Fault monitoring data early warning processing method and system for underground trackless rubber-tyred vehicle

Abstract

The invention provides a fault monitoring data early warning processing method and a fault monitoring data early warning processing system for an underground trackless rubber-tyred vehicle, and relates to the technical field of fault monitoring of underground trackless rubber-tyred vehicles. Detecting the mechanical response of a hinge point bearing according to the steering angle data, converting the mechanical response into a magnetic characteristic sequence, carrying out adaptive frequency control processing on the magnetic characteristic sequence to output a three-dimensional characteristic pattern, and taking the steering angle, the dynamic stress data and the three-dimensional characteristic pattern as joint observation signals; a digital twinborn model of a built-in hinge point bearing physical model is input, equivalent abrasion loss and abrasion distribution characteristics are reversely deduced by calculating a difference value of a joint observation signal and a prediction signal, and finally hinge point bearing loosening state early warning information is generated based on the equivalent abrasion loss and the abrasion distribution characteristics, so that underground trackless rubber-tyred vehicle hinge point bearing loosening state early warning can be realized. Wear feature achievement is deduced through multi-data acquisition, processing and a digital twinning model.

Description

Technical Field

[0001] This application relates to the field of fault monitoring technology for trackless rubber-tired vehicles in underground mines, and in particular to a fault monitoring data early warning processing method and system for trackless rubber-tired vehicles in underground mines. Background Technology

[0002] Articulated trackless rubber-tired vehicles are core equipment for material transportation and personnel transfer in underground mines. Their articulation point bearings are subjected to radial, axial, and impact loads during steering, making them prone to increased clearance and uneven wear due to accumulated complex stresses. In severe cases, this can lead to steering jamming, vehicle deviation, or even rollover accidents. Given the confined space, high dust levels, and humidity in underground environments, coupled with the high cost of equipment downtime for maintenance, there is an urgent need for a technical solution that can accurately detect early wear conditions of the articulation point bearings in real time and issue early warnings to ensure safe equipment operation and production continuity.

[0003] Currently, the most commonly used method for monitoring this type of bearing wear is a monitoring scheme based on vibration signal analysis. This scheme involves installing vibration sensors near the hinge point to continuously collect vibration signals during bearing operation. Algorithms such as Fourier transform are then used to perform spectral analysis on the signals, extracting characteristic frequency components. These components are then compared with preset normal state thresholds to determine whether abnormal wear exists in the bearing.

[0004] The scheme has obvious limitations: First, there are many sources of interference in the underground environment, such as equipment vibration and ore collision, which result in a low signal-to-noise ratio of vibration signals and are prone to misjudgment. Second, the vibration signals are not sensitive enough to early minor wear and gap changes, and can often only be identified when the wear is more severe, resulting in a strong lag in early warning. Third, it can only reflect the overall abnormal state of the bearing and cannot accurately infer the amount of wear and wear distribution characteristics, making it difficult to support targeted maintenance decisions. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for fault monitoring data early warning processing of trackless rubber-tired vehicles in underground mines, so as to solve the problems of low signal-to-noise ratio, insufficient sensitivity to early wear, and inability to accurately infer the amount and distribution of bearing wear in existing vibration signal monitoring schemes.

[0006] To address the aforementioned technical problems, in a first aspect, this application provides a method for processing fault monitoring data and early warning for underground trackless rubber-tired vehicles, comprising:

[0007] Obtain the steering angle data of the underground trackless rubber-tired vehicle and the dynamic stress data of the bearing seat at the hinge point in the underground trackless rubber-tired vehicle;

[0008] Based on the steering angle data, the mechanical response generated by the hinge point bearing is detected, and the mechanical response is converted into a magnetic feature sequence, which is a time-series electrical signal characterizing the magnetic field change caused by the mechanical response;

[0009] The magnetic feature sequence is subjected to adaptive frequency control processing. The bandwidth and center frequency of the phase-locked loop are dynamically adjusted according to the spectral characteristics, and a three-dimensional feature map is output. The phase-locked loop refers to the internal phase synchronization component on which the adaptive automatic frequency control processing depends. The three-dimensional feature map is used to describe the frequency response and displacement relationship of the hinge bearing over time.

[0010] The steering angle data, dynamic stress data, and three-dimensional feature map are used as joint observation signals and input into a digital twin model with a built-in physical model of the hinge point bearing. By calculating the difference between the joint observation signal and the predicted signal generated by the digital twin model, the equivalent wear amount and wear distribution characteristics inside the hinge point bearing are inferred in reverse.

[0011] Based on the equivalent wear amount and wear distribution characteristics, early warning information is generated for the loose state of the bearing at the hinge point.

[0012] Optionally, the step of inputting the steering angle data, the dynamic stress data, and the three-dimensional feature map as joint observation signals into a digital twin model containing a physical model of the hinge point bearing, and inversely inferring the equivalent wear amount and wear distribution characteristics inside the hinge point bearing by calculating the difference between the joint observation signal and the predicted signal generated by the digital twin model, includes:

[0013] The steering angle data, the dynamic stress data, and the three-dimensional feature map are combined into a joint observation signal;

[0014] The joint observation signal is input into a digital twin model, which has a built-in physical model of the hinge point bearing. The physical model includes rigid body dynamics equations describing the motion of the hinge mechanism and wear physical models characterizing the material removal process of the bearing.

[0015] The steering angle data in the joint observation signal is used as input to drive the digital twin model to generate a prediction signal;

[0016] Calculate the difference between the dynamic stress data and three-dimensional feature map in the joint observation signal and the predicted signal;

[0017] Based on the difference value, the equivalent wear amount and wear distribution characteristics inside the hinge bearing are inferred by inversely solving the physical model in the digital twin model.

[0018] Optionally, the adaptive frequency control processing of the magnetic feature sequence dynamically adjusts the bandwidth and center frequency of the phase-locked loop based on the spectral characteristics, outputting a three-dimensional feature map. The phase-locked loop refers to the internal phase synchronization component upon which the adaptive automatic frequency control processing relies. The three-dimensional feature map describes the frequency response and displacement relationship of the hinge point bearing over time, including:

[0019] Spectral analysis is performed on the magnetic feature sequence to obtain the instantaneous frequency components and energy distribution;

[0020] Based on the instantaneous frequency components and energy distribution, the bandwidth and center frequency of the phase-locked loop are dynamically adjusted so that the phase-locked loop adaptively tracks the main frequency changes in the magnetic feature sequence;

[0021] The magnetic feature sequence is demodulated in frequency and phase by the phase-locked loop, and the demodulated frequency and phase signals are output.

[0022] The demodulated frequency and phase signals are combined with the time series to generate a three-dimensional feature map. The first dimension of the three-dimensional feature map represents time, the second dimension represents frequency, and the third dimension represents the equivalent displacement calculated based on the phase signal.

