Intelligent Contamination Identification and Calibration Method for Infrared Flame Detectors in Tunnel Environments

Abstract

This application discloses an intelligent identification and calibration method for contamination of infrared flame detectors in tunnel environments, relating to the field of intelligent control identification. It addresses the technical problems of existing technologies, such as the inability to identify the type and spatial distribution of optical window contamination, the reliance on a single calibration strategy, and the difficulty in maintaining a stable alarm distance in complex tunnel environments. The method includes: acquiring data within the current detection cycle; calculating the transmittance deviation between the transmittance characterization value and a pre-calibrated reference transmittance characterization value; acquiring spatial response deviation information; acquiring at least one type of tunnel operating condition information within the tunnel environment; determining the contamination mode of the optical window based on the transmittance deviation, spatial response deviation information, and tunnel operating condition information; selecting a calibration strategy matching the contamination mode of the optical window from a set of preset calibration strategies based on the contamination mode of the optical window; and using the selected calibration strategy to perform sensitivity compensation calibration on the infrared flame detection signal.

Description

Technical Field

[0001] This application relates to the field of intelligent identification technology, and in particular to an intelligent identification and calibration method for contamination of infrared flame detectors in tunnel environments. Background Technology

[0002] Tunnels, as key enclosed nodes in highway and railway transportation networks, are high-risk fire scenarios. Infrared flame detectors are crucial equipment for early warning of tunnel fires, identifying fires by detecting infrared radiation from flames. However, in the semi-enclosed environment of tunnels, pollutants such as vehicle dust, exhaust fumes, oil mist, mud splashes, and condensation caused by temperature and humidity changes gradually accumulate on the surface of the detector's optical window, attenuating the transmittance of the flame signal. This significantly shortens the detector's effective alarm distance and, in severe cases, can even lead to missed alarms, posing a significant threat to tunnel operational safety. To address the issue of decreased sensitivity due to window contamination, existing technologies primarily employ a built-in light source combined with an external reflector for optical path self-checking. The degree of window contamination is determined by the energy change of the reflected light received by the sensor. Alternatively, a transmission-type optical path design or a photoelectric switch differential method is used to calculate the contamination rate based on the attenuation of the detected light after passing through the window, thereby uniformly compensating and calibrating the detection signal. However, structures relying on external reflectors have low reliability in tunnel vibration and dust environments, and secondary contamination of the reflector surface can easily lead to detection failure. Summary of the Invention

[0003] This application provides an intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments. It addresses the technical problems of existing technologies, such as the inability to identify the type and spatial distribution of pollution in the optical window, the reliance on a single calibration strategy, and the difficulty in maintaining a stable alarm distance in complex tunnel environments.

[0004] To achieve the above objectives, this application adopts the following technical solution: Firstly, a method for intelligent identification and calibration of contamination in infrared flame detectors for tunnel environments is provided, comprising: acquiring the active optical path detection signal and multiple infrared flame detection signals corresponding to the optical window of the infrared flame detector within the current detection cycle; the active optical path detection signal is the signal received by the reference optical path detector after the detection light emitted by the built-in light source of the infrared flame detector is transmitted through the optical window; the infrared flame detection signals are the environmental infrared radiation signals received by each of the multiple flame detection channels; determining the transmittance characterization value of the optical window within the current detection cycle based on the active optical path detection signal, calculating the transmittance characterization value and the pre-calibrated reference transmittance characterization value; and acquiring the transmittance of multiple flame detection channels for the same... Spatial response deviation information between response signals in an environmental infrared radiation scene; wherein, the spatial response deviation information is used to characterize the non-uniformity of contamination distribution on the optical window; at least one tunnel condition information in the tunnel environment is acquired, including traffic flow information and environmental humidity information; based on transmittance deviation, spatial response deviation information, and tunnel condition information, the contamination mode of the optical window is determined; the contamination mode includes at least uniform deposition type contamination, local occlusion type contamination, and interface adhesion type contamination; based on the contamination mode of the optical window, a calibration strategy matching the contamination mode of the optical window is selected from multiple preset calibration strategies, and the selected calibration strategy is used to perform sensitivity compensation calibration on the infrared flame detection signal.

[0005] The technical solution of this application combines active optical path detection with multi-channel flame detection to extract transmittance deviation and spatial response deviation information, and incorporates tunnel environmental conditions information. It can accurately identify the physical morphology of pollutants on the optical window, and then adaptively select calibration strategies according to different pollution modes, thereby achieving precise compensation for the sensitivity of flame detection signals and effectively maintaining the alarm distance of the detector in the complex pollution environment of the tunnel.

[0006] In one possible embodiment, the contamination mode of the optical window is determined based on transmittance deviation, spatial response deviation information, and tunnel condition information, including: determining a contamination attenuation coefficient based on transmittance deviation, spatial response deviation information, and tunnel condition information; the contamination attenuation coefficient is used to characterize the degree of contamination and contamination distribution characteristics of the optical window; and the contamination mode of the optical window is determined based on the matching relationship between the contamination attenuation coefficient and preset contamination mode determination conditions.

[0007] In one possible embodiment, the pollution attenuation coefficient is determined based on transmittance deviation, spatial response deviation information, and tunnel operating condition information, including: performing time-series statistics on the transmittance deviation in the current detection cycle and multiple consecutive historical detection cycles to obtain an overall transmittance attenuation component, which reflects the overall decrease in the transmittance capability of the optical window; measuring the dispersion of the spatial response deviation information corresponding to multiple flame detection channels within the same detection cycle to obtain a spatial non-uniformity component, which reflects the non-uniformity of the spatial distribution of pollution on the optical window; performing hysteresis analysis on the temporal changes of at least one operating parameter in the tunnel operating condition information and the temporal changes of transmittance deviation to obtain an environmental contribution component, which reflects the degree of contribution of the tunnel environment to the formation of pollution in the optical window; and combining the overall transmittance attenuation component, the spatial non-uniformity component, and the environmental contribution component to obtain the pollution attenuation coefficient.

[0008] In one possible embodiment, the contamination mode of the optical window is determined based on the matching relationship between the contamination attenuation coefficient and preset contamination mode determination conditions, including: determining the contamination mode of the optical window as uniform deposition type contamination based on the overall transmission attenuation component and spatial non-uniformity component in the contamination attenuation coefficient satisfying a first determination relationship; the first determination relationship is that the overall transmission attenuation component indicates a decrease in transmission capability, and the spatial non-uniformity component is lower than a preset non-uniformity threshold; determining the contamination mode of the optical window as partial occlusion type contamination based on the overall transmission attenuation component and spatial non-uniformity component in the contamination attenuation coefficient satisfying a second determination relationship; the second determination relationship is that the spatial non-uniformity component exceeds a preset non-uniformity threshold, and the overall transmission attenuation component indicates a step change in attenuation; determining the contamination mode of the optical window as interface adhesion type contamination based on the environmental contribution component and overall transmission attenuation component in the contamination attenuation coefficient satisfying a third determination relationship; the third determination relationship is that the environmental contribution component exceeds a preset environmental-related threshold, and the change amplitude of the overall transmission attenuation component during continuous detection cycles exceeds a preset rapid change threshold.

[0009] In one possible embodiment, based on the contamination pattern of the optical window, a calibration strategy matching the contamination pattern of the optical window is selected from a plurality of preset calibration strategies. This includes: extracting calibration requirement features corresponding to the contamination pattern based on the contamination pattern of the optical window; the calibration requirement features include compensation spatial distribution features and compensation temporal variation features; matching the calibration requirement features with the applicable scope features of the preset plurality of calibration strategies; the plurality of calibration strategies include at least an overall gain compensation strategy, a partitioned differential compensation strategy, and a dynamic tracking compensation strategy; and selecting a calibration strategy whose applicable scope features and current calibration requirement features satisfy preset matching conditions based on the matching results.

[0010] In one possible embodiment, sensitivity compensation calibration of the infrared flame detection signal is performed using the selected calibration strategy, including: determining the calibration gain parameter corresponding to each flame detection channel based on the contamination attenuation coefficient and the selected calibration strategy; obtaining the calibration constraints corresponding to the calibration strategy, and constraining the calibration gain parameter of each channel based on the calibration constraints; constraining the signal-to-noise ratio of the calibrated flame detection signal to be no less than a preset minimum signal-to-noise ratio threshold; and applying the constrained calibration gain parameter to the infrared flame detection signal output by the corresponding flame detection channel to obtain the calibrated flame detection signal.

