A magnetoelectric-piezoelectric effect sensor array current measurement device

Abstract

The application discloses a magnetoelectric-piezoelectric effect sensing array current measuring device, and relates to the technical field of power system measurement, wherein the sensing array is a mixed L type co-prime sparse array, and comprises a plurality of sensing units based on the magnetoelectric-piezoelectric effect; the sensing array is used for collecting the magnetic field signals of the measured conductor at multiple time points and converting the magnetic field signals into multiple voltage signals; an array signal processing module is used for performing signal conditioning, analog-digital conversion and digital signal processing on the multiple voltage signals at the multiple time points, so as to obtain magnetic field space distribution data; a temperature compensation module is used for obtaining the environmental temperature of the measured conductor and performing temperature compensation on the magnetic field space distribution data; and a current inversion module is used for determining the current value of the measured conductor by using a built-in current inversion model and according to the temperature-compensated magnetic field space distribution data. The application can solve the contradiction between the sensitivity of the magnetoelectric sensing unit and the miniaturization of the device, the angle estimation ambiguity of the sparse array and the problem that the temperature drift influences the measurement accuracy.

Description

Technical Field

[0001] This application relates to the field of power system measurement technology, and in particular to a magnetoelectric-piezoelectric effect sensor array current measurement device. Background Technology

[0002] With the deepening of smart grid construction and the rapid development of power Internet of Things technology, increasingly higher demands are being placed on the comprehensive perception, real-time monitoring, and accurate measurement of the operating status of transmission lines. As one of the most critical operating parameters of a power system, the advancement of current measurement technology directly affects the safe, stable operation and intelligent level of the power grid. Looking back at the development history of current measurement technology, it has roughly gone through three stages: The first generation was the traditional electromagnetic current transformer (CT) based on the principle of electromagnetic induction. While technologically mature and highly accurate, it suffered from inherent defects such as requiring series connection to the line, complex insulation structure, large size and weight, and susceptibility to ferroresonance and magnetic saturation. Furthermore, it could not achieve non-contact measurement, making it difficult to meet the needs of flexible deployment and mobile inspection.

[0003] The second generation consists of non-invasive current sensors based on the Hall effect, magnetoresistive effect, or Rogowski coils. These technologies can measure current without disconnecting the circuit, which solves the installation problems of traditional CT scanners to some extent. However, Hall sensors suffer from problems such as large temperature drift, poor linearity, and limited sensitivity, and usually require complex temperature compensation and signal conditioning circuits. Although magnetoresistive sensors have improved sensitivity, their measurement range is narrow and they are susceptible to interference from external magnetic fields. While Rogowski coils have the advantages of no magnetic saturation and a wide measurement range, their output is a differential voltage signal, which requires integration circuit processing, and the measurement accuracy is greatly affected by the coil geometry, installation position, and relative position of the wires, making it difficult to guarantee the consistency and accuracy of the measurement.

[0004] In recent years, novel magnetoelectric sensors based on the magnetoelectric coupling effect have attracted widespread attention from academia and industry due to their unique advantages. Magnetoelectric sensors utilize the composite structure of magnetostrictive and piezoelectric materials to directly convert magnetic field energy into electrical energy output. They are characterized by being passive, highly sensitive, and having a wide bandwidth, providing a new technical path for non-invasive current measurement. However, most existing magnetoelectric current measurement technologies are still in the laboratory research stage and face many challenges in practical engineering applications: (1) The sensitivity of a single magnetoelectric sensing unit is limited. To achieve long-distance detection of weak magnetic fields, it is often necessary to increase the size of the sensing unit, which contradicts the need for device miniaturization; (2) Although sparse array technology can achieve a larger array aperture with fewer sensing units, the inherent ambiguity problem of angle estimation seriously restricts the measurement accuracy; (3) The key parameters of magnetoelectric materials are significantly affected by temperature, causing the sensor output to drift with temperature changes, which reduces the measurement accuracy.

[0005] Therefore, there is an urgent need for a magnetoelectric-piezoelectric effect sensor array current measurement device that can solve the above problems. Summary of the Invention

[0006] The purpose of this application is to provide a current measurement device for a magnetoelectric-piezoelectric effect sensing array to solve the problems of the contradiction between the sensitivity of the magnetoelectric sensing unit and the miniaturization of the device, the ambiguity of the angle estimation of the sparse array, and the impact of temperature drift on the measurement accuracy.

[0007] To achieve the above objectives, this application provides the following solution.

[0008] This application provides a magnetoelectric-piezoelectric effect sensing array current measurement device, comprising: a sensing array, an array signal processing module, a temperature compensation module, and a current inversion module connected in sequence; the sensing array includes multiple sensing units based on the magnetoelectric-piezoelectric effect, and the sensing array is a hybrid L-type coprime sparse array; The sensor array is used to acquire the magnetic field signal of the conductor under test at multiple times and convert it into multiple voltage signals. The array signal processing module is used to perform signal conditioning, analog-to-digital conversion, and digital signal processing on multiple voltage signals at multiple times to obtain magnetic field spatial distribution data; the digital signal processing is implemented using an improved multi-signal classification algorithm. The temperature compensation module is used to obtain the ambient temperature of the conductor under test, perform temperature compensation on the magnetic field spatial distribution data, and obtain the temperature-compensated magnetic field spatial distribution data. The current inversion module is used to determine the current value of the conductor under test based on the temperature-compensated magnetic field spatial distribution data using the built-in current inversion model.

[0009] In one embodiment, the sensing unit includes an upper protective layer, a magnetostrictive layer, an intermediate adhesive layer, a piezoelectric layer, and a lower protective layer stacked sequentially from top to bottom; The magnetostrictive layer is used to generate mechanical strain under the action of an alternating magnetic field when current flows through the conductor under test and an alternating magnetic field is generated. The intermediate adhesive layer is used to transfer the mechanical strain generated by the magnetostrictive layer to the piezoelectric layer; The piezoelectric layer is used to generate a polarization voltage under the received mechanical strain, and outputs a voltage signal that is proportional to the magnetic field strength of the alternating magnetic field.

[0010] In one embodiment, the hybrid L-type coprime sparse array includes a triangular electromagnetic vector sensor located at the origin, a first scalar quantum array arranged along the x-axis, and a second scalar quantum array arranged along the y-axis; both the first and second scalar quantum arrays are composed of multiple sensing units. Each sensing unit is arranged along the x-axis and y-axis with a spacing between sensing units that are coprime integers, for the purpose of synchronously acquiring the multiple voltage signals in space.

