Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
FLAME INSTABILITY DETECTOR
Document Type and Number:
WIPO Patent Application WO/2014/099477
Kind Code:
A1
Abstract:
Systems and methods for detecting an instability associated with at least one burner are disclosed. A detector (202) measures a signal associated with a characteristic of the at least one burner. The signal is converted (204) to a time-varying signal spectrum using at least one processor. An instability is detected (206) based at least in part on the time-varying signal spectrum. The instability can be detected based on an instability indicator calculated based at least in part on the time-varying signal spectrum. A threshold can be associated with the instability indicator such that an instability is detected when the instability indicator is greater than the threshold.

Inventors:
LI WEICHANG (US)
DOBBS GARY T (US)
GRENDA JEFFREY M (US)
HERHOLD AMY B (US)
CHHOTRAY SAN (US)
MCGREGOR DUANE R (US)
SONG LIMIN (US)
Application Number:
PCT/US2013/074057
Publication Date:
June 26, 2014
Filing Date:
December 10, 2013
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
EXXONMOBIL RES & ENG CO (US)
International Classes:
G05B23/02
Foreign References:
EP2309186A22011-04-13
JPH0814550A1996-01-19
US20100151397A12010-06-17
Other References:
ZENGSHOU DONG ET AL: "Flame stability recognizing of HTAC based on texture and structure analysis", INTELLIGENT CONTROL AND AUTOMATION, 2008. WCICA 2008. 7TH WORLD CONGRESS ON, IEEE, PISCATAWAY, NJ, USA, 25 June 2008 (2008-06-25), pages 5040 - 5044, XP031301713, ISBN: 978-1-4244-2113-8
POINSOT T ET AL: "ACTIVE CONTROL : AN INVESTIGATION METHOD FOR COMBUSTION INSTABILITIES", JOURNAL DE PHYSIQUE III, EDITIONS DE PHYSIQUE, PARIS, FR, vol. 2, no. 7, 1 July 1992 (1992-07-01), pages 1331 - 1357, XP000307090, ISSN: 1155-4320, DOI: 10.1051/JP3:1992103
Attorney, Agent or Firm:
SHANLEY, Matthew, T. et al. (1545 Route 22 EastP.O. Box 90, Annandale NJ, US)
Download PDF:
Claims:
CLAIMS

What is claimed is:

1. A method for detecting an instability associated with at least one burner based on the spectral entropy of a signal measuring a characteristic of the at least one burner.

2: A method for detecting an instability associated with at least one burner, comprising:

obtaining a signal from a detector, the detector measuring at least one characteristic associated with the at least one burner;

converting, using at least one processor, the signal into a time-varying signal spectrum; and

detecting, based at least in part on the time-varying signal spectrum, an instability associated with the at least one burner.

3 : The method of Claim 1 or Claim 2, wherein the instability is detected based on a single channel of data.

4: The method of Claim 1 , 2, or 3, wherein the characteristic comprises a pressure metric.

5: The method of any of the foregoing claims wherein the detector comprises a dynamic pressure sensor.

6: The method of any of the foregoing claims, wherein the characteristic comprises a fluctuation metric.

7: The method of any of the foregoing claims, wherein the detector comprises a device that captures video frames,

8: The method of any of the foregoing claims, further comprising converting a video frame into a single scalar to produce a scalar times series signal.

9: The method of any of the foregoing claims, wherein the characteristic comprises a vibration metric. 10: The method of any of the foregoing claims, wherein the detector comprises an accelerometer.

1 1 : The method of any of the foregoing Claims, wherein the time- varying signal spectrum comprises a spectrogram.

12: The method of any of the foregoing Claims, wherein converting the signal comprises using time-frequency analysis.

13: The method of Claim 12, wherein the time-frequency analysis comprises a short-time Fourier transform.

14: The method of any of the foregoing Claims, wherein detecting an instabilit '- associated with the at least one burner comprises computing a spectral entropy based on the time-varying signal spectrum.

15: The method of any of the foregoing Claims, wherein detecting an instability associated with the at least one burner comprises determining an instability indicator.

16: The method of Claim 15, wherein the instability detector corresponds to a probability of instability.

17: The method of Claims 15 or 16, wherein determining an instability indicator comprises converting the time-varying signal spectrum into the instability indicator.

