Synchrosqueezing transform-based oscillating combustion fault detection method

Abstract

The present invention relates to the technical field of signal processing and fault diagnosis. The method of the present invention comprises: collecting a flame chemiluminescence signal, and performing adaptive filtering noise reduction processing; using an improved proper orthogonal decomposition method and a high-order mode to obtain comprehensive combustion feature information; developing a real-time synchrosqueezing transform algorithm to perform time-frequency analysis on data, and updating energy distribution in real time; taking into account sensor data to perform multi-parameter joint analysis to obtain a combustion state evaluation, using a plurality of high-speed cameras to acquire flame images from different angles, reconstructing three-dimensional flame morphology by means of a computer vision technology, and providing fault detection information; and constructing a fault prediction model to give an early warning of a combustion fault. The present method implements nondestructive and rapid detection of oscillating combustion faults, keeps rich information of flame images, reflects overall flame pulsation characteristics, increases the speed and accuracy of detection, and compensates for the defect of low frequency resolution of conventional time-frequency analysis methods.

Description

Oscillatory combustion fault detection method based on synchronous compressive transform TECHNICAL FIELD

[0001] The present application relates to the technical field of signal processing and fault diagnosis, and particularly relates to an oscillatory combustion fault detection method based on synchronous compressive transform. BACKGROUND

[0002] Gas turbines are widely used in power generation, ship propulsion, and mechanical power generation due to their high efficiency as energy conversion devices. However, the operational stability of the combustion chamber has a decisive influence on the performance and safety of the entire system. Gas turbine combustion chambers usually adopt lean premixed swirling combustion, which can reduce combustion temperature and NOx emissions, but is prone to induce oscillatory combustion faults. This fault can seriously affect the operational efficiency and combustion efficiency of the gas turbine and its combined cycle, and may cause fluctuations in important parameters such as combustion chamber temperature and NOx concentration, posing a threat to equipment and personnel safety. Gas turbines may experience combustion instability during operation, such as oscillatory combustion, which can lead to a decrease in equipment performance and even cause equipment damage. Therefore, it is particularly important to quickly detect and diagnose oscillatory combustion faults in gas turbines.

[0003] Traditional oscillatory combustion fault monitoring methods mainly rely on pressure sensors to analyze time-domain pressure signals in the frequency domain. However, this method has some obvious shortcomings. First, since the sensor is usually installed far from the flame, there is an error between the actual combustion chamber pressure and the sensor measured pressure, resulting in data monitoring deviation and lag. Second, due to the limitation of sensor response speed, fault phenomena are not identified in time, and the fault is often discovered after serious losses occur, which is extremely disadvantageous for actual production.

[0004] Existing gas turbine oscillatory combustion fault detection techniques usually use acoustic detection, pressure fluctuation analysis, and other methods. These methods can detect combustion oscillation to some extent, but have the following shortcomings:

[0005] Low sensitivity: Acoustic detection and pressure fluctuation analysis are affected by environmental noise and mechanical vibration, and the detection effect of weak combustion oscillation signals is not good, which can easily lead to false positives or false negatives.

[0006] Difficult to locate: Existing detection techniques cannot accurately locate the specific position of combustion oscillation, and cannot provide accurate information for fault handling.

[0007] Poor timeliness: Traditional detection methods require a long time for signal acquisition and processing, and cannot achieve real-time monitoring and rapid diagnosis. SUMMARY

[0008] In view of the above existing problems, the present application is proposed.

[0009] Therefore, the present application provides an oscillation combustion fault detection method based on synchronous compression transformation, which can reduce the time cost of fault detection, improve the accuracy of fault detection, and reduce the false alarm rate.

[0010] To solve the above technical problems, the present application provides the following technical scheme, an oscillation combustion fault detection method based on synchronous compression transformation, comprising: collecting flame chemiluminescence signals, and performing adaptive filtering and noise reduction processing on the collected flame chemiluminescence signals; using an improved proper orthogonal decomposition method to extract 1st order modal time coefficients from the flame image, using high order modes to obtain comprehensive combustion feature information; developing a real-time synchronous compression transformation algorithm to perform time-frequency analysis on the data, and updating the energy distribution in real time to respond to small changes in the combustion state; combining sensor data to perform multi-parameter joint analysis to obtain combustion state evaluation, using multiple high-speed cameras to obtain flame images from different angles, reconstructing the three-dimensional flame morphology through computer vision technology, and providing fault detection information; constructing a fault prediction model to automatically complete the entire fault detection process, and realizing early warning of combustion faults.

[0011] As a preferred scheme of the oscillation combustion fault detection method based on synchronous compression transformation, the adaptive filtering and noise reduction processing comprises distinguishing noise and actual combustion oscillation signals during oscillation combustion fault detection, selecting an adaptive filter based on the steepest descent method, and combining noise estimation and signal enhancement in the filter.