[0011] In one possible embodiment, determining the transmittance characterization value of the optical window within the current detection cycle based on the active optical path detection signal, and calculating the transmittance deviation between the transmittance characterization value and a pre-calibrated reference transmittance characterization value, includes: performing time-domain segmentation on the active optical path detection signal acquired within the current detection cycle to obtain stable segment signals and interference segment signals in the active optical path detection signal; determining the transmittance characterization value of the optical window within the current detection cycle based on the amplitude statistics of the stable segment signals; performing feature analysis on the interference segment signals to obtain interference type marking information; the interference type marking information is used to distinguish between transient particulate matter occlusion interference and optical window interface state change interference in the active optical path detection signal; obtaining the reference transmittance characterization value, comparing the transmittance characterization value with the reference transmittance characterization value to obtain the initial transmittance deviation; and performing effectiveness correction on the initial transmittance deviation to obtain the final transmittance deviation.

[0012] In one possible embodiment, an intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments further includes: predicting the trend of transmittance deviation to obtain an estimated value of the transmittance attenuation of the optical window; based on the estimated value of transmittance attenuation and the currently determined pollution mode, estimating the change in the effective alarm distance of the infrared flame detector after a preset time, and outputting early warning information.

[0013] In one possible embodiment, before acquiring the spatial response deviation information between the response signals of multiple flame detection channels to the same environmental infrared radiation scene, the method further includes: performing consistency calibration on the reference response signals output by multiple flame detection channels when the optical window is in a clean state to obtain the calibration deviation value of each flame detection channel; when acquiring the spatial response deviation information, using the calibration deviation value to preprocess the response signal of each channel; the preprocessing is used to eliminate the inherent differences between channels; the inherent differences are the inconsistencies in the responses of multiple flame detection channels at the hardware level. Attached Figure Description

[0014] Figure 1 A flowchart illustrating an intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments, provided as an embodiment of this application; Figure 2 A flowchart illustrating another intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments, provided as an embodiment of this application; Figure 3 A flowchart illustrating another intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments, provided as an embodiment of this application; Figure 4 A flowchart illustrating a method for determining transmittance deviation provided in an embodiment of this application; Figure 5 A flowchart of a tunnel fire alarm method provided in an embodiment of this application. Detailed Implementation

[0015] To enable those skilled in the art to better understand the technical solutions of this disclosure, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0016] It should be noted that the terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0017] It should also be understood that the term "comprising" indicates the presence of the described feature, whole, step, operation, element and / or component, but does not exclude the presence or addition of one or more other features, wholes, steps, operations, elements and / or components.

[0018] The technical solutions in 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, and not all embodiments.

[0019] It should be noted that the network system described in the embodiments of this application is for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and does not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network systems and the emergence of other network systems, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0020] The intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments provided in this application can be applied to infrared flame detectors installed in tunnels. The detector includes an optical window, multiple flame detection channels for receiving infrared radiation from flames, a built-in light source for emitting detection light, and a reference optical path detector for receiving the detection light transmitted through the optical window.

[0021] Among them, the reference optical path detector and the flame detection channel are independent photoelectric detection elements, and the reference optical path detector is not sensitive to the characteristic wavelength of flame radiation in order to avoid the flame signal from interfering with the self-test optical path.

[0022] Specifically, the reference optical path detector can use a silicon photodiode in conjunction with a filter to make it respond only to the emission band of the built-in light source.

[0023] It should be understood that the flame detection channel refers to the complete photoelectric sensing link inside the detector used to receive the infrared radiation of the flame. Each channel contains an infrared sensitive element and a preamplifier circuit coupled to it.

[0024] Alternatively, the infrared sensing element may be a thermopile sensor or a pyroelectric sensor.

[0025] The aforementioned thermopile sensor operates based on the Seebeck effect, converting incident radiation energy into thermoelectric potential output.

[0026] The aforementioned pyroelectric sensor operates based on the pyroelectric effect and responds to the rate of temperature change.

[0027] It should be understood that both sensors have high responsivity in the characteristic bands of flame radiation.

[0028] The characteristic band of flame radiation mentioned above refers to the specific infrared emission peak produced when hydrocarbons burn, with a typical value of 4.3 μm. This band corresponds to the characteristic vibrational frequency of carbon dioxide molecules.

[0029] For example, in actual deployment, the optical axis directions of multiple flame detection channels can be spatially symmetrically distributed, such as a four-quadrant layout (up, down, left, and right), so that each channel receives incident radiation from different directional areas of the window, providing a hardware basis for subsequent analysis of the spatial distribution differences of contaminants on the window.

[0030] Optionally, the built-in light source can be an infrared LED.

[0031] Its center wavelength is 850nm or 940nm, which is located in the near-infrared region and is within the sensitive response range of silicon photodiodes. The built-in light source emits one or more detection light pulses lasting several milliseconds in each detection cycle. When the light pulse is transmitted through the optical window, it undergoes optical processes such as transmission, scattering and reflection. Contaminants attached to the optical window will change the energy distribution ratio of these optical processes.

[0032] Specifically, dust pollutants are mainly absorbed and scattered, oil pollutants are mainly absorbed and surface scattered, and water film pollutants are mainly reflected and absorbed. Therefore, the changes in the signal received by the reference optical path detector not only reflect the overall decrease in transmittance, but also contain information about the physical form of the pollutants.

[0033] In one possible implementation, the reference optical path detector is arranged inside the optical window, adjacent to the outgoing optical path of the built-in light source.

[0034] It should be understood that the above layout enables the reference optical path detector to effectively receive part of the detection light that returns after being reflected or scattered by the inner surface of the window, forming an active optical path detection signal.

[0035] The changing characteristics of this signal directly reflect the evolution of the window surface state.

[0036] This embodiment provides an intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments, such as... Figure 1 As shown, the method includes the following steps: Step S110: Collect the active optical path detection signal and multiple infrared flame detection signals corresponding to the optical window of the infrared flame detector within the current detection cycle.

[0037] The aforementioned active optical path detection signal is the signal received by the reference optical path detector after the detection light emitted by the built-in light source of the infrared flame detector is transmitted through the optical window.

[0038] The aforementioned infrared flame detection signals are the environmental infrared radiation signals received by each of the multiple flame detection channels.

[0039] In one possible implementation, the built-in light source emits a detection light pulse lasting several milliseconds at the beginning of each detection cycle. This light pulse is transmitted through an optical window and then received by a reference optical path detector and converted into an active optical path detection signal.

[0040] The center wavelength of the built-in light source is selected as 850nm or 940nm. This band is within the sensitive response range of the silicon photodiode and is separated from the characteristic band of flame radiation. The reference optical path detector, in conjunction with a narrowband filter, only responds to the emission band of the built-in light source and is insensitive to the characteristic band of flame radiation. Thus, it can accurately acquire the window self-test signal in a flame combustion scenario without being interfered with by flame light. Within the same detection cycle, four flame detection channels synchronously acquire ambient infrared radiation in their respective fields of view. Each channel contains an infrared sensitive element and a preamplifier circuit. The optical axes of the four channels are spatially symmetrically distributed in the up, down, left, and right directions, allowing each channel to receive ambient radiation incident from different areas of the window. This provides a hardware basis for subsequently determining the spatial distribution differences of contamination on the window.

[0041] Step S120: Determine the transmittance characterization value of the optical window in the current detection cycle based on the active optical path detection signal, and calculate the transmittance deviation between the transmittance characterization value and the pre-calibrated reference transmittance characterization value.

[0042] It should be understood that the aforementioned transmittance value refers to a numerical indicator that can quantify the current transmittance capability of the optical window. The higher the value, the stronger the light transmittance of the window and the less surface contamination.

[0043] The baseline transmittance characterization value is a reference value that is pre-measured and stored under the same testing conditions and with the window clean, providing a reference benchmark for subsequent judgment of the degree of window contamination.

[0044] Transmittance deviation refers to the difference between the current transmittance value and the reference transmittance value. The larger the deviation, the more severe the degree of contamination and obstruction of the window.

[0045] The aforementioned spatial response deviation information refers to a quantitative indicator used to characterize the degree of spatial non-uniformity in the distribution of contaminants on the optical window.

[0046] When the contamination on the window surface is evenly distributed, the response of each flame detection channel to the same radiation scenario should be basically consistent.

[0047] Once a window is partially blocked by pollutants such as mud or oil stains, the radiation received by the corresponding channel in the blocked area will be lower than that in the corresponding channel in the unblocked area. This results in a difference in response between multiple channels. The quantitative value of this difference is the spatial response deviation information. The larger the value, the more uneven the spatial distribution of pollution on the window.

