SAE TECHNICAL PAPER SERIES

1999-01-0921

Numerical Simulation of a Two-Stroke Linear Engine-Alternator Combination

Christopher M. Atkinson, Sorin Petreanu, Nigel N. Clark, Richard J. Atkinson, Thomas I. McDaniel, Subhash Nandkumar and Parviz Famouri
West Virginia University

Reprinted From: Hybrid Vehicle Engines and Fuel Technology (SP-1422)

International Congress and Exposition Detroit, Michigan March 1-4, 1999

400 Commonwealth Drive, Warrendale, PA 15096-0001 U.S.A. Tel: (724) 776-4841 Fax: (724) 776-5760

The appearance of this ISSN code at the bottom of this page indicates SAEs 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 $7.00 per article copy fee through the Copyright Clearance Center, Inc. Operations Center, 222 Rosewood Drive, Danvers, MA 01923 for copying beyond that permitted by Sections 107 or 108 of the U.S. Copyright Law. This consent 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 Sales and Satisfaction Department. Quantity reprint rates can be obtained from the Customer Sales and Satisfaction Department. To request permission to reprint a technical paper or permission to use copyrighted SAE publications in other works, contact the SAE Publications Group.

All SAE papers, standards, and selected books are abstracted and indexed in the Global Mobility 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 1999 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 Group. 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 manuscript to: Secretary, Engineering Meetings Board, SAE.

Printed in USA

1999-01-0921

Numerical Simulation of a Two-Stroke Linear Engine-Alternator Combination

Christopher M. Atkinson, Sorin Petreanu, Nigel N. Clark, Richard J. Atkinson, Thomas I. McDaniel, Subhash Nandkumar and Parviz Famouri
West Virginia University
Copyright 1999 Society of Automotive Engineers, Inc.

ABSTRACT
Series hybrid electric vehicles (HEVs) require powerplants that can generate electrical energy without specifically requiring rotary input shaft motion. A small-bore working prototype of a two-stroke spark ignited linear engine-alternator combination has been designed, constructed and tested and has been found to produce as much as 316W of electrical energy. This engine consists of two opposed pistons (of 36 mm diameter) linked by a connecting rod with a permanent magnet alternator arranged on the reciprocating shaft. This paper presents the numerical modeling of the operation of the linear engine. The piston motion of the linear engine is not mechanically defined: it rather results from the balance of the in-cylinder pressures, inertia, friction, and the load applied to the shaft by the alternator, along with history effects from the previous cycle. The engine computational model combines dynamic and thermodynamic analyses. The dynamic analysis performed consists of an evaluation of the frictional forces and the load (in this case the alternator load) across the full operating cycle of the engine. The thermodynamic analysis consists of an evaluation of each process that characterizes the engine cycle, including scavenging, compression, combustion and expansion, based on the first law of thermodynamics. Since the modeled engine was crankshaftless, a time-based Wiebe function (as opposed to a conventional crank angle-based approach) was used to express the mass fraction burned for the combustion process, while the combustion model used was a single-zone model. To render the model useful, the parameters used were based on experimental data obtained from the working example, including instantaneous shaft position, velocity and in-cylinder pressure. Also, a parametric study was performed to predict the behavior of the engine over a wide operating range, given variations in fuel combustion properties, the reciprocating mass of the piston shaft assembly, frictional load and the externally applied electrical load.

INTRODUCTION AND LITERATURE REVIEW

Free-piston engines have been a subject of research and development for several decades. A recent review of freepiston engine concepts has been conducted by Achten [16]. Free-piston engines utilizing internal combustion (as opposed to the external combustion Stirling engine, which suffers from poor power density) have their origin in the 1920s when R. Pescara [1] patented their use as air compressors. Junkers in Germany developed a freepiston engine for use in German submarines in World War II. The French SIGMA free-piston gasifier saw service for decades in stationary power generation. The use of free-piston engines in automotive application was most heavily promoted in the period 1952 to 1961, when both General Motors and Ford Motor Company produced running prototypes [2], [3]. In both cases these engines were two-stroke, opposed piston spark ignited engines with combustion bounce/compression chambers. These engines where used as gasifiers to generate hot gases to drive exhaust turbines through which energy would be extracted. Development efforts largely ceased by the 1960s as turbine powered vehicles were increasingly viewed as not commercially viable. The elimination of the crankshaft mechanism in free-piston engines provides potential for reduction in mechanical losses. Due to the fact that the piston is not constrained in a free-piston engine, the piston motion is not prescribed, and it varies from one operating regime to another. Characteristic of the free-piston engines is the fact that they do not have a flywheel. As a result, they do not accumulate energy from the previous cycles for the subsequent cycles, except in terms of achieving a greater or lesser stroke associated with greater or lesser gas compression energy. A linear engine operates somewhat similarly to a free-piston engine, the only difference being that the reciprocating assembly consists of two pistons connected by a common connecting rod, with each piston operating in its own cylinder.

