A road axle load monitoring system and method based on off-axis polarization piezoelectric effect

Abstract

This invention belongs to the field of highway traffic monitoring and relates to a road axle load monitoring system and method based on off-axis polarization piezoelectric effect. A piezoelectric sensing unit is embedded within the road surface structure layer. The piezoelectric sensing unit comprises a piezoelectric material with off-axis polarization characteristics, and the residual polarization vector direction of the piezoelectric material has a preset angle with its force direction. A signal processing microsystem is electrically connected to the output of the piezoelectric sensing unit to determine axle load information. The input of a power management circuit is electrically connected to the output of the piezoelectric sensing unit, and its output is electrically connected to the power supply of the signal processing microsystem. The power management circuit supplies power to the signal processing microsystem and stores surplus energy in an energy storage element. This enables the system to possess high sensitivity, high durability, adaptive extraction of nonlinear features, and energy self-balancing during axle load detection.

Description

Technical Field

[0001] This invention relates to the field of highway traffic monitoring and intelligent transportation systems, and specifically discloses a road axle load monitoring system and method based on the off-axis polarization piezoelectric effect. Background Technology

[0002] Accurate monitoring of traffic loads is fundamental to achieving smart highways, overload control, and digital twins of infrastructure. Existing road surface dynamic weighing and axle load monitoring solutions mainly fall into two categories: The first category is traditional bending plate and quartz crystal dynamic weighing systems (Weigh-in-Motion, WIM). While these systems offer high measurement accuracy, their sensors and associated data acquisition equipment rely on mains power or high-capacity batteries, requiring the laying of power and data transmission cables along the highway. This results in high construction costs, significant damage to the road structure during construction, and difficulties in maintaining the cables and power supply equipment, making large-scale deployment on already open road sections challenging. The second category comprises axle load detectors based on common piezoelectric materials, such as piezoelectric cable sensors and single-layer piezoelectric ceramic sheet sensors, which have emerged in recent years. The advantages of this type of sensor lie in its relatively simple structure, low cost, and lack of external power supply. However, it generally suffers from the following technical defects in practical engineering applications: Conventional piezoelectric sensors employ axial polarization (polarization direction parallel to the force direction). Although the output voltage is high, the mechanical coupling quality factor is low, making them prone to brittle fracture under dynamic heavy loads due to internal stress concentration, resulting in a sensor lifespan far shorter than that of the road surface. Existing technologies mostly rely on time-domain peak values ​​or simple integration for axle load estimation, ignoring the nonlinear dynamic characteristics of vehicle-road coupling. This leads to a sharp drop in the accuracy of identifying complex axle types (such as double-axle and triple-axle), especially with a significant increase in error when the vehicle is changing speeds. Most research remains focused on single functions such as "energy harvesting" or "sensing," lacking a system-level design that deeply integrates energy management with intelligent signal processing, thus failing to achieve truly maintenance-free long-term deployment.

[0003] In view of this, the present invention provides a road axle load monitoring system and method based on the off-axis polarization piezoelectric effect, which has high sensitivity, high durability, adaptive extraction of nonlinear features and the ability to achieve self-balancing of electrical energy during axle load detection. Summary of the Invention

[0004] The purpose of this invention is to provide a road axle load monitoring system and method based on the off-axis polarization piezoelectric effect, addressing the problem of how to achieve high sensitivity, high durability, adaptive extraction of nonlinear features, and self-balancing of electrical energy during axle load detection. The specific solution is as follows: A road axle load monitoring system based on off-axis polarization piezoelectric effect includes at least one piezoelectric sensing unit, a signal processing microsystem, a power management circuit, and a wireless communication unit. The piezoelectric sensing unit is embedded within the road surface structure layer. It comprises a piezoelectric material with off-axis polarization characteristics, and the residual polarization vector direction of the piezoelectric material has a preset angle with its applied force direction, used to convert the mechanical stress generated by vehicle loads into electrical signals through the off-axis polarization piezoelectric effect. The signal processing microsystem is electrically connected to the output of the piezoelectric sensing unit for collecting electrical signals. The signal is processed, and the axle load information is determined based on the processed signal and the preset load-voltage correspondence. The input terminal of the power supply management circuit is electrically connected to the output terminal of the piezoelectric sensing unit, and its output terminal is electrically connected to the power supply terminal of the signal processing microsystem. The power supply management circuit is used to rectify, stabilize, and distribute part of the electrical energy generated by the piezoelectric sensing unit to power the signal processing microsystem, and to store the surplus electrical energy in the energy storage element. The wireless communication unit is electrically connected to the data output terminal of the signal processing microsystem and is used to send the axle load information to the cloud data center.

