Detector with cooling device and flame identification method of detector

Abstract

The invention relates to the technical field of gas turbine safety, in particular to a detector with a cooling device and a flame recognition method of the detector. The invention provides a detector with a cooling device, which comprises an outer shell, an inner shell arranged in the outer shell and a flame detector arranged in the inner shell, and an interlayer is formed between the outer shell and the inner shell; the air pipe joint extends into the inner shell from the outer part of the outer shell, the inner shell is provided with a ventilation hole, and the silencer extends into the interlayer from the outer part of the outer shell; the positions, facing the flame detector probe, of the outer shell and the inner shell are each provided with an installation hole, and light-transmitting mirror faces are arranged on the installation holes. And flame monitoring is realized through a flame identification method, and generation of false alarms is reduced. The technical problem that in the prior art, flame monitoring reliability is not high is solved.

Description

Technical Field

[0001] This invention relates to the technical field of gas turbine safety, and in particular to a detector with a cooling device and a flame identification method for the detector. Background Technology

[0002] The detectors between the gas turbines are flame probes used to detect whether a flame has occurred. However, the detector's detection is affected by various factors, including the detector itself and the system's detection capabilities.

[0003] The highest temperature that can be measured by commonly used air purifiers is 125 degrees Celsius. However, the operating temperature in the turbine compartment of a gas turbine is around 140 degrees Celsius. During gas turbine operation, the fuel ring pipes, combustion chamber, and outer cylinders of the turbine are not allowed to be wrapped with insulation. The temperature inside the turbine compartment can only be controlled by two ventilation fans, and the adjustment methods are limited. It is impossible to continuously control the temperature inside the gas turbine compartment below 125 degrees Celsius, which could cause detectors to malfunction or even burn out due to high temperatures, making it impossible to effectively monitor the flame situation in the turbine compartment. Therefore, an external cooling medium is needed to maintain the temperature.

[0004] With the development of industrial safety technology, flame monitoring methods have become more efficient and intelligent. In high-risk environments such as gas turbine workshops, traditional flame detection methods mainly rely on a single sensor to detect flame or smoke signals, which can achieve early warning to a certain extent. However, in complex environments, single sensors are easily affected by environmental interference, leading to high false alarm rates or prolonged response times. While delayed alarms can cause significant harm, false alarms causing shutdowns can also result in substantial losses. To improve the accuracy of flame detection, a common improvement measure is to combine data from multiple sensors for comprehensive analysis, such as using infrared and visible light detectors together. However, this approach still has certain limitations, especially when the detection capability in the violet band is weak, making it difficult to comprehensively capture the characteristic signals of the flame, thus affecting the timeliness and reliability of identification.

[0005] Therefore, it is necessary to have a detector with a cooling device that can identify flames based on a violet-infrared composite detector. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the prior art and provide a detector with a cooling device and a flame identification method for the detector, which aims to solve the technical problem of low reliability of flame monitoring in the prior art.

[0007] To achieve the above objectives, in a first aspect, the present invention proposes a detector with a cooling device, comprising a housing, an inner shell disposed within the housing, and a flame detector disposed within the inner shell, wherein a sandwich layer is formed between the housing and the inner shell; a gas pipe connector extends from the outside of the housing into the inner shell, the inner shell is provided with a ventilation hole, and a silencer extends from the outside of the housing into the sandwich layer; each of the housing and the inner shell is provided with a mounting hole facing the flame detector probe position, and a light-transmitting mirror is provided on the mounting hole.

[0008] Preferably, both the outer shell and the inner shell are rectangular shells, and the inner shell is provided with an air intake guide plate, which is located at the air outlet of the air pipe connector; the air exchange hole is located on one side of the corresponding surface of the air pipe connector; the air inlet of the muffler and the air pipe connector are on the same side.

