Development and Field Test of a SOLO 161 Stirling Engine based

Micro-CHP unit with Ultra-Low Emissions

Magnus Pålsson

Dep’t of Heat & Power Engineering, Lund University

P.O. Box 118, S-221 00 Lund, Sweden

[e-mail]

http://burn.at/stirling

Abstract

For the last decade, work has been made at Lund University, Sweden, on developing a new sort of
natural gas combustion chamber for the V160/SOLO161 Stirling. It is a lean premix combustion
chamber with internal combustion gas recirculation and a metallic flame holder for flame stabilisation,
and it has produced extremely low emissions that are comparable to the emissions of catalytic
combustion. The combustion chamber is considered ready for market introduction by the end of 2001.

To combine the task of adapting this combustor for the market with the need to demonstrate small-
scale Stirling engine CHP technology, a project has been started with the purpose to demonstrate and
evaluate the operation of a Stirling engine unit based on the SOLO 161 Stirling engine equipped with
the Lund combustion chamber. The evaluation program should give information regarding operation
costs, efficiencies, emissions and running characteristics.

In November 2000 the engine was transferred from the Lund University laboratory to its final location
in Gothenburg. The engine is now installed and all necessary adaptation of engine, gas system and
water heating system is made. The unit is running unattended in normal everyday operation. Current
operating time is approx. 1200 hours, and delivered electric output is approx. 6000 kWh (July 2001).

Introduction

Decentralised combined heat and power (CHP) is an area that gets more and more attention. Gas
engine-based decentralised CHP have been installed in sizes of approximately 0.5 to 10MW. Other
sizes occur, both bigger and smaller, but the main part of the units installed is within this power range.
As further experience of installing and running these units is made, the interest in making use of
smaller heat demands as well, e.g. schools, apartment blocks, sports centres and even private houses,
is increasing.

The Stirling engine technique has been developed over several years and today there are a number of
different models and sizes. What they all have in common is that they are in a phase of development.
Thus they have not been demonstrated in a greater extent as small-scale power heating units. A brief
summary of the present situation with regard to the Stirling engine is given by the Swedish Gas Center
in [Ref. 3]. To this should be added the work that Gasunie does to adapt a Whispertech 800 W Stirling
engine unit to the market for use in small houses as a CHP unit, and also Ocean Power
Corporation/Sigma Elektroteknisk’s plan of developing and introducing a 3kW unit to the market.

At Lund University, Sweden, the Stirling engine competence is solid and work is constantly made on
developing combustion systems for Stirling engines. During the last decade extensive research has
been made on a lean premix prevaporized combustion concept with recirculation of the inert burnt gas
and a metallic flame holder. From this concept a new lean premixed natural gas combustion chamber
with internal combustion gas recirculation (CGR) has been developed for the V160/SOLO 161 Stirling
engines. This combustor has ultra-low emission levels, comparable to those of catalytic combustion.

At the start of the current project the combustor was ready to be adapted for production, with expected
market introduction in 2001. In Table 1 below the emissions from the old V160 swirling diffusion
combustion chamber are compared with the design goals for the new combustion system.

“Old” V160 combustor

“New” SOLO161 combustor (design goal)

HC

< 1 ppm

< 1 ppm

NO

x

220 – 650 ppm

< 10 ppm

CO

150 – 800 ppm

30 – 50 ppm

Table 1 – Emissions (design goal) for the new combustor compared with the old combustor.

Project goal

The purpose of this project is to demonstrate and evaluate the operation of a Stirling engine unit based
on the V160/SOLO 161 engines equipped with a combustion chamber with extremely low emissions
as described in the last paragraph.

The project can be divided into three different tasks:
• Selection of field test location and of operation strategy

• Installation in laboratory and laboratory testing

• Running/demonstration/evaluation of the CHP-unit in a commercial installation

Selection of field test location and of operation strategy

The unit had to be placed in an environment, which on one hand suits the engine power level and on
the other hand suits the unique running qualities that the unit possesses, i.e. low sound level, minimal
vibrations, low emissions.

