Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

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
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

The appearance of this 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 $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 Sec-
tions 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
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

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 INTRODUCTION AND LITERATURE REVIEW

Series hybrid electric vehicles (HEVs) require power- Free-piston engines have been a subject of research and
plants that can generate electrical energy without specifi- development for several decades. A recent review of free-
cally requiring rotary input shaft motion. A small-bore piston engine concepts has been conducted by Achten
working prototype of a two-stroke spark ignited linear [16]. Free-piston engines utilizing internal combustion (as
engine-alternator combination has been designed, con- opposed to the external combustion Stirling engine,
structed and tested and has been found to produce as which suffers from poor power density) have their origin
much as 316W of electrical energy. This engine consists in the 1920’s when R. Pescara [1] patented their use as
of two opposed pistons (of 36 mm diameter) linked by a air compressors. Junkers in Germany developed a free-
connecting rod with a permanent magnet alternator piston engine for use in German submarines in World
arranged on the reciprocating shaft. This paper presents War II. The French SIGMA free-piston gasifier saw ser-
the numerical modeling of the operation of the linear vice for decades in stationary power generation. The use
engine. The piston motion of the linear engine is not of free-piston engines in automotive application was most
mechanically defined: it rather results from the balance of heavily promoted in the period 1952 to 1961, when both
the in-cylinder pressures, inertia, friction, and the load General Motors and Ford Motor Company produced run-
applied to the shaft by the alternator, along with history ning prototypes [2], [3]. In both cases these engines were
effects from the previous cycle. The engine computa- two-stroke, opposed piston spark ignited engines with
tional model combines dynamic and thermodynamic combustion bounce/compression chambers. These
analyses. The dynamic analysis performed consists of an engines where used as gasifiers to generate hot gases to
evaluation of the frictional forces and the load (in this drive exhaust turbines through which energy would be
case the alternator load) across the full operating cycle of extracted. Development efforts largely ceased by the
the engine. The thermodynamic analysis consists of an 1960’s as turbine powered vehicles were increasingly
evaluation of each process that characterizes the engine viewed as not commercially viable. The elimination of the
cycle, including scavenging, compression, combustion crankshaft mechanism in free-piston engines provides
and expansion, based on the first law of thermodynam- potential for reduction in mechanical losses. Due to the
ics. Since the modeled engine was crankshaftless, a fact that the piston is not constrained in a free-piston
time-based Wiebe function (as opposed to a conventional engine, the piston motion is not prescribed, and it varies
crank angle-based approach) was used to express the from one operating regime to another. Characteristic of
mass fraction burned for the combustion process, while the free-piston engines is the fact that they do not have a
the combustion model used was a single-zone model. To flywheel. As a result, they do not accumulate energy from
render the model useful, the parameters used were the previous cycles for the subsequent cycles, except in
based on experimental data obtained from the working terms of achieving a greater or lesser stroke associated
example, including instantaneous shaft position, velocity with greater or lesser gas compression energy.
and in-cylinder pressure. Also, a parametric study was
A linear engine operates somewhat similarly to a free-pis-
performed to predict the behavior of the engine over a
ton engine, the only difference being that the reciprocat-
wide operating range, given variations in fuel combustion
ing assembly consists of two pistons connected by a
properties, the reciprocating mass of the piston shaft
common connecting rod, with each piston operating in its
assembly, frictional load and the externally applied elec-
own cylinder.
trical load.

