Homologous vibration reference-based flame detection signal anti-interference processing method, system, medium and product

Abstract

This application discloses a method, system, medium, and product for anti-interference processing of flame detection signals based on homogeneous vibration reference in the field of flame detection and early warning. The method includes: S100, simultaneously acquiring raw electrical signals from a main signal channel, at least one auxiliary signal channel, and a vibration sampling signal channel; S200, performing filtering preprocessing on the input signals and extracting vibration features from the filtered vibration sampling signal channel; S300, configuring an adaptive filter based on the vibration feature parameter P, estimating vibration noise components through the adaptive filter, and filtering out noise components from the main signal channel and auxiliary signal channel to obtain a clean signal; S400, outputting a judgment signal and an early warning signal for flame detection based on the clean signal. This application achieves accurate flame detection judgment by processing homogeneous vibration signals and performing high-confidence filtering of vibration noise.

Description

Technical Field

[0001] This application relates to a method, system, medium, and product for anti-interference processing of flame detection signals based on homogeneous vibration reference in the field of flame detection and early warning. Background Technology

[0002] Pyroelectric flame detectors detect flame signals using pyroelectric sensors sensitive to infrared radiation of specific wavelengths. Their core sensing element is a piezoelectric material, which converts changes in incident light radiation into electrical signals. However, it also converts externally applied mechanical stress or structural vibrations into electrical signals. In applications involving mechanical vibration, these vibration-induced electrical signals can overlap with the flame signal frequency band, creating interference known as vibration noise or piezoelectric noise. This noise can severely contaminate or even drown out the photoelectric signal originating from the flame, causing false alarms where the detection system mistakes noise for flame, or missed alarms where the noise drowns out the valid signal, thus reducing the reliability of flame detection.

[0003] A conventional design approach employs multi-channel detection combined with signal processing algorithms. This approach sets up a main detection channel and an auxiliary reference channel. The main channel sensor is equipped with an optical filter targeting the characteristic wavelength of flames, while the auxiliary channel sensor uses a filter that is insensitive to flames but sensitive to ambient background radiation. The flame signal is distinguished from steady-state background radiation interference by comparing the differences between the two signals. However, this approach's core design objective is to differentiate signals generated by different types of radiation sources and determine whether they are fire source signals. It is ineffective against piezoelectric noise generated by mechanical vibrations in the sensor. Vibration noise originates from the intrinsic piezoelectric effect of the sensor material and is independent of the incident light signal. Therefore, the optical filter in the auxiliary channel cannot selectively target it. Both the main and auxiliary channel sensors, when subjected to the same mechanical vibration, will generate interference signals from the same source, rendering the comparative analysis algorithm based on optical filter differences ineffective.

[0004] To improve the vibration resistance of pyroelectric flame detectors, some technical solutions employ seismic-resistant designs for the detector's mounting bracket, housing, or sensor module, reducing the direct excitation of the pyroelectric sensor by vibration at the physical level. However, this approach has fundamental limitations in the application scenario of wind turbine nacelles. The vibration of wind turbine nacelles originates from their own operation, such as the operation of the gearbox and generator, the rotation of the blades and aerodynamic loads, as well as varying environmental wind loads. These vibrations are the result of the interaction between the equipment's inherent operating mode and the external environment, characterized by wide frequency range, randomness, large amplitude variations, and unpredictability. Eliminating or significantly isolating these vibrations that coexist with equipment operation at their source is extremely costly and almost impractical in engineering.

[0005] Some cutting-edge research attempts to introduce more complex signal analysis and artificial intelligence methods, constructing correlation learning models from the raw outputs of vibration sensors and detectors through feature fusion and modeling to learn the complex mapping relationship between vibration and noise from historical data. However, such machine learning-based solutions suffer from significant accuracy and reliability deficiencies in practical engineering applications, especially in the complex scenarios of wind turbine nacelles. The constantly changing operating conditions of wind turbines and ambient wind speeds result in strong time-varying and non-stationary vibration characteristics, making it difficult for correlation models trained on static historical data to effectively track and adaptively adjust. Furthermore, the vibration-noise correlation learning of deep models has a "black box" characteristic; its decision-making logic lacks a clear physical explanation, and the model's generalization ability and robustness drop sharply when facing unknown disturbances, potentially leading to misjudgments that are difficult to trace and understand. Summary of the Invention

[0006] The purpose of this application is to overcome the shortcomings of the prior art and provide a flame detection signal anti-interference processing method, system, medium and product based on the same source vibration reference. By processing the same source vibration signal, high confidence filtering of vibration noise is performed, thereby achieving accurate flame detection judgment.

[0007] Firstly, the anti-interference processing method for flame detection signals based on homogeneous vibration reference provided in this application is applied to the nacelle fire alarm monitoring of wind turbines. The technical solution adopted includes the following steps: S100, synchronously acquires raw electrical signals from the main signal channel, at least one auxiliary signal channel, and one vibration sampling signal channel; wherein, the main signal channel, the auxiliary signal channel, and the vibration sampling signal channel use the same pyroelectric sensor, and the pyroelectric sensors of different signal channels input different filter signals; S200, for the main signal channel S main Auxiliary signal channel S aux and vibration sampling signal channel S ref Perform filtering preprocessing to obtain the filtered main signal channel S main Auxiliary signal channel S aux 'and vibration sampling signal channel S ref ';For the filtered vibration sampling signal channel S ref 'Perform vibration feature extraction, and extract vibration feature parameters P that characterize the vibration state;' S300 dynamically adjusts the parameters of the adaptive filter based on the vibration characteristic parameter P, and then processes the filtered vibration sampling signal channel S through the adaptive filter. ref 'Conduct vibration noise component V' est The estimation will be applied to the filtered main signal channel S. main 'and auxiliary signal channel Saux 'Input an independent adaptive filter, through the noise component V est After filtering, the pure main signal channel S after vibration filtering is obtained. main-clean and pure auxiliary signal channel S aux-clean ; S400, based on the pure main signal channel S main-clean and pure auxiliary signal channel S aux-clean The comparison calculation outputs a judgment signal and a warning signal for flame detection.

