Winterbone - 1990.pdf

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

The Engineering Society For Advancing Mobility and sea Air and Space,

INTERNATIONAL^

400 COMMONWEALTH DRIVE, WARRENDALE, PA 15096-0001 USA.

The Accuracy of Calculating Wave Action in Engine Intake Manifolds D. E. Winterbone Depart. of Mechanical Engineering

UMIST, Manchester, England

M. Yoshitomi Komatsu Limited Oyarna-shi, Japan

MAR 7 1990

SAE

fJeRAaW International Congress and Exposition Detroit, Michigan February 26 - March 2,1990

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

The appearance of the ISSN code at the bottom of this page indicates SAE's consent that copies of the paper may be made for personal or internal use of specific clients. This consent is given on the condition, however, that the copier pay a $5.00 per article copy fee through the Copyright Clearance Center, Inc., Operations Center, 27 Congress St., Salem, MA 01970 for copying beyondthat permitted by Sections 107or 108of the U.S. Copyright Law. Thisconsent does not extend to other kinds of copying such as copying for general distribution, for advertising or promotional purposes, for creating new collective works, or for resale. SAE routinely stocks printed papers for a period of three years following date of publication. Direct your orders to SAE Customer Service Department. To obtain quantity reprint rates, permission to reprint a technical paper or permission to use copyrighted SAE publications in other works, contact the SAE Publications Group.

---- -----.- -. ---- -.-----. --. --. - -- - .---.--= , --.---- - - - - - . -.. ------- - - -. -. --------- -------.-.

ABSAEp~psrsare abstracted and rndexed m the SAE G W MoinlityDatabase.

SAE GLOBAL MOBlLrrY DATABASE

No part of this publication may be reproduced in any form, in an electronic retrieval system or otherwise, without the prior written permission of the publisher. +

'

,

*

ISSN 0148-7191 Copyright 1990 Society of Automotive Engineers, Inc. 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. A process is available by which discussions will be printed with the paper if it is published in SAE Transactions. For permission to publish this paper in full or in part, contact the SAE Publications Division.

Persons wishing to submit papers to be considered for presentation or publication through SAE should send the manuscript or a 300 word abstract of a proposed manuscriptto: Secretary, Engineering Activity Board, SAE. Printed in USA

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

The Accuracy of Calculating Wave Action in Engine Intake Manifolds D. E. Winterbone Depart. of Mechanical Engineering UMIST, Manchester, England

M. Yoshitomi Komatsu Limited Oyama-shi, Japan

This paper describes a comparison between calculated and measured pressure traces and air mass flows through a family of inlet manifold geometries. It is shown that a non-linear wave action calculation technique, based on the method of characteristics, can accurately predict the detailed variation of pressure in the manifold over a broad range of engine speed: it can also accurately predict the mass flow. It is shown that it is necessary to include end effects for the various pipes in order to obtain realistic predictions. The mass flow can be predicted to better than 2% over the majority of the engine operating speed, although the accuracy decreases slightly at the tuning speeds. This reduction in accuracy is probably due to the increased losses resulting from the higher velocities and flow reversals occurring at the tuned speeds. It is also shown that tuning of the system occurs when the natural frequency of the manifold causes an increase in cylinder pressure when the inlet valve closes. THE TUNING OF INTAKE MANIFOLDS has been of interest to engine designers for many years: the first reference to such tuning is claimed to be in 1904. room el describes a number of equations for designing intake systems, based on both wave and inertia effects of the gases in the manifold. These equations provided a useful yardstick for design at that time, but some deficiencies were readily seen. These included the number of empirical parameters required to tune the equations to the problem. Other investi ators including ~ n ~ e l m a n ~ h, o m ~ s o n vorum4, ~, Cse , etc., have proposed similar equations with varying degrees of success: these approaches have been based on evaluating the tuned frequencies of the manifolds. In a companion paper Pearson and winterbone6 describe a linear method for evaluating the variation of volumetric efficiency of engines over the whole speed range. Their approach is based on solving the linearized wave equations using a rapid step-by-step method. It has the advantage over analytical solutions of the wave equations of readily allowing for interference between intake strokes of various cylinders: this interference can be extremely important in certain circumstances,