18: The method of Claim 17, wherein converting the time-varying signal spectrum comprises normalizing the time-varying signal spectrum,

19: The method of Claims 17 or 18, wherein converting the time- varying signal spectrum comprises calculating a probability mass function based on the time-varying signal spectrum,

20: The method of Claims 17, 18, or 19, wherein converting the time -varying signal spectrum comprises computing a Shannon entropy of a probability mass function as a function of time.

21 : The method of Claims 15, 16, 17, 18, 19, or 20, wherein the instability indicator comprises an inverse of a spectral entropy,

22: The method of any of the foregoing claims, wherein detecting an instability comprises comparing the instability indicator to a threshold value. 23: The method of Claim 22, wherein the instability is detected when the instability indicator exceeds a threshold value.

24: The method of any of the foregoing Claims, further comprising providing an alarm when the instability is detected.

25: The method of any of the foregoing Claims, further comprising taking corrective action when the instability is detected.

26: The method of Claim 25, wherein the corrective action comprises adjusting an operating property of the at least one burner.

27 : The method of Claim 25 or 26, wherein the corrective action comprises disabling the at least one burner.

28: The method of any of the foregoing Claims, wherein the at {east one burner comprises a plurality of burners.

29: The method of any of the foregoing Claims, wherein the detector measures at least one characteristic of the plurality of burners as a whole.

30: The method of any of the foregoing Claims, wherein detecting an instability comprises detecting an instability of the plurality of burners as a whole.

31 : A system for detecting an instability associated with at least one burner, comprising:

a detector configured to obtain a signal measuring at least one characteristic associated with at least one burner;

a converter, coupled to the detector, comprising at feast one processor and configured to receive the signal from the detector and convert the signal into a time- varying signal spectrum; and

an instability detector, coupled to the converter, comprising the at least one processor and configured to detect an instability associated with the at least one burner based at least in part on the time-varying signal spectrum.

32: The system of Claim 31, configured for use in accordance with any of the methods of claims 1-27.

Description:
FLAME INSTABILITY DETECTOR

FIELD

[00011 The invention is generally related to flame instability detectors. Particularly, the present application relates to monitoring a flame state and identifying an instability using a single channel detector.

BACKGROUND

[0002] Furnace monitoring is becoming an increasingly important problem in refinery operations. Industrial furnaces, fired heaters, and boilers are used extensively across multiple refinery processes such as process heating and steam production, and are generally responsible for the largest proportion of the total refinery fuel consumption. The proper operation of these furnaces is particularly relevant for safety, environmental, and energy efficiency concerns.

[0003] In addition, industrial furnaces can contribute substantially to total refinery NOx emissions. NOx emissions can be reduced through lowering the adiabatic flame temperature while maintaining safe operation, which can be achieved by, e.g., enhancing fuel gas recirculation, steam injection, or use of technologies such as premixed flames and ultra-low NOx s. However, these technologies are often more prone to flame instability than tradition processes. It therefore is necessary to monitor the burner stability and provide feedback signals to control fuel and'or diluent flow when instabilities occur,

[00041 Traditionally, flame monitoring in industrial furnaces has been accomplished through visual inspection, analyzer-based monitoring, and phoiodetector devices. Visual inspection can readily identify flame blowoff, but is generally inadequate for identifying instability prior to blowoff. Analyzer-based monitoring typically has long latency and lacks the dynamic coverage needed for reliable detection. Photodetector devices such as flame eye are mainly burner based and expensive for wide-deployment Furthermore, the practical use of line-of-sight techniques, such as Tunable Diode Laser-based monitoring, can be restricted due to their design.

[00051 ew flame monitoring strategies have been introduced, but are limited in various ways. For example, variance-based approaches have been proposed, but are limited due to their low output signal-to-noise ratio, which requires an operator to choose between early detection and a low false positive rate, in addition, draft pressure fluctuation approaches have been reported in the past, but these techniques have been limited to a specific frequency range.

SUMMARY

[0006] The purpose and advantages of the present application will be set forth in and apparent from the description that follows, as well as will be learned by practice of the disclosed subject matter. Additional advantages of the disclosed subject matter will be realized and attained by the method and apparatus particularly pointed out in the written description and claims hereof, as well as from the appended drawings.

[0007] To achieve these and other advantages and in accordance with the purpose of the application, as embodied and broadly described, the disclosed subject matter includes a method for detecting an instability associated with at least one burner. The method can include the steps of obtaining a signal from a detector, the detector measuring at least one characteristic associated with at least one burner; converting, using at least one processor, the signal into a time-varying signal spectrum; and detecting, based at least in part on the time- varying signal spectrum, an instability associated with the at least one burner. The detector can be, for example, a single- channel detector.