In the last two decades there has been resurgence in interest in these engines for hybrid electric vehicle applications. Several linear engines have been designed but most of them have a complicated mechanical configuration. This complication is the result of the need for tight control of the length of the stroke of the piston assembly, to avoid having the piston contact the cylinder head at its furthermost outer position. Some notable recent work in the development of linear engines is reviewed below. Bock [4] has developed a compression ignition linear engine. The engine incorporates a hydraulic pump cylinder arranged in the central part of the engine in order to provide useful work. The engine is gas cushioned and it uses a nitrogen filled elastic annular body serving as a shock absorbing body. A free-piston engine pump system has been disclosed by Heintz [5]. The device consists of a pair of opposite pistons connected by a common rod operating a hydraulic pump. The free-piston engine pump has double acting power pistons and pumping pistons attached as a main reciprocating member and movable in a housing. The housing itself moves for mass balancing purposes. The engine is spark ignited, and stroke control is provided by a valving system. Air is supplied and the exhaust products are removed from the two combustion chambers by common intake and common valves. These valves are operated by a common actuator in response to the position of the main reciprocating member. Rittmaster et al. [6] have presented a spark ignition linear engine connected to hydraulic power system. Hydraulic fluid is stored and maintained under pressure in two pressurization chambers on the opposite side of the pistons forming an internal combustion engine. In order to time the operation of the engine, a set of proximity detectors located around the connecting rod are used. The hydraulic fluid flow in the two pressurization chambers is controlled by a set of cross-over valves. A hydraulic motor is interposed downstream of the cross-over valves to convert the flow of the fluid into the rotation of a shaft. The cylinders of the engine have intake and exhaust valves electronically operated. A flywheel is attached to the shaft of the hydraulic motor to dampen the pulsation induced by the shifting of the cross-over valves and to store the energy of the pistons between reciprocation. Iliev et al. [7] have presented a linear engine coupled to an electrical alternator. The engine operates on a twostroke cycle and is spark ignited. The engine is controlled by an electronic module, which also controls the linear alternator. The spark timing is regulated based on the quality of fuel and the electrical load. According to the inventors the engine is designed to operate at high frequencies thereby attaining a high thermal efficiency. Galitello [8] has proposed a linear engine controlled by a computer. The engine is a two-stroke compression ignition engine. According to the inventor the engine operates at ultra high speeds and is vibration free. During engine starting, ignition assist is provided by spark plugs. 2

The engine can be connected to an electric generator or to a hydraulic power system. An interesting design has been presented by Kos [9]. The disclosed engine is a two-stroke cycle linear engine employing a reciprocating piston in conjunction with an electromagnetic transducer for control and power output. The engine can be spark or compression ignited (with multi-fuel operation possible). The engine is designed to be controlled by computer, with the engine stroke varied by tailoring the magnetic field. The inventor suggests an alternative arrangement of the engine by using a pair of opposite pistons linked by a connecting rod. Widener and Ingram [10] recently described a numerical model of a free-piston linear generator for a hybrid vehicle modeling study. The model addressed the use of a free-piston engine coupled with a linear generator as a potential auxiliary power unit in hybrid electric vehicles. The feasibility of such a model was analyzed with regards to power output and efficiency of the unit with reference to conversion of mechanical power output of the linear engine to electrical power output. The study was conducted on a two-stroke cycle engine and a reciprocating rig was developed to study and characterize the operation of the generator.

PREVIOUS WORK
The most recent work done in the area of the development of linear engines has been presented by Clark et al. [11],[12] and in the thesis of Nandkumar [15]. The authors have designed, constructed and tested a prototype spark-ignited linear engine coupled with a linear alternator (Figure 1). An idealized numerical model based on an air standard Otto cycle operation has also been presented [11]. This idealized model has revealed the relationship between stroke, compression ratio and operating parameters. The prototype engine has been tested extensively under several different operating modes, with the load being provided by a friction brake, as well as with the load applied by a permanent magnet linear alternator. In-cylinder pressure profiles for different applied loads were experimentally determined through extensive engine testing. The results of the engine testing, using a friction brake to provide the load, are presented in Table 1. A further analysis of the experimental operation of the linear engine in combination with the linear alternator has been presented [12]. The linear alternator is of the permanent magnet-type, and has been shown to produce as much as 316W from the prototype engine-alternator combination. Based on the in-cylinder pressure data gathered during operation of the enginealternator combination, an analysis of the cycle to cycle variation of integrated mean effective pressure (IMEPg) was performed. It was shown that there are significant cycle-to-cycle variations of the in-cylinder pressure versus time for different operating regimes. The result of the performance tests of the engine-alternator prototype combination are shown in Table 2.