s

particularly with six-cylinder engines. The linearized solution is a very rapid one taking about 1 minute on a main-frame computer to evaluate the volumetric efficiency-engine speed characteristic for an engine. However, such a method embodies a number of important approximations and is not capable of predicting accurately the ubsoZute levels of volumetric efficiency, but it is excellent at predicting trends. A method for predicting absolute values of volumetric efficiency must include the non-linear equations of wave action, and make allowance for heat transfer, friction, compressible flow through the intake valve, and any interference with exhaust manifold effects. ~ e n s o n 7 describes such an approach based on solving the non-homentropic equations of compressible flow using the mesh method of characteristics. Many others have applied similar techniques to engine calculations and a summary of this work is included in ~interbone8.9. The data preparation and results processing from such methods are substantially more complex than for the program described by Pearson and winterboned and it recommended that the non-linear method be used to undertake detailed analysis of the system only after most of the parameters have been defined within fairly close limits. This paper describes a series of rig tests which were performed to check the accuracy of the non-linear computer program in calculating the variation of both instantaneous pressure at a number of points in the manifold, and the mean mass flow rate over a cycle. Winterbone, Worth and ~ i c h o l s l o described a similar series of tests performed on a running engine, but they were not able to record accurately the variation of mass flow through the engine. They showed that the prediction of inlet manifold pressure over the whole operating range was good, although some deficiencies were apparent, but the evaluation of exhaust manifold pressure caused more difficulty. No comparisons were given of mass flows because of difficulties in measuring this parameter without interfering with the actual engine performance: attempts to calculate back from air-fuel ratio and fuel flow rate proved to be too inaccurate to engender much faith in the values. As a result of the project reported in ref 10 it was decided to isolate the intake system from the remainder of the engine and to analyse this separately. This was

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

achieved by removing the manifold, cylinder head and camshaft assembly and driving the camshaft to simulate cylinder operation. The intake manifold was supplied with air at 1.5 bar via a large receiver and the steady, mean airflow rate was measured upstream of the receiver by an orifice plate. Detailed pressure variations in the manifold were measured by pressure transducers. The tests were undertaken both with open cylinders and when pistons were fitted although only the latter will be discussed here. This arrangement enabled the intake system to be evaluated in isolation from the exhaust system and the piston motion. The whole rig was simulated using a wave action simulation program, TSPARK, and the results compared with the measured values.

PRIMARY P I E S

&EN,%!

h

\

/

INJECTOR W I N O

M R I m REtElMR TANK 1.5 bar

DUMMY PISTONS

9-J

Fig 2

Diagram of manifold showing location of transducers

EXPEXIMENTAL P R O G R A M M E EXPERMFATAL RIG Figure 1 shows a schematic of the whole apparatus. Air was supplied by a large capacity compressor and the receiver pressure was maintained at 1.5 bar. The receiver tank (volume = 1.5 m 3 ) was large enough to damp out pressure fluctuations initiated by the intermittent flow through the inlet valves: the air mass flow could be measured accurately using an orifice plate (to BS 1042) mounted upstream of the receiver tank. Figure 2 shows the structure of the intake manifold together with the location of the pressure transducers. A typical "rake" arrangement was used for the manifold which was constructed in a modular manner. The manifold comprised primary and secondary pipes and a plenum chamber: the dimensions of these components are defined in Table 1. Different sized manifolds could be constructed by selecting various modules and connecting them by threaded joints. Table 2 gives the salient details of the Austin Rover 1.6 litre 'S' series engine which was used for the test. The cylinder head, block and camshaft were mounted on a base plate and the camshaft was driven by a variable speed electric motor arrangement. The crankshaft was not used and the pistons were replaced by dummy ones at fixed positions in the cylinders. This enabled the effects in the manifolds to be assessed without the additional complexity of variable cylinder volume influencing the results.

