.0 cm diameter pipe was at a higher speed than the 4.0 cm pipe except for the shortest pipe where the tuning peaks for both diameters were nearly identical.

Later, Engelman and his coworkers developed an intake tuning prediction tool based on a lumped Helmholtz resonator; they modeled the intake and cylinder by treating the air in the cylinder as a spring with no inertia and the air in the runner as a mass with no compressibility. The tuning peak is predicted to occur when the resonant frequency of the Helmholtz system is around twice the piston frequency. For a single cylinder with an intake runner open to atmosphere, Engelman (1973) suggested for the speed of the tuning peak

2/1955][

efft V

AcK

RPMN , (3)

where c [m/s] is the speed of sound, A [cm²] is the effective cross-sectional area of the intake, [cm] is the effective length of the inlet system, K is a factor usually equal to 2.1, but that varies from 2.0 to 2.5 depending upon valve timing and other factors, and,

)1(2)1(

][ 3

c

cdeff r

rVcmV , (4)

where Vd [cm3] is the displacement volume and rc is the compression ratio. This method provides a simple tool for calculating the location of a single intake tuning peak, yet gives no insight into its magnitude. Even though it incorporates extra parameters over the organ pipe method such as the pipe area and cylinder volume, which Thompson (1968) showed gives a predictive advantage over the organ pipe method, the value for K is slightly ambiguous without experimental data. Intakes of non-constant cross-section, such as tapers, can be handled by using (Engelman, 1973)

L N

j j

j

effective AAdx

A 0 1 . (5)

In 1968, Thompson investigated the effects of not only different lengths and areas of primary runners, but also of curved runners, pipes with and without well-rounded inlets, non-circular tubes, and runners of varying cross-sectional area. His experiments were performed on individual cylinders of a motored Cummins V6 diesel engine and compression pressure was acquired from 500 to 3000 RPM. The main purpose of his thesis was to validate Engelman’s approach to intake tuning prediction.

In a precursor to the present study, Howard (2003) studied the effects of tapered runners on a motored single cylinder research engine. His tapers varied from TARs of 1.5 to 3 and the taper length was varied from 25% to 100% of the overall length of the test piece. He examined the tapers’ effects on v as well as intake and exhaust pressure. He found that, as TAR increased, the tuning peaks moved to higher speeds, which correlated well with the predicted speeds given by the lumped Helmholtz intake approximation. He also observed that, in general, the peak intake pressure near the valve before intake valve closing (IVC) was indicative of v trends, but the correlation was sensitive to the timing of the peak pressure.

Although the literature is extensive on the effects of primary intake runner geometry, most of it has focused on changing lengths of constant-area tubes. Little is available on the effects that taper and inlet

2Author:Gilligan-SID:12324-GUID:50206991-164.107.10.90

3

radius of the primary runner have on the v, brake power, and intake pressure of an IC engine. Thus, the objective of the present study is to systematically study the effect of these geometries on engine performance. Seven tapered runners with TAR varying from 1.5 to 3 and taper length varying from 25% to 100% of the overall length of the test piece and four bellmouth runners varying from Ri/D of 0.05 to 1 were tested on a single cylinder engine at speeds from 1000 to 5500 RPM.

EXPERIMENTAL SETUP

ENGINE - All experiments for this work were performed on a single cylinder research engine designed by Ford Motor Company. This is a 4-valve spark-ignition engine designed to mimic one cylinder of the Jaguar 3.0L V6 X200 engine, including the combustion chamber, bore and stroke, piston geometry, valve timing, and intake and exhaust ports. Although the Jaguar engine has variable cam phasing, the single cylinder engine does not. The only engine accessory is a dry sump oil pump which is run off the timing belt; the water pump is electric and not powered by the engine. The camshaft follower is a direct-acting mechanical bucket. The overall engine specifications are given in Table 1.

Bore 8.90 cm Stroke 7.95 cm Rod Length 13.81 cm Compression Ratio 10.5:1Clearance Volume 47.10 cm3

Maximum Valve Lift Intake 0.914 cm

Exhaust 0.937 cm Valve Timing

Intake Open 308.0 CAD Intake Duration 286.0 CAD Exhaust Open 86.5 CAD

Exhaust Duration 326.0 CAD

Table 1. Single Cylinder Research Engine Specifications.

The fuel and spark delivery to the engine was controlled by the Haltech E6K programmable engine control unit. It was connected to the engine through a custom-built wiring harness and programmed through a PC. The spark timing was set at each speed to a conservative amount close to MBT of the baseline and remained constant for each speed during all experiments. The air-fuel ratio was set to 12.5:1 for all experiments and monitored with a Horiba Mexa-110AFR Analyzer.

