Recent studies point at the significant decrease in NOx emanations in thin premixed gas turbine combustors. However, combustors of this type pose the built-in hazard of thin runaway ( LBO ) . This calls for an online early warning system that can be integrated with the combustor controls to guarantee efficient and uninterrupted operation. High dependability and a speedy response clip of such systems are necessary for most applications, and are peculiarly important in aircraft propulsion applications. In this survey we propose and investigate fresh techniques aimed at developing an efficient LBO anticipation system. The work characterizes the behaviour of a whirl stabilized, LPG-air fueled, dump plane combustor fire near its LBO bound in footings of CH* chemiluminiscence ( optical ) emanation strength, entire emanation strength and fire colour analysis. We investigate the influence of fuel-air equality ratio and the commixture distance ( Lfuel ) in the premixer on the efficaciousness of the anticipation methods. The acquired information points at the happening of assorted burning manners in conformity with the equality ratio and fuel supply status. Several techniques based on Fourier spectral analysis, statistical analysis and direct natural signal thresholding are presented for different burning manners which analyze optical signal and transform it to some suited quantitative index for feeling the propinquity of thin runaway. It has been observed that the different detection technique discussed here behave otherwise with regard to the commixture length provided for fuel and air. A elaborate priori function of LBO for a fixed combustor constellation is required for make up one’s minding the quantitative index for peculiar detection technique. A new RGB analysis technique has been explored for feeling the signature of LBO based on the image processing technique carried on the direct digital picture taking of the fire. It is observed that for all the different ports conditions which gives different commixture distances, the Red/ Blue ( R/B ) colour strength ratio decreases significantly towards LBO. This ratio could be a more suited LBO precursor for feeling the inchoate tilt runaway.
Keywords: Lean blow out ( LBO ) , Combustor, Image processing, Chemiluminescence, Statisticss
The usage of uninterrupted burning processes is inevitable in the present twenty-four hours industries and transit systems. The turning usage of burning put a heavy punishment to the environment in the signifier of more pollutant emanation ( NOx ) . To accomplish the lowest possible NOx, modern combustors are designed to run in a thin premixed or partly premixed manner [ 1 ] that operate at low temperature. However, both premixed and partial premixed combustor are prone to built-in hazard of thin blow out ( LBO ) as combustors are made to run at an equality ratio ( I¦ ) in the propinquity of their LBO bound ( I¦LBO ) . When sufficiently thin fires are capable to power scene alterations, flow perturbations or fluctuation in fuel composing, the ensuing equality ratio ( I¦ ) disturbances may do loss of burning. Such a runaway could do loss of power and expensive down clip in stationary turbine, which increases care cost and cut down productiveness and engine life. The job is peculiarly terrible for aircraft engines where burning is the ultimate beginning of engine push, thin runaway ( LBO ) causes loss of push and therefore posses a important safety jeopardy. This is particularly of import when power is reduced during attack and landing. To avoid such scenarios, current systems are typically operated with a broad border above the unsure LBO bound. Enhanced public presentation will necessitate a decrease of this border. Therefore, the ability to feel the LBO signature can supply important final payments. For illustration, a detection system with appropriate sensitiveness, dependability and clip response could be used in an active control system that would let combustors to run at thin conditions compared to the present combustors, thereby cut downing NOx emanation without compromising safety.
In the past much more extended surveies were conducted for the stableness of combustors. Stability of a combustor can be loosely classified into inactive and dynamic stablenesss. Dynamic stability trades with the ego excited oscillations produced due to matching between heat release fluctuations with the force per unit area oscillations [ 2 ] . A combustor is considered to be statically unstable when the fire can non be held inside the combustor. Therefore at Flashback or runaway, the fire becomes statically unstable. Inactive stableness boundaries are normally dependent on runing parametric quantities like flow rate, temperature and equality ratio ( I¦ ) . Dynamic instability can do fluctuations in some of these parametric quantities and can therefore drive the combustor runing conditions towards its inactive stableness bounds. Likewise, inactive instability can do alterations in fire form and therefore impact the heat release distribution inside the combustor and therefore drive dynamic instabilities. There is much research in the field of commanding dynamic instability [ 3, 4 ] , there is less work in understanding complex combustors near their inactive stableness bounds ( flashback or blowout bounds ) . Although some old surveies shed the visible radiations on fire kineticss near inactive stableness bounds ( LBO bound ) like there have been several surveies in simple fire that indicate there is a local loss of fire that grows to finish loss of fire [ 5, 6 ] . De Zilwa et. Al. [ 7 ] investigated fire kineticss near to runaway in shit stabilized combustors with and without low whirl. They noticed really low frequence oscillations as the rich and thin extinction bounds were approached. Murugandam and Seitzman [ 8 ] observed the extinction and reignition events prior to LBO in whirl stabilized, methane fueled gas turbine combustor. There are some more surveies which reported flame instability or transeunt behaviour near to blowout [ 9, 10 ] .