[0008j For example, the characteristic can be a pressure metric, a fluctuation metric, or a vibration metric. The detector can be a dynamic pressure sensor, a device that captures video frames, or an accelerometer.

[0009] Tn accordance with one embodiment of the disclosed subject matter, converting the signal into a time-varying signal spectrum comprises using a time- frequency analysis, such as a short-time Fourier transform. The time -varying signal can be represented as a spectrogram.

[0010] In accordance with one representative embodiment of the disclosed subject matter, detecting an instability associated with the at least one burner can comprise computing a spectral entropy based at least in part on the time-varying signal spectrum. [0011] in accordance with another embodiment of the disclosed subject matter, an instability indicator can be used to detect an instability. The instability indicator can correspond to a probability of instability. Additionally or alternatively, the instability indicator can correspond to the temporal-spectral structure of the signal obtained from the detector. The time-varying signal spectrum can be converted into the instability indicator. For example, the time-varying signal spectrum can be normalized to obtain a probability ma ss function. The Shannon entropy of the probability mass function as a function of time can be computed. The inverse of the Shannon entropy of the probability m ss function can then be used as an instability indicator.

[0012] As disclosed herein, the instability indicator can be compared to a threshold value. An instability is detected when the instability indicator exceeds a threshold value.

[0013] In accordance with another embodiment, an alarm is provided when an instability is detected. Corrective action can be taken when the instability is taken. For example, the correcti ve action can include adjusting an operating property of the at least one burner or disabling the at least one burner.

[0014] Also disclosed herein is a system to detect an instability associated with at least one burner. The system can include a detector for obtaining a signal, the detector measuring a characteristic associated with at least one burner; a converter comprising at least one converting processor, the converter configured to convert the signal into a time- varying signal spectrum; and an instability detector comprising at least one instability detector processor, the instability detector configured to detect, based at least in part on the time-varying signal spectrum, an instability associated with the at least one burner. Additional aspects and features of the system are described in conjunction with the method.

BRIEF DESCRIPTION OF THE DRAWINGS

[0015] FIGURE 1 is a graph showing the draft pressure measured by two pressure detectors over time as a flame is driven from stable to unstable combustion and approaches blowoff. [0016] FIGURE 2 is a flow chart describing a representative embodiment of a method for detecting an instability associated with at least one burner in accordance with the disclosed subject matter.

[0017] FIGURE 3 is a graph showing the draft pressure measured by a pressure detector over time as a flame is driven from stable to unstable combustion and approaches blowoff.

[0018] FIGURE 4 is a series of video frames showing the flames of three burners over time. The video frame rate for the video frames in Figure 4 is around 6.4 frames per second which, with the oscillation cycle having 1 1 frames, leads to an approximately 1.72 second oscillation cycle, or equivalent!}/ 0.58 Hz peak frequency.

[0019] FIGURE 5 is a flow chart describing a representative embodiment of a method for converting a series of video frames into a scalar time series signal in accordance with the disclosed subject matter.

[0020] FIGURE 6 is a graph of a scalar time series calculated based on a series of video frames in accordance with the disclosed subject matter.

[0021] FIGURE 7 A is a spectrogram showing the frequency of a draft pressure signal over time as a flame is driven from stable to unstable combustion and approaches blowoff.

[0022] FIGURE 7B is a simplified version of Figure 7A and is presented for purposes of explanation only.

[0023] FIGURE 8 is a graph showing a room mean square plot of a vibration measurement on furnace piping over time as a flame is driven from stable to unstable combustion and approaches blowoff.

[0024] FIGURE 9 is a flow chart describing a representative embodiment of a method for determining an instability indicator in accordance with the disclosed subject matter.

[0025] FIGURE 10 is a graph illustrating a comparison between an instability indicator determined in accordance with the disclosed subject matter and an instability indicator determined based on a variance-based approach. [0026] FIGURE 1 1 is an illustration of a representative embodiment of a system for detecting an instability associated with at least one burner in accordance with the disclosed subject matter

DETAILED DESCRIPTIONS OF THE PREFERRED EMBODIMENTS

Overview

[0027] Generally, the disclosed subject matter is directed to a method of detecting an instability associated with at least one burner, the method comprising obtaining a signal from a detector, the detector measuring at least one characteristic associated with at least one burner, converting, using at least one processor, the signal into a time- varying signal spectrum, and detecting, based at least in part on the time-varying signal spectrum, an instability associated with the at least one burner. Additionally, a system is provided herein. The system generally includes a detector configured to obtain a signal measuring at least one characteristic associated with at least one burner, a converter, coupled to the detector, comprising at least one processor and configured to receive the signal from the detector and convert the signal into a time-varying signal spectrum, and an instability detector, coupled to the converter, comprising at least one processor and configured to detect an instability associated with the at least one burner based at least in part on the time -varying signal spectrum.

