An acoustic non-destructive coal bunker three-dimensional temperature field imaging method and system

Abstract

This invention relates to the field of coal bunker safety monitoring and industrial measurement technology, specifically to an acoustic non-destructive method and system for three-dimensional temperature field imaging in coal bunkers. The method involves deploying a multi-layered ring-shaped acoustic transducer array on the coal bunker wall to acquire multipath acoustic signals penetrating the coal and their transit times; simultaneously, it acquires the air temperature and relative humidity inside the coal bunker. Combined with the coal's pore structure parameters, a dual-domain coupled equivalent acoustic slow-motion model is established. A conversion factor is used to correct the free-space acoustic slow-motion to accurately characterize the propagation of sound waves in the porous medium. Using this model as a physical constraint, and the measured transit time as the observation value, an iterative inversion solution is performed using a regularization constraint consistent with the coal compaction anisotropy to obtain the reconstructed three-dimensional temperature field. The system includes a sensing layer, an edge layer, and a platform layer. This invention achieves non-contact, full-volume, three-dimensional real-time imaging and hotspot localization of the internal temperature of a coal bunker.

Description

Technical Field

[0001] This invention relates to the field of coal bunker safety monitoring and industrial measurement technology, specifically to an acoustic non-destructive three-dimensional temperature field imaging method and system for coal bunkers. Background Technology

[0002] Coal storage and transportation inherently carry risks of low-temperature oxidation and heat accumulation. Loose coal is a poor conductor of heat, and the slow heat exchange between the air and coal particles within its pores can easily lead to the formation of hidden "hot spots" or heat nuclei within the coal pile. If these hot spots are not detected and addressed in time, they can develop into spontaneous combustion, causing fires, production disruptions, and significant economic losses. Currently, temperature monitoring in coal bunkers primarily relies on point-based or cable-based contact temperature measurement, resulting in discrete monitoring information and significant blind spots. When hot spots are not close to the sensor path, missed detections or delayed reports often occur. Furthermore, installation and maintenance are difficult, and stress and dust corrosion lead to a high failure rate. Medium and large-sized coal bunkers urgently require a non-contact, full-volume-coverage, online-operable three-dimensional temperature detection method to achieve early warning and precise location.

[0003] Existing technologies for temperature detection in coal bunkers include: one is a contact temperature measurement method, which uses several cables arranged along the bunker height to achieve multi-point temperature measurement. This method is technically mature and has high single-point accuracy, but has low spatial resolution, sparse distribution of points, and fixed positions; on-site installation and maintenance are complex, and it is easily affected by compaction, pulling, moisture, and dust; for large coal bunkers with dynamic changes in volume, blind spots and insufficient representativeness often occur. Another method is non-contact temperature measurement such as infrared / thermal imaging and other optical temperature measurement, but this method can only obtain the temperature of the surface or visible area, and is easily affected by dust obstruction, smoke, light path contamination, and changes in emissivity. It cannot penetrate the bunker to obtain internal temperature information, and its ability to identify endogenous hot spots in the early stage is limited. There is also a distributed fiber optic temperature measurement method. However, this method requires burying the optical fiber inside the coal pile, which is limited by the difficulty of burial, mechanical damage, bending radius, and contact thermal conduction hysteresis. The stability of wireless sensor networks is limited in high dust, metal enclosure, and complex electromagnetic environments, and battery replacement and intrinsic safety explosion protection are also challenges.

[0004] While acoustic temperature measurement technology has applications in boilers and other similar scenarios, directly applying this method to loose coal in coal bunkers still faces a series of differences in media mechanisms and engineering implementation. The main reasons are as follows: Loose coal can be regarded as a rigid porous medium, where sound waves mainly propagate along the pore air rather than the solid skeleton; the pore structure is tortuous and has a variety of scale distributions, which leads to dispersion and additional attenuation (viscous / thermal boundary layer loss) related to the effective sound velocity frequency. If the free field air sound velocity model is directly used for temperature inversion, systematic biases are likely to occur. The low-temperature oxidation and seepage process of coal causes dynamic changes in pore gas composition and relative humidity, which can affect sound velocity on the same order of magnitude as the temperature effect. Without composition / humidity compensation, the uniqueness and accuracy of temperature inversion are difficult to guarantee. Coal piles are highly absorbing media, and high-frequency sound propagation attenuates quickly and scatters strongly. Noise from equipment inside the silo, material impact, and structural conducted noise make it difficult to accurately measure the transit time, requiring the combination of low-frequency broadband excitation and robust time delay estimation algorithms. The coal bunker is a closed structure, and the transducers can only be arranged in a limited position along the bunker wall. The number and geometric angle of the projection (survey lines) are limited, which leads to the ill-conditioning and blind zone problems of the three-dimensional inversion. It is necessary to comprehensively design the device layout, scheduling method and mathematical regularization.

[0005] Dust, humidity, corrosion, and temperature differences pose challenges to the long-term stability of sensors; intrinsic safety limits, explosion-proof wiring, and anti-static requirements in coal dust explosion hazardous environments increase the difficulty of system integration.

[0006] In summary, the existing technology lacks a systematic solution that can simultaneously solve the problems of accurate modeling of porous media, robust measurement in noisy environments, and stable 3D reconstruction under finite projection, which makes it impossible to achieve online monitoring and accurate early warning of the 3D temperature field inside coal bunkers in an engineered manner. Summary of the Invention

[0007] To address the problems existing in the prior art, this invention provides a method and system for non-destructive acoustic imaging of a three-dimensional temperature field in a coal bunker. This method acquires multi-path acoustic wave signals penetrating the coal body and their transit times by deploying a multi-layered ring-shaped acoustic transducer array on the coal bunker wall; simultaneously, it acquires the air temperature and relative humidity inside the coal bunker. Combining this with the coal body's pore structure parameters, a dual-domain coupled equivalent acoustic slow-motion model is established. A conversion factor is used to correct the free-space acoustic slow-motion to accurately characterize the propagation of sound waves in the porous medium. Using this model as a physical constraint and the measured transit time as the observation value, an iterative inversion solution is performed using a regularization constraint consistent with the coal body's compaction anisotropy to obtain the reconstructed three-dimensional temperature field. The system includes a sensing layer, an edge layer, and a platform layer. This invention achieves non-contact, full-volume, three-dimensional real-time imaging and hotspot localization of the temperature inside a coal bunker.