[0011] In one embodiment, the array signal processing module includes a signal conditioning circuit, an analog-to-digital conversion circuit, and a digital signal processing unit connected in sequence. The signal conditioning circuit is used to condition the multiple voltage signals to obtain conditioned analog signals; the signal conditioning includes amplification and filtering. The analog-to-digital conversion circuit is used to perform analog-to-digital conversion on multiple conditioned analog signals to obtain multiple digital signals. The digital signal processing unit is used to process multiple digital signals at multiple times using an improved multi-signal classification algorithm to obtain the magnetic field spatial distribution data.

[0012] In one embodiment, an improved multi-signal classification algorithm is used to process multiple digital signals at multiple times to obtain the magnetic field spatial distribution data, including: Construct an array signal model based on multiple digital signals at multiple times; Based on the array signal model, construct the array covariance matrix; The array covariance matrix is ​​decomposed into eigenvalues ​​to obtain the signal subspace and the noise subspace. Tensor decomposition is performed on the array covariance matrix to reconstruct the virtual comatrix covariance matrix; By utilizing the orthogonality between the noise subspace and the steering vector of the sensing array, a spatial spectrum function is constructed based on the virtual covariance matrix, and a spectral peak search is performed on the spatial spectrum function to determine the estimated value of the direction of arrival of the tested conductor. Based on the coprime array configuration of the sensor array, the direction of arrival (DOA) estimate is de-ambigued to eliminate false spectral peaks and obtain an unambiguous DOA estimate. Based on the direction of arrival estimation results, the spatial distribution data of the magnetic field is generated.

[0013] In one embodiment, the temperature compensation module includes a surface acoustic wave temperature sensor, an RF read / write unit, and a temperature-magnetic field decoupling unit; the surface acoustic wave temperature sensor and the RF read / write unit are wirelessly connected, and the RF read / write unit is electrically connected to the temperature-magnetic field decoupling algorithm unit. The radio frequency read / write unit is used to transmit a radio frequency query signal to the surface acoustic wave temperature sensor and receive the reflected echo signal generated by the surface acoustic wave temperature sensor according to the radio frequency query signal. The temperature-magnetic field decoupling unit has a built-in temperature-magnetic field decoupling algorithm, which is used to determine the ambient temperature of the conductor under test based on the reflected echo signal, and to perform temperature compensation on the magnetic field spatial distribution data according to the ambient temperature, so as to obtain the temperature-compensated magnetic field spatial distribution data.

[0014] In one embodiment, the process of determining the current inversion model includes: Obtain the training dataset; the training dataset includes sample magnetic field spatial distribution data and corresponding true current values; A deep learning inversion network is constructed; the deep learning inversion network includes three convolutional neural networks and two fully connected layers connected in sequence; each convolutional neural network is constructed based on a 3×3 filter, a batch normalization layer and a ReLU activation function; The deep learning inversion network is trained on the training dataset using a loss function that incorporates physical constraints to obtain the current inversion model; the physical constraints include Ampere's circuital law constraints, current amplitude limits, and L2 regularization terms.

[0015] In one embodiment, the sensor array is provided with at least one differential anti-interference structure; The differential anti-interference structure includes two sensing units and a differential operation circuit; the two sensing units are symmetrically arranged along a direction perpendicular to the conductor being measured, and both sensing units are connected to the differential operation circuit. The differential operation circuit is used to subtract the voltage signals output by the two sensing units to obtain a differential voltage signal, and output the differential voltage signal as one voltage signal of the sensing array.

[0016] In one embodiment, the magnetoelectric-piezoelectric effect sensing array current measurement device further includes: a main control and communication module; the main control and communication module includes a microcontroller and a wireless communication unit connected in sequence, the microcontroller is connected to the current inversion module, and the wireless communication unit is wirelessly connected to a remote external device; The microcontroller is used to receive the current value; The wireless communication unit is used to send the current value to an external device.

[0017] In one embodiment, the magnetoelectric-piezoelectric effect sensing array current measuring device further includes: a protective housing; The sensor array, the array signal processing module, the temperature compensation module, the current inversion module, and the main control and communication module are all encapsulated within the protective housing; the protective housing has an IP65 or higher protection rating.

[0018] According to the specific embodiments provided in this application, the following technical effects are disclosed: This application discloses a magnetoelectric-piezoelectric effect sensing array current measurement device. It acquires the magnetic field signal of the measured conductor at multiple moments using a sensing array and converts it into multiple voltage signals. Multiple sensing units based on the magnetoelectric-piezoelectric effect are arranged in a hybrid L-shaped coprime sparse array, enabling equivalent aperture expansion within a miniaturized size with a smaller number of sensing units. This effectively resolves the contradiction between the sensitivity of a single magnetoelectric sensing unit and the miniaturization of the device, achieving high-sensitivity acquisition of weak magnetic field signals over long distances. The array signal processing module uses an improved multi-signal classification algorithm to perform digital signal processing on the multiple voltage signals. This effectively suppresses multipath interference, solves the ambiguity problem in sparse array angle estimation, and obtains accurate magnetic field spatial distribution data, improving direction finding accuracy and spatial resolution. The temperature compensation module acquires the ambient temperature of the measured conductor and performs temperature compensation on the magnetic field spatial distribution data, eliminating the influence of magnetoelectric material temperature drift on measurement accuracy. The current inversion module inverts the current value based on the temperature-compensated magnetic field spatial distribution data, realizing long-distance non-contact current measurement of high-voltage transmission lines and improving the accuracy and reliability of current measurement. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the embodiments 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.

[0020] Figure 1 This is a schematic diagram of the structure of a magnetoelectric-piezoelectric effect sensor array current measurement device provided in an embodiment of this application; Figure 2 This is a schematic diagram of the sensor unit structure; Figure 3 A schematic diagram of a hybrid L-shaped coprime sparse array structure; Figure 4 This is a schematic diagram of the differential anti-interference structure; Figure 5 A schematic diagram of the overall structure of a miniaturized long-distance voltage and current measuring instrument with a magnetoelectric-piezoelectric effect sensor array; Figure 6 This is an interface diagram of a miniaturized long-distance voltage and current measuring instrument based on a magnetoelectric-piezoelectric effect sensor array.

[0021] Figure label: Sensing array—1, array signal processing module—2, temperature compensation module—3, current inversion module—4, main control and communication module—5, upper protective layer—6, magnetostrictive layer—7, intermediate adhesive layer—8, piezoelectric layer—9, lower protective layer—10, first sensing unit—11, second sensing unit—12, differential operation circuit—13, wire under test—14. Detailed Implementation

[0022] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0023] The purpose of this application is to provide a current measurement device for a magnetoelectric-piezoelectric effect sensing array, which aims to solve the contradiction between the sensitivity of the magnetoelectric sensing unit and the miniaturization of the device, the ambiguity of the angle estimation of the sparse array, and the problem of temperature drift affecting the measurement accuracy.