This passage behaviour, which appears to predate blow out could be used in the yesteryear for bring forthing the LBO precursors. The pick of non-intrusive detector is best suitable which is outside from the rough environment of combustor and captured electromagnetic and acoustic radiation produced within the combustor. Muruganandam et. Al. [ 11 ] used OH* chemiluminescence to place the LBO precursors which are short continuance, localized fire extinction and reignition events in premixed, whirl stabilized, methane fueled combustor. They used those precursor events to develop the detection methods which are incorporated in active control system to extenuate the impended runaway. Yi and Gutmark [ 12 ] used the optical emanation for feeling the LBO precursors in partial premixed, liquid fueled, multiswirl stabilized gas turbine combustor. There are few more surveies which used acoustic radiation instead than optical emanation like Mukhopadhyay et. Al. [ 13 ] identified traveling kurtosis and threshold based techniques for early sensing of fire extinction by utilizing clip series informations of combustor force per unit area. Shashvat prakash et. Al. [ 14 ] incorporated acoustic based detection technique with active control system for extenuation of impended runaway. It is observed that in all the old surveies the focal point is chiefly on the word picture of fire kineticss towards blowout bound for fixed constellation of combustor either premixed or partial premixed. A study of old surveies [ 3, 15 ] reported that the equality ratio fluctuation and the unmixedness of fuel and air influence a heat release fluctuation which is the primary cause of dynamic instability and later which affects the inactive bound ( LBO bound ) of fire. As per our observation no survey has been reported yet where the consequence of fuel unmixedness has been seen on the behaviour of feeling techniques, which motivate us to set about experimental survey on a whirl stabilized, LPG ( 60 % propane and 40 % butane ) and air fueled, dump plane combustor, in order to clarify the effects of fuel and air unmixedness on the fire kineticss near blowout bound and later asses the behaviour and effectivity of different feeling techniques in the following subdivision.
Fig.1 illustrates the experimental apparatus, which consists of three major elements: a combustor, mass flow control devices for fuel and air and fire imagination system consist of PMT and camera. The non-vitiated burning air is supplied at ambient temperature from compressor to the bottom port provided on premixing tubing and metered upstream of the combustor utilizing graduated mass flow accountant ( MFC ) ( Aalborg range 0-500 LPM ) . The fuel ( LPG ) is supplied from a pressurized cylinder fitted with needle valve to command the flow rate and metered upstream of the combustor by graduated Aalborg mass flow accountant ( RANGE 0-10 LPM. ) . The fuel is injected to the different side ports provided at different locations on premixing tube gives different blending lengths ( Lfuel ) . The injected fuel mixes with air as it passes through the commixture distance ( Lfuel ) . Finally fuel-air mixture entered the combustor through an recess swirler in the ring around a Centre organic structure, located merely prior to dump plane in the premixing subdivision. The interior diameter of the premixing tubing is 2.3 centimeter, and the diameter of the centre organic structure is 0.8 centimeter. The recess swirler has six vanes positioned at 600 to the flow axis. The vitreous silica tubing is provided in a burning zone holding internal diameter 6 centimeter. and length 20 centimeter. The quartz tubing facilitates the optical nosologies of the burning.
The line of sight heat release measuring of burning was obtained by mensurating the chemiluminescence strength from CH* groups of fire with photomultiplier tubing ( PMT ) of Hamamatsu make ( pattern 931B ) , PMT is fitted with a 10nm bandwidth filter centred at 432nm to let emanations from CH* groups ( 432nm ) entirely to make the PMT. This PMT has a built in amplifier ( bandwidth of 20 kilohertzs ) to change over the current to voltage and operates from a 12VDC beginning. The PMT end product signal ( electromotive force ) is conditioned by utilizing NI SCXI 1050 08-channel signal conditioner, later the learned signal from detector was acquired to a personal computing machine utilizing a PXI 8050 National InstrumentsA® 16 spot informations, 06-channel A/D card. A sum of 32768 informations ( N ) ( i.e. , 215 points ) were acquired in each experiment utilizing NI package LABVIEW 7.0A® at a trying rate of 2000 Hz. High-speed images of the fire under stable and close blow-off status were obtained in order to better understand the phenomenology of the flame blow-off procedure, and therefore better capablenesss for construing the optical signature. A camcorder ( SONY, DCR-TRV 300 ) is used to capture the fire picture, subsequently the images are captured from picture at 33 frames per second ( FPS ) . The planetary characteristics of fire were examined utilizing a color digital camera. The aperture, exposure times and focal length of camera could be independently controlled to acquire the right consequences. The colour phenomenology of fire was recorded with the camera.
Figure 1. Conventional diagram of the experimental setup
The combustor was approached to LBO stepwise by commanding the equality ratio ( I¦ ) , which is adjusted via the fuel flow rate while maintaining the air flow fixed. A larger incremental decrease in I¦ of about 0.05 is used when the burning is good above the LBO, whereas a smaller decrease in I¦ of about 0.02 is used when burning is close to LBO.
Five different ports as Port1, Port 2, Port 3, Port 4 and Port 5 has been used for shooting the fuel which later gives different blending lengths ( Lfuel ) as 35cm, 30cm, 25 centimeter, 20cm and 15cm severally for fuel and air. The experiments were carried out by utilizing three different air flow status which gives three different Reynolds Numberss mentioned in table1 of experimental status.
Table 1. Experimental Condition
Number of fuel injecting ports
( 05 ) Port1, Port 2, Port 3, Port 4, Port 5
Blending lengths ( Lfuel )
35cm, 30cm, 25cm, 20cm, 15cm. respective to
Port1, Port 2, A_ _ _ _ _Port 5.