[0008] This invention employs the following technical solution: a method for acoustic non-destructive imaging of a three-dimensional temperature field in a coal bunker, comprising: A multi-layered, ring-shaped array of acoustic transducers is deployed in the coal bunker in a predetermined manner. The acoustic transducer array emits acoustic signals that penetrate the coal body through multiple acoustic paths, and the measured flight time of the acoustic signal in each acoustic path is obtained. Acquire internal environmental data of the coal bunker and establish a dual-domain coupled equivalent acoustic-slow-motion model; the internal environmental data of the coal bunker includes: air temperature, relative humidity, frequency, pore structure parameters and ambient gas parameters; The dual-domain coupled equivalent acoustic-slow-motion model specifically includes: Based on air temperature and relative humidity, the baseline acoustic slowness of the free space inside the coal bunker is calculated using a wet air sound velocity model. Conversion factors are constructed based on pore structure parameters and ambient gas parameters; By multiplying the baseline acoustic slowness by the conversion factor, a dual-domain coupled equivalent acoustic slowness model is constructed; The three-dimensional temperature field inside the coal bunker is reconstructed by iterative inversion solution based on the dual-domain coupled equivalent acoustic-slow-motion model and the measured flight time. The iterative inversion solution includes: calculating the theoretical flight time using the dual-domain coupled equivalent acoustic-slow-motion model based on the assumed temperature field value in the current coal bunker. The flight time residual is obtained by comparing the theoretical flight time with the measured flight time. With the goal of minimizing the time-lapse residual, a regularization constraint consistent with the anisotropy of coal compaction is constructed to update the assumed temperature field value inside the coal bunker. The three-dimensional temperature field inside the coal bunker is output through iterative inversion solution.

[0009] Furthermore, a conversion factor is constructed based on pore structure parameters and ambient gas parameters, including: The fundamental components of the conversion factor that are independent of frequency are obtained based on the pore structure parameters. Based on pore structure parameters and ambient gas parameters, the frequency-negative correlation relaxation components characterizing viscous loss and heat exchange loss in the conversion factor are obtained. Based on the anisotropy of the coal body due to its own weight compaction, the direction correction component related to the direction of sound wave propagation in the conversion factor is obtained. An isotropic conversion factor is constructed based on the fundamental component and the frequency negative correlation relaxation component, and the isotropic conversion factor is anisotropically corrected using the direction correction component to obtain the conversion factor.

[0010] Furthermore, the expression for the conversion factor is: In the formula, Indicates the conversion factor. Represents absolute temperature. Indicates frequency, Indicates pore structure parameters, Parameters representing the compaction state of the coal body. A unit vector representing the direction of a sound ray; The transformation factor represents isotropic properties. Represents the anisotropic gain coefficient. Represents the unit vector in the vertical direction. This indicates the compaction index.

[0011] Furthermore, the isotropic conversion factor is expressed as: In the formula, The transformation factor represents isotropic properties. Represents absolute temperature. Indicates frequency, Indicates pore structure parameters, For high-frequency tortuosity, Porosity and This represents the amplitude coefficient of the two branches: viscous loss and heat exchange loss. and The characteristic frequencies represent the two branches of viscous loss and heat exchange loss. and It represents the relaxation index of the two branches: viscous loss and heat exchange loss.

[0012] Furthermore, the characteristic frequencies of the two branches, viscous loss and heat exchange loss, are respectively expressed as: In the formula, The characteristic frequency representing viscous loss, Represents absolute temperature. Represents the mole fraction vector of gas components. Indicates pore structure parameters, The dimensionality coefficient representing viscous loss. This represents the flow resistance of a static viscous channel. Porosity Indicates the viscous characteristic length, For high-frequency tortuosity, Indicates the viscosity of moist air. Indicates the density of moist air; The characteristic frequency representing heat exchange loss, The dimensioning coefficient represents the heat exchange loss. This indicates the flow resistance of the static thermal channel. Indicates the length of the thermal characteristic. This represents the Prandtl number.

[0013] Furthermore, after applying anisotropic correction to the isotropic conversion factor using the aforementioned direction correction component to obtain the conversion factor, the process further includes: Obtain the microclimate correction factor, and multiply the microclimate correction factor by the conversion factor to obtain the linearly corrected conversion factor; The expression for obtaining the microclimate correction factor is: in, As a microclimate correction factor, Represents absolute temperature. Represents the mole fraction vector of gas components. Indicates pore structure parameters, This indicates the current relative humidity inside the coal bunker. Indicates reference relative humidity. Indicates the humidity sensitivity coefficient. Represents the reference gas component mole fraction vector. Indicates the sensitivity coefficient of gas components, superscript This represents the vector transpose operator.

[0014] Furthermore, based on the assumed temperature field value within the current coal bunker, the theoretical flight time is calculated using the aforementioned dual-domain coupled equivalent acoustic-slow-motion model, expressed as: In the formula, Indicates the first A sound wave path at frequency The theoretical flight time is as follows. The geometric traversal length weight represents the weight of the first traversal. The sound wave path is at the 1st The length of the straight line segment that passes through the interior of a voxel. Indicates that the baseline sound is slow. Indicates the conversion factor. Indicates the first The absolute temperature of a voxel. Indicates the first The relative humidity of individual elements, Indicates the first The mole fraction vector of gas components of an individual element. Indicates the first Pore ​​structure parameters of individual elements, Indicates the first Compaction state parameters of individual elements Indicates the first The unit vector representing the direction of propagation of a sound wave path.

[0015] Furthermore, the iterative inversion solution process includes: The flight time residual is converted into an equivalent length domain residual, and the equivalent length domain residual is weighted and fused according to the temperature sensitivity and measurement noise at each frequency point to obtain the observation residual vector used for inversion. An optimization problem is established with the goal of minimizing the observed residual vector, and a regularization constraint consistent with the anisotropy of coal compaction is introduced; The update of the three-dimensional temperature field is obtained by solving the optimization problem with regularization constraints. The current temperature field assumptions are corrected using the update amount of the three-dimensional temperature field to obtain the updated three-dimensional temperature field. The process iterates until the preset convergence condition is met, and then outputs the reconstructed three-dimensional temperature field.