[0048] In one possible implementation, the process of obtaining the transmittance characterization value is as follows: Statistical processing is performed on the sampled values ​​of multiple active optical path detection signals continuously acquired in the current detection cycle. First, outliers caused by transient interference are removed. The removal standard is outliers exceeding three times the standard deviation. Then, the arithmetic mean or median of the remaining sampled values ​​is taken as the transmittance characterization value for the current cycle. Using the median or truncated mean instead of the simple mean can further resist the contamination of the characterization value by residual interference pulses.

[0049] The above transmittance characterization values ​​are compared with the pre-stored reference transmittance characterization values.

[0050] Specifically, transmittance deviation can be obtained through differential calculation, which is to calculate the absolute difference between the reference value and the current value. Alternatively, transmittance deviation can be obtained through ratio calculation, which is to calculate the ratio of the current value to the reference value. The closer the ratio is to 1, the cleaner the window is; the closer it is to 0, the more serious the window contamination is.

[0051] Within the same detection cycle, the process of acquiring spatial response deviation information is as follows: The amplitude of the ambient infrared radiation signal output from each flame detection channel is acquired. After correction using the calibration deviation values ​​of each channel, the maximum difference between the corrected amplitudes of each channel is calculated, or the coefficient of variation of the corrected amplitudes of each channel is calculated. The coefficient of variation is the ratio of the standard deviation to the mean, possessing dimensional normalization characteristics, facilitating horizontal comparisons under different conditions. This maximum difference or coefficient of variation serves as the spatial response deviation information for the current cycle.

[0052] Step S130: Obtain spatial response deviation information between the response signals of multiple flame detection channels to the same environmental infrared radiation scene.

[0053] The aforementioned spatial response deviation information is used to characterize the non-uniformity of contaminant distribution on the optical window.

[0054] It should be understood that, under the condition that the optical window is clean and each channel has been calibrated for consistency, the response of each flame detection channel to the same radiation scenario should be basically consistent.

[0055] When unevenly distributed contamination appears on the window surface, the amount of ambient infrared radiation received by the channel in the contaminated and blocked area will be lower than that of the channel in the unblocked area, resulting in differences in the output signals of each channel. The quantitative measure of this difference is the spatial response deviation information. The larger the value of this information, the more uneven the spatial distribution of contamination on the window.

[0056] In one possible implementation, the amplitude of the ambient infrared radiation signal output by each flame detection channel is acquired synchronously within the same detection cycle. This amplitude represents the response of each channel to the current infrared radiation background of the tunnel environment; it can be the original voltage value or a digital value after analog-to-digital conversion. The amplitude output by each channel is corrected using the calibration deviation value of each channel. The correction method involves subtracting the calibration deviation value of the corresponding channel from the output amplitude of each channel, ensuring that each channel has a consistent response reference under clean conditions, eliminating inherent deviations caused by individual differences in detection elements, differences in optical path positions, or differences in circuit parameters between channels. The dispersion of the corrected amplitude of each channel is then measured.

[0057] One measurement method is to calculate the coefficient of variation of the corrected amplitude of each channel, which is the ratio of the standard deviation to the mean. The coefficient of variation is a dimensionless quantity, which eliminates the influence of the absolute amplitude of the signal and facilitates lateral comparison under different environmental radiation intensity conditions, making the judgment of spatial non-uniformity consistent. Another measurement method is to calculate the maximum difference between the corrected amplitudes of each channel, that is, the difference between the channel with the highest amplitude and the channel with the lowest amplitude. This method has low computational load and fast response speed, and is suitable for application scenarios with high real-time requirements.

[0058] The aforementioned coefficient of variation or maximum difference is output as the spatial response deviation information for the current detection cycle, for use in subsequent pollution mode determination.

[0059] In another possible implementation, the sampled values ​​of multiple active optical path detection signals continuously acquired within the current detection cycle are statistically processed. First, outlier sampled values ​​exceeding three times the standard deviation are removed, and then the arithmetic mean of the remaining sampled values ​​is taken as the transmittance characterization value for the current cycle.

[0060] The reference transmittance value is a reference value that is pre-calibrated and stored in the detector under the same detection conditions and in a clean state. The clean state is obtained after factory calibration or on-site maintenance and cleaning.

[0061] The transmittance deviation is obtained by performing a difference or ratio calculation between the transmittance characterization value and the reference transmittance characterization value.

[0062] Within the same detection cycle, the output signal amplitudes of four flame detection channels are acquired, and the maximum difference or coefficient of variation between the amplitudes of each channel is calculated, i.e., the ratio of the standard deviation to the mean, as spatial response deviation information. This information reflects the degree of difference in contamination obstruction in different areas of the window. The larger the spatial response deviation, the more uneven the spatial distribution of contamination on the window.

[0063] Step S140: Obtain at least one type of tunnel condition information in the tunnel environment.

[0064] The tunnel condition information mentioned above includes traffic flow information and ambient humidity information.

[0065] In one possible implementation, the detector obtains real-time traffic flow statistics and ambient humidity sensor readings from the central monitoring system of the tunnel via a fieldbus interface such as a CAN bus.

[0066] Traffic flow information is the number of vehicles passing through per unit time, which represents the intensity of exhaust emissions and dust generation. The greater the traffic flow, the faster the pollutant generation rate inside the tunnel.

[0067] The ambient humidity information is the relative humidity value inside the tunnel. When the ambient humidity is high, water vapor in the air may condense on the surface of the optical window to form a water film.

[0068] Alternatively, the detector can also have a built-in temperature and humidity sensor to directly collect ambient humidity information, thereby reducing reliance on external systems.

[0069] Step S150: Determine the contamination mode of the optical window based on transmittance deviation, spatial response deviation information, and tunnel condition information.

[0070] The aforementioned pollution modes include at least uniform deposition pollution, localized shading pollution, and interfacial adhesion pollution.

[0071] In one possible implementation, the transmittance deviation change trend is analyzed over multiple consecutive detection cycles to obtain transmittance attenuation behavior characteristics. The spatial response deviation information is measured for dispersion to obtain spatial consistency characteristics reflecting the unevenness of pollution spatial distribution. The temporal changes of at least one working parameter in the tunnel working condition information are correlated with the temporal changes of transmittance deviation to obtain environmental correlation characteristics reflecting the contribution of the tunnel environment to the formation of window pollution. Based on the combination of preset judgment conditions satisfied by the above three characteristics, the pollution mode of the current window is determined from the preset pollution modes.

[0072] Specifically, uniform deposition-type pollution is characterized by a monotonically slow increase in transmittance deviation over multiple consecutive detection cycles, with the spatial response deviation information indicating a generally consistent response attenuation magnitude across all channels, corresponding to a scenario of uniform dust settling in tunnels. Localized shading-type pollution is characterized by significant differences in response attenuation between different channels, indicated by spatial response deviation information, and a step-like jump in transmittance deviation, corresponding to scenarios of vehicle mud splashing or foreign object impact. Interface adhesion-type pollution is characterized by a synchronous fluctuation relationship between transmittance deviation changes and environmental humidity information, or a positively correlated but lagging relationship with traffic flow information, corresponding to scenarios of water film condensation or long-term oil adhesion.

[0073] Step S160: Based on the contamination mode of the optical window, select a calibration strategy that matches the contamination mode of the optical window from a number of preset calibration strategies, and use the selected calibration strategy to perform sensitivity compensation calibration on the infrared flame detection signal.

[0074] In one possible implementation, the detector has multiple preset calibration strategies, including at least an overall gain compensation strategy, a zone-specific compensation strategy, and a dynamic tracking compensation strategy.

[0075] When the contamination mode is uniform deposition type, an overall gain compensation strategy is selected. Based on the transmittance deviation, the signal gain of all flame detection channels is uniformly adjusted so that the compensated signal is equivalent to the window clean state level. When the contamination mode is local occlusion type, a zone-based differentiated compensation strategy is selected. Based on the spatial response deviation information, the relative attenuation degree of each channel is determined. Higher gain is applied to channels with large attenuation, while the original gain is maintained or a lower gain is applied to channels with small attenuation, achieving differentiated compensation that matches the occlusion degree of each channel.

[0076] When the pollution mode is interface adhesion type pollution, a dynamic tracking compensation strategy is selected to make the compensation parameters adjust in real time according to the dynamic changes in transmittance deviation, so as to avoid compensation lag or overcompensation caused by the periodic formation and dissipation of water film or the continuous accumulation of oil.