[0024] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0025] In one exemplary embodiment, such as Figure 1 As shown, a current measurement device based on a magnetoelectric-piezoelectric effect sensing array 1 is provided, comprising: a sensing array 1, an array signal processing module 2, a temperature compensation module 3, and a current inversion module 4 connected in sequence; the sensing array 1 includes multiple sensing units based on the magnetoelectric-piezoelectric effect, and the sensing array 1 is a hybrid L-shaped coprime sparse array. The sensor array 1 is used to acquire the magnetic field signal of the conductor under test at multiple times and convert it into multiple voltage signals; The array signal processing module 2 is used to perform signal conditioning, analog-to-digital conversion, and digital signal processing on multiple voltage signals at multiple times to obtain magnetic field spatial distribution data; the digital signal processing is implemented using an improved multi-signal classification algorithm. Temperature compensation module 3 is used to acquire the ambient temperature of the conductor 14 under test, perform temperature compensation on the magnetic field spatial distribution data, and obtain the temperature-compensated magnetic field spatial distribution data. The current inversion module 4 is used to determine the current value of the conductor under test based on the temperature-compensated magnetic field spatial distribution data using the built-in current inversion model.

[0026] As an optional implementation method, such as Figure 2As shown, the sensing unit includes an upper protective layer 6, a magnetostrictive layer 7, an intermediate adhesive layer 8, a piezoelectric layer 9, and a lower protective layer 10 stacked from top to bottom. The magnetostrictive layer 7 is used to generate mechanical strain under the action of an alternating magnetic field when current flows through the conductor being tested and an alternating magnetic field is generated; The intermediate adhesive layer 8 is used to transfer the mechanical strain generated by the magnetostrictive layer 7 to the piezoelectric layer 9; The piezoelectric layer 9 is used to generate a polarization voltage under the received mechanical strain, and outputs a voltage signal that is proportional to the magnetic field strength of the alternating magnetic field.

[0027] Specifically, the sensing unit uses PZT / Metglas composite magnetoelectric-piezoelectric material, wherein the magnetostrictive layer 7 is made of Metglas amorphous alloy, with an intrinsic magnetic noise as low as 10. -13 The piezoelectric layer 9 is made of lead zirconate titanate (PZT) piezoelectric ceramic, which offers performance improvements of more than 5 times compared to traditional piezoelectric materials. The working principle is as follows: when current flows through the conductor being measured, an alternating magnetic field is generated. The magnetostrictive layer 7 undergoes mechanical strain under the influence of the magnetic field, and this strain is transmitted to the piezoelectric layer 9, causing it to generate a polarization voltage. This polarization voltage is proportional to the magnetic field strength (i.e., the measured current). The sensing unit utilizes a mechanical resonance design to construct an acoustic resonant cavity within the composite structure of the magnetostrictive layer 7 and the piezoelectric layer 9. The resonant frequency is designed to be near an integer multiple of the power frequency current frequency, and the quality factor Q value is greater than 10 in the resonant state. 4 The equivalent magnetic sensitivity can be improved by two orders of magnitude, enabling effective detection of weak magnetic field signals at the pT level.

[0028] As an optional implementation method, such as Figure 3 As shown, the hybrid L-type coprime sparse array includes a triangular electromagnetic vector sensor located at the origin, a first scalar quantum array arranged along the x-axis, and a second scalar quantum array arranged along the y-axis; both the first and second scalar quantum arrays are composed of multiple sensing units. The sensing units are arranged along the x-axis and y-axis with a spacing between them that is a multiple of coprime integers, for the purpose of synchronously acquiring multiple voltage signals in space.

[0029] Specifically, sensor array 1 employs a hybrid L-shaped coprime sparse array, combining two-dimensional spatial coverage with the advantages of sparse aperture expansion. Sensor array 1 includes a triangular electromagnetic vector sensor located at the origin and two scalar quantum arrays, a first scalar quantum array and a second scalar quantum array, placed along the x-axis and y-axis respectively. The virtual aperture is expanded by the spacing between the coprime sensing units, achieving an equivalent aperture expansion of 3.2 times while maintaining miniaturization. The dimensions of sensor array 1 are no greater than 6.5cm × 6.5cm × 7cm.

[0030] As an optional implementation, the array signal processing module 2 includes a signal conditioning circuit, an analog-to-digital conversion circuit, and a digital signal processing unit connected in sequence; Signal conditioning circuits are used to condition multiple voltage signals to obtain conditioned analog signals; signal conditioning includes amplification and filtering. Analog-to-digital converter circuits are used to convert multiple conditioned analog signals into multiple digital signals. The digital signal processing unit is used to process multiple digital signals at multiple times using an improved multi-signal classification algorithm to obtain magnetic field spatial distribution data.

[0031] Specifically, the signal conditioning circuit employs a charge amplifier and phase-locked loop (PLL) detection technology to reduce interface circuit noise to 0.3 nV / √Hz; the analog-to-digital conversion circuit uses a 24-bit high-precision analog-to-digital converter (ADC) with a data acquisition frequency of no less than 4 kHz; the digital signal processing unit runs an improved Multiple Signal Classification (MUSIC) algorithm, integrating tensor decomposition theory and virtual co-array reconstruction technology to solve the ambiguity problem in sparse array angle estimation, achieving multipath component separation and multipath effect suppression. The digital signal processing unit adopts a floating-point arithmetic architecture with a main frequency of no less than 400 MHz, a built-in matrix operation coprocessor, and a single spatial spectrum estimation time of no more than 50 ms, meeting real-time processing requirements.

[0032] As an optional implementation, an improved multi-signal classification algorithm is used to process multiple digital signals at multiple times to obtain magnetic field spatial distribution data, including: Construct an array signal model based on multiple digital signals at multiple times; Based on the array signal model, construct the array covariance matrix; Eigenvalue decomposition is performed on the array covariance matrix to obtain the signal subspace and noise subspace; Tensor decomposition is performed on the array covariance matrix to reconstruct the virtual comatrix covariance matrix; By utilizing the orthogonality between the noise subspace and the steering vector of sensor array 1, a spatial spectrum function is constructed based on the virtual covariance matrix, and the spectral peaks of the spatial spectrum function are searched to determine the estimated value of the direction of arrival of the conductor under test. Based on the coprime array configuration of sensor array 1, the direction of arrival (DOA) estimate is de-ambigued to eliminate false spectral peaks and obtain an unambiguous DOA estimate. Based on the direction of arrival estimation results, the spatial distribution data of the magnetic field is generated.