[0008] By adopting the above technical solution, and through a collaborative architecture of homogeneous vibration reference sampling and feature-driven adaptive noise cancellation, a technical path is provided for sampling, modeling, and real-time filtering of vibration noise from its source. This solves the core problems of inaccurate filtering and poor adaptability in existing technologies due to the lack of direct noise references, providing a methodological guarantee for achieving high accuracy and high reliability in flame detection under complex vibration environments.

[0009] Preferably, in S100, the main signal channel uses a pyroelectric sensor that inputs a filtered signal in the infrared band, the auxiliary signal channel includes at least one pyroelectric sensor that inputs a filtered signal in the ultraviolet band, and the vibration sampling signal channel uses a fully enclosed, opaque pyroelectric sensor.

[0010] By adopting the above technical solution, the main signal channel and the auxiliary signal channel achieve the traditional flame-background light discrimination function through optical filtering differences; the fully enclosed design of the vibration sampling signal channel ensures that the sensor only responds to mechanical vibration and does not respond to any light signal at all, thereby outputting a pure vibration noise reference signal and improving the accuracy of noise component estimation.

[0011] Preferably, in S200, the main signal channel S is acquired. main Auxiliary signal channel S aux and vibration sampling signal channel S ref For signals within a preset time window, the specific steps include: S201, for the main signal channel S main Auxiliary signal channel S aux and vibration sampling signal channel S ref A low-pass filter is used to perform low-pass filtering, resulting in the filtered main signal channel S. main Auxiliary signal channel S aux 'and vibration sampling signal channel S ref '; S202: Acquire the fan speed information and convert it into frequency information. Based on the fan frequency information, configure a narrowband bandpass filter with adjustable center frequency, and filter the vibration sampling signal channel S.ref Input narrowband bandpass filter, output regular vibration component V reg ; S203, for the regular vibration component V reg Extracting the dominant frequency F of regular vibration reg and regular vibration amplitude A reg ; S204, the filtered vibration sampling signal channel S ref 'Through a frequency F that is dominated by regular vibrations reg A dynamic notch filter with a center frequency is used to filter out regular vibration components V. reg The irregular vibration component V was obtained. irr ; S205, for irregular vibration component V irr Extracting the root mean square of vibration intensity A irr Broadband vibration intensity variance A var ; The vibration characteristic parameter P is the dominant frequency F of the regular vibration. reg、 Regular vibration amplitude A reg Root mean square of vibration intensity A irr and broadband vibration intensity variance A var A set of.

[0012] By adopting the above technical solution, the input signal is low-pass filtered to remove stray signals generated by the circuit itself; the periodic vibrations that are strongly correlated with the rotational speed in the wind turbine vibration are extracted and classified separately from other random vibrations, the complex mixed noise problem is decomposed and processed, and multi-dimensional feature parameters are extracted from it, providing a reliable basis for the accurate and differentiated configuration of the subsequent adaptive filter.

[0013] Preferably, several narrowband bandpass filters are configured in parallel. The center frequency of one of the narrowband bandpass filters is configured based on the frequency information of the wind turbine, and the center frequencies of the other narrowband bandpass filters are configured based on the harmonic frequencies of the wind turbine frequency. The filtered vibration sampling signal channel S ref After filtering by each parallel narrowband bandpass filter, the outputs of each narrowband bandpass filter are weighted and summed according to harmonic weights to obtain the regular vibration component V. reg .

[0014] By adopting the above technical solution and processing the fundamental and harmonic components in parallel, the entire regular vibration component can be captured and reconstructed more comprehensively and accurately. This prevents incomplete separation of irregular vibrations caused by ignoring important harmonic components, resulting in residual noise estimation. This makes the modeling of regular vibrations closer to the real physical situation and improves the vibration filtering effect.

[0015] Preferably, in S300, the adaptive filter adopts a normalized minimum mean square filter, including a regular vibration filter and an irregular vibration filter, specifically including the following steps; S301, based on the regular vibration dominant frequency F reg The length L1 of the tap weight vector W1 of the regular vibration filter and the length L2 of the tap weight vector W2 of the irregular vibration filter are constrained. The tap weight vectors are initialized to zero vectors, based on the regular vibration amplitude A. reg Configure the step size convergence factor μ1 of the regular vibration filter based on the broadband vibration intensity root mean square A. irr and broadband vibration intensity variance A var Configure the step size convergence factor μ2 of the irregular vibration filter; S302, at each discrete time point n, respectively from the regular vibration component V reg and irregular vibration component V irr Extract the first L1 or L2 samples to construct the reference input vector X: X reg (n)=[V reg (n), V reg (n-1), ..., V reg [(n-L1+1)]; X irr (n)=[V irr (n), V irr (n-1), ..., V irr [(n-L2+1)]; S303, perform a dot product between the weight vector of the adaptive filter and the reference input vector to calculate the current time-to-time response of the regular vibration component V. reg and irregular vibration component V irr Estimated value of vibration noise V est-reg and V est-irr : V est-reg =W1*X reg (n); V est-irr =W2*X irr (n); Vibration noise component V est =V est-reg +V est-irr ; S304, Obtain the filtered main signal channel S main ', Subtract the vibration noise component V from the signal channel est The signal deviation is obtained, that is, the pure main signal channel S. main-clean =S main '-Vest ; S305, based on the current moment's pure main signal channel S main-clean The tap weight vectors of the regular vibration filter and the irregular vibration filter are updated using the reference input vector X, respectively: W1(n+1)=W1(n)+[μ1 / (δ+||X reg (n)|| 2 ]*S main-clean *X reg (n); W2(n+1)=W2(n)+[μ2 / (δ+||X irr (n)|| 2 ]*S main-clean *X irr (n); Where δ is the regularization parameter, and is a minimum positive number to prevent division by zero; S306, time index n increments, return to S302, repeat the entire process, at each time n, the vibration noise component V output from S303 is... est Used for filtering the main signal channel S main The vibration noise is filtered out, and the pure main signal channel S304 is obtained simultaneously. main-clean Output the signal and update the tap weight vector; obtain the pure auxiliary signal channel S in the same way. aux-clean External output is achieved through the clean main signal channel S. main-clean and pure auxiliary signal channel S aux-clean To make subsequent judgments on the flame detection.

