Operation method of underground coal mine fire area information detection emergency robot

Abstract

The invention discloses an operation method of an underground coal mine fire area information detection emergency robot, and relates to the technical field of coal mine emergency rescue, and the operation method comprises the following steps: the robot determines fire area arrival conditions, and matches an entry mode; the robot plans a fire area entering route based on an entering mode in combination with a Dijkstra algorithm, and arrives at a specified detection position; the specified detection position is obtained based on terahertz imaging and multi-line laser radar fusion imaging positioning; the robot collects data at the detection position, judges the explosion risk and sends an alarm signal; the ground control center receives the alarm signal to carry out emergency response, fire source determination and fire extinguishing work, and the problems that an existing robot cannot quickly enter a fire area in the emergency disposal operation process, cannot pass nearby a working face support without obstacles, and can replace rescue and workers to go deep into the fire area to carry out fire point detection and explosiveness prediction are solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of coal mine emergency rescue, and particularly discloses a working method of an emergency robot for detecting information of a fire zone in a coal mine. BACKGROUND

[0002] Due to coal spontaneous combustion, equipment failure or illegal operation, etc., once the fire cannot be effectively controlled after a fire accident occurs in the mine, the conventional method is to close the working face. When the closure wall is constructed, the entire area will be disconnected from power and signal transmission will be blocked, and there is a risk of explosion due to accumulation of harmful gases in the high-temperature area. After the working face is closed, it is necessary to use injection of inert gas or pressure injection of filling materials for targeted management of the fire zone, and the judgment of the fire point is extremely important in this process. After the fire zone management is completed and the unsealing standard specified in the Coal Mine Safety Regulations is reached, the working personnel need to perform the working face unsealing operation, and in this process, the gas concentration in the closed wall decreases due to ventilation, fresh air enters to provide sufficient O2, and once the fire zone reignites, it is easy to cause explosion and other secondary disasters. At present, gas explosions caused by the closure of the fire zone, fire extinguishing and unsealing process account for more than 70% of the number of deaths in gas explosions. Detecting gas information in the fire zone, determining the high-temperature fire source point and preventing gas explosions during emergency disposal are the urgent safety problems to be solved in the event of coal mine fires and other disasters. Among them, the "3·11" gas explosion accident in Xieqiao Coal Mine of Anhui Huaihe Energy Holding Group caused 9 deaths, 15 injuries and direct economic losses of 1637.73 million yuan. The main reason for this accident is that the working personnel did not have relevant equipment to detect and transmit the information such as gas composition and concentration and fire point position in the closed area in a timely manner during the implementation of the closure and fire extinguishing process of the working face, and the ground command center could not dynamically master the gas explosion, combustion and other conditions of the coal mining face in real time, could not make explosive prediction, and therefore could not notify the underground operating personnel to retreat to the safe area in time before the gas explosion, resulting in extremely heart-wrenching life and death.

[0003] In summary, whether the working face is closed, unsealed or extinguished, it is urgent to determine the gas information and high temperature point position in the fire area to implement targeted ground or underground work and prevent the occurrence of secondary accidents such as explosion. In recent years, a coal mine tracked inspection robot KRXD51.2C(A) developed by Shandong Guoxing Intelligent Technology Co., Ltd. has many functions of detecting toxic and flammable gases and dust, but it cannot detect fire points and give explosion warning; it carries a signal relay station for wireless communication, but uses a sliding mechanism to lower it, which cannot guarantee accurate sliding and stable communication of the base station; and due to its large size, it is difficult to pass through the working face support area without obstacles, and cannot adapt to the special environment during the "closure-treatment-unsealing" operation of the fire area. The KQR48 mine detection robot developed by CMC Intelligence Equipment Co., Ltd. and the ZR mine detection robot developed by China University of Mining and Technology both use optical fiber communication and have a large size, so their detection functions are limited. The coal mine emergency rescue robots developed so far are generally large in size and mostly use optical fiber communication, which cannot quickly enter the fire area during emergency disposal operations and pass through the working face support area without obstacles, replacing rescue and workers to enter the fire area for fire point detection and explosion prediction. SUMMARY

[0004] In order to overcome the above-mentioned defects of the prior art, the embodiments of the present application provide a working method of a coal mine underground fire area information detection emergency robot, which quickly enters the fire area in the form of direct entry when the fire area is not closed, and then detects the temperature, gas composition and concentration in real time, predicts the explosion risk and fire point position of the area, and prevents gas explosion, so as to solve the problem that the existing robots cannot obtain the gas concentration, high temperature point position and explosion risk warning in real time during the closure, treatment and unsealing process of the fire area.