[0033] As an optional implementation, the temperature compensation module 3 includes a surface acoustic wave temperature sensor, an RF read / write unit, and a temperature-magnetic field decoupling unit; the surface acoustic wave temperature sensor and the RF read / write unit are wirelessly connected, and the RF read / write unit is electrically connected to the temperature-magnetic field decoupling algorithm unit. The radio frequency (RF) read / write unit is used to transmit RF query signals to the surface acoustic wave (SAW) temperature sensor and receive the reflected echo signals generated by the SAW temperature sensor based on the RF query signals. The temperature-magnetic field decoupling unit incorporates a temperature-magnetic field decoupling algorithm. This algorithm is used to determine the ambient temperature of the conductor under test based on the reflected echo signal, and to perform temperature compensation on the spatial distribution data of the magnetic field based on the ambient temperature, thereby obtaining the temperature-compensated spatial distribution data of the magnetic field.

[0034] Specifically, the surface acoustic wave (SAW) temperature sensor employs a single-port resonator differential measurement structure, integrating the temperature-sensitive unit and the environmental reference unit on a shared substrate. It utilizes the intrinsic temperature sensitivity of the AT-cut quartz substrate and the temperature response delay characteristics of the insulation layer to construct a frequency drift difference calculation model. Its frequency-temperature relationship is Δf / f0 = a0(T-T0) + a1(T-T0). 2 +a2(T-T0) 3 Where Δf is the frequency offset, f0 is the reference frequency, T is the current ambient temperature, a0 is the linear temperature coefficient, a1 and a2 are higher-order correction coefficients, and T0 is the reference temperature. A differential structure eliminates common-mode frequency drift, retaining only the difference signal proportional to temperature. A gradient composite magnetothermal isolation material and a three-dimensional isolated sensor structure are employed, combined with a vertically layered micro-electro-mechanical system (MEMS) thermal sensor / air bridge / MTJ magnetic sensor three-dimensional integration, controlling magnetic-temperature crosstalk to within 1%. The temperature-magnetic field decoupling algorithm adopts a data-mechanism fusion intelligent decoupling strategy: a baseline coupling relationship is constructed based on the COMSOL multiphysics model, and the measured residuals are learned using a Long Short-Term Memory (LSTM) network. The prediction accuracy for extreme conditions is improved through training with a physical constraint loss function. This module can achieve wireless passive temperature compensation over distances exceeding 3 meters, with a compensation accuracy of ±1%, and a temperature measurement range covering -40℃ to 120℃.

[0035] As an optional implementation method, the process of determining the current inversion model includes: Obtain the training dataset; the training dataset includes sample magnetic field spatial distribution data and corresponding true current values; Construct a deep learning inversion network; the deep learning inversion network consists of three convolutional neural networks and two fully connected layers connected in sequence; each convolutional neural network is constructed based on a 3×3 filter, a batch normalization layer, and a ReLU activation function; By using a loss function with embedded physical constraints, a deep learning inversion network is trained on a training dataset to obtain a current inversion model; the physical constraints include Ampere's circuital law constraints, current amplitude limits, and L2 regularization terms.

[0036] A demodulation system employing finite element modeling, differential anti-interference, and intelligent algorithm inversion is adopted. First, a three-dimensional electromagnetic model of the power line and its surrounding environment is constructed using COMSOL to simulate the magnetic field distribution generated by the current and external magnetic field interference. Then, multi-dimensional input features such as spatial gradient and frequency domain harmonics are automatically extracted from the magnetic field data. Finally, combining deep learning and prior physical knowledge, Ampere's circuital law, amplitude constraints, and L2 regularization terms are embedded into the objective function to address the nonlinearity, overfitting, and lack of physical plausibility issues in the inversion problem. The inversion network adopts a two-branch architecture of feature extraction and physical constraint regression. The feature extraction branch uses a three-layer convolutional neural network (CNN) to extract multi-scale features of the magnetic field spatial distribution, while the physical constraint regression branch embeds Ampere's circuital law into the loss function L1=L_data+λ1·L_Ampere+λ2·‖I‖ in the form of an equality constraint. 2 Where L_data is the data fitting loss, L_Ampere is the penalty term for violating Ampere's circuital law, and ||I|| 2 λ1 and λ2 are L2 regularization terms.

[0037] As an optional implementation, the sensor array 1 is provided with at least one differential anti-interference structure; The differential anti-interference structure includes two sensing units and a differential operation circuit; the two sensing units are symmetrically arranged along the direction perpendicular to the conductor being measured, and both sensing units are connected to the differential operation circuit. The differential operation circuit is used to subtract the voltage signals output by the two sensing units to obtain the differential voltage signal, and then output the differential voltage signal as one voltage signal of the sensing array.

[0038] Specifically, the sensor array employs a differential anti-interference sensing structure with 180° phase inversion modulation, such as... Figure 4As shown, any differential anti-interference structure includes a first sensing unit 11, a second sensing unit 12, and a differential operation circuit 13. The first sensing unit 11 and the second sensing unit 12 are symmetrically arranged along a direction perpendicular to the measured conductor 14 to construct a symmetrical magnetoelectric coupling system of opposite polarities, utilizing dual-channel reverse phase compensation to suppress nonlinear effects. The magnetostrictive layers 7 of the first sensing unit 11 and the second sensing unit 12 have opposite polarization directions, resulting in strains in opposite directions under the same magnetic field. This achieves zero response to external common-mode interference magnetic fields and a multiplied response to differential-mode measured magnetic field signals at the differential output. By adopting a symmetrical dual-sensor differential architecture, the external interference magnetic field manifests as a common-mode signal (equal amplitude and phase), while the magnetic field generated by the current in the measured conductor 14 is a differential-mode signal (amplitude difference). Common-mode interference is eliminated through differential operation, while simultaneously achieving a sensitivity multiplication effect on odd-order harmonic response characteristics. This differential structure can effectively suppress more than 99% of common-mode electromagnetic noise and mechanical interference.

[0039] As an optional implementation, the magnetoelectric-piezoelectric effect sensing array 1 current measurement device further includes: a main control and communication module 5; the main control and communication module 5 includes a microcontroller and a wireless communication unit connected in sequence, the microcontroller is connected to the current inversion module 4, and the wireless communication unit is wirelessly connected to a remote external device; The microcontroller is used to receive the current value; The wireless communication unit is used to send the current value to an external device.

[0040] Specifically, the wireless communication unit includes a Bluetooth 5.0 module (communication range of 40 meters) and / or a 4GCat.1 remote communication module. The device uses a high-energy-density lithium battery module power supply system, supporting a handheld terminal non-contact mobile inspection mode.

[0041] As an optional implementation, the magnetoelectric-piezoelectric effect sensing array 1 current measuring device further includes: a protective housing; The sensor array 1, array signal processing module 2, temperature compensation module 3, current inversion module 4, and main control and communication module 5 are all encapsulated in a protective housing; the protective housing has an IP65 or higher protection rating.

