Preliminary Analysis of Combustion Noise Generated by Burners in Cracking Furnaces

Paper from the AFRC 2013 conference titled "Preliminary Analysis of Combustion Noise Generated by Burners in Cracking Furnaces" by Grazyna Petela

Grazyna Petela, Karen Draves, Les Benum NOVA Chemicals, Calgary/Joffre, Canada Preliminary Analysis of Combustion Noise Generated by Burners in Cracking Furnaces AFRC 2013 Industrial Symposium, Hawaii, Sept. 22-25, 2013 2 1.0 OBJECTIVE 2.0 BACKGROUND: Mechanism Of Combustion Noise Generation 2.1 Aerodynamic Noise 2.2 Combustion Noise 2.3 Burner Noise in Free Field 2.4 Noise from a Burner Enclosed in a Furnace 3.0 MEASUREMENTS of Noise Generated by Burners in a Cracking Furnace 3.1 Furnace-Burner Configuration 3.2 Acoustic Measurement Procedure 3.3 "Scouting" Analysis of Noise from Burners in "Routine" Operation (2010) 3.4 Effect of Operation Parameter Changes on Burner Noise Spectrum (>>1 year later) 4.0 CONCLUDING, at this stage… CONTENT : 3 To investigate if measurement of combustion noise generated by multiple burners in a cracking furnace provide any information about combustion characteristics, (such as e.g. flame extinction, stability, fuel/air ratio, burner loading, etc.) Can acoustic measurements be used as a diagnostic tool to help monitoring safety or quality of combustion process in a furnace ? 1.0 OBJECTIVE 4 • Quadropole in nature, broadband, relatively high frequency range, strongly directional , (fundamentals by Rayleigh, Lighthill) 2.0 BACKGROUND: Mechanism Of Combustion Noise Generation Noise spectra measured at different angles for the jet of 195 m/s velocity Directionality of jet noise, [Bogey, Bailly, Juv´e, , 2002] 0 30 60 90 120  (deg) 118 114 110 106 102 OSPL, dB Jet noise spectra dB 20 50 100 200 500 1000 2000 5000 10,000 20,000 f, (Hz) 15o 45o 90o • Depends on jet velocity, characteristic dimensions and turbulence • Peak frequency is Strouhal (St = f djUj ) related • Sound power of sub and supersonic jets Pa  djUj 8 2.1 Aerodynamic Noise of Non-Reacting Jets 5 • Monopole type, broadband, low frequency range, flatter than jet noise spectrum, weak directionality, (Bragg, Strahle, Putnam, Gupta, Beer) • Related to chemically controlled reaction rate per unit volume : • Acoustic power of a diffusion flame : f (Hz) dB (Z) 100 500 1000 5000 f, Hz dB Combustion roar ∝    ′ ∝ 2 2  ∝ ℎ ℎ • Noise/roar from large industrial combustion systems: thermo acoustic efficiency  (power of firing rate )1.8-2 2.2 Combustion Noise f, Hz 6 Acoustic energy (spectrum) Jet: Turbulent mechanics (turbulence, shear, pressure) Flame: Chemically controlled reaction intensity (reaction rate, heat release rate, diffusivity) = Superposition of aerodynamic and combustion noise Diagnostic possibility based on a burner noise measured in free field conditions A) Indication of flame instabilities B) Related to swirl and turbulence level C) Depends on combustion air/fuel ratio •Aerodynamic jet noise amplified by flame (reaction generates additional turbulence) •Chemically generated flame noise influenced by jet turbulence •Industrial burners: jet noise can coincided with chemical reaction noise, for sufficiently high flow rates 2.3 Burner Noise in a Free Field 7 Acoustic signatures, (i) - spectrum, ii) - in time domain), from a piloted burner approaching lean blowout , [Suraj Nair, 2006] Flame sound power efficiency vs. swirl number Flame noise level as a function equivalence ratio [Bertrand, Michelfelder, 2000] 0.4 0.8 1.2 1.6  < >102 ( Pa) 1.8 1.4 1.0 0.6 0.2 Re=5800 Re=2770 Re=3850 C3H8+air 0 0.5 1.0 1.5 2.0 S 1MW 2MW 