Experimental Study of Ignition in a Pilot Flame System

Transcription

1 Experimental Study of Ignition in a Pilot Flame System Ricardo Alexandre da Fonseca Rato Center IN+ - Lab. of Themofluids and Energy Systems Dept. Mech. Engineering Instituto Superior Técnico Av. Rovisco Pais, 19-1 Lisboa, Portugal Technical University of Lisbon (UTL) rato_ricardo@hotmail.com Abstract In new water-heater units, called intelligent, the pilot flame only turns on when the hot water tap is opened, instead of what happens in the older units where the pilot flame is always lit. Therefore, these units must have a reliable pilot flame ignition system. In this context, the objective of the present work was to study in detail a pilot flame system (commercially available) in order to identify the causes that may contribute to the non success of ignition and to propose a new pilot flame system with a higher ignition ability. In order to accomplish this objective, a current pilot flame system was submitted to an experimental characterization. Secondly, an experimental study was performed to evaluate from the fundamental point of view the effect of mixtures properties and electrodes conditions parameters on the success of spark discharge (occurrence of a spark discharge) and on the success of ignition (sustained flame propagation after a spark discharge). These experiments were performed controlling the electrodes spacing, d, to determine ds (critical spark distance) and di (critical ignition distance), for a fixed voltage/energy supply, varying equivalence ratio, temperature, humidity of the air, mean velocity, and electrodes diameter. Finally, based on all these results a new pilot flame ignition system was proposed and experimental characterized/tested. The experimental characterization of the current and the proposed pilot flame system included: measurements of the velocity field at the pilot tube exit using LDV technique, determination of the primary equivalence ratio and high-speed cinematography recordings of spark and flame development. A new pilot flame system was proposed based on a new electrodes arrangement and on a new pilot tube geometry. In this system the spark is dischargenside the pilot jet (in contrast with the current system) which has a primary equivalence ratio, prim, of 1.7, while the current system works with prim=.7. The proposed pilot flame has 1% of ignition probability using a single spark discharge. Key-words: Ignition, Pilot Flames, Ignition Improvement. 1

2 1. Introduction In a commercial water-heater unit the aim of the pilot flame is to ignite several burner flutes that exist inside, as shown in Fig. 1a-b). In it is presented the main components of a pilot flame ignition system, which are: the pilot burner tube, the fuel injector, the electrode and the spark discharge unit, as shown in Fig. 1c-d). Primary Air- Mixture Pilot tube ( a) Secondary Air Entrainment Air Injector Primary Air- Mixture Electrode Secondary Air Entrainment discharge unitair ( Injector d) ( b ) ( c ) Electrode discharge unit Fig.1: Pilot burner tube location on the water-heater unit ants geometry. a) Water-heater unit, b) Location of the pilot burner in the water-heater, c) Pilot burner geometry, d) Schematic representation of the pilot flame ignition system. The fuel injected entrains primary ambient air until entering into the pilot tube, where mixing occurs. At the exit of the pilot tube, the mixture forms a free jet where a secondary ambient air is entrained. The spark discharge unit supplies several successive spark discharges with a predefined time interval of 1ms until a flame is established. To ignite a mixture using spark ignition it is necessary, firstly, to have a spark discharge and secondly the energy added by the spark to the mixture must be enough to cause a self-sustained flame. With these two conditions a successful ignition is accomplished. For particular mixture condition, there is a lowest voltage that will cause a spark to be established between two electrodes, the breakdown voltage, Vb []. Paschen s law states that, Vb=f(Pd), which means that the breakdown voltage in a uniform field gap (d/d ) is a unique function of the product of pressure and the electrodes spacing for a particular gas mixture and electrode material [3]. Now, the influence of non-uniform electric field on Vb, i.e. (d/d ), was experimental tested by [, ], reviling that the decrease the nonuniformity leads to a decrease of Vb. In order to ignite a mixture it is necessary to add a certain amount of energy to it, which in this case is provided by a spark. According with [1], the energy supplied by the spark must to be higher than a threshold value to the spark produces a sustained flame. This minimum ignition energy, Emin, is function of experimental variables of the reactant mixture, the electrodes characteristics, the spark discharge characteristics, pressure, temperature and flowing characteristics [1,, -11]. The equivalence ratio is an important factor that affects the value of the minimum ignition energy of a combustible mixture [1,,-]. In jet flames, the equivalence ratio varies spatially, suggesting that the success of ignition is influenced by the electrode position and arrangement. In addition, when the jet flames are partially premixed the value of the local equivalence ratio depends on the upstream conditions such as fuel injector and pilot tube geometry. The flow properties, as the mean velocity, also has an effect on the amount of energy required [, ]. The ambient air conditions as temperature and humidity may be other important factors that influence the success of ignition (sustained flame propagation after a spark discharge).