[00281 Reference will now be made in detail to representative embodiments of the disclosed subject matter, examples of which are illustrated in the accompanying drawings. The methods and systems disclosed herein will be described in conjunction with each other for clarity.

[0029] With reference to Figure 1, a time series of draft pressure measurements from two detectors is shown as a flame is driven from a stable condition or phase to an unstable condition or phase and approaches blowoff. The time series typically has three phases. The first phase is the stable phase. (See 102) Stable combustion generates more or less random broadband pressure fluctuations. In contrast, flame instability, as represented by the oscillating phase, is typically coherent, as manifested by harmonic pressure oscillations. (See 104) The final phase is blowoff. (See 106)

[0030] Accordingly, the disclosed subject matter generally relates to the detection of an instability associated with at least one burner based on the spectral structure in the measured signal , Although the disclosed systems and methods are generally discussed as utilizing pressure measurements, it will be understood by those having ordinary skill in the art that other measurements can also be used, as explained in greater detail herein. Moreover, while the disclosed systems and methods are generally discussed as detecting an instability associated with a single burner, those having ordinary skill in the art will understand that the disclosed subject matter can be used to detect an instability associated with more than one burner. For example, a single detector can measure a draft pressure associated with a two burner system and detect an instability associated with the two burner system as a whole.

Definitions

[0031] in the discussion herein, the phrase "detecting an instability" refers to identifying an anomaly from stable combustion. However, the phrase does not include determining whether a flame is present.

[0032] In the discussion herein, the term "time-varying signal spectrum" refers to the characteristics of the frequency over time. For example, in one embodiment ihai time-varying signal spectrum refers to the signal spectral density estimated continuously over time that characterizes both the spectral structure (e.g. flat over broadband or spiky with peak frequencies) and the time evolution of these spectral structures, including the time trajectory of the peak frequency positions and the magnitude of the associated spectral components. The time-varying signal spectrum can be represented, for example, as a spectrogram.

[0033] in the discussion herein, the term "single channel instability detector" refers to an apparatus that detects an instability using only a single stream of data (e.g., an apparatus including a single pressure sensor ihai provides updated pressure measurements based on the sampling rate of the sensor).

[0034] In the discussion herein, the term "spectral entropy" refers to entropy calculated based on the time-varying signal spectrum. A wide variety of methods for calculating entropy can be used as known in the art. For example, the spectral entropy can be the Shannon entropy of the time-varying signal spectrum. [0035] As used herein, the term "coupled" means operatively in communication with each other, either directly or indirectly, using any suitable techniques, including through hard wire, connectors, and remote communicators.

Detectors

[0036] With reference to Figure 2, an exemplary method of def ecting an instability is shown. First, a signal is obtained from a detector that measures a characteristic associated with the burner. (See 202). The detector can be disposed in any manner that allows it to measure the characteristic of interest. For example, the detector may be disposed within the furnace (e.g., in the case of a pressure sensor) or outside of the furnace (e.g., in the case of a video camera for recording the flame).

[0037] The instability detector can be a single -channel instability detector such that only one detector (e.g., a pressure sensor) is needed. . Two or more detectors can be used in case one of the detectors malfunctions, but the identification of an instability will be based on input from only a single channel of data. In another embodiment, multiple detectors can be used and the measured signals can be combined. While this may improve, for example, the signal-to-noise ratio, the use of a single detector can improve deployment feasibility as compared with the multi-channel format.

[0038] Many characteristics of the burner can be measured by the detector without departing from the scope of the disclosed subject matter. For example, the characteristic can be a pressure metric such as the draft pressure. The associated detector can be a dynamic pressure sensor, such as a pressure probe, that can capture a high frequency signal. Figure 3 illustrates an exemplary draft, pressure measurement as a function of time as the flame approaches blowoff.

[003 1 In another embodiment, the characteristic can be a fluctuation metric. The associated detector can be a device, such as a video camera, that captures video frames. The detector captures a series of video frames. For example, with reference to Figure 4, the detector captures video frames 402.-424. Each of the video frames in Figure 4 shows the flames associated with three burners at various times.