1.5M  x106 0.5 1.0 0.75 0.25 0 Image of a flame approaching blowout (pilot 0.6%) B) C) A) i) ii) 8 • Interaction between aerodynamic noise, combustion noise and enclosure • Components of the frequency spectrum - attenuated or amplified at enclosure natural frequencies • Amplification/attenuation depends on the location of the burner in the furnace, internal shape of the furnace, acoustic properties of the furnace lining 0 30 60 90 120  (deg) 118 114 110 106 102 SPL, dB 10 20 50 100 200 500 1000 2000 f(Hz) 100 90 80 70 60 SPL, dB 300 kW burner in furnace No1 6 burners ( 300 kW) in furnaces No 2 25 100 1000 5000 f, Hz dB typical combustion roar enclosure response burner tile response 25 100 1000 5000 f, Hz dB • Noise interaction from multiple burners Noise spectra indicating burners /enclosure interactions, [Putnam, Faulkner, Fied, 1993] 2.4 Noise from a Burner Enclosed in a Furnace 9 3.0 MEASUREMENTS of Burner Noise in a Cracking Furnace  Challenge: • 48 burners enclosed in a furnace • Confined flames; high temperature environment • Damping of certain frequencies in noise spectra by the furnace enclosure • Possibility of multi-modal acoustic responses of a furnace at the following frequencies: f=c/(c-..) ; (c-c) = 2Lx,y,z /n, (c-o) = 4Lx,y,z/n, n=1,3,5..; Lx,y,z - furnace characteristic dimensions) Can measurements capture a noise change for a single burner? or is a furnace acting like one acoustic source? Burners in two cracking furnaces in Joffre Plant E1 selected for noise analysis / measurements 10  Furnaces: Natural Draft (H106) and Induced Draft (H108) 3.1 Furnace - Burners Configuration • Draft controlled by a stack damper in H106 and by a fan in H106 (lower stack) • Identical geometry of radiant sections: 3 decks with 2x8 burners • O2 analyzer in the stack  Burners: Preheated fuel H2/CH4 , Air introduced at the burner tip via adjustable air register 4 coils; 48 burners; 1 coil heated by 12 burners: 6 in cold box, and 6 in hot box feed products radiant section 1st coil "cold" box "hot" box 4th coil 3rd coil 2nd coil ~14m ~7 m ~1.5 m burners (48) top deck (3rd) middle deck (2nd) bottom deck (1st ) observation windows W E 11 3.2 Acoustic Measurement Procedure  Measured parameters: • Noise: Overall Sound Pressure level, (OSPL) Fast Fourier Transform (FFT) noise spectrum (20Hz -10,000Hz); • Locations: microphone in open observation windows, facing burners; all 3 furnace decks, West and East sides; • Furnace operating parameters (fuel flow, draft, O2, coil loadings, etc.)  Acoustic Equipment • B&K Analyser 2250 ( + FFT Analysis Software BZ 7230) • B&K Microphone 4182 (operating temp. 700oC) 12  Noise recorded from multiple burners in a cracking furnace to verify: • Accuracy / consistency of measured acoustic spectra • Effect of microphone location • Background noise 3.3 "Scouting" Analysis of Noise from Burners in "Routine" Operation, (2010) Burner noise spectra measured on the 2nd deck (E &W sides of the deck) Burner noise spectra measured on all 3 decks ( East side of each deck )  Burners in Natural Draft Furnace 100 1000 f, (Hz) 10000 (dB ) 50 60 70 80 90 100 EAST background (3rd_deck) EAST, 2E(3rd_deck) EAST, 2E(2nd deck) EAST, 2E(1st deck) bottom (the 1st) deck background noise Noise spectra for the natural draft furnace 06, measured in cold box; all DECKS, @ EAST Window 2 top (the 3rd ) deck middle (the 2nd) deck Cold box, H106 furnace acoustic frequencies 100 1000 f, (Hz) 10000 (dB) 50 60 70 80 90 100 WEST, 1W WEST, 2W EAST background EAST, 1E EAST, 2E background noise the 2nd deck, EAST the 2nd deck, WEST Cold box, H106 furnace