[0005] To achieve the above object, the present application provides the following technical scheme: A working method of a coal mine underground fire area information detection emergency robot, comprising the following steps: the robot determines the fire area accessibility condition and matches the entering mode; the fire area accessibility condition is determined based on the environment perception layer, and the entering mode is matched; the entering route into the fire area is planned based on the entering mode combined with Dijkstra algorithm, and the specified detection position is reached, which is obtained based on terahertz imaging and multi-line laser radar fusion imaging positioning; data collection is performed at the detection position, explosion risk is judged, and an alarm signal is sent; the alarm signal is received by the control center for emergency response and positioning.

[0006] In a preferred embodiment, the environment perception layer includes obtaining a fire temperature gradient map by a dual-spectrum thermal imager, detecting obstacle distribution by a millimeter wave radar, and extracting channel attributes by analyzing a building information modeling (BIM) model.

[0007] In a preferred embodiment, the determining the fire area access condition and matching the access mode specifically comprises: obtaining real-time sensing data through the environment perception layer, the sensing data at least including thermal imaging data collected by the dual-spectrum thermal imager and gas concentration data collected by the gas sensor; when it is detected that there is high temperature behind the continuous solid obstacle and the carbon monoxide concentration rises, the smoke concentration rises, and the oxygen concentration decreases as shown by the gas sensor, it is determined that the fire area is closed; when it is detected that there is an open gap and there is no high temperature anomaly in the thermal imaging data and no harmful gas exceeds the standard in the gas concentration data, it is determined that the fire area is not closed; if the fire area is closed, the closed fire area access program is executed, and if the fire area is not closed, the robot directly enters the fire area after the environmental indicators continuously meet the safety threshold.

[0008] In a preferred embodiment, if the fire area is closed, the closed fire area access program is executed, and the closed fire area access program further comprises the execution of the wall breaking program, specifically comprising: scanning the surrounding roadway of the fire area through the laser radar and the geological radar, synchronously calling the coal mine geological data and the roadway drawing to generate a first data set containing the obstacle position, internal structure and material characteristics; detecting the temperature and gas concentration of the surrounding environment of the wall breaking in real time, and dynamically avoiding the dangerous area whose temperature or gas concentration exceeds the preset threshold; based on the material characteristics and structure characteristics in the first data set, the hardness evaluation model is used to calculate the hardness grade of the fire area wall; if the hardness grade is less than or equal to a threshold H, a hydraulic breaking hammer is used to break the obstacle; if the hardness grade is greater than the threshold H, water jet cutting and laser cutting are used for collaborative operation to break the hard structure.

[0009] In a preferred embodiment, the specified detection position is obtained based on terahertz imaging and multi-line laser radar fusion imaging positioning, specifically comprising: initializing the mounted terahertz imaging device and multi-line laser radar device; performing a penetrating fault scan on the roadway rock wall and coal body through the terahertz imaging device to obtain thermal radiation intensity distribution data in a smoke dust environment, identifying temperature anomaly areas and outputting two-dimensional thermal spot coordinates; constructing a three-dimensional point cloud map of the roadway through the multi-line laser radar, extracting the spatial coordinates of the roadway wall and obstacles, and calculating the robot pose data in real time; based on the robot pose data, a coordinate conversion matrix is established through the hand-eye calibration method to map the two-dimensional thermal spot coordinates to the three-dimensional point cloud space to generate a fire source candidate area with a confidence score; the fire source candidate area is screened according to the confidence score to determine the final detection position.

[0010] In a preferred embodiment, the entering mode-based Dijkstra algorithm is used to plan the entering route into the fire zone, specifically: based on the three-dimensional map of the fire zone roadway, obstacle distribution and gas concentration anomaly area data obtained by the information detection system, combined with the historical first data set, the theoretically optimal route is planned by Dijkstra algorithm; during the robot's travel, the laser radar constructs a point cloud map in real time and matches the environmental features, the vision sensor identifies fixed obstacles and dynamic risk targets, and the inertial measurement unit monitors the real-time attitude of the robot; when an obstacle is detected, a dynamic window method and an artificial potential field method fusion algorithm are used to locally correct the theoretically optimal route to generate an obstacle avoidance path as the second route; if a high-temperature area or a harmful gas exceeding area is detected, the second route is forcibly adjusted to generate the final entering route into the fire zone.