[0016] This invention further proposes an acoustic non-destructive three-dimensional temperature field imaging system for coal bunkers, which performs the acoustic non-destructive three-dimensional temperature field imaging method for coal bunkers as described above. The system includes: The sensing layer includes a multi-layer ring-shaped acoustic transducer array arranged along the coal bunker wall. The multi-layer ring-shaped acoustic transducer array contains multiple transducer nodes for transmitting low-frequency broadband acoustic signals in a predetermined sequence and receiving acoustic signals after penetrating the coal body. The edge layer, connected to the perception layer, includes a control module, a signal processing module, and an inversion calculation module; The control module is used to control the transmission and reception timing of the multilayer ring acoustic transducer array; The signal processing module is used to process the received acoustic wave signal to extract the flight time data of each path. The inversion calculation module is used to construct and update the dual-domain coupled equivalent acoustic-slow-motion model, and execute the iterative inversion and reconstruction algorithm of the three-dimensional temperature field to obtain the reconstructed three-dimensional temperature field. The platform layer, connected to the edge layer, includes a data storage unit, a visualization unit, and an early warning unit; The data storage unit is used to store real-time reconstructed three-dimensional temperature field data and historical three-dimensional temperature field data. The visualization unit is used to render and display the reconstructed three-dimensional temperature field; The early warning unit is used to identify hot spots in the coal bunker based on the reconstructed three-dimensional temperature field and trigger an alarm.

[0017] The beneficial effects of this invention are as follows: By constructing a dual-domain coupled equivalent acoustic-slow-motion model representing "humid air and porous air," this invention accurately simulates the propagation law of sound waves in complex porous coal bodies, solving the model mismatch problem from a physical perspective. The sound wave transmission and reception employ a multi-layer ring array and composite path design, combined with low-frequency broadband signals and robust time delay estimation, achieving high-precision acquisition of full-field sound wave data under strong noise conditions. Furthermore, it innovatively proposes an inversion algorithm that integrates multi-frequency data weighted fusion, model sensitivity, and regularization constraints consistent with anisotropic physics, thereby stably reconstructing a high-resolution three-dimensional temperature field under limited sensor conditions. This meets the requirements of industrial environments such as explosion-proof and dust-proof environments, enabling non-contact, full-coverage, three-dimensional precise positioning and early warning of hidden hotspots inside coal bunkers, providing a reliable and efficient intelligent monitoring method for coal storage safety. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is a schematic diagram of a method for three-dimensional temperature field imaging of a coal bunker using acoustic non-destructive imaging, according to an embodiment of the present invention. Figure 2 This is a schematic diagram of the topology of an acoustic transducer array according to an embodiment of the present invention; Figure 3 This is a schematic diagram of the arrangement of a ring array of acoustic transducers and the acoustic wave propagation path according to an embodiment of the present invention. Figure 4 This is a schematic diagram of an acoustic transducer structure according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of an acoustic non-destructive coal bunker three-dimensional temperature field imaging system according to an embodiment of the present invention.

[0020] In the diagram, 1. Coal bunker wall; 2. Warehouse wall vibration isolation layer; 3. Acoustic transducer; 4. Short waveguide; 5. Acoustic window; 6. Accelerometer; 7. Miniature temperature, humidity, and pressure probe; 8. Power amplifier; 9. Analog-to-digital converter; 10. Industrial-grade computer; 11. Integrated power supply and cabinet; 12. Labyrinth-type sealed connection; 13. Sound-absorbing coating; 14. Sound wave transmission channel; 15. Dustproof and sound-permeable membrane; 16. Metal protective mesh; 17. Heating film. Detailed Implementation

[0021] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] A schematic diagram of an acoustic non-destructive coal bunker three-dimensional temperature field imaging method and system according to an embodiment of the present invention is shown below. Figure 1 As shown, it includes: A multi-layered, ring-shaped array of acoustic transducers is deployed in the coal bunker in a predetermined manner. In this embodiment of the invention, the acoustic transducer array is arranged using a geometric topology of "multi-ring, multi-layer + oblique projection through", such as... Figure 2 and Figure 3 As shown in the image on the left, three to five layers of ring arrays are set along the height of the coal bunker. When setting this number of layers, it is necessary to ensure that at least the top, middle and bottom of the coal bunker are included. Furthermore, several transducers are evenly arranged around the circumference of each ring, while adjacent rings are staggered by half a pitch in the circumferential direction, in order to increase the diversity of the incident azimuth of the measuring line and reduce the ill-conditioned nature of the geometric matrix.

[0023] In one specific embodiment of the present invention, the number of transducer nodes evenly distributed along the circumference of the coal bunker can be set based on the circumference of the coal bunker and engineering experience. Simultaneously, at each set installation height, a corresponding number of equally spaced installation center points can be marked on the bunker wall using measuring tools such as a total station. Furthermore, to obtain more incident angles in the horizontal direction for the sound wave path and reduce the ill-conditioned nature of the inversion geometry matrix, the nodes of adjacent annular arrays are staggered in the circumferential direction. After the layout is completed using the above settings, as shown... Figure 3 As shown in the image on the right, three types of sound wave propagation paths with rich spatial orientation can be generated: Same-layer oblique projection path: For example, the path formed between two adjacent nodes at approximately 90° radians on the bottom ring is used to detect the horizontal distribution of that height section.

[0024] Cross-layer through-path: For example, a near-vertical path is formed between a node in the bottom ring and a node in the top ring that is roughly diametrically opposite each other, which enhances longitudinal sensitivity.

[0025] Cross-layer oblique path: For example, an oblique long path formed between a node in the middle ring and a non-opposite node in the upper ring can effectively penetrate the central area of ​​the warehouse and reduce the central blind spot.

[0026] The above-mentioned layout method enables a limited number of wall-mounted nodes to construct thousands of acoustic wave paths with different lengths, azimuth angles, and elevation angles, providing a superior geometric observation basis for subsequent three-dimensional temperature field reconstruction.

[0027] The acoustic transducer array emits acoustic signals that penetrate the coal body through multiple acoustic paths, and the measured flight time of the acoustic signal in each acoustic path is obtained. In this embodiment of the invention, the excitation signal adopts a linear frequency sweep (LFM) waveform with a duration of 0.1 seconds. The main frequency band is set to 400-900 Hz according to the type of coal stored. At the same time, in order to improve efficiency, multiple non-overlapping sub-bands are allowed to be excited simultaneously within a measurement frame. According to the optimized timing sequence, the measurement frame is divided into multiple time slots. In the first time slot, node A located at the bottom ring is called to transmit an LFM signal with a center frequency of 500 Hz. At the same time, nodes B and C located in the middle ring, which form a "cross-layer oblique transmission path" with it, are called to transmit LFM signals with center frequencies of 650 Hz and 800 Hz, respectively. This is the form of "multiple pairs of paths transmitting in parallel (FDM) within one time slot". In the next time slot, the system switches to another set of transducer pairs to transmit the same or different sub-band signals until all preset critical paths are covered.