[0016] By adopting the above technical solution and using the NLMS filter, a good balance is achieved between convergence speed, steady-state error, and computational complexity. The dual-filter architecture corresponds to regular vibration and irregular vibration noise respectively, and the filter parameters are adjusted and adaptively adjusted accordingly to ensure that the filter can respond quickly to changes in vibration state. The tracking and filtering performance of vibration is optimized, and the robustness, convergence speed, and final filtering accuracy of the vibration filtering scheme are enhanced in complex and time-varying vibration environments.

[0017] As a preferred embodiment, S400 specifically includes: S401, from the clean main signal channel S main-clean Extracting flame intensity features within the frequency band from the pure auxiliary signal channel S aux-clean Extract the corresponding background light intensity features; S402, calculate the characteristic ratio R between the flame intensity feature and the background light intensity feature; S403, compare the intensity of the feature ratio R with the first dynamic fire detection threshold to determine whether it exceeds the threshold; S404, compare the duration for which the intensity of the characteristic ratio R exceeds the first dynamic fire detection threshold with the second dynamic fire detection threshold to determine whether the threshold is exceeded; S405, a flame is determined to exist only when both the conditions of the first dynamic fire detection threshold and the second dynamic fire detection threshold are met.

[0018] By adopting the above technical solution, after obtaining a clean signal, the system uses dual threshold judgment to prevent false alarms caused by transient interference or signal fluctuations, thus forming a highly reliable fire alarm output logic. This ensures that the system only alarms when the signal strength is sufficient and lasts for a certain period of time, effectively reducing the false alarm rate.

[0019] Preferably, in S400, a vibration dynamic weighting coefficient is obtained based on the vibration characteristic parameter P, and the values ​​of the first dynamic fire detection threshold and the second dynamic fire detection threshold are adjusted according to the vibration dynamic weighting coefficient.

[0020] By adopting the above technical solutions, the system has the ability to automatically balance sensitivity and reliability according to the real-time vibration environment, and the accuracy of the final judgment of fire alarm is further optimized and improved based on noise filtering.

[0021] Secondly, the flame detection signal anti-interference processing system based on the same source vibration reference provided in this application adopts the following technical solution, including a signal acquisition module, a filtering preprocessing module, an adaptive filtering module, a flame signal judgment module, and an early warning module; The signal acquisition module includes a main pyroelectric sensor, at least one auxiliary pyroelectric sensor, and a vibration sampling pyroelectric sensor to acquire monitoring input signals. The filtering preprocessing module is connected to the signal acquisition module and performs filtering preprocessing on the monitored input signal. It includes a low-pass filter submodule and a narrowband bandpass filter submodule. The narrowband bandpass filter submodule is connected to the control module of the wind turbine. The adaptive filtering module is connected to the filtering preprocessing module, performs adaptive filtering on the filtering preprocessing signal, and outputs a pure main signal channel and a pure auxiliary signal channel after vibration filtering. The flame signal determination module determines the flame detection based on the characteristic ratio of the pure main signal channel and the pure auxiliary signal channel; The early warning module issues an early warning signal based on the judgment signal from flame detection.

[0022] Thirdly, this application provides a computer-readable storage medium storing a computer program thereon. In the technical solution adopted, when the computer program is executed by a processor, it implements the steps of the above-mentioned anti-interference processing method for flame detection signals based on homogeneous vibration reference.

[0023] Fourthly, this application provides a computer program product, which includes a computer program or instructions, and the technical solution adopted enables the computer program or instructions to implement the steps in the above-mentioned anti-interference processing method for flame detection signals based on homogeneous vibration reference.

[0024] In summary, this application includes at least one of the following beneficial technical effects: 1. This application obtains a pure vibration noise sample that is completely consistent with the physical characteristics of the main and auxiliary interference channels and is not contaminated by any light signals by establishing an independent, fully enclosed, opaque vibration sampling channel using the same pyroelectric sensor. This overcomes the shortcomings of existing technologies that can only indirectly estimate noise from different modal vibration signals or cannot obtain noise references, and lays the physical and data foundation for the subsequent realization of high-precision noise filtering.

[0025] 2. This application realizes the refined modeling, separation, and dynamic tracking filtering of complex vibration noise. By introducing the fan speed, the noise is divided into regular and irregular components and multi-dimensional feature parameters are extracted for each. A dual NLMS filter architecture with dynamically adjustable parameters is adopted to optimize the estimation and cancellation of the two types of noise respectively. This enables the system to deeply understand and accurately decompose complex vibration noise and perform filtering in real time adaptively, which greatly improves the accuracy and adaptability of noise filtering.

[0026] 3. This application constructs a closed-loop adaptive system of "perception-filtering-judgment", introduces dual dynamic criteria of intensity and duration, and feeds back the vibration feature parameters extracted in real time to the judgment threshold itself, realizing the dynamic adjustment of the threshold and intelligently achieving the best balance between detection sensitivity and reliability. This forms a closed-loop optimization from environmental perception to final decision-making, which significantly improves the overall reliability and environmental intelligence of the final flame detection decision. Attached Figure Description

[0027] Figure 1 This is a flowchart illustrating the anti-interference processing method for flame detection signals based on homogeneous vibration reference in the embodiments of this application. Figure 2 This is a flowchart illustrating step S200 of the flame detection signal anti-interference processing method based on homogeneous vibration reference in an embodiment of this application. Figure 3 This is a flowchart illustrating step S300 of the flame detection signal anti-interference processing method based on homogeneous vibration reference in an embodiment of this application. Figure 4 This is a flowchart illustrating step S400 of the flame detection signal anti-interference processing method based on homogeneous vibration reference in the embodiments of this application. Figure 5 This is a schematic diagram of the architecture of the flame detection signal anti-interference processing system based on the same source vibration reference in the embodiments of this application. Detailed Implementation

[0028] This specific embodiment is merely an explanation of this application and is not intended to limit it. After reading this specification, those skilled in the art can make modifications to this embodiment without contributing any inventive step, but such modifications are protected by patent law as long as they are within the scope of this application.

[0029] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that in the optional embodiments of this application, the object information and other related data involved require the permission or consent of the object when the embodiments of this application are applied to specific products or technologies, and the collection, use, and processing of related data must comply with the relevant laws, regulations, and standards of the relevant countries and regions. That is to say, if the embodiments of this application involve data related to the object, it needs to be obtained with the authorization and consent of the object, the authorization and consent of the relevant departments, and in compliance with the relevant laws, regulations, and standards of the country and region. If personal information is involved in the embodiments, the acquisition of all personal information requires the consent of the individual. If sensitive information is involved, the separate consent of the information subject is required, and the embodiments also need to be implemented with the authorization and consent of the object.

