Infrared flame identification and false alarm suppression method and system based on three detectors

Abstract

The invention relates to the technical field of fire detection, in particular to an infrared flame recognition and false alarm suppression method and system based on three detectors, and the system integrates a fixed infrared detector, a movable infrared detector and a micro-amplitude vibration excitation source. After a suspicious signal is triggered, active diagnosis is started, fusion analysis of space contour and excitation coupling characteristics is carried out on a dynamic signal collected by a detector under the background of scanning and vibration excitation, and deterministic judgment is carried out according to the fusion analysis so as to identify real flames and suppress interference. The invention can solve or at least alleviate the problem of static space and dynamic spectrum fuzziness that the sensitivity and the false alarm rate are difficult to balance in the existing infrared flame detection, and provides the infrared flame identification and false alarm suppression method and system based on the three detectors, which can eliminate the fuzziness without sacrificing the detection sensitivity and improve the detection accuracy. And accurate identification and effective suppression of a high-confusion interference source are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of fire detection technology, and in particular to an infrared flame identification and false alarm suppression method and system based on three detectors. BACKGROUND

[0002] With the rapid development of petrochemical industry, power transmission, large-scale storage, transportation hub and other industries, higher requirements are put forward for the safety and timeliness of early fire warning, and the performance of infrared flame detector as the core component of early fire warning system is directly related to the safety of personnel life and property. With the progress of science and technology, in order to improve the system robustness of early fire warning system in the variable electromagnetic environment, thermal radiation environment and meteorological environment, the main design idea is to use multiple detectors for redundant configuration and mutual judgment. For example, three detectors or three waveband, multi-waveband infrared sensors are used, through the mutual correlation between different spectral signals, some non-flame radiation signals can be eliminated or reduced to affect the early fire warning system, compared with single detector, the reliability of system detection is significantly improved.

[0003] Under this technical background, the present application provides many flame identification methods based on multi-channel infrared signal processing. For example, see Chinese patent CN106408843B discloses a three waveband infrared flame detector, the technical points of which are to set three infrared sensors sensitive to flame characteristic waveband, artificial interference source waveband and background radiation waveband respectively, to compare and analyze the signal intensity of three channels by microprocessor, and to perform frequency spectrum analysis on the signal flicker frequency, so as to judge whether it is flame. In the design of the above scheme, the starting point is to use the information redundancy of multi-spectral dimension to distinguish the target radiation source from other general background heat sources, and to exclude the common frequency high temperature heat source by using the low frequency flicker (1-20Hz) characteristics of flame. In a specific historical period and application scene, the above analysis paradigm based on spectral + time domain dual characteristics solves the general false alarm problem of background thermal radiation in the initial detection technology, and represents an important direction of the then technology development.

[0004] However, with the continuous development of related technologies and the continuous improvement of performance requirements, the above technical solutions have some inherent characteristics in principle, which gradually expose their inherent defects when facing some new and high-level interference sources. The reason for this phenomenon is that the above technical paradigm is a passive analysis method based on the macro statistical characteristics of point source signals, and the two indeterminate zones that cannot be eliminated by macro analysis will cause some confusion that should not occur. The first fuzzy zone is static spatial fuzziness. No matter what spectral configuration is used, the signal obtained by the traditional detector is actually the simple addition of all infrared radiation energy within its field of view, and it cannot perceive the spatial form of the radiation source. Therefore, if the traditional detector faces a static high-temperature interference source with a larger size and lower temperature (such as a high-temperature steam pipe far away from the detector, a larger area of the heating furnace wall of the heating furnace, etc. line source or surface source), the total radiation flux obtained on the sensor will have a very close signal level to that emitted by a real flame source (which can be regarded as a point source) with a smaller size and a higher temperature on the sensor. Only relying on signal strength for logical confusion judgment is easy to confuse, leading to misjudgment of large-area industrial heat source as fire, resulting in false alarm. The second fuzzy zone is dynamic spectral fuzziness. The traditional scheme relies on spectral analysis tools such as fast Fourier transform to determine whether the signal contains a flicker frequency component of 1-20Hz, but cannot determine the physical cause of the frequency signal. The flicker of real flame is due to the instability of the combustion process, i.e. turbulence, which is essentially a random process with continuous distributed energy in the characteristic frequency band; another common dynamic interference source, such as a rotating mechanical warning light, a periodically on-off heating device or an electric welding arc, its flicker is caused by deterministic physical motion or electrical modulation, its spectral characteristics are discrete, periodic signals with stable fundamental frequency and harmonic frequency. Different signals analyzed by spectral analysis method will have obvious energy peaks in the characteristic frequency band, causing similar spectral characteristics of various signals, making it difficult for the system to distinguish between random flicker of flame and periodic motion of equipment, causing another major false alarm.