[0077] The beneficial effects achievable through the aforementioned technical means are that by combining active optical path detection with multi-channel flame detection, transmitting power deviation and spatial response deviation information can be extracted and integrated with tunnel environmental conditions. This allows for precise identification of the physical morphology of pollutants on the optical window, and adaptive selection of calibration strategies based on different pollution modes. This achieves accurate compensation for the sensitivity of the flame detection signal and effectively maintains the alarm distance of the detector in the complex pollution environment of the tunnel.

[0078] This embodiment provides yet another intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments, such as... Figure 2 As shown, the method includes the following steps: Step S210: Acquire active optical path detection signal and infrared flame detection signal during the current detection cycle.

[0079] The aforementioned detection cycle refers to the time interval between a complete data acquisition, contamination assessment, and calibration update performed by the detector.

[0080] It should be understood that the detection cycle can be set according to the rate of change of pollution in the tunnel environment and the power consumption requirements of the detector.

[0081] For example, it can be set to collect data once every 5 seconds, every 10 seconds, or every 30 seconds.

[0082] The aforementioned active optical path detection signal refers to the signal received by the reference optical path detector after the detection light emitted by the built-in light source of the detector is transmitted through the optical window. It directly reflects the overall transmittance of the optical window to the detection light. When there is contamination on the surface of the window, the detection light will be absorbed, scattered or reflected by the contaminants, resulting in a decrease in the light energy received by the reference optical path detector. Therefore, the strength of the active optical path detection signal can be used as a basis for characterizing the degree of window contamination.

[0083] The aforementioned infrared flame detection signals refer to the environmental infrared radiation signals received by each of the multiple flame detection channels.

[0084] It should be understood that the flame detection channel refers to the photoelectric sensor and its signal processing circuit inside the detector used to receive infrared radiation from the flame.

[0085] Specifically, each channel includes an infrared sensing element and a preamplifier circuit, which has a high response to the characteristic bands of flame radiation.

[0086] The aforementioned infrared sensing element can be a thermopile sensor or a pyroelectric sensor.

[0087] The ambient infrared radiation signal includes not only possible flame radiation, but also infrared radiation from the background environment inside the tunnel. Together, they constitute the original signal output by each channel.

[0088] In one possible implementation, the detector's built-in light source emits one or more detection light pulses lasting several milliseconds each detection cycle. The center wavelength of the built-in light source can be selected as 850nm or 940nm, which is a commonly used near-infrared band in infrared detectors.

[0089] The detection light is received by a reference optical path detector arranged inside the optical window. The reference optical path detector and the flame detection channel are independent photoelectric detection elements.

[0090] It should be understood that the reference optical path detector is not sensitive to the characteristic wavelengths of flame radiation.

[0091] The typical value of the characteristic band of flame radiation is 4.3 μm, which belongs to the characteristic radiation peak of hydrocarbon combustion.

[0092] The reference optical path detector, in conjunction with a filter, responds only to the emission band of the built-in light source, thereby avoiding interference from flame signals to the self-test optical path. At the same time, multiple flame detection channels receive infrared radiation from the environment.

[0093] For example, the detector can be equipped with four flame detection channels, each facing a different area of ​​the window. The optical axes of the four channels can be symmetrically distributed vertically and horizontally to fully perceive the transmission of light in each area of ​​the window.

[0094] Step S220: Obtain spatial response deviation information between the response signals of each flame detection channel to the same environmental infrared radiation scene.

[0095] Spatial response deviation information refers to a quantitative indicator used to characterize the degree of spatial unevenness in contamination distribution within a radiation window. Under conditions where the window is clean and all channels have undergone consistent calibration, the responses of each channel to the same radiation scenario should be essentially identical. If a certain area of ​​the window is partially obscured by mud or oil stains, the received signal of the corresponding channel in that area will be lower than that of other channels, resulting in response differences among multiple channels. The greater the difference, the more uneven the contamination distribution.

[0096] The aforementioned infrared radiation scenario in the same environment can be the normal background radiation of a tunnel in a fire-free state, or it can be the same standard test fire source.

[0097] In one implementation, the signal amplitudes of the four channels are acquired within the same detection period, and the maximum difference between the amplitudes of each channel is calculated, or the coefficient of variation of the amplitude is calculated, which is the ratio of the standard deviation to the mean. The maximum difference or coefficient of variation is used as spatial response deviation information for subsequent analysis.

[0098] Step S230: Determine the contamination mode of the optical window based on the transmittance deviation and spatial response deviation information.

[0099] Pollution patterns refer to categories classified based on the physical form of pollutants and their distribution characteristics on the window surface. This embodiment presupposes two modes: uniform deposition pollution and partial obstruction pollution. Uniform deposition pollution is characterized by pollutants covering the window relatively evenly, with a generally consistent attenuation of transmittance across different areas; a typical cause is the continuous settling of fine dust. Partial obstruction pollution is characterized by pollutants covering only a portion of the window, with significant differences in transmittance at different locations; a typical cause is the impact of mud or gravel splashed up by vehicles.

[0100] In one implementation, a trend analysis is performed on the transmittance deviation over the most recent five detection cycles.

[0101] If the deviation increases monotonically and slowly in each cycle, and the spatial response deviation information indicates that the response differences between channels are small and below the preset non-uniform threshold, then it is determined to be uniform deposition type pollution. The specific meaning of monotonically and slowly increasing is that the increase in each cycle does not exceed a first preset proportion of the historical maximum change; the first preset proportion can be 30%.

[0102] If the spatial response deviation information indicates that the response difference of each channel exceeds the preset non-uniform threshold, and the transmittance deviation changes significantly in a certain period and the single-period increase exceeds the second preset proportion of the benchmark value; the second preset proportion can be 5%, then it is determined to be local shading pollution, and this change is the step characteristic.

[0103] Optionally, a preset non-uniform threshold can be triggered when the ratio of the maximum difference to the mean of the signals in each channel reaches 15%.

[0104] Step S240: Select a matching scheme from the preset calibration strategy and perform sensitivity compensation calibration according to the contamination mode.

[0105] The calibration strategy refers to the gain adjustment scheme reserved to compensate for signal attenuation caused by window contamination.

[0106] It should be understood that different strategies correspond to different contamination modes, with the aim of restoring sensitivity while avoiding the introduction of noise due to excessive amplification.

[0107] It should be understood that when the window surface is uniformly contaminated, the signal attenuation amplitude of each flame detection channel is basically the same. If the same compensation gain is applied to each channel, the output signal of all channels can be equivalently restored to the level of the clean window state. The compensation method is simple and does not introduce relative deviation between channels.

[0108] When the window surface is partially obstructed by contamination, the signal attenuation amplitude of each channel varies significantly. If a uniform gain is still used for compensation, two adverse consequences will occur. On the one hand, the channel with less obstruction has less signal attenuation, and the signal will be over-amplified after applying the uniform gain, and the background noise will be amplified simultaneously, resulting in a decrease in the signal-to-noise ratio of the channel. On the other hand, the channel with severe obstruction has greater signal attenuation, and the uniform gain may not be sufficient to fully compensate for it, resulting in insufficient detection sensitivity of the channel.

[0109] In one implementation, when the pollution is determined to be uniform deposition type, an overall gain compensation strategy is selected.

[0110] Specifically, a uniform compensation gain coefficient is calculated based on the transmittance deviation of the current cycle.

[0111] The calculation method is to divide the reference transmittance value by the current transmittance value, and the resulting ratio is the gain coefficient. For example, if the transmittance value drops to 70% of the reference value, the gain coefficient is 1 divided by 0.7. After multiplying the output signals of each channel by this coefficient, the response level is equivalent to that of a clean state.

[0112] The calculated uniform gain coefficient is applied to the output signal of each flame detection channel to complete the proportional signal amplification of all channels.

[0113] In one implementation, when the pollution is determined to be localized shading, a zone-specific differential compensation strategy is selected.

[0114] Specifically, the relative attenuation level of each channel is identified based on spatial response deviation information. The channel with the highest response signal amplitude in the current period is used as the reference channel, corresponding to the area with the least contamination on the window. The response amplitudes of the remaining channels are compared with the response amplitude of the reference channel to calculate the attenuation ratio of each channel relative to the reference channel.

