Unmanned aerial vehicle acoustic monitor based on metamaterial reinforcement and fault detection method

Abstract

The invention discloses an unmanned aerial vehicle acoustic monitor based on metamaterial reinforcement and a fault detection method. The invention comprises an unmanned aerial vehicle acoustic monitor based on metamaterial reinforcement and a fault detection method. The unmanned aerial vehicle acoustic monitor based on metamaterial reinforcement comprises an acoustic sensing system and a trumpet-shaped acoustic metamaterial structure installed in a load area below an unmanned aerial vehicle, and the acoustic metamaterial structure comprises a plurality of circular plates with the diameters linearly increased from bottom to top and a bar penetrating through the centers of all the circular plates. The acoustic sensing system comprises an acquisition card and a plurality of acoustic sensors, one acoustic sensor is mounted in a gap between every two adjacent circular plates, and all the acoustic sensors are electrically connected with the acquisition card. The method is applied to the technical field of unmanned aerial vehicle remote sensing.

Description

Technical Field

[0001] The present invention is applied to the technical field of unmanned aerial vehicle remote sensing, and in particular relates to an unmanned aerial vehicle acoustic monitor and a fault detection method based on metamaterial reinforcement. Background Art

[0002] Drone remote sensing technology has become a transformative approach that enhances traditional condition monitoring and inspection methods by providing unprecedented speed, efficiency, and versatility over a variety of terrains. Drone remote sensing technology is increasingly being used in areas such as infrastructure health monitoring, railway and traffic surveillance, forest and fire management, photovoltaic energy system inspection, and wind turbine evaluation. Current applications mainly rely on optical sensing modes such as carrying high-resolution cameras, thermal imaging, multispectral sensors, and laser scanning, but these methods face limitations in situations involving visual obstructions or internal defects of the monitored object. In such cases, acoustic-based drone remote sensing technology offers a promising alternative, such as in detecting internal mechanical faults such as wind turbine drivetrains.

[0003] UAV acoustic remote sensing technology has great potential, but its widespread adoption in industrial environments is severely limited. The main problem is the strong background noise generated by the propulsion system of the drone during flight, which is characterized mainly by the tones and broadband components of the propeller blade pass frequency. Moreover, this complex noise environment is more serious in industrial drones, because industrial drones usually have four or more motors and propeller units, which makes it difficult for airborne acoustic sensors to isolate the target sound source from the noise generated by the drone itself. This usually results in the signal-to-noise ratio of the collected signal being less than -15dB, making traditional sound analysis algorithms completely ineffective, which seriously restricts the practical application of UAV acoustic remote sensing technology.

[0004] To overcome these limitations, a large number of technical solutions are currently devoted to optimizing the signal processing of multi-microphone arrays using advanced beamforming techniques. Specifically, this method uses the spatial distribution and phase relationship between microphone array elements to enhance sound source estimation and mitigates uncorrelated noise through phase difference correction to improve the signal-to-noise ratio. In the past decade, some substantial progress has been made in developing complex airborne microphone array systems and improving the related signal processing algorithms, but some shortcomings still exist.

[0005] According to the theory of beamforming methods, increasing the number of microphones and the array aperture can significantly improve the sensing performance. Therefore, it is beneficial to improve the signal-to-noise ratio by expanding the design of a larger number of microphones or by expanding the array framework to maximize the phase difference. However, these technologies bring new limitations, especially the increased computational requirements for real-time multi-channel audio processing, the reduced payload capacity of drones, and the physical limitations of aerodynamics. The corresponding alternative strategy is to expand the heterogeneous microphone array configuration or special connection mechanism so that the microphone system is relatively far away from the drone noise source, thereby directly improving the signal-to-noise ratio of the target source signal. Although these technologies have a certain enhancement effect, they often change the center of gravity of the drone, require complex flight control compensation, and potentially increase the risk of flight operations due to unstable center of gravity. In addition, there are also microphone array signal calculation methods assisted by machine learning that are also expected to enhance the signal-to-noise ratio, but this method also faces related limitations such as model generalization and further airborne computing resource constraints. Moreover, to date, most drone acoustic remote sensing technologies are also mainly limited to the application of sound source positioning, and cannot obtain target source signals with high signal-to-noise ratios, and cannot provide effective solutions for actual industrial monitoring, fault inspection and other application scenarios. Summary of the invention