[0005] The two types of judgment ambiguity intersect with each other, resulting in the prior art falling into a vicious cycle: in order to reduce the false alarm threshold of the static high-temperature surface source, it is often necessary to increase the alarm threshold, which will lead to the missed alarm of the early small fire source and increase the risk of missed alarm; in order to avoid the dynamic periodic interference source, a narrower band-pass filter or a more complex algorithm may be used, and at the same time, the real signal of some flames may be filtered out due to excessive suppression. The performance indicators of the above are this and that, which is rooted in the inherent limitation of passive sensing method for information dimension, that is, passive participation in detection, which cannot obtain deep information to reveal the physical nature of the target. Therefore, a new flame detection method is designed to break through the limitations of passive signal analysis method, fundamentally diagnose and distinguish those interference sources with different physical natures but similar apparent characteristics, that is, to solve the problems of static spatial ambiguity and dynamic spectral ambiguity, which is the technical bottleneck and technical problem to be solved for those skilled in the art. SUMMARY

[0006] The technical problem to be solved by the present application is that the existing infrared flame detection technology based on multiple detectors essentially relies on passive, macroscopic statistical characteristic analysis of the target radiation source. This analysis paradigm cannot actively obtain deep information to reveal the physical nature of the target, so when facing interference sources with similar apparent radiation characteristics but different physical causes, there are problems of static spatial ambiguity and dynamic spectral ambiguity. Static spatial ambiguity is that the system cannot effectively distinguish between small-size, high-temperature real flame point sources and large-area, medium-low-temperature industrial heat surface sources; dynamic spectral ambiguity is that the system cannot fundamentally distinguish between random turbulent flicker of flames and periodic signal flicker generated by mechanical and electrical equipment. The two ambiguities superimpose on each other, resulting in the existing technology being difficult to balance between reducing false alarm rate and maintaining high detection sensitivity.

[0007] In order to achieve the above application purpose, the present application provides an infrared flame recognition and false alarm suppression method and system based on three detectors, which integrates active spatial scanning and active physical excitation two diagnosis mechanisms, and performs fusion analysis and collaborative decision on the multi-dimensional data obtained in the diagnosis process, aiming to fundamentally eliminate the above static spatial ambiguity and dynamic spectral ambiguity, so as to realize accurate recognition and effective suppression of high-deception interference sources without sacrificing detection sensitivity.

[0008] In order to achieve the above object, the present application provides the following technical scheme: An infrared flame identification and false alarm suppression method based on three detectors, a sensing module including three infrared detectors, one of which is a verification detector, comprising the following steps: S1, initial monitoring and triggering, passive monitoring of the monitoring area by at least one infrared detector, when the monitoring signal meets the preset composite trigger condition, determining that a suspicious flame signal is monitored, and generating a diagnostic trigger instruction; S2, composite diagnosis, responding to the diagnostic trigger instruction, starting an active diagnosis program, which includes: a) synchronous activation mechanism: driving a micro-amplitude vibration excitation source to apply a preset diagnostic frequency mechanical vibration to the sensing module, and driving a micro-rotary actuator to drive the verification detector to perform spatial scanning; b) data acquisition and analysis: during the synchronization of scanning and vibration, the output signal of the verification detector is collected, and the spatial profile features representing the spatial distribution form of the radiation source and the excitation coupling features representing the response characteristics of the radiation source to the mechanical vibration are extracted from the signal; c) fusion decision: according to the combination result of the spatial profile features and the excitation coupling features, the preset fusion decision logic is executed to make a final judgment on the nature of the signal source.

[0009] In order to further realize the present application, the following technical scheme can be preferred: Preferably, in the initial monitoring and triggering step, the composite trigger condition includes an amplitude determination condition and a flicker frequency determination condition; The amplitude determination condition is that among the three initial infrared radiation signals output by the three infrared detectors, the intensity values of at least two signals simultaneously exceed the preset first signal amplitude threshold value; The flicker frequency determination condition is that the frequency spectrum energy peak values of at least two signals meeting the amplitude determination condition obtained by a spectrum analysis algorithm are all within a preset flame characteristic frequency interval.

[0010] Preferably, in the synchronous activation mechanism, the micro-amplitude vibration excitation source is driven to work; and the micro-rotary actuator is driven to perform the spatial scanning with the azimuth angle of the suspicious flame signal as the center.

[0011] Preferably, the preset diagnostic frequency of the mechanical vibration applied to the sensing module is higher than the frequency of the flame characteristic frequency interval; and the spatial scanning of the verification detector is reciprocating scanning within a preset angle range at a preset angular velocity.

[0012] Preferably, in the data acquisition and analysis, the process of extracting the spatial profile feature comprises: correlating the acquired output signal of the verification detector with the simultaneously recorded real-time angular position data of the detector to construct a spatial intensity distribution profile representing the variation of infrared radiation intensity with scanning angle; using a mathematical fitting algorithm to model analyze the spatial intensity distribution profile, and according to the result of the modeling analysis, determining the spatial profile feature of the radiation source as a point source feature or a surface or line source feature.