Air flow rate ( Q )
70, 80 and 85 LPM
Flow Reynolds figure ( Re. no. )
6645, 7480, & A ; 7948 severally with air flow rate ( Q )
Sampling rate from PMT
A/D digitisation declaration
16 spot ( +- 5V )
In this subdivision, we describe the phenomenology of the fire runaway procedure. High-speed images are presented in concurrence with coincident ocular informations to help the development of data-analysis strategies described in the following subdivision. Prior survey have observed [ 7, 8 ] towards lean blowout bound, there are random cases where the fire exhibit oscillation in the combustor. These oscillations produce near fire loss event in combustor. The similar fire loss event we observed in the status of port1, port 2 and port 3. One such flame loss event is elucidated here in Fig. 2. which plots a sequence of images captured at 33 frames per second, here port1 is used for shooting the fuel, air flow rate used as Q= 80 LPM and in this instance of experiment blow out occurs at equality ratio 0.75 ( I¦LBO= 0.75 ) . From Fig. 2, it is seen that, the combustor ab initio has a spatially burning zone. Then the fire detaches from the Centre organic structure, demoing weak reaction and moves farther downstream from the combustor recess and stabilise at that place. The fire begins to vanish from field of position and there is about complete loss of fire, proposing extinction event. From the down watercourse the fire packages are convicted back to recess, which reignites the unburned fuel which entered the combustor in old period. The fuel is quickly consumed which exhibit intense burning and the fire is re-establish in the combustor ( reignition event ) . These alone extinction and reignition events span a period of several msecs, and they occur indiscriminately in clip, prior to LBO. As the combustor approached towards LBO bound, the frequence of these events additions and therefore the clip between two such events decreases closer to LBO.
Fig.2. Sequence of high velocity images of planetary fire with I¦= 0.75 for port1. The images are separated by 30 millisecond. The clip sequence of images is from left to compensate on the top followed by left to compensate on the undermentioned rows. The first image shows the screening window for optical detection
Fig.3. Sequence of high velocity images of planetary fire with I¦= 0.51 for port 4. The images are separated by 30 millisecond. The clip sequence of images is from left to compensate on the top followed by left to compensate on the undermentioned rows.
Time ( s )
Figure 2. Time series informations of CH
chemiluminescence signal ( port1 )
Fig.5. Time series informations of CH chemiluminescence signal for for port2, port3, port4, and port5 for off and nearer to LBO status ( Read Top to Bottam )
Fig.4 secret plans typical measured clip dependance of the CH* chemiluminiscence at the equality ratios i.e. I¦ = 0.81 and I¦=0.75 stand foring the fire status far from and one near to LBO severally. As expected, the average strength decreased with lower equality ratio, due to reduced heat release as the fuel flow is reduced. Besides for the thin instance i.e. I¦=0.75, the chemiluminescence strength shows some comparatively high amplitude explosions ( unsteady events ) with the signal traveling both below and good above the average value and on occasion beads to near zero value. To exemplify this behavior in item, a expanded secret plan of these events are besides shown. Often, these explosions or events are characterized by an about complete loss of chemiluminiscence signal quickely followed by intense emanation from a imaged part. Thease bursts or events coincide with the happening of fire loss ( exctinction ) and re-ignition events described in the high-velocity picture images. Similarly Fig. 3 shows high velocity images towards blowout bound for port 4 for the same experimental status used in depicting high velocity images in port 1. The similar manner we observed in port 5 instance besides. From Fig. 3 it is observed that the fire in port 4 and port 5 does non oscillates in the full combustor and instead the oscillations are limited to dump plane so that the fire loss are non so important in this instance.
Figure 5 plots the CH* chemiluminiscence fluctuations for port 2, port 3, port 4 and port 5 for rich and one stopping point to tilt runaway equality ratio. From Fig. 5, it is clearely seen that for port 2 and port 3 instance, the strength of chemiluminiscence explosions ( unstedy events ) is more intense relation to intend and on occasion beads to zero value ( extinction ) as compared to the CH signal for port 4 and port 5 instance. These explosions lucifers with the fire loss and reignition events due to intense oscillations of fire in full combustor seen in port 2 and port 3 instance which is non so important in port 4 and port 5 instance.
5.LBO Feeling Schemes:
Developing LBO feeling schemes with maximal sensitiveness, velocity and hardiness requires a thorough apprehension of the fire features prior to runaway. The high-velocity fire images obtained and analyzed in concurrence with coincident ocular informations and the passage of coloring material phennomelogy prior to blow off were used to help the development of such scheme. A simple attacks are considered like statistical, Fourier spectral and direct natural signal ( Thresholding ) in assorted confrugation of combustor obtained by altering the commixture length available for fuel and air. Apart from this a new image processing based technique, RGB analysis has been established for feeling the signature of blowoff in the following subdivision.