Natural Gas

Furnace

Stirling Engine

Return
temperature

Supply (forward) temperature

ADDED

Ambient
outdoor
temperature

Secondary
hot water circuit

Figure 1 – Secondary hot-water circuit after installation of Stirling engine unit

A suitable location was established together with Göteborg Energi, the local utility company in
Göteborg, Sweden. The engine was placed in a small boiler station that supplies heat to a combined
office, workshop and warehouse in Marieholm, Göteborg. Furthermore, the building belongs to the
utility company. The supplied building has a total area of 2500 m

3

and the heat is supplied from a 250

kW hot water boiler. The yearly energy demand is 284 MWh and a SOLO 161 engine with an output
of 20 kW heat can cover approximately half of that heat demand.
A hot-water scheme like in Figure 1 was decided on, and the supply and return temperatures for this
scheme are shown in Figure 2. It was decided the Stirling engine unit was to run constantly at 12MPa
cycle pressure during the winter season when the heating demand is great, and run in on-off mode in
the summer season when the heat demand is less than the units heat output at 12MPa continuous
operation.

−20

−15

−10

−5

0

5

10

15

0

10

20

30

40

50

60

70

80

90

Outdoor temperature [

°

C]

Water temperature [

°

C]

Supply temp
Return temp, day
Return temp, night

Figure 2 – Secondary hot-water circuit temperatures

Installation in laboratory and laboratory testing

SOLO Kleinmotoren and Intersol have supplied a SOLO161 Stirling engine unit to the project. The
unit was first installed in the Lund University (LU) laboratories. LU together with Intersol have
designed and installed the combustion and heater system for the engine. The combustion system is
based on a LPP combustion system that has been developed at LU [Ref. 1, 2]. The engine was
provided with equipment necessary for the supervision and computer communication of the project.
Initial tests along different development lines were made on the Lund University United Stirling
V160F laboratory engine, which was equipped with an experimental burner (Figure 3). At this stage
tests concerning general combustor layout were made, including tests with different flame holders,
with varied mixing tube length, with varied number of air nozzles, with varied air nozzle diameter and
with different methods of natural gas injection [Ref. 2]. Also, tests aimed at lowering system

stirling
engine
heater

flame
holder

fuel

cooled

exhaust gases

pre-
heated

air

inlet
air

combustion gases

CGR

hot combustion
gases

flow guide
body

mixing tube

preheater

Figure 3 – Schematic view of the Lund V160 experimental combustion chamber

Figure 4 – The new combustion chamber mounted on the SOLO 161 unit.

pressure losses were made, including tests with varied lambda and combustion gas recirculation
(CGR) rates, so that the standard SOLO161 air blower could be used, and tests aiming at adapting the
burner from the lab 3.6 bar natural gas grid absolute pressure to a natural gas pressure of 0.1 bar. The
evolution of a suitable start sequence was initiated. When the general combustor design was decided
on, a prototype combustor was designed and manufactured for the SOLO 161 engine by Intersol. The
SOLO 161 engine and control system hardware (Figure 4) was adapted to the new prototype
combustor, which was fitted to the engine for further tests (Figure 4). The combustion system tests led
to the decision to re-design the flame holder and its internal support. However, this was the only
change made from the original prototype design.

Thermocouple

Helium temperature

Thermocouple

Mixtube temperature

Pressure transducer

Cycle pressure

RPM counter

Oil pressure guard

Coolant temperature

guard

Electronic control unit

Main switch

generator

Pressure transducer

Bottle pressure

Ignition coil

Supply valve

Helium

Dump valve

Helium

Air blower

Spark plug

380 V
50 Hz

Secondary

main NG

valve

Choke valve

NG

NG pressure

regulator

Combustion

chamber

Primary main

NG valve

Electr(on)ic
Air
NG

Figure 5 – Engine and combustion control hardware - schematic

1.2

1.3

1.4

1.5

1.6

1.7

1.8

0

5

10

15

20

25

30

SOLO 161, Lund combustor, p

Cycle

= 120 bar (001103

1

)

λ

(calculated from exhaust composition)

C

3

H

8

[ppm]

NO

x

[ppm]

CO/100 [ppm]

Figure 6 - Combustor emissions for the unit at full load, measured just prior to delivery to Göteborg.

A suitable control strategy had to be determined, along with the selection of an “exact enough” fuel
control valve. In the end a Kromschröder air/gas ratio control was chosen, that regulates inlet gas
pressure to the same pressure as that of the inlet air (c.f. Figure 5, NG pressure regulator). An on-off
gas valve (c.f. Figure 5, Choke valve) in series with a restricting orifice was mounted on a parallel fuel
line for cold start gas flow control. Final cold and hot start-up control sequences were decided on and
programmed to the electronic control unit’s EPROM memory, see also [Ref. 2].