[0005] Furthermore, the piezoelectric sensing unit includes a piezoelectric-polymer composite functional body, an upper electrode layer, and a lower electrode layer: the piezoelectric-polymer composite functional body is composed of piezoelectric ceramic fibers and an epoxy resin matrix, and the residual polarization vector direction of the piezoelectric-polymer composite functional body is formed by off-axis polarization process; the upper electrode layer is disposed on the upper surface of the piezoelectric-polymer composite functional body; and the lower electrode layer is disposed on the lower surface of the piezoelectric-polymer composite functional body.

[0006] Furthermore, the signal processing microsystem includes a charge amplification unit, an analog-to-digital conversion unit, and an embedded edge computing unit: the charge amplification unit is used to amplify and filter the electrical signal; the analog-to-digital conversion unit is electrically connected to the output of the charge amplification unit and is used to convert the processed electrical signal into a digital signal; the embedded edge computing unit is electrically connected to the output of the analog-to-digital conversion unit and internally contains a time-frequency feature extraction algorithm and a pre-trained classification model; the time-frequency feature extraction algorithm is used to perform time-frequency decomposition and feature extraction on the digital signal to obtain a feature vector containing instantaneous frequency and energy distribution; the classification model is used to identify the axle load type and axle load value based on the extracted feature vector.

[0007] Furthermore, the power supply management circuit includes a rectification-boost management circuit, a voltage regulation circuit, and an energy storage element; the input terminal of the rectification-boost management circuit is electrically connected to the output terminal of the piezoelectric sensing unit, and is used to convert the AC power generated by the piezoelectric sensing unit into DC power; the voltage regulation circuit is electrically connected to the output terminal of the rectification-boost management circuit, and is used to stably supply the voltage of the rectified DC power to the signal processing microsystem and the wireless communication unit; the energy storage element is electrically connected to the output terminal of the voltage regulation circuit, and is used to store surplus electrical energy.

[0008] This invention also provides a road axle load monitoring method based on the off-axis polarization piezoelectric effect, applied to any of the road axle load monitoring systems based on the off-axis polarization piezoelectric effect described above, comprising: Step 1: Sensing vehicle load using a piezoelectric sensing unit, and converting mechanical stress into an electrical signal through the off-axis polarization piezoelectric effect; the residual polarization vector direction of the piezoelectric material in the piezoelectric sensing unit has a preset angle with its force direction; Step 2: Acquiring the electrical signal and performing signal processing to obtain the processed signal; Step 3: Determining axle load information based on the processed signal; the axle load information includes the number of axles, axle group type, and axle load level of the vehicle.

[0009] Furthermore, the preset angle between the residual polarization vector direction of the piezoelectric material and its applied force direction in the piezoelectric sensing unit is achieved through an off-axis polarization process, including: Step 11: heating the piezoelectric material to a polarization temperature of 110℃~140℃; Step 12: applying an external strong electric field with a preset direction to the piezoelectric material to polarize it along that predetermined direction; the polarization electric field strength is 2.5kV / mm~4.0kV / mm; the polarization time is 20min~40min; Step 13: cooling the piezoelectric material to room temperature to fix the residual polarization vector along the preset direction, thereby forming a preset angle with the actual applied force direction of the piezoelectric sensing unit.

[0010] Furthermore, the preset included angle ranges from 30° to 60°.

[0011] Further, step 2 includes: step 21: performing charge amplification and filtering on the acquired electrical signal to obtain a processed electrical signal; step 22: performing analog-to-digital conversion on the processed electrical signal to obtain a digital signal; step 23: performing time-frequency decomposition on the digital signal to obtain multiple modal components; step 24: extracting feature parameters from the multiple modal components to form a feature vector; step 25: inputting the feature vector into a pre-trained classification model to identify the vehicle axle load type and axle load value.