The single cylinder engine was connected to a General Electric DC engine dynamometer (dyno) capable of both motoring the engine and absorbing the load while firing at a specific speed. It is capable of delivering 134 kW as a motor and absorbing 142 kW as a dyno; the maximum speed is 6300 RPM. The dyno was computer-controlled by Horiba Systems EDTCS-1000. Each intake was tested from 1000 to 5500 RPM in 500 RPM increments from 1000 to 2000 RPM, then in 250 RPM increments from 2000 to 5500 RPM. Additional points were tested around the v peaks for each intake configuration. All experiments were performed at wide-open throttle (WOT), and the oil and cooling water were held to 93°C.

INTAKE AND EXHAUST - The intake of the single cylinder engine, shown schematically in Fig. 1, consists of the intake port, a fuel rail block, an adapter section, a barrel-style throttle, and the test piece. The intake port is a split-port design with a diameter of 3.02 cm at each valve; the port stays separated for 6.48 cm, then the two branches come together in an oval of 14.32 cm² for 3.52 cm until the head face. Although all experiments are done at WOT, a barrel throttle is used in case the engine needed to be throttled due to unexpected knocking or other engine-damaging phenomena. The throttle is made such that at WOT, there would be no obstruction in the flow to the pressure waves. The only component of the intake that varied for each experiment is the test piece, which is described in detail in the next section.

The exhaust, shown schematically in Fig. 2, is designed to lead the exhaust gas outside as efficiently as possible while keeping external noise to an acceptable level. It is made of 6.03 cm diameter straight pipe and mandrel-bent 90° elbows with bend radii of 19.4 cm. The exhaust port is a split-port design with a diameter of 2.54 cm at each valve; the port stays separated for 4.86 cm, and then comes together into a 10.18 cm² oval for a length of 3.48 cm until the head face. A 16.5 cm long adapter, shown in Fig. 3, provides a smooth transition from the oval exhaust port of 10.18 cm² to the rest of the exhaust with a diameter of 6.03 cm. A silencer is placed between the exit of the 6.03 cm diameter section of the exhaust and the large tube that routes the exhaust gas outside.

TEST PIECES – Twelve configurations consist of a baseline case and two groups of geometries: tapers and bellmouths. One characteristic common to all test pieces is the large flange at the entrance to the piece, which ensures a hemispherical propagation of the pressure waves leaving the duct. The baseline, designated intake #1 and shown in Fig. 4a, is a straight duct 26.45 cm long with an inner diameter (D) of 4.20 cm. At the inlet of the baseline piece is a 1.45 cm radius bellmouth and the large flange.


Fuel Rail Block

Barrel Throttle

Adapter Section Test Piece

Intake Valves

Pressure Transducer (i1)



10.0

5.14

12.0

8.679.51

4.20 DIA4.20 DIA4.31 equivalent DIA

Dimensions are in cm

Engine Head Face

20.0

Figure 1. Single Cylinder Intake.

16.5

R 19.4

R 19.4

R 19.4

90.2

263

55.8Muffler

to outside

Pressure Transducer (e1)

Dimensions are in cm

In-Cylinder Pressure Transducer (c1)

6.35

12.7

Engine

Figure 2. Single Cylinder Exhaust.


5

9.53

3.18

8.57

16.5

27.3

Exhaust Valves

Pressure Transducer

Thermocouple

Emissions Tap

Head Face

Oxygen Sensor

6.03 DIA

8.34

3.60 equivalent DIA

Dimensions are in

(e1)

cm

Figure 3. Adapter Section of Single Cylinder Exhaust.

4.20 DIA

Lt

Dt

20.0

6.45

Ri=1.45


All Dimensions are in cm

(b)

4.20 DIA

20.0 5.00


Ri

4.20 DIA

All Dimensions are in cm

(a)

Figure 4. Intake Test Pieces: (a) baseline and bellmouth group, (b) taper group.


The tapers, referred to as intakes #2 – 8, are shown in Fig. 4b along with the key dimensions given in Table 2. They have overall lengths and inlet radii (Ri)equal to the baseline. The TAR varies from 1.5 to 3. For TAR = 1.5, the length of the tapered section (Lt)varies from 25% to 100% of the overall length and for TAR = 2, Lt varies from 50% to 100%. The others with TAR = 2.5 and 3 are tapered over their entire length.

Intake # Lt (cm) Dt (cm) TAR2 6.6133 13.234 26.45

5.14 1.5

5 13.236 26.45 5.94 2

7 6.62 2.58 26.45 7.26 3

6

Table 2. Taper Group Dimensions.

The bellmouth group, labeled as intakes #9 – 12, is shown in Fig. 4a along with the key dimensions given in Table 3. Each bellmouth piece retains the length from the start of the test piece to Ri and D of the baseline, and Ri/D varies from 0.05 to 1.0.

Intake # Ri (cm) Ri/D Overall Length (cm)

1 (baseline) 1.45 0.35 26.459 0.21 0.05 25.21

10 0.84 0.20 25.8411 2.10 0.50 27.1012 4.20 1.00 29.20

Table 3. Bellmouth Group Dimensions.