At the end of the laboratory test period the engine was mechanically inspected and the PL seals were
renewed before it was transported to the commercial site in Göteborg, in order to ascertain that the
engine was in an “as-good-as-new” condition at delivery.

Also, emission tests at full load for different

lambda values were made for comparison with later field test results (Figure 6). The development and
testing of the combustion system is described further in [Ref. 2], presented at this conference.

Figure 7 – Installation of the SOLO161 unit in a small boiler station in Marieholm, Göteborg.

Running and demonstration of the CHP-unit in a commercial installation

The engine was moved to its chosen location in Marieholm, Göteborg, early in November, 2000
(Figure 7). The goal was to operate the engine during the heating season Dec 2000 - Dec 2001 in order
to follow the behaviour of the engine in a commercial application. Electricity is delivered to the local
grid or consumed internally in the building. Göteborg Energi with sub-contractors have been
responsible for supplying facilities for data collection and remote monitoring. The measuring program
in Göteborg is focused on parameters that relates to the operation and maintenance costs for the
engine, but also emission measurements are scheduled for the late summer plus winter 2001.
Continuous data acquisition is made of the following parameters and data are available online in real
time (Figure 8):
• Accumulated gas consumption (VP1-GM2)
• Instantaneous electric power

• Accumulated electric energy consumption (from grid) (SM1-EM2)

• Accumulated electric energy generation (to grid) (SM1-EM1)

• Instantaneous heat power

• Accumulated heat delivery (to hot-water grid) (VP1-EM2)

• Total accumulated heat consumption of building inc. heat delivered from boiler (VP1-EM1)

• Temperature of water to and from engine (GT4-2/GT4-1)
• Accumulated running hours of engine

• (External hot-water grid) control parameters of Stirling engine

• Ambient room temperature (LB05-GT5-1)

• Ambient outdoor temperature (VP1-GT-UTE)
Abbreviations in parenthesis refer to meters indicated in figure 8. The monitoring system also allows
plotting of trend charts, see figure 8 - small window.

Figure 8 – Online data monitoring system, presentation window view

(INU-vision v 2.3A by Honeywell INUcontrol AB, Sweden).

Evaluation of the CHP-unit

The evaluation program will result in documented data of the following subjects:
• Operation costs

• Efficiencies

• Emissions (sound/vibrations/exhausts)

• Running characteristics (controllability/availability/supervision/maintenance)

The unit will be evaluated by Sydkraft/Sycon AB. At two different occasions a mobile high precision
emission measuring equipment will be used to measure emissions. Efficiencies will be measured
online. Operation costs and running characteristics will be evaluated at the completion of the project in
springtime 2002.

Results and experiences so far

One of the main experiences is that the installation and certification procedures take more time than
anticipated, and that there is immense paperwork involved. However, the work done can be of use in
future projects. The education of the maintenance personnel involved in the daily inspections of the
unit might be an important factor for making engine operation run smoothly.
A lot of problems were caused by interference between the “external” secondary hot-water grid
control including control of the furnace, and the “internal” Stirling engine unit control. Badly scaled
control parameters in the “external” control system resulted in unsuitable control signals being sent to
the Stirling internal control system. E.g. the outer system tried to run the Stirling engine outside its
operating range, resulting in shutdown of the Stirling unit. A well-defined interface between the
Stirling embedded control system and outer hot-water grid control system is needed. However, the
control situation has been improved greatly by reprogramming of the outer control system. Another
solution could be to equip the Stirling engine with primary hot-water circuit temperature control.
The operation and maintenance of the unit would be much simplified if the Stirling engine’s control-
and errorcodes could be monitored and reset online.

Apart from the above, the following operation disturbances were experienced:
• Boiling of engine cooling water (primary hot-water circuit) was experienced due to a

malfunctioning water pump in the secondary hot-water circuit, and by badly scaled control
parameters for the outer control system (see above). These problems have been solved.

• A water-to-atmosphere o-ring for the sliding gas cooler had to be exchanged, either it was

damaged by insufficient cooling during the incidents above, or it was damaged during engine
assembling. The engine has an old design of the sliding gas cooler which makes it susceptible to
o-ring malfunction. A new improved design is already implemented in later versions.

• The coolant water reservoir vessel had to be exchanged. In later versions of the engine a new

improved design is used.

• A gas tube leakage in the helium return tube (from the engine’s compressor to its bottle), due to

fatigue in a brazing joint. The tube was exchanged and the design was improved to minimise the
force acting on the joint.