AI R 8 3

11

A IR

10

2 1 9 6

Figure 1. Geometric Layout of the Prototype Two-Stroke Cycle Linear Engine-Alternator Combination modeled in this work. (1. Cylinder, 2. Piston, 3. Connecting Rod, 4. Pulsed Solenoid Fuel Injector, 5. Intake Port, 6. Exhaust Port, 7. Motoring Coil, 8. Linear Alternator, 9. Frame, 10. Spark Plug, 11. Throttle)

Table 1.

Experimental Data from the Linear Engine Operation using Load applied by a Friction Brake [11] Average positive indicated Work Output per stroke Average positive Power Output [W] 81.3 262.5 438.0 804.0

Average Stroke [mm] 44.3 35.3 37.6 47.0

Average Frequency [cycles/min] 1574 1260 1197 1470

Load No Yes Yes Yes Table 2.

[ J] 1.55 6.25 10.79 16.40

Experimental Data from the Linear Engine using Load applied by an Integrated Permanent Magnet Alternator [12] Load Test 1 2 3 4 5 6 7 8 9 10 11 [ohms] Open Circuit 156.0 130.0 104.0 78.0 52.0 26.0 24.0 23.4 19.5 17.3 Voltage [V] 132.0 120.0 119.0 115.0 111.0 103.0 90.0 88.5 87.5 79.0 74.0 Current [A] 0.00 0.75 0.88 1.07 1.38 1.92 3.30 3.54 3.58 3.90 4.15 Load Power Output [W] 0 92 104 124 153 200 300 312 313 316 309 Frequency [Hz] 25.0 24.6 24.4 23.4 24.1 26.6 23.6 23.6 23.6 23.1 22.7

DESCRIPTION OF THE MODELED ENGINE

The modeled engine is a two-stoke spark ignition engine connected to a linear alternator (see Figure 1). The engine consists of two opposed pistons, connected by a common connecting rod that is allowed to oscillate back and forth between the two end-mounted cylinders. The cylinders are ported such a manner to utilize a loop scavenging process, although this process was not optimized. The fuel is supplied to each cylinder by two pulse widthmodulated gasoline fuel injectors. In order to keep the engine temperature within a reasonable operating range, water is forced through the cylinder heads. An electronic control device allows the adjustment of the ignition timing and fuel injection timing and quantity. The engine stroke is controlled on a stroke-by-stroke basis by the ignition timing and the amount of fuel injected. The engine is equipped with two motoring coils (also connected to the electronic control device) used as a starting device. These are automatically disengaged after the engine exceeds a certain reciprocation frequency. The motoring coils are also used in case of misfire in one of the cylinders, to reverse the motion of the piston assembly and to assist in restarting. A permanent magnet linear alternator connected to the engine shaft transforms the kinetic energy of the reciprocating motion into electrical energy [11],[12].

where x represents the displacement of the piston assembly, and

d 2x dt 2

is the acceleration of the piston

The right hand side of equation (1) represents the summation of the forces that act in the plane of motion. The only forces are considered to act on the moving assembly are the resultant pressure forces given by the difference between the pressures in the two cylinders, a frictional force, the inertial force, and the load. Equation (1) can be written as

d x m 2 = FP F f L dt

(2)

FP = ( p1 p 2 )

D 2 4

(3)

where FP is the resultant of the in-cylinder pressure forces

Ff L D p1

is the friction force, is the load applied to the shaft, is the piston diameter, and and

NUMERICAL SIMULATION OF A TWO-STROKE LINEAR ENGINE

OBJECTIVE OF THE SIMULATION The objective of the present work is to develop a numerical model that simulates the operation of the two-stroke linear engine. The numerical model was developed for a spark ignited linear engine but can be easily adapted for the case of a compression ignition linear engine. The numerical analysis also allows a parametric study of the operation of this type of engine. The engine modeling has been validated using results from the existing working linear engine. The numerical model represents an idealized case due to the assumptions made, while allowing a parametric study to be performed. THE DYNAMIC MODEL The modeling starts with a dynamic analysis of the linear engine. Consider the case of a linear engine with two reciprocating pistons linked by a solid shaft, that oscillate back and forth in a left-to-right motion with fixed port timing through the use of port valves. A system of coordinates was chosen having the origin at the outermost point of the left cylinder. Considering a mechanical system represented by the piston assembly in motion, this system obeys Newton's second law.