[0028] In this embodiment of the invention, while any transducer node transmits an acoustic signal, high-precision data acquisition is simultaneously initiated through all transducer nodes (typically multiple nodes on the same or adjacent layers) that may receive the signal, according to a preset receiving path list. The acquisition clocks of each transducer node are kept synchronized at the microsecond level using the IEEE 1588 (PTP) precision clock protocol. Figure 4 As shown, the received signal is digitized by the power amplifier 8 and the 24-bit analog-to-digital converter 9 within the node at a sampling rate of no less than 96 kS / s. The system prioritizes measurement during the intervals of coal bunker material falling or strong mechanical impact to avoid strong interference. For each acoustic wave path, the received acoustic signal is processed as follows to obtain a high-precision measured transit time: The first step is preprocessing, which involves digital bandpass filtering of the acoustic signal to preserve the effective frequency band and suppress out-of-band noise. Then, a wavelet thresholding denoising algorithm is applied to further reduce the persistent broadband background noise within the coal bunker. The preprocessed acoustic signal is then cross-correlated with the stored original transmitted LFM reference signal. The position of the main peak with the largest amplitude is identified in the calculated cross-correlation function to obtain a coarse estimate of the transit time. Subsequently, parabolic interpolation or cubic spline interpolation is used near the main peak to improve the accuracy of the transit time estimation to the sub-sampling interval level, ultimately obtaining the measured transit time value with microsecond-level accuracy.

[0029] Acquire internal environmental data of the coal bunker and establish a dual-domain coupled equivalent acoustic-slow model; In this embodiment of the invention, the internal environmental data of the coal bunker includes at least: air temperature, relative humidity, frequency, pore structure parameters, and ambient gas parameters; such as Figure 2 and Figure 4 As shown, in each layer of the ring array, at least two miniature temperature, humidity and pressure probes 7 are deployed to monitor the air temperature and relative humidity inside the coal bunker at that height level in real time. These probes are connected to nearby acoustic transducer nodes through digital interfaces, and the data is uploaded along with the acoustic signal as an instantaneous input for calculating the sound velocity of humid air in free space.

[0030] Environmental gas parameters mainly refer to the mole fraction of key gases such as carbon dioxide and methane in pore air that can affect the speed of sound. There are usually two ways to obtain them. One is to install equipment such as infrared gas analyzers at the top of the coal bunker or at the ventilation opening for online monitoring to obtain the macroscopic gas concentration trend. The other is to use them as state parameters of the model and estimate them through an inversion process. That is, by setting a set of reference values, in subsequent iterations, when the multi-frequency residuals of the sound wave data show systematic characteristics related to the changes in gas composition, the reference values ​​are updated.

[0031] Pore ​​structure parameters, including porosity, high-frequency tortuosity, flow resistance, and characteristic length, are inherent properties of coal and do not change in real time with temperature, but vary with coal type and compaction state. These parameters can be determined in the following ways: Laboratory calibration: This involves performing industrial analysis, CT scans, and acoustic tests on stored coal samples to fit and obtain representative pore structure parameters for that coal type. On-site zero-field calibration: This involves conducting a dedicated "zero-field calibration" during a period when the coal temperature in the coal bunker is uniform and stable. Using the large amount of acoustic path data collected at this time, combined with the known ambient temperature and humidity, a set of optimal pore structure parameters is inverted and solved as the initial calibration values ​​for that coal type in that bunker.

[0032] Frequency is an input variable of the model, not environmental data. It directly corresponds to the frequency of the emitted acoustic wave linear sweep signal and can be substituted as a known quantity during model calculation.

[0033] Based on the data obtained above, the dual-domain coupled equivalent acoustic-slow model can be established as follows: Based on air temperature and relative humidity, the baseline acoustic slowness of the free space inside the coal bunker is calculated using a wet air sound velocity model. In this embodiment of the invention, the real-time air temperature and relative humidity are substituted into a mature engineering formula for the speed of sound in humid air (such as the Cramer formula) for calculation. First, the speed of sound in humid air is obtained, and then its reciprocal, i.e. the baseline sound slowness, is obtained. This calculation is performed in real time in each measurement cycle.

[0034] Conversion factors are constructed based on pore structure parameters and ambient gas parameters; The conversion factor is the core innovation of the dual-domain coupled equivalent acoustic manifold model in this embodiment of the invention. It is a dimensionless multiplier that corrects the free-space acoustic manifold to the porous medium acoustic manifold. Its construction consists of three core parts, all of which depend on the obtained parameters, as follows: The fundamental components of the conversion factor that are independent of frequency are obtained based on the pore structure parameters. In this embodiment of the invention, the fundamental component can be obtained by calculating the high-frequency asymptotic term based on the calibrated pore structure parameters. This part is frequency-independent and characterizes the most fundamental influence of the pore set on sound wave propagation. Its expression in the conversion factor establishment process is as follows: ,in, For high-frequency tortuosity, Porosity.

[0035] Based on pore structure parameters and ambient gas parameters, the frequency-negative correlation relaxation components characterizing viscous loss and heat exchange loss in the conversion factor are obtained. In this embodiment of the invention, relaxation terms characterizing viscous loss and heat exchange loss are constructed, which are expressed as follows during the establishment of the conversion factor: In the formula, This represents the frequency-negative correlation relaxation component. Represents absolute temperature. Indicates frequency, Indicates pore structure parameters, and This represents the amplitude coefficient of the two branches: viscous loss and heat exchange loss. and The characteristic frequencies represent the two branches of viscous loss and heat exchange loss. and The relaxation index represents the two branches of viscous loss and heat exchange loss; among which, and Two amplitude coefficients and and The two relaxation exponents are calibration parameters, while and The two characteristic frequencies are calculated based on the current gas properties and pore structure parameters, thus coupling the environmental gas effect into the dispersion model. Their calculation expression is as follows: In the formula, The characteristic frequency representing viscous loss, Represents absolute temperature. Represents the mole fraction vector of gas components. Indicates pore structure parameters, The dimensionality coefficient representing viscous loss. This represents the flow resistance of a static viscous channel. Porosity Indicates the viscous characteristic length, For high-frequency tortuosity, Indicates the viscosity of moist air. Indicates the density of moist air; The characteristic frequency representing heat exchange loss, The dimensioning coefficient represents the heat exchange loss. This indicates the flow resistance of the static thermal channel. Indicates the length of the thermal characteristic. This represents the Prandtl number.