[0013] Preferably, the modeling analysis of the spatial intensity distribution profile specifically comprises: fitting the data of the spatial intensity distribution profile with a standard Gaussian function model and a standard trapezoidal function model respectively, and calculating the fitting goodness of each; if the fitting goodness of the spatial intensity distribution profile with the Gaussian function model is higher than a preset first fitting threshold, and the standard deviation parameter representing the spatial scale of the radiation source calculated through the fitting is less than a preset point source scale threshold, then the spatial profile feature is determined as a point source feature; otherwise, if the fitting goodness of the spatial intensity distribution profile with the trapezoidal function model is higher, or the standard deviation parameter of the Gaussian fitting is greater than the point source scale threshold, then the spatial profile feature is determined as a surface or line source feature.

[0014] Preferably, in the data acquisition and analysis, the process of extracting the excitation coupling feature comprises: cutting a data segment corresponding to the peak region of the spatial intensity distribution profile from the output signal of the verification detector; performing high-resolution fast Fourier transform on the data segment to generate a detailed amplitude spectrum graph; detecting whether there is a significant energy peak at the preset diagnostic frequency point and its harmonic frequency point in the amplitude spectrum graph, and determining the presence of a significant excitation coupling feature or no excitation coupling feature accordingly.

[0015] Preferably, to realize the quantitative judgment of the energy peak, the method calculates an excitation coupling degree index by the following way: obtaining the spectral amplitudes at the preset diagnostic frequency point and its second harmonic frequency point in the amplitude spectrum graph; summing the two spectral amplitudes to obtain an excitation signal energy representation value; obtaining the average noise amplitude in a frequency band near the diagnostic frequency point and not containing the known signal as a reference noise level; calculating the ratio between the excitation signal energy representation value and the reference noise level to obtain the excitation coupling degree index; if the calculated excitation coupling degree index is greater than a preset coupling determination threshold, it is determined that there is a significant excitation coupling feature; otherwise, it is determined as no excitation coupling feature.

[0016] Preferably, the fusion decision logic includes the following rules: Decision logic one: when the spatial profile feature is determined as a point source feature, and the excitation coupling feature is determined as no excitation coupling feature, it is confirmed as a real flame; Decision logic two: when the spatial profile feature is determined as a surface or line source feature, it is confirmed as a static high temperature interference source regardless of the determination result of the excitation coupling feature; Decision logic three: when the spatial profile feature is determined as a point source feature, and the excitation coupling feature is determined as a significant excitation coupling feature, it is confirmed as a dynamic periodic interference source.

[0017] An infrared flame recognition and false alarm suppression system based on three detectors, the system comprising: a sensing module comprising a fixed first infrared detector, a fixed second infrared detector, and a verification detector; a micro rotary actuator mechanically connected with the verification detector and used to drive the verification detector to perform a precise angular scanning motion; a micro-vibration excitation source rigidly connected with the sensing module and used to apply a micro-mechanical vibration of a preset diagnostic frequency to the entire sensing module; a signal conditioning and acquisition unit electrically connected with the first infrared detector, the second infrared detector, and the verification detector, and used to condition and digitally acquire the signals output by each detector; a central processing and control unit in data and control connection with the signal conditioning and acquisition unit, the micro rotary actuator, and the micro-vibration excitation source; wherein the central processing and control unit stores instruction codes that, when executed, configure the central processing and control unit to perform the method of any one of claims 1 to 9, specifically: in the initial monitoring phase, processing the signals acquired by the signal conditioning and acquisition unit to determine whether the composite trigger condition is met; when the condition is met, entering the active diagnosis phase, and controlling the micro rotary actuator to perform scanning and the micro-vibration excitation source to apply vibration in parallel; during diagnosis, processing the dynamic signals from the verification detector to extract the spatial profile feature and the excitation coupling feature in parallel; finally, based on the combination of the two features, performing fusion decision logic to output the final judgment conclusion.

[0018] The beneficial effects of the present application are: The present application aims to fundamentally eliminate the static spatial ambiguity and dynamic spectrum ambiguity by integrating the active space scanning and active physical excitation two diagnostic mechanisms, and performing fusion analysis and collaborative decision on the multi-dimensional data obtained in the diagnostic process, can effectively distinguish the real flame and high-deception interference source with similar apparent radiation characteristics but completely different physical nature, realizes the accurate identification and effective suppression of the high-deception interference source without sacrificing the detection sensitivity, greatly reduces the false alarm rate. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 is the method flow chart of the present application.

[0020] Figure 2 is the system structure block diagram of the present application.

[0021] Figure 3 is the space intensity distribution profile schematic diagram for space profile feature extraction in the present application.