[0012] The adaptive filter update rule is as follows:

[0013] Wherein, w(n) represents the filter weight vector, n represents the discrete time step, represents the noise power estimation at the nth moment, represents a constant, x(n) represents the input flame signal at the nth moment, e(n) represents the error signal at the nth moment, and represents the forgetting factor.

[0014] As a preferred scheme of the oscillation combustion fault detection method based on synchronous compression transformation, the improved proper orthogonal decomposition method comprises: performing POD analysis on the flame image set processed by the filter to obtain eigenmodes and eigenvalues, selecting the M eigenmodes with the highest energy according to the size of the eigenvalues, forming a new mode set, and representing the most significant spatial features in the data.

[0015] The time coefficient is calculated for each selected mode, and the time coefficient represents the projection of the flame image on each eigenmode over time, and the calculation formula is as follows:

[0016] where c i (m) represents the time coefficient, I(m) represents the mth flame image frame, represents the mean value of the image set, φ i (x,y) represents the intrinsic mode of the flame image, i is the index of the mode;

[0017] The calculated time coefficient is stored in the set:

[0018] where, represents the set containing the high-order mode time coefficient, the set reveals the fine fluctuations in the combustion process and the early signs of failure.

[0019] As a preferred scheme of the oscillation combustion fault detection method based on synchronous compression transformation, the development of the real-time synchronous compression transformation algorithm includes: rearranging the coefficients of the time-frequency transformation through the synchronous compression operator, moving the time-frequency coefficient of the signal at any point in the time-frequency plane to the center of gravity of the energy, enhancing the energy concentration degree of the instantaneous frequency, applying a window function on the basis of the Fourier transform, dividing the signal time domain, obtaining the frequency distribution under the window at different times through the window function sliding, and arranging these short-time spectra in time sequence to describe the time-varying law of the frequency components of the signal.

[0020] For a certain time-varying signal The window function of the short-time Fourier transform is g(t), and after the STFT transformation, there is:

[0021] G(t,f)=∫ R s(τ)g(τ-t)e-i2πf(τ-t)dτ

[0022] Where G(t,f) represents the result of the Fourier transform, s(τ) represents the original signal, τ represents the time variable of the original signal, g(τ-t) represents, t represents the time variable, f represents the frequency variable, and i represents the imaginary unit.

[0023] For the time-varying harmonic signal and the window function signal, the Fourier transform is performed to obtain the estimation result of the instantaneous frequency in the STFT time-frequency representation, which is represented as:

[0024] Where θ(t,ω) represents the instantaneous frequency, ω represents the frequency variable, arg(G(t,f)) represents the argument, and Re represents the real part, represents the partial derivative with respect to time t.

[0025] The STFT coefficients with the same frequency are collected by using the above instantaneous frequency estimation result, and according to the Dirac function property, a synchronous compression operator function is given as ∫ R δ(η-ω0(t,ω)), and a synchronous compression transform is obtained, which is expressed as:

[0026] T s (t,η)=∫ R G(t,f)δ(η-θ(t,ω))dω

[0027] Wherein, T s (t,η) represents an instantaneous frequency signal, δ(η-θ(t,ω)) represents a Dirac function, and η represents an instantaneous frequency variable. The time-frequency coefficients are rearranged in the frequency direction, and the synchronous compression transform process at time t is the frequency point rearrangement and throughout all time points.

[0028] As a preferred scheme of the oscillation combustion fault detection method based on the synchronous compression transform, wherein: the multi-parameter joint analysis includes analyzing the data collected by the sensor to understand the distribution and characteristics, adjusting the parameters through an optimization algorithm, and finding the best parameter combination through multiple iterations;

[0029] The parameter combination is substituted into the multi-parameter joint analysis formula to calculate the weighted contribution of each sensor, the weight considers the different sensitivities of the sensor to the angle, speed and light intensity, and a normalization function is applied to ensure the comparability between different sensor data, and the specific formula is:

[0030] Wherein, G(θ,v,I) represents a multi-parameter joint analysis main function, θ represents a flame tilt angle, v represents a flame combustion speed, I represents a flame radiation intensity, n represents a number of sensors, Ω i represents an effective measurement domain of the i th sensor, f i (θ i ,v i ,I i ) represents a weighted contribution, α i represents an angle parameter of the i th sensor, θ i represents an angle value measured by the i th sensor, θ 0i represents a reference angle value of the i th sensor, β i represents an adjustment coefficient of a speed parameter of the i th sensor, γ i represents an adjustment coefficient of a light intensity parameter of the i th sensor, respectively represent a mean value and a standard deviation of the angle measurement value of the i th sensor, respectively represent a mean value and a standard deviation of the speed measurement value of the i th sensor.

[0031] As a preferred scheme of the oscillation combustion fault detection method based on synchronous compressive transform, the reconstructed three-dimensional flame shape comprises reconstructing the three-dimensional shape of the flame by using a computer vision algorithm in combination with the results of multi-parameter joint analysis.