[0042] Specifically, the protective shell is made of high-strength engineering plastic with an IP65 or higher protection rating, enabling stable and reliable operation in environments with relative humidity of 10%–100% and temperatures ranging from -20°C to 50°C. The overall volume of the device does not exceed 5dm. 3 It is easy to carry and deploy on site.

[0043] Furthermore, this application also provides a miniaturized long-distance current measurement system for a magnetoelectric-piezoelectric effect sensor array, comprising: at least one magnetoelectric-piezoelectric effect sensor array current measurement device as a mobile inspection terminal; a handheld terminal for receiving, displaying, and analyzing measurement data on-site; and a back-end server for storing, displaying, and analyzing measurement data, supporting multi-device data aggregation and trend analysis. Further, for different voltage levels (10kV, 35kV, 110kV, 220kV, 500kV), this application provides differentiated array configuration designs: 10kV: linear coprime array; 35kV: ordinary L-type coprime array; 110kV: hybrid L-type coprime array; 220kV: multi-scale L-type coprime array; 500kV: multi-layer multi-scale L-type coprime array. Through gradient array design, cost optimization is achieved while ensuring measurement accuracy.

[0044] Furthermore, specific applications of the apparatus described in this application have been implemented.

[0045] 1. Miniaturized long-distance voltage and current measuring instrument with magnetoelectric-piezoelectric effect sensor array.

[0046] like Figure 5 and Figure 6 As shown, this embodiment provides a miniaturized long-distance voltage and current measuring instrument with a magnetoelectric-piezoelectric effect sensing array. Its dimensions are 180mm × 150mm × 160mm, and its overall volume is approximately 4.2dm. 3 Net weight 1.8kg.

[0047] Designed for long-distance, non-contact measurement, the device mainly includes a sensor array, an array signal processing module, a differential anti-interference structure, a wireless passive temperature compensation module, a current inversion module, a main control and communication module, and a protective housing.

[0048] The protective housing is injection molded from ABS engineering plastic to form a sealed cavity. The array signal processing module, current inversion module, main control and communication module, and power management circuit are fixed within this sealed cavity. The Bluetooth antenna and 4G antenna are located near the main control and communication module and electrically connected to it. The sensor array is arranged in the magnetic field-sensitive area on the upper part of the protective housing, and its output is electrically connected to the input of the array signal processing module via flexible ribbon cable or circuit board traces. The SAW sensor antenna of the wireless passive temperature compensation module is arranged on the side or top of the housing for receiving and transmitting radio frequency signals. The device is powered by a high-energy-density lithium battery module, with the battery compartment located at the bottom of the housing for easy replacement and charging.

[0049] In terms of signal connectivity, the sensor array converts the alternating magnetic field generated by the conductor current into multiple analog voltage signals, which are then sent to the array signal processing module. After amplification, filtering, and analog-to-digital conversion, the array signal processing module sends the digital sampled values ​​to the main control and communication module. The main control and communication module runs the array signal processing algorithm and the current inversion algorithm to calculate current, harmonics, temperature, and state parameters, and transmits these parameters to the handheld terminal or backend server via Bluetooth 5.0 or 4GCat.1 communication units. Regarding power supply connectivity, the lithium battery module, after being regulated by the power management circuit, forms a DC power supply bus, which supplies power to the sensor array, array signal processing module, wireless passive temperature compensation module, current inversion module, main control and communication module, and communication unit.

[0050] The magnetostrictive layer is made of Metglas 2605S3A amorphous alloy sheet, with a thickness of approximately 25 μm and a saturation magnetostriction coefficient of ≈27 × 10⁻⁶. -6 The intrinsic magnetic noise is as low as 10. -13 The temperature is on the order of T / √Hz, with a Curie temperature of approximately 340℃. The piezoelectric layer uses a Sr-Mg co-doped PZT-5H piezoelectric ceramic sheet with a thickness of approximately 0.3mm, a piezoelectric constant ≥600pC / N, an electromechanical coupling coefficient ≥0.75, a dielectric constant ≈3400, and a mechanical quality factor ≈65. The intermediate bonding layer is achieved using an Au-In gradient eutectic bonding process, with the bonding temperature controlled at 135℃-145℃, a bonding pressure of 0.8MPa-1.2MPa, a bonding time of 15min-20min, a bonding strength ≥18MPa, and a piezoelectric performance retention rate ≥95%. This bonding process achieves efficient stress transfer between the magnetostrictive layer and the piezoelectric layer through the Au-In gradient alloy transition layer, with an interfacial strain transfer efficiency ≥92%, while avoiding the creep aging problem of traditional epoxy resin bonding layers. When current flows through the conductor under test, the resulting alternating magnetic field causes magnetostrictive strain in the magnetostrictive layer. This strain is transmitted to the piezoelectric layer through the intermediate adhesive layer, thereby generating a voltage signal in the piezoelectric layer that is proportional to the magnetic field strength. The upper and lower protective layers are made of polyimide film, approximately 50 μm thick, serving as insulation and mechanical support, while ensuring that the external magnetic field can penetrate to the magnetostrictive layer without attenuation. Silver electrodes, approximately 10 μm thick, are deposited on the upper and lower surfaces of the piezoelectric layer using a coating process to collect piezoelectric charges and output voltage signals.

[0051] The sensing array adopts a hybrid L-shaped coprime sparse array architecture. A triangular electromagnetic vector sensor is set at the origin of the array, which can simultaneously detect the x, y, and z components of the magnetic field. A first scalar quantum array is set along the x-axis, containing N1 sensing units with a spacing of d1 between the sensing units. A second scalar quantum array (23) is set along the y-axis, containing N2 sensing units with a spacing of d2 between the sensing units. Wherein, N1 and N2 are coprime integers, d1=N2·λ / 2, d2=N1·λ / 2, and λ is the operating wavelength. Through the coprime sensing unit spacing design, a virtual co-array can be constructed to achieve an equivalent aperture expansion of 3.2 times. In this embodiment, N1=5, N2=6, the length of the x-axis subarray is about 6cm, the length of the y-axis subarray is about 6.5cm, and the overall array size is about 6.5cm×6.5cm×7cm.

[0052] The differential anti-interference structure employs a symmetrical dual-unit differential design. The first and second sensing units are symmetrically arranged along a direction perpendicular to the conductor under test, with a spacing of Δd. The magnetic field generated by the conductor under test manifests as differential-mode voltage signals with different amplitudes at the two sensing units, while environmental interference magnetic fields (such as the Earth's magnetic field or distant interference sources) manifest as approximately equal-amplitude and in-phase common-mode signals. Let the voltage signal output by the first sensing unit be V1(t) = V dm (t)+V cm The voltage signal output by the second sensing unit is V2(t) = -V(t). dm (t)+V cm (t), where V dm (t) represents the differential mode signal component, V cm If (t) is the common-mode signal component, then the differential-mode voltage signal V diff (t)=V1(t)-V dm (t)=2V dm (t) The common-mode signal is completely canceled, and the differential-mode signal sensitivity is doubled. By subtracting the output signals of the two sensing units through a differential operation circuit, common-mode interference can be effectively canceled while retaining the differential-mode signal. This differential structure can suppress more than 99% of common-mode electromagnetic noise and mechanical vibration interference, while improving the odd-order harmonic response sensitivity by about 2 times.