[0036] Based on the anisotropy of the coal body due to its own weight compaction, the direction correction component related to the direction of sound wave propagation in the conversion factor is obtained. In this embodiment of the invention, the orientation correction component is constructed based on the anisotropic characteristics of the coal body due to its own weight compaction, and is expressed as follows: In the formula, Indicates the direction correction component. Represents absolute temperature. Indicates frequency, Parameters representing the compaction state of the coal body. A unit vector representing the direction of a sound ray; Represents the anisotropic gain coefficient. Represents the unit vector in the vertical direction. The compaction index is represented by the component , which varies with height and can be obtained through calibration. The anisotropic gain coefficient is also a calibration parameter, which enables the model to distinguish the speed differences of sound waves propagating in different directions.

[0037] An isotropic conversion factor is constructed based on the fundamental component and the frequency negative correlation relaxation component, and the isotropic conversion factor is anisotropically corrected using the direction correction component to obtain the conversion factor.

[0038] In this embodiment of the invention, the isotropic conversion factor is expressed as: In the formula, The transformation factor represents isotropic properties. Represents absolute temperature. Indicates frequency, Indicates pore structure parameters, For high-frequency tortuosity, Porosity and This represents the amplitude coefficient of the two branches: viscous loss and heat exchange loss. and The characteristic frequencies represent the two branches of viscous loss and heat exchange loss. and It represents the relaxation index of the two branches: viscous loss and heat exchange loss.

[0039] Therefore, the expression for the anisotropic-corrected transformation factor is obtained as follows: In the formula, Indicates the conversion factor. Represents absolute temperature. Indicates frequency, Indicates pore structure parameters, Parameters representing the compaction state of the coal body. A unit vector representing the direction of a sound ray; The transformation factor represents isotropic properties. Represents the anisotropic gain coefficient. Represents the unit vector in the vertical direction. This indicates the compaction index.

[0040] In another specific embodiment of the present invention, after obtaining the anisotropically corrected conversion factor, a microclimate correction factor is obtained, and the conversion factor is multiplied by the microclimate correction factor to obtain the linearly corrected conversion factor. The expression for obtaining the microclimate correction factor is: in, As a microclimate correction factor, Represents absolute temperature. Represents the mole fraction vector of gas components. Indicates pore structure parameters, This indicates the current relative humidity inside the coal bunker. Indicates reference relative humidity. Indicates the humidity sensitivity coefficient. Represents the reference gas component mole fraction vector. Indicates the sensitivity coefficient of gas components, superscript This represents the vector transpose operator.

[0041] The aforementioned conversion factor already includes the basic component, the frequency negative correlation relaxation component, and the direction correction component. This conversion factor can be used when constructing the dual-domain coupled equivalent acoustic-slow-motion model. In this embodiment of the invention, the dynamically changing gas composition and humidity parameters within the coal bunker are further considered, thereby constructing a microclimate correction factor to correct the conversion factor in real time. That is, the microclimate correction factor is designed for complex environments such as dust and airflow disturbances within the coal bunker. A dynamic compensation mechanism is established based on real-time parameters collected in the current coal bunker environment. By linearly correcting the conversion factor, the subsequently constructed dual-domain coupled equivalent acoustic-slow-motion model can be further refined, ensuring that the model can still output stably when environmental parameters fluctuate drastically. In this embodiment of the invention, this linear correction process is an optional optimization step, and the relevant content is not reflected in the subsequent formulas. In practical applications, it can be selected according to the environmental requirements within the coal bunker.

[0042] By multiplying the baseline acoustic slowness by the conversion factor, a two-domain coupled equivalent acoustic slowness model is constructed, and the model expression is as follows: In the formula, This is the dual-domain coupled equivalent acoustic-slow-motion model. For baseline slow sound, As the conversion factor, Represents absolute temperature. Represents the mole fraction vector of gas components. Indicates pore structure parameters, This indicates the current relative humidity inside the coal bunker. Indicates frequency, Parameters representing the compaction state of the coal body. A unit vector representing the direction of a sound ray.

[0043] The three-dimensional temperature field inside the coal bunker is reconstructed by iterative inversion solution based on the dual-domain coupled equivalent acoustic-slow-motion model and the measured flight time. In this embodiment of the invention, the specific steps of the iterative inversion solution include the following: First, the effective volume occupied by the coal in the coal bunker is discretized into a cubic voxel grid with a set side length, which can be adjusted according to the size of the bunker. At the same time, each voxel is assigned a set of initial state values, including temperature assumptions, pore structure parameters, compaction index, and environmental parameters. Then, based on the precise geometric coordinates of the acoustic transducer array and the voxel grid, the length of the straight line segment through each voxel for each acoustic path is calculated, thus forming a geometric weight matrix. This matrix establishes a mapping relationship between spatial distribution (voxel) and integral observation (path).

[0044] First, based on the assumed temperature field value within the coal bunker, the theoretical flight time is calculated using a dual-domain coupled equivalent acoustic slow-motion model. In this embodiment of the invention, using the assumed temperature value, i.e., the environmental parameters, the dual-domain coupled equivalent acoustic slow-motion model is invoked to perform forward calculations for each sound wave path at each emitting frequency point. The expression is as follows: In the formula, Indicates the first A sound wave path at frequency The theoretical flight time is as follows. The geometric traversal length weight represents the weight of the first traversal. The sound wave path is at the 1st The length of the straight line segment that passes through the interior of a voxel. Indicates that the baseline sound is slow. Indicates the conversion factor. Indicates the first The absolute temperature of a voxel. Indicates the first The relative humidity of individual elements, Indicates the first The mole fraction vector of gas components of an individual element. Indicates the first Pore ​​structure parameters of individual elements, Indicates the first Compaction state parameters of individual elements Indicates the first The unit vector representing the direction of propagation of a sound wave path.

[0045] Then, the flight time residual is obtained based on the theoretical flight time and the measured flight time, which is the difference between the measured flight time and the theoretical flight time.

[0046] With the goal of minimizing the time-of-flight residual, a regularization constraint consistent with the anisotropy of coal compaction is constructed to update the assumed temperature field values ​​within the current coal bunker. Through iterative inversion, the three-dimensional temperature field inside the coal bunker is output. This invention further converts the flight time residual into an equivalent length residual: In the formula, Indicates the first The sound wave path at the frequency point The equivalent length residual is below. To measure the actual flight time, Indicates the first The temperature assumption for the next iteration is... , No. The sound wave path at the frequency point The theoretical flight time is calculated below. Represents the speed of sound in the free domain. This is the conversion factor.