EXPERIMENTAL MEASUREMENTS - The intake of the single cylinder engine is outfitted with three Kistler 4045A2 piezoresistive pressure transducers with a 0-2 bar range. Each of these transducers is connected through a Kistler 4603A Amplifier to a Concurrent high-speed data acquisition system (model MC68040). This system is capable of acquiring data at a sampling rate of 2 MHz from 32 channels simultaneously. The pressure transducers labeled i2 and i3 are both located in the adapter section of the intake, 14.65 cm from the head face as shown in Fig. 1. Both transducers are flush with the inner wall of the adapter piece, and both are perpendicular to and point toward the centerline of the duct. The pressure transducer labeled i1 is located in

the test piece. An aluminum bushing is made to push into each test piece and fish-mouthed to fit flush with the inner wall of the duct. The transducer threads tightly into this bushing, also is flush with the inner wall of the duct and is perpendicular to and points toward its centerline. For each piece, the i1 transducer is located 20.00 cm upstream from the surface that mates with the barrel throttle. The pressures acquired at locations i1, i2, and i3 have a resolution of 1 CAD and are averaged for 64 cycles. Above the inlet of the intake duct is a thermocouple for measuring ambient air temperature.

The exhaust of the single cylinder engine is fitted with a pressure transducer, an emissions tap, a thermocouple, and a wide-band oxygen sensor (recall Fig. 3). The pressure transducer, e1, is a Kistler 4045A5 piezoresistive pressure transducer with a 0-5 bar range. This transducer threads into a water jacket, which is supplied with cooling water. The water jacket is threaded into its bung in the adapter section such that it is flush with the inner wall of the duct. The e1 transducer is located 9.53 cm from the head face. As with the intake pressure traces, the exhaust pressure traces at e1 have a 1 CAD resolution and are averaged for 64 cycles. The thermocouple is located 8.57 cm from the head face. The wide-band oxygen sensor is located 27.3 cm from the head face and connected to a Horiba MEXA-110 AFR Analyzer. The emissions tap is connected first through stainless steel tube and then macroline to a Horiba MEXA-7100 Motor Exhaust Gas Analyzer. The engine brake and motoring torques are measured with a Revere Transducers load cell (model 606555-24). The fuel flow is metered with the Pierburg Instruments PII401*11863 Fuel Measurement System. Emissions, torque, and fuel flow data are acquired by the Horiba dyno controller computer and are averaged over a minute at each engine speed.

A PCB 145A07 Piezotronics in-cylinder pressure transducer, labeled c1 in Fig. 2, is located in the top center section of the combustion chamber, fitting flush with its inner surface, between the spark plug and the rearward-most intake and exhaust valves. The output, amplified via a Kistler 5010 dual-mode amplifier, is recorded by the Concurrent data-acquisition system.

Using the air-fuel ratio obtained from the MEXA-7100 and the fuel flow rate, v is calculated for each engine speed by

NVmdia

av

,

2 , (6)

where [g/s] is the mass flow rate of air (determined from the measured fuel flow rate and the air-fuel ratio) into the cylinder, Vd [cm3] is the displacement volume, N[RPS] is the engine speed, and a,i [g/cm3] is the inlet air density. The inlet air density is determined from the

am


ideal gas relationship by using the barometer and temperature readings.

1000 1500 2000 2500 3000 3500 4000 4500 5000 550075

82

89

96

103

110

117

Vol

umet

ric E

ffici

ency

(%)

Engine Speed (RPM)1000 1500 2000 2500 3000 3500 4000 4500 5000 5500

3

6

9

12

15

18

21

Cor

rect

ed B

rake

Pow

er (k

W)

1000 1500 2000 2500 3000 3500 4000 4500 5000 55003

6

9

12

15

18

21

solid = run 1dashed = run 2

Volumetric Efficiency

Power

RESULTS

This section presents v, brake power, and intake pressures for each test piece. For brevity, only intake pressures at i2 are presented since they are more indicative of v trends than i1, which is further upstream. Locations i2 and i3 were the same distance from the back of the intake valves, and pressures measured there proved to be identical. For further information on pressure measurements from the i1, e1, and c1 locations, the reader is referred to Mariucci (2006).

BASELINE – The v and brake power for repeated runs of the baseline are presented in Fig. 5. The level of repeatability observed in this figure is similar in other test pieces. There are three distinct v peaks, at 3000, 3750, and 4750 RPM. With c = 348 m/s and = 0.6226 m (the total geometric length of the intake), the intake tuning prediction method developed by Morse et al. [recall Eq. (1)] gives the tuning peak speeds shown in Table 4. This method over-predicts the tuning peak speeds for all three values of q. Using Engelman’s lumped-parameter Helmholtz model [recall Eq. (3)] with c = 348 m/s, = 62.26 cm, A = 13.85 cm2, and a preliminary guess of K = 2.1, a tuning peak location was predicted at 4314 RPM, a 15% difference when compared to the largest experimental tuning peak at 3750 RPM. To accurately predict this tuning peak, K = 2.42, within the range prescribed by Engelman.

Figure 5. Experimental v and Brake Power for the Baseline.