[0011] In a preferred embodiment, the theoretically optimal route is planned by Dijkstra algorithm, specifically: (1) the three-dimensional map of the fire zone roadway is extracted from the information detection system and the first data, the nodes, paths, obstacles and dangerous areas are marked, and a network graph containing path length and danger degree weight is constructed; (2) the current position of the robot is taken as the starting node, the detection position is specified as the target node, and the cumulative weight value of the starting node is initialized; (3) starting from the starting node, the search is diffused outward, and the cumulative weight value of each adjacent node of the current node is calculated; (4) in the network graph, the search is diffused outward from the starting node; (5) if the cumulative weight value is less than the recorded cumulative weight value of the node, the cumulative weight value and the predecessor node are updated, and steps (1)-(4) are repeated until the target node is searched; (6) the predecessor node sequence of the target node is traced back to the starting node to generate the theoretically optimal route.

[0012] In a preferred embodiment, the roadway rock wall and coal body are penetrated by a terahertz imaging device for tomographic scanning to obtain thermal radiation intensity distribution data in a smoke dust environment, specifically: the environmental thermal radiation signal in the absence of heat sources is obtained before scanning as background noise; during scanning, the actual collected thermal radiation data is differentially processed with the background noise data to eliminate environmental interference; the terahertz wave of a specific frequency band is emitted to the roadway rock wall and coal body by a Hertz source; the reflected and transmitted terahertz waves are received by a detector array, the detection area is discretized into a pixel point array, and the received signal is converted into a digital signal representing the thermal radiation intensity of each pixel point; the pose information of the scanning device is recorded in real time; the pose information, time of flight and phase difference data of the echo signal are fused; the digital signal is processed by a back projection algorithm and an algebraic reconstruction algorithm to generate a three-dimensional thermal radiation intensity distribution model.

[0013] In a preferred embodiment, the two-dimensional thermal spot coordinate set of the temperature anomaly area is identified, specifically: a wavelet transform and median filtering hybrid denoising processing is performed on the thermal radiation intensity data; based on the denoised data, a suspected abnormal area is obtained by primary segmentation using the Otsu algorithm; statistical outlier detection is applied to the suspected abnormal area to realize secondary segmentation to extract a high temperature area; morphological erosion and expansion operation is performed on the high temperature area to optimize the boundary, and neighborhood connectivity analysis is used to filter false hot spots; based on geometric features, effective high temperature areas are screened, and the centroid coordinates are calculated using a gray weighted method; a Gaussian surface fitting algorithm is applied to the centroid coordinates for sub-pixel level optimization; the thermal spot edge temperature gradient is calculated, and the positioning accuracy is verified by comparison with the characteristics of the real heat source; the thermal spot pixel coordinates are converted to the global coordinate system by fusing the terahertz internal participation and hand-eye calibration parameters; the laser radar point cloud depth is used to correct the coordinate distortion, and the two-dimensional thermal spot global coordinates are output.

[0014] The technical effects and advantages of the operation method of the coal mine underground fire area information detection emergency robot of the present application are as follows: 1. The present application effectively avoids the danger of directly entering the fire area by scientifically determining the accessibility condition and setting fire and explosion prevention measures, and ensures the safety of the robot operation.

[0015] 2. The present application uses the positioning and obstacle avoidance capability of the walking system to enable the robot to quickly and accurately reach the target point, thereby improving the detection efficiency.

[0016] 3. The present application uses the information system and laser radar to calculate the explosion risk index in real time and compare and feed back, thereby providing timely and accurate data support for emergency decision-making and assisting rescue personnel in scientifically dealing with fire danger. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 The flowchart of the operation method of the coal mine underground fire area information detection emergency robot of the present application is shown. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be described in detail below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0019] Embodiment 1, Figure 1 The operation method of the coal mine underground fire area information detection emergency robot of the present application is given, which includes the following steps: S1, the robot determines the accessibility condition of the fire area and matches the entry mode; S2, the robot plans an entering route into the fire area based on the entering mode combined with Dijkstra algorithm, and reaches the designated detection position; The designated detection position is located based on terahertz imaging and multi-line laser radar fusion imaging positioning.

[0020] S3, the information system collects data, combines with laser radar assisted navigation detection, and loads software to calculate explosion risk index; S4, compare the risk index with the safety threshold, and make timely feedback.

[0021] This example effectively avoids the risk of directly entering the fire area by scientifically determining the access conditions and setting fire and explosion prevention measures, ensures the safety of the robot operation; using the positioning and obstacle avoidance capability of the walking system, the robot can quickly and accurately reach the target point, improving the detection efficiency; with the help of the information system and laser radar, real-time calculation of explosion risk index and comparison feedback are provided to provide timely and accurate data support for emergency decision-making, helping rescue personnel to scientifically deal with fire area danger.