[0032] A dynamic three-dimensional flame model is obtained by considering the change of the flame over time through a time correlation function:

[0033] Fx,y,z,t)=∫ Ω Gθ,v,I)·htdt

[0034] h(t)=sin(ωt+φ)·e -λt

[0035] wherein x, y, z represent the coordinate axes of the three-dimensional space, ht represents the change function of the flame over time, Ω represents the spatial domain of integration, φ represents the phase offset, and λ represents the attenuation coefficient; the three-dimensional flame model gives the complete shape of the flame in the three-dimensional space, including the change over time.

[0036] As a preferred scheme of the oscillation combustion fault detection method based on synchronous compressive transform, the fault prediction model comprises comparing the calculated flame state and flame shape with the normal behavior threshold, and if the calculation result exceeds the preset threshold range, it is considered that there is abnormal behavior.

[0037] When the flame temperature is lower than 1000K or higher than 2500K, it is determined that the abnormal behavior is insufficient combustion or clogging of the burner, and the solution measure is to adjust the oxygen / fuel ratio of the burner, check and clean the internal carbon deposition and deposits of the burner;

[0038] When the flame oscillation frequency is lower than 0.5Hz or higher than 2Hz, and the flame shape is lower than 0.5 or higher than 1.5, it is determined that the abnormal behavior is damage to the internal structure of the burner or unstable fuel supply, and the solution measure is to check the fuel supply system and keep the fuel flow stable; check whether the nozzle and combustion chamber of the burner are clogged, and check whether the internal structure is damaged;

[0039] When the contrast of the flame image is lower than 0.6 or higher than 1.5, it is determined that the abnormal behavior is caused by improper camera settings or the characteristics of the flame itself, and the solution measure is to adjust the parameters of the camera to improve the image quality, check whether the lens of the camera is contaminated or damaged; check the brightness and color characteristics of the flame;

[0040] The detected abnormal behaviors and causes are summarized as fault detection information and recorded in a database for subsequent analysis and reporting, and a fault detection report is generated regularly to monitor the health status and performance of the combustion process.

[0041] Another object of the present application is to provide an oscillating combustion fault detection system based on synchronous compressive transform, which can improve the fault detection accuracy and diagnosis speed through the oscillating combustion fault detection algorithm.

[0042] As a preferred scheme of the oscillating combustion fault detection system based on synchronous compressive transform, the system comprises a data acquisition module, an intrinsic orthogonal decomposition method module, a synchronous compressive transform module, a multi-parameter joint analysis module, a three-dimensional flame shape reconstruction module, and a fault detection module.

[0043] The data acquisition module adopts a high-sensitivity flame chemiluminescence sensor to accurately capture the light signals generated by chemical reactions in the flame.

[0044] The intrinsic orthogonal decomposition method module accurately predicts faults by analyzing and identifying change patterns in the combustion process.

[0045] The synchronous compressive transform module uses real-time SCT algorithms to perform time-frequency analysis on sensor data, captures instantaneous changes in the combustion process, reflects subtle changes in the combustion state, and improves the sensitivity of fault detection.

[0046] The multi-parameter joint analysis module uses multi-parameter joint analysis algorithms to evaluate the flame state based on the results of synchronous compressive transform.

[0047] The three-dimensional flame shape reconstruction module uses computer vision technology to reconstruct the three-dimensional flame shape from flame images obtained from different angles.

[0048] The fault detection module analyzes the results of multi-parameter joint analysis and three-dimensional flame shape reconstruction to detect abnormal behavior of the flame.

[0049] A computer device comprising a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of an oscillating combustion fault detection method based on synchronous compressive transform when executing the computer program.

[0050] A computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the steps of an oscillating combustion fault detection method based on synchronous compressive transform.

[0051] The beneficial effects of the present application: in the aspect of data acquisition, the use of high-speed photography technology to collect flame images can avoid the data lag problem caused by the traditional pressure monitoring method, and can realize non-destructive and rapid detection of oscillating combustion failure; intrinsic orthogonal decomposition can reduce the dimension of flame image data and extract features, retain the rich information of flame image, reflect the overall pulsation characteristics of flame, and improve the detection speed and accuracy; synchronous compression transformation can realize energy rearrangement and instantaneous concentration through synchronous compression operator, make up for the low frequency resolution of traditional time-frequency analysis method, and realize accurate diagnosis and rapid detection of oscillating combustion failure. BRIEF DESCRIPTION OF DRAWINGS

[0052] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed in the embodiment description will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0053] Fig. 1 is a flowchart of an oscillating combustion failure detection method based on synchronous compression transformation provided by an embodiment of the present application.

[0054] Fig. 2 is a detection result of an oscillating combustion failure detection method based on synchronous compression transformation provided by an embodiment of the present application.

[0055] Fig. 3 is a pressure signal SST result graph of an oscillating combustion failure detection method based on synchronous compression transformation provided by an embodiment of the present application under two working conditions of flame oscillation and stability.