[0012] Furthermore, obtaining multiple modal components involves: adaptively decomposing the digital signal into multiple intrinsic mode function components with different center frequencies using a variational mode decomposition algorithm; each intrinsic mode function component represents the oscillation mode of the signal at different frequency scales; constructing the feature vector involves: extracting the instantaneous amplitude and center frequency from each intrinsic mode function component obtained by decomposition, and constructing a multidimensional feature phase space as the feature vector; the classification model is a one-dimensional convolutional neural network, which sequentially includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a flattening layer, a fully connected layer, and a Softmax output layer; the first convolutional layer has 32 3×1 convolutional kernels; the second convolutional layer has 64 3×1 convolutional kernels; and the fully connected layer has 64 neurons.

[0013] Furthermore, the piezoelectric material in the piezoelectric sensing unit is composed of piezoelectric ceramic fibers and epoxy resin.

[0014] The present invention has the following advantages and beneficial effects: This invention breaks away from the conventional design of axial polarization in piezoelectric sensors and adopts off-axis polarization (at a specific angle, such as 45°), which enables the piezoelectric composite functional body to generate a combined stress mode of shear and compression when subjected to vertical wheel loads. This not only significantly improves the dynamic charge output sensitivity, but also changes the internal stress distribution path, greatly improving the fatigue fracture resistance.

[0015] This invention introduces variational mode decomposition, an advanced adaptive signal processing method, to decompose complex axle load signals into narrowband intrinsic mode functions with clear physical meaning, and extract instantaneous frequency and energy entropy as features, thus solving the problem of poor robustness of traditional time-domain features under variable speed and multi-axis interference.

[0016] This invention designs a composite energy storage unit with a wide input range rectification-boost management circuit and a miniature solid-state lithium capacitor, and combines it with a system-level energy management strategy, achieving for the first time a leap from "milliwatt-level instantaneous output" to "microwatt-level long-term self-sustaining" for piezoelectric surface sensors. Attached Figure Description

[0017] Figure 1 This is an exemplary structural diagram of a road axle load monitoring system based on the off-axis polarization piezoelectric effect provided by the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0019] Example 1 This embodiment provides a road axle load monitoring system based on the off-axis polarization piezoelectric effect, such as... Figure 1 As shown, the system has self-sensing, self-powering and self-identification capabilities, including at least one intelligent piezoelectric sensing unit, a signal processing microsystem, a power management circuit and a wireless communication unit, as well as a stress-adaptive packaging shell.

[0020] The intelligent piezoelectric sensing unit is embedded within the road surface structure layer. The intelligent piezoelectric sensing unit includes a piezoelectric sensing unit, which comprises a piezoelectric material with off-axis polarization characteristics, specifically a piezoelectric-polymer composite functional body composed of piezoelectric ceramic fibers and an epoxy resin matrix. For example, a type 1-3 piezoelectric ceramic fiber (PZT-5H) / epoxy resin composite material is used, with a fiber diameter of 250 μm and a volume fraction of 55% in the composite material. This composite material is fabricated into a block with dimensions of 30 mm × 30 mm × 5 mm, and an upper electrode layer and a lower electrode layer are respectively disposed on its upper and lower surfaces.

[0021] The piezoelectric-polymer composite functional body is processed using an off-axis polarization process, which creates a preset angle between the direction of its residual polarization vector and the preset force direction. In some embodiments, the specific parameters of the off-axis polarization process are a polarization electric field strength of 2.5 kV / mm to 4.0 kV / mm, a polarization temperature of 110°C to 140°C, and a polarization time of 20 min to 40 min, so that the direction of its residual polarization vector forms an angle of 30° to 60° (e.g., 45°) with the preset force direction. For example, the specific steps of the polarization process include: heating the piezoelectric-polymer composite functional body to a polarization temperature of 120°C; applying an external strong electric field with a polarization electric field strength of 3.5 kV / mm, the direction of which forms a 45° angle with the preset perpendicular force direction, and continuing polarization for 30 minutes; then cooling to room temperature, so that the residual polarization vector is fixed along the preset direction, thereby forming a preset angle of 45° with the actual force direction of the piezoelectric sensing unit.

[0022] The piezoelectric sensing unit is embedded at the interface between the bottom of the road surface layer and the base layer. The planar embedding density of the intelligent piezoelectric sensing unit is one unit every 1.5m to 2.5m (e.g., 2m) along the driving direction, and one unit is placed at each wheel track in the cross-sectional direction. The longitudinal spacing error between adjacent units does not exceed ±5%.