[0022] S1, the robot determines the access conditions of the fire area and matches the entering mode; The robot determines the access conditions of the fire area and matches the entering mode, specifically: The access conditions of the fire area include fire area closure and fire area unclosure; The entering mode includes breaking the wall to enter the fire area and directly entering the fire area; Scan the surrounding tunnel of the fire area by laser radar; When the dual-spectrum thermal imager detects high temperature behind the continuous solid obstacle, and the gas sensor shows that the concentration of carbon monoxide and smoke rises and the oxygen content decreases, it is determined that the fire area is closed; When an open gap is detected and there is no high temperature and abnormal harmful gas, it is preliminarily determined that the fire area is not closed; When the fire area is closed, determine the wall breaking position and method and set fire and explosion protection measures, at the same time, use professional drilling equipment for operation, and monitor the drilling state in real time and lay detection equipment, break the wall to enter the fire area; When the fire area is not closed, directly enter the fire area after confirming that the environmental indicators meet the safety threshold.

[0023] The following is a feasible embodiment: Start the 32-line laser radar (scanning range 360°, ranging accuracy ±2cm, frequency 10Hz) carried by the emergency robot, emit 905nm wavelength laser beams, scan the tunnel within 50 meters around the fire area, construct a 0.1m×0.1m resolution three-dimensional point cloud map, and mark the positions and distances (such as 8 meters) of the obstacles such as the tunnel width of 3.5 meters, the height of 2.8 meters, and the gangue pile with a length of 2m, a width of 1.5m, and a height of 1.2m; During the scanning process, if the dual-spectrum thermal imager detects that the temperature behind the obstacle reaches ≥60°C, and the temperature gradient is ≥5°C / m, while the gas sensor data shows that the CO concentration is ≥24 ppm (with an increase rate of ≥5 ppm / min), the smoke concentration is ≥0.5% obs / m, and the O2 content is ≤18%, all three indicators are triggered at the same time, and the system determines that the fire area is closed. At this time, the hydraulic drilling machine (drill bit diameter 100 mm, maximum depth 10 meters, rotation speed 800 r / min) is started, the area with a crack rate ≥5% is selected for drilling, and the drilling resistance is monitored in real time (normal ≤5 kN, abnormal ≥8 kN stop), the depth is fed back by the laser range finder (accuracy ±1 cm), and sensors are arranged at the drilling hole to collect fire area CO, CH4 data every 10 seconds; If the laser radar detects a gap with a width ≥1.2 meters, and the temperature at the gap is ≤30°C, the CO concentration is <24 ppm, and the CH4 concentration is <1%, it is preliminarily determined that the fire area is not closed. Further confirm the gas concentration (CH4 <0.5%, CO <16 ppm, O2 ≥19.5%) and detect the crack width of the roadway roof and sidewall <3 mm and no obvious loosening through the visual sensor, when these safety indicators are up to standard, the robot enters the fire area according to the planned route.

[0024] Start the 32-line laser radar (scanning range 360°, ranging accuracy ±2 cm, scanning frequency 15 Hz) and the center frequency 100 MHz geological radar (detection depth up to 20 meters, resolution 0.2 meters) to scan the surrounding roadway of the fire area comprehensively. Synchronously call the three-dimensional geological model of the coal mine, the roadway CAD drawing, combine the scanning data, use GIS spatial analysis technology to accurately locate the obstacle coordinates (error ≤5 cm), analyze the internal crack distribution and cavity position through the geological radar echo signal, determine the material properties such as concrete strength grade (C20-C40) and rock hardness (Mohs hardness 3-7 grade) according to the density inversion algorithm; Deploy infrared temperature sensors (temperature measurement range -40°C-500°C, accuracy ±1°C), catalytic combustion gas sensors (range 0-100% LEL, resolution 0.1% LEL, response time ≤15 s), and electrochemical carbon monoxide sensors (range 0-1000 ppm, accuracy ±1 ppm) to collect data in real time at a frequency of 10 times per second. When detecting that the gas concentration is ≥1% LEL, the carbon monoxide concentration is ≥24 ppm, and the temperature is ≥50°C, the system automatically marks the dangerous area, generates a detour path plan, and ensures the safety of the working area environment; According to the results of obstacle analysis, if the material Mohs hardness is less than or equal to 4 (such as ordinary sandstone, gypsum), a hydraulic breaking hammer (working pressure 25-35 MPa, impact frequency 400-800 times / min, maximum impact force 80 kN) is used to break; for hard structures such as reinforced concrete and granite with hardness greater than or equal to 5, an ultra-high pressure water jet cutting system (pressure 400-600 MPa, flow rate 2-5 L / min, cutting accuracy ±0.5 mm) and a pulsed fiber laser cutting device (power 2-5 kW, wavelength 1064 nm, cutting speed 0.5-2 m / min) are used to achieve efficient and safe non-contact cutting.