[0056] Fig. 4 is a working module schematic diagram of an oscillating combustion failure detection system based on synchronous compression transformation provided by an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the above-mentioned purposes, features and advantages of the present application more apparent and easy to understand, the specific embodiments of the present application will be described in detail below with reference to the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor should be within the scope of protection of the present application.

[0058] In the following description, many specific details are set forth in order to provide a thorough understanding of the present application, but the present application can also be implemented in other ways different from those described herein, and those skilled in the art can make similar generalizations without departing from the connotation of the present application, therefore the present application is not limited to the specific embodiments disclosed below.

[0059] Secondly, the "one embodiment" or "embodiment" referred to herein is intended to represent a specific feature, structure, characteristic, or combination of features and characteristics described herein that can be included in at least one implementation of the present application. The appearance of the phrase "in one embodiment" in various places in the specification is not intended to be construed as an indication that each of the features, structures, or characteristics, so described is required in all implementations or that actual implementations only include features, structures, or characteristics so described.

[0060] The present application is described in detail below in conjunction with the schematic drawings, and in the detailed description of the embodiments of the present application, the sectional view of the device structure is partially enlarged without the general proportion for the convenience of description, and the schematic drawings are only examples which should not limit the scope of protection of the present application. In addition, the three-dimensional spatial dimensions of length, width and depth should be included in actual production.

[0061] Meanwhile, in the description of the present application, it should be noted that the orientation or positional relationship indicated by the terms "upper, lower, inner and outer" is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, and does not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application. In addition, the terms "first, second or third" are only for descriptive purposes and cannot be understood as indicating or implying relative importance.

[0062] In the present application, unless otherwise explicitly specified and limited, the terms "mounting, connecting, connection" should be understood broadly, for example: it can be fixed connection, detachable connection or integral connection; it can also be mechanical connection, electrical connection or direct connection, it can also be indirectly connected through an intermediate medium, or it can be the internal communication of two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0063] Embodiment 1

[0064] Referring to FIG. 1, the first embodiment of the present application provides an oscillating combustion fault detection method based on synchronous compression transformation, which comprises:

[0065] S1: Collecting flame chemiluminescence signals, and performing adaptive filtering and noise reduction processing on the collected flame chemiluminescence signals.

[0066] Further, the adaptive filtering and noise reduction processing comprises distinguishing noise and actual combustion oscillation signals when performing oscillating combustion fault detection, selecting an adaptive filter based on the steepest descent method, and the filter combines noise estimation and signal enhancement.

[0067] The adaptive filter update rule is as follows:

[0068] Wherein, w(n) represents the filter weight vector, and n represents the discrete time step. where n represents the noise power estimation at the nth moment, ε represents a constant, x(n) represents the input flame signal at the nth moment, e(n) represents the error signal at the nth moment, and α represents a forgetting factor.

[0069] S2: A first-order modal time coefficient is extracted from the flame image by using an improved proper orthogonal decomposition method, and a comprehensive combustion feature information is obtained by using a high-order modal.

[0070] Further, the improved proper orthogonal decomposition method comprises: performing POD analysis on the flame image set processed by the filter to obtain eigenmodes and eigenvalues; and selecting M eigenmodes with the highest energy according to the size of the eigenvalues to form a new modal set representing the most significant spatial features in the data.

[0071] A time coefficient is calculated for each selected modal, and the time coefficient represents the projection of the flame image on each eigenmode over time, and the calculation formula is as follows:

[0072] where c i (m) represents the time coefficient, I(m) represents the mth flame image frame, represents the mean value of the image set, φ i (x, y) represents the eigenmode of the flame image, and i is the index of the modal.

[0073] The calculated time coefficient is stored in the set:

[0074] wherein, represents a set containing high-order modal time coefficients, and the set reveals subtle fluctuations and early signs of failure in the combustion process.

[0075] S3: A real-time synchronous compression transform algorithm is developed to perform time-frequency analysis on the data and update the energy distribution in real time to respond to small changes in the combustion state.

[0076] Further, the real-time synchronous compression transform algorithm comprises: rearranging the coefficients of the time-frequency transform by a synchronous compression operator to move the time-frequency coefficients of the signal at any point on the time-frequency plane to the center of gravity of the energy, enhance the energy concentration degree of the instantaneous frequency, and apply a window function on the basis of the Fourier transform to divide the signal time domain, obtain the frequency distribution under the window at different moments by sliding the window function, and arrange these short-time spectra in time sequence to describe the time-varying law of the frequency components of the signal.