[0006] The technical problem to be solved by the present invention is to overcome the shortcomings of the prior art and provide a UAV acoustic monitor and fault detection method based on metamaterial reinforcement. The UAV acoustic monitor reinforced by metamaterial uses an artificially designed acoustic metamaterial structure to adjust the sound field around the UAV, and obtains frequency selectivity enhancement of the target source signal by adopting a highly anisotropic acoustic metamaterial structure. At the same time, the strong directional sensitivity of the acoustic metamaterial structure is used to suppress incoherent noise, especially severe noise interference from the UAV. High signal-to-noise ratio sound signal collection can be achieved in UAV acoustic sensing, and efficient fault detection can be achieved by using only a single-channel signal and a simple processing algorithm.

[0007] The technical solution adopted by the present invention is: the present invention includes a UAV acoustic monitor based on metamaterial reinforcement and a fault detection method. The UAV acoustic monitor based on metamaterial reinforcement includes an acoustic sensing system and a horn-shaped acoustic metamaterial structure installed in the load area below the UAV, the acoustic metamaterial structure includes a plurality of circular plates with linearly increasing diameters from bottom to top and a lever passing through the center of all the circular plates, the acoustic sensing system includes an acquisition card and a plurality of acoustic sensors, each of the gaps between adjacent circular plates is equipped with an acoustic sensor, and all the acoustic sensors are electrically connected to the acquisition card.

[0008] Furthermore, the circular plate is made of polylactic acid material by 3D printing.

[0009] Furthermore, the intervals between adjacent circular plates are equal.

[0010] Furthermore, the acquisition card is electrically connected to an integrated control board of the UAV, and the integrated control board is electrically connected to an onboard computer system interface of the UAV.

[0011] Furthermore, the number of the circular plates is 18, the thickness of the circular plates is 3 mm, the spacing between adjacent circular plates is 10 mm, the radius of the largest circular plate is 110 mm, and the radius of the smallest circular plate is 10 mm.

[0012] A fault detection method applied to the aforementioned metamaterial-reinforced UAV acoustic monitor comprises the following steps:

[0013] S1. Set the inspection path of the drone, and set a number of inspection points on the inspection path to cover the required inspection object. When the drone flies to the set inspection point, it hovers and collects the audio at the current inspection point through the drone acoustic monitor based on metamaterial reinforcement to judge the state of the inspection object;

[0014] S2. At the detection point of the current detection object, the drone is set to hover within a cone range of 30 degrees above the detection object to collect audio data, so as to make full use of the enhanced performance of the metamaterial, and a number of acoustic sensors of the drone acoustic monitor based on the metamaterial enhancement simultaneously collect the acoustic signals on the detection object to form the audio to be analyzed;

[0015] S3 extracts audio point data on the audio to form an audio data set corresponding to different enhanced frequency bands of the metamaterial;

[0016] S4. Select the audio data of the required frequency band for time-frequency pre-analysis according to the physical characteristics of the detection object;

[0017] S5. Further filter and envelope analyze the current audio data group, by using a bandpass filter to perform fixed-band filtering and then perform Hilbert transform to obtain an envelope spectrum, and identify the fault characteristic frequency through the envelope spectrum to determine the fault state of the detection object.

[0018] The beneficial effects of the present invention are: 1. Unlike traditional microphone (i.e., acoustic sensor) array technology, the UAV acoustic monitor based on metamaterial reinforcement achieves both compact design and acoustic enhancement. This innovative design structure effectively reduces the serious interference of noise caused by the UAV, while eliminating the need to increase the number and spacing of microphones to improve the signal-to-noise ratio. It also optimizes the center of gravity of the UAV and enhances the overall flight stability and payload.

[0019] 2. It has a special frequency-selective enhancement characteristic, which can effectively extract subtle mechanical fault features even in high noise conditions with a signal-to-noise ratio as low as -20dB; in addition, its powerful directional enhancement characteristics reduce the interference of irrelevant signals, providing a solid foundation for fault identification and positioning during UAV detection.