[0053] The array signal processing module, which supports signal conditioning and acquisition for array signal processing, can be implemented using existing low-noise amplification, filtering, and analog-to-digital conversion circuits. It includes a multi-channel low-noise charge amplifier, a filtering circuit, a 24-bit high-precision analog-to-digital converter, and a digital signal processor (DSP) connected in sequence. The multiple weak analog voltage signals output from the sensor array first enter the low-noise charge amplifier to improve the signal-to-noise ratio, with an equivalent input noise voltage density ≤0.3nV / √Hz; then, they enter the filtering circuit for power frequency bandpass or anti-aliasing filtering; finally, they are sampled by the 24-bit ADC at a sampling frequency of not less than 4kHz / channel; in this embodiment, the sampling frequency can be set to 8kHz. The digital sampled values ​​output by the ADC are sent to the DSP via SPI or a parallel data bus. The DSP runs an improved MUSIC algorithm and a chaotic genetic algorithm to complete direction-of-arrival estimation and multipath suppression.

[0054] The specific steps of array signal processing are as follows: Step S1: Multi-channel synchronous sampling to acquire the output voltage signal of each sensing unit; Step S2: Signal preprocessing, including filtering, amplification, and analog-to-digital conversion; Step S3: Construct the array covariance matrix and perform eigenvalue decomposition; Step S4: Construct a virtual co-array using tensor decomposition theory to expand the array aperture; Step S5: Run the improved MUSIC algorithm to perform spatial spectrum estimation and obtain the direction of the magnetic field source (i.e., the estimated direction of arrival of the measured conductor). The specific process of this step is as follows: Step S51: Construct the array signal model. Let the nth output vector received by the array (i.e., the multi-channel digital signal at the nth moment) be X[n]=A(θ)·s[n]+N[n], where A(θ) is the array steering matrix, s[n] is the nth K×1 signal vector, N[n] is the nth M×1 noise vector, θ is the signal direction of arrival, K is the number of signal sources, and M is the total number of sensing units; Step S52: Estimate the array covariance matrix. According to the formula R̂=(1 / M)·Σ{X[n]·X H [n]} Construct an M×M dimensional array covariance matrix R̂, where H represents the conjugate transpose; Step S53: Perform eigenvalue decomposition on the covariance matrix. Decompose R̂ into R̂ = Us·Σs·Us H +Un·Σn·Un H Where Us is the signal subspace, consisting of eigenvectors corresponding to K large eigenvalues; Un is the noise subspace, consisting of eigenvectors corresponding to MK small eigenvalues. The number of signal sources K is estimated using the AIC criterion or the MDL criterion. Step S54: Construct a virtual co-array based on tensor decomposition theory. Expand the covariance matrix R̂ into a third-order tensor T along different dimensions, and use Tucker decomposition to obtain the core tensor G and factor matrices U1, U2, and U3. Reconstruct the virtual co-array covariance matrix Rv through Khatri-Rao product, thereby expanding the virtual aperture to the order of O(N1·N2), thus significantly expanding the array aperture without increasing the number of physical sensing units. Step S55: Construct the improved MUSIC spatial spectrum function. Utilizing the orthogonality between the noise subspace Un and the guiding vector a(θ), construct the spatial spectrum function P_MUSIC(θ) = 1 / [a H (θ)·Un·Un H ·a(θ)], where a(θ) is the array's steering vector at direction θ. When θ equals the true signal direction, a(θ) is orthogonal to Un, the denominator approaches zero, and a spectral peak appears in P_MUSIC(θ); Step S56: Perform spatial spectral peak search. Discretize P_MUSIC(θ) in the range of [-90°, 90°] with a step size of 0.1°, and extract the angles θ1, θ2, ..., θK corresponding to the first K spectral peaks, which are the estimated values ​​of the direction of arrival of the magnetic field source. At the same time, calculate the 3dB bandwidth of each spectral peak as the uncertainty of the angle estimation; Step S57: Sparse Array Deambiguity Processing. For the inherent ambiguity in angle estimation of coprime arrays, a difference joint array manifold is constructed using the coprime property. By searching for the unique solution in the virtual sensing unit difference set {d1·m1-d2·m2|0≤m1≤N1, 0≤m2≤N2} that matches the candidate Direction of Arrival (DOA), spurious spectral peaks are eliminated, resulting in an unambiguous DOA estimation. Simultaneously, by combining the higher-order statistical information obtained from tensor decomposition, a direction-finding accuracy of 0.8° can still be maintained under low signal-to-noise ratio (<5dB) conditions.

[0055] Step S6: Optimize the array configuration using an improved chaotic genetic algorithm and iteratively update the array parameters; Step S7: Output the direction of arrival estimation results and spatial spectrum data.

[0056] The optimization process for the improved chaotic genetic algorithm includes: Step S61: Parameter initialization, setting parameters such as population size, crossover probability, and mutation probability; Step S62: Generate an initial population based on the Logistic chaotic mapping to ensure that the initial solutions are uniformly distributed in the solution space; Step S63: Calculate the fitness value of each individual. The fitness function takes into account the main lobe width, side lobe level, and number of sensing units. Step S64: Employ an elite tournament selection strategy to prioritize individuals with high fitness for the next generation; Step S65: Perform point-to-point crossover operation, and perform adaptive crossover based on the neighborhood topology; Step S66: Perform random addition and subtraction correction mutation operations, and control the mutation intensity through chaotic parameters; Step S67: Determine whether the termination condition (maximum number of iterations or convergence threshold) has been met. If not, return to step S63 to continue iterating. Step S68: Output the optimal array configuration parameters.

[0057] In the temperature compensation module, the SAW temperature sensor employs a single-port resonator differential structure, integrating the temperature sensing unit and reference unit onto the same AT-cut quartz substrate. The substrate dimensions are approximately 3mm × 3mm × 0.5mm, the operating frequency is approximately 433MHz, and the quality factor Q≥8000. The temperature sensing unit has approximately 50 interdigital transducer (IDT) cycles and an acoustic aperture of approximately 200μm. The reference unit is covered by a thermal insulation layer (SiO2 thin film, approximately 2μm thick), exhibiting a temperature response delay approximately 10 times that of S1, eliminating common-mode frequency drift through differential calculation. The RF read / write unit transmits RF query signals to the SAW sensor via an antenna and receives reflected echo signals, calculating the temperature value by measuring changes in the resonant frequency. The temperature measurement range covers -40℃ to 120℃, with a temperature compensation accuracy of ±1% at a distance of 3m.