[0030] The embodiments of this application will now be described in further detail with reference to the accompanying drawings.

[0031] National safety regulations and fire protection standards for wind turbine generators clearly stipulate that flame detection devices must be installed inside the nacelle to enable early fire warnings and ensure the safety of power facilities. During actual deployment and application, technicians have found that pyroelectric flame detectors installed in wind turbine nacelles generally have a high false alarm rate. In-depth research revealed that the core reason for this problem lies in the continuous and complex mechanical vibrations within the nacelle, which induce significant vibration noise in the pyroelectric sensors of the detectors. This noise signal may overlap with the actual flame flicker signal in the frequency domain, thus severely interfering with flame detection judgment.

[0032] The aforementioned interfering vibrations mainly originate from the wind turbine's own operating mode and the external wind load environment, and can be specifically categorized into two types: first, regular vibrations strongly correlated with rotational speed, generated by the rotation of the turbine rotor, gearbox, and generator; second, irregular broadband vibrations caused by randomly varying wind pressure acting on the nacelle structure. The combined effect of these two types of vibrations makes the vibration conditions under which the detector operates complex and severe. Traditional detectors struggle to effectively distinguish and filter out the resulting electrical noise, leading to false alarms or missed alarms and reducing the reliability of the monitoring system.

[0033] The flame detection signal anti-interference processing method based on homogeneous vibration reference in this application aims to efficiently and accurately filter out the vibration noise unique to wind turbine nacelles, thereby improving the accuracy and reliability of flame detectors in this scenario. Please refer to... Figure 1 Specifically, it includes the following steps.

[0034] S100 synchronously acquires raw electrical signals from the main signal channel, an auxiliary signal channel, and a vibration sampling signal channel. The main signal channel, auxiliary signal channel, and vibration sampling signal channel all use the same pyroelectric sensor, with different wavelength filters installed in front of the pyroelectric sensors in different signal channels to input the corresponding filtered signals.

[0035] More specifically, in the embodiments of this application, the main signal channel uses a 4.3μm infrared pyroelectric sensor to acquire the characteristic infrared light emitted by the vibrational energy level transitions of CO2 molecules generated during combustion; the auxiliary signal channel uses a 5.0μm pyroelectric sensor to acquire light wavelengths far from the flame combustion, which are more sensitive to background blackbody radiation sources and are used to shield interference; the vibration sampling signal channel uses a fully enclosed pyroelectric sensor to completely shield the light signal and only collect signals caused by environmental vibrations. Using the same pyroelectric sensor ensures that the sensors in the three channels have completely consistent response characteristics to mechanical vibrations, maintaining consistency in amplitude, phase, and frequency responses caused by vibration signals. This allows the noise samples collected by the vibration sampling channel to describe the actual noise present in the main signal channel and the auxiliary signal channel. Simultaneously, the specific installation positions of the sensors in the three channels need to be close to each other in physical space to ensure that the vibration interference experienced by the three channels has the characteristic of "same source and same phase".

[0036] S200, for the main signal channel S main Auxiliary signal channel S aux and vibration sampling signal channel S ref Perform filtering preprocessing to obtain the filtered main signal channel S main Auxiliary signal channel S aux 'and vibration sampling signal channel S ref';For the filtered vibration sampling signal channel S ref 'Perform vibration feature extraction, extracting vibration feature parameters P that characterize the vibration state. Please refer to...' Figure 2 Specifically, it includes the following steps.

[0037] S201, for the main signal channel S main Auxiliary signal channel S aux and vibration sampling signal channel S ref A low-pass filter is used to perform low-pass filtering, resulting in the filtered main signal channel S. main Auxiliary signal channel S aux 'and vibration sampling signal channel S ref '.

[0038] The raw signals from the three signal channels are first low-pass filtered to perform signal preprocessing and preliminary frequency band shaping, in order to eliminate irrelevant interference such as high-frequency circuit noise and digital sampling noise introduced by the sensor and circuit itself, retain the baseband signal, and simplify subsequent signal processing.

[0039] S202: Acquire the fan speed information and convert it into frequency information. Based on the fan frequency information, configure a narrowband bandpass filter with adjustable center frequency, and filter the vibration sampling signal channel S. ref Input narrowband bandpass filter, output regular vibration component V reg .

[0040] Vibrations within a wind turbine nacelle are not entirely random; they include periodic vibration components strongly correlated with the operation of rotating components such as the turbine rotor, gearbox, and generator. The frequency of these vibrations is directly related to the rotational speed (and its harmonics). Acquiring the rotational speed and converting it into frequency information is crucial for accurately guiding the setting of the center frequency of subsequent filters. Using this frequency information, regular vibration components operating at the same frequency as the wind turbine can be identified and extracted from the wide-band vibration signal, achieving target separation of the primary vibration source. Rotational speed information can typically be obtained by directly reading real-time speed data from the wind turbine's main control or monitoring system, or by measuring speed sensors installed inside the nacelle or on the turbine's main shaft.

[0041] S203, for the regular vibration component V reg Extracting the dominant frequency F of regular vibration reg and regular vibration amplitude A reg .

[0042] S204, the filtered vibration sampling signal channel S ref 'Through a frequency F that is dominated by regular vibrations reg A dynamic notch filter with a center frequency is used to filter out regular vibration components V. reg The irregular vibration component V was obtained.irr .

[0043] Among them, the irregular vibration component V irr This refers to vibrations induced by non-periodic excitations such as random wind pressure turbulence, gust impacts, and random structural responses. It has a wide spectrum, strong randomness, no fixed dominant frequency, and large amplitude fluctuations. Regular components V are extracted using a narrowband bandpass filter. reg Then, a frequency F with regular vibration as the dominant frequency is adopted. reg A dynamic notch filter with a center frequency acts to create a deep stopband at and around that specific frequency, thereby blocking strong, regularly occurring vibrational components with known frequencies from the vibration sampling signal channel S. ref The system precisely filters out and extracts irregular vibration components. By employing a decomposition strategy of first extracting and then filtering, it enables independent and targeted analysis and subsequent processing of two types of noise with vastly different characteristics, greatly improving the accuracy of noise modeling.