1
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

In the last two decades there has been resurgence in The engine can be connected to an electric generator or
interest in these engines for hybrid electric vehicle to a hydraulic power system.
applications. Several linear engines have been designed
An interesting design has been presented by Kos [9]. The
but most of them have a complicated mechanical configu-
disclosed engine is a two-stroke cycle linear engine
ration. This complication is the result of the need for tight
employing a reciprocating piston in conjunction with an
control of the length of the stroke of the piston assembly,
electromagnetic transducer for control and power output.
to avoid having the piston contact the cylinder head at its
The engine can be spark or compression ignited (with
furthermost outer position. Some notable recent work in
multi-fuel operation possible). The engine is designed to
the development of linear engines is reviewed below.
be controlled by computer, with the engine stroke varied
Bock [4] has developed a compression ignition linear by tailoring the magnetic field. The inventor suggests an
engine. The engine incorporates a hydraulic pump cylin- alternative arrangement of the engine by using a pair of
der arranged in the central part of the engine in order to opposite pistons linked by a connecting rod.
provide useful work. The engine is gas cushioned and it
Widener and Ingram [10] recently described a numerical
uses a nitrogen filled elastic annular body serving as a
model of a free-piston linear generator for a hybrid vehi-
shock absorbing body.
cle modeling study. The model addressed the use of a
A free-piston engine pump system has been disclosed by free-piston engine coupled with a linear generator as a
Heintz [5]. The device consists of a pair of opposite pis- potential auxiliary power unit in hybrid electric vehicles.
tons connected by a common rod operating a hydraulic The feasibility of such a model was analyzed with regards
pump. The free-piston engine pump has double acting to power output and efficiency of the unit with reference
power pistons and pumping pistons attached as a main to conversion of mechanical power output of the linear
reciprocating member and movable in a housing. The engine to electrical power output. The study was con-
housing itself moves for mass balancing purposes. The ducted on a two-stroke cycle engine and a reciprocating
engine is spark ignited, and stroke control is provided by rig was developed to study and characterize the opera-
a valving system. Air is supplied and the exhaust prod- tion of the generator.
ucts are removed from the two combustion chambers by
common intake and common valves. These valves are PREVIOUS WORK
operated by a common actuator in response to the posi-
tion of the main reciprocating member. The most recent work done in the area of the develop-
ment of linear engines has been presented by Clark et al.
Rittmaster et al. [6] have presented a spark ignition linear
[11],[12] and in the thesis of Nandkumar [15]. The
engine connected to hydraulic power system. Hydraulic
authors have designed, constructed and tested a proto-
fluid is stored and maintained under pressure in two pres-
type spark-ignited linear engine coupled with a linear
surization chambers on the opposite side of the pistons
alternator (Figure 1). An idealized numerical model
forming an internal combustion engine. In order to time
based on an air standard Otto cycle operation has also
the operation of the engine, a set of proximity detectors
been presented [11]. This idealized model has revealed
located around the connecting rod are used. The hydrau-
the relationship between stroke, compression ratio and
lic fluid flow in the two pressurization chambers is con-
operating parameters. The prototype engine has been
trolled by a set of cross-over valves. A hydraulic motor is
tested extensively under several different operating
interposed downstream of the cross-over valves to con-
modes, with the load being provided by a friction brake,
vert the flow of the fluid into the rotation of a shaft. The
as well as with the load applied by a permanent magnet
cylinders of the engine have intake and exhaust valves
linear alternator. In-cylinder pressure profiles for different
electronically operated. A flywheel is attached to the
applied loads were experimentally determined through
shaft of the hydraulic motor to dampen the pulsation
extensive engine testing. The results of the engine test-
induced by the shifting of the cross-over valves and to
ing, using a friction brake to provide the load, are pre-
store the energy of the pistons between reciprocation.
sented in Table 1. A further analysis of the experimental
Iliev et al. [7] have presented a linear engine coupled to operation of the linear engine in combination with the lin-
an electrical alternator. The engine operates on a two- ear alternator has been presented [12]. The linear alter-
stroke cycle and is spark ignited. The engine is controlled nator is of the permanent magnet-type, and has been
by an electronic module, which also controls the linear shown to produce as much as 316W from the prototype
alternator. The spark timing is regulated based on the engine-alternator combination. Based on the in-cylinder
quality of fuel and the electrical load. According to the pressure data gathered during operation of the engine-
inventors the engine is designed to operate at high fre- alternator combination, an analysis of the cycle to cycle
quencies thereby attaining a high thermal efficiency. variation of integrated mean effective pressure (IMEPg)
was performed. It was shown that there are significant
Galitello [8] has proposed a linear engine controlled by a
cycle-to-cycle variations of the in-cylinder pressure ver-
computer. The engine is a two-stroke compression igni-
sus time for different operating regimes. The result of the
tion engine. According to the inventor the engine oper-
performance tests of the engine-alternator prototype
ates at ultra high speeds and is vibration free. During
combination are shown in Table 2.
engine starting, ignition assist is provided by spark plugs.