[0023] The stress-adaptive enclosure seals the piezoelectric sensing unit, signal processing microsystem, power management circuitry, and wireless communication unit within its casing. This enclosure features a multi-layered stress-adjusting structure: a 3mm thick, curved, hard alloy force-transmitting plate (150mm radius curvature) at the top to bear vehicle wheel loads and uniformly transfer stress to the piezoelectric-polymer composite functional body; a 2mm thick nitrile rubber viscoelastic damping layer at the bottom to absorb residual vibration and impact energy; and 304 stainless steel sidewalls. The overall dimensions after encapsulation are 40mm × 40mm × 10mm.

[0024] The signal processing microsystem is electrically connected to the output of the piezoelectric sensing unit and includes a charge amplification unit, an analog-to-digital converter (ADC), and an embedded edge computing unit. The charge amplification unit uses an AD8605 operational amplifier to amplify and filter the electrical signal output from the piezoelectric sensing unit. The ADC has 16-bit resolution and a sampling frequency of at least 2kHz (e.g., 2.5kHz), converting the processed analog electrical signal into a digital signal. The embedded edge computing unit uses an STM32F407 microcontroller, internally containing a time-frequency feature extraction algorithm and a pre-trained classification model. Specifically, the embedded edge computing unit uses a variational mode decomposition algorithm to decompose the original signal, extracting the instantaneous amplitude and center frequency of the first three intrinsic mode functions to construct a multidimensional feature phase space. It should be noted that the load-voltage correspondence refers to the mapping relationship between the mechanical load applied to the piezoelectric material and its output voltage. According to the piezoelectric constitutive equation, the charge Q generated by a piezoelectric material under force is proportional to the applied force F. For a piezoelectric sensing unit with a specific capacitance C, its output voltage V = Q / C∝d×F, meaning that under ideal conditions, the load and voltage V have a linear relationship. Here, d is the piezoelectric strain constant. In existing technologies, this relationship is usually determined through calibration experiments. That is, a standard vehicle with a known axle load passes through the piezoelectric sensing unit, and the voltage response signal is simultaneously collected. After signal processing, a load-voltage calibration curve or calibration coefficient is obtained by fitting the signal.

[0025] The time-frequency feature extraction algorithm is based on variational mode decomposition (VMD), which adaptively decomposes a digital signal into multiple intrinsic mode functions (EMFs). In some embodiments, the penalty factor α in the VMD algorithm is 2000~4000, the number of mode decompositions K is 3~6, ​​and the convergence tolerance τ is 1e-6~1e-7. For example, the penalty factor α is 3000, the number of mode decompositions K is 4, and the convergence tolerance τ is 1e-6. This algorithm decomposes the original shaft-loaded signal into multiple narrowband EMF components with different center frequencies. Each EMF component represents the oscillation mode of the signal at different frequency scales. The instantaneous amplitude and center frequency of the first three EMFs are extracted to construct a multidimensional feature phase space, which serves as a feature vector containing instantaneous frequency and energy distribution.

[0026] The classification model is a lightweight one-dimensional convolutional neural network, whose structure includes, in sequence, an input layer, a first convolutional layer (with 32 3×1 convolutional kernels), a first pooling layer, a second convolutional layer (with 64 3×1 convolutional kernels), a second pooling layer, a flattening layer, a fully connected layer (with 64 neurons), and a softmax output layer. During model training, K-fold cross-validation and early stopping are used to prevent overfitting, and the training set to validation set ratio is 7:3.

[0027] The power management circuit includes a rectification-boost management circuit, a voltage regulation circuit, and an energy storage element. The rectification-boost management circuit uses an LTC3331 chip, whose input is electrically connected to the output of the piezoelectric sensing unit. It accepts a wide range of AC input from ±2V to ±50V, converting the AC power generated by the piezoelectric sensing unit into DC power and boosting it to a usable voltage level. The voltage regulation circuit stably outputs 3.3V or 5.0V DC power to supply the signal processing microsystem and wireless communication unit. The energy storage element uses a miniature solid-state lithium capacitor. In some embodiments, the capacity of the miniature solid-state lithium capacitor is 50mAh to 200mAh, and the operating voltage is 3.3V to 5.0V. The wide-input-range rectification-boost management circuit accepts AC input from ±2V to ±50V and stably outputs 3.3V or 5.0V DC power. For example, a 100mAh capacitor with an operating voltage of 3.8V (FPC3.8V100mAh) is electrically connected to the output of the voltage regulation circuit to store surplus energy.