5.1 Spectral Aanalysis
Fig. 6 shows the Fourier power spectrum of the CH* chemiluminiscence signal for I¦= 0.81, 0.77 and 0.75 after excepting the D.C. signal. These curves have been normalized at each I¦ to hold the same entire power. Here air flow rate Q= 80LPM, N= 32768, I¦LBO = 0.75. From Fig. 6, it has been observed that the fractional energy in low frequences appears to increase as LBO bound is approached. A cumulative distribution map I? ( I ) is defined to quantify the energy distribution with the frequence [ 12 ]
( K= 1,2, aˆ¦aˆ¦aˆ¦.1000 ) ( 1 )
Where Hk refers to the amplitude of CH* chemiluminiscence at the kth pin in fast Fourier transform ( FFT ) analysis. Fig. 7 shows that the low frequence constituent of CH* chemiluminescence have a much higher energy per centum at I¦= 0.75 and 0.77 than that at I¦ = 0.81. It was observed that the fraction of low frequence content additions quickly as the combustor approached towards LBO. Thus the propinquity to LBO can be sensed by quantifying the comparative sum of low frequence content in the signal. Using two filters, a set base on balls filter for mensurating the low frequence content and high base on balls filter for taking the DC signal can carry through the practical execution of this attack. However, supervising the lower frequence content require longer times, which will halter the clip response of the control system.
Fig.6. Fourier power spectrum of CH emanation signal for different I¦ for port1
Fig.7.Percentage of cumulative energy with frequence for port 1 trial
In order to asses the consequence of fuel unmixedness on the behaviour of intensified low frequence content which decides the feasibleness of implementing this attack for feeling the propinquity of thin runaway in different constellation of combustor obtained so far. Trials were carried on utilizing different ports for shooting fuel which give different commixture distances for fuel and air. Figure 8 shows the different secret plans for per centum of the cumulative energy for port 2, port 3, port 4 and port 5. From Fig. 8, it is observed that for port 2 and port 3 the tendency of increasing fractional energy in low frequence constituent is more or less similar to port 1 consequences. For port 4 and port 5 the overall fractional energy of CH* chemiluminiscence increased towards thin blowout bound ( I¦LBO ) but low frequence constituent is non so intensified. The ground behind this may be, remember the consequences of high velocity fire images where we observed that towards LBO bound for port 4 and port 5 the fire does non hover significantly in the full combustor instead the oscillations are limited to dump plane merely. For port 1 and port 2 conditions towards LBO bound, the important oscillation of fire in full combustor creates low frequence fire extinction and reignition events which later give intensified low frequence constituent. The above observation is similar to anterior surveies. For illustration Nicholson and Field noted big graduated table, low frequence fire oscillations every bit good as periodic withdrawal of the fire from its fire holder near the LBO bound. Muruganandam et.al. [ 16 ] observed low frequence power content of both acoustic and OH spectra were higher in a thin fire.
The above analysis reveled that spectral attack for foretelling the propinquity of thin runaway is more suited in more premixed constellation combustor like port 1, port 2 and port 3 where the distances provided for blending fuel and air is more.
Fig.8. Percentage of cumulative energy with frequence for port 2, port3, port4 and port5 place ( read left to compensate )
As an option to frequence sphere analysis, here the clip series informations of CH* fluctuation captured on PMT is proved to be utile to examine the near-LBO burning kineticss. Fig. 4 shows a typical CH* chemiluminiscene fluctuations of fire for the one stable equality ratio far from LBO i.e. I¦= 0.81, and one stopping point to LBO i.e. I¦ = 0.75. As discussed earlier, the average emanation signal decreases with decreased I¦ , as the fuel is reduced. To quantify the high amplitude, short continuance explosions ( unsteady events ) , which matches with the fire extinction and reignition events near thin blow off status which nowadays in the CH* signal shown in Fig. 4, a statistical analysis is used. In the past Murugandam et. Al. [ 16 ] sucessesfully used the statistical minutes like standard divergence, 2nd minute, 4th minute and kurtosis for foretelling the incipient runaway in a premixed, whirl stabilized, methane fueled dump combustor in existent clip utilizing OH* chemiluminiscence.
Similarly Yi and Gutmark [ 12 ] used normalized root mean square divergence for early anticipation of thin runaway in a multiswirl stabilized, turpentine fueled atmospheric gas turbine combustor. The same attack has been explored here in more inside informations so that to detect the consequence of unmixedness in the undermentioned subdivision.
The normalized chemiluminescence RMS which is called as coefficient of scattering ( C.D. ) in statistics. Which is the ratio of standard divergence ( RMS ) to the
C.D. = RMS/ I? ( 2 )
chemiluminescence mean ( I? ) is plotted with the equality ratio normalized by its value at blowoff i.e. ( I¦/I¦LBO ) . Fig. 9 compares C.D. for different flow conditions, correspondingly gives different Reynolds Numberss when port 1 is used for shooting the fuel. C.D. increases easy with diminishing equality ratio until I¦/I¦LBO & gt ; 1.15. This tendency has been confirmed by other researches [ 17, 18 ] , who observed that the heat release RMS was approximately relative to the mean heat release rate. C.D. alterations little in the big scope of I¦ and hit up as the combustor approached towards LBO. Fig.11 shows the fluctuation of chance denseness map ( PDF ) utilizing normal distribution calculated for chemiluminiscence which is normalized by its mean, it has been observed that the PDF is about invariant for port1 until I¦/I¦ LBO & gt ; 1.15 ( I, vitamin E. rich status ) which makes C.D. about changeless for I¦ / I¦ LBO & gt ; 1.15.