[0115] The compensation gain coefficient for each channel is determined according to the attenuation ratio of each channel. A higher gain is applied to the channel with a large attenuation ratio so that its output level after compensation is close to that of the reference channel. The original gain is maintained or a lower gain is applied to the channel with a small attenuation ratio so that its output is kept at a level comparable to that of the reference channel without being over-amplified. The differential gain coefficients corresponding to each channel are applied to the output signal of the corresponding channel.

[0116] The aforementioned zone-based differentiated compensation method ensures that the output of channels with different degrees of obstruction tends to be balanced after compensation. This avoids the unobstructed channels being over-amplified and introducing noise due to uniform compensation, while ensuring that the severely obstructed channels receive sufficient compensation and restore their proper flame detection capabilities.

[0117] The advantage of this embodiment is that by introducing multi-channel spatial response deviation information, it achieves the classification and identification of uniform pollution and local pollution, solving the problem that traditional methods treat the window as a uniform whole and cannot distinguish the pollution type. Uniform deposition pollution is compensated as a whole, which is simple and effective. Partially obstructed pollution is compensated in a zoned differentiated manner, and is processed according to the actual degree of obstruction of the channel. This avoids the dual drawbacks of noise amplification in unobstructed channels and insufficient compensation in obstructed channels under uniform compensation. No external reflectors or other mechanical parts are required, and the equipment maintains the simplicity of the equipment structure and the reliability of operation in tunnel vibration and dust environments.

[0118] This embodiment provides yet another intelligent pollution identification and calibration method for infrared flame detectors in tunnel environments, such as... Figure 3 As shown, the method includes the following steps: Step S310: Collect active optical path detection signals, infrared flame detection signals, and tunnel working condition information during the current detection cycle.

[0119] It should be understood that the aforementioned tunnel operating condition information refers to data reflecting the environmental operating status related to pollution formation within the tunnel, which can be obtained through the communication interface with the tunnel monitoring system or collected through auxiliary sensors built into the detector.

[0120] The traffic flow information mentioned above refers to the number of vehicles passing through the tunnel per unit time, used to characterize the intensity of exhaust emissions and dust generation. The higher the traffic flow, the faster the pollutants are generated inside the tunnel.

[0121] The above-mentioned ambient humidity information refers to the relative humidity value inside the tunnel, which is used to characterize the water vapor content. When the ambient humidity is high, water vapor inside the tunnel may condense on the surface of the optical window to form a water film.

[0122] In one possible implementation, the detector obtains real-time traffic flow statistics and ambient humidity sensor readings from the central monitoring system of the tunnel via a fieldbus interface such as RS-485 or CAN bus, and uses these as tunnel condition information for subsequent processing.

[0123] Optionally, the detector can also have a built-in temperature and humidity sensor to directly collect ambient humidity information.

[0124] Step S320: Determine the transmittance characterization value and calculate the transmittance deviation based on the active optical path detection signal.

[0125] For details, please refer to S220; this embodiment will not elaborate further.

[0126] Step S330: Obtain spatial response deviation information for multiple flame detection channels.

[0127] For details, please refer to S230; this embodiment will not elaborate further.

[0128] Step S340: Determine the contamination mode of the optical window based on transmittance deviation, spatial response deviation information, and tunnel condition information.

[0129] It should be understood that this embodiment further expands the identifiable contamination modes based on uniform deposition type contamination and localized occlusion type contamination, adding interface adhesion type contamination.

[0130] The aforementioned interface-adhesive contamination refers to one or more layers of deposits formed on the surface of the optical window, whose transmittance characteristics fluctuate over time.

[0131] For example, typical interfacial fouling includes water film aggregation fouling and oil adhesion fouling.

[0132] The aforementioned water film condensation pollution refers to the thin water film formed by the condensation of water vapor in the air on the window surface due to temperature and humidity differences between the inside and outside of the tunnel. The characteristic of this type of pollution is that the transmittance changes periodically with the ambient humidity. The periodic fluctuations in humidity (such as condensation caused by temperature differences between day and night) will be directly reflected in the fluctuations in transmittance.

[0133] The aforementioned oil-adhesive pollution refers to the long-term adhesion of unburned oil mist, asphalt volatiles, and other substances from vehicle exhaust to the window surface. This type of pollution is characterized by a slow but continuous decrease in light transmittance. Furthermore, due to the adhesive properties of oil, dust tends to adhere more easily, leading to an accelerated decay trend. The long-term trend of light transmittance changes is positively correlated with traffic flow information.

[0134] In one possible implementation, the determination process in step S340 includes the following sub-steps: A preliminary judgment is made based on the short-term variation characteristics of transmittance deviation.

[0135] If the transmittance deviation changes by more than the preset rapid change threshold in a short period of time, indicating that the attenuation has a rapid change characteristic, it is initially judged to be interface adhesion type pollution, because uniform deposition type and local shading type usually change more slowly or have a step.

[0136] (2) Analyze the relationship between tunnel working condition information and transmittance deviation.

[0137] Specifically, the temporal changes in environmental humidity information are correlated with the temporal changes in transmittance deviation, or the temporal changes in traffic flow information are correlated with the long-term trend of transmittance deviation.

[0138] The above correlation analysis can be performed by methods such as calculating correlation coefficients, cross-correlation functions, or observation time series synchronicity.

[0139] For example, if the transmittance deviation fluctuates synchronously with the ambient humidity information in a short period of time, that is, the deviation increases when the humidity increases and decreases when the humidity decreases, and the spatial response deviation information indicates that the attenuation of each channel is relatively uniform, then it is determined to be water film coagulation type pollution.

[0140] If the trend of transmittance deviation lags behind traffic flow information over a longer time span, that is, after a period of time after the peak traffic flow period, the transmittance begins to decrease rapidly, and the active optical path detection signal is accompanied by scattering enhancement characteristics, such as the reference optical path detector detecting additional scattered light components while receiving transmitted light, then it is determined to be oil stain adhesion type pollution.

[0141] Step S350: Select a calibration strategy based on the contamination mode and perform sensitivity compensation calibration.

[0142] It should be understood that this embodiment adds a dynamic tracking compensation strategy on the basis of the overall gain compensation strategy and the partition-differentiated compensation strategy.

[0143] The aforementioned dynamic tracking compensation strategy refers to a system where the compensation parameters are not fixed but are adjusted in real time to follow the dynamic changes in transmittance deviation in order to adapt to rapid fluctuations in the degree of pollution.

[0144] In one possible implementation, if the pollution pattern is determined to be uniform deposition type pollution, the overall gain compensation strategy is still adopted; if the pollution pattern is determined to be local shading type pollution, the zone-specific differential compensation strategy is still adopted.

[0145] If the contamination mode is determined to be interface adhesion type contamination, then a dynamic tracking and compensation strategy is selected.

[0146] Specifically, the processor continuously monitors the real-time trend of transmittance deviation, removes measurement noise using moving average filtering or low-pass filtering, calculates the required compensation gain parameters in real time, and updates the gain value of each channel once every detection cycle or every few cycles.

[0147] For example, for water film condensation-type pollution, when an increase in humidity and a decreasing trend in transmittance are detected, gain compensation is increased in advance or simultaneously; when humidity drops and transmittance begins to recover, gain compensation is reduced in a timely manner to avoid overcompensation leading to false alarms.

[0148] For oil-adhesive pollution, the compensation parameters are dynamically adjusted to follow the long-term, slow decline in transmittance.

[0149] The beneficial effects of this embodiment are as follows: by incorporating tunnel working condition information, it can effectively identify interface adhesion pollution such as water film agglomeration and oil stain adhesion, overcoming the problem that relying solely on optical signals cannot distinguish the cause of pollution. The dynamic tracking compensation strategy can adjust the compensation parameters in real time according to the dynamic changes in transmittance deviation, solving the sensitivity drift problem caused by the periodic fluctuation of transmittance in water film condensation scenarios under traditional fixed parameter compensation, avoiding the risk of false alarms due to overcompensation during transmittance recovery, and is suitable for areas with large temperature and humidity changes such as tunnel entrances and exits, significantly improving the adaptability of the detector under various environmental conditions.

[0150] This embodiment provides a method for determining transmittance deviation, such as... Figure 4 As shown, the method includes the following steps: This embodiment focuses on improving the calculation method of transmittance deviation. By performing time-domain segmentation and interference type labeling on the active optical path detection signal, the accuracy and reliability of the transmittance characterization value are improved, thereby enhancing the accuracy of subsequent contamination pattern recognition.

[0151] Step S410: Perform time-domain segmentation on the active optical path detection signal acquired within the current detection period to obtain the stable segment signal and the interference segment signal.