[0047] By converting the time-of-flight residual into an equivalent length residual, the inherent scale differences caused by medium dispersion at different frequencies are eliminated. Furthermore, for each acoustic path, the equivalent length residuals at each frequency point are weighted and fused to obtain the convergence residual of that path. The frequency weights used in the weighted fusion are... In this embodiment of the invention, frequency point weights are calculated by simultaneously weighting temperature sensitivity and measurement noise to form dimensionless and interpretable frequency point weights, expressed as follows: In the formula, Indicates the first The sound wave path at the frequency point Frequency weights, Indicates the first The sound wave path at the frequency point The confidence level of the flight time. This indicates the sensitivity of the equivalent acoustic slowness to temperature. Indicates the first The sound wave path at the frequency point The standard deviation of the flight time.

[0048] In this embodiment of the invention, for each voxel in the iteration, it is necessary to calculate the overall influence intensity of its temperature change on the equivalent acoustic slowness of all acoustic paths at all excitation frequencies, i.e., the global sensitivity amplitude. The calculation method is as follows: For each frequency point along each acoustic path, based on the assumed current temperature field, current environmental parameters, and pore structure parameters, the partial derivative of the equivalent acoustic slowness with respect to the voxel temperature at that frequency point is calculated using a dual-domain coupled equivalent acoustic slowness model, yielding its local sensitivity. Then, for all paths and all frequencies, the global sensitivity of the voxel is obtained by weighted summing the absolute values ​​of these local sensitivities, expressed as: In the formula, Indicates the first Global sensitivity of individual elements Represents the global weight of frequency points. This indicates the sensitivity of the equivalent acoustic slowness to temperature. Indicates the first Hypothetical temperature field values ​​for individual elements.

[0049] Based on the calculated global sensitivity amplitudes of all voxels, an N×N column scaling matrix is ​​constructed, represented as: In the formula, This represents a column scaling matrix.

[0050] Therefore, based on the linearized approximation of the model in the current state, a linear inversion equation is established: In the formula, For the unknowns after reparameterization Represents a column scaling matrix. This represents the desired voxel temperature increment.

[0051] In solving the linear inversion equation, this embodiment of the invention further constructs a direction-sensitive regularization term based on the anisotropic characteristics of coal compaction due to its own weight, and incorporates it as a constraint into the solution of the optimization problem. The specific steps are as follows: First, two key parameters need to be obtained: the compaction exponential function and the anisotropic gain coefficient. The compaction exponential function is a dimensionless function that describes the degree of coal compaction as a function of the height within the chamber. This function is usually obtained through calibration experiments. The anisotropic gain coefficient is a calibrable dimensionless parameter used to control the overall intensity of the anisotropic effect. It can be obtained through historical data inversion or set empirically.

[0052] The connection between all adjacent voxel pairs in a 3D voxel mesh is defined as an edge. Let its two voxel indices be... and For each edge, its anisotropic weight is calculated as follows: In the formula, Represents the anisotropic weights. For the set weights, The compaction index, Let the edge direction be the unit vector. Let be a unit vector in the vertical direction; where, for the horizontal side, Then the corresponding anisotropic weights This means that the smoothing constraint applied to that edge is weakened, which aligns with physical facts: in the horizontal direction, coal compaction is weaker, the pore structure is relatively uniform, and the actual temperature field may change gradually, thus allowing the algorithm to perform stronger smoothing to suppress noise in the horizontal direction; for the vertical edge, Then the corresponding anisotropic weights This means the smoothing constraint is strengthened. This is because in the vertical direction, the coal body experiences stronger compaction, and rising hot air easily forms a natural temperature gradient, meaning the true temperature field may vary significantly in the vertical direction. Strengthening the constraint means the algorithm penalizes temperature changes more heavily in this direction, thus protecting any potential true vertical gradients from being over-smoothed away. The constructed regularization direction is represented as: In the formula, Indicates the anisotropic regularization direction, To absorb scale weights, This represents the anisotropic weights.

[0053] During the iterative inversion solution, a two-stage decoupling operation of orthogonal complement space is performed simultaneously with solving the scaling domain inversion equation and applying anisotropic regularization constraints. The purpose is to decompose the parameter update space into a subspace strongly correlated with temperature and a subspace strongly correlated with environmental parameters, and to update different categories of parameters in stages and in an orderly manner in the two orthogonal subspaces, as follows: First, it is necessary to quantify the sensitivity of the equivalent length residual in the observation data to changes in environmental parameters. An environmental parameter vector containing relative humidity and gas component mole fractions for all voxels is defined. Similar to the temperature sensitivity calculation method, for each acoustic path, the gradient of its equivalent length residual with respect to environmental parameters is calculated, forming an environmental parameter sensitivity matrix. Each row of this matrix corresponds to a path, and each column corresponds to an environmental parameter. Then, an orthogonal complement projection matrix is ​​constructed, projecting the observation residual vector onto a subspace orthogonal to the space spanned by the columns of the environmental parameter sensitivity matrix. The temperature increment is solved within this subspace, ensuring that the solved temperature change does not attempt to "fit" the portion of the data that should be explained by changes in environmental parameters. The expression for the orthogonal complement projection matrix is: In the formula, Denotes the projection matrix of the orthogonal complement space. This represents the equivalent length-domain observation weight matrix. This represents the sensitivity matrix of the length domain to environmental parameters. It is the identity matrix. This indicates the environmental parameter damping.

[0054] In this embodiment of the invention, during each iterative inversion solution, based on the assumed temperature field value and corresponding environmental parameters of the current iteration, the current value set of the conversion factors of all relevant voxels and paths is calculated. At the same time, combined with the sensitivity matrix of the conversion factors to temperature obtained during the solution of the inversion equation, the linear approximate prediction value of the change of the conversion factors under the current temperature increment is calculated. Then, the mean or norm of the conversion factors of all relevant voxels and paths is selected as the normalized baseline conversion factor from the current value set of the conversion factors.

[0055] A trial step size is set to calculate the trial temperature field, and the conversion factor value under the trial temperature field is recalculated based on the dual-domain coupled equivalent acoustic-slow-motion model. This establishes a consistency trust domain step size control model for the conversion factor, calculates the linearization error under the trial step size, and dynamically adjusts the step size according to a preset threshold to ensure the physical validity and numerical stability of the iteration. The expression is: In the formula, Indicates a trial step size Consistency error of the conversion factor under the given conditions Normalized benchmark conversion factor, For the error threshold, This represents the set of current values ​​of the transformation factors for all relevant voxels and paths. This represents a linear approximate prediction of the change in the conversion factor under the current temperature increment. Indicates the trial step size The sensitivity matrix of the conversion factor to temperature, which has been calculated during the process of solving the inversion equation.