p2

are

the

corresponding

in-cylinder

pressures. In order to determine the solution of this differential equation, is necessary to integrate it twice with respect to time. The analytic integration is somewhat complicated to evaluate due to the complex variation of the three forces in space and time. The piston assembly does not follow a prescribed motion but rather its resultant motion is established as a result of the net balance in the applied forces. THE APPLIED LOAD In experimental testing of the engine, a friction brake provided a retarding force on the shaft in order to obtain an approximate simulation of the load that a linear alternator would provide. According to the measurements made, the frictional drag was roughly constant across the full range of motion of the piston assembly with an average value of about 130 N. THE FRICTION FORCE Compared to conventional crankshaft mechanism internal combustion engines, the friction force for the linear piston engine seems to be less complicated, due to the fact that the elimination of the crankshaft mechanism reduces certain sources of friction. The friction force for the linear engine is a result of the friction between the piston rings, piston skirt and cylinder wall. The explicit determination of the friction forces for the experimental model is elusive, and an existing correlation from the literature is used. 4

d 2x m 2 = Fi x i dt

(1)

The resultant friction force in this model is taken as having a constant value throughout the length of the stroke, in order to simplify the calculation. The value for the frictional force has been determined by using a correlation determined by Blair [13] (for the friction caused by the rings and the pistons). Blair suggests an empirical relation for calculating the friction mean effective pressure for two-stroke engines. In the calculation of the mean effective frictional pressure, it is assumed that the motion of the piston assembly is the maximum theoretical stroke length.

the frictional force is not significant and so all future references to the load will include the frictional force. In this numerical model we have assumed several different functional forms for the shape, or profile, of the load as a function of the piston assembly position, all of which are assumed to average to the same absolute value. THE RESULTANT PRESSURE FORCE This force is the resultant of the pressure difference between the two cylinders, and is determined by evaluating the pressure variation in each cylinder through the engine cycle. THERMODYNAMIC ANALYSIS OF THE ENGINE CYCLE This analysis consists of an evaluation of the thermodynamic parameters for each process that characterizes the engine cycle, based on the first law of thermodynamics. The thermodynamic model used is a singlezone model. In single-zone models the mixture composition, pressure, and temperature of the combustion chamber are assumed to be uniform, and the energy released by the combustion of the fuel is specified or calculated after the fact from the measured in-cylinder pressure versus time or cylinder volume data. THE SCAVENGING PROCESS For the purposes of this numerical model, the scavenging process is assumed to be a perfect process. The cylinder pressure when the exhaust port closes is assumed to have the same value as the applied intake pressure. The exhaust blowdown process is also considered to be perfect, so that the cylinder pressure will instantaneously drop to the value of the intake pressure when the exhaust port opens. COMPRESSION CALCULATION In this calculation, the compression process is considered to be governed by a polytropic equation with a constant polytropic coefficient. The compression process is considered to take place between the time when the exhaust port is closed, and when the spark occurs. The pressure and temperature at the closing of the exhaust port (the beginning of the compression process) have the same values as the corresponding parameters at the end of the scavenging process. The pressure during the compression process is given as a function of the piston assembly position.

fmep = A L n
2 1

(4)

where A = 150kg.m .s , L is the length of the stroke in meters, and n is the engine speed, in our case the frequency of reciprocation of the piston assembly in cycles/ min. (For the purposes of clarification, a stroke refers to a single left-to-right or right-to-left movement of the piston assembly, while a cycle means a back and forth movement involving two individual strokes). For the experimental prototype engine at a representative speed of 1426 cycles / min and a maximum possible stroke length of 0.05 m it follows

fmep = 150 50 10 3 1426 = 10695 Pa

Assuming a constant friction force during the cycle, this can be calculated as follows:

fmep =
where force,

Wf Vd

(5)

Wf

is the work required to overcome the friction

Vd

is the displaced volume,

Vd =

D2 L 4

(6) (7)

W f = Ff 2L
fmep = 8F f

By substituting equations (5) and (6) into equation (4)

Ff = Ff =

b fmep 8
2

X ep 1 p 1 (t ) = p a X (t ) 1
(8)

mc

(9.a)

0.0364 2 10695 = 5.56 N 8

Similarly, according to the chosen coordinate system, the pressure variation during the compression process for the right cylinder is

This calculation confirms that, compared to the typical loads applied by the friction brake or the linear alternator,

X max X ep 2 p 2 (t ) = p a X (t ) X ep 2 2
5

mc

(9.b)

where

pa

is the intake port air pressure,

X 1 ( t ) is the left piston position, X 2 (t ) is the right piston position, X ep1 is the left cylinder exhaust port closing coordinate, X ep 2 is the right cylinder exhaust port closing coordinate,

X max X ep 2 (10.b) T2 (t ) = Ta X (t ) X 2 2 ep where Ta is the mixture temperature at the beginning of

the compression process that is assumed to be the same as the intake air temperature. The values for the compression polytropic exponent are taken from statistical data, see Table 3. COMBUSTION PROCESS CALCULATION The combustion calculation seeks to express the pressure and temperature variation in each combustion cylinder during the combustion process. The first assumption made is that the start of the combustion process coincides with the moment at which the spark occurs, which implies that the ignition delay is ignored. Another assumption made is that the heat input during the combustion process is the resultant heat input, after heat transfer to the walls and piston. Based on the assumption that the heat transfer is neglected, the in-cylinder pressure variation is given by:

m c 1

X max is the maximum theoretical stroke length, and m c is the compression polytropic coefficient (adopted
based on empirical data). The relationship between the two piston coordinates is