2
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

4 AI R 11 A IR

8
10
7
3

2
1 6
9

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 Average positive Power
per stroke Output Average Stroke Average Frequency
Load [ J] [W] [mm] [cycles/min]
No 1.55 81.3 44.3 1574
Yes 6.25 262.5 35.3 1260
Yes 10.79 438.0 37.6 1197
Yes 16.40 804.0 47.0 1470

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

3
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

DESCRIPTION OF THE MODELED ENGINE where x represents the displacement of the piston
assembly, and
The modeled engine is a two-stoke spark ignition engine
connected to a linear alternator (see Figure 1). The d 2x
engine consists of two opposed pistons, connected by a is the acceleration of the piston
common connecting rod that is allowed to oscillate back dt 2
and forth between the two end-mounted cylinders. The The right hand side of equation (1) represents the sum-
cylinders are ported such a manner to utilize a loop scav- mation of the forces that act in the plane of motion.
enging process, although this process was not optimized.
The fuel is supplied to each cylinder by two pulse width- The only forces are considered to act on the moving
modulated gasoline fuel injectors. In order to keep the assembly are the resultant pressure forces given by the
engine temperature within a reasonable operating range, difference between the pressures in the two cylinders, a
water is forced through the cylinder heads. An electronic frictional force, the inertial force, and the load.
control device allows the adjustment of the ignition timing Equation (1) can be written as
and fuel injection timing and quantity. The engine stroke
is controlled on a stroke-by-stroke basis by the ignition 2
d x
timing and the amount of fuel injected. The engine is m 2 = FP − F f − L (2)
equipped with two motoring coils (also connected to the dt
electronic control device) used as a starting device.
πD 2
These are automatically disengaged after the engine FP = ( p1 − p 2 ) (3)
exceeds a certain reciprocation frequency. The motoring 4
coils are also used in case of misfire in one of the cylin-
where FP is the resultant of the in-cylinder pressure
ders, to reverse the motion of the piston assembly and to
forces
assist in restarting.
A permanent magnet linear alternator connected to the Ff is the friction force,
engine shaft transforms the kinetic energy of the recipro-
cating motion into electrical energy [11],[12]. L is the load applied to the shaft,

NUMERICAL SIMULATION OF A TWO-STROKE D is the piston diameter, and

LINEAR ENGINE
p1 and p2 are the corresponding in-cylinder
OBJECTIVE OF THE SIMULATION – The objective of pressures.
the present work is to develop a numerical model that
simulates the operation of the two-stroke linear engine. In order to determine the solution of this differential equa-
The numerical model was developed for a spark ignited tion, is necessary to integrate it twice with respect to
linear engine but can be easily adapted for the case of a time. The analytic integration is somewhat complicated to
compression ignition linear engine. The numerical analy- evaluate due to the complex variation of the three forces
sis also allows a parametric study of the operation of this in space and time. The piston assembly does not follow a
type of engine. The engine modeling has been validated prescribed motion but rather its resultant motion is estab-
using results from the existing working linear engine. The lished as a result of the net balance in the applied forces.
numerical model represents an idealized case due to the
assumptions made, while allowing a parametric study to THE APPLIED LOAD – In experimental testing of the
be performed. engine, a friction brake provided a retarding force on the
shaft in order to obtain an approximate simulation of the
THE DYNAMIC MODEL – The modeling starts with a load that a linear alternator would provide. According to
dynamic analysis of the linear engine. Consider the case the measurements made, the frictional drag was roughly
of a linear engine with two reciprocating pistons linked by constant across the full range of motion of the piston
a solid shaft, that oscillate back and forth in a left-to-right assembly with an average value of about 130 N.
motion with fixed port timing through the use of port
valves. A system of coordinates was chosen having the THE FRICTION FORCE – Compared to conventional
origin at the outermost point of the left cylinder. crankshaft mechanism internal combustion engines, the
friction force for the linear piston engine seems to be less
Considering a mechanical system represented by the complicated, due to the fact that the elimination of the
piston assembly in motion, this system obeys Newton's crankshaft mechanism reduces certain sources of fric-
second law. tion. The friction force for the linear engine is a result of
the friction between the piston rings, piston skirt and cyl-
d 2x inder wall. The explicit determination of the friction forces
m 2 = ∑ Fi x (1) for the experimental model is elusive, and an existing cor-
dt i relation from the literature is used.