acoustic frequencies 13 Analyzing: Very similar noise spectra (amplitudes and frequency content) recorded on all decks in the natural draft furnace(106), at all microphone locations: either the burners of the same type, operating with the same fuel loading, generated practically identical noise spectra, or noise from a single burner could not be distinguished as the firebox radiated noise like a single source Furnace acoustic responses in the low frequency (>350Hz) part of the noise spectrum No effect of background noise 14  Burners in Induced Draft Furnace Burners on the 3rd deck had the reduced fuel loading ! Burner noise spectra measured on the middle ( 2nd ) deck, at all locations (observation windows) Burner noise spectra measured on all decks, at East location 100 1000 f (Hz) 10000 (dB) 50 60 70 80 90 100 WEST background (2_deck) (wind!) WEST, 1W WEST, 2W EAST background (2_nd deck) EAST, 1E EAST, 2E background noise (E / W) Noise spectra for the balanced draft furnace 08, measured in cold box; 2nd DECK, @ EAST & WEST windows the 2nd deck, WEST the 2nd deck, EAST Cold box, H108 Furnace acoustic frequencies 100 1000 f, (Hz) 10000 (dB) 50 60 70 80 90 100 EAST background (3rd_deck) EAST, 2E(3rd_deck) EAST, 2E(2nd deck), EAST, 2E(1st deck) background noise top (the 3rd ) deck Noise spectra for the balanced draft furnace 08, measured in cold box; all DECKS, @ EAST Window 2 middle (the 2nd) deck bottom (the 1st) deck Cold box, H108 furnace acoustic frequencies 15  Noise measurements after imposing the following operational changes: •Single burner : A) No fuel supply (flame extinguished) B) No combustion air supply (the air register closed) C) Maximum air supply (maximum opening of the air register) 3.4 Effect of Operation Parameter Changes on Burner Noise Spectrum (>>1year later) Approach:  "Reference" acoustic signatures, i.e. noise spectra for burners in both furnaces operating in baseline conditions, (typical coil loadings; nominal air/fuel supply to burners) • Furnace: D) Change in the coil load and furnace draft "Baseline" furnace parameters Natural Draft, H106 Induced Draft, H108 fuel mass flow rate, m F, (kg/h) 1280.0 2380.0 draft, P, (Pa) - 35.0 - 32.0 O2, in STACK, (%) ~ 5.7 ~2.5 Analysis / comparison of the acquired spectra with the reference signatures 16 3.4 A) Single Burner: No Fuel Supply (Flame Extinct) :  Natural Draft Furnace (middle deck) mF (=1260kg/h), P  ref (-35Pa), O2, STACK (=6%)  Induced Draft Furnace (middle deck) mF (=2360kg/h), P  ref (-32Pa), O2, STACK  (=2.8%) 100 1000 f (Hz) 10000 50 60 70 80 90 100 REFERENCE operation (baseline) NO FUEL to 1 burner (dB) 100 1000 f (Hz) 10000 (dB) 50 60 70 80 90 100 REFERENCE operation (baseline) NO FUEL to 1 burner 100 1000 f (Hz) 10000 (dB) 50 60 70 80 90 100 NO FUEL to 1 burner REFERENCE 100 1000 f (Hz) 10000 (dB) 50 60 70 80 90 100 NO FUEL to 1 burner REFERENCE 10th order regression 10th order regression on the middle deck on the middle deck 17 3.4 B) Single Burner: No Air Supply, (Register Closed)  Natural Draft Furnace mF  ref (1280kg/h), P  ref (-35Pa), O2, STACK ref (5.7%)  Induced Draft Furnace mF  ref (2380 kg/h), P  ref (-32Pa), O2, STACK  ref (2.5%) 100 1000 f (Hz) 10000 50 60 70 80 90 100 REFERENCE (baseline) NO AIR to 1 burner (dB) 100 1000 f (Hz) 10000 50 60 70 80 90 100 REFERENCE (baseline) NO AIR to 1 burner (dB) 100 1000 10000 50 60 70 80 90 100 REFERENCE NO AIR to 1 burner f (Hz) (dB) 100 1000 10000 50 60 70 80 90 100 REFERENCE NO AIR to 1 burner f (Hz) (dB) 10th order regression 10th order regression 18 3.4 C) Single Burner: Max. Air Supply, (Register