Table 1 Dimensions of Modular Manifold Primary pipe dimensions

Short Med i um Long Combined Port length

+

Length L l (mm) 76.2 152.4 304.8 534.0 adaptor lengths

Secondary pipe dimensions

Short Long

Length L2(mm) 415 914

engine

Fig 1 Schematic diagram of airflow rig

Diameter D2(mm) 48 48

Plenum chamber dimensions Small Med i um Large

Volume V(cm3 ) 773.5 1443.0 2882.7

S u r f a c e Area F(cmZ) 651.4 925.1 1363.7

Table 2 Engine dimensions Cylinder dimensions: bore x stroke No. o f c y l i n d e r s Engine c a p a c i t y Compression r a t i o i vo Valve t i m i n g i vc evo evc

c i n e tuning valve

Diameter D l (mm> 30 30 30 30 = 165 mm.

76.2mm x 87.6mm 4 1.598 l i t r e 9.6:l 27" b t d c 69" abdc 123" a t d c 19" a t d c .

THE MODULAR MANIFOLD - The experimental tests were divided into two areas: single-cylinder and four-cylinder tests. In the single-cylinder engine tests the primary pipes listed in Table 1 were used. In the four-cylinder tests the manifolds were constructed from the primary, and secondary pipes and plenum chambers. A total of twenty-four different manifolds were constructed from these components.

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

The pressure was INSTRUMF,NTATION - presswe measured at the points indicated in Figure 2 using Kistler type 4075A10 piezo-resistive absolute pressure transducers. These were used in a Wheatstone Bridge with a Fylde type 402 bridge driver; the signals were amplified by a Fylde type 254GA amplifier before data collection by computer. The pressure transducers were placed as close as possible to the surface of the measuring region to minimize passage effects. The pressure transducer in the cylinder was located in the spark plug hole. The pressure traces were averaged over 8 cycles to reduce the effects of noise on the signals. meramre Temperatures were measured using Nickel-Chromiurn/Nickel-Aluminium thermocouples. These were located 3.7m upsteam of the receiver tank (to measure the temperature of the air at inlet to the system) and in the test cell ambient region. Air Mass Flow Rm - The air mass flow rate was measured using an orifice meter situated upstream of a damping vessel. The plate was designed to BS.1042 and the differential pressure was measured using a manometer. Data Acauisition S v s t a The data acquisition system was based on a Z80A microcomputer system fitted with a 20 Mbyte Winchester disk. The data could also be stored on 5.25 inch floppy disks. Data was collected from all pressure points at one degree intervals.

TEST PROCEDURE - The experimental tests were split into two distinct sections: single-cylinder and four-cylinder tests. The four-cylinder tests were further sub-divided into those in which the intake valves to all four cylinders were activated and a further group in which only the intake valve to cylinder no.1 was activated. The last set of tests enabled the method proposed by Matsumoto and ohatall to be simulated and examined. In this arrangement there can be no interference between the cylinders. The tests in which pressures were measured were conducted at wide open throttle over a speed range from 1000 to 4500 revlmin (engine speed) in 500 revlmin steps. The overall mass flow rate was measured when all four cylinders were activated over a speed range from 1000 to 5750 revlmin in 250 revlmin steps.

SIMULATION OF ENGINE RIG The group at UMIST has broad experience of simulating flows in intake and exhaust manifolds of engines, together with the flows into and out of the cylinder. The calculation methods, based on solving the flow equations for compressible flow, have been described in detail in ~ e n s o n 7 and Horlock and winterbone12. In this project a computer program named TSPARK, normally used for simulating turbocharged spark-ignited engines, was used to model the rig. This was achieved by data manipulation rather than program modification. The fixed piston arrangement was defined by: (i) (ii) (iii)

setting the stroke to 0.1 mm; making the clearance volume equal to the fixed volume of the cylinder; setting the compression ratio equal to 1.01.

EVO

. .