[0077] For a determined time-varying signal The window function of the short-time Fourier transform is g(t), and after STFT transformation, there is: G(t, f) = ∫R s(τ)g(τ-t)e-i2πf(τ-t)dτ

[0078] Wherein, G(t,f) represents the result of Fourier transform, s(τ) represents the original signal, τ represents the time variable of the original signal, g(τ-t) represents, t represents the time variable, f represents the frequency variable, i represents the imaginary unit;

[0079] The Fourier transform is performed on the time-varying harmonic signal and the window function signal, and the estimation result of the instantaneous frequency in the STFT time-frequency representation is obtained, which is represented as:

[0080] Wherein, θ(t,ω) represents the instantaneous frequency, ω represents the frequency variable, arg(G(t,f)) represents the argument, Re represents the real part, The partial derivative with respect to time t is represented as:

[0081] The above instantaneous frequency estimation result is used to collect STFT coefficients with the same frequency, and according to the Dirac function property, the synchronous compression operator function ∫ R δ(η-ω0(t,ω)) is given, and the synchronous compression transform is represented as: T s (t,η)=∫ R G(t,f)δ(η-θ(t,ω))dω

[0082] Wherein, T s (t,η) represents the instantaneous frequency signal, δ(η-θ(t,ω)) represents the Dirac function, η represents the instantaneous frequency variable, the time-frequency coefficients are rearranged in the frequency direction, and the synchronous compression transform process at time t is that the frequency points are rearranged and go through all time points.

[0083] S4: Combining sensor data, performing multi-parameter joint analysis to obtain combustion state evaluation, using multiple high-speed cameras to obtain flame images from different angles, reconstructing three-dimensional flame morphology through computer vision technology, and providing fault detection information.

[0084] Furthermore, the multi-parameter joint analysis includes analyzing the data collected by the sensor to understand the distribution and characteristics, adjusting the parameters through an optimization algorithm, and finding the best parameter combination through multiple iterations;

[0085] The parameter combination is substituted into the multi-parameter joint analysis formula to calculate the weighted contribution of each sensor, the weight considers the different sensitivities of the sensor to angle, speed and light intensity, and a normalization function is applied to ensure the comparability between different sensor data, and the specific formula is:

[0086] Wherein, G(θ, v, I) represents a multi-parameter joint analysis main function, θ represents a flame tilt angle, v represents a flame burning speed, I represents a flame radiation intensity, n represents a sensor number, Ω i represents an effective measurement domain of the i-th sensor, f i (θ i ,v i ,I i ) represents a weighted contribution, α i represents an i-th sensor angle parameter, θ i represents an angle value measured by the i-th sensor, θ 0i represents a reference angle value of the i-th sensor, β i represents an adjustment coefficient of the i-th sensor speed parameter, γ i represents an adjustment coefficient of the i-th sensor light intensity parameter, respectively represent a mean value and a standard deviation of the i-th sensor angle measurement value, respectively represent a mean value and a standard deviation of the i-th sensor speed measurement value.

[0087] Further, the reconstructed three-dimensional flame shape includes, in combination with the result of the multi-parameter joint analysis, reconstructing the three-dimensional shape of the flame by using a computer vision algorithm;

[0088] By considering the change of the flame over time through a time correlation function, a dynamic three-dimensional flame model is obtained: Fx,y,z,t)=∫ Ω Gθ,v,I)·htdt h(t)=sin(ωt+φ)·e -λt

[0089] Wherein, x, y, z represent the coordinate axes of the three-dimensional space, ht represents a function of the change of the flame over time, Ω represents a spatial domain of integration, φ represents a phase shift, and λ represents an attenuation coefficient; the three-dimensional flame model gives the complete shape of the flame in the three-dimensional space, including the change over time.

[0090] S5: constructing a fault prediction model to automatically complete the entire fault detection process and realize early warning of combustion faults.

[0091] Further, the fault prediction model includes comparing the calculated flame state and flame shape with a normal behavior threshold value, and if the calculation result exceeds the preset threshold value range, it is considered that there is an abnormal behavior;

[0092] When the flame temperature is lower than 1000K or higher than 2500K, it is determined that the abnormal behavior is insufficient combustion or burner blockage, and the solution measure is to adjust the oxygen / fuel ratio of the burner and check and clean the internal carbon deposition and deposits of the burner;

[0093] When the flame oscillation frequency is lower than 0.5 Hz or higher than 2 Hz, and the flame shape is lower than 0.5 or higher than 1.5, it is determined that the abnormal behavior is the damage of the internal structure of the burner, the instability of the fuel supply, and the solution measures are to check the fuel supply system and keep the fuel flow stable; check whether the nozzle and combustion chamber of the burner are blocked, check whether the internal structure is damaged;

[0094] When the contrast of the flame image is lower than 0.6 or higher than 1.5, it is determined that the abnormal behavior is caused by improper camera settings or the characteristics of the flame itself, and the solution measures are to adjust the parameters of the camera to improve the image quality, check whether the lens of the camera is contaminated or damaged; check the brightness and color characteristics of the flame;

[0095] The detected abnormal behaviors and causes are summarized as fault detection information and recorded in the database for subsequent analysis and reporting, and a fault detection report is generated regularly to monitor the health status and performance of the combustion process.