[0020] 3. This method can effectively use single-channel acoustic data to detect mechanical faults without relying on complex signal processing algorithms for noise reduction or optimal frequency band search, which greatly reduces the hardware requirements and computing load of the onboard computer, and provides a feasible solution for online monitoring and fault diagnosis in UAV industrial inspections. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a schematic diagram of the structure of the UAV acoustic monitor;

[0022] Figure 2 (a) is a schematic diagram of the two-dimensional effective medium of the metamaterial. Figure 2 (b) is a schematic diagram of the sound pressure distribution at different incident frequencies in the metamaterial medium;

[0023] Figure 3 (a) is a schematic diagram of the numerical simulation settings and the pressure field distribution obtained at incident frequencies of 1350 Hz, 1490 Hz, and 1710 Hz. Figure 3 (b) is the normalized pressure-frequency response in the 2nd, 4th and 6th gaps, showing the schematic diagram of broadband acoustic enhancement;

[0024] Figure 4 (a) Schematic diagram of the measurement setup of directional enhancement properties through numerical simulation and experiment. Figure 4 (b) Figure 4 (c) and Figure 4 (d) is a schematic diagram showing the normalized directional responses at the 2nd, 4th and 6th gaps respectively;

[0025] Figure 5 (a) is a diagram of an experimental setup using synthetic signals to evaluate the present invention, Figure 5 (b) is a schematic diagram of the UAV acoustic measurement position;

[0026] Figure 6 (a) is a schematic diagram showing the comparison of the time domain waveforms obtained in free space and in the sixth gap of the metamaterial of the present invention, Figure 6 (b) is the spectrum corresponding to the signal obtained inside the metamaterial. Figure 6 (c) provides a comparison diagram of the envelope spectra corresponding to the two groups of signals;

[0027] Figure 7This is the envelope spectrum of the gear fault signal obtained by the UAV acoustic monitor based on metamaterial reinforcement at different distances and incident angles; Figure 7 (a) At a vertical distance of 4 meters directly above the sound source, Figure 7 (b) 4 meters vertical distance from the sound source with an incident angle of 26.6°, Figure 7 (c) At a vertical distance of 9 m directly above the sound source, Figure 7 (d) 4 meters vertical distance from the sound source with an incident angle of 45°, Figure 7 (e) At a vertical distance of 15 m directly above the sound source, Figure 7 (f) The vertical distance from the sound source is 4 meters with an incident angle of 56.3°. DETAILED DESCRIPTION

[0028] In this embodiment, the present invention includes a UAV acoustic monitor and fault detection method based on metamaterial reinforcement. Figure 1 As shown, the UAV acoustic monitor based on metamaterial reinforcement includes an acoustic sensing system and a horn-shaped acoustic metamaterial structure installed in the load area below the UAV 1, the acoustic metamaterial structure includes 18 circular plates 2 with diameters increasing linearly from bottom to top and arranged coaxially and a lever 3 passing through the center of all the circular plates 2, the lever 3 is a solid lever, the acoustic sensing system includes an acquisition card 4 and a plurality of acoustic sensors 5, the gaps between adjacent circular plates 2 are all integrated with an identical micro-electromechanical system acoustic sensor 5, and all the acoustic sensors 5 are electrically connected to the acquisition card 4. The acquisition card 4 is electrically connected to the integrated control board of the drone 1, and the integrated control board is electrically connected to the onboard computer system interface of the drone 1; the sum of the weights of the acoustic metamaterial structure and the acoustic sensor system is approximately 0.77 kg, which only accounts for 22% of the payload capacity of the drone. In addition, since the acoustic performance of the acoustic metamaterial structure is mainly affected by the geometric structure and is limitedly affected by the basic material, the use of a lower density rigid material can further reduce the weight of the system and ensure that the system performance does not change. Preferably, the circular plate 2 is made of polylactic acid material by 3D printing.

[0029] In this embodiment, the intervals between adjacent circular plates 2 are all 10 mm, the thickness of the circular plates 2 is 3 mm, the maximum radius of the circular plates 2 is 110 mm, and the minimum radius of the circular plates 2 is 10 mm.