[0009] Preferably, a cleaning assembly is also included for cleaning the light-transmitting mirror surface; the cleaning assembly includes a cleaning cylinder, which includes a cylinder body, an air intake pipe, and a rotating rod. The air intake pipe is located on the side of the cylinder body near the air intake guide plate, and the air intake pipe is flared. The cylinder body is a hollow cylinder, and the cylinder wall is provided with evenly distributed guide holes. The rotating rod is fixed to the closed end of the cylinder body and installed on the same center line as the cylinder body. The rotating rod is rotatably mounted on a lifting block. The lifting block is mounted on a lead screw, which is driven by a motor. The rotating rod is connected to a first gear, and a rack is fixedly connected to the inner shell. The first gear meshes with the rack, and the rack is located in the inner shell.

[0010] Preferably, the device further includes a mounting cover, a filter screen, and a shaking assembly disposed on the outer casing; the air pipe connector is disposed inside the mounting cover; the filter screen is disposed inside the mounting cover; the shaking assembly, disposed inside the mounting cover, is used to shake the filter screen; the shaking assembly includes a shaking spring, the filter screen is slidably mounted inside the mounting cover, and a connecting block is fixedly connected inside the mounting cover, and a sliding rod is fixedly connected to the filter screen, the sliding rod passes through the connecting block, and the sliding rod is slidably mounted with the connecting block, the shaking spring is sleeved on the sliding rod, and the shaking spring is fixedly connected between the connecting block and the filter screen, and an abutment assembly is provided on the mounting cover.

[0011] Preferably, the abutting component includes an abutting rod, a connecting rod is rotatably mounted on the mounting cover, and the abutting rod is fixed to the connecting rod, corresponding to the filter screen. A driving component is provided between the connecting rod and the lead screw. The driving component includes a connecting shaft, which is rotatably mounted on the housing. A driven bevel gear is fixedly connected to the connecting shaft, and a driving bevel gear is fixedly connected to the lead screw. The driving bevel gear meshes with the driven bevel gear, and a linkage component is provided between the connecting shaft and the connecting rod. The linkage component includes a connecting frame, a connecting strip is rotatably mounted on the connecting frame, a first worm is fixedly connected to the connecting shaft, a first worm wheel is fixedly connected to the connecting strip, and the first worm wheel meshes with the first worm. A second worm is fixedly connected to the connecting strip, and a second worm wheel is fixedly connected to the connecting rod, and the second worm meshes with the second worm wheel.

[0012] To achieve the above objectives, in a second aspect, the present invention proposes a flame identification method using a detector, comprising: The system acquires ultraviolet light, infrared light, and visible light signals in the gas turbine inter-turbine environment, wherein the ultraviolet light signal is acquired by an ultraviolet light detector, the infrared signal is acquired by an infrared detector, and the visible light signal is acquired by a visible light detector. Based on a preset multimodal signal fusion algorithm, the violet light signal, infrared signal and visible light signal are initially fused to generate an initial feature matrix. The multimodal signal fusion algorithm uses a weighted superposition method to normalize the intensity of each signal and combines time series analysis to extract dynamic change trends. Based on the initial feature matrix, a preset flame feature classification model is invoked to determine whether the current environmental state meets the flame warning conditions. The flame feature classification model is built based on a deep learning framework and trained using a historical flame dataset to output a flame probability value. When the conditions for fire warning are met, a fire alarm signal is generated and sent to the monitoring center via the communication module.

[0013] Preferably, the multimodal signal fusion algorithm includes the following steps: The violet light signal, infrared signal, and visible light signal were normalized respectively to obtain the normalized violet light signal intensity, infrared signal intensity, and visible light signal intensity; Sliding window analysis is performed on the normalized signal according to the preset time window to extract the dynamic trend of the signal. The normalized signal strength and the dynamic change trend are weighted and superimposed to generate the initial feature matrix; The flame feature classification model is trained through the following steps: Obtain a labeled historical flame dataset, wherein the historical flame dataset includes violet light signals, infrared signals and visible light signals when the flame occurred and their corresponding flame labels; Supervised learning is performed on the historical flame dataset based on a convolutional neural network structure to optimize the parameters of the flame feature classification model; The trained flame feature classification model is deployed to the central processing unit.