[0022] Figure 4 is the signal spectrum analysis schematic diagram for excitation coupling feature extraction in the present application. DETAILED DESCRIPTION

[0023] In the description of the present application, it also needs to be explained that, unless otherwise explicitly specified and limited, the terms "arrangement", "installation", "connection", "connection" should be understood in a broad sense, for example, it can be fixedly connected, or it can be detachably connected, or integrally connected, it can be mechanically connected, or it can be electrically connected, it can be directly connected, or indirectly connected through an intermediate medium, it can be the communication inside two elements. For ordinary skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely in the following combined with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0025] Embodiment one The present embodiment provides an infrared flame identification and false alarm suppression method based on three detectors, which is executed in a detection system comprising two fixed first infrared detectors and second infrared detectors, a movable verification detector mounted on a micro rotary actuator, a micro vibration excitation source capable of applying micro mechanical vibration to the three detectors as a whole, and a central processing and control unit, as shown in Figure 1 The method comprises the following steps: Step one, initial monitoring and triggering. In this stage, the central processing and control unit controls the micro-rotary actuator and the micro-amplitude vibration excitation source to be in a static state, and synchronously and continuously collects the initial infrared radiation signals output by the first infrared detector, the second infrared detector and the verification detector. The central processing and control unit analyzes the collected three-way initial infrared radiation signals in real time, which includes the dual determination of signal amplitude and signal flicker frequency. Specifically, the central processing and control unit sets a first signal amplitude threshold and a flame characteristic frequency interval. When the intensity values of at least two signals among the three signals simultaneously and continuously exceed the first signal amplitude threshold, and the frequency spectrum energy peaks of the two signals obtained by fast Fourier transform or other frequency spectrum analysis algorithm fall within the preset flame characteristic frequency interval, the central processing and control unit determines that a suspicious flame signal is monitored, and immediately terminates the initial monitoring stage and automatically enters step two of compound diagnosis. The flame characteristic frequency interval is determined to be 1 Hz to 20 Hz.

[0026] Step two, compound diagnosis. In this stage, the central processing and control unit immediately starts the compound active diagnosis program based on the triggering result of step one, which includes three tightly coupled sub-steps of synchronous activation mechanism, data acquisition and analysis and fusion decision.

[0027] As the first sub-step of compound diagnosis, the central processing and control unit performs synchronous activation of the diagnosis mechanism. This activation operation is two actions performed in parallel: one is that the central processing and control unit outputs a preset, high-frequency, continuous sine wave or square wave excitation signal to the driving circuit of the micro-amplitude vibration excitation source, drives the micro-amplitude vibration excitation source to start working, and thus applies a known, stable, small amplitude mechanical vibration to the entire sensing module containing the three detectors. The frequency of the mechanical vibration is set to a specific diagnosis frequency much higher than the flame characteristic frequency interval, and the value is 200 Hz; the physical displacement amplitude of the vibration is controlled in the range of 10 microns to 50 microns. The second is that the central processing and control unit outputs a series of precise step control pulses or a continuous angle control signal to the driving circuit of the micro-rotary actuator, drives the verification detector to perform a small-angle precision scanning motion according to the preset scanning trajectory with the suspicious signal source as the center. The scanning trajectory is a reciprocating arc scanning from -5 degrees to +5 degrees of the central azimuth angle, and the angular velocity is set to 2 degrees per second.

[0028] As the second sub-step of the composite diagnosis, the central processing and control unit performs the spatially resolved data acquisition of the excitation response. During the entire process of verifying the scanning motion of the detector, and in the background of the entire sensing module continuously subjected to high-frequency micro-amplitude vibration excitation, the central processing and control unit specially and continuously acquires and stores the dynamic output signals from the verifying detector at a sampling rate much higher than the diagnostic frequency. The signals obtained in this process constitute a composite time series data stream carrying spatial distribution information of the radiation source and physical excitation response information. The sampling rate is set to be no less than 2000 Hz to ensure that any signal modulation effects caused by the 200 Hz diagnostic frequency can be captured without distortion.

[0029] The central processing and control unit performs integrated multi-dimensional data fusion analysis on the acquired composite time series data stream. This analysis process is divided into two parallel processing calculation tasks: spatial profile feature extraction and excitation coupling feature extraction.

[0030] Further, as shown in Figure 3 , in the process of spatial profile feature extraction, the central processing and control unit first associates the acquired composite time series data stream with the synchronously recorded real-time angular position data of the verifying detector, thereby constructing a discrete data set representing the change of infrared radiation intensity with scanning angle, i.e., a spatial intensity distribution profile. Then, the central processing and control unit uses a mathematical fitting algorithm to fit the shape of the spatial intensity distribution profile to the shape of a standard Gaussian function model and a standard trapezoidal function model, respectively, and calculates the goodness of fit for each. The parameter σ of the standard Gaussian function model is a discriminant parameter for determining the spatial scale of the radiation source. If the goodness of fit of the spatial intensity distribution profile to the standard Gaussian function model is greater than a first fitting threshold, and the calculated parameter σ of the Gaussian function model is less than a point source scale threshold, the central processing and control unit determines that the spatial profile feature of the radiation source is a point source feature; if the goodness of fit of the spatial intensity distribution profile to the standard trapezoidal function model is greater than the first fitting threshold, or the goodness of fit of the spatial intensity distribution profile to the standard Gaussian function model is greater than the first fitting threshold and the calculated parameter σ of the Gaussian function model is greater than the point source scale threshold, the central processing and control unit determines that the spatial profile feature of the radiation source is a surface source or a line source feature.