[0030] In this embodiment, in order to briefly explain the function of the acoustic metamaterial, Figure 2 (a) shows the proposed acoustic metamaterial structure conceptualized as a two-dimensional representation, where the acoustic wave propagates along the positive x-axis. Using the effective medium model, the circular plate array with scattered air can be characterized as a continuous, lossless medium. Therefore, the effective mass density along the x-axis and y-axis is ρx and ρ y , and the bulk modulus K of the acoustic metamaterial structure can be expressed as:

[0031] ρ x =F r ρ plate +(1-F r )ρ air

[0032]

[0033] Among them, F r =t / (g+t) is the filling rate of the circular plate 2, and the material property of the circular plate 2 is defined as the density ρ plate =1250kg / m 3 , the bulk modulus K of circular plate 2 plate =3.8GPa. The density of the surrounding air medium is ρ air =1.2kg / m 3 , the bulk modulus K of the surrounding air air =1.4*10 -5 GPa. This metamaterial model assumes that the entire structure can be decomposed into a series of infinitesimal cross-sections composed of uniform metamaterial plates, and the effective refractive index change of the metamaterial can be further deduced as:

[0034]

[0035] In the above formula, ω = 2πf represents the angular frequency of the incident sound wave, where n air is the refractive index of the incident air, usually approximated by n air =1. y(x) represents the width of the metamaterial plate. Assuming that it is a continuous function along the propagation axis, the gradual increase in the refractive index will lead to wavelength compression, which further leads to the spatial concentration of acoustic energy. Using the change in effective refractive index, the relationship between the amplitude of the acoustic pressure inside the metamaterial and the incident frequency along the x-axis direction can be expressed as:

[0036]

[0037] Based on the above formula of sound pressure amplitude and incident frequency, Figure 2 (b) shows the distribution of acoustic pressure amplitude at different locations within the metamaterial medium within the frequency range of the incident wave, illustrating the ability of this acoustic metamaterial structure to selectively enhance the acoustic pressure amplitude at different frequencies in its spatial domain. This behavior allows for precise frequency-based acoustic manipulation and sensing, which is critical to the proposed method of the UAV acoustic monitor based on metamaterial enhancement.

[0038] According to the above principle, the propagation of sound waves in the designed structure produces compression and enhancement of different frequencies in different intervals of the metamaterial circular plate 2. In order to further illustrate the enhancement characteristics of the proposed acoustic sensing system, a three-dimensional finite element model was established using the commercial finite element software COMSOL Multiphysics v6.0, and the pressure acoustics module was used for numerical simulation calculations. Figure 3 (a) On the left, the model uses a Plane Wave Radiation boundary condition to simulate the incident acoustic wave, while perfectly matched layers are applied on all external boundaries to mitigate reflection interference. Figure 3 The right side of (a) shows the pressure field distribution inside the metamaterial structure at 1350 Hz, 1490 Hz, and 1710 Hz. This result shows that the pressure amplitude inside the metamaterial gap is significantly higher than that in the free field, and the spatial concentration of acoustic energy is related to the incident frequency. In order to quantitatively characterize the frequency selective enhancement characteristics, Figure 3 (b) further gives the normalized pressure-frequency response in the 2nd, 4th and 6th gaps (where the gap refers to the interval between adjacent circular plates 2), which show that the enhanced peaks are obtained at frequencies of 1357 Hz, 1492 Hz and 1708 Hz, respectively. It can be seen from these three representative interval positions that the pressure enhancement caused by the sound wave compression is manifested as broadband sound enhancement, which is particularly beneficial in practical engineering applications.

[0039] In addition, due to the high anisotropy of the acoustic metamaterial structure, it is sensitive to the incident angle of the sound wave. In order to further clarify the directional sensitivity and directional enhancement characteristics of the UAV acoustic monitor, the following studies were conducted: Figure 4 Directional response measurement shown in (a). Based on the above numerical simulation setup, the incident wave angle is fixed and the acoustic metamaterial structure is rotated in 1 degree increments while extracting the change in the sound pressure amplitude in the gap. Figure 4 (b), (c), and (d) show the normalized directional response results in the 2nd, 4th, and 6th gaps, respectively, and the incident wave frequency corresponds to the peak frequency of each gap. Figure 4 As shown in (b), (c), and (d), when the sound wave is incident along the positive direction of the acoustic metamaterial tip, the sound pressure amplitude is the largest, and as the incident angle increases, the sound pressure amplitude decreases rapidly. When the incident angle reaches 45° or 315°, the pressure amplitude drops to about 1 / 5 of the maximum value. Further increase in the incident angle of the sound wave will lead to a continuous decrease in the pressure amplitude, reaching a minimum when the sound wave is incident vertically on the metamaterial circular plate 2 (i.e., 90° or 270°). This is because for vertical sound wave incidence, the equivalent medium of the metamaterial can be approximated as air. In addition, due to the severe impedance mismatch at the metamaterial-air interface, the pressure amplitude of the sound wave incident from the reverse direction (>90° and <270°) is significantly lower than that of the sound wave incident from the positive direction.