[0014] Preferably, the following steps are also included: In the distributed monitoring system, ultraviolet light signals, infrared signals and visible light signals are received from multiple monitoring nodes in the gas turbine turbine room. Each monitoring node is equipped with an ultraviolet light detector, an infrared detector and a visible light detector. Based on a preset multimodal signal fusion algorithm, the signals of each monitoring node are processed independently to generate the initial feature matrix of each node. Based on the initial feature matrix of each node, the preset flame feature classification model is called to determine the flame probability value of each node, and a flame risk heat map is generated by combining the spatial distribution information. The multimodal signal fusion algorithm performs hierarchical and progressive processing on the signals from each monitoring node, including the following steps: The ultraviolet light signal, infrared signal and visible light signal of a single monitoring node are locally fused to generate the local feature matrix of that node; The local feature matrices of multiple monitoring nodes are globally fused to generate a global feature matrix; The flame feature classification model is invoked based on the global feature matrix to determine the flame probability value of each node.

[0015] Preferably, the generation of the flame risk heat map includes the following steps: The fire risk level of each area is determined based on the fire probability value of each monitoring node. In the flame risk heat map, the flame risk level is represented by the intensity of the color, with high-risk areas marked in red and low-risk areas marked in green. High-risk areas are identified, and corresponding fire alarm signals are generated; It also includes the following steps: Multiple monitoring nodes are arranged in the gas turbine room, and each monitoring node is equipped with an ultraviolet light detector, an infrared detector and a visible light detector; The signals from each monitoring node are transmitted to the central processing unit via optical fiber. The signals from each monitoring node are processed in the central processing unit, a flame alarm signal is generated, and the signal is sent to the monitoring center via the communication module.

[0016] To achieve the above objectives, in a third aspect, the present invention proposes a flame identification method for a distributed monitoring system, which uses the aforementioned detector to monitor flames, including: It receives ultraviolet, infrared, and visible light signals from multiple monitoring nodes in the gas turbine turbine room. Each monitoring node is equipped with an ultraviolet detector, an infrared detector, and a visible light detector, and each detector is connected to the central processing unit via optical fiber. Based on a preset multimodal signal fusion algorithm, the signals of each monitoring node are processed independently to generate an initial feature matrix for each node. The multimodal signal fusion algorithm adopts a hierarchical and progressive approach, first performing local fusion on the signals of a single node, and then performing global fusion on the local fusion results of multiple nodes. Based on the initial feature matrix of each node, the preset flame feature classification model is called to determine the flame probability value of each node, and a flame risk heat map is generated by combining the spatial distribution information. High-risk areas are identified in the flame risk heat map, and corresponding flame alarm signals are generated and sent to the monitoring center via a wireless communication module. The spatial distribution information includes the location coordinates of each monitoring node, and the flame risk heat map is generated through the following steps: Based on the flame probability value and location coordinates of each monitoring node, the flame risk level of each area is calculated. In the flame risk heat map, the flame risk level is represented by the intensity of the color, with high-risk areas marked in red and low-risk areas marked in green. High-risk areas are identified, and corresponding fire alarm signals are generated.

[0017] To achieve the above objectives, in a third aspect, the present invention provides a data processing apparatus, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the above-mentioned gas turbine inter-turbine flame identification method based on ultraviolet-infrared composite detector and the gas turbine inter-turbine flame identification method applied to a distributed monitoring system.

[0018] To achieve the above objectives, in a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the above-described gas turbine inter-turbine flame identification method based on an ultraviolet-infrared composite detector and the gas turbine inter-turbine flame identification method applied to a distributed monitoring system.