[0040] To convert the series of video frames into a time-varying signal spec trum, the series of video frames are converted into a scalar time series signal, e.g., each video frame is converted into a single value that can be plotted against time. In one embodiment, a video frame can be converted into a single value based on the intensity associated with each pixel. For example, and with reference to Figure 5, the intensity for each pixel in a first video frame U 0) can be measured (See 502). The intensity for each pixel in a second video frame (t=l) can also be measured (See 504). The first video frame immediately precedes the second video frame (i.e., the first and second video frames are consecutive samples). For each pixel, a change index between the intensity in the first video frame and the intensity at the second video frame is calculated (See 508). The change index for each pixel is then aggregated to calculate a fluctuation index for the second video frame (See 510). The magnitude of the fluctuation index can then be plotted at time t=l as part of the scalar time series signal. This process can be repeated for each subsequent frame, or selected subsequent frames as desired. With reference to Figure 6, a scalar time series signal obtained from a series of video frames is shown.

[0041] Additionally or alternatively, the characteristic can be a vibration metric, such as the oscillation of the furnace piping. The associated detector can be an acceierometer, with measurements processed accordingly.

[0042] Other characteristics and detectors can also be used. For example, optical sensors can be used to measure flicker. Analytic measurements, such as measurements of carbon dioxide and sulfur dioxide levels in a furnace, can also be used.

Time- Varying Si gnal Spectrum

[0043 ] The signal obtained from the detector describes the measured characteristic as a function of time. With further reference to Figure 2, the signal is converted into a time-varying signal spectrum using at least one processor (See 204). The signal can be converted to a time-varying signal spectrum using a time-frequency analysis. The time- frequency analysis can be, for example, a short-time Fourier transform,

[0044] Reference will now be made to a representative method of converting the signal into the time- varying signal spectrum. Given a measured signal x(t), its time- frequency distribution y(t,f) can be computed using the following general form: y(t,f) = F„ f [G(t, T) * T K x (t, T)] (1) where F T - > f denotes the Fourier transform from delay to frequency, K x (t, x) =

x(t+0.5x)x(t-0.5x), and G(t, τ) is a kernel function. For example, in the case of a short- time Fourier transform. Git, τ) = h(t+Q.5x)h(t-0.5x) for some window function. Typically the window function h(i) is a locally supported function with finite squared integral, such as a raised cosine window or Hamming window, so that it effectively computes the spectral density of the signal inside a shaped sliding window ,

[0045] More generally, the signal can he converted into a time-varying signal spectrum by any method that determines the frequency spectrum of the signal over time.

[0046] The time -varying signal spectrum can be represented as a spectrogram. For example, a spectrogram obtained in accordance with the disclosed subject matter is shown in Figure 7 A. A rough approximation of the spectrogram is shown in Figure 7B highlighting the time trajectory of the peak spectral component. The representation in Figure 7B is shown for purposes of explanation only. As illustrated in Figure 7B, the spectrogram shows that the signal does not have a steady frequency until the start of a harmonic oscillation at 702. The spectrogram further shows that the signal maintains a fairly steady frequency until a deterioration point 704. As used herein, the term

"deterioration point" refers to an area where the signal rapidly decreases from a prolonged steady frequency , A lower frequency generally corresponds to a higher amplitude of the signal. For example, with further reference to Figure 3, it can be seen that the amplitude of the draft pressure signal increases as blowoff approaches. Finally, the spectrogram shows that the signal does not have a frequency after blowoff at 706, A lthough the disclosed subject matter is not limited to any particular theory of operation, the time variations of the instability signal spectrum, including decreasing peak frequency and higher spectral magnitude, can be explained by an increasing flame lift-off distance which causes a longer re-attach time, and as a result, more intense pressure fluctuations.

[0047] in another embodiment of the disclosed subject matter, the time-varying signal spectrum can be represented as the root mean square vibration measurement on furnace piping measured by an accelerometer. In one embodiment, the signal measured by the accelerometer can be filtered (e.g., using a low pass filter) before the root mean square representation is calculated. Figure 8 illustrates an exemplary root mean square (RMS) plot of the vibration measurement on furnace piping as the flame approaches blowoff. Instability Detector

With reference now to Figure 2, the disclosed subject matter further includes detecting, based at least in part on the time-varying signal spectrum, an instability associated w th the burner (See 206). An instability indicator can be used to detect the instability. The instability indicator can correspond to a probability that the flame is unstable based at least in part on the time- varying signal spectrum.