[0096] Embodiment 2

[0097] Referring to FIGS. 2 and 3, for an embodiment of the present application, an oscillating combustion fault detection method based on synchronous compression transformation is provided, and scientific demonstration is carried out through experiments to verify the beneficial effects of the present application.

[0098] Fig. 2 is two verification working conditions involved in the embodiment of the application, first, the pressure signals of the verification experimental table in oscillation and stable working conditions are collected, and STFT analysis is performed on the pressure signals, the obvious oscillation main frequency is identified near t=0.75s in the oscillation working condition, but the frequency resolution is poor; and for the stable working condition, no obvious oscillation main frequency appears. Fig. 3 is the synchronous compression transform analysis result of the pressure signals in the two working conditions, for the oscillation working condition, the synchronous compression transform concentrates the energy of the measured signal, superimposes the frequency spectrum in the pseudo-frequency interval, concentrates the energy on the actual instantaneous frequency, improves the time-frequency concentration, and greatly improves the frequency resolution; the STFT analysis result of the flame CH* light intensity signals in the two working conditions, the obvious main frequency response appears near t=0.2s in the oscillation working condition, and no oscillation main frequency appears in the stable working condition. And the flame CH* light intensity signal shows obvious time advantage in the diagnosis speed compared with the pressure signal. The synchronous compression transform analysis result of the flame CH* light intensity signals in the two working conditions, through the synchronous compression transform analysis, the frequency domain resolution of the time-frequency analysis result is greatly improved. The flame image is subjected to data dimension reduction and feature extraction by the intrinsic orthogonal decomposition method, and the first-order modal time coefficient is obtained, which is used as the data input of STFT to diagnose the oscillation combustion fault again. The first-order modal time coefficient of the intrinsic orthogonal decomposition is used as the input of STFT, and the obvious main frequency response appears at t=0.01s, and the diagnosis speed is greatly improved; like the pressure signal and the CH* light intensity signal, there is also the problem of low frequency resolution. Through the synchronous compression transform analysis, the frequency resolution of the frequency spectrum is greatly improved, and the frequency positioning is more accurate.

[0099] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application but not limit the present application. Although the present application is described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced equivalently without departing from the spirit and scope of the technical solutions of the present application, and all should be covered in the scope of the claims of the present application.

[0100] Embodiment 3

[0101] The third embodiment of the present application is different from the first two embodiments:

[0102] If the functions are implemented in the form of software function units and sold or used as independent products, they can be stored in a computer readable storage medium. Based on this understanding, the technical solutions of the present application or the parts of the technical solutions that essentially contribute to the prior art or the parts of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM, Read-Only Memory), a random access memory (RAM, Random Access Memory), a magnetic disk or an optical disk, and various media that can store program codes.

[0103] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a list of executable instructions for implementing logic functions, which can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus or device, such as a computer-based system, a system including a processor or other system that can fetch the instructions from the instruction execution system, apparatus or device and execute the instructions, or in conjunction with these instructions execution systems, apparatus or devices. For the purpose of this specification, the "computer-readable medium" can be any device that can contain, store, communicate, propagate or transport programs for use by or in connection with an instruction execution system, apparatus or device, or in conjunction with these instruction execution systems, apparatus or devices.

[0104] More specific examples (non-exhaustive list) of the computer-readable medium include the following: an electrical connection having one or more wires (electrical devices), a portable computer diskette (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, because the program can be electronically obtained, for example, by optical scanning of the paper or other medium, followed by editing, interpreting or otherwise processing, if necessary, in other suitable ways, to be electronically obtained and then stored in the computer memory.

[0105] It should be understood that various parts of the present application can be implemented in hardware, software, firmware or a combination thereof. In the above embodiments, a plurality of steps or methods can be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and as in another embodiment, it can be implemented using any or a combination of the following technologies, which are known in the art: discrete logic circuitry having logic gates for implementing logic functions on data signals, application specific integrated circuits having appropriate combinational logic gates, programmable gate arrays (PGA), field programmable gate arrays (FPGA), etc.

[0106] Embodiment 4

[0107] Referring to FIG. 4, for one embodiment of the present application, a synchronous compression transform-based oscillatory combustion fault detection system is provided, characterized by comprising a data acquisition module, an intrinsic orthogonal decomposition method module, a synchronous compression transform module, a multi-parameter joint analysis module, a three-dimensional flame morphology reconstruction module, and a fault detection module.

[0108] The data acquisition module uses high-sensitivity flame chemiluminescence sensors to accurately capture the light signals generated by chemical reactions in the flame. The sensor array is arranged at different positions of the burner to obtain the overall chemiluminescence characteristics of the flame.

[0109] The intrinsic orthogonal decomposition method module accurately predicts faults by analyzing and identifying change patterns in the combustion process.

[0110] The synchronous compression transform module uses real-time SCT algorithms to perform time-frequency analysis on sensor data, capturing instantaneous changes in the combustion process, reflecting subtle changes in the combustion state, and improving the sensitivity of fault detection.