[0040] Figure 4(a) Measurement of directional enhancement properties through numerical simulation and experiment: Schematic diagram of the directional response measurement setup. In the simulation, the metamaterial circular plate 2 is rotated in increments of 1 degree; in the experiment, the drone acoustic monitor is rotated in increments of 5 degrees under a fixed sound source.

[0041] This obvious directional sensitivity and directional enhancement characteristics provide favorable conditions for the application of UAV acoustic remote sensing technology. First, according to the UAV acoustic monitor based on metamaterial reinforcement proposed by the present invention, when the UAV collects signals above the monitored target, the target sound source signal is radiated from the positive direction to the metamaterial system and is selectively enhanced. At the same time, the noise generated by the UAV itself or other irrelevant noise is mainly radiated to the metamaterial system from the side or rear, and the gain is significantly reduced or even suppressed compared with the target source signal. This differential enhancement plays a vital role in improving the signal-to-noise ratio of the target source signal. Secondly, the strong directional sensitivity of the metamaterial also ensures that the target source signal can only obtain the maximum gain when it is perpendicular to the bottom of the UAV system. This directional selectivity effectively prevents the enhancement of other irrelevant noises during UAV monitoring, reduces the risk of false detection, and provides an important basis for accurately locating the source of the fault.

[0042] Drawing on the unique wave modulation characteristics of the acoustic metamaterial, it obtains frequency selectivity enhancement characteristics in signal collection at different positions within the structure, and its directional sensitivity characteristics further suppress the interference of drone noise on signal collection. This innovative fault detection method based on metamaterial-enhanced drone acoustic monitor improves the signal-to-noise ratio of the target source signal and can effectively extract useful information from complex background noise without the need for complex signal processing algorithms, such as the detection of mechanical internal fault signals and the extraction of fault features. In contrast, traditional signal measurement methods are severely affected by the noise and environmental interference generated by drones, making it very difficult to extract relevant fault signal features. Next, its beneficial effects are briefly introduced through two embodiments.

[0043] In order to evaluate the effectiveness and fault detection capability of the UAV acoustic monitor based on metamaterial reinforcement, the present invention uses a mechanical fault synthetic signal that can accurately quantify the signal-to-noise ratio. Figure 5 As shown in (a), a synthetic signal is emitted by a high-fidelity speaker to simulate the fault source. The prototype of the UAV acoustic monitor based on metamaterial reinforcement was deployed to fly and hover at an altitude of approximately 4 meters, collecting sound signals directly above the sound source with a sampling frequency of 48kHz. The integrated control board is directly connected to the onboard computer system interface of the UAV, while the ground station accesses the onboard computer and controls the collection of acoustic signals through a wireless network. In addition, the present invention also tests the changes in the fault detection performance of the UAV acoustic monitor based on metamaterial reinforcement at different distances and angles to demonstrate its potential for fault location through directional enhancement characteristics. As shown in Figure 5 As shown in (b), six strategically selected detection positions are established. Positions 1-3 represent the drone hovering directly above the sound source, with vertical distances of 4 meters, 9 meters, and 15 meters, respectively. Positions 4, 5, and 6 have a vertical distance of 4 meters from the sound source, while the horizontal distances are 2 meters, 4 meters, and 6 meters, respectively, corresponding to incident angles of 26.6 degrees, 45 degrees, and 56.3 degrees, respectively.

[0044] Figure 5 (a) Experimental setup for evaluating the present invention using synthetic signals. The setup demonstrates the practical application of fault detection in an outdoor environment. (b) Schematic diagram of the UAV acoustic measurement positions. Six UAV positions (numbered 1-6) are shown at different heights and horizontal distances from the sound source.