[0019] To achieve the above objectives, in a fifth aspect, the present invention provides a computer program product, including a computer program / instruction that, when executed by a processor, implements the steps of the above-described gas turbine inter-turbine flame identification method based on an ultraviolet-infrared composite detector and the gas turbine inter-turbine flame identification method applied to a distributed monitoring system.

[0020] Compared with the prior art, the beneficial effects of the detector with cooling device and the flame recognition method of the detector provided by the present invention are as follows: 1. It realizes the acquisition of ultraviolet light signals, infrared signals and visible light signals in the gas turbine turbine environment, generates an initial feature matrix based on the preset multimodal signal fusion algorithm, calls the preset flame feature classification model to determine the flame early warning conditions, and generates a flame alarm signal to send to the monitoring center. The monitoring center then takes subsequent relief measures, such as notifying personnel to evacuate.

[0021] 2. By introducing a combined configuration of violet, infrared, and visible light detectors, this embodiment of the invention can comprehensively capture multiple characteristic signals in the early stages of a flame, especially signals in the violet band, thereby improving the sensitivity and accuracy of flame identification. The multimodal signal fusion algorithm, by normalizing the intensity of different signals and combining it with time series analysis, effectively reduces the false alarm rate caused by environmental interference from a single sensor. Furthermore, the flame feature classification model, built on a deep learning framework, can continuously optimize flame identification capabilities through learning from historical data, further improving detection efficiency and reliability. In the distributed monitoring system, a flame risk heatmap is generated using spatial distribution information and compared with the original model to further improve accuracy, providing precise support for emergency response.

[0022] 3. A cooling device has been added to each detector, making the detector less susceptible to high temperatures, resulting in better stability and higher reliability during actual use.

[0023] The features and advantages of the present invention will be described in detail through embodiments and in conjunction with the accompanying drawings. Attached Figure Description

[0024] Figure 1 This is an overall schematic diagram of the detector with a cooling device.

[0025] Figure 2 This is a schematic diagram of the detector's internal three-dimensional structure.

[0026] Figure 3 This is a cross-sectional view of the detector.

[0027] Figure 4 This is a schematic diagram of the three-dimensional structure of the outer shell.

[0028] Figure 5 It is a cross-sectional view of the outer shell and the inner shell.

[0029] Figure 6 This is a cross-sectional view of the cleaning cylinder.

[0030] Figure 7 This is a schematic diagram of the lifting block.

[0031] Figure 8 yes Figure 4 Enlarged view of part C.

[0032] Figure 9 yes Figure 5 Enlarged view of part D.

[0033] Figure 10 This is a left view of the cleaning cylinder.

[0034] Figure 11 yes Figure 5 Enlarged view of part F.

[0035] Figure 12 This is a flowchart of the method of the present invention.

[0036] Figure 13 This is a flowchart illustrating the specific processing steps of the multimodal signal fusion algorithm of this invention.

[0037] Figure 14 This is a flowchart illustrating the training and application process of the flame feature classification model of this invention.

[0038] Figure 15 This is a diagram of the distributed monitoring system architecture of the present invention.

[0039] Figure 16 This is a flowchart of the flame risk heat map generation process of the present invention.

[0040] Figure 17 This is a structural block diagram of the data processing device of the present invention.

[0041] in: 1-Outer shell; 2-Connecting rod; 3-Mounting cover; 4-Filter screen; 5-Muffler; 6-Mounting bracket; 7-Mounting plate; 8-Mounting hole; 9-Inlet guide plate; 10-Flame detector; 11-Servo motor; 12-Ventilation hole; 13-Inner shell; 14-Lifting block; 15-Lead screw; 16-Driving bevel gear; 17-Driven bevel gear; 18-Connecting shaft; 19-Probe body; 20-Rack; 21-First gear; 22-Cleaning brush; 23-Abutting rod; 24-Threaded hole; 25-First worm gear; 26-Connecting strip; 27-Second worm wheel; 28-Second worm gear; 29-Air pipe connector; 30-Connecting bracket; 31-First worm wheel; 32-Connecting block; 33-Slide rod; 34-Shaking spring; 35-Rotating rod. Detailed Implementation

[0042] Embodiments of this disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of this disclosure are shown in the drawings, it should be understood that this disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this disclosure. It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure.