• A helium leakage from inside the cycle. The leakage was probably due to a pore in the cast

compressor head, however the reason for the leakage is not yet analysed. The compressor head
was exchanged, and at the same time the sliding gas cooler was exchanged against one having the
new improved design. After the change the leakage was gone.

No running problems related to the combustion system have been experienced, apart from the
exchange of a spark plug and necessary initial adjustments of the control system software.

Most of the above mentioned malfunctions are “first-installation” specific and will not be experienced
in a new unit installation of the same kind. The new combustor has been working fine, as has its
control system after initial adjustments. The operation strategy was changed from on-off operation to
variable cycle-pressure operation with good result. The engine is now operated with a cycle pressure
varying between 5.5Mpa and 12Mpa.
At the time of writing (July 2001) the total operating time in Göteborg is about 1200 hours, and
accumulated electric output is about 6000 kWh.

Continuation

Some details would gain from some extra work put into them when the current project ends. The
control system and software can be adapted for shorter start-up time; at present there is no use for
quick starts so the start-up procedure has deliberately been made relatively slow to spare the engine’s
hot parts from thermal stresses and to ensure reliable engine operation with a new system. However,
during lab tests with the new combustion system fast starts at high thermal power were tried without
start problems. The spark plug seams to have an unsatisfactory life length and can be replaced by a
single electrode. This might lead to increased electronic interference with the control system.
There are also some very interesting observations that can be made from project measurement results.
At stoichiometric and slightly rich conditions emission are extremely low, with the exception of
carbon monoxide [Ref. 2]. Theoretically, system efficiency increases with decreasing

λ, but at close to

stoichiometric conditions (

λ=1) carbon monoxide production will lead to poor combustion efficiency,

which in turn will lower system efficiency. However, with an oxidizing catalytic converter carbon
monoxide emission levels can be minimized. If the catalytic converter is mounted at the exhaust side
of the preheater, the heat generated in the catalytic conversion of CO to CO

2

will be wasted as it will

end up in the exhaust gas (if the converter reaches light-off temperature at all, that is). To reach
highest possible system efficiency the catalytic converter should be situated inside the combustor, just
upstream of the preheater or just downstream the Stirling heater where the temperature is around
800ºC which is a suitable operating temperature for a catalytic converter. Then most of the heat
released in the catalytic converter can be heat-exchanged in the preheater to the inlet air and thus be
used by the engine. In the case of fuel-rich combustion (

λ<1), additional air is needed for complete

oxidation of CO to CO

2

. An extra air inlet can be mounted upstream of the catalyst to supply this air.

An extra benefit of running at stoichiometric conditions is that an ordinary car lambda (O

2

) sensor can

be used for combustor air/fuel control. Such sensors are cheap and their use would simplify Stirling
engine combustor control considerably.
In a continued project, the combustion chamber can be adapted and tested for operation at close to
stoichiometric conditions. It is the firm belief of the current author that such a combustion chamber
will have single-digit ppm emissions, as well as giving high system efficiency. Measures are taken to
find financing for a project continuation.

Project Partners

Project partners are (* denotes active partners):

Intersol* (Sweden)
Sydkraft AB/Sydgas* (Sweden)
Lund University, (Div. of Combustion Engines)* (Sweden)
Göteborg Energi* (Sweden)
SGC - Swedish Gas Center* (Sweden)
Vattenfall Naturgas (Sweden)
Helsingborg Energi (Sweden)
Lunds Energi (Sweden)
Stockholm Energi Gas (Sweden)
STEM - Swedish National Energy Administration (Sweden)
SOLO Kleinmotoren* (Germany)
DGC - Danish Gas Technology Center (Denmark)
EnergieNed (The Netherlands)

References

1. Magnus Pålsson, “Design and Testing of Stirling Engine Premix CGR Combustor for Ultra Low

Emissions”, ISEC97001, Ancona, Italy, 1997

2. Magnus Pålsson, “Development of a LPP CGR Combustion System with Ultra-Low Emissions for

a SOLO161 Stirling Engine based Micro-CHP unit”, submitted and accepted for the 10

th

ISEC,

Osnabrück, Germany, Sept 24-26, 2001

3. Tomas Nilsson, “Mikrokraftvärmeverk med Stirlingmotor”, SGC (Swedish Gas Center) report

080, Jan 1997, ISSN 1102-7371. Written in Swedish.