[0049] With reference to Figure 9, an exemplary method of detecting an instability is described. The time- varying signal spectrum is normalized by the sum of the magnitude of all the spectral components (See 902):

*<■/,) =■ · ' m where y(t,fk) is the estimated signal spectrum at frequency fk and is non-negative. Variations of this normalization in terms of spectral sub-bands instead of the whole frequency band can be used as known in the art.

[00501 With further reference to Figure 9, a probability mass function in frequency can be derived based on the normalized time -varying signal spectrum (See 904):

By converting the time-varying signal spectrum at each time into a probability mass function, entropy can be utilized in capturing the information associated with a probability mass function (PMF). Namely, larger entropy is associated with uncertain distribution and therefore a flatter PMF, because in the absence of any information the distribution will be presumed random. Similarly, smaller entropy is associated with peaky PMF. For a signal associated with stable combustion, its spectrum is typically broadband and relatively flat, and the corresponding PMF is close to a uniform distribution which leads to large entropy. In the presence of instability, steady peak frequencies emerge and signal energy starts to concentrate around those peak frequency points, leading to a "peaky" PMF, or more certain distribution and therefore a smaller value of entropy. [0051] With further reference to Figure 9, the spectral entropy of the signal can be computed (See 906). The spectral entropy can be calculated as the Shannon entropy of the probability mass function as a function of time:

K-\

sit) = - Y P, (i) In P (i) (4)

[0052] As discussed above, stable combustion generally generates random noises that are relatively flat over a broad range of time, which results in a higher entropy value, in contrast, unstable combustion leads to oscillations with clear spectral peaks, resulting in lower spectral entropy. Therefore, the inverse of the spectral entropy can be used as an instability indicator. (See 908) An instability indicator in accordance with the disclosed subject matter is shown in Figure 10 (See 1002).

[0053 ] The use of an instability detector in accordance with the disclosed subject matter can p rovide an improved signal-to-noise ratio and sensitivity of the instability detector. For example, Figure 10 is a comparison between an instability indicator in accordance with the disclosed subject matter (utilizing the spectral entropy approach) and an instability indicator obtained using a variance-based approach. The two

instability indicators have been normalized such that they are at the same level in the absence of an instability. The instability indicator of invention (labeled "spectral entropy" in Figure 10) shows a significantly higher level of sensitivity.

[00541 The instability detector can use a threshold to identify an instability. The threshold can be mathematically derived or based on experimental observations. The identification of the threshold can vary based on several variables, including the detector utilized to obtain the signal, the desired target detection probability, the false positive rate, and the detection delay. For example, if it is desired to detect any instability as soon as possible, the threshold value will be lower and the false positi ve rate will increase. However, if it is desired to minimize the false positive rate (e.g., because incorrect detection of an instability is economically disadvantageous), the threshold can be raised and the detection delay will increase. An improved output signal vs. noise ratio (SN ) obtained by a spectral entropy-based indicator can significantly improve detection performance in the sense that given a fixed false positive rate, it can achieve higher detection probability or a shorter detection delay than a detector with lower SNR such as the approach based on signal variance only.

[0055] As previously noted and with further reference to Figure 7B, the furnace may operate in a mild unstable state for an extended period of time at a steady frequency (as between 702 and 704). Due to the high costs (both economic and environmental) associated with stopping and restarting the furnace, it may be advantageous to keep the burner running during this time period. Therefore, in one embodiment, the threshold can be set so as to detect an instability at the deterioration point 704. Such threshold can be set, for example, based on experimental observation.

[00561 An alarm can be provided when an instability is detected. The alarm can be, for example, an audio alarm such as a siren. The alarm can also be, for example, a visual alarm such as a flashing light or an indication on the monitor of a computer screen. More generally, any method of informing an operator that an instability has been detected can be used as known in the art for its intended purpose.

[0057J Corrective action can be taken when an instability is detected. For example, an operating property of the burner can be adjusted. For example, the amount of steam injected into the furnace can be decreased until the instability is resolved. Similarly, the burner can be disabled, which can prevent an explosion and allow repairs and/or maintenance to be provided to the furnace. In the case of a severe instability with an unknown root cause, the at least one burner can be shut down.