[0043] It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings. Unless otherwise specified, the embodiments and features described in this disclosure can be combined with each other.

[0044] It should be noted that the concepts of "first" and "second" mentioned in this disclosure are used only to distinguish different devices, modules or units, and are not used to limit the order of functions performed by these devices, modules or units or their interdependencies.

[0045] It should be noted that the terms "a" and "a plurality of" used in this disclosure are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0046] This disclosure will now be described in detail with reference to the accompanying drawings and embodiments.

[0047] Reference Figure 1-11 A cooling device for an ultraviolet (UV) probe between gas turbines includes: an outer shell 1, an inner shell 13, a mounting bracket 6, a silencer 5, a mounting hole 8, a mounting plate 7, a probe body 19, a gas pipe connector 29, a cleaning assembly, a mounting cover 3, a filter screen 4, and a vibration assembly. The inner shell 13 is disposed inside the outer shell 1. The mounting bracket 6 is used to fix the inner shell 13 and the outer shell 1. A flame detector 10 is disposed inside the inner shell 13. A threaded hole 24 is provided on the mounting bracket 6. The silencer 5 is disposed between the inner shell 13 and the outer shell 1, and extends to the outside. The mounting hole 8 is formed on the inner shell 13 and the outer shell 1. The mounting plate 7 is disposed between the inner shell 13 and the outer shell 1. The probe body 19 is disposed on the mounting plate 7. The gas pipe connector 29 is disposed between the inner shell 13 and the outer shell 1, and extends to the outside. The cleaning assembly is disposed on the outer shell 1 and the outer shell 1, and is used to clean the lens of the probe body 19. The mounting cover 3 is disposed on the outer shell 1, and the gas pipe connector 29 is disposed inside the mounting cover 3. The filter screen 4 is disposed inside the mounting cover 3. The shaking component is disposed inside the mounting cover 3 for shaking the filter screen 4.

[0048] The cleaning assembly includes a cleaning brush 22, a lifting block 14 slidably mounted on the inner shell 13, rotating rods 35 fixedly connected to both ends of the cleaning brush 22, one of the rotating rods 35 being rotatably connected to the lifting block 14, a lead screw 15 being rotatably mounted on the inner shell 13 and threadedly connected to the lifting block 14, and a servo motor 11 fixedly mounted on the outer shell 1, with the output shaft of the servo motor 11 being fixedly connected to the lead screw 15.

[0049] Another rotating rod 35 is connected to a first gear 21, and a rack 20 is fixedly connected to the inner shell 13. The first gear 21 meshes with the rack 20.

[0050] The shaking assembly includes a shaking spring 34, a filter screen 4 is slidably installed inside a mounting cover 3, and a connecting block 32 is fixedly connected inside the mounting cover 3. A sliding rod 33 is fixedly connected to the filter screen 4, the sliding rod 33 passes through the connecting block 32, and the sliding rod 33 is slidably installed with the connecting block 32. The shaking spring 34 is sleeved on the sliding rod 33 and is fixedly connected between the connecting block 32 and the filter screen 4. An abutment assembly is provided on the mounting cover 3.

[0051] The abutment component includes an abutment rod 23, a connecting rod 2 is rotatably mounted on the mounting cover 3, the abutment rod 23 is fixed to the connecting rod 2, the abutment rod 23 corresponds to the filter screen 4, and a drive component is provided between the connecting rod 2 and the lead screw 15.