3 In order to reduce the occurrence of ignition failures, the present work intends to study a current pilot flame system, analyze the effect of parameters as equivalence ratio,, temperature, T, humidity, mean velocity, U, electrodes spacing, d, and electrodes diameter, d, on the success of spark discharge (occurrence of a spark discharge) and on success of ignition, and based on all these results to propose a new pilot system. To accomplish these objectives, it was firstly submitted the current pilot flame system to an experimental characterization. In this characterization it was taken into account: measurements of the velocity field at the burner exit, determination of the primary equivalence ratio, ignition tests and recordings the earliest moments of the flame ignition. The effects of mixture properties and electrode parameters on success of the spark discharge and on success of ignition the spark ignition were performed by controlling the electrodes spacing, which became a most important variable since it defines the ability of the system (for a fixed voltage/energy supply) to have a spark and a flame. These tests were conducten a model burner, which ensures constant properties of the flowing mixture within electrodes. The mixture was supplied with different temperature and humidity levels, using a developed air and fuel conditioning system. Finally, with all the results obtained, a new pilot flame ignition system was proposed. The new system was experimentally characterized according with the same procedure of the current system.. Experimental Setup During this work two main experimental configurations were used, the configuration of the pilot burner and the configuration of the model burner. 3 The configuration of the pilot burner, Fig., was used to conduct ignition experiments with different pilot tubes and electrodes geometries. The sparks were provided by the spark discharge unit present in current pilot flame system from Bosch. The energy supplied by the sparks is constant, around 5 mj, according to the manufactuer. The position of the injector and of the electrode was controlled by -- micro positioning stages, which allow tri-dimensional movements until.1mm in each direction with μm of accuracy. The combustible flow rate was controlled by an electronic flowmeter (Alicat Scientific MC) with a maximum flow capacity of 1 standard liter per minute (SLPM), being its maximum error ±.SLPM. It was also, used a K-type thermocouple (±1.1 C of typical accuracy). and a HIH-- humidity sensor (±3.5%. of accuracy) Electronic Flowmeter Propane/metane tank Support and Positioning System Pilot tube Injector Electrode Data acquisition and processing Thermocouple & Humidity sensor Fig.: Configuration of the pilot burner Discharge Unit The configuration of the model burner, Fig.3, was created to evaluate the influence of flowing mixture properties and electrode parameters on the success of spark discharge and the success of ignition. The model burner is made of glass and has the exit diameter of.mm and the electrodes are positionen front of the burner exit. This burner has a parabolic curvature, which produces a plug, providing constant properties at the core region, where the electrodes are positioned. The electrodes spacing was controlled using the -- micro positioning stages (the same

