Heinrich J.G., Aldinger F. (Eds.) Ceramic Materials and Components for Engines

Подождите немного. Документ загружается.

PLANAR LIMITING CURRENT

OXYGEN SENSORS

For lean burn applications (GDI, diesel engines, CNG

engines, gas burner) the Nernst measuring principle

with insufficient lambda-range is replaced by an

amperometric principle

[

18, 191.

I---it---siF--------1

Principle

Function

MftusiDn

cathode

Arode

&+4e--..2P

2P-Q+48-

=axrpI.Dc&-co,

04:

Cfdfusbn

me(iiciant

cq:"""bno14

Fig.

Principle of limiting current oxygen sensor

Ic-----X----.CI

At elevated temperatures

(>600

"C)

an active transport

(pumping) of oxygen ions can take place through the

solid electrolyte

ZrO,

ceramic. By applying of an

external pumping voltage (Up) on the electrodes

oxygen ions

(02?

are pumped from the cathode to the

anode (fig.4). The electrical current corresponds to the

oxygen ion current and therefore to the current of

oxygen molecules diffusing to the cathode in the

exhaust gas. With increasing pumping voltage (Up) the

current increases according to the internal cell

resistance. A diffusion barrier in front of the cathode

impedes the

flow

of oxygen molecules to the electrode,

the oxygen concentration at the cathode decreases till

zero and results in a current saturation (IL) beyond a

certain pumping voltage threshold (fig.5).

The resulting limiting current is roughly proportional to

the exhaust gas oxygen concentration (coz):

(1)

=4FD--c,,

Faraday's constant, D

diffusion coefficient,

effective diffusion cross section,

effective diffusion

length).

0.5

1.5

0.5

1.5

8.36

20.9

pumpins

wage

Fig.

Characteristic of limiting current oxygen sensor

Depending on the porosity of the diffusion barrier the

difhsion current is a combination of gas phase and

Knudsen diffusion with different dependences on

temperature and exhaust gas pressure

[20].

This simple

single-cell limiting current oxygen sensor is essentially

used in exclusively lean exhaust gas applications

[8].

A single cell type universal oxygen sensor with a

distinctive measuring range in the rich region can be

realized by applying the anode to the reference air. Such

a sensor is in series production from a Japanese

manufacturer

a conventional thimble type universal

oxygen sensor

[21].

But the measuring range in rich

exhaust gas is limited due to the limited gas supply

from the reference air side. Further disadvantages of

this sensor type are high power consumption and

limitation of the light off time due to the thimble type

construction.

DUAL CELL

PLANAR

UNIVERSAL WIDE

RANGE OXYGEN SENSOR LSU

For the application in the whole lambda range from lean

to rich a universal dual cell wide range oxygen sensor is

preferred

[8,

9, 171.

This sensor consists of

two

cells, a

pumping and a sensing cell separated from each other

by a porous layer with a gap width of

which acts as a diffusion barrier (fig.

6).

3.0

1.0

-1

-2.0

071.0

1.6

1.9

Notmalized

ratio

Fig.

Planar wide range oxygen sensor LSU4 (Bosch)

Depending on the polarity of the applied voltage on the

inner and outer

pumping electrode, oxygen can be

pumped out or into the gap. The sensing cell, consisting

of a sensing electrode inside the gap

and a reference

electrode in a reference air duct, is a Nemst type solid

electrolyte cell measuring the lambda value in the gap.

By regulating the pumping voltage by an electronic

closed loop control circuit a constant lambda value of

the gas in the gap is maintained, corresponding to a

Nernst voltage of the sensing cell of Umf

450 mV. In

lean exhaust gas

is pumped out of, in rich exhaust

gas,

is pumped into the gap which leads to a positive

or negative pumping current depending on the Lambda

value of the exhaust gas. At

h=l

no oxygen needs to be

pumped in any direction and therefore no pumping

current is measured. As for the limiting current sensor

the pumping current is direct proportional to the oxygen

concentration in the lean region (or oxygen requirement

the rich region). The signal characteristic

continuous over the whole lambda range. The sensor

needs an operating temperature of

-700...800

"C.

special electronic control circuit which is integrated in

an ASIC is necessary for operation. Because of the

integrated heater this sensor offers fast light off

behavior

(45

s).

The range of application varies from

slope m of SrTio,6sFeo,3s03 is

1/5

in this oxygen partial

pressure range.

SrTi,-,Fe,O, is chemically stable in the temperature and

partial pressure range which is represented in figure

The ceramics (d

500

pm) respond to a sudden oxygen

partial pressure change within a time constant of a few

seconds, which is in the same range

slightly acceptor

doped SrTiO,

[28],

where the kinetic behavior is also

determined by the diffusion of oxygen vacancies.