[0045] In the first embodiment, the formula of the bearing fault signal can be expressed as:

[0046]

[0047] The above formula can be used to generate a rolling bearing fault simulation signal, where f r1 =8Hz and f n =1700Hz represents the rotation frequency and natural frequency of the bearing respectively. Wherein, A=4 and ζ=100 represent the amplitude coefficient and damping coefficient respectively. 1 =mod(t,0.99 / f d1 ), where f d1 =22.5Hz is the characteristic frequency of the bearing fault. At the same time, Gaussian white noise is used, and n(t) controls the actual signal-to-noise ratio of the signal. In the experimental stage, the UAV acoustic monitor based on metamaterial reinforcement initially collected a 0dB bearing fault signal at position 1 for evaluation, with a sampling time of 5 seconds. Since the reinforcement band of the sixth gap of the metamaterial is consistent with the natural frequency of the bearing, f n The closest distance is conducive to amplifying the carrier signal and further enhancing the detectability of the fault characteristic frequency. Therefore, the sound signal corresponding to gap 6 is used for in-depth analysis. For comparison, an identical acoustic sensor is placed in free space outside the metamaterial to collect a reference signal for performance comparison.

[0048] Figure 6 Comparative analysis of bearing acoustic signals with a signal-to-noise ratio of 0 dB, where the left side is the signal collected in free space, and the right side is the signal collected in a UAV acoustic monitor based on metamaterial reinforcement: (a) time domain waveform, (b) frequency spectrum and (c) envelope spectrum.

[0049] Figure 6(a) shows the time domain waveform comparison between the free space and the metamaterial gap 6 of the present invention. The waveform collected inside the metamaterial not only has a significantly increased sound pressure amplitude, but also has obvious periodic impact characteristics, and the impact frequency corresponds precisely to the fault characteristic frequency of the bearing. The signal spectrum measured in free space, such as Figure 6 (b) As shown on the left, two closely spaced peaks are displayed around 140 Hz, identified as the blade pass frequency of the UAV propeller. Detailed spectral analysis shows that the harmonics of these frequency components are distributed throughout the spectrum, illustrating the extensive impact of the UAV's self-generated noise on the target signal. The presence of two different frequency peaks is mainly attributed to the balancing behavior of the UAV in response to wind disturbances and other unstable factors. This behavior is characterized by differential motor acceleration to produce compensating lift, further complicating the acoustic field and its interference with signal acquisition. The high signal-to-noise ratio of the current sound source ensures that the natural frequency of the bearing and its sideband components can still be observed. The spectrum corresponding to the signal obtained inside the metamaterial, such as Figure 6 (b) As shown in Figure 2, due to the low-frequency acoustic diffraction phenomenon, the blade passing frequency of the drone and its related harmonics can still be observed in the spectrum, but their amplitude has been reduced to a certain extent compared to the natural space. In addition, the signal component centered on the bearing natural frequency and its sideband components show significant amplification.

[0050] According to the conventional bearing fault diagnosis method, the collected signals are further processed, including targeted filtering and demodulation, to obtain the corresponding envelope spectrum. Then the fault diagnosis is performed based on the detailed analysis of the fault frequency components in the envelope spectrum. Considering the serious interference of low-frequency drone sensing elements and the optimized enhanced frequency band provided by metamaterials, a uniform filtering frequency band of 1000-3000Hz is selected. Therefore, Figure 6 (c) provides a comparison of the envelope spectra corresponding to the two sets of signals. It is clear that the bearing fault frequency and its associated harmonics in the signal obtained within the metamaterial are amplified by more than 60 times compared to the signal obtained in free space. In addition, the high-order harmonic components can be clearly observed in the signal within the metamaterial, providing rich fault diagnosis information. It is worth noting that although high-pass filtering is used to remove the low-frequency UAV blade pass frequency, the envelope spectrum of the free space signal is still significantly affected by the high-order harmonics of the blade pass frequency. In sharp contrast, the blade pass frequency component in the signal within the metamaterial is effectively suppressed, and there are no identifiable components in the envelope spectrum, demonstrating the superior performance of the present invention.