4 as the pilot burner configuration). The mixture conditions were conditioned, using a developed air and fuel conditioning system which allows to provide mixture at temperatures between 9 C and 3 C and the humidity ratios between 1.5 g/kg dry air and 7.5 g/kg dry air, ants temperature and the relative humidity were monitored by a developed real-time acquisition system. The primary equivalence ratio prim of different pilot jets was calculated, for different pilot tube geometries, using a method that with the knowledge of the injected fuel mass flow rate and the mixture volume flow (determined by integrating the velocity profiles of the jet from the LDV measurements) it is possible to calculate the mass flow rate of the primary air using equation (1). Support and Positioning System Data acquisition and processing discharge unit (1) Condicioneted and Air Mixture U,,T,HR Glass Burner Fig.3: Configuration of the model burner 3. Diagnostic techniques The Laser Doppler Velocimeter (LDV) technique was used to measure velocity in various pilot tube geometries ann the model burner. The velocity ants characteristics were studied using a dualbeam one component Dantec LDV, in forward scatter mode, based on a W argon-ion laser (Spectra-Physics model Stabilite17) with a wavelength of 51.5nm (green light). The system was used to measure mean quantities, based on 1 valid data points. According to [1], the statistical errors associated with mean values are less than 3% and less than 5% for variance values, due to the high number of occurrences for total time series processed. The visualization of the early moments of the spark ignition process was performed using highspeed digital camera Phantom V., which enables sample rate up to 9 frames per second (fps) and a maximum resolution of 51x51 pixels. The present experimental investigations of the critical spark distance, ds, and the critical ignition distance, di, are sensitivity experiments. The values of ds and di were obtained for the 5% probability of occurrence of either success of spark discharge or success of ignition respectively, using the Up-and-Down method [13].. Results and Discussion The current pilot flame ignition system from Bosch is composed by four parts: the fuel injector, the pilot tube, the electrode and the spark discharge unit, as illustraten Fig.. The injector is a plate orifice type with.35 mm diameter. The pilot tube blended 9, which one is positioned 1mm above the propane injector and has an internal diameter of.mm, ending with a coil. This coil reduces the internal diameter to 3.mm. The electrode, positioned mm below the end of the pilot tube coil, has.mm of diameter ans made of a high temperature iron-chromiumaluminium alloy known as Kantal A. In current pilot flame ignition system from Bosch, the fuel injected entrains primary ambient air until entering into the pilot tube, where mixing occurs. At the exit of the pilot tube, the mixture forms a free jet, which according with velocity

5 measurements taken at.5mm of the burner exit, for the nominal propane flow rate Qpropane=.SLPM, has an quasi-axisymmetrical distribution of axial velocity, with a maximum velocity in the inner zone around m/s, as in shown in Fig.5. The primary a primary equivalence ratio of this pilot jet is.7. In order to ignite this mixture it is required an amount of energy supplien excess of 5mJ according with the data of [1], which corresponds in extremis to the amount of energy supplied. the moment of beginning of the fuel injection and spark discharge, Δt. For a single spark discharge and the nominal volume propane flow rate its higher ignition found probability was 39% for Δt=15s, which decreases to % for Δt=5s. V U W Mean RMS prim=.7 Flowing reactant mixtures Pilot tube. Electrode discharge unit [mm] 1 U [m/s] U [m/s] Mean (a) RMS Injector Fig.: Schematic drawing of the current pilot flame system. Also, the spark is dischargen a region between the electrode and the bottom of the pilot coil exit, as shown by several frames of the transient process of a typical ignition of the current pilot flame presentehown in Fig.. The recordings with the ambient air conditions of T=1 C and RH=5%, using the nominal Qpropane. The pilot jet does not reaches this region, as illustraten Fig., and the propane only reaches that region due to his mass diffusion in surrounding air, assisted by the random fluctuation of radial velocity in between the coil turns. Due to the nature of these processes, it is necessary some time to the propane reaches this region. This situation is critical because the local equivalence ratio in the region where the spark is discharges not known and changes with the time (after the opening of the gas valve), being a not controlled process. Therefore, the ignition probability of the current system has a dependency of the time lag between [mm] (b) Fig. 5: Velocity profiles of the mean and rms of the axial velocity of the current pilot tube. a) Horizontal direction, b) Vertical direction Working conditions: Qpropane=. SLPM, =.5 mm A new pilot flame system was proposed based on two main ideas to improve the ignition ability of the system. The first idea was to insert two electrodes in front of the pilot tube exit in order to the spark being dischargenside the pilot jet. This solution ensures that the spark discharge occurs in a mixture with a known equivalence ratio ants value is constant in time. By this reason, the electrodes arrangement of the proposeystem is composed by two electrodes, vertically opposed, 3mm in front of the pilot tube exit, as illustraten Fig. 7.