[0023] Optionally, the step of inferring the equivalent wear amount and wear distribution characteristics inside the hinge bearing by inversely solving the physical model in the digital twin model based on the difference value includes:

[0024] The difference value is used as the input to the objective function, which is used to quantify the overall deviation between the predicted signal of the digital twin model and the joint observation signal.

[0025] In the parameter space of the wear physics model, the output value of the objective function is minimized by iteratively adjusting the model parameters corresponding to the equivalent wear amount and wear distribution characteristics.

[0026] The model parameter values ​​that minimize the output value of the objective function are determined as the equivalent wear amount and wear distribution characteristics inside the hinge point bearing.

[0027] Optionally, the step of using the steering angle data in the joint observation signal as input to drive the digital twin model and generate a prediction signal includes:

[0028] Using the steering angle data in the joint observation signal as input, the rigid body dynamics equations in the digital twin model are driven to calculate the dynamic mechanical state inside the hinge bearing.

[0029] The dynamic mechanical state is used as input to drive the wear physics model in the digital twin model to generate a prediction signal, which represents the expected value of the dynamic stress data and the three-dimensional feature map under an ideal wear-free state.

[0030] Optionally, the step of generating early warning information for the loosening state of the hinge point bearing based on the equivalent wear amount and wear distribution characteristics includes:

[0031] The equivalent wear amount is compared with a predetermined first-level threshold and a second-level threshold.

[0032] The wear distribution characteristics are analyzed, the dispersion index in the bearing circumferential direction is calculated, and the dispersion index is compared with a predetermined distribution non-uniformity threshold.

[0033] Based on the comparison results of the equivalent wear amount and the threshold, and the comparison results of the dispersion index and the uneven distribution threshold, early warning information for the loose state of the hinge point bearing is generated.

[0034] Optionally, based on the steering angle data, the mechanical response generated by the hinge bearing is detected, and the mechanical response is converted into a magnetic feature sequence, wherein the magnetic feature sequence is a time-series electrical signal characterizing the magnetic field change caused by the mechanical response, including:

[0035] The movement of the articulation mechanism of the trackless rubber-tired vehicle is driven based on the steering angle data, and the mechanical response caused by the change in the bearing state at the articulation point is detected.

[0036] The changes in magnetic field distribution caused by the mechanical response are captured by a magnetic field sensing device.

[0037] By utilizing the physical characteristics of the sensing element in the magnetic field sensing device, the change in magnetic field distribution is converted into a corresponding change in electrical parameters;

[0038] By continuously acquiring and converting the changes in the electrical parameters, a magnetic feature sequence is formed. The magnetic feature sequence is a time-series electrical signal that characterizes the changes in the magnetic field caused by the mechanical response.

[0039] Secondly, this application provides a fault monitoring data early warning processing system for underground trackless rubber-tired vehicles, comprising:

[0040] The acquisition module is used to acquire the steering angle data of the underground trackless rubber-tired vehicle and the dynamic stress data of the hinge point bearing seat in the underground trackless rubber-tired vehicle.

[0041] The conversion module is used to detect the mechanical response generated by the hinge point bearing based on the steering angle data, and convert the mechanical response into a magnetic feature sequence, wherein the magnetic feature sequence is a time-series electrical signal characterizing the magnetic field change caused by the mechanical response;

[0042] The processing module is used to perform adaptive frequency control processing on the magnetic feature sequence, dynamically adjust the bandwidth and center frequency of the phase-locked loop according to the spectral characteristics, and output a three-dimensional feature map. The phase-locked loop refers to the internal phase synchronization component on which the adaptive automatic frequency control processing depends. The three-dimensional feature map is used to describe the frequency response and displacement relationship of the hinge bearing over time.

[0043] The inference module is used to input the steering angle data, the dynamic stress data and the three-dimensional feature map as joint observation signals into a digital twin model with a built-in physical model of the hinge point bearing. By calculating the difference between the joint observation signal and the prediction signal generated by the digital twin model, the equivalent wear amount and wear distribution characteristics inside the hinge point bearing are inferred in reverse.

[0044] The generation module is used to generate early warning information for the loose state of the hinge point bearing based on the equivalent wear amount and wear distribution characteristics.

[0045] Thirdly, this application provides an electronic device, comprising:

[0046] Memory, used to store computer programs;

[0047] A processor is configured to execute the computer program to implement the steps of the fault monitoring data early warning processing method for the trackless rubber-tired vehicle described in the first aspect above.

[0048] Fourthly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, can implement the steps of the fault monitoring data early warning processing method for trackless rubber-tired vehicles in underground mines as described in the first aspect above.

[0049] The fault monitoring data early warning processing method for trackless rubber-tired vehicles provided in this application acquires dynamic stress data of steering angle and hinge point bearing housing, combines it with magnetic feature sequences obtained from mechanical response conversion based on steering angle detection, and generates a three-dimensional feature map that describes the frequency response and displacement relationship of the bearing over time through adaptive frequency control processing. Then, the multi-dimensional data is input as a joint observation signal into a digital twin model of the built-in bearing physical model. Through difference calculation, the equivalent wear amount and wear distribution characteristics are accurately inferred in reverse, and finally targeted early warning information is generated. This method realizes early and accurate monitoring and early warning of the loose state of the hinge point bearing of the trackless rubber-tired vehicle, which not only improves the anti-interference ability and sensitivity of capturing bearing wear characteristics in complex underground environments, but also provides specific wear quantification information for maintenance decisions, effectively ensuring equipment operation safety and production continuity.

[0050] Furthermore, steering angle data, dynamic stress data, and 3D feature maps are combined into a joint observation signal, which is then input into a digital twin model with a built-in physical model of the hinge point bearing. The steering angle data drives the model to generate a prediction signal. The differences between the dynamic stress data and 3D feature map in the joint observation signal and the prediction signal are calculated. Finally, based on these differences, the physical model is solved in reverse to infer the equivalent wear amount and wear distribution characteristics inside the hinge point bearing. By clarifying the specific composition of the physical model built into the digital twin model, using steering angle data to drive the model to generate a prediction signal, and combining multi-dimensional data differences for inverse solving, not only is the accuracy of the digital twin model in simulating the bearing's operating state improved, but the equivalent wear amount and wear distribution characteristics of the bearing can also be inferred more scientifically and meticulously. This provides a more reliable quantitative basis for subsequent early warning of loosening conditions and targeted maintenance. Attached Figure Description

[0051] To more clearly illustrate the technical solutions of the embodiments of this application 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 this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0052] Figure 1 A flowchart illustrating a fault monitoring data early warning processing method for an underground trackless rubber-tired vehicle provided in this application embodiment;

[0053] Figure 2 A flowchart illustrating a specific embodiment of a fault monitoring data early warning processing method for an underground trackless rubber-tired vehicle provided in this application;

[0054] Figure 3A schematic diagram illustrating a specific embodiment of a fault monitoring data early warning processing method for an underground trackless rubber-tired vehicle provided in this application.