[0025] An ExdI explosion-proof and flame-proof device is added to the core control circuit of the robot, the shell is sprayed with a 0.3 mm thick inorganic flame-retardant coating (oxygen index ≥32%), and the battery compartment is equipped with a double-layer heat dissipation flame-proof cover (thermal conductivity 150 W / (m·K), explosion-proof level ExibI). Within a 30-meter range around the work area, high-precision gas sensors (gas detection accuracy 0.01% LEL, carbon monoxide detection accuracy 0.1 ppm) are arranged every 5 meters, and the data is uploaded to the central control system in real time. At the same time, automatic fire extinguishing bombs (fire extinguishing coverage area 5-10 m², response time ≤2 s) and dry powder explosion suppression devices (jet intensity 200-300 g / m³, starting concentration 1000 ppm CO or 1.5% LEL gas) are set up. Once the dangerous parameters exceed the standard, the suppression measures are automatically triggered within 0.5 seconds to block the development of disasters.

[0026] S2, the robot plans an entry route into the fire zone based on the entry mode and Dijkstra algorithm, and reaches the specified detection position; For the terahertz imaging device, its working mode needs to be set, such as continuous scanning mode or pulse scanning mode, and the appropriate scanning frequency is set according to the characteristics of the roadway environment, generally 5-10 times per second, to ensure that the scanning area can be covered quickly and comprehensively; at the same time, the wavelength range is adjusted, usually set to 0.1-10 THz, so that it can effectively penetrate the common smoke and dust in the roadway. In terms of hardware connection, check whether the data transmission cable between the terahertz imaging device and the robot main control system is stable to ensure that the data can be transmitted in real time and accurately.

[0027] It should be noted that for long and narrow roadways or complex structures with many branches, the scanning frequency can be appropriately reduced to 3-5 times per second, and the single scanning time can be extended to improve the data density; in high dust concentration environment, the terahertz wave wavelength needs to be increased to 2-10 THz to enhance the penetration ability, and hardware protection (such as dust cover) is needed to avoid equipment failure. If there is strong electromagnetic interference, the data transmission cable needs to be shielded and an anti-interference protocol (such as CAN bus protocol) is used to ensure data integrity.

[0028] When the multi-line laser radar device is initialized, the laser emission and receiving units thereof need to be calibrated. Through a built-in self-checking program, whether the power of the laser emitter is within a normal range and whether the sensitivity of the laser receiver meets the standard are detected. According to the width and height of the roadway, the scanning angle of the laser radar is reasonably set. For example, the horizontal scanning angle is set to 360°, and the vertical scanning angle is adjusted between -15° and 15° according to actual needs, so as to obtain complete three-dimensional spatial information of the roadway. At the same time, the scanning resolution is determined. Generally, an angular resolution of 0.1°-1° can be selected to balance the data precision and processing efficiency. After the device parameter setting is completed, the terahertz imaging device and the laser radar device are time-synchronized. A high-precision clock synchronization protocol, such as the IEEE1588 protocol, is adopted to ensure that the timestamp error of the data collected by the two devices is controlled within milliseconds, thereby ensuring the accuracy of subsequent data fusion.

[0029] Based on the robot pose data, a conversion matrix T between the terahertz hot spot coordinates and the laser radar point cloud is established by the hand-eye calibration method, the two-dimensional hot spot coordinates are mapped to the three-dimensional space, and a fire source candidate area with confidence is generated. Specifically: The pose data collected by the robot at different positions and attitudes is recorded. These pose data contain the spatial position and direction information of the robot in the roadway. At the same time, in the terahertz thermal image, the corner point detection algorithm such as the Harris corner point detection algorithm is used to extract the corner points of the temperature anomaly area (hot spot) as feature points, and the two-dimensional coordinates of these feature points are recorded. In the laser radar point cloud data, by calculating the curvature of the points, the points with large curvature change are selected as feature points. These points are usually located at the edges or corners of objects, and have potential corresponding relationship with the feature points of the terahertz hot spot.

[0030] A feature matching method based on Euclidean distance is adopted to calculate the Euclidean distance between the terahertz hot spot feature points and the laser radar point cloud feature points. The point pairs with a distance less than a certain threshold are regarded as matched point pairs. In order to improve the accuracy of matching, a feature descriptor such as SIFT (Scale-Invariant Feature Transform) descriptor can also be used to describe and match the feature points more carefully.