I V O EVC ,

r vc

Fig 3 Pressure waves in single cylinder engine without including end-effect in calculation. Engine speed = 2500 revlmin; pipe length = 241.2 mm (a) intake port pressure (b) cylinder presure. TSPARK is a non-homentropic computer program based on the method of characteristics. It contains a two-zone combustion model of spark-ignited combustion, but this was not used in this project. The inlet manifold junctions were simulated using the simple constant- (or equal-) pressure junction model. Models of pressure losses at junctions are included in the program but were not used in this project.

RESULTS The results will be considered in separate sections relating to the different manifold arrangements. First, the results from the single-cylinder arrangement will be discussed, followed by those from the four-cylinder layout with both single and multiple valve operation. PRESSURE VARIATION - Sinde cvlinder operation This is the simplest inlet manifold arrangement because there is no interference between pulses from other cylinders: each cylinder has its own isolated inlet pipe. This arrangement is often used on high performance engines because it enables the maximum volumetric efficiency to be achieved at high engine speeds, due to the avoidance of interference and junction losses. It should be remembered that the "engine" has a fixed cylindcr volume, and this does affect the results obtained. Figure 3 shows the measured and predicted wave action, both in the inlet tract (fig 3a) and the cylinder (fig 3b). There are a number of important features observable in these diagrams:

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

(i) the frequency of the pressure wave in the inlet manifold depends on whether the inlet valve is open or closed; (ii) the frequencies of the measured and predicted pressure waves are different, particularly when the valve is closed; (iii) the predicted pressure waves decay more rapidly than the measured ones. These features allow certain deductions to be made, both about the physical nature of the wave action in the system and about the calculation technique. The first feature, item (i), shows that the mode of vibration of the system varies during the cycle and depends on whether the intake valve is open or closed. When the inlet valve is closed the pipe resonates as a quarter wave organ pipe and the frequency is given by

F3 t a n F1

:[

tan[:

L?] = 1

(4)

If both wLl/a and wL3Ia are much less than unity, then eqn 4 reduces to eqn 2, where the volume term is the product F3L3 Equation 4 gives much more accurate prediction of the frequency when the valve is open, as shown in Table 3. Table 3 Pressure WJVC Frcqucn~icsduring U s priciJ wlicn inlcl v;llvc o p n (A)

Cyl i n d c r I c n g l l i L3=0.0354m Tol;tl printary p i p e I c n g l h i s L ' ] = L ] m. 71ic p o r l Icnglli + a d a p t o r lctigllis = 165 mm. (U)

,

+

165 m

-r---

(rpmj

a = speed of sound in air in manifold, (mls); L = length of the primary pipe, (m); f = frequency (Hz). However, when the valve is open the system resonates as a Helmholtz resonator. This mode of vibration occurs when the whole column of air in the inlet pipe moves in phase against the "stiffness" of the air in the cylinder: this is analogous to a spring-mass system in which the spring is the compressibility of the air in the cylinder and the mass is the mass of the column of air. It can be shown (see ref 13) that the natural frequency of a Helmholtz resonator is given by

Frcqucncy (IIz j

Tcmp. ("Cj

Frcqucncy (11~)

Tcmp.

Frcqucncy ( I

Tcmp. (OC)

where

Eslimalcd 229.3 1rotn c q n . 2 . End ~ r r c c t

where F = area of cross-section of the primary pipe (m2); V = volume of resonator chamber (m3). ~ngeltnan3suggested that the effective volume of the resonator chamber, V, to be used in eqn 2 should be the mean cylinder volume during the induction stroke. Equation 2 then becomes

where r = nominal compression ratio. Such a resonance will result in the engine tuning at certain speeds, when a pressure peak occurs at the valve slightly before the intake valve closes. This point will be returned to later. Application of eqn 2 to the evaluation of the natural frequency of the waves during the period when the valve was open, with the effective volume in this case set equal to the actual fixed volume of the cylinder, resulted in calculated natural frequencies 15 to 25% higher than the measured values. pearson14 suggested that a better model for the system when the valve is open is to simulate the cylinder as a pipe with a closed end (the piston) connected to the primary pipe: the volume term in the Helmholtz resonator equation is then replaced by a length and area term. This results in