[0152] It should be understood that the above-mentioned time-domain segmentation refers to the process of dividing the continuously acquired active optical path detection signals within a detection cycle into multiple analysis intervals according to the chronological order, and classifying each interval into a stable segment signal or an interference segment signal based on the signal statistical characteristics of each interval.

[0153] The aforementioned stable signal segment refers to the portion of the signal amplitude that fluctuates less and is relatively stable, representing the continuous light transmission state of the optical window within this time interval.

[0154] The aforementioned interference segment refers to the portion of the signal that exhibits obvious rapid spikes, steep drops, or abnormal fluctuations, indicating that the detection optical path has been subjected to instantaneous external interference.

[0155] In one possible implementation, active optical path detection signals are continuously acquired at a high sampling rate within a detection cycle to obtain a time-domain sampling sequence. This sequence is then segmented, and the short-time variance or range of each segment is calculated. Segments with short-time variances below a preset fluctuation threshold are marked as stable segments, while segments with short-time variances above the preset fluctuation threshold are marked as interference segments.

[0156] Step S420: Determine the transmittance characterization value within the current detection cycle based on the amplitude statistics of the stable segment signal.

[0157] It should be understood that since the interference segment signal contains instantaneous interference components, its amplitude cannot accurately reflect the true contamination state of the window. Therefore, the statistics of the stable segment signal are used as the transmittance characterization value.

[0158] In one possible implementation, the sampled values ​​of all stable segment signals identified in step S410 are statistically processed, for example, by taking the median or truncated average (i.e., the arithmetic mean after removing the maximum and minimum sampled values), as the transmittance characterization value for the current detection period. Using the median or truncated average instead of a simple average can further resist the influence of residual interference pulses.

[0159] Step S430: Perform feature analysis on the interference segment signal to obtain interference type marking information.

[0160] It should be understood that the above-mentioned interference type marking information refers to the identifier used to distinguish different types of interference in the active optical path detection signal, providing a basis for subsequent targeted correction of transmittance deviation.

[0161] The aforementioned transient particulate matter obstruction interference refers to the interference caused by particles such as smoke plumes and dust that quickly drift through the tunnel and momentarily rise up, which momentarily obstruct the detection optical path. Its characteristics are that the interference pulse width is extremely narrow and the amplitude is peak-shaped.

[0162] The aforementioned optical window interface state change interference refers to the slow signal change caused by the physical state change of the optical window surface, characterized by a slowly varying baseline drift of the signal over a longer time scale.

[0163] In one possible implementation, the waveform characteristics of each interference segment signal are analyzed. These waveform characteristics may include the duration of the interference pulse, the steepness of its amplitude change, and its frequency of occurrence. If the fall and recovery times of the interference segment signal are extremely short and the pulse has a sharp shape, it is labeled as transient particulate matter obstruction interference. If the interference segment signal exhibits a slow-changing process over a relatively long period, it is labeled as optical window interface state change interference.

[0164] Step S440: Obtain the reference transmittance characterization value, calculate the initial transmittance deviation, and perform an effectiveness correction on the initial transmittance deviation based on the interference type labeling information to obtain the final transmittance deviation.

[0165] It should be understood that the above-mentioned effectiveness correction refers to the reasonable adjustment of the transmittance deviation calculated from the stable segment signal according to the type and intensity of the interference, so that the corrected transmittance deviation more accurately reflects the actual degree of contamination of the window and avoids misjudgment of the contamination status due to various interference factors.

[0166] In one possible implementation, the transmittance characterization value obtained in step S420 is compared with the reference transmittance characterization value to obtain an initial transmittance deviation. Then, based on the interference type labeling information obtained in step S430, the initial deviation is corrected. If the interference is marked as transient particulate matter obstruction, and the number and intensity of interference pulses detected in the interference segment signal exceed the preset interference tolerance limit, the weight of the transmittance deviation in the current cycle in subsequent trend judgment should be appropriately reduced, or the transmittance deviation value of the previous effective cycle should be directly used to replace the current value, in order to avoid sudden interference such as instantaneous dust being misjudged as a sudden increase in window contamination. If the interference is marked as optical window interface state change, it indicates that the window surface may be undergoing physical changes. In this case, the detected deviation change can be appropriately retained, and the response weight of the deviation in the current cycle in subsequent contamination mode judgment can be moderately increased, so that the system remains sensitive to interface state changes.

[0167] The beneficial effects of this embodiment are as follows: By performing time-domain segmentation and interference type identification on the active optical path detection signal, the interference caused by transient particulate matter obstruction in the tunnel on the transmittance deviation calculation is effectively eliminated. At the same time, it distinguishes and rationally utilizes the signal drift information caused by changes in the physical state of the window surface, improving the accuracy and stability of the transmittance characterization value. It also significantly reduces misjudgments of pollution status and misadjustments of calibration parameters caused by sudden factors such as instantaneous dust from vehicles. This embodiment has particularly significant applicability to tunnel environments with high traffic volume and frequent dust. This embodiment provides a tunnel fire alarm method, such as... Figure 5 As shown, the specific steps include: Step S510: Perform baseline response consistency calibration on each flame detection channel.

[0168] It should be understood that the above-mentioned benchmark response consistency calibration refers to the process of pre-measuring and recording the response differences of each flame detection channel to the same standard radiation source when the optical window is in a clean state. Its purpose is to eliminate the inherent response deviation between channels caused by hardware differences.

[0169] The aforementioned inherent response deviation refers to the inconsistency in response between different flame detection channels caused by at least one of the following factors: differences in the individual characteristics of the detection element, differences in the optical path position, or differences in circuit parameters. For example, the responsivity of different thermopile sensors exhibits batch-specific variability, and even slight differences in the installation angle of the channel inside the detector can lead to slight variations in the received radiation.

[0170] In one possible implementation, during factory calibration or after on-site installation and optical window cleaning, a reference response consistency calibration is performed on each flame detection channel. Specifically, with the optical window clean, each channel is aligned with the same stable standard infrared radiation source, such as a built-in standard blackbody, or a relatively stable background area selected in a tunnel environment, and the reference response signal of each channel is collected. The difference between the reference response signal of each channel and the average value of the reference response signals of all channels is calculated as the calibration deviation value of each channel, and the calibration deviation value of each channel is stored in the detector's internal memory.

[0171] Before acquiring spatial response deviation information each time, the real-time response signal of each channel is corrected using the calibration deviation value corresponding to each channel. This ensures that the response benchmark of each channel is consistent under clean conditions, and that the spatial response deviation in subsequent analysis fully reflects the non-uniformity of window contamination, rather than the hardware differences between channels.

[0172] Step S520: Determine the pollution attenuation coefficient based on transmittance deviation, spatial response deviation information, and tunnel operating condition information.

[0173] It should be understood that the aforementioned pollution attenuation coefficient is a comprehensive parameter composed of multiple components, used to comprehensively characterize the degree of pollution and pollution distribution characteristics of the optical window, and to provide a unified basic parameter for subsequent pollution mode determination.

[0174] The aforementioned overall transmission attenuation component refers to a quantitative indicator reflecting the decrease in the overall transmission capability of the optical window, which is derived from the transmission deviation through time series statistics.

[0175] The aforementioned spatial non-uniformity component refers to a quantitative indicator reflecting the degree of non-uniformity in the spatial distribution of pollution on the optical window, which is derived from spatial response deviation information through a measure of dispersion.

[0176] The aforementioned environmental contribution component refers to a quantitative indicator reflecting the degree of contribution of tunnel environmental factors to the formation of optical window pollution. It is derived from the lag analysis between the temporal changes of tunnel operating conditions and the temporal changes of transmittance deviation.

[0177] In one possible implementation, step S520 includes the following sub-steps: (1) Obtain the overall transmission attenuation component.

[0178] The transmittance deviations in the current detection cycle and multiple consecutive historical detection cycles are statistically analyzed over time, and their mean or weighted moving average is calculated as the overall transmittance attenuation component. The larger the value of this component, the worse the overall light transmittance of the window.

[0179] (2) Obtain the spatial non-uniformity component.

[0180] The spatial response deviation information of each flame detection channel within the same detection cycle is measured for dispersion. Specifically, the coefficient of variation (ratio of standard deviation to mean) of the response signal of each channel after calibration deviation correction in step S410 is calculated. This coefficient of variation is the spatial non-uniformity component.

[0181] (3) Obtain the environmental contribution component.

[0182] A lag analysis was performed on the temporal variation of at least one working parameter in the tunnel working condition information and the temporal variation of the transmittance deviation.