Fig.9. Variation of C.D. for different flow conditions for port 1.
Fig.10. Effectss of N on C.D.
Figure 11. Variation of PDF for normalized chemiluminiscence/ mean for port 1 and port 4 ( L to R )
For the computation of C.D we need sample N. Because of the helter-skelter and apparently irregular characteristic of close LBO burning kineticss, a larger N is preferred. However, a larger N will necessarily decelerate down the observing velocity. Fig.10 shows that even a smaller sample window is capable of accomplishing the sensible balance between the truth and the observing speed.. To originate the LBO warning signal which feed to active control system for commanding the at hand runaway, one needs some quantitative index. After carefully detecting the fluctuation of C.D. for port 1 instance, the undermentioned status found to be more suited. On fulfilling the below status the LBO warning should be initiated [ 12 ]
( 3 )
It is found that the value of I± =1.75 is more suited in this type of premixed combustor confrugation where port 1 is used supplying maximal mixing length ( L fuel= 35 centimeter. ) .
Trials were carried on the different ports which subsequently gives different blending distances reference in table1. The same experimental conditions were used merely the I¦ LBO value has been changed for port to port. The fluctuation of C.D with I¦ / I¦ LBO has been plotted for port 2, port 3, port 4 and port 5 shown in Fig. 12. It has been seen that the tendency of fluctuation of C.D. in port1 instance has non been repeated in other ports. The PDF of normalized chemiluminiscence which about invariant in port1 status which makes the C.D is about changeless until I¦ / I¦ LBO & gt ; 1.15 is no longer seen in other ports. Fig 11 shows the fluctuation of PDF utilizing normal distribution for normalized chemiluminiscence by its mean for port 4 where it is vary and the standard divergence is continuously increased in rich instance i.e until I¦ / I¦ LBO & gt ; 1.15.
Hong et. Al. [ 15 ] by experimentation demonstrated that the commixture length is a important parametric quantity for triping dynamic instability and observed different manners of force per unit area fluctuation with regard to unmixedness of fuel which is controlled by blending length. So mixing length is of import parametric quantity which affects the amplitude of force per unit area fluctuation and later heat release rate. That may be the ground to obtained different behaviour of C.D. for different ports. Please note the focal point of our paper is chiefly on the behaviour of feeling technique instead than the survey of dynamic instabilities triggered due to plain cape.
The status which we mention for port 1 to originate the LBO warning signal and the threshold value 1.15 up to which the C.D. is about changeless and the value of I± =1.75 would non be suited for other ports where the commixture distance available for fuel and air is less. In order to do this detection technique suited for partial premixed constellation, item priori LBO function of the combustor should be needed so that to happen the suited value for threshold and I± .
Fig.12.Variation of C.D. for different flow conditions for port 2, port3, port4 and port5 ( Read L to R )
5.3. Direct signal analysis ( Thresholding attack )
The major restriction of predating statistical method was, it gives some mean step of divergence of clip trace signal of chemiluminiscence from its mean. Here, the clip localised jaunts of CH* signal is used to place the precursor events. The simple threshold attack is used [ 11 ] . The precursor event is acknowledge when the CH* signal beads below some threshold value. For illustration 35 % of CH* mean is taken for the current experiment. The concluding behind this attack is that, the precursor signature is initiated by a local extinction event that temporally lowers the chemiluminiscence. Therefore, the low threshold provides the easiest sensing of the event. Choice of optimal value of threshold for sensing can change depending on the combustor design, the detection volume and the needed sensitiveness of the technique.
Fig. 13 shows the fluctuation of frequence of identified precursor events ( norm over sampling clip. ) over I¦ . It is seen that, this parametric quantity has about zero value for the higher I¦ and increases as the LBO bound is approached. The event frequence parametric quantity gives reasonable quantitative blow out index for control system to feel the propinquity to LBO.
As discussed earlier, these precursor events occurs indiscriminately both in frequence and continuance. To take in to account this two consequence, Yi and Gutmark [ 12 ] computed and used normalized cumulative continuance of LBO precursor events. The same attack has been used here.
I? = Nt / N ( 4 )
Where Nt denotes entire precursor events observed in sample N. Fig. 14 shows the fluctuation of I? with I¦/ I¦LBO for different flow conditions for more premixed constellation ( port1 ) . Similarly to C.D shown in Fig..9, I? increases really easy until I¦/ I¦LBO & gt ; 1.15 and increases significantly as the LBO bound is approached. , the ground behind this is that the C.D. and the I? are mathematically mutualist, as we have seen that the fluctuation of PDF for chemiluminescence /mean ( ( normalized ) is invariant until I¦ / I¦ LBO & gt ; 1.15. The I? is the cumulative chance for chemiluminiscence / mean & lt ; 0.35. For bring forthing LBO warning signal, the same status given below which used earlier in statistical method found to be more suited.
( 5 )
The value of I± = 1.75 is more suited for port1. Similar attack has been tried for other ports in order to see the consequence of fuel unmixedness on behavior I? . Fig. 15 shows the fluctuation for I? index for port 2, port 3, port 4 and port 5 for different flow conditions. As I? and C.D. are mathematically mutualist. The fluctuation of I? lucifers with the fluctuation of C.D. plotted earlier in statistical attack. The above status mentioned for port 1 for triping the LBO warning signal and the threshold value 1.15 up to which the I? is about changeless and the value of I± =1.75 could non be found suited in other ports. Similar to statistical attack here besides, in order to do this attack suited for partial premixed constellation where commixture distances for fuel and air is less, we need a elaborate prioro LBO function of the combustor in order to happen out the new value for threshold and I± .