4
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

The resultant friction force in this model is taken as hav- the frictional force is not significant and so all future refer-
ing a constant value throughout the length of the stroke, ences to the load will include the frictional force.
in order to simplify the calculation. The value for the fric-
In this numerical model we have assumed several differ-
tional force has been determined by using a correlation
ent functional forms for the shape, or profile, of the load
determined by Blair [13] (for the friction caused by the
as a function of the piston assembly position, all of which
rings and the pistons). Blair suggests an empirical rela-
are assumed to average to the same absolute value.
tion for calculating the friction mean effective pressure for
two-stroke engines. In the calculation of the mean effec-
THE RESULTANT PRESSURE FORCE – This force is
tive frictional pressure, it is assumed that the motion of
the resultant of the pressure difference between the two
the piston assembly is the maximum theoretical stroke
cylinders, and is determined by evaluating the pressure
length.
variation in each cylinder through the engine cycle.
fmep = A ⋅ L ⋅ n (4)
THERMODYNAMIC ANALYSIS OF THE ENGINE
−2 −1
where A = 150kg.m .s , L is the length of the stroke CYCLE – This analysis consists of an evaluation of the
in meters, and n is the engine speed, in our case the fre- thermodynamic parameters for each process that charac-
quency of reciprocation of the piston assembly in cycles/ terizes the engine cycle, based on the first law of thermo-
min. dynamics. The thermodynamic model used is a single-
zone model. In single-zone models the mixture composi-
(For the purposes of clarification, a “stroke” refers to a tion, pressure, and temperature of the combustion cham-
single left-to-right or right-to-left movement of the piston ber are assumed to be uniform, and the energy released
assembly, while a “cycle” means a back and forth move- by the combustion of the fuel is specified or calculated
ment involving two individual strokes). after the fact from the measured in-cylinder pressure ver-
For the experimental prototype engine at a representative sus time or cylinder volume data.
speed of 1426 cycles / min and a maximum possible
stroke length of 0.05 m it follows THE SCAVENGING PROCESS – For the purposes of
this numerical model, the scavenging process is
fmep = 150 × 50 × 10 −3 ⋅ 1426 = 10695 Pa assumed to be a perfect process. The cylinder pressure
when the exhaust port closes is assumed to have the
Assuming a constant friction force during the cycle, this same value as the applied intake pressure. The exhaust
can be calculated as follows: blowdown process is also considered to be perfect, so
that the cylinder pressure will instantaneously drop to the
Wf value of the intake pressure when the exhaust port
fmep = (5) opens.
Vd
COMPRESSION CALCULATION – In this calculation,
where Wf is the work required to overcome the friction the compression process is considered to be governed
force, by a polytropic equation with a constant polytropic coeffi-
cient. The compression process is considered to take
Vd is the displaced volume,
place between the time when the exhaust port is closed,
and when the spark occurs. The pressure and tempera-
π ⋅ D2
Vd = L (6) ture at the closing of the exhaust port (the beginning of
4 the compression process) have the same values as the
corresponding parameters at the end of the scavenging
W f = Ff ⋅ 2L (7)
process.
8F f The pressure during the compression process is given as
fmep = a function of the piston assembly position.
π ⋅D 2

mc
 X ep 1 
p 1 (t ) = p a ⋅ 
By substituting equations (5) and (6) into equation (4)
 (9.a)
 X 1 (t ) 
π ⋅b 2
Ff = fmep (8) Similarly, according to the chosen coordinate system, the
8
pressure variation during the compression process for
π ⋅ 0.0364 2 the right cylinder is
Ff = 10695 = 5.56 N
8 mc
 X max − X ep 2 
This calculation confirms that, compared to the typical p 2 (t ) = p a ⋅  