[0055] Figure 4 This is a schematic diagram of the structure of a fault monitoring data early warning processing system for an underground trackless rubber-tired vehicle provided in an embodiment of this application. Detailed Implementation

[0056] For the wear monitoring of bearings at the articulation points of articulated trackless rubber-tired vehicles in underground mines, existing solutions based on vibration signal analysis have significant shortcomings. Interference sources such as vibration of the underground equipment itself and ore collisions result in low signal-to-noise ratios of vibration signals, making misjudgments easy. At the same time, this solution is not sensitive enough to early minor wear and clearance changes in bearings, often only identifying them when wear is more severe, resulting in a strong lag in early warning. Furthermore, it can only reflect overall bearing anomalies and cannot accurately infer the amount and distribution characteristics of wear, making it difficult to support targeted maintenance decisions and seriously affecting equipment operation safety and production continuity.

[0057] To address the aforementioned issues, this application proposes a fault monitoring data early warning processing method for underground trackless rubber-tired vehicles. The core of this method involves acquiring steering angle data and dynamic stress data of the bearing housing at the hinge point. This data, combined with steering angle detection of the bearing's mechanical response, is converted into a magnetic feature sequence. After adaptive frequency control processing, a three-dimensional feature map is output. This multi-dimensional data is then used as a joint observation signal input to a digital twin model with an embedded bearing physical model. Through difference calculation, the equivalent wear amount and wear distribution characteristics are inferred in reverse, generating an early warning. This method improves signal anti-interference capability and early wear sensitivity in complex environments through the magnetic feature sequence and adaptive frequency control processing. The digital twin model, combined with multi-source data, achieves accurate quantification of wear amount and distribution, fundamentally solving the problems of low signal-to-noise ratio, delayed early warning, and inaccurate wear information inference in existing solutions, thus ensuring equipment safety and continuous production.

[0058] To enable those skilled in the art to better understand the present application, the present application will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are merely some embodiments of the present application, and not all embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0059] The core of this application is to provide a method for fault monitoring and early warning processing of trackless rubber-tired vehicles in underground mines. A flowchart of one specific implementation is shown below. Figure 1 As shown, the method includes:

[0060] S101. Obtain the steering angle data of the underground trackless rubber-tired vehicle and the dynamic stress data of the hinge point bearing seat in the underground trackless rubber-tired vehicle.

[0061] In the above scheme, the steering angle data reflects the magnitude and variation of the rotation angle of the steering system of the underground trackless rubber-tired vehicle. Dynamic stress data refers to the stress data of the articulation point bearing housing over time due to the various forces it bears during vehicle operation. The articulation point bearing is a key component connecting the front and rear frames of the underground trackless rubber-tired vehicle, enabling relative rotation.

[0062] In this embodiment, a displacement sensor fixed to the piston rod of the steering cylinder first collects the extension and retraction of the cylinder. Then, based on the geometric relationship between the cylinder hinge point and the steering knuckle, a time-sequential sequence of steering angle data is calculated, which serves as the steering angle data for the trackless rubber-tired vehicle. For example, if the extension and retraction of the steering cylinder of a trackless rubber-tired vehicle changes by 1 centimeter, the steering angle changes by 2 degrees according to the geometric relationship. After the displacement sensor collects the change in extension and retraction over time, a sequence of steering angle changes over time can be obtained.

[0063] Secondly, by using a fiber optic strain sensor array arranged in a matrix at key stress points on the surface and inside the bearing housing at the hinge point, spatially distributed dynamic stress waveform data generated by the bearing housing during vehicle steering and driving are simultaneously acquired. Each sensor node corresponds to a specific spatial coordinate. For example, if the sensor array is arranged in a 3×3 matrix, each node corresponds to a different position on the bearing housing, and stress change waveforms at each position can be acquired separately.

[0064] Finally, ensure that the acquisition trigger signals of the displacement sensor and the fiber optic strain sensor array are synchronized, so that the steering angle data sequence and the dynamic stress waveform data correspond strictly in timestamps. For example, both should start acquiring data simultaneously at a frequency of 100 times per second to ensure that the steering angle and stress data at the same point in time can be accurately matched.

[0065] S102. Based on the steering angle data, detect the mechanical response generated by the hinge point bearing, and convert the mechanical response into a magnetic feature sequence, wherein the magnetic feature sequence is a time-series electrical signal characterizing the magnetic field change caused by the mechanical response;

[0066] Optionally, step S102 may specifically include the following steps:

[0067] S1021. Drive the articulation mechanism of the trackless rubber-tired vehicle in the well based on the steering angle data, and detect the mechanical response caused by the change in the bearing state at the articulation point;

[0068] S1022. Using a magnetic field sensing device, capture the change in magnetic field distribution caused by the mechanical response;

[0069] S1023. Utilizing the physical characteristics of the sensing element in the magnetic field sensing device, the change in magnetic field distribution is converted into a corresponding change in electrical parameters;

[0070] S1024. By continuously acquiring and converting the changes in the electrical parameters, a magnetic feature sequence is formed. The magnetic feature sequence is a time-series electrical signal that characterizes the changes in the magnetic field caused by the mechanical response.

[0071] In the above scheme, mechanical response refers to the relative displacement and yaw between the front and rear frames caused by changes in the bearing condition at the hinge point. The magnetic field sensing device is used to capture changes in the magnetic field, and includes components such as magnets and AMR sensors. Changes in magnetic field distribution refer to the gradient changes in the strength and direction of the magnetic field caused by the mechanical response. Changes in electrical parameters refer to the changes in electrical parameters such as resistance of the sensing element in the magnetic field sensing device due to changes in the magnetic field. The magnetic characteristic sequence is a time-series electrical signal characterizing the changes in the magnetic field caused by the mechanical response; that is, a set of electrical signals arranged in chronological order.

[0072] In this embodiment, step S1021 first uses steering angle data to drive the articulation mechanism of the underground trackless rubber-tired vehicle to perform corresponding movements. During the movement of the articulation mechanism, a detection device monitors the relative position and attitude changes of the front and rear frames, capturing the mechanical response caused by changes in the bearing condition at the articulation point. For example, when the steering angle data shows that the vehicle turns 30 degrees, the articulation mechanism rotates accordingly. If the bearing is worn, a mechanical response will be detected where the relative displacement of the front and rear frames increases by 2 mm and the yaw angle increases by 1 degree.