[0028] When the power management circuit and energy storage element work together, the following power management strategy is executed: when the capacitor voltage is higher than 4.2V, the rectifier-boost management circuit automatically cuts off the charging circuit and starts discharging to supply power; when the capacitor voltage is lower than 3.3V, the rectifier-boost management circuit activates the low power mode, only maintaining the core real-time clock and wake-up listening function, and resumes normal operation after the next vehicle passes by and generates power.

[0029] The wireless communication unit is electrically connected to the data output terminal of the signal processing microsystem and is used to send axle load information to the cloud data center.

[0030] Example 2 This invention also provides a road axle load monitoring method based on off-axis polarization piezoelectric effect, the specific working process of which is as follows: Step 1: When the vehicle passes by, the stress is transmitted through the arc-shaped hard alloy force transmission plate to the off-axis polarized piezoelectric-polymer composite functional body. The composite functional body converts the mechanical stress into an electrical signal with rich frequency response characteristics through the off-axis polarized piezoelectric effect. At the same time, the AC power generated by the piezoelectric sensing unit is input into the power supply management circuit.

[0031] Step 2: The signal processing microsystem acquires the electrical signal and performs signal processing. Specifically, the charge amplification unit amplifies and filters the electrical signal; the analog-to-digital conversion unit performs analog-to-digital conversion at a sampling frequency of not less than 2kHz (e.g., 2.5kHz) to obtain a digital signal. The embedded edge computing unit calls the variational mode decomposition algorithm to decompose the original signal, extracting the instantaneous amplitude and center frequency of the first three intrinsic mode functions, and constructing a multidimensional feature phase space. Based on the processed signal, axle load information is determined; the axle load information includes the number of axles, axle group type, and axle load level of the vehicle. The embedded edge computing unit calls the variational mode decomposition algorithm to perform time-frequency decomposition on the digital signal, obtaining multiple intrinsic mode function components; the instantaneous amplitude and center frequency are extracted from each intrinsic mode function component to form a multidimensional feature vector containing instantaneous frequency and energy distribution; this feature vector is input into a pre-trained one-dimensional convolutional neural network classification model to identify the number of axles, axle group type, and axle load level of the vehicle.

[0032] Step 3: Determine axle load information based on the processed signal, and send the identification results and feature data to the cloud data center through the wireless communication unit; the axle load information includes the number of axles, axle group type and axle load level of the vehicle.

[0033] During the execution of the above steps, the power supply management circuit converts the AC power generated in step 1 into stable DC power through the rectification-boost management circuit, and supplies it to the signal processing and communication links in steps 2 and 3 through the voltage regulation circuit. The remaining power is stored in the miniature solid-state lithium capacitor to achieve system power self-balancing.

[0034] Example 3 This embodiment is basically the same as Embodiment 1, except that in the off-axis polarization process, the preset angle between the polarization direction and the force direction is 30°. The rest of the structure, process parameters and workflow are the same as in Embodiment 1.

[0035] To verify the technical effectiveness of the present invention, the following comparative examples were set up for performance comparison tests: Comparative Example 1 (Conventional Axial Polarization): It is basically the same as Example 1, except that the polarization mode of the piezoelectric-polymer composite functional body is changed to conventional axial polarization (the polarization direction is parallel to the force direction), and the other parameters are the same.

[0036] Comparative Example 2 (Traditional Time-Domain Peak Algorithm): The system hardware is exactly the same as in Example 1, but the signal processing part abandons variational mode decomposition and convolutional neural network, and uses conventional time-domain peak and waveform width for axis load fitting.

[0037] Comparative Example 3 (existing commercially available piezoelectric axle load cell): A piezoelectric quartz load cell (typical product) of a certain imported brand was selected, and it was installed according to its standard installation process, with an external data acquisition instrument and power supply connected.