[0031] Further, as shown in Figure 4As shown, the central processing and control unit extracts the spatial intensity profile data segment containing the peak of the spatial intensity profile profile from the composite time series data stream in the above excitation coupling feature extraction step, and performs high-resolution fast Fourier transform on the data segment to generate the amplitude spectrum of the spatial intensity profile profile. The central processing and control unit detects whether there is a significant energy peak at the diagnostic frequency point (200 Hz) and its second harmonic frequency point in the above amplitude spectrum. The central processing and control unit quantitatively determines the method as follows: The central processing and control unit determines an excitation coupling degree index C, C = [A(f_d) + A(2*f_d)] / A_n, where A(f_d) is the spectral amplitude of the diagnostic frequency point; A(2*f_d) is the spectral amplitude of the second harmonic of the diagnostic frequency point; A_n is the average noise amplitude in a frequency band near the diagnostic frequency point that does not contain any known signals. If the above excitation coupling degree index C is greater than the preset coupling determination threshold, the central processing and control unit determines that there is a significant excitation coupling feature; if the above excitation coupling degree index C is less than the preset coupling determination threshold, the central processing and control unit determines that there is no excitation coupling feature.

[0032] As the third sub-step of the composite diagnosis, the central processing and control unit implements a deterministic fusion decision logic based on the fusion analysis results of the above spatial profile feature and excitation coupling feature, and outputs the final decision result. The fusion decision logic is as follows: Decision logic one: if the spatial profile feature is a point source feature and the excitation coupling feature is no excitation coupling feature, it is confirmed to be a real flame, and a fire alarm signal of the highest priority is issued. Only the combination of the above two features can uniquely point to a real flame source without entity, whose radiation is not affected by the slight physical vibration of the detector, i.e., the above diagnostic result points to a real flame source without entity, whose radiation mode is not affected by the slight physical vibration of the detector.

[0033] Decision logic two: if the spatial profile feature is determined to be a surface source or a line source feature, it is confirmed to be a static high-temperature interference source regardless of the analysis result of the excitation coupling feature, and any alarm signal output is actively suppressed. This feature points to large-area industrial heat sources such as high-temperature steam pipes and heating furnace walls, which fundamentally do not have the point source form of flame.

[0034] Decision logic three: if the spatial profile feature is determined to be a point source feature, but the excitation coupling feature is determined to be a significant excitation coupling feature, it is confirmed to be a dynamic periodic interference source, and the alarm is also actively suppressed. This combination of features accurately points to physical interference sources with entity structure, whose radiation field will be intensity-modulated due to the slight angle jitter of the detector, such as rotating warning lights using mirrors or lenses, and periodic on-off electric arc welding.

[0035] The fourth judgment logic: if the signal intensity that triggers the diagnosis decreases to below the first signal amplitude threshold during the scanning process of the verification detector, the system determines that the transient noise or interference has disappeared, immediately terminates the current diagnosis process, resets all actuators to the initial state, and returns to the initial monitoring mode of step one.

[0036] Embodiment two The embodiment provides a three-detector-based infrared flame identification and false alarm suppression system, which realizes all the technical solutions of the above method, as shown in the figure, the system comprises a sensing module, a micro-rotary actuator, a micro-amplitude vibration excitation source, a signal conditioning and acquisition unit, and a central processing and control unit. Figure 2

[0037] The sensing module is packaged in a metal shell with high dimensional stability and thermal stability, and the shell material is 6061-T6 aluminum alloy. The sensing module internally integrates three infrared detectors, namely a first infrared detector, a second infrared detector, and a verification detector. The first infrared detector and the second infrared detector are firmly fixed on the internal base of the shell, and their optical axes are parallel to each other. The first infrared detector is an indium antimonide infrared detector, whose spectral response range covers 3-5 microns, and the peak response wavelength is located at 4.3 microns, which is used to match the characteristic emission peak of carbon dioxide generated by flame combustion. The second infrared detector is a lead sulfide infrared detector, whose spectral response range is 1-3 microns, which is used as a reference channel for background radiation and interference monitoring. The verification detector is exactly the same as the first infrared detector in terms of model and specification, so as to ensure the comparability of the signals in the diagnosis process.

[0038] The micro-rotary actuator is a high-precision stepping motor, whose stepping angle precision is not less than 0.1 degree. The output shaft of the micro-rotary actuator is connected to a specially designed bracket through a precision coupling, and the verification detector is firmly installed on the bracket. The micro-rotary actuator is fixed inside the shell of the sensing module, and its rotation axis is perpendicular to the optical axis of the verification detector, so as to drive the verification detector to perform accurate angle scanning in the vertical plane of the optical axis.

[0039] The micro-amplitude vibration excitation source is a stacked piezoelectric ceramic actuator, which is rigidly installed between the bottom of the sensing module shell and the main installation base plate of the entire detection system. When a specific alternating voltage is applied to the piezoelectric ceramic actuator, it can generate high-frequency, micro-amplitude reciprocating displacement perpendicular to the optical axis of the detector, thereby driving the entire sensing module to vibrate synchronously. The working frequency of the piezoelectric ceramic actuator is precisely controlled by the central processing and control unit, and the vibration amplitude is adjusted by the amplitude of the excitation voltage.