2 9 8.8rntn

25.4

201.9 6.3m

167.0

26.4

9 .O I I I ~

Moving now to item (ii) in the list, namely the difference between the measured and predicted frequencies of the pressure waves. Figure 3 was calculated using the geometric dimensions of the various components. It is well known that the effective length of pipe in which there is wave action is longer than the geometric length due to the "end effect", caused by the inertia of the gas entering and leaving any end open to a plenum. Thompson et a13 suggest using an additional length of 7cd,/4, but other authors suggest different values. In this case it was decided to evaluate the end effect by using TSPARK to estimate how much length must be added to the pipe to obtain the measured frequency. First, the pressure trace was divided into the open and closed period and each was treated separately. The end effect calculated for the mode when the valve was closed was almost constant for all the pipe lengths over the whole range of speed - the average value was 36.2mm, approximately 1.2 times the pipe diameter. The end effect during the valve open mode varied much more across the range of lengths and speeds considered but the average value was only 12.lmm; ie substantially less than in the previous case. It is suggested that the mean inflow velocity during the valve open period reduces effectiveness of the flow leaving the open end of the pipe. It is recommended that an end effect be added to the length of the pipe when performing such calculations; this is particularly important when attempting to evaluate the natural frequencies of manifolds with short pipes. The end effect used in the wave action calculation was a

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

composite value obtained by matching the pressure traces at one condition and an "average" value of 15mm (0.5 D) was used. The final feature of these results is the "damping" of the predicted pressure waves. It might be thought that the friction factor used in the momentum equation is too high, but this is unlikely to be the case here. The more likely reason is the inherent numerical damping introduced by the linear interpolation process in the mesh method of characteristics. Such damping is more prominent in simple systems because the pressure waves are more pronounced. It is possible that calculation techniques based on finite difference techniques (see refs 15,16) might reduce the amount of damping introduced, although if the boundary conditions are calculated using the method of characteristics then it likely that the improvement will be negligible, or very small. Figure 4 shows the measured and simulated results for a single cylinder engine with the medium length intake pipe, running at 3000 revlmin. It can be seen that in these results, which include the end effect, the prediction is extremely accurate: the only features missing are some of the minor waves in the period when the valves are open. These results are typical of those obtained over the whole range of manifold pipes and engine speeds. A number of features of these results, which are most apparent in the cylinder pressure values, are the differences between pressures at various points in the cycle. It should be recognized that in this simulation the power stroke of the engine was not calculated (because it was not relevant) and the pressure at EVO was simply stipulated. This results in the calculated pressure trace dropping instantaneously from about 1.63 bar at IVC to atmospheric pressure. The measured trace shows what actually happens in the rig: first, around IVC, the efrect of leakage becomes important and the pressure deviates from the calculated value; the pressure continutes to drop through leakage until EVO. Second, at EVO the trapped charge passes through the exhaust valve and again reduces the pressure whilst establishing some pressure waves in the short exhaust ports. In later diagrams the pressure at EVO was set to the measured value and better agreement of the cylinder pressure between EVO and IVO is achieved. Hence, the predicted cylinder pressure only mirrors that of the experiment during the period when a valve is open, and comparison should only be made in this region. It can then be seen that the prediction is in good agreement with the measurements. The pressure variation in the cylinder displays the same frequency as the pressure in the pipe, indicating that the system is acting like a Helmholtz resonator.

FOUR CYLINDER MANIFOLDS - Many of the features of single-cylinder manifolds cany over into the four-cylinder manifold results. The results for these cases come as a set of four diagrams: (a) the pressure near intake port no. 1; (b) the pressure near intake port no. 3; (c) the pressure in the plenum; (d) the pressure in the cylinder.

EVO

l.e+

1 V O EVC

IVC

Fig 4 Pressure waves in single cylinder engine with end-effect included in calculation. Engine speed = 3000 revlmin; pipe length = 317.4 mm (a) intake port pressure (b) cylinder pressure.