[0073] Secondly, in step S1022, by installing components of a magnetic field sensing device (magnets and AMR sensors) at both ends of the hinge pin of the hinge point bearing of the trackless rubber-tired vehicle underground, it is ensured that the positions of both are relatively fixed and capable of sensing changes in the magnetic field. When a mechanical response is detected, i.e., relative displacement and sway of the front and rear frames, it will cause changes in the position or orientation of the hinge pin and the magnets at both ends, thereby changing the surrounding magnetic field environment. The AMR sensor will sense this change in the magnetic field environment and capture the change in magnetic field distribution caused by the mechanical response.

[0074] For example, at the hinge bearing of the C-type trackless rubber-tired underground vehicle, workers installed a pair of magnets at both ends of the hinge pin, and the corresponding AMR sensors were installed on the frame near the magnets. When the vehicle's hinge mechanism experienced a mechanical response of 2 mm relative displacement and 1 degree yaw due to bearing wear, the magnets changed position and orientation with the pin. The AMR sensors captured data showing a magnetic field distribution change, with the magnetic field strength gradient from 500 Gs to 530 Gs and a magnetic field direction gradient deflection of 1.5 degrees.

[0075] Then, in step S1023, utilizing the physical characteristic that the resistance value of the AMR sensor changes with the direction of the magnetic field, when a change in the magnetic field distribution, especially a gradient change in the magnetic field direction, is captured, the AMR sensor converts the input gradient change in the magnetic field direction into a continuous change in the resistance value based on its own resistance changing with the direction of the magnetic field. For example, when the magnetic field direction deflects by 4 degrees, the sensor's resistance value continuously changes from 980Ω to 1030Ω, realizing the conversion from a change in magnetic field distribution to a change in electrical parameters.

[0076] Finally, through step S1024, the continuous changes in resistance value are continuously acquired and converted into signals by the signal conditioning circuit at the back end of the sensor to form a magnetic feature sequence. For example, the resistance value is acquired once every 0.1 seconds and converted into a voltage signal. Arranging these voltage signals in chronological order yields a time-series electrical signal sequence that characterizes the changes in the magnetic field.

[0077] S103. Perform adaptive frequency control processing on the magnetic feature sequence, dynamically adjust the bandwidth and center frequency of the phase-locked loop according to the spectral characteristics, and output a three-dimensional feature map. The phase-locked loop refers to the internal phase synchronization component on which the adaptive automatic frequency control processing depends. The three-dimensional feature map is used to describe the frequency response and displacement relationship of the hinge point bearing over time.

[0078] Optionally, step S103 may specifically include the following steps:

[0079] S1031. Perform spectral analysis on the magnetic feature sequence to obtain the instantaneous frequency components and energy distribution;

[0080] S1032. Based on the instantaneous frequency components and energy distribution, dynamically adjust the bandwidth and center frequency of the phase-locked loop so that the phase-locked loop adaptively tracks the main frequency changes in the magnetic feature sequence;

[0081] S1033. The magnetic feature sequence is demodulated in frequency and phase by the phase-locked loop, and the demodulated frequency and phase signals are output.

[0082] S1034. Combine the demodulated frequency and phase signals with the time series to generate a three-dimensional feature map. The first dimension of the three-dimensional feature map represents time, the second dimension represents frequency, and the third dimension represents the equivalent displacement calculated based on the phase signal.

[0083] In the above scheme, the magnetic characteristic sequence is a time-series electrical signal characterizing the magnetic field changes caused by the mechanical response. Adaptive frequency control processing is a signal processing method that dynamically adjusts processing parameters based on the signal's spectral characteristics. Spectral characteristics refer to the distribution and energy distribution of different frequency components in the signal. The phase-locked loop (PLL) is the internal phase synchronization component upon which adaptive frequency control processing relies, used to track and synchronize the signal frequency and phase. The bandwidth is the width of the frequency range that the PLL can process, and the center frequency is the center value of the frequency processed by the PLL. The three-dimensional feature map is a graph used to describe the relationship between the frequency response and displacement of the hinge point bearing over time. The first dimension is time, the second dimension is frequency, and the third dimension is equivalent displacement.

[0084] In this embodiment of the application, the magnetic feature sequence is first processed by the spectrum analysis method in step S1031. The sliding time window technique is introduced, and an appropriate window duration and sliding step size are set. Then, the magnetic feature sequence segment in each sliding time window is decomposed by frequency, the instantaneous frequency value contained in the signal in that time period is extracted, and the signal energy corresponding to each instantaneous frequency component is calculated to obtain the energy distribution of different frequency components.

[0085] Secondly, based on the instantaneous frequency components and signal energy distribution in step S1032, the main frequency components in the magnetic feature sequence, i.e., the frequency values ​​with the highest energy proportion, are identified. Then, a time-varying spectral characteristic model of the magnetic feature sequence is established to describe the variation law of the main frequency components over time. According to the variation law of the time-varying spectral characteristic model, the bandwidth and center frequency of the phase-locked loop are dynamically adjusted so that the tracking range of the phase-locked loop can cover the main frequency changes, ensuring that it can adaptively track the changes in the main frequency. For example, when the model shows that the main frequency increases from 20Hz to 30Hz, the center frequency of the phase-locked loop is adjusted from 20Hz to 30Hz, and the bandwidth is widened from 10Hz to 15Hz to cover the main frequency change.

[0086] Then, in step S1033, the magnetic feature sequence is input into the phase-locked loop (PLL) after parameter configuration. The PLL starts operating based on the adjusted bandwidth and center frequency. Through internal phase comparison and loop filtering mechanisms, the magnetic feature sequence undergoes frequency tracking and phase synchronization processing, separating frequency and phase information from the complex signal. The separated frequency and phase information is then organized, and a corresponding timestamp is added to each data point, outputting a frequency demodulated signal and a phase demodulated signal containing the timestamp. For example, the output signal sequence has a frequency of 25Hz and a phase of 90 degrees at 10:00:01.000, and a frequency of 26Hz and a phase of 100 degrees at 10:00:01.010.

[0087] Finally, in step S1034, the demodulated frequency and phase signals are correlated and matched with the corresponding time series to ensure that each frequency and phase value corresponds to an accurate time point. An integral operation is performed on the phase demodulated signal to obtain the equivalent displacement, and the angular displacement is converted into a linear equivalent displacement value through integration. Finally, using the time series as the first dimension (X-axis), the amplitude of the frequency demodulated signal as the second dimension (Y-axis), and the calculated equivalent displacement as the third dimension (Z-axis), the three data are combined to create a three-dimensional feature map that visually displays the relationship between the frequency response and displacement of the hinge point bearing over time.