X 2 (t ) = X 1 (t ) + l
where l is the distance between the piston crowns. The temperature variation during the compression stroke is given by

X (t ) T1 (t ) = Ta 1 X ep1
Table 3.

m c 1

(10.a)

Q d dp p dV = + ( 1) in dt V dt V dt

(11)

Engine Parameters used in this Numerical Simulation. 2 36.4 [mm] 50 [mm] 1426 [cycles/min] 2.37 [m/s] 0.135 [kPa] 341 [K] Variable 1.28-1.37

NUMBER OF CYLINDERS BORE MAXIMUM POSIBLE STROKE FREQUENCY OF MOTION PISTON MEAN SPEED INTAKE PRESSURE INTAKE TEMPERATURE COMPRESSION POLYTROPIC EXPONENT mc

EXPANSION POLYTROPIC EXPONENT md

Variable 1.20-1.30

EXHAUST PORT OPENING INTAKE PORT OPENING EXHAUST PORT HEIGHT INTAKE PORT HEIGHT

19 [mm] from the end of the maximum theoretical stroke 21 [mm] from the end of the maximum theoretical stroke 10 [mm] 10 [mm]

where

=
stant,

cp cv

is the ratio of specific heats, assumed con-

cess, also known as the power stroke, is considered to be governed by a polytropic equation with a constant polytropic coefficient. The pressure and the temperature variation respectively are expressed as follows

1.33

V p
Q in
d dt

is the instantaneous cylinder volume. is the instantaneous cylinder pressure.

X p d (t ) = p ec ec X (t ) X Td (t ) = Tec ec X (t )
where

md

(15)

is the total heat energy added to the cycle during the combustion process. is the mass fraction burned rate.

(md 1 )

(16)

In order to solve this equation, it is necessary to assume a certain function for the mass fraction burned equation; in this approach we use the Wiebe function.The Wiebe function is usually expressed as the mass fraction burned as a function of an instantaneous crank angle. Since the linear engine has no crankshaft, it is not possible to express the mass fraction burned in this way. Instead, the mass fraction burned is expressed as a function of time explicitly rather than implicitly as a function of crank angle.
b +1 t ts = 1 exp a Cd

p d is the pressure during the expansion process

p ec is the pressure at the end of combustion process

Td is the temperature during the expansion process Tec is the temperature at the end of the combustion
process

X ec is the piston position at the of the combustion process

m d is the polytropic coefficient for the power stroke chosen based on statistical data.
Then for the right piston, the pressure and the temperature variation during the expansion process can be written as

(12)

is the mass fraction burned

Cd is the combustion duration t s is the start of the combustion a and b are functional shape
adjustable.

X p1 (t ) = p ec1 ec1 X (t )
parameters and are

md

(17.a)

According to Heywood [14], actual mass fraction burned curves are well fitted with a = 5 and b = 2 , while varying a and b changes the shape of the curve significantly. Then the mass fraction burned rate has the following form
b b +1 d b + 1 t ts t ts =a exp a dt Cd Cd Cd

X T1 (t ) = Tec1 ec1 X (t )

m d 1

(18.a)

Similarly for the right hand side cylinder, the pressure and temperature for the expansion process are

X X ec 2 p 2 (t ) = p ec 2 max X X (t ) max X X ec 2 T2 (t ) = Tec 2 max X X (t ) max

md

(17.b)

(13)

m d 1

(18.b)

The heat release rate can be written as a function of time as

dQin b + 1 t ts t ts = a exp a dt Cd Cd Cd
b

b +1

Q in (14)

The values for the expansion polytropic exponent are taken from statistical data and are given in Table 3.

RESULTS
Using the numerical modeling, we attempted to match the in-cylinder pressure profiles obtained from the experiments using the prototype engine. The information obtained from the experimental operation of the engine

THE EXPANSION PROCESS CALCULATION The expansion process starts at the end of the combustion process (once the mass fraction burned equals unity) and ends when the exhaust port opens. The expansion pro-