( )
(9.b)
loads applied by the friction brake or the linear alternator,
 2X t − X ep 2 

5
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

where m c −1
pa  X max − X ep 2 
is the intake port air pressure,
T2 (t ) = Ta ⋅  

( )
(10.b)
X 1 ( t ) is the left piston position,  X 2 t − X ep 
2

X 2 (t ) is the right piston position, where Ta is the mixture temperature at the beginning of
X ep1 is the left cylinder exhaust port closing coordinate, the compression process that is assumed to be the same
as the intake air temperature. The values for the com-
X ep 2 is the right cylinder exhaust port closing coordi- pression polytropic exponent are taken from statistical
nate, data, see Table 3.
X max is the maximum theoretical stroke length, and
COMBUSTION PROCESS CALCULATION – The com-
m c is the compression polytropic coefficient (adopted bustion calculation seeks to express the pressure and
based on empirical data). temperature variation in each combustion cylinder during
The relationship between the two piston coordinates is the combustion process. The first assumption made is
that the start of the combustion process coincides with
X 2 (t ) = X 1 (t ) + l the moment at which the spark occurs, which implies that
the ignition delay is ignored. Another assumption made is
where l is the distance between the piston crowns. that the heat input during the combustion process is the
resultant heat input, after heat transfer to the walls and
The temperature variation during the compression stroke piston. Based on the assumption that the heat transfer is
is given by neglected, the in-cylinder pressure variation is given by:

 X (t ) 
m c −1
dp p dV Q dχ
T1 (t ) = Ta ⋅  1  (10.a)
= −γ ⋅ ⋅ + (γ − 1) in (11)
dt V dt V dt
 X ep1 

Table 3. Engine Parameters used in this Numerical Simulation.

NUMBER OF CYLINDERS 2

BORE 36.4 [mm]

MAXIMUM POSIBLE STROKE 50 [mm]

FREQUENCY OF MOTION 1426 [cycles/min]

PISTON MEAN SPEED 2.37 [m/s]

INTAKE PRESSURE 0.135 [kPa]

INTAKE TEMPERATURE 341 [K]

COMPRESSION POLYTROPIC EXPONENT mc Variable

1.28-1.37

EXPANSION POLYTROPIC EXPONENT md Variable

1.20-1.30

EXHAUST PORT OPENING 19 [mm] from the end of the maximum theoretical stroke

INTAKE PORT OPENING 21 [mm] from the end of the maximum theoretical stroke

EXHAUST PORT HEIGHT 10 [mm]

INTAKE PORT HEIGHT 10 [mm]

6
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

where cess, also known as the power stroke, is considered to be

governed by a polytropic equation with a constant poly-
cp
γ = is the ratio of specific heats, assumed con-
tropic coefficient. The pressure and the temperature vari-
ation respectively are expressed as follows
cv
stant, γ ≅ 1.33 md
 X 
V is the instantaneous cylinder volume. p d (t ) = p ec  ec  (15)
 X (t ) 
p is the instantaneous cylinder pressure.
(md −1 )
Q in  X 
Td (t ) = Tec  ec 
is the total heat energy added to the cycle during
the combustion process. (16)
dχ  X (t ) 
is the mass fraction burned rate.
dt where

In order to solve this equation, it is necessary to assume p d is the pressure during the expansion process
a certain function for the mass fraction burned equation; p ec is the pressure at the end of combustion process
in this approach we use the Wiebe [Link] Wiebe
function is usually expressed as the mass fraction burned
Td is the temperature during the expansion process
as a function of an instantaneous crank angle. Since the Tec is the temperature at the end of the combustion
linear engine has no crankshaft, it is not possible to process
express the mass fraction burned in this way. Instead, the
mass fraction burned is expressed as a function of time X ec is the piston position at the of the combustion
explicitly rather than implicitly as a function of crank process
angle. m d is the polytropic coefficient for the power stroke cho-
sen based on statistical data.
  t − ts  
b +1