[0111] The multi-parameter joint analysis module uses multi-parameter joint analysis algorithms to evaluate the flame state based on the results of the synchronous compression transform.

[0112] The three-dimensional flame morphology reconstruction module uses computer vision technology to reconstruct the three-dimensional flame morphology from flame images obtained from different angles.

[0113] The fault detection module analyzes the results of multi-parameter joint analysis and three-dimensional flame morphology reconstruction to detect abnormal behavior of the flame.

[0114] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not limiting. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the present application, and they should be included in the scope of the claims of the present application.

Claims

1. A synchronous compression transform based oscillatory combustion fault detection method characterized by: comprising, acquiring a flame chemiluminescence signal, and performing adaptive filtering and noise reduction on the acquired flame chemiluminescence signal; extracting a first-order modal time coefficient from a flame image by using an improved proper orthogonal decomposition method, and obtaining comprehensive combustion feature information by using a high-order modal; developing a real-time synchronous compression transform algorithm to perform time-frequency analysis on data, and updating energy distribution in real time to respond to slight changes in the combustion state; combining sensor data to perform multi-parameter joint analysis to obtain a combustion state evaluation, using multiple high-speed cameras to obtain flame images from different angles, reconstructing a three-dimensional flame shape by using computer vision technology, and providing fault detection information; constructing a fault prediction model to automatically complete the entire fault detection process, and realizing early warning of combustion faults.

2. A synchronous compression transform based oscillatory combustion fault detection method as recited in claim 1, characterized by: The adaptive filtering and noise reduction includes, when detecting an oscillating combustion fault, distinguishing noise from an actual combustion oscillation signal, selecting an adaptive filter based on the steepest descent method, and combining noise estimation and signal enhancement in the filter. The adaptive filter update rule is as follows: where w(n) represents a filter weight vector and n represents a discrete time step, represents a noise power estimation at the n th moment, ε represents a constant, x(n) represents an input flame signal at the n th moment, e(n) represents an error signal at the n th moment, and α represents a forgetting factor.

3. A synchronous compression transform based oscillatory combustion fault detection method as recited in claim 2, characterized by: The improved proper orthogonal decomposition method includes, performing POD analysis on a set of flame images processed by a filter to obtain eigenmodes and eigenvalues, selecting M eigenmodes with the highest energy according to the size of the eigenvalues, and forming a new modal set representing the most significant spatial features in the data. A time coefficient is calculated for each selected mode, the time coefficient representing the change in the projection of the flame image onto each eigenmode over time, calculated as follows: where c i (m) denotes a time coefficient, I(m) denotes the mth flame image frame, denotes the mean of the set of images, φ i (x,y) denotes the eigenmodes of the flame image, i is the index of the mode; The calculated time coefficients are stored in a set: wherein, represents a set containing high-order modal time coefficients, and the set reveals subtle fluctuations and early signs of faults in the combustion process.

4. A synchronous compression transform based oscillatory combustion fault detection method as recited in claim 3, characterized by: The development of the real-time synchronous compression transform algorithm includes, rearranging the coefficients of a time-frequency transform by using a synchronous compression operator to move the time-frequency coefficients of a signal at any point in the time-frequency plane to the center of gravity of the energy, enhancing the energy concentration degree of the instantaneous frequency, applying a window function to the Fourier transform to divide the signal time domain, obtaining the frequency distribution under different time windows by sliding the window function, arranging these short-time spectra in time sequence to describe the time-varying law of the signal frequency components, and For a deterministic time-varying signal The window function of the short-time Fourier transform is g(t), and after STFT transformation, we have: G(t,f) = ∫ R s(τ)g(τ-t)e-i2πf(τ-t)dτ where G(t,f) represents the result of the Fourier transform, s(τ) represents the original signal, τ represents the time variable of the original signal, g(τ-t) represents, t represents the time variable, f represents the frequency variable, and i represents the imaginary unit; The Fourier transform is performed on the time-varying harmonic signal and the window function signal to obtain an estimation result of the instantaneous frequency in the STFT time-frequency representation, which is represented as: where θ(t,ω) denotes the instantaneous frequency, ω denotes the frequency variable, arg(G(t,f)) denotes the argument, and Re denotes the real part, represents the partial derivative with respect to time t. The STFT coefficients with the same frequency are collected by using the above instantaneous frequency estimation result, and according to the Dirac function property, a synchronous compression operator function ∫ R δ(η-ω0(t,ω)) is given, and a synchronous compression transform is obtained, which is represented as: T s (t,η) = ∫ R G(t,f) δ(η - θ(t,ω)) dω where T s (t,η) denotes the instantaneous frequency signal, δ(η-θ(t,ω)) denotes the Dirac function, η denotes the instantaneous frequency variable, the time-frequency coefficients are rearranged in the frequency direction, and the synchronous compression transform process at time t, i.e., the frequency point rearrangement and the duration of all time points.