[0088] Specifically, in the fault monitoring signal processing system of the Type A trackless rubber-tired underground vehicle, the staff obtained an 8-second magnetic feature sequence. First, a spectral analysis was performed using a sliding time window to extract the instantaneous frequency components fluctuating in the 8-45Hz range and their corresponding energy distribution. Then, based on this data, a time-varying spectral characteristic model was established, dynamically adjusting the phase-locked loop (PLL) center frequency from 15Hz to 25Hz and the bandwidth from 8Hz to 12Hz. Next, the configured PLL was used to demodulate the magnetic feature sequence, obtaining a timestamped frequency and phase demodulated signal. Finally, the demodulated signal was combined with the time sequence, and the phase signal was integrated to obtain the equivalent displacement, generating a three-dimensional feature map with time, frequency, and equivalent displacement for subsequent bearing condition analysis.

[0089] The overall scheme of S103 described above can perform precise adaptive frequency control processing on the magnetic feature sequence, dynamically track signal frequency changes, demodulate to obtain clear frequency and phase signals, and then generate a three-dimensional feature map that can intuitively reflect the relationship between bearing time, frequency and displacement. This map presents the bearing state information in a multi-dimensional visualization form, providing rich and accurate feature basis for subsequent inference of bearing wear state by combining other data.

[0090] S104. The steering angle data, the dynamic stress data, and the three-dimensional feature map are used as joint observation signals and input into a digital twin model with a built-in physical model of the hinge point bearing. By calculating the difference between the joint observation signal and the prediction signal generated by the digital twin model, the equivalent wear amount and wear distribution characteristics inside the hinge point bearing are inferred in reverse.

[0091] Optionally, step S104 may specifically include the following steps:

[0092] S1041. Combine the steering angle data, the dynamic stress data, and the three-dimensional feature map into a joint observation signal;

[0093] S1042. Input the joint observation signal into the digital twin model. The digital twin model has a physical model of the hinge point bearing. The physical model includes rigid body dynamics equations describing the motion of the hinge mechanism and wear physical model characterizing the bearing material removal process.

[0094] S1043. Using the steering angle data in the joint observation signal as input, drive the digital twin model to generate a prediction signal;

[0095] S1044. Calculate the difference between the dynamic stress data and three-dimensional feature map in the joint observation signal and the predicted signal;

[0096] S1045. Based on the difference value, the equivalent wear amount and wear distribution characteristics inside the hinge bearing are inferred by inversely solving the physical model in the digital twin model.

[0097] Specifically, step S1043 includes the following process: using the steering angle data in the joint observation signal as input, driving the rigid body dynamics equation in the digital twin model to calculate the dynamic mechanical state inside the hinge bearing; using the dynamic mechanical state as input, driving the wear physics model in the digital twin model to generate a prediction signal, wherein the prediction signal represents the expected value of the dynamic stress data and the three-dimensional feature map under an ideal wear-free state.

[0098] Specifically, step S1045 includes the following process: using the difference value as input to the objective function, which is used to quantify the overall deviation between the predicted signal of the digital twin model and the joint observation signal; in the parameter space of the wear physics model, by iteratively adjusting the model parameters corresponding to the equivalent wear amount and wear distribution characteristics, minimizing the output value of the objective function; and determining the model parameter value corresponding to the minimum output value of the objective function as the equivalent wear amount and wear distribution characteristics inside the hinge bearing.

[0099] In the above scheme, the digital twin model is a virtual model simulating the bearing's operating state, with a built-in physical model of the bearing's hinge point. The physical model includes rigid body dynamics equations describing the motion of the hinge mechanism and a wear physical model characterizing the material removal process of the bearing. The rigid body dynamics equations describe the mechanical laws governing the motion of the hinge mechanism, while the wear physical model characterizes the process and laws of material removal due to wear. The predicted signal is the expected value of the dynamic stress and three-dimensional feature map output by the digital twin model under an ideal, wear-free state. The difference value is the numerical deviation between the jointly observed signal and the predicted signal. The equivalent wear amount is a quantitative value reflecting the overall wear degree of the bearing. The wear distribution characteristics are the wear conditions at different locations inside the bearing.

[0100] In the embodiments of this application, such as Figure 2 As shown, firstly, step S1041 uses timestamps as a reference to associate and match the steering angle data, dynamic stress data, and three-dimensional feature map. Then, all the associated multi-dimensional data are arranged in chronological order to form a joint observation signal, ensuring the temporal consistency and integrity of the data.

[0101] Secondly, the joint observation signal is input to the built-in hinge point bearing physical model through step S1042. The physical model includes rigid body dynamics equations describing the motion of the hinge mechanism and wear physical models characterizing the bearing material removal process.

[0102] For example, in the virtual simulation system of the C-type trackless rubber-tired underground vehicle, the staff constructed a digital twin model of the hinge point bearing. The model incorporated rigid body dynamics equations describing the force and displacement relationships during hinge arm rotation and a wear physics model describing the relationship between wear and material removal rate. After converting the integrated joint observation signals into a format recognizable by the model, the model began simulating its operational state based on the input data and its internal physical model.

[0103] Next, in step S1043, steering angle data is extracted from the joint observation signal and used as input conditions to drive the operation of the digital twin model. The rigid body dynamics equations in the driving model are calculated, and the motion process of the articulated mechanism is simulated according to the change of steering angle, thereby obtaining the dynamic mechanical state inside the bearing at the articulation point. The calculated dynamic mechanical state is used as input to drive the wear physical model in the digital twin model. Since it is simulating an ideal wear-free state, the wear parameter is set to zero, thereby generating a prediction signal, that is, the expected value of the dynamic stress data and the three-dimensional feature map under this state.

[0104] Then, in step S1044, the dynamic stress data, three-dimensional feature map data and prediction signals in the joint observation signal are matched one-to-one with the timestamps to ensure that the actual observation data at each time point can find the corresponding prediction data. The numerical difference calculation method is used to calculate the difference between the dynamic stress data and the corresponding predicted stress, as well as the difference between the data of each dimension in the three-dimensional feature map and the corresponding prediction data. Finally, all the calculated difference values ​​are sorted out to form a difference value sequence, which provides data for subsequent reverse inference of wear state.

[0105] Finally, in step S1045, the difference values ​​calculated in step S1044 are used as input to the objective function. The objective function performs comprehensive calculations on these difference values ​​and outputs the overall deviation between the predicted signal and the joint observation signal. Within the parameter space of the wear physics model, model parameters corresponding to the equivalent wear amount and wear distribution characteristics are selected, and an initial set of parameter values ​​is substituted into the model. Then, these model parameters are iteratively adjusted, and the output value of the objective function is calculated after each adjustment. The changing trend of the output value is observed. Finally, the iterative adjustment continues until the output value of the objective function reaches its minimum. The corresponding model parameter values ​​at this point represent the equivalent wear amount and wear distribution characteristics inside the hinge point bearing.