Fig.13. Dependence of figure of CH based events upon I¦ for port1
Fig.14.Variation of I? with I¦/I¦LBO for port 1
Fig.15.Variation of I? with I¦/I¦LBO for port 2, port3, port4 and port5 ( Read L to R )
5.4 Image processing technique ( RGB analysis )
For finding the propinquity of combustor to tilt runaway. Here a new attack has been explored based on the image processing technique carried on the direct digital picture taking of the fire. Bing a camera located outside the rough environment of combustor, this attack besides fulfils the demand of non intrusive imagining. For researching this detection technique in more item so that to recite the consequence of fuel unmixedness on the feasibleness of implementing this technique in assorted constellation of combustor. Similar to predating discussed methods here besides five different fuel injecting ports have been used which later gives different commixture distances ( Lfuel ) for air and fuel.
Figure 16. Variation of Red & A ; Blue strength with fire manners for different Equivalence ratio ( port 1 )
For each port the experiment has been carried out for three different air flow rates ( Q ) correspondingly gives three different Reynolds nos. mentioned in table 1 of experimental status. The equality ratio ( I¦ ) is varied stepwise by altering the fuel flow rate while maintaining the air flow rate ( Q ) invariable. A larger incremental decrease in I¦ of about 0.05 is used when burning is good above the LBO, whereas a smaller incremental decrease in I¦ of about 0.02 is used when burning is close to LBO. The feeling methodological analysis developed here is based on the image processing technique. At each equality ratio ( I¦ ) status four exposure of the fire modes has been taken. The algorithm was developed to read the images of fire at each fluctuations of I¦ . The mean strength of Red and Blue coloring material is calculated for the corresponding equality ratio ( I¦ ) . Figure16 shows the different manners of fire obtained on changing the I¦ from rich to tilt blowout status. Here the air flow rate used is Q=80 LPM and port 1 is used for shooting the fuel which gives maximal commixture length ( Lfuel= 35 cm. ) . In the same secret plan the fluctuation of Red and Blue colour strength is besides plotted. From Fig.16, it is observed that the ruddy strength is bit by bit decreased while the bluish strength is increased as the combustor approached towards blowout from rich status. The ground behind this is towards thin conditions the less fuel with fixed air flow gives more premixed fire. In premixed fire the bluish coloring material is dominant, besides soot per centum is about range to zero, which is indicated by ruddy coloring material. The passage of coloring material phenomenology which precedes the runaway could be used to give the LBO precursor for feeling the propinquity of runaway.
Fig.17 plots the fluctuation of Red/Blue ( R/B ) colour strength ratio with equality ratio ( I¦ ) for port 1 where Lfuel= 35 centimeter for three different air flow rates giving different Reynolds figure at the recess subdivision of combustor ( Re. no.= 6545, 7480, 7948 ) . Again for one individual air flow rates three different tally has to be taken so that to gauge the per centum mistakes. Mistake bars are plotted for all the different air flow rates conditions seen in Fig.17. From the per centum mistake appraisal, it is found that the mistake per centum for most of the readings lies in the scope of + – 2 to 8 % for all the three different flow conditions. From Fig. 17, it is observed that, as the combustor attack towards LBO in all the three variable air flow rate conditions, this ( R/B ) colour strength ratio drops significantly, which could gives a suited quantitative index for feeling the propinquity of combustor to LBO.
Fig.17. Variation of R/B ratio with different ( I¦ ) for port 1 ( Lfuel= 35 centimeter. )
To happen out the bead in per centum of the R/B coloring material strength ratio from rich ( stochiometric ) status to tilt blowout bound. The R/B ratio at each I¦ is normalised by the maximal value of R/B ratio brush in that tally which is normally seen near to stochiometric status ( I¦=1.0 ) . Figure 18 secret plans such fluctuation of normalized R/B ratio with normalized equality ratio by I¦LBO i.e. ( I¦/ I¦LBO ) for port 1 instance. The same experimental information has been considered which we used earlier for plotting Fig.17. From Fig.18, it is observed that when the combustor range to near LBO limit the per centum bead in R/B ratio is 85 % or instead in other words we can state near LBO the R/B ratio get 15 % of the maximal value of R/B ratio brush in that tally. It is seen that in all the three different air flow conditions the R/B ratio get more or less same value i.e. 15 % of ( R/B ) Max.
The above observation could give a value of threshold degree which can be used for feeling the propinquity of combustor to tilt runaway with active LBO accountant. The more suited threshold found in this instance of port1 consequence is 20 % ( R/B ) MAX, which fulfil the standards of neither it is really close to LBO bound nor it is excessively off from it. So foretelling the incipient runaway in existent clip the undermentioned status is recommended. The LBO warning be initiated on fulfilling the below status. For ciphering the R/B coloring material strength ratio really few calculation is required so the existent clip demand is satisfied.