[0044] S205, for irregular vibration component V irr Extracting the root mean square of vibration intensity A irr Broadband vibration intensity variance A var .

[0045] The vibration characteristic parameter P is the dominant frequency F of the regular vibration. reg Regular vibration amplitude A reg Root mean square of vibration intensity A irr and broadband vibration intensity variance A var A set of.

[0046] Among them, the dominant frequency F of regular vibration reg The most prominent periodic frequency characterizing regular vibrations typically corresponds to the fundamental frequency of the fan speed. This parameter directly determines the core operating frequency of adaptive filters or notch filters used to remove regular noise; the amplitude of the regular vibration, A... reg The intensity or energy level of the regular vibration component reflects the strength of the regular disturbance and is used to dynamically adjust convergence parameters such as the step size of the subsequent adaptive filter; the root mean square of the vibration intensity A irr Characterizing the irregular vibration component V irr The average energy level describes the intensity of random vibrations, and the broadband vibration intensity variance A var Characterizing the irregular vibration component V irr The magnitude or dispersion of the amplitude is used to adjust the convergence parameters such as the step size of the subsequent adaptive filter in real time.

[0047] It should be noted that the acquired main signal channel S main Auxiliary signal channel S aux and vibration sampling signal channel Sref The signal is within a preset time window. Within different sliding time windows, the regular vibration component V... reg Irregular vibration component V irr The vibration characteristic parameter P is obtained through calculation.

[0048] In another embodiment, several narrowband bandpass filters are configured in parallel. The center frequency of one narrowband bandpass filter is configured based on the frequency information of the wind turbine, while the center frequencies of the remaining narrowband bandpass filters are configured based on the harmonic frequencies of the wind turbine frequency. The filtered vibration sampling signal channel S... ref After filtering by each parallel narrowband bandpass filter, the outputs of each narrowband bandpass filter are weighted and summed according to harmonic weights to obtain the regular vibration component V. reg .

[0049] In actual wind turbine operation, the spectral composition of the regular vibrations generated by rotating components and the transmission chain is not a single pure sine wave. Besides the fundamental frequency component directly corresponding to the rotational speed, it also includes harmonic components that are integer multiples of the fundamental frequency, as well as possible sideband components, such as harmonics generated by the rotation of multiple blades. These harmonic components are also an important part of the regular, predictable vibrations, and their energy cannot be ignored. In the embodiments of this application, by constructing a filter bank by setting several narrowband bandpass filters in parallel, the fundamental frequency and all higher-order harmonic components can be extracted simultaneously from the original signal, and a more accurate total regular vibration component V can be synthesized according to their contribution ratio in the actual vibration energy. reg The harmonic weights are derived from prior parameters obtained through engineering experience and experimental data. This is achieved through the regular vibration component V. reg It can optimize irregular vibration components V irr This improves the stripping effect and enhances the accuracy and reliability of the input signal for subsequent adaptive filters.

[0050] S300 dynamically adjusts the parameters of the adaptive filter based on the vibration characteristic parameter P, and then processes the filtered vibration sampling signal channel S through the adaptive filter. ref 'Conduct vibration noise component V' est The estimation will be applied to the filtered main signal channel S. main 'and auxiliary signal channel S aux 'Input an independent adaptive filter, through the noise component V est After filtering, the pure main signal channel S after vibration filtering is obtained. main-clean and pure auxiliary signal channel S aux-clean .

[0051] In the embodiments of this application, the adaptive filter employs a normalized least mean square (NLMS) filter, including a regular vibration filter and an irregular vibration filter. See also... Figure 3 Specifically, it includes the following steps.

[0052] S301, based on the regular vibration dominant frequency F reg The length L1 of the tap weight vector W1 of the regular vibration filter and the length L2 of the tap weight vector W2 of the irregular vibration filter are constrained. The tap weight vectors are initialized to zero vectors, based on the regular vibration amplitude A. reg Configure the step size convergence factor μ1 of the regular vibration filter based on the broadband vibration intensity root mean square A. irr and broadband vibration intensity variance A var Configure the step size convergence factor μ2 of the irregular vibration filter.

[0053] The tap weight vector length of the normalized minimum mean square filter determines the memory depth that the filter can model. The optimal memory depth requirement differs for different types of noise. Regular vibration component V reg Given the strong periodicity of a signal, to effectively model and predict it, the filter needs to cover at least one or more complete cycles of the signal. Therefore, the constraint is to ensure that the time window length corresponding to L1 covers at least one or more main cycles of the regular oscillation, such that L1 = k*(Fs / F...). reg ), Fs / F reg The number of sampling points corresponds to the fundamental frequency period, and k is the period coefficient, which can be an integer from 2 to 5. When the dominant frequency of the regular vibration is F... reg As the fan speed changes, L1 is also dynamically adjusted to ensure that the filter always has the ability to effectively model the signal of the current cycle.

[0054] For the irregular vibration component V irr Without a dominant single period, the constraint of L2 is not related to the regular oscillation dominant frequency F. reg It is directly linked. However, L2 is constrained by L1 and needs to be set longer than L1 in order to capture enough random signal samples for effective statistics.

[0055] The step size convergence factor determines the convergence speed and steady-state error of the NLMS adaptive filter. A larger step size convergence factor results in a larger filter weight update amplitude and a faster convergence to the optimal solution, but it also causes larger fluctuations near the optimal solution, leading to a larger steady-state error.

[0056] In the embodiments of this application, the vibration intensity characteristics of the vibration signal include a regular vibration amplitude A. reg Root mean square of vibration intensity A irrand broadband vibration intensity variance A var It directly reflects the current noise amplitude level and the degree of fluctuation, and is the direct basis for dynamically adjusting the step size convergence factor.

[0057] In the embodiments of this application, a monotonically increasing function is used to configure the step size convergence factor.

[0058] μ1=α*A reg ; μ2=β / (A irr +γ*A var ); Where α is the proportionality coefficient, and β and γ are the normalized weighting coefficients.