[0056] In this embodiment of the invention, after the iterative inversion solution converges, an approximate lower bound of the temperature estimation standard deviation is calculated for each voxel, thereby characterizing the reliability of the temperature reconstruction result at that location. The expression for this lower bound is: In the formula, Indicates the first Approximate lower bound of the standard deviation of temperature for individual voxels. Indicates the first Individual elements in the first The length of the sound wave path. The standard deviation of the flight time. This is the temperature sensitivity matrix. This is the equivalent regularization strength.

[0057] Finally, the approximate lower bounds of the temperature standard deviations of all voxels are combined into a confidence map isomorphic to the three-dimensional temperature field, which is then overlaid on the three-dimensional temperature field and displayed, for example, using transparency or contour lines, to intuitively understand which regions have more reliable temperature inversion results.

[0058] In another specific embodiment of the present invention, an acoustic non-destructive coal bunker three-dimensional temperature field imaging system is also proposed, the structural schematic diagram of which is shown below. Figure 5 As shown, the system includes: The sensing layer includes a multi-layer ring acoustic transducer array arranged along the coal bunker wall. The multi-layer ring acoustic transducer array contains multiple transducer nodes for transmitting low-frequency broadband acoustic signals in a predetermined sequence and receiving acoustic signals after penetrating the coal body. In embodiments of the present invention, such as Figure 2 and Figure 3 As shown, a multi-layer annular acoustic transducer array is arranged along the concrete bunker wall 1. A wall vibration isolation layer 2 is also installed between the coal bunker wall 1 and the transducer mounting base to attenuate the vibration transmitted by the bunker structure. The multi-layer annular acoustic transducer array can be set as a 3-layer annular array, located at the top, middle and bottom positions respectively. Each layer has 20 acoustic transducers 3 evenly installed on the circumference of the bunker wall. The nodes of adjacent layers are staggered by half a spacing in the horizontal direction to optimize the geometric diversity of the acoustic path.

[0059] The structure of the acoustic transducer 3 is as follows Figure 4 As shown, its core is a low-frequency broadband transducer unit (operating frequency band 400-1200Hz, directivity half-angle ≥45°). The front of the transducer is connected to an acoustic interface via a flange. This interface can be an acoustic window 5 composed of a dustproof sound-permeable membrane 15 and a metal protective mesh 16, or a short waveguide 4. The sound wave propagation channel 14 is coated with a sound-absorbing coating 13 to achieve acoustic coupling with the medium inside the chamber and isolate dust. The transducer housing integrates a power amplifier 8, an analog-to-digital converter 9, a microprocessor unit, and built-in auxiliary sensors, including a miniature temperature, humidity, and pressure probe 7 and an acceleration pickup 6. The transducer is installed through the chamber wall vibration isolation layer 2. The interface adopts a labyrinth-type sealed connection port 12 and a heating film 17 design, with an IP66 protection level, meeting the requirements of explosive dust environments.

[0060] The edge layer is connected to the sensing layer. Its core consists of an industrial-grade computer 10 and a supporting synchronous control and data acquisition unit located in a protective cabinet near the coal bunker. This unit includes a power amplifier 8 and an analog-to-digital converter 9. The power amplifier 8 receives instructions from the edge layer control module and generates a high-power excitation signal to drive the transducer to transmit. Then, the analog-to-digital converter 9 converts the analog signal from the power amplifier 8 into a digital signal with high resolution and high sampling rate, providing raw materials for subsequent digital signal processing. At the same time, its preamplifier section amplifies the received microvolt-level acoustic signal with extremely low noise. In addition, a comprehensive power supply and cabinet 11 are provided to provide a centralized, stable power supply that meets intrinsically safe explosion-proof requirements for the entire edge layer. The edge layer specifically includes a control module, a signal processing module, and an inversion calculation module. The control module is used to control the transmission and reception timing of the multi-layer ring acoustic transducer array. Based on the IEEE 1588v2 (PTP) precision clock protocol, it provides hardware-level time synchronization for all acoustic transducer nodes through industrial Ethernet, ensuring that the timing error across nodes is better than ±1µs. It also executes a pre-optimized "FDM / TDMA" hybrid scheduling algorithm to orderly excite different nodes to transmit linear sweep signals within a measurement frame of 0.5-1.0 seconds.

[0061] The signal processing module is used to process the received acoustic wave signal to extract the time-of-flight data of each path. This module extracts the time-of-flight (ToF) and confidence level of each acoustic wave path by running the "wavelet denoising + generalized cross-correlation (GCC-PHAT)" algorithm.

[0062] The inversion calculation module is used to construct and update the dual-domain coupled equivalent acoustic-slow-motion model and execute the iterative inversion reconstruction algorithm of the three-dimensional temperature field to obtain the reconstructed three-dimensional temperature field. The platform layer, which connects to the edge layer, includes data storage units, visualization units, and early warning units; Data storage unit, used to store real-time reconstructed three-dimensional temperature field data and historical three-dimensional temperature field data; The visualization unit is used to render and display the reconstructed three-dimensional temperature field; The early warning unit is used to identify hot spots in the coal bunker based on the reconstructed three-dimensional temperature field and trigger an alarm.

[0063] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An acoustic non-destructive coal bunker three-dimensional temperature field imaging method, characterized in that, The method comprises the following steps: arranging a multi-layer annular distributed acoustic transducer array in a coal bunker in a set manner; transmitting acoustic signals penetrating the coal body through the acoustic transducer array in multiple acoustic wave paths to obtain the measured flyover time of the acoustic signals in each acoustic wave path; obtaining coal bunker internal environment data to establish a dual-domain coupled equivalent acoustic slow model; the coal bunker internal environment data includes: air temperature, relative humidity, frequency, pore structure parameters and environmental gas parameters; the dual-domain coupled equivalent acoustic slow model specifically comprises: calculating the baseline acoustic slow of the free space inside the coal bunker according to the air temperature and the relative humidity by using a wet air sound speed model; constructing a conversion factor according to the pore structure parameters and the environmental gas parameters; constructing the dual-domain coupled equivalent acoustic slow model by multiplying the baseline acoustic slow and the conversion factor; iterative inversion solving according to the dual-domain coupled equivalent acoustic slow model and the measured flyover time to reconstruct the three-dimensional temperature field inside the coal bunker; the iterative inversion solving comprises: calculating the theoretical flyover time by using the dual-domain coupled equivalent acoustic slow model according to the temperature field assumption value in the current coal bunker; obtaining the flyover time residual according to the theoretical flyover time and the measured flyover time; minimizing the flyover time residual as the target, constructing a regularization constraint consistent with the compaction anisotropy of the coal body to update the temperature field assumption value in the current coal bunker, and outputting the three-dimensional temperature field inside the coal bunker through iterative inversion solving.