Fully Open)  Natural Draft Furnace mF  ref (1280 kg/h), P (=-32 Pa), O2, STACK  ref (5.7%)  Induced Draft Furnace mF  (=2350 kg/h), P  ref (-35 Pa), O2, STACK (=2.7%) 100 1000 f (Hz) 10000 50 60 70 80 90 100 REFERENCE (baseline) MAXIMUM AIR to 1 burner (register max open) (dB) 100 1000 f (Hz) 10000 50 60 70 80 90 100 REFERENCE MAXIMUM AIR (baseline) to 1 burner (register max open) (dB) 100 1000 10000 50 60 70 80 90 100 REFERENCE MAX AIR to 1 burner f (Hz) (dB) 100 1000 10000 50 60 70 80 90 100 REFERENCE MAX AIR to 1 burner f (Hz) (dB) 10th order regression 10th order regression 19 3.4 D) The Furnace: Reduced Draft and Coil Loadings  Natural Draft Furnace P  (= -20Pa), O2, STACK  (=5%), mF (=1260kg/h  Induced Draft Furnace P  (=-24Pa), O2, STACK  (=4.0%), mF (=2150kg/h) f (Hz) (dB) 100 1000 10000 50 60 70 80 90 100 REFERENCE REDUCED DRAFT & LOAD to coils background 100 1000 f (Hz) 10000 50 60 70 80 90 100 LOWER DRAFT in the furnace REFERENCE (baseline) (dB) background 10th order regression 10th order regression 20 "Reference" noise spectrum, emitted by burners during baseline operation in both cracking furnaces, had the following characteristics: •Extended in the frequency range 100-10,000Hz; •Had consistent, steady spectral components, (high correlation coefficients indicating high similarity between reference spectra measured ~2 years apart: 0.9889 in the ND furnace H106 and 0.9862 in the ID furnace H108); •The 1st mode and higher acoustic responses of the furnace were always present in spectra, in frequency range < 350Hz; 4.0 CONCLUDING, at this stage of the project Changes in a burner operation resulted in a shift of the burner noise spectrum from the reference spectrum, (e.g. after loss of a flame by a burner, autocorrelation between the burner noise spectrum and the reference spectrum decreased to 0.9689 and to 0.9500, in furnaces H106 and H108, respectively) 21 Based on spectral analysis of the burner noise, it was possible to detect the following changes in the furnace operation: •Loss of a flame and air starvation by a single burner •Reduced fuel supply to a group of burners •Change in the furnace draft and load If the results are further verify, a simple non-intrusive method can be developed, to assist in improving safety of furnace operation Concluding..cont… Detectable by noise measurements carried out in the "neighborhood" of the affected burner, i.e. on the same furnace deck Detectable by noise measurements at any furnace location The information contained herein is provided for general reference purposes only. By providing the information contained herein, NOVA Chemicals makes no guaranty or warranty and does not assume any liability, with respect to the accuracy or completeness of such information, or product results in any specific instance, and hereby expressly disclaims any implied warranties of merchantability or fitness for a particular purpose or any other warranties or representations whatsoever, expressed or implied. Nothing contained herein shall be construed as a license to use the products of NOVA Chemicals in any manner that would infringe any patent. Nothing herein shall be copied, reproduced, distributed or otherwise used without the express written permission of NOVA Chemicals. is a registered trademark of NOVA Brands Ltd.; authorized use. www.novachemicals.com Responsible Care® is a registered trademark of the Chemistry Industry Association of Canada (CIAC) in Canada and is a registered service mark of the American Chemistry Council (ACC) in the United States. PERFORMANCE DRIVEN. CUSTOMER INSPIRED.