[0059] For the step-size convergence factor μ1, when the regular oscillation amplitude A reg A larger value indicates severe regular noise interference. In this case, obtaining a larger μ1 allows the regular vibration filter to converge quickly and generate a strong cancellation signal in a timely manner. When the regular vibration amplitude A... reg By obtaining a smaller μ1 within a few hours, a smaller steady-state error can be obtained while ensuring the filtering effect, thus avoiding unnecessary disturbance to the weak flame signal.

[0060] For the step-size convergence factor μ2, A irr A reflects the average energy of irregular vibrations. var This reflects the severity of its amplitude fluctuations, (A) irr +γ*A var The comprehensive interference intensity and risk index are formed. When the average energy of irregular vibrations is large or the instantaneous fluctuations are severe, the index value increases. In this case, if the step-size convergence factor μ2 also increases, the filter will attempt to track every random spike due to aggressive updates, which can easily lead to severe weight oscillations, algorithm misalignment, or even divergence, completely losing its filtering ability. Therefore, when the random interference is strong or changes rapidly, by making the step-size convergence factor μ2 inversely proportional to the comprehensive risk index, the filter update becomes more conservative and slower, and the weight changes are smoother, thus maintaining stability under strong random interference. Although the convergence speed is reduced, the main statistical components of noise can be reliably filtered out.

[0061] By employing differentiated regular and irregular vibration filters, parameter configurations can be tailored based on the differences in time-frequency characteristics, predictability, and filter requirements between regular and irregular vibrations, thereby improving the modeling accuracy and convergence performance of noise filtering. When a noise characteristic changes, such as a sudden change in fan speed leading to F... reg The changes will only affect the corresponding filters and their parameters, and will not cause a drastic impact on filters that process other types of noise, making the overall system more stable.

[0062] S302, at each discrete time point n, respectively from the regular vibration component V reg and irregular vibration component V irr Extract the first L1 or L2 samples to construct the reference input vector X: X reg (n)=[V reg (n), V reg (n-1), ..., V reg [(n-L1+1)]; X irr (n)=[V irr (n), V irr (n-1), ..., V irr (n-L2+1)].

[0063] The reference input vector X represents the local waveform segments of the two types of vibration noise in the current and recent period.

[0064] S303, perform a dot product between the weight vector of the adaptive filter and the reference input vector to calculate the current time-to-time response of the regular vibration component V. reg and irregular vibration component V irr Estimated value of vibration noise V est-reg and V est-irr : V est-reg =W1*X reg (n); V est-irr =W2*X irr (n); Vibration noise component V est =V est-reg +V est-irr .

[0065] By summing the instantaneous estimates of regular and irregular vibration noise, the optimal current estimate of the same-source vibration signal by the adaptive filter is obtained, which is the cancellation signal for noise filtering.

[0066] S304, Obtain the filtered main signal channel S main ', Subtract the vibration noise component V from the signal channel est The signal deviation is obtained, that is, the pure main signal channel S. main-clean =S main '-V est .

[0067] S305, based on the current moment's pure main signal channel S main-clean The tap weight vectors of the regular vibration filter and the irregular vibration filter are updated using the reference input vector X, respectively: W1(n+1)=W1(n)+[μ1 / (δ+||X reg (n)|| 2 ]*S main-clean *X reg (n); W2(n+1)=W2(n)+[μ2 / (δ+||X irr (n)|| 2 ]*S main-clean *X irr (n); Where δ is the regularization parameter, and is a minimum positive number that prevents division by zero.

[0068] Updating the tap weight vector of the adaptive filter is the core mechanism for the adaptive filter to learn, based on the feedback of the current filtering result (here, S). main-clean According to the NLMS algorithm rules, the tap weight vectors W1 and W2 are dynamically adjusted so that the filter can continuously approach the optimal weights and continuously improve the noise estimation accuracy in subsequent time steps.

[0069] S306, time index n increments, return to S302, repeat the entire process, at each time n, the vibration noise component V output from S303 is... est Used for filtering the main signal channel S main The vibration noise is filtered out, and the pure main signal channel S304 is obtained simultaneously. main-clean Output the data and update the tap weight vector.

[0070] The pure auxiliary signal channel S is obtained in the same way. aux-clean The output ensures that vibration noise from both signal channels is filtered out synchronously and uniformly, providing a consistent purity basis for subsequent flame judgment logic based on dual signal channels, so as to ensure that the pure main signal channel S... main-clean and pure auxiliary signal channel S aux-clean To make subsequent judgments on the flame detection.

[0071] It should be noted that the solution in this application employs a normalized least mean square filter. Those skilled in the art will understand that any adaptive filter capable of receiving vibration sampling signals or their processed signal reference input, as well as signals from the main signal channel and auxiliary signal channel, and dynamically adjusting its own parameters under the guidance of the characteristic parameter P to estimate noise, can be incorporated into the system framework of this application's technical solution.

[0072] S400, based on the pure main signal channel S main-clean and pure auxiliary signal channel S aux-clean The comparison calculation outputs a judgment signal and a warning signal for flame detection. Please refer to... Figure 4 Specifically, it includes the following steps.

[0073] S401, from the clean main signal channel S main-clean Extracting flame intensity features within the frequency band from the pure auxiliary signal channel S aux-clean Extract the corresponding background light intensity features.

[0074] S402, calculate the characteristic ratio R between the flame intensity feature and the background light intensity feature.

[0075] S403, compare the intensity of the feature ratio R with the first dynamic fire detection threshold to determine whether it exceeds the threshold.

[0076] S404, compare the duration for which the intensity of the characteristic ratio R exceeds the first dynamic fire detection threshold with the second dynamic fire detection threshold to determine whether the threshold is exceeded.

[0077] S405, a flame is determined to exist only when both the conditions of the first dynamic flame detection threshold and the second dynamic flame detection threshold are met.

[0078] This application employs a dual-threshold judgment scheme, applying both intensity and duration thresholds to the characteristic ratio R sequentially, to suppress transient interference. Based on obtaining a clean signal channel, the implemented flame detection judgment scheme is not the main technical improvement of this application and will not be elaborated upon here.

[0079] In another embodiment, in S400, a vibration dynamic weighting coefficient is obtained based on the vibration characteristic parameter P, and the values ​​of the first dynamic fire detection threshold and the second dynamic fire detection threshold are adjusted according to the vibration dynamic weighting coefficient.