[0052] The drive assembly includes a connecting shaft 18, which is rotatably mounted on the housing 1. A driven bevel gear 17 is fixedly connected to the connecting shaft 18, and a driving bevel gear 16 is fixedly connected to the lead screw 15. The driving bevel gear 16 meshes with the driven bevel gear 17, and a linkage assembly is provided between the connecting shaft 18 and the connecting rod 2.

[0053] The linkage assembly includes a connecting frame 30, a connecting bar 26 rotatably mounted on the connecting frame 30, a first worm 25 fixedly connected to the connecting shaft 18, a first worm wheel 31 fixedly connected to the connecting bar 26, the first worm wheel 31 meshing with the first worm 25, a second worm 28 fixedly connected to the connecting bar 26, and a second worm wheel 27 fixedly connected to the connecting rod 2, the second worm 28 meshing with the second worm wheel 27.

[0054] Two sets of vibration springs 34, slide rods 33, and connecting blocks 32 are provided and arranged symmetrically. Multiple ventilation holes 12 are evenly opened on the inner shell 13. An air intake guide plate 9 is provided inside the inner shell 13, and the air intake guide plate 9 corresponds to the air pipe connector 29.

[0055] Working process: Compressed air for the gas turbine area is introduced into the inner shell 13 of the probe cooling device. Compressed air for the probe body 19 is introduced into the interlayer of the inner shell 13 through the vent 12, forming a heat insulation layer. Simultaneously, the probe body 19 is cleaned, and the radiant heat from the outer shell 1 is cooled by the flow of compressed air. The cooled compressed air is then discharged through the silencer 5. The air inlet guide plate 9 is installed at the air inlet of the compressed air pipe connector 29. Compressed air at the inlet is directly introduced into the inner shell 13, and the interlayer of the inner shell 13 is sealed to prevent the formation of a compressed air bypass and avoid reducing the cooling effect. The inner and outer vent holes in this device are opposite to the air inlet, allowing the compressed air used for cooling to cool the probe body 19 to the maximum extent. The exhaust silencer 5 is installed opposite the vent holes of the inner shell 13, ensuring that the interlayer of the inner shell 13 is filled with compressed air, achieving optimal heat insulation and cooling effects. Simultaneously, by setting up the filter screen 4, dust can be prevented at the air inlet of the air pipe connector 29. The servo motor 11 can be started, so that the output shaft of the servo motor 11 drives the lead screw 15 to rotate, thereby causing the lifting block 14 to move up and down, which in turn causes the cleaning brush 22 to move up and down. Since the gear and rack 20 mesh, the cleaning brush 22 rotates back and forth and cleans the lens of the probe body 19, thereby further improving the cleanliness. At the same time, when the active bevel gear 16 rotates, the driven bevel gear 17 rotates, which causes the connecting shaft 18 to rotate, which causes the first worm 25 to rotate, which causes the first worm wheel 31 to rotate, which causes the connecting bar 26 to rotate, which causes the second worm 28 to rotate, which causes the second worm wheel 27 to rotate, which causes the connecting rod 2 to rotate, which causes the abutment rod 23 to rotate. Under the action of the shaking spring 34, the filter screen 4 shakes back and forth, thereby shaking off the dust on the filter screen 4, thus preventing dust from adhering to the filter screen 4 and affecting the air intake effect of the filter screen 4.

[0056] The above description is merely a selection of preferred embodiments of this disclosure and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in the embodiments of this disclosure is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of this disclosure.

Claims

1. A detector with a cooling device, characterized in that: It includes an outer shell, an inner shell disposed within the outer shell, and a flame detector disposed within the inner shell, wherein a sandwich is formed between the outer shell and the inner shell; The air pipe connector extends from the outside of the outer shell into the inner shell, and the inner shell is provided with a ventilation hole. The silencer extends from the outside of the outer shell into the interlayer. The outer shell and the inner shell are each provided with a mounting hole facing the flame detector probe, and the mounting hole is provided with a light-transmitting mirror.