6 E min [mj] (a) μs (c) μs (b) μs (d) μs constant, limited by the spark discharge unit, this decrease in equivalence ratio defines a working condition where the required energy to ignite the mixture is lower than the available, resulting in a more favourable conditions to obtain a successful ignition. In this context, it was proposed a new pilot tube geometry with a uniform internal diameter of.5mm. This pilot tube geometry enhances the air entrainment, reducing the value primary equivalence ratio to 1.7. In Fig. it is shown a comparison between the primary equivalence ratio of the proposed and the current system, in a graph of minimum ignition energy as a function of equivalence ratio. The pilot flame produced by the new tube geometry at the nominal volume propane flow rate (Qpropane=.SLPM) has stability problem. Therefore a stability study had been made, suggesting that the Qpropane should be within the interval.15slpm and.slpm. (e) 1 μs (f) μs prim Proposed System prim Current System Figure : Frames of the earliest moments of a current 5 pilot flame system typical ignition process ants relative to the image (b).working conditions: Qpropane=.SLPM, T=1 C, RH=5%. 1 prim= Flowing reactant mixtures Fig. : Representation of the prim of the current and proposeystem in a graph of Emin as function of, for Pilot tube propane-air quiescent mixtures [1]. Injector 1.5 Figure 7: Schematic drawing of the proposed pilot flame system The secondea was to reduce the primary equivalence ratio of the pilot jet, approximating to it to near stochiometry, by increasing the entrainment of air. Since the energy supplies The choice of the electrodes diameter, d, and the electrode spacing, d, for the proposeystem were based on the experimental analysis performed to evaluate the effect of the mixture and electrode parameters on the ability of the system to have spark/ignition. In this analysis were created two new variables to evaluate the influence of different parameters (, T, w, U, d) on the spark production and on the success of ignition, due to the usepark discharge unit to provide a constant value of the initial voltage and of the

7 [mm] supplied energy. These variables are the critical spark distance, ds, which corresponds to the maximum electrode spacing that allow a spark discharge to occur, for a given constant initial voltage and the critical ignition distance di which corresponds to the minimum electrode spacing that allow a sustained flame propagation after a spark discharge, for a given constant energy supplied. Thus, to ignite a mixture using a spark ignition system, first the spark must be discharged within the electrodes (d<ds ) anecondly the heat losses for the electrodes must be small enough in order to the dischargepark energy to ignite the reactant mixture (d>di). Fig.9 it is shown a graph with the curves of ds and di as function of that explains graphically the two conditions expressed above, for a propane-air mixture with: T, RH, U, d fixed parameters. No + Ignition = Flame No Ignition Fig. 9: Critical spark distance and critical ignition distance.working conditions: Propane-air mixtures, U=.9m/s T=.5 C, RH=%, w=. gwater/kgdryair, d=.mm. The chosen electrodes diameter was.5mm, instead of.mm of the current system, because has been shown that for a particular mixture, the decrease of electrode leads to a significant increase of ds, i.e., gives more favourable conditions to have a spark discharge as shown in Fig. 1. Additionally, this electrodes diameter gives more favourable conditions to have an ignition success (di decreases slightly with d) w [g water /kg dry air ] =.75 ; T=13 [ C] =1.9 ; T=13 [ C] =.75 ; T=7 [ C] =1.9 ; T=7 [ C] =.75 ; T=3 [ C] =1.9 ; T=3 [ C] Exponential Fit =.75 ; T=19 [ C] =1.9 ; T=19 [ C] =.75 ; T= [ C] =1.9 ; T= [ C] =.75 ; T=35 [ C] =1.9 ; T=35 [ C] Exponential Fit Fig.1: The influence of the humidity of the air, mixture temperature, equivalence ratio and electrodes diameter on the critical spark distance. Working conditions: Propane-air mixtures, U=.9m/s. For the choice of the electrode spacing was used the criterion that in order to ignite a mixture using a spark ignition system, the electrodes spacing must be di<d<ds. The experiments conducted at different temperatures and humidity ratios of the air revealed that the humidity ratio have a significant effect on the ability of the system to have a spark discharge. It was observed that by decreasing the humidity ratio, ds decreases significantly, for a fixed mixture velocity of.9m/s and electrode diameter of.mm, causing an important reduction of the range of the electrodes spacing requiregniting a mixture, as shown in Fig.11. Therefore, in order to prevent spark failure in dry mixtures (worst condition) using the actual parameters of the proposeystem, d=.5mm and prim=1.7, the electrode spacing of the proposeystem is.mm. Concluding, the ignition probability of proposed pilot flame system found to be 1% for range of Qpropane from.1 to.slpm, ant is independent of the time lag between the moment of beginning of the fuel injection anpark discharge, Δt. Fig. 1 presents the comparison between the ignition probability of the proposed and current systems as a function of Δt. d=.mm d=.5mm 7