[0040] ​The signal conditioning and acquisition unit includes three independent signal conditioning channels corresponding to the three detectors respectively and a multi-channel synchronous analog-digital converter. Each signal conditioning channel includes a low-noise trans-impedance amplifier for converting the weak current signal output by the detector into a voltage signal and performing preliminary amplification. The sampling resolution of the multi-channel synchronous analog-digital converter is not less than 24 bits, and the highest synchronous sampling rate is not less than 5000 Hz, ensuring high-fidelity and phase-difference-free data acquisition of the three detector signals.

[0041] The central processing and control unit is the core of the whole system, which is realized as an embedded computing platform integrated with a 32-bit or higher microprocessor, a hardware floating-point operation unit, a non-volatile flash memory with sufficient capacity and a random access memory. The non-volatile flash memory stores all instruction codes for implementing the method of the application, including initial monitoring algorithms, active diagnosis programs, spatial profile feature extraction algorithms, excitation coupling feature extraction algorithms and final fusion decision logic library. The data input end of the central processing and control unit is electrically connected with the output end of the signal conditioning and acquisition unit to receive the digitized three-channel infrared signal data stream. The control output end of the central processing and control unit is connected with the electrodes of the micro rotary actuator and the micro amplitude vibration excitation source through driving circuits to accurately send motion control instructions and vibration excitation signals. In addition, the central processing and control unit is also provided with an alarm signal output interface connected with an external fire alarm system or monitoring center to output standardized alarm level signals or communication protocol data packets when the decision result is a real flame.

[0042] In the workflow of the system, the coordination between components is precisely defined. In the initial monitoring phase, the central processing and control unit only processes data from the signal conditioning and acquisition unit and makes judgments according to the built-in trigger logic. Once the trigger condition is met, the central processing and control unit immediately enters the active diagnosis state, and the internal timer and controller start to work synchronously. On the one hand, the 200 Hz sine wave data stream is generated by the digital signal generator module and applied to the piezoelectric ceramic actuator through the digital-to-analog converter and power amplifier. On the other hand, the preset pulse sequence is sent to the stepper motor driver through the motion control module. During this period, the data acquisition module focuses on the channel of the verification detector, applies the internal angular position counter timestamp to the high-speed collected data stream, aligns each sample and stores it in a dedicated data buffer area in the random access memory. After the diagnosis is completed, the microprocessor of the central processing and control unit calls the mathematical calculation library and signal processing library stored in the flash memory to perform the Gaussian trapezoidal function fitting and high-resolution FFT calculation, and compares the results (σ value and coupling degree C value) with the threshold value stored in the flash memory. Then, through the lookup table or conditional judgment structure, the decision-making logic is executed and it is determined whether to change the output state through the alarm signal output interface. The total time required for the entire diagnosis process, from triggering to outputting the final decision result, is limited to within 2 seconds, ensuring a rapid response to fire.

[0043] Example Three In order to better enable those skilled in the relevant art to fully understand and implement the present application, the specific implementation principles of the present application are further supplemented below in conjunction with a specific application scenario The above system is deployed in a large industrial plant for actual scene testing. The test target is a standard heptane fire source with a diameter of 15 cm, placed 10 meters away from the detection system.

[0044] After the system is powered on, it enters the initial monitoring phase of step one. The radiation generated by the burning of the heptane fire is captured by the first infrared detector and the verification detector, and the signal strength after conditioning far exceeds the preset first signal amplitude threshold. At the same time, the central processing and control unit performs STFT analysis on the signal, and the energy peak of the flicker frequency is located at about 4.5 Hz, falling within the flame characteristic frequency interval of 1-20 Hz. Both the amplitude and frequency conditions are met, and the system determines it as a suspicious flame signal and immediately enters step two.

[0045] In the diagnosis stage, the micro-vibration excitation source is started at a frequency of 200 Hz, while the verification detector starts to perform a scan from -5 degrees to +5 degrees. After the data acquisition is completed, the central processing and control unit performs analysis in parallel. For spatial profile feature extraction, the algorithm fits the scan data to a Gaussian function model and a trapezoidal function model. The results show that the goodness of fit R² with the Gaussian function model reaches 0.98, which is significantly better than the trapezoidal model, and the calculated Gaussian standard deviation σ is 0.43 degrees, which is less than the preset 0.5-degree point source size threshold. Therefore, the spatial profile feature is determined as a point source feature.

[0046] For excitation coupling feature extraction, the algorithm performs a 4096-point FFT analysis on the original data in the signal peak area. On the generated frequency spectrum, the spectral amplitude at the 200 Hz diagnosis frequency point and its 400 Hz harmonic frequency point is 1.2 nanovolts / √Hz and 1.1 nanovolts / √Hz, respectively, which has no significant difference compared to the average noise floor amplitude of 1.15 nanovolts / √Hz in the nearby 220-250 Hz frequency band. Calculate the excitation coupling index C = (1.2 + 1.1) / 1.15 ≈ 2.0. This value is much smaller than the preset coupling determination threshold of 10.0. Therefore, the excitation coupling feature is determined as a non-excitation coupling feature.