includes pressure-volume profiles (P-V diagrams), in-cylinder pressure profiles, the engine speed (or frequency of reciprocation), and the ignition timing. Matching the experimentally derived and numerically obtained in-cylinder pressure profiles is a complicated task due to the number of unknown variables, namely the absolute value of the heat addition, the combustion duration, and the actual load on the engine. There are an infinite number of combinations of these unknowns that can match the given shape of the in-cylinder pressure profiles. Nevertheless an important piece of information in the combustion process was that the spark timing was known. In these engine experiments, there is no explicit information regarding the combustion duration, so matching the experimental and numerical P-V diagrams is extremely complicated. In the simulations we chose base values for the reciprocating mass, heat addition, combustion duration and engine load. The results shown here for the most part are for the variation in one of these variables, with all other parameters held constant. In most cases the simulations assume the heat input per stroke to be 25 J, which corresponds to the combustion of 2.7 mg of gasoline per stoke with an apparent efficiency of 25%, the reciprocating mass to be 2.5 kg, which is the mass of the piston-alternator assembly in the prototype engine, the combustion duration to be 5.85 ms, and the engine load to average 130 N across the cycle, although with varying functional forms of the load. Figure 2 shows typical experimental results from the tests on the prototype engine, which we attempted to match with the simulations [12]. CASE I Constant Load Simulation The first case simulated was one in which the applied load of 130N was taken to be constant throughout the stroke (this being the simplest case). The same P-V diagram (peak pressure and integrated area) was obtained for the experiments as for the numerical simulation with a constant load but with long combustion durations (Figure 3). From Figure 4, it can be observed that for the same load and for decreasing values of combustion duration, the peak pressure increases, while not affecting the overall shape of the P-V diagram to any great extent. Also it can be observed that the higher the peak pressure, the higher the engine speed, or frequency of reciprocation (Figure 4). A further important observation is that the length of the stroke varies in proportion with the engine speed (Figure 5), with higher engine speeds corresponding in longer strokes due to the greater inertial forces developed by the piston assembly.

CASE II Linearly Varying Load The second case used in matching the P-V diagram was that of a triangular shaped load, that is a load that increases as the piston assembly moves from one extremity to its center position and then decreases linearly with displacement as the assembly approaches its outermost limit of the stroke. The total integrated area under the load curve was kept equal to the area under the load curve taken in the first case (for an average 130 N load). The idea was to determine the sensitivity of the P-V results to a variation in the application of the load, albeit with the same integrated net load (Figure 6). It was observed that there was neither considerable variation of the shape of the P-V curve (Figure 7) or of the engine speed versus the case with a constant load (Figure 8). The parameters used in the case of the triangular shaped load had the same values as for the case with constant load (notably the spark timing, the heat input, and the combustion duration). The motion of the piston assembly is then dominated by the balance of inertial and compressive forces. Both in case one and case two, the data obtained from the experiment was matched for a relatively long combustion duration. An interesting conclusion derives from the analysis of the previous two cases, namely the fact that the P-V diagram was apparently not sensitive to the shape of the load for the two cases considered. In order to prove the supposition that the operation of the linear engine is not sensitive to the shape of the load, a more complex semi-sinusoidal load profile was then investigated. Case III More Complex Load Functions Extending the analysis, the third case of more complex load functions was considered. Figure 9 illustrates the different shapes or profiles of the load that were used in the numerical model. These shapes were chosen to represent the cases of high load near the extremes of the piston stroke, or highest load in the center of the stroke. The case when the load is concentrated towards the end of the stroke corresponds to greater values of the shape factor k (where k varies from 0 to 3.55). From the data obtained from the simulation (Figure 10), it can be observed that the in-cylinder peak pressure varies markedly for the 3 different load profiles chosen. The frequency variation is, however, very small so that it can be concluded that the shape of the load does not influence the engine speed greatly. From Figure 11, it can be observed that the velocity profile changes fairly significantly with shape of the load. In this case the velocity profile is relatively constant in the middle portion of the stroke, while the peak pressure increases as the shape factor k increases. It was mentioned before that the higher the peak pressure, the higher the frequency of reciprocation of the engine and the longer the stroke.

60 00

50 00

40 00 Pressure [kPa]

30 00

20 00

10 00

0 0 5 10 15 20 25 D is p lac em en t [m m ] 30 35 40 45 50

Figure 2. Experimental Data derived from the Operation of the Prototype Linear Engine-Alternator Combination [12]

6000

5000

C d=0.00585 s freq=29.2 H z H eat input Q= 25 J/stroke The spark occ urs at X =45.35 m m

4000 Pressure [kPa]

3000

2000

1000

0 0 5 10 15 20 25 D isplacem ent [m m ] 30 35 40 45 50

Figure 3. In-Cylinder Pressure vs. Piston Assembly Displacement for a Constant Load (Case I) By concentrating the load in the middle of the stroke (corresponding to lower values of the shape factor k), the velocity profile rotates clockwise in the velocity-displacement domain. In this case the peak pressure has a lower value and the frequency and the engine stroke decrease correspondingly. of a prototype unit. The simulations were used to show the effect of the total heat input, the combustion duration, the reciprocating mass and the load on the operation of this engine-alternator combination. The experimental testing performed previously on the linear engine [12] showed that the linear alternator introduces a load that has a roughly sinusoidal shape throughout the stroke here considered to be of second order (corresponding to k=0 in Figure 9). Using this load profile, a sensitivity analysis was performed to determine the influence upon the engine operation of variation in one or more parameters.