2. A detector with a cooling device as described in claim 1, characterized in that: Both the outer shell and the inner shell are rectangular shells. An air inlet guide plate is provided on the inner shell and is located at the air outlet of the air pipe connector. The air exchange hole is located on one side of the corresponding surface of the air pipe connector. The air inlet of the muffler and the air pipe connector are on the same side.

3. A detector with a cooling device as described in claim 1, characterized in that: It also includes a cleaning assembly for cleaning the light-transmitting mirror surface; the cleaning assembly includes a cleaning cylinder, which includes a cylinder body, an air intake pipe, and a rotating rod. The air intake pipe is located on the side of the cylinder body near the air intake guide plate, and the air intake pipe is flared. The cylinder body is a hollow cylinder, and the cylinder wall is provided with evenly distributed guide holes. The rotating rod is fixed to the closed end of the cylinder body and installed on the same center line as the cylinder body. The rotating rod is rotatably mounted on a lifting block. The lifting block is mounted on a lead screw, which is driven by a motor. The rotating rod is connected to a first gear, and a rack is fixedly connected to the inner shell. The first gear meshes with the rack, and the rack is located in the inner shell.

4. A detector with a cooling device as described in claim 3, characterized in that: It also includes a mounting cover, a filter screen, and a shaking assembly mounted on the outer casing; an air pipe connector is located inside the mounting cover; the filter screen is located inside the mounting cover; the shaking assembly, located inside the mounting cover, is used to shake the filter screen; the shaking assembly includes a shaking spring, the filter screen is slidably mounted inside the mounting cover, a connecting block is fixedly connected inside the mounting cover, and a sliding rod is fixedly connected to the filter screen, the sliding rod passes through the connecting block and is slidably mounted with the connecting block, the shaking spring is sleeved on the sliding rod and is fixedly connected between the connecting block and the filter screen, and an abutment assembly is provided on the mounting cover.

5. A detector with a cooling device as described in claim 4, characterized in that: The abutment assembly includes an abutment rod, a connecting rod is rotatably mounted on the mounting cover, and the abutment rod is fixed to the connecting rod, corresponding to the filter screen. A drive assembly is provided between the connecting rod and the lead screw. The drive assembly includes a connecting shaft, which is rotatably mounted on the housing. A driven bevel gear is fixedly connected to the connecting shaft, and a driving bevel gear is fixedly connected to the lead screw. The driving bevel gear meshes with the driven bevel gear, and a linkage assembly is provided between the connecting shaft and the connecting rod. The linkage assembly includes a connecting frame, a connecting strip is rotatably mounted on the connecting frame, a first worm is fixedly connected to the connecting shaft, a first worm wheel is fixedly connected to the connecting strip, and the first worm wheel meshes with the first worm. A second worm is fixedly connected to the connecting strip, and a second worm wheel is fixedly connected to the connecting rod, and the second worm meshes with the second worm wheel.

6. A flame identification method for a detector, characterized in that: Monitoring flames using any one of the detectors described in 1-4, including: The system acquires ultraviolet light, infrared light, and visible light signals in the gas turbine inter-turbine environment, wherein the ultraviolet light signal is acquired by an ultraviolet light detector, the infrared signal is acquired by an infrared detector, and the visible light signal is acquired by a visible light detector. Based on a preset multimodal signal fusion algorithm, the violet light signal, infrared signal and visible light signal are initially fused to generate an initial feature matrix. The multimodal signal fusion algorithm uses a weighted superposition method to normalize the intensity of each signal and combines time series analysis to extract dynamic change trends. Based on the initial feature matrix, a preset flame feature classification model is invoked to determine whether the current environmental state meets the flame warning conditions. The flame feature classification model is built based on a deep learning framework and trained using a historical flame dataset to output a flame probability value. When the conditions for fire warning are met, a fire alarm signal is generated and sent to the monitoring center via the communication module.