1000 1500 2000 2500 3000 3500 4000 4500 5000 550075

80

85

90

95

100

105

110

115

Vol

umet

ric E

ffici

ency

(%)

Engine Speed (RPM)

intake # 1intake # 2intake # 3intake # 4

(a)

1000 1500 2000 2500 3000 3500 4000 4500 5000 55003

6

9

12

15

18

21

Cor

rect

ed B

rake

Pow

er (k

W)

Engine Speed (RPM)

intake # 1intake # 2intake # 3intake # 4

(b)

q PredictedPeak (RPM)

ExperimentalPeak (RPM)

% Difference

3 5590 4750 17.74 4192 3750 11.85 3353 3000 11.8

Table 4. Morse et al. Method for Intake Tuning Prediction.

TAPERS - The measured v and brake power of the engine for each taper (intakes #2-8) are compared to the baseline. Figure 6 shows all intakes with TAR = 1.5 (intakes #2-4) and Fig. 7 with TAR = 2.0 (intakes #5 and 6). All intakes that are tapered over their full length (intakes #4, 6, 7, and 8) are compared in Fig. 8 to show the overall trend of increasing TAR. The location of the peak v and its magnitude are given in Table 5 for each taper.

As predicted by Eq. (3), the tuning peaks move toward higher engine speeds as /Aeffective decreases. The /Aeffective for each taper, calculated using Eq. (5), along with the predicted tuning peaks using the lumped-

Figure 6. Experimental (a) v and (b) Brake Power for Intakes with TAR = 1.5 Compared to the Baseline.


1000 1500 2000 2500 3000 3500 4000 4500 5000 550075

80

85

90

95

100

105

110

115

Vol

umet

ric E

ffici

ency

(%)

Engine Speed (RPM)

intake # 1intake # 5intake # 6

(a)

1000 1500 2000 2500 3000 3500 4000 4500 5000 55003

6

9

12

15

18

21

Cor

rect

ed B

rake

Pow

er (k

W)

Engine Speed (RPM)


(b)

Figure 7. Experimental (a) v and (b) Brake Power for Intakes with TAR = 2.0 Compared to the Baseline.

1000 1500 2000 2500 3000 3500 4000 4500 5000 550075

80

85

90

95

100

105

110

115

Vol

umet

ric E

ffici

ency

(%)

Engine Speed (RPM)

intake # 1intake # 4intake # 6intake # 7intake # 8

(a)

1000 1500 2000 2500 3000 3500 4000 4500 5000 55003

6

9

12

15

18

21

Cor

rect

ed B

rake

Pow

er (k

W)

Engine Speed (RPM)

intake # 1intake # 4intake # 6intake # 7intake # 8

(b)

Figure 8. Experimental (a) v and (b) Brake Power for Intakes Tapered over Entire Length Compared to the Baseline.

Intake #

Peak vlocation(RPM)

Peak vmagnitude

(%)

% difference from baseline

1 3750 111.5 --2 3850 112.1 0.533 3875 110.5 0.894 4000 111.6 0.095 4000 110.7 0.726 4125 109.6 1.707 4250 108.5 2.698 4375 107.5 3.59

8

Table 5. v Characteristics of the Taper Group.

Intake#

/A|effective (cm-1)

PredictedPeak

(RPM)

ExperimentalPeak (RPM)

%Difference

2 4.407 3787 3850 1.63 4.319 3825 3875 1.34 4.145 3905 4000 2.35 4.214 3873 4000 3.16 3.935 4008 4125 2.87 3.800 4078 4250 4.08 3.689 4139 4375 5.4

Table 6. Taper Intake Tuning Predictions using the Lumped-Parameter Helmholtz Approximation.


parameter Helmholtz approximation are presented in Table 6. Experimentally, the location of peak vincreases by 125 RPM as TAR increases by 0.5 for each intake tapered over its entire length; Eq. (3) did not predict the increase to be so large, thus the prediction error associated with this method generally increases with TAR. The overall magnitudes of peak v remainsimilar to the baseline for intakes #2 – 5, while the peak

v begins to diminish for intakes #6 – 8. This may be due to the decrease in /Aeffective, which decreases the inertial effect of the air due to weaker reflected compression waves. Another contributor may be the flow losses having a more pronounced effect as the engine speed of the peaks increases (Heywood, 1988).

9

The brake power for each taper generally follows v, with spikes in power occurring near v peaks.Peak power and its location are given in Table 7 for each taper. For intakes #2 – 5, the peak brake power occurs near the highest-speed v peak and is of similar magnitude to the baseline. The peak power for intakes #6 – 8 are near the maximum v speeds, and substantial increases in peak power are seen w.r.t. the baseline. High-speed brake power generally increases for the tapered intakes compared to the baseline, with the largest belonging to intake #6 (30.0% at 5500 RPM), intake #7 (35.8% at 5500 RPM), and intake #8 (33.6% at 5500 RPM).