[0106] Specifically, in the wear inference module of the C-type underground trackless rubber-tired vehicle, the operator inputs the difference value into the objective function, initially setting the equivalent wear amount to 0.1 mm and the wear distribution to be uniform. Through iterative adjustments, when the equivalent wear amount increases to 0.28 mm, and the wear distribution is 0.32 mm on the inner side and 0.24 mm on the outer side, the overall deviation of the objective function output decreases from the initial 45 to a minimum of 12. At this point, the equivalent wear amount and wear distribution characteristics are determined to be the actual wear state inside the bearing.

[0107] In practical applications, within the fault analysis system of the Type A trackless rubber-tired underground vehicle, staff first combine the collected steering angle data, bearing housing dynamic stress data, and 3D feature maps according to timestamps into a joint observation signal, which is then input into a digital twin model with a built-in bearing physical model. Using the steering angle data to drive the model, the dynamic mechanical state of the bearing is first calculated through rigid body dynamics equations, and then a prediction signal is generated through the wear physical model. Calculations revealed that the observed dynamic stress was on average 15 MPa higher than the predicted value, and the average displacement deviation in the 3D feature map was 0.4 mm. These differences were input into the objective function, and the wear parameters were iteratively adjusted. When the equivalent wear amount was set to 0.25 mm and the wear distribution was concentrated on the inner side of the bearing, the objective function value was minimized, thus inferring the equivalent wear amount and wear distribution characteristics of the bearing.

[0108] The overall solution of S104 mentioned above combines multi-source observation data with a digital twin model. By using the difference between model prediction and actual observation, the wear state can be inferred in reverse. This enables accurate quantification of the equivalent wear amount and wear distribution characteristics of the hinge point bearing. It breaks through the limitation of traditional methods that make it difficult to directly obtain the internal wear information of the bearing, and provides scientific and reliable wear data support for subsequent targeted early warning and maintenance.

[0109] S105. Based on the equivalent wear amount and wear distribution characteristics, generate early warning information for the loose state of the hinge point bearing.

[0110] Optionally, step S105 may specifically include the following steps:

[0111] S1051. Compare the equivalent wear amount with a predetermined first-level threshold and a second-level threshold.

[0112] S1052. Analyze the wear distribution characteristics, calculate its dispersion index in the bearing circumferential direction, and compare the dispersion index with a predetermined distribution non-uniformity threshold.

[0113] S1053. Based on the comparison results of the equivalent wear amount and the threshold, and the comparison results of the dispersion index and the uneven distribution threshold, generate early warning information for the loose state of the hinge point bearing.

[0114] In the above scheme, the first and second level thresholds are pre-set standard values ​​used to judge the severity of wear. The dispersion index is a parameter that measures the uniformity of wear distribution along the bearing circumference. The non-uniformity threshold is a pre-set standard value for judging whether the wear distribution is uniform. The warning information consists of different levels of alerts generated based on the amount and distribution of wear, indicating the bearing's looseness.

[0115] In this embodiment of the application, the specific value of the equivalent wear amount of the hinge point bearing obtained in step S1051 is first used to retrieve the first-level threshold and the second-level threshold preset in the system. The equivalent wear amount is compared with the first-level threshold and the second-level threshold respectively to determine which range the equivalent wear amount is in.

[0116] Secondly, the wear distribution characteristics are obtained through step S1052. These characteristics include specific wear data at multiple locations along the bearing circumference. A statistical method is used to calculate the dispersion index of this wear data, for example, by calculating the standard deviation of the data. The calculated standard deviation of the wear at the above eight locations is 0.06 mm, meaning the dispersion index is 0.06. A preset non-uniformity threshold is retrieved, such as 0.08 mm. The dispersion index is compared with this threshold. For example, if 0.06 mm is less than 0.08 mm, then it is determined that the dispersion index does not exceed the non-uniformity threshold.

[0117] Finally, S1053 clarifies the comparison results between the equivalent wear amount and the threshold, and the comparison results between the dispersion index and the uneven distribution threshold. Based on preset warning rules, a judgment is made: if the equivalent wear amount exceeds the first-level threshold but not the second-level threshold, and the dispersion index does not exceed the uneven distribution threshold, a first-level warning is generated; if the equivalent wear amount exceeds the second-level threshold and the dispersion index does not exceed the threshold, a second-level warning is generated; if the equivalent wear amount exceeds the first-level threshold and the dispersion index exceeds the threshold, a third-level warning is generated; if the equivalent wear amount exceeds the second-level threshold and the dispersion index exceeds the threshold, a fourth-level warning is generated. For example, if the comparison result is that the equivalent wear amount exceeds the first-level threshold but not the second-level threshold, and the dispersion index exceeds the threshold, then a third-level warning is generated. Finally, the generated warning information is output to the display or alarm device to alert the staff.

[0118] In practical application, in the early warning system of the A-type trackless rubber-tired underground vehicle, the staff preset the first-level threshold to 0.25 mm, the second-level threshold to 0.45 mm, and the uneven distribution threshold to 0.35 mm. The equivalent wear of the bearing was obtained as 0.3 mm, and the dispersion index of the wear distribution was calculated to be 0.4. Comparing the equivalent wear with the thresholds, it exceeded the first-level threshold but did not exceed the second-level threshold; the dispersion index exceeded the uneven distribution threshold. According to the rules, a third-level early warning message indicating localized abnormal wear was generated, prompting the staff to pay attention to the localized wear of the bearing.

[0119] The above-mentioned S105 overall solution can combine the equivalent wear amount and wear distribution characteristics of the bearing to generate different levels of early warning information according to preset thresholds, realizing accurate graded warning of bearing loosening status. This allows staff to clearly understand the severity and distribution of bearing wear, thereby taking targeted measures to effectively avoid equipment damage or safety accidents caused by bearing failure.

[0120] The following is a complete example for steps 101-105, such as Figure 3 As shown, in the fault monitoring system of the A-type trackless rubber-tired vehicle in the well, the displacement sensor fixed on the piston rod of the steering cylinder first collects the cylinder extension and retraction amount and calculates the steering angle data sequence in time order based on the geometric relationship between the cylinder hinge point and the steering knuckle. At the same time, fiber optic strain sensor arrays are arranged in a 2×3 matrix on the surface of the hinge point bearing seat and the key stress points inside to synchronously collect spatially distributed dynamic stress waveform data, and ensure that the acquisition trigger signals of the displacement sensor and the sensor array are synchronized, so that the two types of data strictly correspond in timestamp.

[0121] Next, the articulated mechanism is driven to move based on the acquired steering angle data. When the vehicle turns, mechanical responses such as relative displacement and yaw of the front and rear frames caused by changes in the state of the articulated bearing are detected. The changes in magnetic field strength and direction gradient caused by this mechanical response are captured by magnets and AMR sensors installed at both ends of the articulated bearing. The magnetic field gradient change is converted into a continuous change in resistance by utilizing the characteristic that the resistance value of the AMR sensor changes with the direction of the magnetic field. The signal conditioning circuit at the back end of the sensor continuously collects and converts the resistance value change to form a magnetic feature sequence characterizing the magnetic field change.