7. The flame identification method for a detector as described in claim 6, characterized in that: The multimodal signal fusion algorithm includes the following steps: The violet light signal, infrared signal, and visible light signal were normalized respectively to obtain the normalized violet light signal intensity, infrared signal intensity, and visible light signal intensity; Sliding window analysis is performed on the normalized signal according to the preset time window to extract the dynamic trend of the signal. The normalized signal strength and the dynamic change trend are weighted and superimposed to generate the initial feature matrix; The flame feature classification model is trained through the following steps: Obtain a labeled historical flame dataset, wherein the historical flame dataset includes violet light signals, infrared signals and visible light signals when the flame occurred and their corresponding flame labels; Supervised learning is performed on the historical flame dataset based on a convolutional neural network structure to optimize the parameters of the flame feature classification model; The trained flame feature classification model is deployed to the central processing unit.

8. The flame identification method for a detector as described in claim 6, characterized in that: It also includes the following steps: In the distributed monitoring system, ultraviolet light signals, infrared signals and visible light signals are received from multiple monitoring nodes in the gas turbine turbine room. Each monitoring node is equipped with an ultraviolet light detector, an infrared detector and a visible light detector. Based on a preset multimodal signal fusion algorithm, the signals of each monitoring node are processed independently to generate the initial feature matrix of each node. Based on the initial feature matrix of each node, the preset flame feature classification model is called to determine the flame probability value of each node, and a flame risk heat map is generated by combining the spatial distribution information. The multimodal signal fusion algorithm performs hierarchical and progressive processing on the signals from each monitoring node, including the following steps: The ultraviolet light signal, infrared signal and visible light signal of a single monitoring node are locally fused to generate the local feature matrix of that node; The local feature matrices of multiple monitoring nodes are globally fused to generate a global feature matrix; The flame feature classification model is invoked based on the global feature matrix to determine the flame probability value of each node.

9. A flame identification method for a detector as described in claim 8, characterized in that: The generation of the flame risk heat map includes the following steps: The fire risk level of each area is determined based on the fire probability value of each monitoring node. In the flame risk heat map, the flame risk level is represented by the intensity of the color, with high-risk areas marked in red and low-risk areas marked in green. High-risk areas are identified, and corresponding fire alarm signals are generated; It also includes the following steps: Multiple monitoring nodes are arranged in the gas turbine room, and each monitoring node is equipped with an ultraviolet light detector, an infrared detector and a visible light detector; The signals from each monitoring node are transmitted to the central processing unit via optical fiber. The signals from each monitoring node are processed in the central processing unit, a flame alarm signal is generated, and the signal is sent to the monitoring center via the communication module.

10. A flame identification method applied to a distributed monitoring system, characterized in that, Monitoring flames using any one of the detectors described in 1-4, including: It receives ultraviolet, infrared, and visible light signals from multiple monitoring nodes in the gas turbine turbine room. Each monitoring node is equipped with an ultraviolet detector, an infrared detector, and a visible light detector, and each detector is connected to the central processing unit via optical fiber. Based on a preset multimodal signal fusion algorithm, the signals of each monitoring node are processed independently to generate an initial feature matrix for each node. The multimodal signal fusion algorithm adopts a hierarchical and progressive approach, first performing local fusion on the signals of a single node, and then performing global fusion on the local fusion results of multiple nodes. Based on the initial feature matrix of each node, the preset flame feature classification model is called to determine the flame probability value of each node, and a flame risk heat map is generated by combining the spatial distribution information. High-risk areas are identified in the flame risk heat map, and corresponding flame alarm signals are generated and sent to the monitoring center via a wireless communication module. The spatial distribution information includes the location coordinates of each monitoring node, and the flame risk heat map is generated through the following steps: Based on the flame probability value and location coordinates of each monitoring node, the flame risk level of each area is calculated. In the flame risk heat map, the flame risk level is represented by the intensity of the color, with high-risk areas marked in red and low-risk areas marked in green. High-risk areas are identified, and corresponding fire alarm signals are generated.

Images