χ = 1 − exp − a ⋅    (12)
Then for the right piston, the pressure and the tempera-
ture variation during the expansion process can be writ-
  Cd   ten as
χ is the mass fraction burned md
X 
Cd is the combustion duration p1 (t ) = p ec1  ec1 
 X (t ) 
(17.a)
t s is the start of the combustion
m d −1
a and b are functional shape parameters and are
X 
adjustable. T1 (t ) = Tec1  ec1 
 X (t ) 
(18.a)
According to Heywood [14], actual mass fraction burned
curves are well fitted with a = 5 and b = 2 , while
Similarly for the right hand side cylinder, the pressure and
varying a and b changes the shape of the curve
temperature for the expansion process are
significantly.
md
Then the mass fraction burned rate has the following
 X − X ec 2 
form p 2 (t ) = p ec 2  max 
( )
(17.b)
b +1  max
X − X t 
dχ b + 1  t − ts 
b
  t − ts  
=a⋅ ⋅  exp  − a ⋅    (13) m d −1
Cd  Cd   Cd    X − X ec 2 
T2 (t ) = Tec 2  max
dt   (18.b)
The heat release rate can be written as a function of time  max
X − X (t ) 
as
The values for the expansion polytropic exponent are
taken from statistical data and are given in Table 3.
dQin b + 1  t − ts 
b
  t − ts 
b +1

= a⋅ ⋅  exp  − a ⋅    ⋅ Q in (14)
dt Cd  Cd    Cd   RESULTS

THE EXPANSION PROCESS CALCULATION – The ex- Using the numerical modeling, we attempted to match
pansion process starts at the end of the combustion pro- the in-cylinder pressure profiles obtained from the experi-
cess (once the mass fraction burned equals unity) and ments using the prototype engine. The information
ends when the exhaust port opens. The expansion pro- obtained from the experimental operation of the engine

7
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

includes pressure-volume profiles (P-V diagrams), in-cyl- CASE II – Linearly Varying Load – The second case
inder pressure profiles, the engine speed (or frequency of used in matching the P-V diagram was that of a triangular
reciprocation), and the ignition timing. shaped load, that is a load that increases as the piston
assembly moves from one extremity to its center position
Matching the experimentally derived and numerically
and then decreases linearly with displacement as the
obtained in-cylinder pressure profiles is a complicated
assembly approaches its outermost limit of the stroke.
task due to the number of unknown variables, namely the
The total integrated area under the load curve was kept
absolute value of the heat addition, the combustion dura-
equal to the area under the load curve taken in the first
tion, and the actual load on the engine. There are an infi-
case (for an average 130 N load). The idea was to deter-
nite number of combinations of these unknowns that can
mine the sensitivity of the P-V results to a variation in the
match the given shape of the in-cylinder pressure pro-
application of the load, albeit with the same integrated
files. Nevertheless an important piece of information in
net load (Figure 6).
the combustion process was that the spark timing was
known. In these engine experiments, there is no explicit It was observed that there was neither considerable vari-
information regarding the combustion duration, so match- ation of the shape of the P-V curve (Figure 7) or of the
ing the experimental and numerical engine speed versus the case with a constant load (Fig-
ure 8). The parameters used in the case of the triangular
P-V diagrams is extremely complicated.
shaped load had the same values as for the case with
In the simulations we chose base values for the recipro- constant load (notably the spark timing, the heat input,
cating mass, heat addition, combustion duration and and the combustion duration). The motion of the piston
engine load. The results shown here for the most part are assembly is then dominated by the balance of inertial and
for the variation in one of these variables, with all other compressive forces.
parameters held constant.
Both in case one and case two, the data obtained from
In most cases the simulations assume the experiment was matched for a relatively long com-
bustion duration. An interesting conclusion derives from
• the heat input per stroke to be 25 J, which corre-
the analysis of the previous two cases, namely the fact
sponds to the combustion of 2.7 mg of gasoline per
that the P-V diagram was apparently not sensitive to the
stoke with an apparent efficiency of 25%,
shape of the load for the two cases considered. In order
• the reciprocating mass to be 2.5 kg, which is the to prove the supposition that the operation of the linear
mass of the piston-alternator assembly in the proto- engine is not sensitive to the shape of the load, a more
type engine, complex semi-sinusoidal load profile was then
• the combustion duration to be 5.85 ms, investigated.
• and the engine load to average 130 N across the
cycle, although with varying functional forms of the Case III – More Complex Load Functions – Extending
load. the analysis, the third case of more complex load func-
tions was considered. Figure 9 illustrates the different
Figure 2 shows typical experimental results from the tests shapes or profiles of the load that were used in the
on the prototype engine, which we attempted to match numerical model. These shapes were chosen to repre-
with the simulations [12]. sent the cases of high load near the extremes of the pis-
ton stroke, or highest load in the center of the stroke. The
CASE I – Constant Load Simulation – The first case sim- case when the load is concentrated towards the end of
ulated was one in which the applied load of 130N was the stroke corresponds to greater values of the shape
taken to be constant throughout the stroke (this being the factor k (where k varies from 0 to 3.55).
simplest case). The same P-V diagram (peak pressure
and integrated area) was obtained for the experiments as From the data obtained from the simulation (Figure 10), it
for the numerical simulation with a constant load but with can be observed that the in-cylinder peak pressure varies
long combustion durations (Figure 3). From Figure 4, it markedly for the 3 different load profiles chosen. The fre-
can be observed that for the same load and for decreas- quency variation is, however, very small so that it can be
ing values of combustion duration, the peak pressure concluded that the shape of the load does not influence
increases, while not affecting the overall shape of the P-V the engine speed greatly. From Figure 11, it can be
diagram to any great extent. Also it can be observed that observed that the velocity profile changes fairly
the higher the peak pressure, the higher the engine significantly with shape of the load. In this case the veloc-
speed, or frequency of reciprocation (Figure 4). A further ity profile is relatively constant in the middle portion of the
important observation is that the length of the stroke var- stroke, while the peak pressure increases as the shape
ies in proportion with the engine speed (Figure 5), with factor k increases. It was mentioned before that the
higher engine speeds corresponding in longer strokes higher the peak pressure, the higher the frequency of
due to the greater inertial forces developed by the piston reciprocation of the engine and the longer the stroke.
assembly.