[0183] For example, cross-correlation analysis can be performed on the time series of ambient humidity values ​​and the time series of transmittance deviation over multiple consecutive detection cycles to calculate the cross-correlation coefficient between the two and identify the time lag of humidity change relative to transmittance deviation change.

[0184] If the cross-correlation coefficient exceeds the preset correlation threshold and the humidity change precedes the deviation change by a set time, it indicates that the ambient humidity has a significant contribution to window pollution, and the environmental contribution component is the product of the cross-correlation coefficient and the lag weight; otherwise, the environmental contribution component is zero or a small value.

[0185] (4) Combine the overall transmission attenuation component, the spatial non-uniformity component and the environmental contribution component to obtain the pollution attenuation coefficient.

[0186] One possible combination is a weighted summation, i.e., pollution attenuation coefficient = w1 × overall transmission attenuation component + w2 × spatial non-uniformity component + w3 × environmental contribution component, where w1, w2, and w3 are preset weight coefficients, and the sum of the three is 1. The optimal weight can be determined through offline experimental calibration.

[0187] It should be understood that the above weighted summation is performed after preprocessing each parameter to the same dimension, and the preprocessing method can be normalization.

[0188] Step S530: Determine the contamination mode of the optical window based on the matching relationship between the contamination attenuation coefficient and the preset contamination mode determination conditions.

[0189] In one possible implementation, the detector is pre-set with multiple contamination mode determination conditions. Each determination condition defines the combination relationship that each component in the contamination attenuation coefficient must satisfy. The contamination attenuation coefficient calculated in step S420 is compared and matched with each pre-set determination condition one by one to determine the contamination mode that best matches the current data.

[0190] Specifically, the preset pollution mode determination conditions include: The first determination relationship is that when the overall transmission attenuation component indicates a decrease in transmission capability (i.e., the overall transmission attenuation component exceeds the preset attenuation threshold) and the spatial non-uniformity component is lower than the preset non-uniformity threshold, the pollution mode is determined to be uniform deposition type pollution if this determination relationship is satisfied.

[0191] The second determination relationship is: if the spatial non-uniformity component exceeds the preset non-uniformity threshold and the overall transmission attenuation component indicates a step change in attenuation (i.e., between adjacent detection cycles, the increase of the overall transmission attenuation component exceeds the preset step threshold), then the contamination mode is determined to be local occlusion type contamination.

[0192] The third determining relationship is that if the environmental contribution component exceeds the preset environmental-related threshold and the overall transmission attenuation component changes more than the preset rapid change threshold during multiple consecutive detection cycles, the pollution mode is determined to be interface adhesion type pollution.

[0193] Step S540: Select a calibration strategy that matches the current contamination status from a set of preset calibration strategies.

[0194] It should be understood that the selection of calibration strategies in this embodiment is not based on a one-to-one matching of the contamination mode name, but rather on feature matching based on the calibration requirement characteristics extracted from the contamination attenuation coefficient and the applicable scope characteristics of each preset calibration strategy, thereby improving the flexibility and adaptability of the matching.

[0195] The aforementioned calibration requirement characteristics refer to the description of the spatial distribution and temporal variation requirements of the compensated signal, derived from the values ​​of each component in the current contamination attenuation coefficient. These calibration requirement characteristics include spatial distribution characteristics and temporal variation characteristics. The spatial distribution characteristics describe the degree of difference in compensation required for each channel, while the temporal variation characteristics describe the rate of change of the compensation amount.

[0196] The above-mentioned scope of application characteristics refer to the description of the applicable scenarios declared for each preset calibration strategy.

[0197] In one possible implementation, step S540 includes the following sub-steps: (1) Extract calibration requirement characteristics based on pollution attenuation coefficient.

[0198] For example, the spatial non-uniformity component is compared with a preset threshold. If it is higher than the threshold, the compensation spatial distribution characteristic is determined to be "non-uniform"; otherwise, it is "uniform". The recent rate of change of the overall transmission attenuation component is compared with a preset rate threshold. If it exceeds the threshold, the compensation temporal change characteristic is determined to be "rapid change"; otherwise, it is "stable change".

[0199] (2) Match the calibration requirement characteristics with the applicable scope characteristics of the preset calibration strategy.

[0200] The detector is equipped with at least three calibration strategies: overall gain compensation strategy (the applicable range of the strategy is uniform in spatial distribution and stable in temporal variation), zone-specific differential compensation strategy (the applicable range of the strategy is non-uniform in spatial distribution), and dynamic tracking compensation strategy (the applicable range of the strategy is rapid in temporal variation).

[0201] (3) Select a calibration strategy whose applicable scope characteristics and current calibration requirements meet the preset matching conditions.

[0202] The aforementioned preset matching conditions refer to the matching or full agreement of the feature values ​​of the two.

[0203] Step S550: Using the selected calibration strategy, determine the calibration gain parameters of each channel based on the contamination attenuation coefficient and perform constraint processing.

[0204] It should be understood that due to the high background noise in the tunnel environment, excessive signal amplification due to severe pollution may also amplify the background noise, leading to a deterioration in the signal-to-noise ratio and consequently affecting the reliability of flame detection. Therefore, it is necessary to impose constraints on the calibration gain parameters of each channel.

[0205] The aforementioned calibration constraints refer to preset gain limits, such as a minimum signal-to-noise ratio (SNR) threshold, to ensure the quality of the flame detection signal after calibration. The SNR refers to the ratio of effective signal power to background noise power in the flame detection channel output signal; a higher SNR indicates a stronger ability for the detector to distinguish flames.

[0206] In one possible implementation, calibration gain parameters are determined for each flame detection channel based on the contamination attenuation coefficient and the selected calibration strategy.

[0207] For example, for an overall gain compensation strategy, the calibration gain parameter of each channel is related to the overall transmission attenuation component; for a zone-specific differential compensation strategy, the calibration gain parameter of each channel is also related to spatial response deviation information. The specific calculation method can be achieved by looking up a preset compensation parameter mapping table, which contains pre-calibrated relationships between different pollution attenuation coefficient ranges and their corresponding calibration gain parameters.

[0208] After determining the initial calibration gain parameters, obtain the calibration constraints corresponding to the selected calibration strategy.

[0209] For example, the preset minimum signal-to-noise ratio (SNR) threshold is 10 dB. The evaluation assesses whether the output SNR of each channel remains above this threshold after amplifying the signal at the current calculated gain. If a channel experiences severe attenuation requiring extremely high gain, resulting in proportional amplification of background noise and a predicted SNR drop below the threshold, then the gain of that channel is constrained to the maximum allowable range that ensures the SNR does not fall below the threshold. The constrained calibration gain parameters are then applied to the corresponding channel's output infrared flame detection signal to obtain the calibrated flame detection signal.

[0210] Step S560: Transmittance deviation trend prediction and graded early warning.

[0211] It should be understood that the above trend prediction refers to the use of mathematical methods to estimate the transmittance decay trend in the future period based on existing historical data of transmittance deviation.

[0212] The aforementioned tiered early warning refers to outputting alarm signals of different levels based on preset threshold levels, in order to provide tiered reminders for maintenance and emergency response.

[0213] One possible implementation involves curve fitting of the overall transmission attenuation component sequence recorded over multiple consecutive detection cycles, or using time series prediction algorithms such as exponential smoothing or ARIMA to estimate the degree of transmission attenuation after a set time.

[0214] Based on the currently determined pollution pattern, the change in the effective alarm distance of the infrared flame detector after a set time is estimated. When the estimated distance will drop below the safe limit, an early warning message is output.

[0215] For example, the system may display a message such as "Window contamination is worsening, and the alarm distance is expected to drop to X meters. Maintenance is recommended as soon as possible."

[0216] When the transmittance deviation exceeds the preset severe contamination threshold in any detection cycle, an optical window contamination fault alarm signal is output. At this time, the sensitivity compensation calibration function for the flame detection signal is still maintained, so that the detector can continue to work.

[0217] When the transmittance deviation exceeds the preset failure threshold, it indicates that the window contamination is extremely serious and compensation can no longer reliably restore the detection performance. At this time, a detector failure alarm signal is output, and the output of the flame detection signal can be stopped to avoid false alarms in an unreliable state.