. . Sin~lecvlinder achvahoq - In this case the camshaft only actuated the valves of cylinder 1. This situation was considered because one of the seminal papers in this area, by Matsurnoto and ohatall, considered the evaluation of manifold characteristics based on an analytical method in which only one valve was actuated: this avoided possible effects of interference between cylinders. Figure 5 shows the pressure variations at the various stations in the system for the engine with a manifold constructed from medium primary pipes (L1=305.4mm), long secondary pipe (L2=992.0mm and a large plenum chamber (V=2882.7cm ) with an engine speed of 1500 revlmin. It can be seen that the prediction of the pressure waves is good, but the waves are "overdamped" during the period when the valves are closed. It is apparent that the manifold only receives excitations from one cylinder and the wave action in between perturbation is small. The pressure variation at the valves of other cylinders (see fig 5b) is extremely small, and unrepresentative of that in the active pipe. However, the pressure variation close to cylinder no 1 , during the period the intake valve is open, is similar to that which occurs during normal operation. Hence, it is likely that single valve operation gives a reasonable estimate of the variation of volumetric efficiency for a manifold in which there is only a small amount of interference. This would suggest that the method proposed by Matsumoto and 0hatal1 is satisfactory for certain manifolds.

d

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

Four cvlinder o~eratipa Figure 6 shows the pressure variation in a manifold similar to that in the previous section (medium primary pipes, long secondary pipe, large plenum chamber), again operating at 1500 revlmin. Figures 5 and 6 are directly comparable and show the difference between single and four cylinder operation. The most obvious difference is the active period in fig 6b, due to the activation of that cylinder. Otherwise the pressure diagrams are extremely similar to those in fig 5, again supporting the approach of Matsumoto and Ohata. Comparing fig 7 and fig 4, both for manifolds incorporating medium length primary pipes with an engine speed of 3000 revlmin, it is qnite apparent that the residual wave action is greater for a single cylinder engine than one fitted with a multi-cylinder manifold. This results in the pressure being different in the valve open period, with a lower pressure from the multi-cylinder manifold at the time of valve closure in this case: this would give a lower trapped mass in the cylinder, and hence lower volumetric efficiency. Examination of the diagrams over the whole speed range shows that the waves occurring from organ pipe tuning are never as large with a multi-cylinder manifold as when individual intake pipes are fitted - presumably explaining why high performance engines are always arranged in the latter manner. However, when the manifold acts as a Helmholtz resonator it is possible for the volumetric efficiency to be enhanced to a level above the organ pipe value - and specifically at a lower natural frequency, or engine speed. Figure 8 shows how the pressures at each transducer come into phase during this resonance mode, with the pressure variation in the plenum chamber becoming almost equal to that in the pipes.

EVO

I V O EVC

IVC

Fig.5 Pressure waves in four-cylinder engine manifold with only one cylinder activated. Engine speed = 1500 revlmin. Primary pipe length = 305.4 mm. pipe length = 992.0 mm. Plenum volume = 2882.7 cm (a) intake port of cylinder 1 (b) intake port of cylinder 3. (c) plenum chamber (d) cylinder.

SecOndar

MASS FLOW VARIATION - While the pressure variation in the manifold is of interest it is more important to assess its effect on the air mass flow through the engine, which is often referred to in terms of volumetric efficiency. If the volumetric efficiency of an engine remained constant then the air flow through the engine would be linearly related to engine speed. In the case of a petrol engine the volumetric efficiency is deliberately reduced by throttling to provide load control, but even at full throttle the volumetric efficiency of the l ~ the engine varies with engine speed. ~ a ~ l o rrelated variation of volumetric efficiency to the Mach Index of flow through the inlet valve, and Fukutani and watanabe18 modified this approach by applying a Mean Inlet Mach Number. Both of these parameters, based on an engine without an inlet manifold, showed that the volumetric efficiency of the engine would drop as the flow velocity through the inlet valve increased. If an inlet manifold is fitted then it is possible to modify this inherent characteristic: manifold tuning is applied to achieve the highest torque at the point where it is required. By using manifolds with more than one degree of freedom, that is more complex than a single pipe, it is possible to achieve more than one tuning point: variable geometry manifolds give even more scope for tailoring the engine torque curve. The results presented here show the manner in which the mass flow through the rig varied with engine speed. The upstream pressure was maintained at 1.5 bar to provide the driving force to pass the air through the engine, and the camshaft was actuated over the speed