[0047] Finally, the central processing and control unit performs fusion decision logic. The two input features are point source feature and non-excitation coupling feature, which completely match the decision logic one. The system finally makes a decision to confirm the real flame, and sends out a fire alarm signal through the alarm signal output interface.

[0048] Through the above embodiment, it can be clearly seen that the infrared flame recognition and false alarm suppression method and system based on three detectors proposed by the present application can effectively distinguish real flames and high-deception interference sources with similar apparent radiation characteristics but completely different physical natures by introducing two diagnostic mechanisms of active spatial scanning and active physical excitation, and performing fusion analysis and collaborative decision on the obtained multi-dimensional data, thereby greatly reducing the false alarm rate while maintaining high detection sensitivity, and solving the key technical bottleneck existing in the prior art.

[0049] Embodiment Four Before the start of the composite diagnosis step, the method further includes a diagnostic parameter adaptive adjustment step, specifically: the central processing and control unit first analyzes the preliminary features of the suspicious flame signal obtained in the initial monitoring and triggering step, the preliminary features at least including its main flicker frequency or signal intensity; then, the central processing and control unit dynamically selects or adjusts at least one diagnostic parameter for the composite diagnosis step from a preset parameter library according to the analysis result of the preliminary features, the diagnostic parameters including the preset diagnosis frequency or the angular velocity of spatial scanning; thereby realizing adaptive optimization of the diagnosis process for different types of suspicious signals.

[0050] Introducing the ability of intelligent self-adaptation. The technical solution of embodiment one uses fixed diagnostic parameters, while the solution of this embodiment enables it to tailor the treatment to the symptoms. For example, if the preliminary judgment is that the flame flickering frequency is high, the system can automatically select a higher or more appropriate excitation frequency to achieve the best decoupling effect. This solves a deeper problem: how to make the diagnostic process itself more efficient and more accurate.

[0051] When the final judgment conclusion of the fusion decision logic is to confirm as a real flame, the method further includes a flame precise positioning step, specifically: the central processing and control unit extracts the spatial intensity distribution profile constructed in the execution of the multi-dimensional fusion analysis step; determines the real-time angular position data corresponding to the point where the signal intensity reaches the maximum value from the profile; and outputs or records the angular position data as the accurate orientation information of the flame source.

[0052] Based on the judgment of true and false flame in the solution of embodiment one, the function of precise positioning is added. It makes use of the spatial intensity distribution profile already generated in the diagnostic process, without increasing additional cost to realize a new and highly valuable function. From recognition to recognition + positioning, not only can it alarm, but also can provide accurate guidance for the subsequent automatic fire extinguishing system.

[0053] The data acquisition and analysis in the composite diagnosis step further includes an active background noise suppression process, specifically: after the synchronous activation mechanism is started and before the output signal used for analysis is formally collected, the central processing and control unit first controls the verification detector to perform a "silent" spatial scan, that is, to complete a scan under the condition that the micro-vibration excitation source remains inactive, to collect the background vibration noise spectrum of the target area; then, when performing frequency spectrum analysis on the output signal superimposed with the active excitation vibration to extract the excitation coupling characteristics, the influence of the previously collected background vibration noise spectrum is subtracted or excluded in other algorithmic forms from the frequency spectrum of the signal; thereby improving the accuracy and robustness of the excitation coupling characteristic detection under strong environmental vibration interference.

[0054] Solves a potential weakness of the solution of embodiment one: how to deal with strong vibration interference that exists in the environment itself. By adding a silent scan to actively learn and eliminate background noise, the anti-interference ability and reliability of the entire system in harsh working conditions are greatly improved.

[0055] The above is only the preferred specific implementation of the present application, but the protection scope of the present application is not limited thereto, any skilled person in the art can make equivalent replacement or change according to the technical solution and the inventive concept of the present application within the technical range disclosed by the present application, which should be covered within the protection scope of the present application.

Claims

1. A three-detector based infrared flame recognition and false alarm suppression method, the sensor module comprises three infrared detectors, one of which is a verification detector, characterized in that, The method comprises the following steps: S1, initial monitoring and triggering, passive monitoring of a monitoring area by at least one infrared detector, when the monitoring signal meets the preset composite trigger condition, determining that a suspicious flame signal is monitored, and generating a diagnostic trigger instruction; S2, composite diagnosis, in response to the diagnostic trigger instruction, starting an active diagnosis program, which includes: a) a synchronous activation mechanism: driving a micro-amplitude vibration excitation source to apply a preset diagnostic frequency mechanical vibration to the sensing module, and driving a micro-rotary actuator to drive the verification detector to perform spatial scanning; b) data acquisition and analysis: during the synchronization of scanning and vibration, the output signal of the verification detector is collected, and the spatial profile features representing the spatial distribution form of the radiation source and the excitation coupling features representing the response characteristics of the radiation source to the mechanical vibration are extracted from the signal; c) fusion decision: according to the combination result of the spatial profile features and the excitation coupling features, a preset fusion decision logic is executed to make a final judgment on the nature of the signal source.