8
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

60 00

50 00

40 00
Pressure [kPa]

30 00

20 00

10 00

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

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

6000

C d=0.00585 s
5000 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 30 35 40 45 50
D isplacem ent [m m ]

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 (cor- of a prototype unit. The simulations were used to show
responding to lower values of the shape factor k), the the effect of the total heat input, the combustion duration,
velocity profile rotates clockwise in the velocity-displace- the reciprocating mass and the load on the operation of
ment domain. In this case the peak pressure has a lower this engine-alternator combination. The experimental
value and the frequency and the engine stroke decrease testing performed previously on the linear engine [12]
correspondingly. showed that the linear alternator introduces a load that
has a roughly sinusoidal shape throughout the stroke –
CONCLUSIONS here considered to be of second order (corresponding to
k=0 in Figure 9). Using this load profile, a sensitivity anal-
A numerical model of a spark-ignited two-stroke cycle lin- ysis was performed to determine the influence upon the
ear engine-alternator combination has been developed engine operation of variation in one or more parameters.
and validated using experimental data from the operation

9
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

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

C d=0 .0 04 85 s
12 00 0 fre q=3 4.0 H z

10 00 0 C d=0 .0 05 85 s
fre q=2 9.2 H z
Pressure [kPa]

80 00

60 00

40 00

20 00

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

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

C d=0.00585s
2

1
Velocity [m/s]

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

-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 influ- engine stroke both increase. The frequency varies in an
ences the peak pressure, the frequency of the engine, inverse proportional relationship with the mass of the
and also the stroke length (Figures 12, 13 and 14). Varia- shaft.
tion in the combustion duration also influences the peak
A further observation from the numerical simulation is
pressure, the frequency of the engine, and the displace-
that, for very low piston assembly masses, the operation
ment in the same manner as for the heat input variation.
of the linear engine becomes possible only if the heat
By varying the mass of the moving piston assembly, the input is increased significantly, due to the low inertial
peak pressure and maximum displacement, vary propor- forces associated with low piston speeds.
tionally. For a greater mass, the peak pressure and the