[0058] The temperature-magnetic field decoupling algorithm employs a data-mechanism fusion intelligent decoupling strategy: it constructs a baseline coupling relationship based on the COMSOL multiphysics model, combines it with an LSTM neural network to learn the measured residuals, and improves the prediction accuracy for extreme conditions through training with a physical constraint loss function. The LSTM model input features include temperature measurements, historical temperature sequences (sliding window length 64), magnetic field strength, and ambient humidity, and the output is the temperature-compensated magnetic field correction value. The LSTM network uses a three-layer stacked structure with hidden layer dimensions of 128, 64, and 32 respectively. A Dropout strategy (dropout rate 0.2) is used to prevent overfitting. Training employs the Adam optimizer with a learning rate of 0.001, a batch size of 32, and 200 training epochs. The physical constraint loss function is L2 = L_MSE + o·L_phys, where L_MSE is the mean squared error between the predicted and measured values, L_phys is the penalty term for the prediction bias of the COMSOL model, and o = 0.05 is the physical constraint weight coefficient.

[0059] The current inversion module receives magnetic field spatial distribution data from the array signal processing module and obtains the measured current value through a regularized framework embedded with physical constraints. Its workflow includes: Step 1: Establish a COMSOL 3D electromagnetic simulation model to simulate the magnetic field distribution under different currents, distances, and environments; Step 2: Generate a training dataset, including samples of the spatial distribution of the magnetic field and the corresponding true values ​​of the current; Step 3: Construct a deep learning inversion network. Three convolutional neural networks (CNNs) are used to extract spatial features, combined with two fully connected layers for regression prediction. Each CNN is built based on 3×3 filters, batch normalization layers, and the ReLU activation function. The first CNN uses 32 3×3 filters in each convolutional layer, the second uses 64, and the third uses 128. The fully connected layers have dimensions of 256 and 1, respectively. The final output is the effective current value I_rms. The total number of network parameters is approximately 50,000, and the inference time is less than 5ms. Step 4: Embed physical constraint terms into the loss function, including Ampere's circuital law constraints, current amplitude limits, and L2 regularization terms; Step 5: Train the network using the training data and optimize the network parameters; Step 6: Input the actual measured spatial distribution data of the magnetic field into the trained network, and output the inverted current value.

[0060] Main Control and Communication Module: The main controller is an ESP32 series microcontroller (MCU) with a clock frequency of 240MHz. The MCU receives digital signals from the array signal processing module, calculates parameters such as current value, harmonics, and direction of arrival using built-in algorithms, and packages the data. The communication module includes a Bluetooth 5.0 chip and a 4GCat.1 module. Bluetooth 5.0 is used for on-site debugging, parameter configuration, and short-range data reading, with a communication distance of up to 40 meters; 4GCat.1 is used for remote reporting. The device is powered by a 10000mAh high-energy-density lithium battery, which can operate continuously for more than 8 hours on a single charge. The complete algorithm processing flow of the main control and communication module is as follows: (a) System Initialization and Self-Test: After power-on, the MCU performs system initialization, including loading calibration parameters (zero point D0, sensitivity SB, temperature coefficient α), configuring the ADC sampling rate (4kHz), initializing the Bluetooth 5.0 and 4GCat.1 communication modules, performing sensor array health checks (checking whether the output of each sensing unit is within the normal range), and generating a self-test status report. If an abnormal sensing unit is detected, the MCU automatically marks it as offline and removes the data from that channel in subsequent processing; (b) Multi-channel synchronous data acquisition: The MCU synchronously reads the 24-bit ADC digital sample value D[n] output by the array signal processing module via the SPI bus. Parallel data from all sensing units are acquired in each sampling cycle to construct the array output vector X[n]. The sampling rate is not less than 4kHz, meeting the Nyquist sampling requirements for the 50Hz fundamental wave and its 25th harmonic. (c) Signal quality assessment: The MCU calculates the signal-to-noise ratio (SNR), signal amplitude stability, and spectral purity of each channel in real time. When the SNR of a certain channel is lower than the preset threshold (10dB), the MCU automatically reduces the weight coefficient of that channel in the covariance matrix to avoid low-quality data affecting the array signal processing results; (d) Array signal processing and direction-of-arrival estimation: The MCU runs the improved MUSIC algorithm to perform array covariance matrix construction, eigenvalue decomposition, virtual co-array reconstruction, and spatial spectrum estimation, outputting the direction-of-arrival θ and spatial spectrum data of the magnetic field source. This process utilizes the DSP coprocessor to accelerate matrix operations, with a single spectrum estimation taking no more than 50ms; (e) Current Inversion Calculation: The MCU calls the current inversion module, inputs the temperature-compensated magnetic field spatial distribution data and the direction-of-arrival estimation results into the pre-trained deep learning inversion network, and outputs the effective value of the conductor current, Irms. The specific calculation steps are as follows: 1. Zero-point correction and voltage conversion: The MCU acquires the zero-point sampling value D0 under no-current or known reference conditions, and performs zero-point subtraction on the nth ADC sampling value D[n] during operation. The voltage is converted to the equivalent voltage of the sensor array output terminal according to V_ME[n]=(D[n]-D0)×V_ref / (2^N×G_A), where V_ME[n] is the equivalent output voltage of the sensor array after deducting bias and gain, D[n] is the nth ADC digital sampling value; D0 is the zero-point or DC bias sampling value; V_ref is the ADC reference voltage; N is the effective number of bits of the ADC, N=24 in this embodiment; G_A is the equivalent voltage gain of the low-noise amplifier circuit and the front-end conditioning circuit. 2. Magnetic flux density conversion from voltage: For the i-th sensing unit, the magnetic flux density at its location can be calculated as B_i[n]=V_ME_i[n] / S_B, where B_i[n] is the magnetic flux density at the location of the i-th sensing unit; S_B is the equivalent magnetic field-voltage sensitivity of the magnetoelectric-piezoelectric sensing unit, in units of V / T, which can be obtained by calibration using a standard magnetic field source.

[0061] 3. Array signal processing and direction of arrival estimation: Based on multi-channel magnetic field sampling data, the array covariance matrix R̂ is constructed, according to the formula R̂=(1 / M)×ΣX[n]×X H[n] Calculation. Eigenvalue decomposition is performed on R̂ to separate the signal subspace and noise subspace. Spatial spectrum estimation is performed using the improved MUSIC algorithm to obtain the direction of arrival θ of the magnetic field source.