Instability Detection System

[0058] For purpose of explanation and illustration, and not limitation, an exemplary embodiment of the system for detecting an instability associated with a burner in accordance with the application is shown in Figure 1 1. The instability detection system 1 100 can include a detector 1 102, a converter 1 104, and an instability detector 1 106.

[0059] The detector 1 102 is disposed within or near a furnace i 108 having at least one burner 1 1 10, The detector 1 102 can be disposed in any manner that allows it to measure the characteristic of interest. For example, the detector may be disposed within the furnace 1 108 (e.g., in the case of a pressure sensor) or outside of the furnace 1 108 (e.g., in the case of a video camera for recording the flame). [0060] The converter 1104 is coupled to the detector 1 102 and is configured to receive the signal from the detector 1 102 and convert the signal into a time-varying signal spectrum. As discussed above, the converter can implement this functionality in a number of ways including, for example, by using a short- time Fourier transform.

[0061 ] The instability detector 1106 is coupled to the converter and is configured to receive the time-varying signal spectrum from the converter 1 104 and detect an instability based on the time-varying signal spectrum. The instability detector 1 108 can include an instability indicator generator 1 1 12 that is configured to determine an instability indicator in accordance with the disclosed subject matter. Additional functional units can be used to perform other functions of the method as disclosed herein.

[0062] The converter 1 104, the instability detector 1 106, the instability indicator generator 1 1 12, and other functional units of the instability detection system 1 100 can be implemented in a variety of ways as known in the art. For example, each of the functional units can be implemented using an integrated single processor. Alternatively, the each functional unit can be implemented on a separate processor. Therefore, the instability detection system 1 100 can be implemented using at least one processor and/or one or more processors.

[0063] The at least one processor comprises one or more circuits. The one or more circuits can be designed so as to implement the disclosed subject matter using hardware only. Alternatively, the processor can be designed to can out the instructions specified by computer code stored in a hard drive, a removable storage medium, or any other storage media. Such non-transitory computer readable media can store instructions th t, upon execution, cause the at least one processor to perform the methods in accordance with the disclosed subject matter.

[00641 The system can further include at least one burner 1 1 10. The at least one burner 1 1 10 can be a part of a furnace 1 108. The term "furnace," as used herein, refers to a wide variety of equipment that includes at least one burner, including, for example, industrial furnaces, fired heaters, and boilers. The furnace 1 108 can be located at a refinery or similar location. The at feast one burner 1 1 10 or another functional element of the furnace 1 108 (e.g., a steam injector) can be coupled to the instability detector 1 106 and a corrective action processor in order to automatically institute a corrective action when an instability is detected. The corrective action processor can include one or more processors comprising one or more circuits as discussed above.

[0065] The instability detection system 1 100 can further include additional components in accordance with the disclosed subject matter. For example, the system can include an alarm coupled to the instability detector that is activated when an instability is detected. The alarm can be, for example, a siren, a flashing light, an alarm on a computer console (preferably a manned distributed control console), or any other alarm.

Additional Embodiments

[0066] Additionally or alternately, the invention can include one or more of the following embodiments

[00671 Embodiment 1 . A method for detecting an instability associated with at least one burner based on the spectral entropy of a signal measuring a characteristic of the at least one burner.

[0068] Embodiment 2: A method for detecting an instability associated with at least one burner, comprising: obtaining a signal from a detector, the detector measuring at least one characteristic associated with the at least one burner; converting, using at least one processor, the signal into a time-varying signal spectrum; and detecting, based at least in part on the time-varying signal spectrum, an instability associated with the at least one burner.

[0069] Embodiment 3: The method of Embodiment 1 or Embodiment 2, wherein the instability is detected based on a single channel of data.

[0070] Embodiment 4: The method of Embodiment 1, 2, or 3, wherein the characteristic comprises a pressure metric.

[0071] Embodiment 5: The method of any of the foregoing embodiments wherein the detector comprises a dynamic pressure sensor.

[0072] Embodiment 6: The method of any of the foregoing embodiments, wherein the characteristic comprises a fluctuation metric. [0073] Embodiment 7: The method of any of the foregoing embodiments, wherein the detector comprises a device that captures video frames.

[0074] Embodiment 8: The method of any of the foregoing embodiments, further comprising converting a video frame into a single scalar to produce a scalar times series signal

[00751 Embodiment 9: The method of any of the foregoing embodiments, wherein the characteristic comprises a vibration metric.

[00761 Embodiment 10: The method of any of the foregoing embodiments, wherein the detector comprises an accelerometer.