[0038] Experimental Example: Performance Comparison Test Sensors or systems from Examples 1, 3, and Comparative Examples 1 to 3 were embedded in a full-scale asphalt pavement accelerated loading test track. Standard axle loads (10t for a single axle, 20t for a double axle) and various vehicle speeds (30km / h, 60km / h, 80km / h) were applied, and calibration was performed using a high-precision weighbridge. The main test results are as follows: Table 1: Comparison of Energy Harvesting and Power Management Performance

[0039] Note: Comparative Example 2 uses the same hardware as Example 1, so the energy collection is the same; Comparative Example 2 has no energy storage unit, and Comparative Example 3 requires an external power supply, so it cannot be self-sustaining.

[0040] Table 2: Comparison of Dynamic Axle Load Recognition Performance

[0041] The results in Tables 1 and 2 show that: Example 1 (45° off-axis polarization, nonlinear time-frequency characteristics and hybrid energy storage) is significantly better than the comparative examples in terms of energy harvesting efficiency (4.8 mJ / cycle), fatigue life (energy decay of only 8% after 500,000 loading cycles) and identification accuracy (97% for shaft type and 95% for shaft load level), which verifies the synergistic advantages of off-axis polarization in mechanical and electrical performance.

[0042] Compared to Comparative Example 1 (axial polarization), off-axis polarization not only improves energy harvesting efficiency by 4 times, but also significantly increases fatigue life, proving that it solves the technical bias of traditional piezoelectric sensors being prone to fatigue fracture.

[0043] Compared to Comparative Example 2 (traditional algorithm), under the same hardware conditions, the variational mode decomposition + lightweight convolutional neural network algorithm of this invention reduces the axle load estimation error under variable speed from ±20% to ±6%, demonstrating the superiority of nonlinear time-frequency feature extraction in processing complex signals.

[0044] The system in Example 1 can start at vehicle speeds as low as 5 km / h and can sustain itself for up to 120 days in the absence of traffic, truly achieving "maintenance-free" and "self-powered" road monitoring nodes.

[0045] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A road axle load monitoring system based on off-axis polarization piezoelectric effect, characterized in that, It includes at least one piezoelectric sensing unit, a signal processing microsystem, a power management circuit, and a wireless communication unit: The piezoelectric sensing unit is embedded in the road surface structure layer; the piezoelectric sensing unit contains a piezoelectric material with off-axis polarization characteristics, and the residual polarization vector direction of the piezoelectric material has a preset angle with its force direction, which is used to convert the mechanical stress generated by the vehicle load into an electrical signal through the off-axis polarization piezoelectric effect. The output terminals of the signal processing microsystem and the piezoelectric sensing unit are electrically connected to collect electrical signals and perform signal processing. Based on the processed signal and the preset load-voltage correspondence, the axle load information is determined. The input terminal of the power management circuit is electrically connected to the output terminal of the piezoelectric sensing unit, and its output terminal is electrically connected to the power supply terminal of the signal processing microsystem. The power management circuit is used to rectify, regulate and distribute part of the electrical energy generated by the piezoelectric sensing unit to power the signal processing microsystem, and store the surplus electrical energy in the energy storage element. The wireless communication unit and the data output terminal of the signal processing microsystem are electrically connected and used to send axle load information to the cloud data center.

2. The road axle load monitoring system based on off-axis polarization piezoelectric effect according to claim 1, characterized in that, The piezoelectric sensing unit includes a piezoelectric-polymer composite functional body, an upper electrode layer, and a lower electrode layer: The piezoelectric-polymer composite functional body is composed of piezoelectric ceramic fibers and epoxy resin matrix, and the residual polarization vector direction of the piezoelectric-polymer composite functional body is formed by off-axis polarization process. The upper electrode layer is disposed on the upper surface of the piezoelectric-polymer composite functional body; The lower electrode layer is disposed on the lower surface of the piezoelectric-polymer composite functional body.

3. The road axle load monitoring system based on off-axis polarization piezoelectric effect according to claim 1, characterized in that, The signal processing microsystem includes a charge amplification unit, an analog-to-digital conversion unit, and an embedded edge computing unit. The charge amplification unit is used to amplify and filter electrical signals; The analog-to-digital conversion unit is electrically connected to the output terminal of the charge amplification unit, and is used to convert the processed electrical signal into a digital signal. The embedded edge computing unit is electrically connected to the output of the analog-to-digital conversion unit, and internally contains a time-frequency feature extraction algorithm and a pre-trained classification model. The time-frequency feature extraction algorithm is used to perform time-frequency decomposition and feature extraction on digital signals to obtain feature vectors containing instantaneous frequency and energy distribution. The classification model is used to identify axle load type and axle load value based on the extracted feature vectors.