Intake #

Peak power location(RPM)

Peak powermagnitude

(kW)

% difference from

baseline1 4750 18.66 --2 4750 18.87 1.133 4750 18.61 0.274 5000 18.56 0.545 5000 18.31 1.886 4250 19.06 2.147 4375 19.35 3.708 4500 19.77 5.95

Table 7. Brake Power Characteristics of the Taper Group.

For brevity, only the intake pressure at the i2 location (Pi2) are compared for a select set that includes intakes #1, 4, 6, 7, and 8. These intakes were chosen because of their significant differences in v with respect to one another. Comparisons are presented for speeds corresponding to the main tuning peak of the baseline case, 3750 RPM, and the main tuning peak of each of the other intakes. While examining the impact of pressure waves at the intake valve in the following comparisons, the distance between the transducer location i2 and the valve should be kept in mind. The

approximate time in CADs required for the compression wave to travel from i2 to the back of the intake valves may be estimated from

cN

CAD6

, (7)

where is the distance from i2 to the back of the valves (= 0.2465 m for this study), N [RPM] is the engine speed, and c [m/s] is the speed of sound. Equation (7) gives CAD = 16° at 3750 RPM, 17° at 4000 RPM, and 18° at 4250 RPM. Each pressure trace is presented in the crank-angle resolved time domain, as well as the engine-order resolved frequency domain.

Figure 9 shows Pi2 at 3750 RPM for intakes #1 and 4. For intake #4, the compression wave returning to i2 near IVC is reduced in magnitude by 0.166 bar (12.6%) at its peak, which decreases the effectiveness of intake tuning; at the same speed, Fig. 6 shows a reduction in v of 9.4% from intake #1 to intake #4. The dominant frequency of the “quasi standing wave” (QSW), defined as the pressure wave in the intake during the intake valve (IV) closed period, has increased for intake #4, from about 136 Hz for the baseline to 150 Hz. Note that any dominant frequency of the QSW presented in this paper is calculated by measuring the CAD between peaks of the QSW. For the straight baseline, the dominant frequency of the QSW is similar to the first resonance frequency of a quarter-wave silencer,

4c

f , (8)

where c [m/s] is the speed of sound and [m] is the effective length of the intake duct. With c = 348 m/s and = 0.6226 m, ƒ becomes 140 Hz, a 3% difference with

respect to the measured frequency. The increase in average intake area for the tapered case has a similar effect on the frequency of the QSW as shortening the effective length of the intake. For intake #4, there is a 5.5 dB reduction in sound pressure level (SPL) at order 2, the dominant order. Figure 10 shows Pi2 at the main tuning peak speed for intake #4, 4000 RPM. At this speed, the peak magnitude of the compression wave near IVC is larger for the baseline than for intake #4; however, it occurs 18 CAD later for intake #1 and right at IVC. Thus, it seems that this compression wave is slightly late in getting from i2 to the back of the valves, while the compression wave for intake #4 is early enough to affect positive intake tuning, giving a rise of 11% in v over the baseline case. At 4000 RPM, the dominant frequency of the QSW has increased for the taper case just as it had for 3750 RPM. The peak SPL of intake #4 has decreased by 0.5 dB compared to intake #1, while those of neighboring orders have increased.


0 90 180 270 360 450 540 630 7200.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #4

IVCEVCIVO(a)

0 1 2 3 4 5 6 7 8 9 10100

110

120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #4

c/4L (baseline) (b)

Figure 9. Experimental Pi2 at 3750 RPM for Intakes #1 and 4 in the (a) Time Domain and (b) Frequency Domain.

0 90 180 270 360 450 540 630 720

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #4

IVCEVCIVO (a)

0 1 2 3 4 5 6 7 8 9 10120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #4

c/4L (baseline) (b)


Figure 11 compares Pi2 at 3750 RPM for intakes #1 and 6. For intake #6, the compression wave arriving at i2 near IVC is drastically reduced, by 0.182 bar (13.8%), at its peak. At the same speed, Fig. 8 shows that v is reduced by 11% compared to the baseline. As with intake #4, an increase in the dominant frequency of the QSW is observed for intake #6 compared to the baseline: about 150 Hz for intake #6 vs. 136 Hz for the baseline. The peak SPL for intake #6 has moved from order 2 to 2.5 and reduced by 5.9 dB compared to intake #1. A direct comparison of pressure at the intake #6 tuning peak speed of 4125 RPM is not made since that speed is not acquired for the baseline case; instead 4250 RPM is presented in Fig. 12. The peak magnitudes of the compression waves occurring during the IV open period are similar for both intakes #6 and 1, but the peak occurs 10° sooner for intake #6, while the peak for the baseline case arrives at i2 right at IVC and thus its effect on intake tuning has decreased. Also, the

average pressure between intake valve opening (IVO) and exhaust valve closing (EVC) (during the valve overlap period) is higher for intake #6. This may mean the difference between pulling fresh charge into the cylinder (increasing v) and back flow of exhaust gas (decreasing v) during the overlap, and may be another reason for the 13% increase in v for intake #6 over intake #1 at 4250 RPM. Compared to the baseline, the frequency spectrum shows a decrease in SPL from order 0.5 to 1.5 and an increase from orders 2 to 3 for the taper, with the largest increase (15.5 dB) at order 2.5.