In the following, it will be shown that the temperature

independence and the slope m of the characteristic

curve of the ceramics (x

0.35) can be transferred to

thick films of the same material, as shown in figure

Fig.

SrTio,6,Feo,3,03 thick film sensor on a Al,O,

substrate with Pt bottom side contacts.

square

mm2.

For the preparation of SrTio,,,Feo,,,O, thick films a

screen printing paste prepared of the same powders as

above was printed onto an alumina substrate

(96

%).

The thick films were dried and fired at 1050 "C. The

resulting layer thickness was

pm, with grain sizes of

0.5

pm and an open porosity of

All examined

thick films were equipped with Pt bottom side contacts,

that were bumed in at

1300

"C.

Sensors manufactured that way show very short

response times due to their small layer thickness

pm). The response times of an oxygen thick film

sensor (d

15 pm) are shown in figure

a function

of temperature. The fast kinetic behavior (response

times) of the thick films were investigated by a method

developed by Tragut

[29,

301.

The total air gas pressure

and the corresponding oxygen partial pressures p02

were changed periodically from

2.104

4.104

Pa. The

modulation frequency f can be varied from f

0.02

kHz. The magnitude IAl and phase

of the sensor

signal (conductivity) are determined in dependence on

the simultaneously measured pressure modulation. This

allows the evaluation of sensor kinetics in the frequency

domain. From the magnitude a limiting 3

dE3

frequency

fg can be determined which serves

basis for

conversion into the response time

1/(3fJ. Unlike

temperature independence

conductivity, the response

time shows a dependence on the sensor temperature.

This kinetic behavior of the Sr(Tio,65Feo,35)03 thick film

is comparable with the short response times of very fast

oxygen sensors based on thin film technique

11.

,o~~

11po lop0

970

oo]

thin

film

porous

thick

film

SrTiO,

SrFe0.35Ti0.6503

*---

CeO,

0.70

0.80

0.90

.oo

1000

tvr

Fig.

10.

Response times of a Sr(Tio,6sFeo,3s)03 thick film

pm, grain size

0.5

pm) compared with fast thin

film sensors (d

pm).

Figure

shows a measurement result of the thick film

sensor under lean-bum conditions in a car engine

(Daimler-Benz,

2.0

cylinder) with preceding fuel

injection. The sensor was heated by a printed Pt

substrate heater to its working temperature. It was

protected only by a metallic housing with bores. It was

mounted a few centimeters behind the outlet valve of

the engine in the exhaust gas stream. Because of this

installation close to the engine the sensor detects only

one single cylinder.

The aidfuel ratio was changed via the injection duration

(4.2

5.1

ms) stepwise from

h=1.4

to h=l.l. The

sensor was operated at

via a serial resistor of

3.9

kn.

This leads to a sensor signal between 1

and

(sensor resistance:

1.2

1.4).

Considering the

delayed outflow process, the sensor can detect the very

fast PO,-changes on account of the varied injection

times ti.

fuel

injection cycle

1.4

1.2-7

=1.4

1,42

1,l

=1,1

.n!

I."

100

150

200

Fig.

11.

Profile of the sensor signal after change of the

fuel injection times

of a car engine. The dashed line

shows the injection process. On account of the delayed

outflow process, the sensor signal is time-shifted.

Exhaust gas temperature:

820

engine speed:

3500 r.p.m.

CONCLUSIONS

ZrO, based oxygen sensors have been proved for a long

time as reliable sensors for engine control systems. For

lean burn engines universal wide range oxygen sensors

ZrO,

planar technology are preferred in current series

applications. Temperature independent resistive type

oxygen sensors as presented might be a future

alternative for distinctive applications in lean exhaust

gases and offer high performance for fast lambda

control.

ACKNOWLEDGMENTS

We would like to thank the Institut fir

Kolbenmaschinen of the Universitiit Karlsruhe (TH) for

the measurements on the car engine, and the

Keramikverbund Karlsruhe Stuttgart

(KKS)

of the State

Government of Baden-Wiirttemberg for the financial

support.

REFERENCES

[

Bosch: Kraftfahrtechnisches Taschenbuch 23,

Auflage, Vieweg-Verlag

(

1999).

[2]

Antonius, A. Gamer,

Garrett, Honda first to

have gasoline engine verified at ULEV exhaust

levels, Honda Press Release, Jan. (1995).

[3]

K. Winkler, M. Kusell, E. Schnaibel, W. Strehlau,

Gobel, J. Hohne, W. Muller, The Development

of an aftertreatment System for Gasoline Direct

Injection Passenger Cars, F98T2 18, FISITA

World Automotive Congress Proceedings, Sept.

[4] M. Kiisell, W. Moser, M. Philipp, Motronic MED7

for Gasoline Direct Injection Engines: Engine

Management System and Calibration Procedures,

SAE 1999-01-1284, Detroit (1999).

[5]

Sojima and

Mase, Multi-Layered Zirconia

Oxygen Sensor for Lean Bum Engine Application,

SAE 850378, Detroit (1985).

[6] T. Takeuchi, Oxygen Sensors, Sensors and

Actuators, 14, 109-124 (1988).

[7] T. Yamada, N. Hayakawa, Y. Kami, T. Kawai,

Universal Air-Fuel Ratio Heated Exhaust Gas

Oxygen Sensor and Further Applications, SAE

920234, Detroit (1992).

[8] H.-M. Wiedenmann, G. Hotzel, H. Neumann, J.

Riegel, and H. Weyl, Exhaust Gas Sensors, in

Automotive electronics Handbook, edited by

Ronald K. Jurgen, McGraw Hill Inc. (1995).

[9]

Riegel,

Hotzel, H. Neumann, Advanced

Electrochemical Exhaust Gas Sensors for

Automotive Application, Proc. 44" Meeting

the

International Society of Electrochemistry, Berlin,

Sept. 5-10 (1993).

Science, Proceedings 1, Westerville, Ohio, 28 1

301 (1980).

(1992).

Oct. 1 (1 998) Paris.

[

101 E.M. Logothetis in Ceramic Engineering

[l 11 P.T. Moseley, Sensors and Actuators B 6, 149-156

[

12) K. Park, E.M. Logothetis, J. Electrochem. SOC.:

Solid-state Science and Technology 124 (9),

1443-46 (1977).

[13] C. Yu, Y. Shimizu, H. Arai, Chemistry Letters,

[

141 P.T. Moseley, D.E. Williams, Polyhedron 8

[15] D.E. Williams, B.C. Tofield, P. McGeehin,

European Patent Specification EP0062994 (1 982).

[16] R. E. Mistler, D. J. Shanefield, R. B. Runk, Foil

Casting of Ceramics, in Ceramic Processing before

Firing, edited by

Y. Onada and L. L. Hench,

John Wiley and Sons, Inc., New York ,411-448

(1978).

[

171 H. Neumann, G. Hotzel, G. Lindemann, Advanced

Planar Oxygen Sensors for Future Emission

Control Strategies,

SAE

970459, Detroit (1997).

Electrolyte Oxygen Sensors, Solid State Ionics 6,

563-566 (1986).

(13/14), 1615-18 (1989).

[

181 H. Dietz, Gas-Diffusion-Controlled Solid-

175-183 (1982).

[

191 H.-M. Wiedenmann, Characteristics of Oxygen

Sensors for Lean Exhaust Gas, VDI-Berichte 578,

129- 15 1, VDI-Verlag, Dusseldorf (1 985)

[20] K. Saji, Characteristics of Limiting Current Type

Oxygen Sensors,

J. Electrochem. SOC.:

Electrochemical Science and Technology,

Vo1.134, No.10, Oct. (1987).

[21] K.Mizusawa, K. Katoh,

Hamaguchi, H.

Hayashi,

Hocho, Development of Air Fuel

Ratio Sensor for 1997 Model1 Year LEV Vehicle,

SAE

970843, Detroit (1997).

HZirdtl, Solubility of Lanthanum in Strontium

Titanate in Oxygen-Rich Atmospheres,

Mater.

Sci. 32 4247-52 (1997).

Norby, J. Phys. Chem. Solids 58 (6), 969-76

(1997).

[24] R. Moos, K.H. HZirdtl, J. Am. Ceram. SOC. 80 (lo),

[25] W. Menesklou, H.-J. Schreiner, K.H. Hiirdtl and

E. Ivers-Tiffke, Sensors and Actuators B 5912-3,

[26]

Menesklou, H.-J. Schreiner, R. Moos,

[22] R.

Moos,

T. Bischoff, W. Menesklou,

[23]

Steinsvik, R. Bugge, J. Gjonnes, J. Tafto,

2549-62 (1997).

184- 189 (1 999).

HZirdtl,

Ivers-Tiffke, Sr(Ti,Fe)O,: Materials for

temperature independent resistive oxygen sensors,

be published in

MRS

fall meeting proceedings,

Materials for Smart Systems 111, Volume 604

(1999).

[27] H.-J. Schreiner, Temperaturunabhhgige resistive

Sauerstoffsensoren auf der Basis von Sr(Ti,Fe)O,.

Ph.D. thesis, Karlsruhe (1999).

(1991).

Fortschr.-Ber. VDI 8 (291), Diisseldorf, Ph.D.

thesis

(

1992).

[28] R. Waser, J. Am. Ceram. SOC. 74 [8], 1934-40

[29] Ch. Tragut, Kinetik schneller Sauerstoffsensoren,

[30] Ch. Tragut, K.H. Hiirdtl, Sensors and Actuators B

[3 13 U. Lampe

J. Gerblinger, H. Meixner, Sensors and

4,425-429 (1991).

Actuators B 7,787-91 (1992).

CERAMIC GAS TURBINE "CGT302"

DEVELOPMENT SUMMARY

Tetsuo Tatsumi", Isashi Takehara, Yoshihiro Ichikawa

Kawasaki Heavy Industries Ltd., Gas 'hrbine Research

Development Center

673-8666

Akashi, Japan

'88

ABSTRACT

'89

'90

'91

'92

'93

'94

'95

'96

Ceraiic Coiponet!t Fab!icatior! T&ch!-iology

'97 '98

In Japan, a 300kW Ceramic Gas Turbine (CGT)

research and development program was begun in 1988

as a part of "the New Sunshine Project" promoted by

the Japanese Ministry of International Trade and

Industry. The development target was to demonstrate

the thermal efficiency of over 42% and low "Ox

emission at 1350 "C of turbine inlet temperature (TIT).

Kawasaki Heavy Industries, Ltd. (KHI) has been taking

a part of this project and developed a regenerative

two-shaft CGT, which was named CGT302.

We achieved 42.1% of thermal efficiency and

31.7

ppm

NOx

emission, and also we conducted 2,100 cumulating

operating hours at 1200 "C TIT, which is considered to

be reasonable temperature for commercial use. This

project was finished March 1999. For the next step

development, MITI started "Research and

Development on Practical Industrial Co-generation

Technology". The project objective is to encourage

prompt industrial applications of co-generation

technology by confirming its reliability and soundness

of Hybrid Gas Turbines (HGT).

Component Technology

Basic Design

900°C

MGT

1200°C

INTRODUCTION

(Turbine. Compressor, Heat Exchan

Interim AppraisLl

Basic CGT

1350°C

Pilot

CGT

Efficiency of gas turbines has been improved by

utilizing higher TIT. In the field of large sized gas

turbines, TIT has been steadily increased owing to

progress of high temperature resistant alloys and turbine

blade cooling technologies.

For small gas turbines, however, it

difficult to

manufacture air-cooled blades because of their size

limitations.

The purpose of using ceramics is to increase TIT

without cooling air by utilizing their superior heat

resistant characteristics.

New Energy and Industrial Technology Development

Organization (NEDO) started a CGT project in 1988

develop three different types of CGT, namely CGT301,

CGT302 and CGT303 under a contract with the Agency

of Industrial Science and Technology of MITI.

Kawasaki Heavy Industries Ltd. proposed the

development of CGT302, two-shaft CGT with

recuperator, in co-operation with Kyocera Corporation

and Sumitomo Precision Products.

DEVELOPMENT TARGET

The development target is shown in Table

This target

was set to 42% ambitiously to have comparable

efficiency with that of diesel engines expected at the end

of this project.

Table

Development target

Target

Thermal Efficiency

Turbine Inlet Temperature 1350 "C

Output Power

300 kW class

Satisfy the national

regulation

Exhaust Emission

DEVELOPMENT SCHEDULE

The development schedule is shown in Table 2.

The development phases were as follows.

(Ceramic gas turbine) Phase

Basic design

Phase 2: Basic MGT (900°C Metal gas turbine)

Phase

Basic CGT (1200°C ceramic gas turbine)

Phase

Pilot CGT (1350% ceramic gas turbine)

The each phase corresponds to the part load condition of

the final target of Phase 4, because the basic designs for

all phases are the same.

Table

Development schedule

ENGINE CONFIGURATION AND

DESIGN FEATURES

The design policy of CGT302 is as follows.

To adopt conventional component layout to

concentrate the development effort for ceramic

utilization avoiding other difficulties.

To adopt the metallic plate-fin type recuperator as a

heat exchanger instead of a ceramic rotary

regenerator which has been commonly used in the

CGT for automobile use.

To adopt module configuration to facilitate

assemble and disassemble at numbers of test

operation.

Composite-component structure ('hrbine nozzle)

The turbine nozzle is a large component consisting of

complicated shaped nozzle vanes and an inner and an

outer shroud. It is exposed in the highest temperature

among the turbine components in which a large thermal

stress may occur due to the temperature distribution on

it.

Accordingly, the segment configuration has been

adopted to reduce thermal stress. This method has the

merit of easy manufacturing but has demerits of stress

concentration at the joining position and gas leakage

between the segments.

a countermeasure for these

demerits, we developed composite components.

This concept is shown in Fig. 2.

The fabricating process of the composite components is

schematically shown in Fig.

CGT302 consist of single stage centrifugal compressor,

single stage axial gas generator turbine, single stage

axial power turbine, single can combustor, a plate fin

recuperator, speed reduction gearbox for output shaft

and other auxiliary equipment.

Cutaway view of CGT302 is shown in Fig.

Heat

Exchanger

Gas

Generator

/turbine

Compressor

Scroll

Combustor

Fig.

Cutaway view

CGT302

The main design features of the CGT302 are described

below.

ADVANCED UTILIZATION OF

CERAMICS

Si3N4, a candidate ceramic material, is much more

brittle than metals and its thermal expansion coefficient

is almost 1/4 of that of typical heat resistant alloys.

Considering these characteristics, we have developed

advanced utilization technologies for ceramics contrary

to the conventional component layout. These

technologies are as follows.

Composite-component structure

(Monolithic

FRC*

hybrid component)

*Fiber

Reinforced Composite

Stress-free elastic supporting structure

Stress-free independent supporting structure

Binding

Impregnation

(organwilimn

Polymer)

Conversion

Repeat

Coating

(To

prevent

separauon

fibers)

Composite Compoment

Fig.

Turbine nozzle concept Fig.

Fabrication process

We can obtain a composite component that has a

monolithic ceramic gas path with high heat resistance

and a FRC backup ring with high fracture toughness.

We applied this technology to the gas generator turbine

(GGT) nozzle and the power turbine

(PT)

nozzle. By

this technology, GGT and

nozzle segments can be

assembled tightly enough to seal gas leakage between

the nozzle segments and also the centering function can

be kept precise enough to minimize the turbine tip

clearance.

For example of this composite component technology,

the GGT turbine nozzle is shown in Fig.

Fig.

GGT turbine nozzle

(Composite component)

Stress-free structure

A stress-free elastic supporting structure and

independent supporting structure were developed to

protect ceramic components from excessive force

derived from collision of components and loss of

tightening force due to the thermal expansion difference

between ceramic and metallic components and

asymmetric deformation or expansion of metallic

supporting structures.

We did not adopt the conventional structure such as

stacking up the ceramic components. We designed to

support ceramic components independently from each

other and to keep appropriate clearance between them.

Accordingly, ceramic components can be displaced

freely avoiding interference of each other when the

metallic support is deformed asymmetrically.

Ceramic coil springs and wave rings are properly

utilized in the CGT302. The clearance between each

ceramic component is sealed by ceramic elastic

components such as coil springs or wave rings. These

ceramic elastic components are also used for absorbing

difference of thermal expansion between metallic

support and ceramic components. Fig.

shows the

configuration of the stress-free supporting structure.

eolL

SPRllBS

Fig.

Configuration of stress-free supporting

Combustor

The combustor is a rather simple shaped component

among the ceramic components for CGT, but it is

exposed directly in the combustion flame that has

extremely large temperature distribution along both the

radial and axial direction. The single can configuration

we adopted inevitably causes asymmetry of the casing.

Accordingly, careful design is required to support it

freely from the excessive force caused by the

interference of asymmetrical deformation of the

combustor casing and associated casings while there is

difference in thermal expansion between the metal

casing and the ceramic combustor.

We designed this component as a simple one-piece

cylindrical combustor liner to facilitate a reliable

support system and manufacturing. It is supported by

the coil springs arranged to push

against the scroll gas

inlet port. This configuration is shown in Fig.

Also

excellent low NOx emission characteristics of dry

low emission

(DLE)

combustion system can be

expected by adopting the ceramic liner in which the

maximum gas temperature decreases as the ceramic

does not need cooling air.

Fig.

Schematic Drawing of

DLE ceramic combustor

BLISK type turbine rotor

There are two types of ceramic turbine configuration.

One is "BLISK type that the blades and disk are made

in one-piece, and the other is hybrid type, which is made

by using a slotted metal disk and inserted ceramic

blades.

these two types have advantages and

disadvantages, we have to choose the better one

considering application requirements.

The BLISK type is difficult to make a large size turbine

and has to be joined with the metal shaft, but it has

better properties than the hybrid type in heat resistance,

oxidation resistance, strength-weight ratio and

manufacturing cost. We introduced a BLISK type

turbine rotor for a GGT that is relatively small. And we

introduced a hybrid type for

rotor that has large

diameter and could not be manufactured at the start

point of the development. According to the development

progress, manufacturing technology was improved and

BLISK type PT rotor has become able to be

manufactured. We changed PT design to BLISK from

hybrid.

Manufactured GGT and PT rotor are shown in Fig.

The outer diameter of this PT rotor is 192 mm and must

be the biggest BLISK ceramic rotor in the world.

Fig.

BLISK

tvDe

GGT

and

rotor

The turbine tip clearance greatly affects turbine

efficiency especially in these small size gas turbines.

Ceramic BLISK turbine is more suitable than hybrid or

conventional metal turbine to minimize the tip clearance,

because the coefficient of thermal expansion of

ceramics

low to minimize the tip clearance in various

operation conditions. BLISK turbine also contributes to

obtain better performance with no gas leakage between

the blades.

ENGINE TEST RESULTS

900

Metal GT

The MGT is manufactured as a test bed for the CGT and

was tested to confirm the performance ability and

propriety of the design concept of the CGT. At the

MGT design,

all ceramic components of the basic

design are replaced by metal parts that have the same

aerodynamic geometry as the CGT. The efficiency at

the initial stage was only 17.3% and 42 kW output at

913°C TIT, and the final stage performance was

improved to 23% efficiency and 59 kW output at 899°C

TIT. The performance improvement

shown in Fig.

compared with the later CGT development progress.

The cumulated operation time of MGT was

Hr and

the number of start stop was 282 times.

1200

BasicCGT

The high temperature metallic components in MGT

were replaced with ceramics for the 1200°C Basic CGT.

Ceramic material for Basic CGT was SN252

Si3N4.

Before installing ceramic components, we conducted

component tests such as the cyclic thermal-shock test

for stationary components and the cold and hot spin test

for turbine rotors.

Before the all-ceramic configuration test, we conducted

the confirmation test by installing single or minimum

ceramic component step by step.

We experienced unexpected obstacles through these

tests, but after appropriate countermeasures, we have

proved its durability through 40 hours of operation at

1200°C TIT, and also obtained 33% of thermal

efficiency and 164 kW at 1190°C

TIT

as shown in Fig.

13.

This

R&D

result of Metal GT and Basic CGT was

evaluated by the neutral committee of industrial

technology deliberation, and progressing to Pilot CGT

stage was decided.

1350

Pilot CGT

Fortunately the Basic CGT development was

successfully demonstrated and we could step forward to

Pilot CGT stage without a wholesale design change.

Ceramic material was changed to SN281 for GGT rotor

and SN282 for all other ceramic components from

SN252. The difference between SN281 and SN282 is

HIP process that is applied for SN281 only. And the

turbine tip speed at rated condition was reduced to 480

dsec from 570dsec as a countermeasure against FOD.

We tested Pilot CGT increasing TIT stepwise by

50°C

from 1200

at which successful operation was

confirmed by the Basic CGT, and in these each step we

confirmed stable operation and performance. To

operate the engine safely according to increased

TIT

through these steps, we needed some modifications such

as cooling of the joining part of GGT rotor and the

clearance of roller bearings in the hot section. After

these steps, we succeeded in

hrs of 1350°C

TIT

operation, but the target efficiency was not achieved.

To obtain further performance improvement, we

redesigned the compressor to increase the airflow rate to

get better matching to the turbine, because we found

that the turbine nozzle throat area was slightly larger

than the optimum value. We also improved the

manufacturing accuracy of the turbine nozzle throat area

because it had become evident from the engine test and

dimensional inspection that the throat area fluctuated in

each product and in each throat.

Also

we redesigned

the GGT rotor blade shape to match the reduced

rotational speed. But even after these efforts, the

efficiency was staying around

40%,

still slightly short

from the target.

To achieve the 42% target efficiency, improvements

such

a larger heat exchanger with increased stack

layer, an abradable

shroud to minimize the tip

clearance, and a

rotor with thinner blades for

improved turbine efficiency, and an improved

compressor, were implemented. After several attempts,

we achieved 42.1% at 1396 "C TIT with 322 kW and

41.5% at 1359 "C TIT with 302 kW in test #4-39. The

pilot CGT performance improvement is shown in Fig.

CGT302 Performance

Progress

Target bilot CGT)

I I

1050 1150 1250 13M 1450

Turbine Inlet Temp.(oC)

Fig.

Pilot CGT Performance Improvement

The performance comparison in the range of power

output and efficiency improvement history is shown in

Fig. 9 and Fig.

10,

respectively.

10'

10' 1

. .

....

Fig.

CGT302 Efficiency Compared with

Conventional GT

The achieved efficiencies are comparable or superior to

those of the largest gas turbines for power generation

that have almost

lo00

times the output power of the

CGT.

The cumulative operating time for the Basic CGT and

the Pilot CGT was 334 hrs and the number of start-

stops was 645.

i~oa

45,

I I

Items

............

.'.S'YFS

115

401

Targets

c-I

'mFl

Output Power

Thermal Efficiency

Turbine Inlet Temperature

Operation period

Pilot

CGT

Basic

CGT

Year

Fig 10 Efficiency Improvement History

8,000

kW class

34% or more

1,250oC

4,000

hours

Concerning to the NOx emission, we confirmed

ppm

by the rig test and

31.7

ppm (02=16%) by the engine

test at

1350

TIT.

Measured NOx emission data is

shown in

Fig.

and

Fig.

12.

100

CCI-901(ReaI Press)

250?

cumulative operating hours by total of three engines.

The longest endurance operation hours by the same

engine is 782 hours and the high time component is the

combustor liner, which was operated to

1,OOO

hours. We

experienced three times of unexpected termination of

the test by the engine damage. We supposed that it was

initiated by the GGT rotor failure from the investigation

(

Research on soundness and reliability

)

Kyocera

Corp.

(Development and evaluation test

the

heat-resistant ceramic comuonents

)

Tokyo Gas

Co.,

Ltd.

Osaka Gas

Co.,

Ltd.

100

results of damaged engine. There are several suspects

Toho

Gas

Co.,

Ltd.

8.1

.ta

717

15.6

valve

Pilot

50X

251

t-6-

10%

-cc)-

TUlLt

ffl

100

120

A/F

kdkg

0.40.350.3

0.25

0.2

0.15

Eguivalsnce ratio

Fig.

NOx

emission in rig test

...............

Main

Fuel

Ratio

100%

Prim.

Fuel

Ratio

....................................

I I

.....................................

1100

1150

1200

1250

1300

1350

Turbine Inlet Temperature (deg-C)

Fig. 12

NOx

emission in engine test

deterioration, turbine blade resonance and foreign or

domestic object damage, but we could not determine

what was the real cause of the GGT rotor damage.

The endurance test results are as follows.

Target=1,000 Hrs (Weekly Start Stop)

Cumulative operating time over 1,200 OC

2,117 Hr 57 Min

1st.Try

2nd.Try

3rd.T~

4th. Try

Other

7 Hr, Check and adjustment

Total cumulative operating time

2,146

10 Min

Longest time

the same assembly

782 Hr

Longest time

for

the same component

(Combustor Liner)

1,000 Hr

Operation time

Heat Exchanger

2,118 Hr

592

Hr,

GGT rotor damaged

519 Hr, DITTO,

the

combustor support

ring survived

782 Hr,

DITTO,

the

combustor

liner and

support

ring

218 Hr, Combustor liner=1,000 Hr, Stop the test

survived

RESEARCH AND DEVELOPMENT

ON PRACTICAL INDUSTRIAL

CO-GENERATION TECHNOLOGY

Based on the successful results of

300

kW ceramic gas

turbine development, MITI and

NED0

started

"Research and Development on Practical Industrial

Co-generation Technology" project.

The purpose

to encourage prompt industrial

applications of co-generation technology that employs

hybrid gas turbines (using both metal and ceramic parts

in its high-temperature section) by confirming its

reliability and soundness.

The development activities are performed through

material evaluation tests and long-term operation tests

for a hybrid gas turbine

the medium size

(8,000

class). It is expected that the development can realize

low pollution by reducing the emission of

C02

with

highly efficient use

energy.

Exhaust gas properties

Standard regulation values

less

1999

ODevelopment and

Evaluation tests of

Ceramic Components

@Research on

Reliability

Soundness and

REFERENCES

2000 2001 2002 2003

Interim Final

Evalu- Evalu-

(1) Yamagishi, K., Yamada, Y., Echizenya, Y., Ishiwata,

ation ation

S.,

1989 "Current Status of Ceramic Gas Turbine

R&D in Japan", ASME, 89-GT-114

(2) Nagamatu,

S.,

Mizuhara, K., Matsuda, Y., Iwanaga,

A., Ishiwata,

S.,

"Current Status of Industrial and

Automotive Ceramic Gas Turbine R&D in Japan",

(3) Shimada, K., Ushijima, H., Yabe, A., Ogiyama,

H.,

ASME, 91-GT-10

Basic design

Detailed design

Manufacturing

Operation

@General

Investigation of the

System

Tsutsui Y. "Advanced Ceramic Technology Developed

For Industrial

300

kW CGT (Ceramic Gas Turbine)

Research And Development Project In Japan", ASME,

93-GT-188

(4) Arakawa, H, Suzuki, T, Saito, K, Tamura,

Kishi,

"Research and Development of 300kW class Ceramic

Gas Turbine Project in Japan", ASME, 97-GT-87

PREPARATION OF PLANAR SOFC-COMPONENTS VIA TAPE CASTING

OF AQUEOUS SYSTEMS, LAMINATION AND COFIRING

Bernd Bitterlich*’, Christiane Lutz, Andreas Roosen

University of Erlangen-Nuremberg, Department of Materials Science,

Glass

and Ceramics,

D-91054

Erlangen, Germany

Present address: Technical University of Clausthal, Institute for Nonmetallic Materials,

D-38678

Clausthal-Zellerfeld, Germany

ABSTRACT

The Solid Oxide Fuel Cell (SOFC) gains a high eco-

nomical efficiency but its manufacturing costs

are

too

high to compete successfully at the market. Tape casting

via the doctor-blade method is used as a low cost proc-

ess to produce flat components for the planar SOFC.

decrease of the manufacturing costs could be achieved if

the different green sheets would be laminated and then

cofired in a single firing step.

A further advantage for low cost manufacturing would

be the replacement of organic solvents by water in the

tape

casting process. Organic solvents which are widely

used in the industry are flammable, explosive, and a

hazard to environment and health.

this paper anode and electrolyte of the SOFC were

prepared by

tape

casting of aqueous slurries. To achieve

a highly porous anode graphite was used as a filler

which burns out after binder burnout. The anode/ elec-

trolyte composite structure is formed by lamination by

the thermo-compression method. The laminates were

cofired. Defect-free connections between the porous

anode and the dense electrolyte were obtained. The

paper discusses the results in detail.

INTRODUCTION

Due to its electrochemical energy conversion the solid

oxide fuel

cell

(SOFC) offers the advantages of high

efficiency and low pollution in comparison with con-

ventional power stations

[l].

But its spreading is still

hindered by the high manufacturing

costs

[2].

The main functional components of the SOFC

are

the

electrolyte and its two electrodes. Fuel (e.g. hydrogen)

is oxidized on the anode side while the oxidant (oxygen

air)

is reduced on the cathode side. Oxygen ions

diffuse through the dense electrolyte via oxygen vacan-

cies in the lattice which results in an electrical voltage.

Components of a planar SOFC

can

be produced by the

economic

tape

casting process

[3,

41. Lamination and

cofiring of the single green sheets would reduce the

production costs drastically. Cofiring of components for

the SOFC has been carried out by some authors but on

different processing routes [5-71.

this work tape cast-

ing on the basis of aqueous slurries was used to fabri-

cate the SOFC components.

the industry organic

solvents are widely used. They are flammable, explo-

sive, and a hazard to environment and health

[8].

using water as solvent safety considerations during

processing can be kept low [9]. But it is more difficult

to obtain green sheets of high quality. This is due to the

different physical and chemical properties of water

compared to organic solvents, which have a strong im-

pact on the drying behavior. This has to be taken into

account if defect-free green tapes of the electrolyte and

the anode

are

developed

water based slurries. The

lamination of the different

tape

cast sheets is a fast

processing step and can easily introduced in a produc-

tion line. Energy and time are saved by cofiring of the

laminated green sheets.

the common concept of planar SOFCs the electrolyte

is the substrate for the electrodes and must have a cer-

tain thickness

guarantee the mechanical stability

[

10,

111. Reducing the thickness of the electrolyte would

reduce its electrical resistance making it possible to

lower the operating temperatures while gaining the

same power density [12].

In addition the lower operat-

ing temperature would reduce the complex material

problems of the SOFC-periphery which cause high

costs. This concept is realized in the design proposed by

the Research Center of Jiilich, which consists of a thick

porous anode supporting a thin 20 pm electrolyte layer

and the second electrode (Fig. 1) [13].

Cathode

-50

Anode

-50pm

a) Electrolyte

substrate

Fig.

Planar SOFC designs: Substrate is either the

electrolyte (a)

the anode

(b).

Anode

substrate!

The single components must fulfill several require-

ments:

The electrolyte, which consists of fully yttria stabilized

zirconia

(YSZ), must

impervious to gases.

order to

reduce the electrical resistance enabling a lower work-

ing temperature

the SOFC, the electrolyte must

very thin

pm).

The anode consists of a Ni/YSZ-Cermet, which is

formed

a reducing step, when NiO is transferred to

Nickel. The anode has various functions: it is the load

bearing substrate for electrolyte and cathode and there-

fore it must have a sufficient thickness (appr.

1,5

mm)