4. The road axle load monitoring system based on off-axis polarization piezoelectric effect according to claim 1, characterized in that, The power supply management circuit includes a rectifier-boost management circuit, a voltage regulation circuit, and an energy storage element; The input terminal of the rectifier-boost management circuit is electrically connected to the output terminal of the piezoelectric sensing unit, and is used to convert the AC power generated by the piezoelectric sensing unit into DC power. The voltage regulation circuit is electrically connected to the output of the rectifier-boost management circuit, and is used to stably supply the voltage of the rectified DC power to the signal processing microsystem and the wireless communication unit. The energy storage element is electrically connected to the output terminal of the voltage regulation circuit and is used to store surplus electrical energy.

5. A method for monitoring road axle loads based on off-axis polarization piezoelectric effect, characterized in that, The road axle load monitoring system based on off-axis polarization piezoelectric effect as described in any one of claims 1-4 includes: Step 1: The vehicle load is sensed using a piezoelectric sensing unit, and the mechanical stress is converted into an electrical signal through the off-axis polarized piezoelectric effect; the residual polarization vector direction of the piezoelectric material in the piezoelectric sensing unit has a preset angle with the direction of the force it is subjected to; Step 2: Acquire electrical signals and perform signal processing to obtain the processed signal; Step 3: Determine axle load information based on the processed signal; the axle load information includes the number of axles, axle group type, and axle load level of the vehicle.

6. The road axle load monitoring method based on off-axis polarization piezoelectric effect according to claim 5, characterized in that, The preset angle between the residual polarization vector direction of the piezoelectric material and its applied force direction in the piezoelectric sensing unit is achieved through an off-axis polarization process, including: Step 11: Heat the piezoelectric material to its polarization temperature; the polarization temperature is 110℃~140℃; Step 12: Apply a strong external electric field with a preset direction to the piezoelectric material to polarize the piezoelectric material along the preset direction; the polarization electric field strength is 2.5kV / mm~4.0kV / mm; the polarization time is 20min~40min; Step 13: Cool the piezoelectric material to room temperature, so that the remaining polarization vector is fixed along the preset direction, thereby forming a preset angle with the actual force direction of the piezoelectric sensing unit.

7. The road axle load monitoring method based on off-axis polarization piezoelectric effect according to claim 6, characterized in that, The preset included angle ranges from 30° to 60°.

8. The road axle load monitoring method based on off-axis polarization piezoelectric effect according to claim 5, characterized in that, Step 2 includes: Step 21: Perform charge amplification and filtering on the acquired electrical signal to obtain the processed electrical signal; Step 22: Perform analog-to-digital conversion on the processed electrical signal to obtain a digital signal; Step 23: Perform time-frequency decomposition on the digital signal to obtain multiple modal components; Step 24: Extract feature parameters from multiple modal components to form a feature vector; Step 25: Input the feature vector into the pre-trained classification model to identify the vehicle axle load type and axle load value.

9. The road axle load monitoring method based on off-axis polarization piezoelectric effect according to claim 8, characterized in that, The multiple modal components are obtained by: adaptively decomposing the digital signal into multiple intrinsic mode function components with different center frequencies using a variational mode decomposition algorithm; each intrinsic mode function component represents the oscillation mode of the signal at different frequency scales; The aforementioned feature vector is constructed by extracting the instantaneous amplitude and center frequency from each intrinsic mode function component obtained from the decomposition, and then constructing a multidimensional feature phase space as the feature vector. The classification model is a one-dimensional convolutional neural network, which includes, in sequence, an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, a flattening layer, a fully connected layer, and a softmax output layer; the first convolutional layer has 32 3×1 convolutional kernels; the second convolutional layer has 64 3×1 convolutional kernels; and the fully connected layer has 64 neurons.

10. The road axle load monitoring method based on off-axis polarization piezoelectric effect according to claim 5, characterized in that, The piezoelectric material in the piezoelectric sensing unit is composed of piezoelectric ceramic fibers and epoxy resin.

Images