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

range (the pistons were fixed as before). Figure 9 shows the variation of mass flow with engine speed for the engine fitted with a manifold comprising combined primary pipes (L1=534mm), short (L2=415mm), and large plenum chambersecondary (V=2882cm ). There are distinct peaks in the mass flow at 2000 and 4500 revlmin, with a marked trough at around 3000 rev/min. This is indicated by comparing the measured mass flow rate to that which would occur if the volumetric efficiency (qv) was 100%. The measured values would give a maximum q v = 130% and a minimum q v = 54%. The predicted results are generally in good agreement with the measurements, being within the limits of experimental error in 12 out of the 19 points. It is noticeable that the greatest difference between prediction and measurement is at the regions where the tuning is most pronounced; this suggests that the actual pressure losses increase above those simulated at these speeds, probably due to higher flow velocities and greater flow reversal at the pipe entries and exits. The diagrams for all the other manifold arrangements are substantially similar and the accuracy is comparable. An interesting feature of these results is the small dependence the tuning shows upon the plenum chamber volume (see fig 10): this measured result is at variance with that by Pearson and winterbone19 who predicted a significant variation in both the tuning speed and magnitude as the plenum chamber volume was changed. However, this effect (or lack of it) has also been noted by ~uelli20.

pY

EVO

I V O EVC

IVC

CONCLUSIONS The simplified engine rig enabled a large number of results to be obtained over a range of similar manifold configurations and a broad range of engine speeds. It also enabled results to be obtained without any question of interference from exhaust manifold effects. Two different modes of tuning occur in all inlet manifolds: organ pipe tuning and Helmholtz resonator effects. These occur during the periods where the inlet valves are closed and open respectively. Both these tuning modes are important but the Helmholtz one seems to have more effect on volumetric efficiency due to its effect on cylinder pressure at the point of intake valve closing. It is possible to calculate both the instantaneous pressure variation in the manifold and the time averaged mass flow with a high accuracy using a computer program based on the method of characteristics. It is necessary to allow for end effects on the pipes to obtain accurate predictions of the flow in the manifold. The losses appear to increase as the manifold tunes, and the program over-predicts mass flows (and hence volumetric efficiency) at these tuning points.

ACKNOWLEDGEMENTS The authors appreciate their discussions with Mr. David Worth and Dr. John Nichols, both of whom made valuable suggestions about this project. They also thank UMIST for providing facilities for undertaking the research, and Komatsu Ltd for supporting one of the authors.

Fig 6 Pressure waves in four-cylinder engine manifold with activation of all cylinders. Engine speed = 1500 revlmin. Manifold as fig.5; locations as fig.5.

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

EVO

i.?

-

..

I.,-. I.>..

- ,..-. K

K Y

a In

1.3..

L 1.2.. K I. I

..

In-.

,, ,

,, ,

,

,

--- ., ,,, ,

.

.

IVC

I V O EVC

--

.

..

:,

. ., ,

7

m4

EVO

I V O EVC

IVC

, ,,,

.,

,.,

,, ,,

,

,, , , ,,

,,, ,

,, ,,

, , ,,

,,

-.

.

, ,

Fig 7 Pressure waves in four-cylinder engine manifold with activation of all cylinders. Engine speed = 3000 revlmin. Manifold as fig.5; locations as fig.5.

Fig 8 Pressure waves in four-cylinder engine manifold with activation of all cylinders: shows total manifold resonance. Engine speed = 1000 revlmin. Manifold as fig.5; locations as fig.5.

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

4.VORUM. P.C. Short pipe manifold design for four-stroke engines. Part I. ASME paper 76-WA-DGP-4, 1975. VORUM, P.C. Short pipe manifold design for four-stroke engines. Part 11. ASME paper 80-DGP-6, 1980.