Pi2 at 3750 RPM for intake #7 is compared to the baseline in Fig. 13. The peak pressure of intake #1 is 0.15 bar (11.4%) higher than intake #7. Also, the average Pi2 during valve overlap is higher for intake #1. Both of these phenomena likely contribute to the 11% reduction in v for intake #7. The frequency spectra


0 90 180 270 360 450 540 630 7200.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #6

IVCEVCIVO(a)

0 1 2 3 4 5 6 7 8 9 10110

120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #6

(b)c/4L (baseline)


0 90 180 270 360 450 540 630 720

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #6

IVCEVCIVO (a)

0 1 2 3 4 5 6 7 8 9 10120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #6

c/4L (baseline)(b)


0 90 180 270 360 450 540 630 7200.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #7

IVCEVCIVO (a)

0 1 2 3 4 5 6 7 8 9 10110

120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #7

(b)c/4L (baseline)



12

show that the peak SPL has shifted from order 2 to 2.5 and reduced by 4.6 dB for intake #7. Figure 14compares the same intakes at the main tuning peak speed of 4250 RPM for intake #7. The peak magnitude of the compression wave near IVC is lower for intake #7 than for intake #1; however, the compression wave for the baseline arrives at i2 later and right at IVC, thus reducing its effect on v and allowing higher v for intake #7 at this speed. Compared to the baseline, the SPL for intake #7 is 15.5 dB higher at order 2.5 and slightly lower from orders 0.5 to 2.

Figure 15 shows Pi2 of the largest tapered intake (Dt = 7.26 cm), intake #8, compared to intake #1 at the main tuning peak of intake #1, 3750 RPM. A reduction of 0.14 bar (10.6%) can be seen in the peak pressure for intake #8, which decreases its effectiveness on intake tuning at this speed, contributing to a 9% reduction in vcompared to baseline. As with the other tapered cases, the dominant frequency of the QSW has increased for intake #8, from about 136 Hz for the baseline to about 161 Hz, an increase of 18.4% over the baseline and 7.3% over intake #6. When compared to intake #1, the frequency spectrum for intake #8 shows a decrease in SPL from orders 0.5 to 2 followed by an increase from orders 2.5 to 4, with the peak SPL moving from order 2 to 2.5. Data is not acquired for the baseline case for the exact v peak speed for intake #8 of 4375 RPM, thus Pi2is presented at 4250 RPM instead in Fig. 16. The peak pressure of intake #8 is 0.132 bar (9.9%) lower than that of the baseline, but it occurs 25° earlier, which is early enough to give the cylinder a supercharging effect before IVC, aiding in the 8% increase in v for intake #8. The SPL from orders 0.5 to 2, have decreased for intake #8, while that of order 2.5 has increased by 16.0 dB compared to intake #1.

BELLMOUTHS – First, the measured v and corrected brake power of the engine for each bellmouth (intakes #9-12) are compared to the baseline. Figure 17 showsintakes #9 and 10 (Ri/D = 0.05 and 0.20, respectively) and Fig. 18 shows intakes #11 and 12 (Ri/D = 0.50 and 1.0, respectively). Table 8 gives the location and magnitude of peak v for each bellmouth.

Intake #

Peak vlocation(RPM)

Peak vmagnitude

(%)

% difference from baseline

1 3750 111.5 --9 3850 107.9 3.2310 3750 111.9 0.3611 3750 110.7 0.7212 3650 111.5 0

Table 8. v Characteristics of the Bellmouth Group.

For intake #9, v at each tuning peak has decreased, with the magnitudes at 3000, 3750, and 4750 RPM, lowered by 3.1%, 3.8%, and 3.5%, respectively. Ri/D = 0.05 results in an inlet loss coefficient of about 0.2 (Miller, 1990), which is measurably detrimental to the overall engine performance. Compared to the baseline, the peak power for intake #9 has reduced slightly, and an increase in power at 5250 to 5500 RPM is noticeable corresponding to the increase in v at those speeds. The v and brake power of intakes #1 and 10 (Ri/D = 0.20) are nearly identical. For an Ri/D > 0.15, the inlet loss is almost negligible (Miller, 1990), which appears to be the case for the bellmouth group, as the vmagnitudes are similar for intakes #10 - 12.

It appears that v of intake #11 has shifted toward lower speeds slightly, increasing at 3650 RPM and decreasing at 3850, 4000, and 4750-5500 RPM. Selamet et al. (2001) have shown that for a duct with flanged bellmouth, as Ri/D increases, the length of the end correction, a hypothetical length of pipe attached to the inlet of the duct to account for inertial effects of the air, increases for frequencies low enough to affect engine tuning. This gives a longer effective length for intake #11 compared to intake #1 despite both intakes having equal distance to the bellmouth, causing the shift in v toward lower speeds for the former. The slight shift in v between the two intakes does not have much influence on the brake power.