[0077] Embodiment 1 1 : The method of any of the foregoing Embodiments, wherein the time- varying signal spectrum comprises a spectrogram.

[0078] Embodiment 12: The method of any of the foregoing Embodiments, wherein converting the signal comprises using time-frequency analysis.

[0079] Embodiment 13: The method of Embodiment 12, wherein the time- frequency analysis comprises a short-time Fourier transform.

[0080] Embodiment 14: The method of any of the foregoing Embodiments, wherein detecting an instability associated with the at least one burner comprises computing a spectral entropy based on the time -varying signal spectrum.

[0081] Embodiment 15: The method of any of the foregoing Embodiments, wherein detecting an instability associated with the at least one burner comprises determining an instability indicator.

[00821 Embodiment 16: The method of Embodiment 15, wherein the instability def ector corresponds to a probability of instability.

[00831 Embodiment 17: The method of Embodiments 15 or 16, wherein determining an instability indicator comprises converting the time-varying signal spectrum into the instability indicator.

[0084] Embodiment 18: The method of Embodiment 17, wherein converting the time-varying signal spectrum comprises normalizing the time-varying signal spectrum. [0085] Embodiment 19: The method of Embodiments 7 or 18, wherem converting the time-varying signal spectrum comprises calculating a probability mass function based on the time- varying signal spectrum.

[00861 Embodiment 20: The method of Embodiments 17, 1 8, or 19, wherein converting the time-varying signal spectrum comprises computing a Shannon entropy of a probability mass function as a function of time.

[0087] Embodiment 21 : The method of Embodiments 15, 16, 17, 18, 19, or 20, wherein the instability indicator comprises an inverse of a spectral entropy.

[0088] Embodiment 22 : The method of any of the foregoing embodiments, wherein detecting an instability comprises comparing the instability indicator to a threshold value.

[0089] Embodiment 23: The method of Embodiment 22 , wherein the instability is detected when the instability indicator exceeds a threshold value.

[0090] Embodiment 24: The method of any of the foregoing Embodiments, further comprising providing an alarm when the instability is detected.

[00911 Embodiment 25: The method of any of the foregoing Embodiments, further comprising taking corrective action when the instability is detected.

[0092] Embodiment 26: The method of Embodiment 25, wherem the corrective action comprises adjusting an operating property of the at least one burner.

[0093] Embodiment 27: The method of Embodiment 25 or 26, wherein the corrective action comprises disabling the at least one burner.

[00941 Embodiment 28: The method of any of the foregoing Embodiments, wherein the at least one burner comprises a plurality of burners.

[00951 Embodiment 29: The method of any of the foregoing Embodiments, wherein the detector measures at least one characteristic of the plurality of burners as a whole.

[0096] Embodiment 30: The method of any of the foregoing Embodiments, wherein detecting an instability comprises detecting an instability of the plurality of burners as a whole. [0097] Embodiment 31 : A system for detecting an instability associated with at least one burner, comprising a detector configured to obtain a signal measuring at least one characteristic associated with at least one burner, a converter, coupled to the detector, comprising at least one processor and configured to receive the signal from the detector and convert the signal into a time-varying signal spectrum, and an instability detector, coupled to the converter, comprising the at least one processor and configured to detect an instability associated with the at least one burner based at least in pari on the time-varying signal spectrum.

[0098] Embodiment 32: The system of Embodiment 31, configured for use in accordance with any of the methods above (i.e., the methods of Embodiments 1 through 27).

[0099] While the present application is described herein in terms of certain preferred embodiments, those skilled in the art will recognize that various modifications and improvements may be made to the application without departing from the scope thereof. Thus, it is intended that the present application include modifications and variations that are within the scope of the appended claims and their equi val ents.

Moreover, although individual features of one embodiment of the application may be discussed herein or shown in the drawings of one embodiment and not in other embodiments, it should be apparent that individual features of one embodiment may be combined with one or more features of another embodiment or features from a plurality of embodiments.

[00100] in addition to the specific embodiments claimed below, the application is also directed to other embodiments having any other possible combination of the dependent features claims below and those disclosed above. As such, the particular features presented in the dependent claims and disclosed above can be combined with each other in other manners within the scope of the application such that the application should be recognized as also specifically directed to other embodiments having any other possible combinations. Thus, the foregoing description of specific embodiments of the application has been presented for purposes of illustration and description. It is not intended to be exhaustive or to limit the application to those embodiments disclosed.