5. CSER, G. Some results of combined charging system. 1.Mech.E. Conference on Turbocharging and Turbochargers. Paper C64178, pp. 127-132. 1978.

1 7 L0 /

I

1

I

I

2000 4000 E NGI NE SPEED ( r.p.m)

L

I

600('

Fig 9 Mass flow variation with engine speed. Combined primary pipes (699 mm) short secondary pipe (415 mm), large plenum (2882 cm3) - measurement o--0 calculation -.calculated mass flow for assumed q, = 100.

6. PEARSON, R and WINTERBONE D E . A rapid wave action simulation technique for intake manifold design. To be presented at the SAE International Congress, Detroit, Michigan, February 1990. 7.BENSON. R.S. The thermodynamics and gas dynamics of internal combustionengines. Volume 1. Oxford University Press, 1982.

8. WINTERBONE, D.E. The theory of wave action applied to reciprocating engines. To be published in Internal Combustion Engines - Science & Technology. Elsevier Science Publishers Ltd. 9. WINTERBONE, D E . The application of wave action techniques to reciprocating engines. To be published in Internal Combustion Engines - Science & Technology. Elsevier Science Publishers Ltd.

10. WINTERBONE, DE., WORTH, DR., and NICHOLS, J.R. A comparison of synthesis and analysis models for wave action manifolds. Instn of Mech Engrs Int Conf on the Small Internal Combustion Engine. London, April 1989. 11. MATSUMOTO, I. and OHATA, A. Variable induction systems to improve volumetric efficiency at low andlor medium engine speeds. SAE Paper 860100, 1986.

ENGINE SPEED ( r.p.m.) Fig 10 Effect of plenum chamber volume on engine mass flow -.large lenum (2882.7 cm3) medium plenum (1443.0 cm3) --- small plenum (773.5 cm3).

-

REFERENCES

l.BROOME, D. Induction ram. Parts 1-3. Automobile Engineer, April pp. 130-133, May pp.180-184, June pp.262-267 1969.

2.ENGELMAN, H.W. Design of a tuned intake manifold. ASME Diesel and Gas Power Conference Proceedings, Houston, Texas, 1974. 3.THOMPSON, M.P, and ENGELMAN. H.W. The two types of resonance in intake tuning. ASME Paper 69-DGP-l l, 1969.

12. HORLOCK. J.H. and WINTERBONE, D E . The thermodynamics and gas dynamics of combustion engines. Vol.11 Oxford Press, 1986.

internal

13. LIGHTHLLL, MJ. Waves in Fluids. Cambridge University Press. 1980. 14. PEARSON, R.J. A linear model for the synthesis of intake manifolds MSc thesis, Univ of Manch 1988.

15. POLONI, M., WNIERBONE, D E . and NICHOLS, J.R. Flow with variable specific heats in the exhaust system of the internal combustion engine. SAE FISITA Congress Paper No.885094 September 25-30th 1988. 16. BLJLATY, T. and NIESSNER, H. Calculation of I-D unsteady flows in pipe systems of I.C. engines. ASME Winter Annual Meeting. New Orleans, Louisiana. 1984.

Downloaded from SAE International by University of Auckland, Sunday, September 23, 2018

17. TAYLOR C.F. The internal combustion engine in theory and practice Vol 1, MIT Press, Camb, MA 1977. 18. FUKUTANI, I. and WATANABE, E. An analysis of the volumetric efficiency characteristics of 4-stroke cycle engines using mean inlet Mach number Mim. SAE Paper 790484 1979.

19. PEARSON, R.J. and WINTERBONE, DE. A rapid synthesis technique for intake manifold design. To be published in The International Journal of Vehicle Design, 1989.

20. DUELLI, H. Berechnungen und versuche zur optimierung mehr-ztkubdernitireb ansaugsystemen fur einzelzylinder-einspritzung. Bericht aus dem Institute fur Verbrennungskraftmaschinen und Kraftfahlzeughau Reihe Technischen Universitat Wien. VerkehrstechnikFakneugtechnik, Nr.85.

von mit der 12: