présentation

ALTERNATIVE METHODS FOR BIOMASS

COMBUSTION CONTROL

Jean-Bernard MICHEL - University of Applied Sciences of Western Switzerland - Geneva

E. Onillon - Swiss Center for Electronics and Microtechnology, Neuchâtel, Switzerland

Outline

• University of Applied Science of Western Switzerland • The problems of biomass use for district heating

• Sensors for biomass combustion control

• Initial results with an optical sensor for excess air control • Conclusion and future perspectives

University of Applied Science of Western Switzerland

• 5 regions of western CH

• 6 job domains • 10’000 students

Health Engineering and architecture

Main R&D activities

Fluid mechanics and energy conversion systems Lab, Geneva

– Building design, simulation, HVAC systems – Aérodynamics of sports.

– Renewable energies

– Thermodynamical cycles

Thermal Engineering Institute, Yverdon

– Phase change materials and ice slurries (IEA working group) – Magnetic cooling

– Combustion

Participation to Shell Eco-Marathon race

Outline

• University of Applied Science of Western Switzerland

• The problems of biomass use for district heating

• Sensors for biomass combustion control

• Initial results with an optical sensor for excess air control • Conclusion and future perspectives

Example: French potential of renewables

0 10 20 30 40 50 60 70 80 B iogas el ect rici ty Soli d Bi om ass elect ric ity B iow a st e el ec tr ic ity G eot her m al el ect ric ity Hydr o l ar ge -s cal e H ydr o sm al l-s ca le P hot ovol ta ic s Solar t her mal elec tr icit y Ti de & W ave Wind onshor e W ind of fs ho re E lec tr ic it y ge n er at ion [ T W h /y ea r]

Achieved potential / Existing plant Additional potential / New plant

0.0 2000.0 4000.0 6000.0 8000.0 10000.0 12000.0 14000.0 B iom ass heat ( gr id) G eo the rm al he at ( gr id) B iom ass heat ( non-gr id ) He at p u mp s So la r c olle ct or s H ea t gene ra ti on [ k to e/ ye ar ] 0 2000 4000 6000 8000 10000 12000 B iof uel s Fu el pro d u c ti on [kto e/ year]

Example: German potential of renewables

0 10 20 30 40 50 60 70 80 Bi ogas el ect ric ity Soli d Biom ass elect rici ty B iow a st e el ec tr ic ity Geot her m al el ect ricit y H ydr o l ar ge -s cale H ydr o sm al l-s ca le Phot ovolt ai cs Solar t her m al el ect ric ity

Tide & Wave _Wind onshor

e W ind of fs ho re E le ct ric it y ge n era ti o n [T W h /y ea r]

Achieved potential / Existing plant Additional potential / New plant

0 2000 4000 6000 8000 10000 12000 Bi o m ass he at ( gr id ) G e ot he rm al he at ( gr id) B io m as s he at (n on -g rid) He a t p u mps So la r c o lle ct or s H ea t gene ra ti on [ k o e/ ye ar ] 0 1000 2000 3000 4000 6000 7000 Bio fue ls Fuel product ion [ kt e/ year ] t o5000

Factors limiting the increase of biomass for heating

• Availability of the biomass in urban areas • Storage

• Investment costs • Emissions

• Variability of the fuel

– Moisture

– Size distribution – Composition

• Reliability – frequent unwanted stops

ÄEconomics very much depending on operating variables ÄNeed for better monitoring & control of the combustion e.g.

excess air, primary/secondary air, CO

Pellet boiler designed to retrofit oil boilers

HOVAL,

Outline

• University of Applied Science of Western Switzerland • The problems of biomass use for district heating

• Sensors for biomass combustion control

• Initial results with an optical sensor for excess air control • Conclusion and future perspectives

Technical and economical requirements for the sensor system

• Reliability for O2 and CO/HC emisssion control

• Low cost : < 1 Euro per kW as a rule of thumb

• Regulate the primary air flow through and along the fuel bed

• No or very low maintenance.

CO/VOC sensor from Lamtec (Escube) – Carbosen 1000

ZrO₂ ceramics

Example of excess air control using CO and O

₂

signals

Outline

• University of Applied Science of Western Switzerland • The problems of biomass use for district heating

• Sensors for biomass combustion control

• Initial results with an optical sensor for excess air control

Flame signature sensor principle for biomass flames

Fuel identification

Signal processing: Average, spectral power CH/OH and CH*/OH*ratio

Local fuel richness

60 seconds samples

Biomass boiler Sensor signals (OH, CH radiation)

Flame signature sensor for gas & oil fired burners

Price target: 100-500 Euros

∅ 12 mm

GaP sensor responses

CH (432 nm)

Dyadic time filtering process

9 frequency bands – 0 to 3000 Hz 0 500 1000 1500 2000 2500 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Gain

Control scheme

• Record « good » flame signature for 10

minutes at constant load

Æ 9 wavelet coefficients corresponding to 9 frequency bands

Æ Average and Standard deviation

Æ Signal/noise ratio W av e le t c o effi ci en ts (a rb it ra ry u n it s) calibration curve +/- 2 std. dev. + + + + + + + + ₊ Instant measurement

•

Monitor deviation from the norm and provide alerts

Sensor response to step changes in excess air- Gas flame

0 50 100 150 200 250 1 2 3 4 5 6 7 8 O2 % O₂ % Samples (1 per 6 s) 0 50 100 150 200 250 -5 samples (1 per 6s) alert1 alert2 distance Alert 1(b) Alert 2 (r) Dist (g) 0 5 10 15 20 25 30 35 40 al er ts a nd di st anc e

Forced-draught gas fired diffusion burner, 180 kW trained for 15% excess air.

Müller AG – Boiler schematic

• Suitable for variety of wood and wood wastes – variable moisture content - 800 to 1300 kW

Initial results

• Signature frequencies much lower than with forced draught burners • Reproducible results with UV/visible sensor

-90 -80 -70 -60 -50 -40 -30 -20 -10 0 Am p lit ud e [d B ]

Wavelet distance as a function of flue-gas oxygen

Baseline O2 = 8% -10 0 10 20 30 40 50 60 70 80 -4 -3 -2 -1 0 1 O2 variation (% ) D ist an ce fr o m b asel in e [ -]

•

200 kW Müller AG Boiler

•

wood chips and saw dust

•

single UV/visible sensor

decreasing

λ

Baseline O2 = 8% 200 300 400 500 600 700 800 ist an ce f ro m b asel in e [-]

increasing

λ

Conclusion and future perspectives

• Several low-cost methods are in development for biomass combustion control:

a) Fuel: In-line moisture indication

b) Flue-gas : micro gas sensors for CO/HC and O2 c) Flame: optical indication of fuel type and excess air

• b) and c) methods already proven and partially implemented in industrial oil burner controllers

• Low cost and direct monitoring of excess air variation • Easy to implement on wood-fired boilers

• Optical sensor features:

– Fast response (~ 5 seconds)

– Potential for air flow adjustment along the grid – Long life time without drift

Acknowledgements

• Financial and technical support of the CREED (the center for environment, energy and waste research) - Veolia

Environnement's research center based in Limay (France) • City of Geneva in providing access to their 1MW