[0122] Subsequently, spectral analysis was performed on the magnetic feature sequence. The instantaneous frequency components fluctuated in the 8-45Hz range and their corresponding energy distribution were extracted using a 50-millisecond sliding time window. Based on these data, a time-varying spectral characteristic model was established, and the bandwidth and center frequency of the phase-locked loop were dynamically adjusted to track the main frequency changes in the magnetic feature sequence. The magnetic feature sequence was then demodulated in terms of frequency and phase using the configured phase-locked loop, outputting frequency demodulated signals and phase demodulated signals containing timestamps. These demodulated signals were combined with the time sequence, and the equivalent displacement was obtained by integrating the phase signal, generating a feature map with time, frequency, and equivalent displacement as three dimensions.

[0123] Then, the steering angle data, dynamic stress data, and three-dimensional feature map are combined into a joint observation signal according to the timestamp, and input into the digital twin model of the built-in hinge point bearing physical model (including rigid body dynamics equations and wear physics model); the dynamic mechanical state of the bearing is calculated by using the rigid body dynamics equations in the model driven by the steering angle data, and then input into the wear physics model to generate a prediction signal under the ideal wear-free state; the difference between the dynamic stress data and three-dimensional feature map in the joint observation signal and the prediction signal is calculated, and it is found that the observed stress is on average 15 MPa higher than the predicted value, and the average displacement deviation is 0.4 mm; the difference value is input into the objective function, and the wear-related model parameters are iteratively adjusted. When the equivalent wear amount is set to 0.25 mm and the wear distribution is concentrated on the inner side of the bearing, the objective function value is minimized, thereby inferring the equivalent wear amount and wear distribution characteristics of the bearing.

[0124] Finally, the first-level threshold is preset to 0.2 mm, the second-level threshold to 0.4 mm, and the uneven distribution threshold to 0.3 mm. The inferred equivalent wear amount of 0.25 mm is compared with the thresholds. It exceeds the first-level threshold but does not exceed the second-level threshold. The wear distribution dispersion index is calculated to be 0.25, which does not exceed the uneven distribution threshold. According to the preset rules, a first-level early warning message suggesting enhanced monitoring is generated and sent to the system console to remind staff to pay close attention to the bearing's operating status.

[0125] Figure 4This is a schematic diagram illustrating a specific implementation of a fault monitoring data early warning processing system for an underground trackless rubber-tired vehicle provided in this application. (Refer to...) Figure 4 The system may include:

[0126] The acquisition module 41 is used to acquire the steering angle data of the underground trackless rubber-tired vehicle and the dynamic stress data of the hinge point bearing seat in the underground trackless rubber-tired vehicle.

[0127] The conversion module 42 is used to detect the mechanical response generated by the hinge point bearing based on the steering angle data, and convert the mechanical response into a magnetic feature sequence, wherein the magnetic feature sequence is a time-series electrical signal characterizing the magnetic field change caused by the mechanical response;

[0128] Processing module 43 is used to perform adaptive frequency control processing on the magnetic feature sequence, dynamically adjust the bandwidth and center frequency of the phase-locked loop according to the spectral characteristics, and output a three-dimensional feature map. The phase-locked loop refers to the internal phase synchronization component on which the adaptive automatic frequency control processing depends. The three-dimensional feature map is used to describe the frequency response and displacement relationship of the hinge bearing over time.

[0129] The inference module 44 is used to input the steering angle data, the dynamic stress data and the three-dimensional feature map as joint observation signals into a digital twin model with a built-in physical model of the hinge point bearing. By calculating the difference between the joint observation signal and the prediction signal generated by the digital twin model, the equivalent wear amount and wear distribution characteristics inside the hinge point bearing are inferred in reverse.

[0130] The generation module 45 is used to generate early warning information for the loose state of the hinge point bearing based on the equivalent wear amount and wear distribution characteristics.

[0131] The fault monitoring data early warning processing system for underground trackless rubber-tired vehicles in this application embodiment is used to implement the aforementioned fault monitoring data early warning processing method for underground trackless rubber-tired vehicles. Therefore, the specific implementation of the fault monitoring data early warning processing system for underground trackless rubber-tired vehicles can be found in the embodiment section of the fault monitoring data early warning processing method for underground trackless rubber-tired vehicles mentioned above. The specific implementation can be referred to the description of the corresponding embodiments, which will not be repeated here.

[0132] This application also provides an electronic device, comprising: a memory for storing a computer program; and a processor for executing the computer program to implement the steps of the fault monitoring data early warning processing method for any of the above-described trackless rubber-tired vehicles.

[0133] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the fault monitoring data early warning processing method for any of the above-described trackless rubber-tired vehicles.

[0134] In one exemplary embodiment, the aforementioned computer-readable storage medium may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, random access memory, portable hard drives, magnetic disks, or optical disks.

[0135] Embodiments of the present invention also provide a computer program product, which includes a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the fault monitoring data early warning processing method for underground trackless rubber-tired vehicles.

[0136] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0137] The above provides a detailed description of the fault monitoring data early warning processing method and system for trackless rubber-tired vehicles in underground mines, as provided in this application. Specific examples have been used to illustrate the principles and implementation methods of this application. The descriptions of the embodiments above are merely for the purpose of helping to understand the method and its core ideas. It should be noted that those skilled in the art can make various improvements and modifications to this application without departing from its principles, and these improvements and modifications also fall within the protection scope of this application.

Claims

1. A fault monitoring data early warning processing method for a trackless rubber-tyred vehicle in a well, characterized in that, The method comprises the following steps: acquiring steering angle data of a trackless rubber-tyred vehicle and dynamic stress data of a hinged point bearing seat in the trackless rubber-tyred vehicle; based on the steering angle data, detecting a mechanical response generated by the hinged point bearing and converting the mechanical response into a magnetic feature sequence, the magnetic feature sequence being a time-series electrical signal representing a change in a magnetic field caused by the mechanical response; performing adaptive frequency control processing on the magnetic feature sequence, dynamically adjusting a bandwidth and a center frequency of a phase-locked loop according to spectral characteristics, and outputting a three-dimensional feature map, the phase-locked loop being an internal phase synchronization component relied on by the adaptive automatic frequency control processing, and the three-dimensional feature map being used to describe a frequency response and displacement relationship of the hinged point bearing changing over time; inputting the steering angle data, the dynamic stress data and the three-dimensional feature map as joint observation signals into a digital twin model in which a physical model of the hinged point bearing is embedded, and reversely inferring an equivalent wear amount and a wear distribution feature inside the hinged point bearing by calculating a difference value between the joint observation signals and a predicted signal generated by the digital twin model; based on the equivalent wear amount and the wear distribution feature, generating early warning information for a loose state of the hinged point bearing.