CONCLUSIONS
A numerical model of a spark-ignited two-stroke cycle linear engine-alternator combination has been developed and validated using experimental data from the operation

16 00 0 Q =2 5 J/stro ke 14 00 0 C d=0 .0 04 85 s fre q=3 4.0 H z C d=0 .0 03 85 s fre q=3 7.1 H z

12 00 0

10 00 0 Pressure [kPa]

C d=0 .0 05 85 s fre q=2 9.2 H z

80 00

60 00

40 00

20 00

0 30 32 34 36 38 40 D is p lac em en t [m m ] 42 44 46 48 50

Figure 4. In-Cylinder Pressure vs. Piston Assembly Displacement for different values of the Combustion Duration for the same Heat Input (Case I)

5 C d=0.00385s 4 C d=0.00485s

3 C d=0.00585s

Velocity [m/s]

0 0 -1 5 10 15 20 25 30 35 40 45 50

-2

-3

-4

-5 D isplacem ent [m m ]

Figure 5. Piston Velocity vs. Displacement for different values of the Combustion Duration (Case I) It was observed that the variation of the heat input influences the peak pressure, the frequency of the engine, and also the stroke length (Figures 12, 13 and 14). Variation in the combustion duration also influences the peak pressure, the frequency of the engine, and the displacement in the same manner as for the heat input variation. By varying the mass of the moving piston assembly, the peak pressure and maximum displacement, vary proportionally. For a greater mass, the peak pressure and the engine stroke both increase. The frequency varies in an inverse proportional relationship with the mass of the shaft. A further observation from the numerical simulation is that, for very low piston assembly masses, the operation of the linear engine becomes possible only if the heat input is increased significantly, due to the low inertial forces associated with low piston speeds.

10

60 00

50 00

40 00 Pressure [kPa] C d = 0 .0 0 5 8 5 s fre q = 3 0 .6 H z H e a t in p u t Q = 2 5 J /s t ro k e T h e s p a rk o c c u rs a t X = 4 5 .3 5 m m

30 00

20 00

10 00

0 0 5 10 15 20 25 D is p la c e m e n t [m m ] 30 35 40 45 50

Figure 6. In-Cylinder Pressure vs. Displacement for a triangular shaped Load and Friction Force (Case II)

60 00

50 00

C d=0 .0 05 85 s fre q=2 9.2 H z Lo ad +F riction are trian gu la r sh ap ed Q = 25 J/stroke C d=0 .0 05 85 s fre q=3 0.6 H z Lo ad =co nstan t F riction = co nstan t Q = 25 J/stroke

40 00 Pressure [kPa]

30 00

20 00

10 00

0 0 5 10 15 20 25 D is p lac em en t [m m ] 30 35 40 45 50

Figure 7. In-Cylinder Pressure vs. Piston Displacement for different profiles of the Friction Force and Load (for the same Combustion Process parameters and the same Heat input Cases I and II) Figures 13 and 14 show the in-cylinder pressure versus displacement and the velocity versus displacement respectively, for different values of the shaft mass. The results of the parametric analysis are shown in Table 4. FUTURE WORK Using the results of this numerical simulation, a second prototype linear engine-alternator combination is currently under development. This unit, which will employ compression ignition with a state-ofthe-art direct injection diesel fueling system, is designed to produce around 10 kW of electrical power, and will be integrated into a series hybrid electric vehicle in a demonstration project. 11

ACKNOWLEDGMENTS
The support of the Department of Defense in funding this research is acknowledged [Grant No. DAAH04-96-10328].

REFERENCES
1. Cleveland Diesel, History and description of the free piston engine-gas turbine power, year unkown. 2. A.F. Underwood,GMR 4-4 Hyprex Free Piston Turbine Engine, SAE Journal, June 1956, pp. 60-66.

5 Constant Load and Friction Force Load+Ff=130 N Q=25 J/stroke

2 Vel oci ty [m/ s]

0 0 -1 5 10 15 20 25 30 35 40 45 50

-2

-3

-4

-5

Triangular shape of the Resultant Force of the Load and the Friction Force Displacement [mm]

Figure 8. Piston Velocity vs. Displacement for different shapes of the Friction Force and Load (for the same Combustion Process Parameters Cases I and II)

800

700

k=3.55

600

500 Lo ad [N ] k=0 400 k=1.7 300

200

100

0 0 -100 Displacement [m] 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05

Figure 9. More Complex Applied Load Profiles (Case III)