8 Ignition Probability [%] (a) (b) (c) (e) (d) Dry bulb Temperature [ C] Figure 11: Effect of the humidity on the working area in propane-air mixtures. a) T=7.5 C, RH=9%, w=19.1 gwater/kgdryair b) T=7.5 C, RH=%, w=1. gwater/kgdryair c) T=.5 C, RH=%, w=. gwater/kgdryair d) T=13. C, RH=%, w=3. gwater/kgdryair e) T=3. C, RH=5%, w=7.1 gwater/kgdryair Working Conditions: Propane-air mixtures U=.9m/s and d=.mm 1 Proposal pilot flame system Current pilot flame system Humidity Ratio [kgwater/kgdryair] Secondly, an experimental study was performed to evaluate the effect of mixture properties and electrode parameters on the success of spark discharge (occurrence of a spark discharge) and on the success of ignition (sustained flame propagation after a spark discharge). Finally, based on all these results a new pilot flame ignition system was proposed and experimental characterized/tested. A new pilot flame system was proposed based on a new electrodes arrangement and on a new pilot tube geometry. In this system the spark is dischargenside the pilot jet (in contrast with the current system) which has a primary equivalence ratio, prim, of 1.7, while the current system works with prim=.7. The proposed pilot flame has 1% of ignition probability using a single spark discharge. The choice of the electrodes diameter, d, and the electrode spacing, d, for the proposeystem were based on the experimental analysis performed to evaluate the effect of the mixture and electrode parameters on the ability of the system to have spark/ignition, where has been shown that the critical spark distance is significantly influenced by the electrodes diameter (decreasing with it) and humidity ratio (increasing monotonically with it) Time lag [s]. References Figure 1: Comparison between the ignition probability in proposed and current systems. 5. Conclusions The objective of the present work was to study in detail a pilot ignition system (commercially available) in order to identify the causes that may contribute to the non success of ignition and to propose a new pilot flame system with a higher ignition ability. In order to accomplish this objective, first the current pilot flame system was submitted to an experimental characterization. 1 Lewis B, and von Elbe G (191) "Combustion, Flames and Explosion of Gases". Academic Press, nd Edition Swett C C (195) " Ignition of flowing gases". NACA Report 17 3 Raizer P (1997)"Gas Discharge Physics". Springer Verlag Ballal D R, and Lefebvre A H (1975) "The influence of spark discharge characteristics on minimum ignition energy in flowing gases. Combustion and Flame, : Kono M, and Kumagai S, Sakai T (197) The optimum condition for ignition of gases by composite sparks. Proceedings of the Combustions Institute, 1: 757-7

9 Ballal D R, and Lefebvre A H (197) "The influence of flow parameters on minimum ignition energy and quencinhg distance". Proceedings of the Combustions Institute, 15: Belles F E, and Swett C C (1957) Ignition and flammability of hydrocarbon fuels". NACA Report 13 Calcote H F, Gregory C A, Barnett C M, and Gilmer R B ignition - Effect of molecular structure. Industrial and Engineering Chemistry, : 5-9 Fenn J B Lean inflammability limit and minimum spark ignition energy. Industrial and Engineering Chemistry, 3: iegler G W, Wagner E P, and Maly R R (19) "Ignition of lean methane-air mixtures by high pressure glow and arc discharges", Proceedings of the Combustions Institute : Arcoumanis C, and Kamimoto T (9) "Flow and Combustion in Reciprocating Engines". Springer- Verlag 1 anta W J, and Smith R A (197) Measurements of Turbulent Transport Properties with a Laser Doppler Velocimetry. 11th AIAA Aerospace Science Meeting 13 Dixon W J, and Massey F J (193) "Introduction to statistical analysis". McGraw-Hill, th edition 9