2. The acoustic non-destructive coal bunker three-dimensional temperature field imaging method according to claim 1, characterized in that: constructing a conversion factor according to the pore structure parameters and the environmental gas parameters, comprising: obtaining a frequency-independent basic component in the conversion factor according to the pore structure parameters; obtaining a frequency-negative correlation relaxation component in the conversion factor representing viscous loss and heat exchange loss according to the pore structure parameters and the environmental gas parameters; obtaining a direction correction component in the conversion factor related to the acoustic wave propagation direction according to the anisotropy of the coal body caused by self-weight compaction; constructing an isotropic conversion factor according to the basic component and the frequency-negative correlation relaxation component, and anisotropically modifying the isotropic conversion factor by using the direction correction component to obtain the conversion factor.

3. The acoustic non-destructive coal bunker three-dimensional temperature field imaging method according to claim 1, characterized in that: the expression of the conversion factor is: wherein denotes a conversion factor, denotes an absolute temperature, denotes a frequency, denotes a pore structure parameter, denotes a compaction state parameter of the coal body, denotes a unit vector of the sound ray direction; denotes an isotropic conversion factor, denotes an anisotropic gain coefficient, denotes a unit vector of the vertical direction, denotes a compaction exponent.

4. The acoustic non-destructive coal bunker three-dimensional temperature field imaging method according to claim 3, characterized in that: the isotropic conversion factor is expressed as: wherein represents the isotropic conversion factor, represents the absolute temperature, represents the frequency, represents the pore structure parameter, is the high frequency tortuosity, is the porosity, and represents the amplitude coefficients of the two branches of viscous and thermal exchange losses, and represents the characteristic frequencies of the two branches of viscous and thermal exchange losses, and represents the relaxation exponents of the two branches of viscous and thermal exchange losses.

5. The acoustic non-destructive coal bunker three-dimensional temperature field imaging method according to claim 4, characterized in that: the characteristic frequencies of the two branches of the viscous loss and the heat exchange loss are respectively expressed as: wherein represents a characteristic frequency of viscous losses, represents an absolute temperature, represents a vector of molar fractions of gas components, represents a parameter of pore structure, represents a dimensioning coefficient of viscous losses, represents a flow resistivity of static viscous channels, is a porosity, represents a viscous characteristic length, is a high-frequency tortuosity, represents a wet air viscosity, represents a wet air density; represents a characteristic frequency of heat exchange losses, represents a dimensioning coefficient of heat exchange losses, represents a flow resistivity of static heat channels, represents a heat characteristic length, represents a Prandtl number.

6. The acoustic non-destructive coal bunker three-dimensional temperature field imaging method according to claim 1, characterized in that: after anisotropically modifying the isotropic conversion factor by using the direction correction component to obtain the conversion factor, it further comprises: obtaining a microclimate correction factor, and multiplying the microclimate correction factor and the conversion factor to obtain a linearly corrected conversion factor; the expression of the microclimate correction factor is: wherein is a microclimate correction factor, denotes the absolute temperature, denotes the gas composition molar fraction vector, denotes the pore structure parameter, denotes the current relative humidity inside the coal bunker, denotes the reference relative humidity, denotes the humidity sensitivity coefficient, denotes the reference gas composition molar fraction vector, denotes the gas composition sensitivity coefficient, superscript denotes the vector transposition operator.

7. The acoustic non-destructive coal bunker three-dimensional temperature field imaging method according to claim 1, characterized in that: calculating the theoretical flyover time by using the dual-domain coupled equivalent acoustic slow model according to the temperature field assumption value in the current coal bunker is expressed as: In the formula, Indicates the first A sound wave path at frequency The theoretical flight time is as follows. The geometric traversal length weight represents the weight of the first traversal. The sound wave path is at the 1st The length of the straight line segment that passes through the interior of a voxel. Indicates that the baseline sound is slow. Indicates the conversion factor. Indicates the first The absolute temperature of a voxel. Indicates the first The relative humidity of individual elements, Indicates the first The gas component mole fraction vector of individual elements. Indicates the first Pore ​​structure parameters of individual elements, Indicates the first Compaction state parameters of individual elements Indicates the first The unit vector representing the direction of propagation of a sound wave path.

8. The acoustic non-destructive coal bunker three-dimensional temperature field imaging method according to claim 1, characterized in that: the iterative inversion solving process comprises: converting the flyover time residual into an equivalent length domain residual, and weighting and fusing the equivalent length domain residual according to the temperature sensitivity of each frequency point and the measurement noise to obtain an observation residual vector for inversion; establishing an optimization problem with the minimum observation residual vector as the target, and introducing a regularization constraint consistent with the compaction anisotropy of the coal body; obtaining the update amount of the three-dimensional temperature field by solving the optimization problem with the regularization constraint. The current temperature field assumption value is corrected by using the updated amount of the three-dimensional temperature field to obtain an updated three-dimensional temperature field; Iterate in sequence until a preset convergence condition is met, and output the reconstructed three-dimensional temperature field.

9. An acoustic non-destructive coal bunker three-dimensional temperature field imaging system, which performs an acoustic non-destructive coal bunker three-dimensional temperature field imaging method according to any one of claims 1-8, characterized in that, Comprise: A perception layer comprising a multilayer annular acoustic transducer array arranged along the wall surface of the coal bunker, the multilayer annular acoustic transducer array comprising a plurality of transducer nodes for transmitting low-frequency broadband acoustic signals and receiving acoustic signals after penetrating the coal body in a predetermined sequence; An edge layer connected with the perception layer, comprising a control module, a signal processing module and an inversion calculation module; The control module is used to control the transmission and reception timing of the multilayer annular acoustic transducer array; The signal processing module is used to process the received acoustic signals to extract the flyover time data of each path; The inversion calculation module is used to construct and update the dual-domain coupled equivalent acoustic slow model, and perform an iterative inversion reconstruction algorithm of the three-dimensional temperature field to obtain the reconstructed three-dimensional temperature field; A platform layer connected with the edge layer, comprising a data storage unit, a visualization unit and a warning unit; The data storage unit is used to store real-time reconstructed three-dimensional temperature field data and historical three-dimensional temperature field data; The visualization unit is used to render and display the reconstructed three-dimensional temperature field; The warning unit is used to identify hot spots in the coal bunker according to the reconstructed three-dimensional temperature field and trigger an alarm.

Images