For intake #12, since the bellmouth is larger, the end correction is larger than that for the baseline and intake #11; therefore the overall length of the intake is longer, and v continues to shift toward lower speeds for intake #12 compared to the baseline or intake #11, with tuning peaks moving from 3750 and 4750 RPM of the baseline to 3650 and 4650 RPM for intake #12. The power, following the trend of v, shifts toward lower engine speeds for intake #12, while the peak power magnitude remains similar.

Figure 19 shows the Pi2 for intakes # 1 and 9 at 3750 RPM. The peak pressure near IVC of intake #9 is 0.063 bar (4.8%) lower than that of the baseline, and the pressures have also decreased during valve overlap. These reductions in pressure contribute to a drop in v of 4% for intake #9. The overall behavior of the frequency spectra is similar for both intakes, with similar or lower SPL for intake #9 for most orders.

Figure 20 shows Pi2 for intakes #1 and 10 at 3750 RPM. Since v for these two intakes is nearly identical, it is expected that the intake pressures should also be nearly identical. This proves to be the case, as only the subtlest of differences can be seen in both the time and frequency domain. The pressure shift seen in the time domain may be attributed partly to the change in barometric pressure between experiments, with that of intake #1 being 0.013 bar higher.


0 90 180 270 360 450 540 630 7200.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #7

IVCEVCIVO(a)

0 1 2 3 4 5 6 7 8 9 10120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #7

(b)c/4L (baseline)


0 90 180 270 360 450 540 630 7200.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #8

IVCEVCIVO(a)

0 1 2 3 4 5 6 7 8 9 10100

110

120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #8

(b)c/4L (baseline)


0 90 180 270 360 450 540 630 7200.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #8

IVCEVCIVO(a)

0 1 2 3 4 5 6 7 8 9 10120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #8

c/4L (baseline)(b)



1000 1500 2000 2500 3000 3500 4000 4500 5000 550075

80

85

90

95

100

105

110

115

Vol

umet

ric E

ffici

ency

(%)

Engine Speed (RPM)


(a)

1000 1500 2000 2500 3000 3500 4000 4500 5000 55003

6

9

12

15

18

21

Cor

rect

ed B

rake

Pow

er (k

W)

Engine Speed (RPM)


(b)

Figure 17. Experimental (a) v and (b) Brake Power for Intakes #9 and 10 compared to the Baseline.

1000 1500 2000 2500 3000 3500 4000 4500 5000 550075

80

85

90

95

100

105

110

115

Vol

umet

ric E

ffici

ency

(%)

Engine Speed (RPM)


(a)

1000 1500 2000 2500 3000 3500 4000 4500 5000 55003

6

9

12

15

18

21

Cor

rect

ed B

rake

Pow

er (k

W)

Engine Speed (RPM)


(b)

Figure 18. Experimental (a) v and (b) Brake Power for Intakes #11 and 12 compared to the Baseline.

0 90 180 270 360 450 540 630 7200.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #9

IVCEVCIVO(a)

0 1 2 3 4 5 6 7 8 9 10110

120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #9

(b)c/4L (baseline)



0 90 180 270 360 450 540 630 7200.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #10

IVCEVCIVO(a)

0 1 2 3 4 5 6 7 8 9 10120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #10

c/4L (baseline)(b)


Pi2 are compared for intakes #1 and 11 in Fig. 21 for the baseline tuning peak speed of 3750 RPM. The peaks and valleys of the wave for intake #11 trail those of the baseline by 2 to 4 CAD at this speed, indicating that the effective length of this intake is slightly longer; the peak magnitudes are similar. At this speed,

v for both intakes are also similar. The frequency spectra for intakes #1 and 11 are almost identical for the dominant engine orders (from 0.5 to 2.5), while the overall behavior of SPL for higher orders are similar.

Figure 22 compares Pi2 for the largest bellmouth (Ri/D = 1.0), intake #12, to the baseline for its v peak speed of 3750 RPM. The dominant frequency of the QSW is noticeably lower for intake #12 than it is for the baseline, by approximately 5 Hz, indicating that the effective length of intake #12 is longer than that of the baseline, despite both intakes having the same length to the bellmouth. The phase lag from intake #1 to intake #12 is also considerably larger (5 to 15 CAD) than for intake #11. Both these aspects suggest that, generally, a longer end correction is needed as Ri/D increases; Mariucci (2006) explores end corrections determined from the present study further. The peak pressure is higher for intake #12 than for intake #1, but it occurs 7 CAD later for intake #12 (and only 10 CAD before IVC), which appears to reduce its effect on intake tuning, as vfor intake #12 is 5% lower than intake #1 at this speed. At this speed, the SPL of these two intakes are within 1 dB from orders 0.5 to 2.5.