3. D.N. Frey, P Klotsch and A. Egli,The Automotive Free-Pis. ton-Turbine Engine, SAE Transactions, Vol. 65, 1957, pp. 629-634. 4. R. Bock, U.S. Patent 4,128,083, December 5,1978,Gas Cushion Free-Piston Type Engine 5. R. P Heintz, U.S. Patent 4,369,021, May 5,1980,Free-Pis. ton Engine Pump 6. P A. Rittmaster et al., U.S. Patent 4,326,380, April . 27,1982,Hydraulic Engine 7. M. D. Iliev et al., U.S. Patent 4,532,431, July 30,1985,Method and Apparatus for Producing Electrical Energy from a Cyclic Combustion Process utilizing Coupled Pistons which Reciprocate in Unison 8. K. A. Galitello, Jr., U.S. Patent 4,876,991, October 31,1989,Two Stroke Cycle Engine 9. J. F. Kos, U.S. Patent 5,002,020, March 26,1991,Computer Optimized Hybrid Engine

12

80 00 k=3 .5 70 00 k=1 .7 60 00 k=0

50 00 Pressure [kPa]

40 00

30 00

20 00

10 00

0 30 32 34 36 38 40 D is p lac em en t [m m ] 42 44 46 48 50

Figure 10. In-Cylinder Pressure vs. Piston Displacement for Different Profiles of the Load (Case III)

5 C d=0.00585 s Qin=25 J/C y cle 4 k=1.7 freq=29.7 H z

2 k=0 freq=32.1 H z 1 Velocity [m/s]

0 0 -1 5 10 15 20 25 30 35 40 45 50

-2

-3

-4 k=3.5 freq=32.7 H z -5 D isplacem ent [m m ]

Figure 11. Piston Velocity vs. Displacement for Different Load Profiles (Case III)
10. S. K. Widener and Ingram, K., Free-Piston Engine Linear Generator Technology Development, Final Report, Under Contract to U.S. Army TARDEC, Mobility Technology Center-Belvoir, Fort Belvoir, Virginia, January 1995. 11. N. N. Clark, T. I. McDaniel, R. J. Atkinson, S. Nandkumar, C. M. Atkinson, S. Petreanu and P Famouri, " Modeling and . Development of a Linear Engine", 1998 Spring Technical Conference, ASME ICE Division, Fort Lauderdale, FL. 12. N. N. Clark, T. I. McDaniel, R. J. Atkinson, S. Nandkumar, C. M. Atkinson, S. Petreanu and P Famouri, " Operation of . a Small Bore Two-Stroke Linear Engine", 1998 13. P Blair, Design and Simulation of Two-Stroke Engines, .G. SAE Inc., Warrendale, Pa., 1996 14. J. B. Heywood, Internal Combustion Engine Fundamentals, John Wiley and Sons, New York, 1986. 15. S. Nandkumar, Modeling of a Linear Engine, MSME Thesis, West Virginia University, 1998 16. P A. J. Achten, A Review of Free-Piston Engine Con. cepts, SAE 941776, 1994.

13

12 00 0 Q = 26 .5 J/stro ke fre q=3 6.1 H z 10 00 0

Q = 26 .0 J/stro ke fre q=3 4.6 H z

80 00 Pressure [kPa]

Q = 25 .8 J/stro ke fre q=3 3.4 H z

60 00

40 00

20 00

0 30 32 34 36 38 40 D is p lac em en t [m m ] 42 44 46 48 50

Figure 12. In-Cylinder Pressure vs. Displacement for varying Heat Input (for the same load profile, k=0, Case III)

3 50 00 M as s=5 .3 kg 3 00 00

2 50 00

Pressure [kPa]

2 00 00

M as s=3 .3 kg

1 50 00

1 00 00 M as s=2 .3 kg

5 00 0

0 30 32 34 36 38 40 D is p la ce m e n t [m m ] 42 44 46 48 50

Figure 13. In-Cylinder Pressure vs. Displacement for different values of the Reciprocating Mass (for the same load profile, k=0, Case III)

14

5 Mass=4.3 kg Mass=3.3 kg 4 Mass=5.3 kg

1 Velocity [m/s]

0 0 -1 5 10 15 20 25 30 35 40 45 50

-2

-3

-4

-5 Displacement [mm]

Figure 14. Piston Velocity vs. Displacement for different values of the Reciprocating Mass (for the same load profile, k=0, Case III)

Table 4.

Parametric Dependence of varying Heat Input, Combustion Duration, Applied Load, and Reciprocating Mass on Engine Performance. PARAMETER VARIED Heat Input Combustion Duration Frictional Force (Load) Mass of Reciprocating Assembly

denotes proportional dependence denotes inverse proportionality

EFFECT SEEN

Frequency Peak pressure Velocity Displacement

CONTACT
Dr. Chris Atkinson, Dept. of Mechanical and Aerospace Engineering, West Virginia University, PO Box 6106, Morgantown, WV 26506, USA

Phone: (304) 293-4111 ext. 333 catkinson@cemr.wvu.edu

15