2. The method of claim 1, wherein, In the initial monitoring and triggering step, the composite trigger condition includes an amplitude determination condition and a flicker frequency determination condition; The amplitude determination condition is that at least two signals of the three initial infrared radiation signals output by the three infrared detectors simultaneously and continuously exceed the preset first signal amplitude threshold; The flicker frequency determination condition is that the frequency spectrum energy peak values of at least two signals meeting the amplitude determination condition obtained by a frequency spectrum analysis algorithm are within the preset flame characteristic frequency interval.

3. The method of claim 1, wherein, In the synchronous activation mechanism, the micro-amplitude vibration excitation source is driven to work; and the micro-rotary actuator is driven to perform the spatial scanning with the azimuth angle of the suspicious flame signal as the center.

4. The method of claim 3, wherein, The preset diagnostic frequency of the mechanical vibration applied to the sensing module is higher than the flame characteristic frequency interval; and the spatial scanning of the verification detector is reciprocating scanning within a preset angle range at a preset angular velocity.

5. The method of claim 1, wherein, In the data acquisition and analysis, the process of extracting the spatial profile features includes: associating the collected output signal of the verification detector with the simultaneously recorded real-time angular position data of the detector to construct a spatial intensity distribution profile representing the change of infrared radiation intensity with scanning angle; using a mathematical fitting algorithm to model and analyze the spatial intensity distribution profile, and according to the results of the modeling analysis, determining the spatial profile features of the radiation source as point source features or surface source or line source features.

6. The method of claim 5, wherein, The modeling analysis of the spatial intensity distribution profile is specifically: fitting the data of the spatial intensity distribution profile with a standard Gaussian function model and a standard trapezoidal function model respectively, and calculating the fitting goodness of each; if the fitting goodness of the spatial intensity distribution profile with the Gaussian function model is higher than a preset first fitting threshold, and the standard deviation parameter representing the spatial scale of the radiation source calculated through the fitting is less than a preset point source scale threshold, the spatial profile features are determined as point source features. Conversely, if the spatial intensity profile has a higher goodness of fit with the trapezoidal function model, or the standard deviation parameter of the Gaussian fit is greater than the point source scale threshold, the spatial profile feature is determined to be a surface or line source feature.

7. The method of claim 1, wherein, In the data acquisition and analysis, the process of extracting the excitation coupling feature includes: From the output signal of the verification detector, a data segment corresponding to the peak region of the spatial intensity profile is intercepted; high-resolution fast Fourier transform is performed on the data segment to generate a detailed amplitude spectrum; at the preset diagnostic frequency point and its harmonic frequency point in the amplitude spectrum, it is detected whether there is a significant energy peak, and whether there is a significant excitation coupling feature or no excitation coupling feature is determined accordingly.

8. The method of claim 7, wherein, In order to realize the quantitative judgment of the energy peak, the method calculates the excitation coupling degree index by the following way: Obtain the spectral amplitude at the preset diagnostic frequency point and its second harmonic frequency point in the amplitude spectrum; sum the two spectral amplitudes to obtain an excitation signal energy representation value; Obtain the average noise amplitude in the frequency band near the diagnostic frequency point without containing known signals as the reference noise level; Calculate the ratio between the excitation signal energy representation value and the reference noise level to obtain the excitation coupling degree index; if the calculated excitation coupling degree index is greater than the preset coupling determination threshold, it is determined that there is a significant excitation coupling feature; Conversely, it is determined that there is no excitation coupling feature.

9. The method of claim 1, wherein, The fusion decision logic includes the following rules: Decision logic one: when the spatial profile feature is determined to be a point source feature and the excitation coupling feature is determined to be no excitation coupling feature, it is confirmed to be a real flame; Decision logic two: when the spatial profile feature is determined to be a surface or line source feature, regardless of the determination result of the excitation coupling feature, it is confirmed to be a static high-temperature interference source; Decision logic three: when the spatial profile feature is determined to be a point source feature and the excitation coupling feature is determined to be a significant excitation coupling feature, it is confirmed to be a dynamic periodic interference source.

10. A three-detector based infrared flame recognition and false alarm suppression system, characterized by, The system includes: A sensing module including a fixed first infrared detector, a fixed second infrared detector, and a verification detector; A micro-rotary actuator mechanically connected to the verification detector and used to drive the verification detector to perform precise angular scanning motion; A micro-amplitude vibration excitation source rigidly connected to the sensing module and used to apply a micro-amplitude mechanical vibration of a preset diagnostic frequency to the entire sensing module; A signal conditioning and acquisition unit electrically connected to the first infrared detector, the second infrared detector, and the verification detector, used to condition and digitize the signals output by each detector; A central processing and control unit connected to the signal conditioning and acquisition unit, the micro-rotary actuator, and the micro-amplitude vibration excitation source for data and control connection; Wherein, the central processing and control unit stores instruction codes, when the instruction codes are executed, the central processing and control unit is configured to execute the method of any one of claims 1 to 9.

Images