[0031] The matched terahertz hot spot feature point coordinates, laser radar point cloud feature point coordinates, and corresponding robot pose data are substituted into the hand-eye calibration model. Here, the classic Tsai two-step hand-eye calibration model is adopted. Through the least squares method or singular value decomposition (SVD) algorithm, the unknown parameters in the model are solved, so as to obtain the conversion matrix T between the terahertz hot spot coordinates and the laser radar point cloud.

[0032] The specified detection position is obtained based on terahertz imaging and multi-line laser radar fusion imaging positioning; In the embodiment, the specified detection position is obtained based on terahertz imaging and multi-line laser radar fusion imaging positioning, specifically: The robot comprises a terahertz imaging device and a laser radar device. The terahertz imaging device and the laser radar device are initialized. The terahertz imaging device is used to perform fault scanning on a roadway rock wall and a coal body, to obtain thermal radiation intensity distribution data after penetrating smoke and dust, and to identify two-dimensional thermal spot coordinates of a temperature abnormal area. A three-dimensional point cloud map of the roadway is constructed by the multi-line laser radar, spatial coordinates of a roadway wall and an obstacle are extracted, and robot pose data are solved in real time. Based on the robot pose data, a conversion matrix of the terahertz thermal spot coordinates and the laser radar point cloud is established by a hand-eye calibration method, the two-dimensional thermal spot coordinates are mapped to a three-dimensional space, and a fire source candidate area with a confidence level is generated. The specified detection position is obtained by screening the fire source candidate area.

[0033] S3, the robot collects data at the detection position, judges an explosion risk, and sends an alarm signal. The information system collects environmental data in real time through a distributed sensor network deployed in the fire area and the surrounding area: a catalytic combustion type gas sensor (range 0-100% LEL, accuracy ±0.1% LEL) detects the gas concentration at a frequency of 2 times per second, an electrochemical carbon monoxide sensor (range 0-2000 ppm, accuracy ±1 ppm) synchronously monitors the CO concentration, a platinum resistance temperature sensor (temperature measurement range -40℃-200℃, accuracy ±0.5℃) collects temperature data in real time, a piezoresistive air pressure sensor (range 80-120 kPa, accuracy ±0.1 kPa) records air pressure changes, and a vibrating string strain gauge (measurement range ±3000με, accuracy ±0.1% F.S.) collects roadway deformation data every 5 seconds.

[0034] At the same time, a 128-line laser radar (scanning frequency 20 Hz, ranging accuracy ±1 cm) continuously scans the roadway space at an angular resolution of 0.05°, constructs three-dimensional environmental data with a point cloud density of 100 points / m², and obtains the position (error ≤5 cm), volume (resolution 0.01 m³) of an obstacle and the size of a roadway cross section (accuracy ±2 cm) in real time.

[0035] The system fuses the above multi-source data through Kalman filtering algorithm (data synchronization error ≤10 ms), and loads the pre-set explosion risk assessment software model. The model sets weight parameters based on the analytic hierarchy process: gas concentration weight 0.4 (explosion lower threshold 5% LEL), CO concentration weight 0.2 (danger threshold 24 ppm), temperature weight 0.2 (ignition point threshold 65°C), oxygen content weight 0.1 (combustion threshold 12%), and space sealing degree weight 0.1 (based on the change rate of roadway cross-sectional area calculated by laser radar scanning). The explosion risk index (value range 0-100) is calculated and generated by fuzzy comprehensive evaluation method, and is mapped to the risk level matrix in real time: index < 30 is safe area, 30-60 is warning area, ≥ 60 is dangerous area, and finally the evaluation results are pushed to the ground control center at a frequency of 0.5 seconds / time.

[0036] S4, the ground control center receives the alarm signal for emergency response, fire source determination and fire extinguishing work; The system compares the real-time calculated explosion risk index (value range 0-100) with the pre-set safety threshold, and the threshold is set to three standards of low threshold 30, medium threshold 60 and high threshold 80 according to the risk characteristics of coal mine fire area. When the risk index is less than 30, the system maintains a data acquisition frequency of 2 times per second, continuously monitors the environmental parameters of the fire area, and feeds back the current safety information (including 12 core indicators such as gas concentration, CO concentration, temperature, etc.) to the central monitoring platform at a frequency of 1 time per minute, and the data transmission delay is controlled within ≤500 ms.