5. A synchronous compression transform based oscillatory combustion fault detection method as recited in claim 4, characterized by: The multi-parameter joint analysis includes, analyzing the data collected by the sensors to understand the distribution and characteristics, adjusting the parameters by using an optimization algorithm, and finding the best parameter combination after multiple iterations; The parameters are combined into the multi-parameter joint analysis formula to calculate the weighted contribution of each sensor, the weight considers the different sensitivity of the sensor to the angle, speed and light intensity, and the normalization function is applied to ensure the comparability between the data of different sensors, and the specific formula is: wherein G(0, v, I) represents a multi-parameter joint analysis master function, 0 represents a flame tilt angle, v represents a flame burning speed, I represents a flame radiation intensity, n represents a number of sensors, Ω i represents an effective measurement domain of the i-th sensor, f i (0 i , v i , I i ) represents a weighted contribution, a i represents an i-th sensor angle parameter, 0 i represents an angle value measured by the i-th sensor, 0 0i represents a reference angle value of the i-th sensor, b i represents an adjustment coefficient of the i-th sensor speed parameter, g i represents an adjustment coefficient of the i-th sensor light intensity parameter, respectively denote the mean and standard deviation of the i-th sensor angle measurement, respectively represent the mean and standard deviation of the speed measurement value of the i th sensor.

6. A synchronous compression transform based oscillatory combustion fault detection method as recited in claim 5, characterized by: The reconstruction of the three-dimensional flame shape includes, combining the results of the multi-parameter joint analysis to reconstruct the three-dimensional shape of the flame by using computer vision algorithms; By considering the change of the flame over time through a time correlation function, a dynamic three-dimensional flame model is obtained: F(x, y, z, t) = ∫ Ω G(θ, v, I) · h(t)dt h(t) = sin(ωt + φ) - e -λt Wherein, x, y, z represent the coordinate axes of three-dimensional space, h(t) represents the function of flame change over time, Ω represents the spatial domain of integration, φ represents the phase shift, and λ represents the attenuation coefficient; The complete form of the flame in three-dimensional space is given by the three-dimensional flame model, including the change over time.

7. A synchronous compression transform based oscillatory combustion fault detection method as recited in claim 6, characterized by: The fault prediction model includes comparing the calculated flame state and flame form with the normal behavior threshold, and if the calculation result exceeds the preset threshold range, it is considered that there is abnormal behavior; When the flame temperature is lower than 1000K or higher than 2500K, it is determined that the abnormal behavior is insufficient combustion or burner blockage, and the solution is to adjust the oxygen / fuel ratio of the burner, check and clean the internal carbon deposition and sediment of the burner; When the flame oscillation frequency is lower than 0.5Hz or higher than 2Hz, and the flame shape is lower than 0.5 or higher than 1.5, it is determined that the abnormal behavior is the damage of the internal structure of the burner or the instability of the fuel supply, and the solution is to check the fuel supply system and keep the fuel flow stable; Check whether the nozzle and combustion chamber of the burner are blocked, check whether the internal structure is damaged; When the contrast of the flame image is lower than 0.6 or higher than 1.5, it is determined that the abnormal behavior is caused by improper camera settings or the characteristics of the flame itself, and the solution is to adjust the camera parameters to improve the image quality, check whether the camera lens is contaminated or damaged; Check the brightness and color characteristics of the flame; The detected abnormal behavior and cause are summarized as fault detection information and recorded in the database for subsequent analysis and reporting, and a fault detection report is generated regularly to monitor the health status and performance of the combustion process.

8. A system employing an oscillating combustion fault detection method based on synchronous compression transform according to any one of claims 1 to 7, characterized in that: It comprises a data acquisition module, an intrinsic orthogonal decomposition method module, a synchronous compression transformation module, a multi-parameter joint analysis module, a three-dimensional flame form reconstruction module, and a fault detection module. The data acquisition module uses high-sensitivity flame chemiluminescence sensors to accurately capture the light signals generated by chemical reactions in the flame, and the sensor array is arranged at different positions of the burner to obtain the overall chemiluminescence characteristics of the flame. The intrinsic orthogonal decomposition method module accurately predicts faults by analyzing and identifying change patterns in the combustion process. The synchronous compression transformation module uses real-time SCT algorithm to perform time-frequency analysis on sensor data, captures transient changes in the combustion process, reflects subtle changes in the combustion state, and improves the sensitivity of fault detection. The multi-parameter joint analysis module uses a multi-parameter joint analysis algorithm to evaluate the flame state based on the results of synchronous compression transformation. The three-dimensional flame form reconstruction module uses computer vision technology to reconstruct the three-dimensional flame form from flame images obtained from different angles. The fault detection module analyzes the results of multi-parameter joint analysis and three-dimensional flame form reconstruction to detect abnormal behavior of the flame. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to realize the steps of the method of any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the method of any one of claims 1 to 7.

Images