[0080] By introducing a dynamic vibration weighting coefficient, when the vibration characteristic parameter P indicates strong vibration, the system automatically infers a high risk of residual noise and raises the first and second dynamic fire detection thresholds through the weighting coefficient. This makes the fire detection criteria more stringent, effectively suppressing false alarms triggered by residual noise and prioritizing system reliability. Conversely, when the vibration characteristic parameter P indicates weak vibration, the system determines that the purity of the signal after noise filtering is high and lowers the dual dynamic thresholds through the weighting coefficient. This makes the fire detection criteria relatively lenient, and the system is more sensitive to weak flame flicker signals, prioritizing system detection sensitivity. This scheme feeds back the vibration characteristic information to the final decision-making stage, forming a complete adaptive closed loop from noise perception to noise filtering to judgment and optimization. The dynamic vibration weighting coefficient can be obtained by calculating each vibration intensity component in the vibration characteristic parameter P at the current moment through a preset mapping function or lookup table.

[0081] It should also be noted that the embodiments in this application are described using a single auxiliary signal channel as an example. Those skilled in the art will understand that pyroelectric sensors using other filter wavelengths can be added to form multiple auxiliary signal channels. The flame detection signal is determined by comparing the main signal channel with multiple auxiliary signal channels. Adding other numbers and filter wavelengths of auxiliary signal channels does not affect the specific implementation of vibration component filtering for each signal channel in the embodiments of this application.

[0082] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0083] Please see Figure 5 The flame detection signal anti-interference processing system based on the same source vibration reference in this application includes a signal acquisition module 1, a filtering preprocessing module 2, an adaptive filtering module 3, a flame signal judgment module 4, and an early warning module 5.

[0084] The signal acquisition module 1 includes a main pyroelectric sensor 11, at least one auxiliary pyroelectric sensor 12, and a vibration sampling pyroelectric sensor 13 to acquire monitoring input signals.

[0085] The filtering preprocessing module 2 is connected to the signal acquisition module 1 to perform filtering preprocessing on the monitored input signal. It includes a low-pass filter submodule 21, a narrowband bandpass filter submodule 22, and a notch filter 23. The narrowband bandpass filter submodule is connected to the wind turbine generator control module 6 to acquire the wind turbine rotation signal.

[0086] The adaptive filtering module 3 is connected to the filtering preprocessing module 2 and includes a regular vibration filter 31 and an irregular vibration filter 32. It performs adaptive filtering on the filtering preprocessing signal and outputs a clean main signal channel and a clean auxiliary signal channel after vibration filtering.

[0087] The flame signal judgment module 4 judges the flame detection based on the characteristic ratio of the pure main signal channel and the pure auxiliary signal channel.

[0088] The early warning module 5 issues an early warning signal based on the judgment signal of flame detection.

[0089] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the flame detection signal anti-interference processing system based on the same source vibration reference described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0090] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0091] In the above embodiments, the methods can be implemented entirely or partially through software, hardware, firmware, or any combination thereof. When implemented in software, they can be implemented entirely or partially in the form of a computer program product. This program can be stored in a computer-readable storage medium, and when executed, it can perform the flow of the above-described embodiment of the flame detection signal anti-interference processing method based on homogeneous vibration reference. The aforementioned storage medium includes various media capable of storing program code, such as ROM or random access memory (RAM), magnetic disks, or optical disks. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function of the flame detection signal anti-interference processing method based on homogeneous vibration reference according to the embodiments of this application are generated.

[0092] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A flame detection signal anti-interference processing method based on homogeneous vibration reference, applied to the nacelle fire alarm monitoring of wind turbine generators, characterized in that... Includes the following steps: S100, synchronously acquires raw electrical signals from the main signal channel, at least one auxiliary signal channel, and one vibration sampling signal channel; wherein, the main signal channel, the auxiliary signal channel, and the vibration sampling signal channel use the same pyroelectric sensor, and the pyroelectric sensors of different signal channels input different filter signals; S200, filtering and preprocessing the main signal channel S main , the auxiliary signal channel S aux and the vibration sampling signal channel S ref to obtain the filtered main signal channel S main ', the filtered auxiliary signal channel S aux ' and the filtered vibration sampling signal channel S ref '; and ref extracting a vibration feature parameter P representing a vibration state by performing vibration feature extraction on the filtered vibration sampling signal channel S S300 dynamically adjusts the parameters of the adaptive filter based on the vibration characteristic parameter P, and then processes the filtered vibration sampling signal channel S through the adaptive filter. ref 'Conduct vibration noise component V' est The estimation will be applied to the filtered main signal channel S. main 'and auxiliary signal channel S aux 'Input an independent adaptive filter, through the noise component V est After filtering, the pure main signal channel S after vibration filtering is obtained. main-clean and pure auxiliary signal channel S aux-clean ; S400, based on the pure main signal channel S main-clean and pure auxiliary signal channel S aux-clean The comparison calculation outputs a judgment signal and a warning signal for flame detection.

2. The flame detection signal anti-interference processing method based on homogeneous vibration reference according to claim 1, characterized in that, In S100, the main signal channel uses a pyroelectric sensor with an input infrared band filtered signal, the auxiliary signal channel includes at least one pyroelectric sensor with an input ultraviolet band filtered signal, and the vibration sampling signal channel uses a fully enclosed, opaque pyroelectric sensor.

3. The flame detection signal anti-interference processing method based on homogeneous vibration reference according to claim 1, characterized in that, In S200, the main signal channel S is acquired. main Auxiliary signal channel S aux and vibration sampling signal channel S ref For signals within a preset time window, the specific steps include: S201, for the main signal channel S main Auxiliary signal channel S aux and vibration sampling signal channel S ref A low-pass filter is used to perform low-pass filtering, resulting in the filtered main signal channel S. main Auxiliary signal channel S aux 'and vibration sampling signal channel S ref '; S202: Acquire the fan speed information and convert it into frequency information. Based on the fan frequency information, configure a narrowband bandpass filter with adjustable center frequency, and filter the vibration sampling signal channel S. ref Input narrowband bandpass filter, output regular vibration component V reg ; S203, for the regular vibration component V reg Extracting the dominant frequency F of regular vibration reg and regular vibration amplitude A reg ; S204, the filtered vibration sampling signal channel S ref 'Through a frequency F that is dominated by regular vibrations reg A dynamic notch filter with a center frequency is used to filter out regular vibration components V. reg The irregular vibration component V was obtained. irr ; S205, for irregular vibration component V irr Extracting the root mean square of vibration intensity A irr Broadband vibration intensity variance A var ; The vibration characteristic parameter P is the dominant frequency F of the regular vibration. reg、 Regular vibration amplitude A reg Root mean square of vibration intensity A irr and broadband vibration intensity variance A var A set of.