( 5 )
As discussed earlier the most suited value we found out for Threshold in port 1 instance is 0.20, one time the LBO warning signal is triggered, the active control system take appropriate action for heightening the stabilisation of fire in combustor so that the incipient runaway could avoided and LBO bound can be extended.
To clarify the consequence of unmixideness of air and fuel ( premixing per centum ) the technique is tested for different ports which gives different commixture distances for air and fuel. Figure 19, 20, 21 & A ; 22 shows the normalised ( R/B ) strength ratio verses the normalised I¦/ I¦LBO for port 2, port 3, port 4 & A ; port 5 severally. The commixture distances i.e ( Lfuel ) for the several ports is mentioned in the table1 of experimental parametric quantities.
Fig18. Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for port 1 ( Lfuel= 35 centimeter. )
From Fig. 19, where the fluctuation of normalized R/B strength is plotted for port 2, it is observed that similar to port 1 consequence, the R/B ratio bead significantly as the combustor attack near to LBO. The lone fluctuation from port1 consequence is in per centum bead of R/B strength, here this value is somewhat decreased. Near LBO bound ( I¦LBO ) , R/B ratio reaches the value in the scope of 20 to 25 % . of the maximal value of R/B ratio encountered in that tally which we found at stochiometric status ( I¦=1.0 ) in all the three different flow conditions.
Similarly when we observed Fig. 20, 21 & A ; 22 we found that the per centum bead in the R/B strength ratio is somewhat decreased as the commixture distances is reduced severally. Near thin runaway bound the R/B strength value obtained in the scope of 20 to 25 % of the maximal R/B ratio value for port 3. Similarly for port 4 it lies in the scope of 30-35 % and port 5 where the commixture distance is minimal it lies in the scope of 35-40 % for all the three different flow conditions which we tried for each port. In Fig. 23, 24 and 25 we tried to exemplify the fluctuation of normalized ( R/B ) ratio for different ports with regard to fixed flow status. The same experimental information has been used which we discussed earlier for different port status.
From this analysis it is clear that the Threshold value = 0.20 we set in port 1 instance is no longer utile in other ports as the R/B strength ratio acquired grater value near thin runaway bound.
The same detection technique could be used in other ports status merely one has to alter the threshold value. Here we observed the best suited threshold value in port 2 and port 3 is = 0.3, for port 4 = 0.4 and for port 5 it is = 0.45. So merely by altering the value of threshold degree for different constellation of combustor obtained by changing the available commixture length for air and fuel make this technique suited for monitering the propinquity of combustor to incipient thin runaway.
Fig.19 Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for port 2 ( Lfuel= 30 centimeter. )
Fig.20 Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for port 3 ( Lfuel= 25 centimeter. )
Fig.21 Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for port 4 ( Lfuel= 20 centimeter. )
Fig.22 Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for port 5 ( Lfuel= 15 centimeter. )
Fig.23.Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for different ports for Re. no.=6545
Fig.24.Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for different ports for Re. no.=7480
Fig.25.Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for different ports for Re. no.=7948.
Fig.26.Variation of normalized R/B ratio for normalized ( I¦/ I¦LBO ) for port 1 with one snap status.
In all the above experiments for finding the R/B strength fluctuations, four exposure were taken for each manner of burning obtained on altering the equality ratio ( I¦ ) . The algorithm developed reads all the four exposure and cipher the average strength for ruddy and bluish colour. To cut down the calculation clip, here alternatively of four exposure we tried to happen out the R/B strength fluctuations for the different I¦ with one catch shown in Fig. 26.
From fig. 26, it is observed that even, one exposure calculation is capable to bring forth the coveted consequences and accurately feel the propinquity of combustor to incipient thin runaway.
In the preceding discussed feeling technique to bring forth the robust step of the propinquity to LBO, one has to supply the clip window or sample length ( N ) . For ciphering the stastical parametric quantities it takes some clip which hampers the clip response characterstic of feeling technique. This major disadvantage of those discussed methods is considerebly reduced in this image treating RGB technique as the calculation clip for merely reading the images is really less.
In all the above discussed feeling techniques, every bit shortly as LBO warning signal is initiated, active control system should be put in to operation to heighten the fire stabilisation in combustor so that the incipient runaway could avoided and LBO bound can be extended. This can be done by either by increasing the pilot fuel [ 11 ] or perchance by utilizing little amplitude fuel transition [ 12 ] .
In this survey, in order to clarify the effects of fuel unmixedness on the kineticss of combustor near blowout bound and asses the behaviour and effectivity of different feeling techniques for the early anticipation of blow out, an experimental survey was conducted in a whirl stabilized, LPG-Air fueled, dump plane combustor.
The consequences of this survey are as follows.
It is found that with regard to the different fuel shooting status, two different manners of fire oscillations has been observed. In more premixed constellation like port 1, port2 and port3 the fire oscillates significantly in the full combustor as the LBO bound is approached. These flame oscillation produces flame loss ( extinction ) and rapid ingestion of unburned fuel ( reignition ) events which lucifers with the short continuance high amplitude explosions ( unstedy events ) seen in CH* chemiluminiscence ( optical ) signal. In port 4 and port 5 instance the fire osciilations are non so important and it is limited to dump plane merely later the unsteady explosions seen in CH signal is non so intense. It is found that apart from the fuel shooting status these unsteady events increases in frequence and continuance in all instances and can be used as LBO precursors.