[0051] In the second embodiment, the formula of the gear fault signal can be expressed as:

[0052]

[0053] In order to further verify the robustness of the present invention, the above formula can be used to generate a gear fault simulation signal, where f r2 =12.9Hz and f m =750Hz represents the rotation frequency and gear meshing frequency respectively. N = (0, 1, 2, 3, 4, 5) is the harmonic number of amplitude modulation and frequency modulation, while M = (1, 2, 3) represents the harmonic number of meshing frequency. m =[0.8, 0.6, 0.4] is the mth harmonic amplitude of the meshing frequency. n is the harmonic order of amplitude modulation and frequency modulation, B n =[0.8, 0.6, 0.4, 0.2, 0] and C n = [0, 0, 0, 0, 0] are the nth harmonic amplitudes of FM and AM respectively. Similarly, the Gaussian white noise n(t) controls the actual signal-to-noise ratio. In addition, it should be emphasized that the meshing frequency f m The modulation sidebands correspond to the rotation frequency f r2 The frequency is the gear fault characteristic frequency, i.e., f d2 =12.9Hz.

[0054] As described in the above experimental setup, in the experimental stage, the present invention uses the gear fault signal to analyze the changes in fault detection performance at different distances and angles. The signal-to-noise ratio is adjusted to -10dB using Gaussian noise. In general, although the amplitude of the higher-order harmonics of the meshing frequency is low, they are not easily affected by noise pollution and may reflect the fault characteristics. Therefore, the enhanced signal at the metamaterial gap 4 is selected for analysis, which is close to the second harmonic of the meshing frequency (2f m =1500Hz).

[0055] The signal envelope spectra obtained by the UAV acoustic monitor based on metamaterial reinforcement at six selected locations are as follows: Figure 7 As shown. Among them, Figure 7 (a), (c) and (e) are the measurement results at 4m, 9m and 15m vertical distances above the sound source, respectively. It can be seen that the distance is inversely proportional to the sound pressure amplitude. However, the gear fault characteristic frequency obtained by the present invention is still significantly higher than the noise, even at a longer distance. In addition, Figure 7 (b), (d), and (f) show the results corresponding to increasing incident angles relative to the sound source. When the incident angle increases to about 26 degrees, a significant attenuation of the sound pressure amplitude by about 5 times is observed, although the vertical distance from the sound source is only 4 meters, but the gear fault characteristic frequency is still clearly visible. As the incident angle further increases to 45 degrees and above, the sound pressure amplitude stabilizes at a low level, but the gear fault characteristic frequency is difficult to distinguish from the background noise.

[0056] Directional sensitivity analysis of a metamaterial-enhanced UAV acoustic monitor reveals operational criteria for fault detection applications. Optimal fault detection performance is achieved when the position of the detected object relative to the UAV is kept within a 30-degree cone during acoustic signal acquisition. Since industrial UAVs typically have real-time kinematic (RTK) positioning systems with centimeter-level accuracy, operators can precisely control the monitoring direction. This operational constraint enhances the detection of potential fault signatures while minimizing false identification from external signal sources, demonstrating the great potential of the invention for precise fault localization in industrial applications.

[0057] Finally, given the remarkable performance of the method of the present invention in mechanical fault detection, it shows good application potential in other fault detection or signal enhancement scenarios. These potential fields include but are not limited to pipeline leakage detection, structural crack identification and marine exploration.

[0058] More specifically, a fault detection method for the UAV acoustic monitor based on metamaterial reinforcement includes the following steps:

[0059] S1. Set the inspection path of the drone 1, and set a number of inspection points on the inspection path to cover the required inspection object. When the drone 1 flies to the set inspection point, it hovers and collects the audio at the current inspection point through the drone acoustic monitor based on metamaterial reinforcement to judge the state of the inspection object;

[0060] S2. At the detection point of the current detection object, the drone 1 is set to hover within a cone range of 30 degrees above the detection object to collect audio data, so as to make full use of the enhanced performance of the metamaterial, and a plurality of acoustic sensors 5 of the drone acoustic monitor based on the metamaterial enhancement are used to simultaneously collect the acoustic signal on the detection object to form the audio to be analyzed; for example, at the current detection point, according to the requirements of frequency resolution, the duration of each audio can be 5-30 seconds;

[0061] S3. Extracting audio point data on the audio to form an audio data group corresponding to different enhanced frequency bands of the metamaterial; taking one of the audio in the metamaterial interval as an example, collecting audio point data on the audio at a certain time interval starting from 0 seconds, and all audio point data are arranged from low frequency to high frequency according to the enhanced frequency band corresponding to the metamaterial gap to form an audio data group;