[0218] The beneficial effects of this embodiment are as follows: Channel consistency calibration eliminates inherent hardware differences, ensuring that spatial response deviation information accurately reflects pollution distribution. By constructing a pollution attenuation coefficient comprising three components—overall transmission attenuation, spatial non-uniformity, and environmental contribution—multi-dimensional and refined quantification of pollution states is achieved. Pollution pattern classification based on multi-component judgment relationships is more robust and accurate than single-parameter judgment. Strategy selection is achieved through a matching mechanism between calibration requirement characteristics and strategy applicability characteristics, making strategy matching more flexible and interpretable. Applying signal-to-noise ratio constraints prevents the risk of false alarms caused by background noise introduced due to overcompensation under high pollution conditions. Trend prediction and tiered early warning enable a shift from passive response to proactive maintenance, allowing maintenance personnel to schedule window cleaning operations in advance based on early warning information, significantly improving the overall reliability and operational efficiency of the tunnel fire protection system.

[0219] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another apparatus, or some features may be ignored or not executed. Furthermore, the mutual coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0220] The units described as separate components may or may not be physically separate. A component shown as a unit can be one or more physical units; that is, it can be located in one place or distributed in multiple different locations. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0221] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0222] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a readable storage medium. Based on this understanding, the technical solutions of the embodiments of this application, essentially, or the parts that contribute to the prior art, or all or part of the technical solutions, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0223] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for intelligent identification and calibration of contamination in infrared flame detectors for tunnel environments, characterized in that, The method includes: The active optical path detection signal and multiple infrared flame detection signals corresponding to the optical window of the infrared flame detector are collected during the current detection cycle. The active optical path detection signal is the signal received by the reference optical path detector after the detection light emitted by the built-in light source of the infrared flame detector is transmitted through the optical window. The infrared flame detection signals are the ambient infrared radiation signals received by each of the multiple flame detection channels. Based on the active optical path detection signal, the transmittance characterization value of the optical window in the current detection cycle is determined, and the transmittance characterization value is calculated to be different from the pre-calibrated reference transmittance characterization value. Spatial response deviation information is obtained among the response signals of the multiple flame detection channels to the same environmental infrared radiation scene; wherein, the spatial response deviation information is used to characterize the non-uniformity of the contaminant distribution on the optical window; Acquire at least one type of tunnel condition information in the tunnel environment, wherein the tunnel condition information includes traffic flow information and ambient humidity information; Based on the transmittance deviation, the spatial response deviation information, and the tunnel operating condition information, the contamination mode of the optical window is determined; the contamination mode includes at least uniform deposition type contamination, local shading type contamination, and interface adhesion type contamination. Based on the contamination pattern of the optical window, a calibration strategy matching the contamination pattern of the optical window is selected from a plurality of preset calibration strategies, and the selected calibration strategy is used to perform sensitivity compensation calibration on the infrared flame detection signal.

2. The method according to claim 1, characterized in that, The determination of the contamination mode of the optical window based on the transmittance deviation, the spatial response deviation information, and the tunnel condition information includes: The pollution attenuation coefficient is determined based on the transmittance deviation, the spatial response deviation information, and the tunnel operating condition information; the pollution attenuation coefficient is used to characterize the degree of pollution and pollution distribution characteristics of the optical window. The contamination mode of the optical window is determined based on the matching relationship between the contamination attenuation coefficient and the preset contamination mode determination conditions.

3. The method according to claim 2, characterized in that, The determination of the pollution attenuation coefficient based on the transmittance deviation, the spatial response deviation information, and the tunnel operating condition information includes: The overall transmission attenuation component is obtained by performing time-series statistics on the transmittance deviation in the current detection cycle and multiple consecutive historical detection cycles. The overall transmission attenuation component reflects the overall decrease in the transmittance capability of the optical window. The spatial response deviation information corresponding to multiple flame detection channels within the same detection cycle is measured for discreteness to obtain a spatial non-uniformity component; the spatial non-uniformity component reflects the degree of non-uniformity of the spatial distribution of contaminants on the optical window; A hysteresis analysis is performed on the temporal changes of at least one working parameter in the tunnel working condition information and the temporal changes of the transmittance deviation to obtain an environmental contribution component, which reflects the degree of contribution of the tunnel environment to the formation of optical window pollution. The pollution attenuation coefficient is obtained by combining the overall transmission attenuation component, the spatial non-uniformity component, and the environmental contribution component.

4. The method according to claim 2, characterized in that, The determination of the contamination mode of the optical window based on the matching relationship between the contamination attenuation coefficient and the preset contamination mode determination conditions includes: Based on the fact that the overall transmission attenuation component and the spatial non-uniformity component in the pollution attenuation coefficient satisfy the first determination relationship, the pollution mode of the optical window is determined to be uniform deposition type pollution; the first determination relationship is that the overall transmission attenuation component indicates a decrease in transmission capability, and the spatial non-uniformity component is lower than a preset non-uniformity threshold. Based on the fact that the overall transmission attenuation component and the spatial non-uniformity component in the pollution attenuation coefficient satisfy the second determination relationship, the pollution mode of the optical window is determined to be local occlusion type pollution; the second determination relationship is that the spatial non-uniformity component exceeds the preset non-uniformity threshold, and the overall transmission attenuation component indicates that there is a step change in attenuation. Based on the fact that the environmental contribution component and the overall transmission attenuation component in the pollution attenuation coefficient satisfy the third determination relationship, the pollution mode of the optical window is determined to be interface adhesion type pollution; the third determination relationship is that the environmental contribution component exceeds a preset environmental-related threshold, and the change amplitude of the overall transmission attenuation component during the continuous detection period exceeds a preset rapid change threshold.

5. The method according to claim 1, characterized in that, The step of selecting a calibration strategy that matches the contamination mode of the optical window from a set of preset calibration strategies, based on the contamination mode of the optical window, includes: Based on the contamination pattern of the optical window, the calibration requirement features corresponding to the contamination pattern are extracted; the calibration requirement features include compensation spatial distribution features and compensation temporal variation features. The calibration requirement characteristics are matched with the applicable scope characteristics of multiple preset calibration strategies; the multiple calibration strategies include at least an overall gain compensation strategy, a zone-specific compensation strategy, and a dynamic tracking compensation strategy. Based on the matching results, select a calibration strategy whose applicable scope features and the current calibration requirement features meet the preset matching conditions.

6. The method according to claim 1, characterized in that, The sensitivity compensation calibration of the infrared flame detection signal using the selected calibration strategy includes: Based on the pollution attenuation coefficient and the selected calibration strategy, determine the calibration gain parameters corresponding to each flame detection channel; Obtain the calibration constraints corresponding to the calibration strategy, and perform constraint processing on the calibration gain parameters of each channel based on the calibration constraints; the constraint processing ensures that the signal-to-noise ratio of the calibrated flame detection signal is not lower than a preset minimum signal-to-noise ratio threshold. The calibration gain parameter after constraint processing is applied to the infrared flame detection signal output by the corresponding flame detection channel to obtain the calibrated flame detection signal.

7. The method according to claim 1, characterized in that, The step of determining the transmittance characterization value of the optical window within the current detection cycle based on the active optical path detection signal, and calculating the transmittance deviation between the transmittance characterization value and the pre-calibrated reference transmittance characterization value, includes: The active optical path detection signal acquired within the current detection period is divided into stable segment signals and interference segment signals in the active optical path detection signal in the time domain. The transmittance characterization value of the optical window in the current detection cycle is determined based on the amplitude statistics of the stable segment signal. The interference segment signal is subjected to feature analysis to obtain interference type marking information; the interference type marking information is used to distinguish between transient particulate matter occlusion interference and optical window interface state change interference in the active optical path detection signal; Obtain the reference transmittance characterization value, and compare the transmittance characterization value with the reference transmittance characterization value to obtain the initial transmittance deviation; The initial transmittance deviation is effectively corrected to obtain the transmittance deviation.

8. The method according to claim 1, characterized in that, Also includes: The transmittance deviation is trend-predicted to obtain the estimated value of the transmittance attenuation of the optical window. Based on the estimated transmission attenuation and the currently determined pollution mode, the change in the effective alarm distance of the infrared flame detector after a preset time is estimated, and a warning message is output.

9. The method according to claim 1, characterized in that, Before acquiring the spatial response deviation information between the response signals of the multiple flame detection channels to the same environmental infrared radiation scene, the method further includes: The reference response signals output by the multiple flame detection channels when the optical window is clean are calibrated for consistency to obtain the calibration deviation value of each flame detection channel; When acquiring the spatial response deviation information, the response signals of each channel are preprocessed using the calibration deviation value; the preprocessing is used to eliminate inherent differences between channels; the inherent differences are the inconsistencies in the responses of multiple flame detection channels at the hardware level.

Images