The consequence of fuel unmixedness on the behaviour and effectivity of different feeling schemes based on frequence analysis, statistical analysis and direct signal thresholding has been presented. It is observed that the fraction energy in low frequence content increases dramatically as combustor attacks LBO in more premixed constellation like port 1, port 2and port 3 so that this frequence analysis based detection technique is more suited in more premixed constellation of combustor. While covering with the stastical and direct thresholding based technique, two indices i.e. C.D and I? found to be more usefull for feeling the propinquity of combustor to incipient thin runaway. The C.D and I? are mathematically mutualist so that their fluctuation is similar with regard to different port status. It has been observed that the C.D and I? behaves otherwise from port to port status. Before utilizing either of this two indices for LBO anticipation, a item priori function of the LBO of combustor is needed. .
A new RGB analysis technique has been explored for feeling the signature of LBO, based on the image processing technique carried on the direct digital picture taking of the fire. It is observed that for all the different ports conditions which gives different commixture distances, the Red/ Blue ( R/B ) strength ratio decreases significantly towards LBO, This ratio could be found more suited quantitative index for feeling the inchoate tilt runaway.
This work was supported by a grant from Defense Research and Development Organization ( DRDO ) of the Government of India. The writer besides would wish to thank the All India Council for Technical Education ( AICTE ) for supplying chance and family to transport the research.
1. Lefebvre, A.H. , The function of fuel readying in low-emission burning. Journal of Engineering for Gas Turbines and Power-Transactions of the Asme, 1995. 117 ( 4 ) : p. 617-654.
2. Ducruix, S. , et al. , Combustion kineticss and instabilities: Elementary yoke and driving mechanisms. Journal of Propulsion and Power, 2003. 19 ( 5 ) : p. 722-734.
3. Cohen, J.M. , J.H. Stufflebeam, and W. Proscia, The consequence of fuel/air blending on propulsion authorization in an active burning instability control system. Journal of Engineering for Gas Turbines and Power-Transactions of the Asme, 2001. 123 ( 3 ) : p. 537-542.
4. Kelsall, G. and C. Troger, Prediction and control of burning instabilities in industrial gas turbines. Applied Thermal Engineering, 2004. 24 ( 11-12 ) : p. 1571-1582.
5. Gutmark, E. , et al. , SIMULTANEOUS OH AND SCHLIEREN VISUALIZATION OF PREMIXED FLAMES AT THE LEAN BLOW-OUT LIMIT. Experiments in Fluids, 1991. 12 ( 1-2 ) : p. 10-16.
6. Bayliss, A. , G.K. Leaf, and B.J. Matkowsky, Pulsating and helter-skelter kineticss near the extinction bound. Combustion Science and Technology, 1992. 84 ( 1-6 ) : p. 253-278.
7. De Zilwa, S.R.N. , J.H. Uhm, and J.H. Whitelaw, Combustion oscillations near to the thin flammability bound. Combustion Science and Technology, 2000. 160: p. 231-258.
8. Muruganandam. , T.M. and J.M. Seitzman. , CHARACTERIZATION OF EXTINCTION EVENTS NEAR BLOWOUT IN SWIRL-DUMP COMBUSTORS. Aiaa Journal, 2005.
9. Kukuck, S. and M. Matalon, The oncoming of oscillations in diffusion fires. Combustion Theory and Modelling, 2001. 5 ( 2 ) : p. 217-240.
10. Joulin, G. and G.I. Sivashinsky, INSTABILITY-INDUCED QUENCHING OF PLANE NON-ADIABATIC FLAMES. Combustion Science and Technology, 1987. 55 ( 1-3 ) : p. 83-88.
11. Muruganandam, T.M. , et al. , Active control of thin runaway for turbine engine combustors. Journal of propulsion and power, 2005. 21: p. 807-814.
12. Lolo, T.X. and E.J. Gutmark, Real-time anticipation of inchoate thin runaway in gas turbine combustors. Aiaa Journal, 2007. 45 ( 7 ) : p. 1734-1739.
13. Mukhopadhyay, A. , S. Datta, and D. Sanyal, Effects of tailpipe clash on the nonlinear kineticss of a thermic pulsation combustor. Journal of Engineering for Gas Turbines and Power-Transactions of the Asme, 2008. 130 ( 1 ) : p. 9.
14. Prakash. , S. , et al. , Acoustic detection and extenuation of thin runaway in premixed fires. Aiaa Journal, 2005.
15. Hong, J.G. , et al. , Effect of the Unmixedness of Unburned Gases on the Pressure Fluctuations in a Dump Combustor. Energy Fuels, 2008. 22 ( 4 ) : p. 2221-2228.
16. T.M.Muruganandam. , et al. , Optical and Acoustic detection of Lean Blowout Precursors. Aiaa Journal, 2002.
17. Lawn, C.J. , Distributions of instantaneous heat release by the cross-correlation of chemiluminescent emanations. Combustion and Flame, 2000. 123 ( 1-2 ) : p. 227-240.
18. Bloxsidge, G.J. , A.P. Dowling, and P.J. Langhorne, Reheat bombilation: an acoustically coupled burning instability. Separate 2. Theory Journal of Fluid Mechanics, 1988. 193: p. 445-473.