2. The method of claim 1, wherein, The method of inputting the steering angle data, the dynamic stress data and the three-dimensional feature map as joint observation signals into a digital twin model in which a physical model of the hinged point bearing is embedded, and reversely inferring an equivalent wear amount and a wear distribution feature inside the hinged point bearing by calculating a difference value between the joint observation signals and a predicted signal generated by the digital twin model, comprises the following steps: combining the steering angle data, the dynamic stress data and the three-dimensional feature map into joint observation signals; inputting the joint observation signals into a digital twin model, the digital twin model having a physical model of the hinged point bearing embedded therein, the physical model including a rigid body dynamics equation describing the motion of a hinged mechanism and a wear physical model representing a material removal process of the bearing; inputting the steering angle data in the joint observation signals into the digital twin model as an input to drive the digital twin model and generate a predicted signal; calculating a difference value between the dynamic stress data and the three-dimensional feature map in the joint observation signals and the predicted signal; based on the difference value, inferring the equivalent wear amount and the wear distribution feature inside the hinged point bearing by reversely solving the physical model in the digital twin model.

3. The method of claim 1, wherein, The method of performing adaptive frequency control processing on the magnetic feature sequence, dynamically adjusting a bandwidth and a center frequency of a phase-locked loop according to spectral characteristics, and outputting a three-dimensional feature map, the phase-locked loop being an internal phase synchronization component relied on by the adaptive automatic frequency control processing, and the three-dimensional feature map being used to describe a frequency response and displacement relationship of the hinged point bearing changing over time, comprises the following steps: performing spectral analysis on the magnetic feature sequence to obtain an instantaneous frequency component and an energy distribution; dynamically adjusting the bandwidth and the center frequency of the phase-locked loop according to the instantaneous frequency component and the energy distribution, so that the phase-locked loop adaptively tracks main frequency changes in the magnetic feature sequence; Frequency and phase demodulation of the magnetic feature sequence by the phase-locked loop, and output of the demodulated frequency and phase signals; Combination of the demodulated frequency and phase signals with the time sequence to generate a three-dimensional feature map, wherein the first dimension of the three-dimensional feature map represents time, the second dimension represents frequency, and the third dimension represents the equivalent displacement calculated based on the phase signal.

4. The method of claim 2, wherein, Based on the difference value, the equivalent wear amount and wear distribution characteristics inside the articulated point bearing are inferred by inversely solving the physical model in the digital twin model, including: The difference value is taken as the input of the target function, which is used to quantify the overall deviation between the predicted signal of the digital twin model and the joint observation signal; In the parameter space of the wear physical model, the model parameters corresponding to the equivalent wear amount and wear distribution characteristics are iteratively adjusted to minimize the output value of the target function; The model parameter value corresponding to the minimum output value of the target function is determined as the equivalent wear amount and wear distribution characteristics inside the articulated point bearing.

5. The method of claim 2, wherein, The steering angle data in the joint observation signal is taken as the input to drive the digital twin model to generate a predicted signal, including: The steering angle data in the joint observation signal is taken as the input to drive the rigid body dynamics equation in the digital twin model to calculate the dynamic mechanical state inside the articulated point bearing; The dynamic mechanical state is taken as the input to drive the wear physical model in the digital twin model to generate a predicted signal, which represents the expected values of the dynamic stress data and the three-dimensional feature map in an ideal no-wear state.

6. The method of claim 1, wherein, Based on the equivalent wear amount and wear distribution characteristics, warning information for the loosening state of the articulated point bearing is generated, including: Comparing the equivalent wear amount with predetermined first and second threshold values; Analyzing the wear distribution characteristics to calculate a dispersion index in the circumferential direction of the bearing, and comparing the dispersion index with a predetermined uneven distribution threshold value; According to the comparison results of the equivalent wear amount and the threshold values and the comparison results of the dispersion index and the uneven distribution threshold value, warning information for the loosening state of the articulated point bearing is generated.

7. The method of claim 1, wherein, Based on the steering angle data, the mechanical response generated by the articulated point bearing is detected, and the mechanical response is converted into a magnetic feature sequence, which is a time sequence electrical signal representing the change in the magnetic field caused by the mechanical response, including: Based on the steering angle data, the articulated mechanism movement of the downhole trackless rubber-tyred vehicle is driven, and the mechanical response caused by the change in the state of the articulated point bearing is detected; With the help of a magnetic field sensing device, the change in the magnetic field distribution caused by the mechanical response is captured; The physical characteristics of the sensing elements in the magnetic field sensing device are used to convert the change in the magnetic field distribution into corresponding changes in electrical parameters; By continuously collecting and converting the electrical parameter changes, a magnetic feature sequence is formed, which is a time sequence electrical signal representing the change in the magnetic field caused by the mechanical response.

8. A fault monitoring data early warning processing system of a trackless rubber-tyred vehicle in a mine, characterized in that, Including: An acquisition module is configured to acquire steering angle data of the underground trackless rubber-tyred vehicle and dynamic stress data of a hinged point bearing seat in the underground trackless rubber-tyred vehicle. A conversion module is configured to detect a mechanical response generated by the hinged point bearing based on the steering angle data and convert the mechanical response into a magnetic feature sequence, which is a time-series electrical signal representing a magnetic field change caused by the mechanical response. A processing module is configured to perform adaptive frequency control processing on the magnetic feature sequence, dynamically adjust a bandwidth and a center frequency of a phase-locked loop according to spectral characteristics, and output a three-dimensional feature map, wherein the phase-locked loop refers to an internal phase synchronization component relied on by the adaptive automatic frequency control processing, and the three-dimensional feature map is used to describe a frequency response and displacement relationship of the hinged point bearing over time. An inference module is configured to input the steering angle data, the dynamic stress data, and the three-dimensional feature map as joint observation signals into a digital twin model having a physical model of the hinged point bearing, and reversely infer an equivalent wear amount and wear distribution characteristics inside the hinged point bearing by calculating a difference value between the joint observation signals and a predicted signal generated by the digital twin model. A generation module is configured to generate early warning information for a loose state of the hinged point bearing based on the equivalent wear amount and the wear distribution characteristics.

9. An electronic device, comprising: The computer program is stored in the computer readable storage medium and is executed by the processor to implement the steps of the fault monitoring data early warning processing method of the underground trackless rubber-tyred vehicle according to any one of claims 1 to 7. The computer program is stored in the computer readable storage medium and is executed by the processor to implement the steps of the fault monitoring data early warning processing method of the underground trackless rubber-tyred vehicle according to any one of claims 1 to 7. ​ 10. A computer-readable storage medium, characterized in that, ​

Images