10
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

60 00

50 00

40 00

C d = 0 .0 0 5 8 5 s
Pressure [kPa]

fre q = 3 0 .6 H z
H e a t in p u t Q = 2 5 J /s t ro k e
30 00 T h e s p a rk o c c u rs a t X = 4 5 .3 5 m m

20 00

10 00

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

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

60 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
50 00 Q = 25 J/stroke

C d=0 .0 05 85 s
fre q=3 0.6 H z
40 00 Lo ad =co nstan t
F riction = co nstan t
Pressure [kPa]

Q = 25 J/stroke

30 00

20 00

10 00

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

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 ACKNOWLEDGMENTS

displacement and the velocity versus displacement
respectively, for different values of the shaft mass. The The support of the Department of Defense in funding this
results of the parametric analysis are shown in Table 4. research is acknowledged [Grant No. DAAH04-96-1-
0328].
FUTURE WORK – Using the results of this numerical
simulation, a second prototype linear engine-alternator REFERENCES
combination is currently under development. This unit,
which will employ compression ignition with a state-of- 1. Cleveland Diesel, “History and description of the free pis-
the-art direct injection diesel fueling system, is designed ton engine-gas turbine power”, year unkown.
to produce around 10 kW of electrical power, and will be 2. A.F. Underwood,”GMR 4-4 Hyprex Free Piston Turbine
integrated into a series hybrid electric vehicle in a dem- Engine”, SAE Journal, June 1956, pp. 60-66.
onstration project.

11
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

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

2
Vel
oci
ty 1
[m/
s]
0
0 5 10 15 20 25 30 35 40 45 50

-1

-2

-3

-4 Triangular shape of
the Resultant Force of the
Load and the Friction
-5 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 k=0
ad 400
[N
] k=1.7
300

200

100

0
0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 0.045 0.05

-100
Displacement [m]

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

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

12
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

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 42 44 46 48 50
D is p lac em en t [m m ]

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 k=1.7
freq=29.7 H z
4

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

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

-1

-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 13. P.G. Blair, “Design and Simulation of Two-Stroke Engines”,
Generator Technology Development,” Final Report, Under SAE Inc., Warrendale, Pa., 1996
Contract to U.S. Army TARDEC, Mobility Technology Cen- 14. J. B. Heywood, “Internal Combustion Engine Fundamen-
ter-Belvoir, Fort Belvoir, Virginia, January 1995. tals”, John Wiley and Sons, New York, 1986.
11. N. N. Clark, T. I. McDaniel, R. J. Atkinson, S. Nandkumar, 15. S. Nandkumar, “Modeling of a Linear Engine”, MSME The-
C. M. Atkinson, S. Petreanu and P. Famouri, " Modeling and sis, West Virginia University, 1998
Development of a Linear Engine", 1998 Spring Technical
16. P. A. J. Achten, “A Review of Free-Piston Engine Con-
Conference, ASME ICE Division, Fort Lauderdale, FL.
cepts”, SAE 941776, 1994.
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
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

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 Q = 25 .8 J/stro ke
fre q=3 3.4 H z
Pressure [kPa]

60 00

40 00

20 00

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

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 42 44 46 48 50
D is p la ce m e n t [m m ]

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

14
 Downloaded from SAE International by University of Auckland, Sunday, August 05, 2018

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

1
Velocity [m/s]

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

-1

-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.
↑ denotes proportional dependence PARAMETER VARIED
↓ denotes inverse proportionality
Heat Input Combustion Frictional Mass of
Duration Force (Load) Reciprocating
Assembly

EFFECT Frequency ↑ ↓ ↓ ↓
SEEN Peak pressure ↑ ↓ ↓ ↑

Velocity ↑ ↓ ↓ ↓

Displacement ↑ ↓ ↓ ↑

CONTACT Phone: (304) 293-4111 ext. 333

catkinson@[Link]
Dr. Chris Atkinson, Dept. of Mechanical and Aerospace
Engineering, West Virginia University, PO Box 6106, Mor-
gantown, WV 26506, USA

15