4. The flame detection signal anti-interference processing method based on homogeneous vibration reference according to claim 3, characterized in that, Several narrowband bandpass filters are configured in parallel. The center frequency of one of the narrowband bandpass filters is configured based on the frequency information of the wind turbine, while the center frequencies of the other narrowband bandpass filters are configured based on the harmonic frequencies of the wind turbine frequency. The filtered vibration sampling signal channel S... ref After filtering by each parallel narrowband bandpass filter, the outputs of each narrowband bandpass filter are weighted and summed according to harmonic weights to obtain the regular vibration component V. reg .

5. The flame detection signal anti-interference processing method based on homogeneous vibration reference according to claim 3, characterized in that, In S300, the adaptive filter adopts a normalized minimum mean square filter, including a regular vibration filter and an irregular vibration filter, specifically including the following steps: S301, based on the regular vibration dominant frequency F reg The length L1 of the tap weight vector W1 of the regular vibration filter and the length L2 of the tap weight vector W2 of the irregular vibration filter are constrained. The tap weight vectors are initialized to zero vectors, based on the regular vibration amplitude A. reg Configure the step size convergence factor μ1 of the regular vibration filter based on the broadband vibration intensity root mean square A. irr and broadband vibration intensity variance A var Configure the step size convergence factor μ2 of the irregular vibration filter; S302, at each discrete time point n, respectively from the regular vibration component V reg and irregular vibration component V irr Extract the first L1 or L2 samples to construct the reference input vector X: X reg (n)=[V reg (n),V reg (n-1),...,V reg (n-L1+1)]; X irr (n)=[V irr (n), V irr (n-1),...,V irr (n-L2+1)]; S303, perform a dot product between the weight vector of the adaptive filter and the reference input vector to calculate the current time-to-time response of the regular vibration component V. reg and irregular vibration component V irr Estimated value of vibration noise V est-reg and V est-irr : V est-reg =W1*X reg (n); V est-irr =W2*X irr (n); Vibration noise component V est =V est-reg +V est-irr ; S304, Obtain the filtered main signal channel S main ', Subtract the vibration noise component V from the signal channel est The signal deviation is obtained, that is, the pure main signal channel S. main-clean =S main '-V est ; S305, based on the current moment's pure main signal channel S main-clean The tap weight vectors of the regular vibration filter and the irregular vibration filter are updated using the reference input vector X, respectively: W1(n+1)=W1(n)+[μ1 / (δ+||X reg (n)|| 2 ]*S main-clean *X reg (n); W2(n+1)=W2(n)+[μ2 / (δ+||X irr (n)|| 2 ]*S main-clean *X irr (n); Where δ is the regularization parameter, and is a minimum positive number to prevent division by zero; S306, time index n increments, return to S302, repeat the entire process, at each time n, the vibration noise component V output from S303 is... est Used for filtering the main signal channel S main The vibration noise is filtered out, and the pure main signal channel S304 is obtained simultaneously. main-clean Output the signal and update the tap weight vector; obtain the pure auxiliary signal channel S in the same way. aux-clean External output is achieved through the clean main signal channel S. main-clean and pure auxiliary signal channel S aux-clean To make subsequent judgments on the flame detection.

6. The anti-interference processing method for flame detection signals based on homogeneous vibration reference according to claim 1, characterized in that, Specifically, the S400 includes: S401, from the clean main signal channel S main-clean Extracting flame intensity features within the frequency band from the pure auxiliary signal channel S aux-clean Extract the corresponding background light intensity features; S402, calculate the characteristic ratio R between the flame intensity feature and the background light intensity feature; S403, compare the intensity of the feature ratio R with the first dynamic fire detection threshold to determine whether it exceeds the threshold; S404, compare the duration for which the intensity of the characteristic ratio R exceeds the first dynamic fire detection threshold with the second dynamic fire detection threshold to determine whether the threshold is exceeded; S405, a flame is determined to exist only when both the conditions of the first dynamic fire detection threshold and the second dynamic fire detection threshold are met.

7. The flame detection signal anti-interference processing method based on homogeneous vibration reference according to claim 6, characterized in that, In S400, a vibration dynamic weighting coefficient is obtained based on the vibration characteristic parameter P, and the values ​​of the first dynamic fire detection threshold and the second dynamic fire detection threshold are adjusted according to the vibration dynamic weighting coefficient.

8. A flame detection signal anti-interference processing system based on homogeneous vibration reference, employing the anti-interference processing method according to any one of claims 1 to 7, characterized in that, It includes a signal acquisition module, a filtering and preprocessing module, an adaptive filtering module, a flame signal judgment module, and an early warning module; The signal acquisition module includes a main pyroelectric sensor, at least one auxiliary pyroelectric sensor, and a vibration sampling pyroelectric sensor to acquire monitoring input signals. The filtering preprocessing module is connected to the signal acquisition module and performs filtering preprocessing on the monitored input signal. It includes a low-pass filter submodule and a narrowband bandpass filter submodule. The narrowband bandpass filter submodule is connected to the control module of the wind turbine. The adaptive filtering module is connected to the filtering preprocessing module, performs adaptive filtering on the filtering preprocessing signal, and outputs a pure main signal channel and a pure auxiliary signal channel after vibration filtering. The flame signal determination module determines the flame detection based on the characteristic ratio of the pure main signal channel and the pure auxiliary signal channel; The early warning module issues an early warning signal based on the judgment signal from flame detection.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the anti-interference processing method for flame detection signals based on the same source vibration reference as described in any one of claims 1 to 7.

10. A computer program product, characterized in that, The computer program product includes a computer program or instructions that enable the computer program or instructions to perform the steps in the flame detection signal anti-interference processing method based on the same source vibration reference as described in any one of claims 1 to 7.

Images