[0037] Once the risk index reaches or exceeds 30, the system immediately triggers a graded alarm mechanism: when the risk index is in the range of 30-60, the first-level sound and light alarm is started, the alarm in the underground operation area is prompted with 85dB sound and red light flashing (frequency 2Hz), and the emergency command terminal is pushed with early warning information containing real-time risk index and threshold parameter list; when the index breaks through 60 and enters the 60-80 medium risk interval, it is upgraded to the second-level alarm, the alarm volume is increased to 95dB, the light flashing frequency is accelerated to 4Hz, and the terminal pushing content is increased risk trend prediction (based on LSTM algorithm, prediction accuracy ± 5%) and local area isolation, enhanced ventilation and other disposal suggestions; if the risk index exceeds 80, the highest third-level alarm is triggered, the alarm is warned with 105dB piercing alarm sound and flashing strong light (frequency 8Hz), the unnecessary power supply in the dangerous area is cut off, the fireproof door is automatically closed, and the emergency command terminal is sent with the emergency plan containing three-dimensional risk heat map and optimal escape route planning (path planning error ≤2m), to ensure the rapid and accurate emergency response.

[0038] The above formulas are all de-dimensioned to calculate the numerical values, the formulas are obtained by collecting a large amount of data to simulate the most recent real situation, and the preset parameters in the formulas are set by a person skilled in the art according to the actual situation.

[0039] The above embodiments can be realized wholly or partially by software, hardware, firmware or any other combination. When realized by software, the above embodiments can be realized wholly or partially in the form of a computer program product.

[0040] Those skilled in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solutions. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0041] In addition, each functional module in each embodiment of the present application can be integrated in one processing module, or each module can exist physically alone, or two or more modules can be integrated in one module.

[0042] The above is merely specific embodiments of the present application, but the protection scope of the present application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0043] Finally: the above is only the preferred embodiment of the present application, and is not used to limit the present application, any modification, equivalent replacement, improvement, etc. within the spirit and principle of the present application should be included in the protection scope of the present application.

Claims

1. An operating method of an emergency robot for detecting fire area information in underground coal mines, characterized in that: include: Determine accessibility to the fire zone based on the environmental perception layer and match entry methods; Planning a route into the fire zone based on the entry method and the Dijkstra algorithm to reach the designated detection position, which is obtained based on terahertz imaging and multi-line lidar fusion imaging positioning; Collect data at the detection location, determine explosion risks, and send alarm signals; Receive alarm signals through the control center for emergency response and positioning.

2. The operating method of the coal mine underground fire area information detection emergency robot according to claim 1, characterized in that: The environmental perception layer includes obtaining a fire scene temperature gradient map through a dual-spectrum thermal imager, detecting obstacle distribution using a millimeter-wave radar, and analyzing the building structure BIM model to extract channel attributes.

3. The method for operating an emergency robot for detecting fire zone information in underground coal mines according to claim 2, characterized in that: The determination of the fire zone accessibility conditions and matching of the entry methods are as follows: Acquire real-time sensor data through the environmental perception layer, wherein the sensor data includes at least thermal imaging data collected by a dual-spectrum thermal imager and gas concentration data collected by a gas sensor; When high temperature is detected behind continuous physical obstacles, and the gas sensor shows an increase in carbon monoxide concentration, an increase in smoke concentration, and a decrease in oxygen concentration, the fire zone is determined to be closed; When an open gap is detected and there is no abnormal high temperature in the thermal imaging data and no excessive harmful gas in the gas concentration data, it is determined that the fire area is not closed; If the fire zone is closed, the closed fire zone entry procedure will be executed. If the fire zone is not closed, the fire zone will be entered directly after the environmental indicators continue to meet the safety threshold.

4. The method for operating an emergency robot for detecting fire zone information in underground coal mines according to claim 3, characterized in that: If the fire zone is closed, the closed fire zone entry procedure is executed, which also includes the execution of the wall breaking procedure, specifically including: Scanning the tunnels around the fire zone using LiDAR and geological radar, while simultaneously retrieving coal mine geological data and tunnel drawings, generates a first dataset containing obstacle locations, internal structures, and material properties. Real-time detection of the temperature and gas concentration of the surrounding environment of the wall breaking, and dynamic avoidance of dangerous areas where the temperature or gas concentration exceeds the preset threshold; Based on the material properties and structural features in the first data set, the hardness level of the fire zone wall is calculated using a hardness assessment model. If the hardness level is ≤ threshold H, a hydraulic breaker is used to break down the obstacle. If the hardness level is greater than threshold H, water jet cutting and laser cutting are used in collaboration to break down the hard structure.