CONCLUSIONS

The volumetric efficiency of the tapers exhibits increasing tuning peak speeds with decreasing /Aeffective.The lumped Helmholtz approximation of the intake system accurately predicts this trend. As the taper area ratio is increased above 1.5, the peak v begins to deteriorate. The shape of the power curves follow a

similar trend to the v, with intake #8 making the highest power of all tapers. The compression wave returning to i2 near IVC is known to cause intake tuning. However, the location of its peak must be taken into account as the wave takes about 17° to travel from i2 to the back of the valves at speeds typical of peak v for this group. If the peak of the compression wave for the baseline was closer to IVC than that of any taper and within 17° of IVC, then a significant decrease in v was observed for the baseline even if the taper had a lower peak pressure magnitude near IVC. For example, intake #4 had a lower peak pressure near IVC at 4000 RPM, but it occurred 18° earlier than that of the baseline, which contributed to an 11% increase in v for the taper. If the compression wave peaks of the baseline and any taper were both an ample time ahead of IVC, then the difference in peak pressures was of the same order of magnitude as the difference in v. For example, the pressures for intakes #6 and 8 were 13.8% and 10.6% lower than the baseline at 3750 RPM and v were 11.0% and 9% lower. For the tapered intakes, the dominant frequency of the QSW at i2 increased compared to the baseline, which is an effect similar to shortening the length of a closed-end quarter-wave silencer.

The volumetric efficiency of the bellmouth group showed slight movement of tuning peaks to lower speeds as Ri/D increased. The lumped Helmholtz approach suggests that increased effective length causes this. The trend of increased length with Ri/D is echoed in the pressures at i2; the phase shifts from the baseline indicates if the end correction for any particular intake configuration in this group increases or decreases with respect to that of the baseline. If the pressures led the baseline, a smaller end correction was needed, and if the pressures trailed the baseline, a larger. A reduction in v was observed at the tuning peaks for intake #9 (Ri/D = 0.05) due to presumably inlet flow losses. When Ri/D > 0.20, no losses were evident, as the tuning peaks v of intakes #10 – 12 were similar.


0 90 180 270 360 450 540 630 720

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #11

IVCEVCIVO(a)

0 1 2 3 4 5 6 7 8 9 10110

120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #11

c/4L (baseline) (b)


0 90 180 270 360 450 540 630 720

0.7

0.8

0.9

1

1.1

1.2

1.3

1.4

CAD

Pre

ssur

e (b

ar)

intake #1intake #12 IVCEVCIVO

(a)

0 1 2 3 4 5 6 7 8 9 10110

120

130

140

150

160

170

180

Engine Order

SP

L (d

B)

intake #1intake #12

(b)c/4L (baseline)


Brake power trends generally followed v. The pressure peak trend at i2 before IVC was generally indicative of the v behavior, although the timing of the peak was an important factor for intake #12.

Extensive engine simulation work has also been performed using the experimental data presented in this paper. For more information, the interested reader is referred to Mariucci (2006).

ACKNOWLEDGMENTS

The authors would like to acknowledge Ford Motor Company for their generous contribution of the single cylinder research engine and test pieces, specifically Dr. Kevin Tallio, Zafar Shaikh, Mike Magnan, Graham Hoare, and Frank Fsadni.

REFERENCES

Downing, E. W., 1957-58, “Petrol Injection: Some Further Developments,” Institution of Mechanical Engineers: Proceedings of the Automobile Division, 6: 161-173.

Engelman, H. W., 1973, “Design of a Tuned Intake Manifold,” ASME Paper 73-WA/DGP-2.

Heywood, J. B., 1988, Internal Combustion Engine Fundamentals. McGraw-Hill, Inc., New York.

Howard, T. M., 2003, “Tapered Intakes on a Single Cylinder Engine,” MS Thesis, The Ohio State University.

Kastner, L. J., 1945, “Induction Ramming Effects in Single-Cylinder Four-Stroke Engines,” Proceedings of the Institution of Mechanical Engineers, 153: 206-220.


Mariucci, V. E., 2006, “An Experimental and Computational Investigation of the Effect of Primary Intake Runner Geometry on the Performance of a Single Cylinder Engine,” MS Thesis, The Ohio State University.

Matthews, R. and Gardiner, A. W., 1924, “Increasing the Compression Pressure in an Engine by Using a Long Intake Pipe,” NACA Technical Memorandum 180.

Miller, D. S., 1990, Internal Flow Systems, 2nd Edition. BHRA Information Services, Cranfield.

Morse, P. M., Boden, R. H., Schecter, H., 1938,“Acoustic Vibrations and Internal Combustion Engine Performance I: Standing Waves in the Intake Pipe System,” Journal of Applied Physics, 9: 16-23.

Selamet, A., Ji, Z. L. and Kach, R. A., 2001, “Wave Reflections from Duct Terminations,” Journal of the Acoustical Society of America, 109: 1304-1311.

Thompson, M. P., 1968, “Non-Mechanical Supercharging of a Four-Stroke Diesel Engine,” MS Thesis, The Ohio State University.


Documents

Effect of Primary Intake Runner Tapers and Bellmouths on ...engine.osu.edu/ASJP/JP/J72.pdf · with a one-cylinder diesel engine, found that the engine would not start unless the compression