[0062] S4. Select the audio data of the required frequency band for time-frequency pre-analysis according to the physical characteristics of the detection object; for example, the enhanced frequency band of the sixth gap of the metamaterial in this embodiment 1 is closest to the bearing natural frequency of the detection object, and has the effect of enhancing weak fault signals. Therefore, the audio data group corresponding to the sixth interval is selected for time-frequency analysis, and the signal-to-noise ratio and effectiveness of the signal are preliminarily determined by identifying and judging the fault signal impact characteristics in the time domain signal and the sideband characteristics in the frequency domain signal;

[0063] S5. Further filter and envelope analyze the current audio data group, perform fixed-band filtering with a bandpass filter and then perform Hilbert transformation to obtain an envelope spectrum, identify the fault characteristic frequency through the envelope spectrum to determine the fault state of the detection object. For example, in this embodiment, the audio data of the sixth gap of the metamaterial is bandpass filtered, the filtering band is the fixed enhanced frequency band corresponding to the metamaterial, and the frequency range is 1000-3000Hz, and the filtered signal is further Hilbert transformed to obtain an envelope signal, and then the envelope signal is subjected to fast Fourier transformation to directly obtain the envelope spectrum. By identifying the bearing fault characteristic frequency and the amplitude of its higher harmonics in the envelope spectrum, it can be effectively determined whether there is a bearing fault and the severity of the fault can be evaluated.

[0064] Although the embodiments of the present invention are described with practical solutions, they do not constitute limitations on the meaning of the present invention. For those skilled in the art, it is obvious to modify the implementation scheme and combine it with other solutions based on this description.

Claims

1. A UAV acoustic monitor based on metamaterial reinforcement, characterized in that: It comprises an acoustic sensing system and a horn-shaped acoustic metamaterial structure installed in the load area below the drone (1), wherein the acoustic metamaterial structure comprises a plurality of circular plates (2) whose diameters increase linearly from bottom to top and a lever (3) passing through the centers of all the circular plates (2), and the acoustic sensing system comprises an acquisition card (4) and a plurality of acoustic sensors (5), wherein each gap between adjacent circular plates (2) is equipped with an acoustic sensor (5), and all the acoustic sensors (5) are electrically connected to the acquisition card (4).

2. The UAV acoustic monitor based on metamaterial reinforcement according to claim 1, characterized in that: The circular plate (2) is made of polylactic acid material by 3D printing.

3. The UAV acoustic monitor based on metamaterial reinforcement according to claim 1, characterized in that: The spacing between adjacent circular plates (2) is equal.

4. The UAV acoustic monitor based on metamaterial reinforcement according to claim 1, characterized in that: The acquisition card (4) is electrically connected to the integrated control board of the drone (1), and the integrated control board is electrically connected to the onboard computer system interface of the drone (1).

5. The UAV acoustic monitor based on metamaterial reinforcement according to claim 3 is characterized in that: The number of the circular plates (2) is 18, the thickness of the circular plates (2) is 3 mm, the spacing between adjacent circular plates (2) is 10 mm, the radius of the largest circular plate (2) is 110 mm, and the radius of the smallest circular plate (2) is 10 mm.

6. A fault detection method for a UAV acoustic monitor based on metamaterial reinforcement as claimed in any one of claims 1 to 5, characterized in that: The following steps are involved: S1. Setting an inspection path of the drone (1), and setting a number of inspection points on the inspection path to cover the required inspection object. When the drone (1) flies to the set inspection point, it hovers and collects the audio at the current inspection point through the drone acoustic monitor enhanced by metamaterials to judge the state of the inspection object. S2. At the detection point of the current detection object, the drone (1) is set to hover within a cone range of 30 degrees above the detection object to collect audio data, so as to fully utilize the enhanced performance of the metamaterial, and simultaneously collect the acoustic signals on the detection object through a plurality of acoustic sensors (5) of the drone acoustic monitor based on the metamaterial enhancement to form the audio to be analyzed; S3 extracts audio point data on the audio to form an audio data set corresponding to different enhanced frequency bands of the metamaterial; S4. Select the audio data of the required frequency band for time-frequency pre-analysis according to the physical characteristics of the detection object; S5. Further filter and envelope analyze the current audio data group, by using a bandpass filter to perform fixed-band filtering and then perform Hilbert transform to obtain an envelope spectrum, and identify the fault characteristic frequency through the envelope spectrum to determine the fault state of the detection object.

Images