5. The operating method of the coal mine underground fire area information detection emergency robot according to claim 4, characterized in that: The designated detection position is obtained based on terahertz imaging and multi-line laser radar fusion imaging positioning, specifically: Initialize the onboard terahertz imaging device and multi-line laser radar device; The terahertz imaging device is used to perform penetrating tomography scanning of the tunnel rock wall and coal body to obtain the thermal radiation intensity distribution data in the smoke and dust environment, identify the temperature abnormality area and output the two-dimensional hot spot coordinates; Use multi-line laser radar to synchronously build a three-dimensional point cloud map of the tunnel, extract the spatial coordinates of the tunnel walls and obstacles, and calculate the robot's posture data in real time; Based on the robot posture data, a coordinate transformation matrix is ​​established through the hand-eye calibration method, the two-dimensional hot spot coordinates are mapped to the three-dimensional point cloud space, and the candidate fire source area with a confidence score is generated; The candidate fire source areas are screened according to the confidence scores to determine the final detection location.

6. The method for operating an emergency robot for detecting fire zone information in underground coal mines according to claim 5, characterized in that: The entry route into the fire zone is planned based on the entry method combined with the Dijkstra algorithm, specifically: Based on the three-dimensional map of the fire zone tunnel, obstacle distribution, and gas concentration abnormality area data obtained by the information detection system, combined with the first historical data set, the theoretically optimal route is planned using the Dijkstra algorithm; During the robot's movement, the laser radar constructs a point cloud map in real time and matches environmental features. The visual sensor identifies fixed obstacles and dynamic risk targets, and the inertial measurement unit monitors the robot's real-time posture. When an obstacle is detected, the dynamic window method and the artificial potential field method fusion algorithm are used to locally correct the theoretical optimal route and generate an obstacle avoidance path as the second route; If a high temperature area or an area with excessive harmful gases is detected, the second route will be forcibly adjusted to generate the final route into the fire area.

7. The method for operating an emergency robot for detecting fire zone information in underground coal mines according to claim 6, characterized in that: The theoretical optimal route is planned by the Dijkstra algorithm, specifically: (1) Extract the three-dimensional map of the fire zone tunnel from the information detection system and the first data, mark the nodes, paths, obstacles, and dangerous areas, and construct a network diagram containing path length and danger level weights; (2) Take the robot’s current position as the starting node, specify the detection position as the target node, and initialize the cumulative weight value of the starting node; (3) Start from the starting node and diffuse the search outward, calculating the cumulative weight value of each adjacent node of the current node; (4) In the network graph, the search spreads outward from the starting node; (5) If the cumulative weight value is less than the cumulative weight value recorded for the node, update its cumulative weight value and predecessor node, and repeat steps (1)-(4) until the target node is found; (6) Trace back to the starting node along the predecessor node sequence of the target node to generate the theoretical optimal route.

8. The method for operating an emergency robot for detecting fire zone information in underground coal mines according to claim 7, characterized in that: The terahertz imaging device is used to perform penetrating tomography scanning on the tunnel rock wall and coal body to obtain thermal radiation intensity distribution data in a smoke and dust environment, specifically: Before scanning, obtain the ambient thermal radiation signal in the absence of heat source as background noise; During scanning, the actual collected thermal radiation data and background noise data are differentially processed to eliminate environmental interference; The terahertz wave of a specific frequency band is emitted to the rock wall and coal body of the roadway through the Hertz source; The reflected and transmitted terahertz waves are received by the detector array, the detection area is discretized into a pixel array, and the received signal is converted into a digital signal representing the thermal radiation intensity of each pixel; Record the position information of the scanning device in real time; fusing the posture information with the flight time and phase difference data of the echo signal; The back-projection algorithm and algebraic reconstruction algorithm are used to collaboratively process digital signals to generate a three-dimensional thermal radiation intensity distribution model.

9. The method for operating an emergency robot for detecting fire zone information in underground coal mines according to claim 8, characterized in that: The two-dimensional hot spot coordinate set for identifying the temperature anomaly area is specifically: Perform hybrid noise reduction processing of wavelet transform and median filtering on thermal radiation intensity data; Based on the denoised data, the Otsu algorithm is used for initial segmentation to obtain the suspected abnormal area; Apply statistical outlier detection to suspected abnormal areas and implement secondary segmentation to extract high-temperature areas; Morphological erosion and dilation operations are performed on high-temperature areas to optimize the boundaries, and pseudo hotspots are filtered through neighborhood connectivity analysis; The effective high-temperature area is screened based on geometric features, and its centroid coordinates are calculated using the grayscale weighted method; Gaussian surface fitting algorithm is applied to perform sub-pixel optimization of centroid coordinates; Calculate the temperature gradient at the edge of the hot spot and verify the positioning accuracy by comparing it with the actual heat source characteristics; Fusing terahertz intrinsic parameters with hand-eye calibration extrinsic parameters, the hot spot pixel coordinates are converted to the global coordinate system; The coordinate distortion is corrected based on the depth of the lidar point cloud and the global coordinates of the two-dimensional hot spot are output.

Images