[0062] 4. Temperature Compensation: The MCU synchronously reads the temperature T output by the temperature sensor and uses the temperature coefficient to compensate for the magnetic field sensitivity, obtaining the compensated magnetic induction intensity: B. T，i[n] =B_i[n]×(1+α×(T-T0)) where, B T，i[n] T represents the temperature-compensated magnetic flux density; T is the currently measured ambient temperature; T0 is the reference temperature; and α is the combined temperature coefficient of the sensor array and front-end circuitry.

[0063] 5. Current Inversion: Based on the temperature-compensated spatial distribution data of the magnetic field and the estimated direction of arrival, a trained deep learning inversion network is used to perform current inversion. Ampere's circuital law constraint is embedded in the loss function to ensure that the inversion results satisfy physical laws. Finally, the effective value of the conductor current is obtained.

[0064] 6. Harmonic Calculation: The MCU performs Discrete Fourier Transform or Fast Fourier Transform on the instantaneous current values ​​within the same sampling window to obtain the fundamental frequency and the amplitudes of each harmonic. For the k-th harmonic, its peak amplitude can be calculated according to A... k =(2 / M)×|ΣI[n]×exp(-j×2π×k×n / M)|, with the effective value calculated according to I. k，rms =A_k / √2, where k is the harmonic order, k=1 represents the 50Hz fundamental frequency, k=2, 3, ... represent the second, third, and higher harmonics; j is the imaginary unit; A k I represents the amplitude of the k-th peak. k，rms This is the effective value of the kth harmonic.

[0065] (f) Harmonic Analysis: The MCU performs an FFT operation on the instantaneous current value sequence within the same sampling window to extract the fundamental frequency (50Hz) and the amplitudes of the 2nd to 25th harmonics, and calculates the total harmonic distortion rate (THD) = sqrt(Σ(I k，rms 2 ,k=2~25))I 1，rms ×100%, and generate harmonic spectrum data for power quality assessment; (g) Status parameter calculation: The MCU synchronously calculates and updates the following status parameters: (i) Battery power percentage, obtained by looking up the table based on the lithium battery open circuit voltage-remaining capacity calibration curve; (ii) Bluetooth / 4G signal strength RSSI value, read from the communication module register; (iii) Sensor array operating temperature, from the output of the SAW temperature sensor; (iv) Online status of each sensing unit, determined by periodic heartbeat detection; (v) Fault diagnosis code, automatically generated based on the built-in anomaly detection algorithm, including fault type codes such as sensor unit open circuit / short circuit, ADC overflow, communication interruption, and temperature over-limit. (h) Data Packaging and Communication: The MCU packages parameters such as RMS current value, harmonic data (amplitude of 25 frequency points), direction of arrival, temperature, battery level, signal strength, and fault diagnostic code according to a custom binary protocol. The data frame format is: frame header (0xAA55) + device ID (2 bytes) + timestamp (4 bytes) + RMS current value (4-byte floating point) + harmonic array (25×4 bytes) + direction of arrival (2 bytes) + temperature (2 bytes) + battery level (1 byte) + signal strength (1 byte) + fault code (2 bytes) + CRC16 checksum (2 bytes). The MCU sends one data frame to the handheld terminal via Bluetooth 5.0 every 200ms for real-time display, and reports one frame to the backend server via 4GCat.1 every 5s for long-term storage and trend analysis. When the detected current value exceeds the preset threshold or the fault code is non-zero, an alarm frame is immediately triggered and sent first.

[0066] Protective housing: Made of ABS engineering plastic injection molding with an anti-slip surface, the protection level reaches IP65, and it can operate stably in ambient temperatures of -20℃ to 50℃ and relative humidity of 10% to 100%. The front of the housing is equipped with an LCD display and operation buttons, which can display information such as current measurement value, temperature, power, and direction of arrival in real time.

[0067] 2. Miniaturized long-distance current measurement device with 500kV magnetoelectric-piezoelectric effect sensor array.

[0068] The difference between this embodiment and Embodiment 1 is that, for the long-distance measurement requirements of 500kV ultra-high voltage lines (safe distance approximately 8.5 meters), the sensing array adopts a multi-layer, multi-scale L-shaped coprime array configuration. The array consists of three layers: the first layer is a coarse-grained L-shaped coprime array (N1=5, N2=7) used to capture large-scale spatial information; the second layer is a medium-grained L-shaped coprime array (N1=4, N2=5) used for medium-resolution positioning; and the third layer is a fine-grained L-shaped coprime array (N1=3, N2=4) used for fine angle estimation. The outputs of the three layers are fused using a multi-scale fusion algorithm (Bayesian hierarchical inference) to further improve spatial resolution and direction-finding accuracy. Simultaneously, the number of array elements is increased to 18, the equivalent aperture expansion factor is increased to 4.5 times, the spatial resolution reaches 0.3°, and the direction-finding accuracy reaches 0.5° (at a 10dB signal-to-noise ratio). The overall dimensions of the device have increased slightly, approximately 200mm × 180mm × 180mm, with a total volume of approximately 6.5dm. 3 It weighs approximately 2.5kg, but still maintains a small and portable design.

[0069] Furthermore, to address the strong electromagnetic environment of 500kV lines, this embodiment adds an electromagnetic shielding layer to the differential anti-interference structure, using a high-permeability permalloy material to create the shielding shell, further enhancing anti-interference capabilities. Measurement accuracy reaches level 3, meeting the requirements for ultra-high voltage line inspection.

[0070] 3. Miniaturized long-distance current measurement device with 10kV magnetoelectric-piezoelectric effect sensor array.

[0071] The difference between this embodiment and Embodiment 1 is that, for short-range measurement scenarios (safe distance approximately 0.7 meters) of 10kV distribution network lines, the sensor array adopts a simplified linear coprime array configuration, with N1=3 and N2=4 array elements arranged along a single axis, reducing the total number of array elements to 7 (including the origin vector sensor). The array size is reduced to 4cm×4cm×5cm, and the overall device volume is reduced to 2.5dm. 3 The weight has been reduced to 1.2kg, making it more portable and flexible. At the same time, the signal processing algorithm has been simplified, reducing computational complexity and further lowering power consumption, increasing the battery life on a single charge to over 12 hours. Measurement accuracy reaches level 2.5, meeting the needs of power distribution network line inspection.

[0072] This application has the following beneficial effects. 1. Achieve true long-distance, non-contact current measurement: Through a high-sensitivity magnetoelectric-piezoelectric composite sensor array and sparse array aperture expansion technology, current measurement can be performed at a safe distance from high-voltage lines without contact with live conductors, ensuring the safety of maintenance personnel. The total device volume is no more than 5dm. 3 It is lightweight, easy to carry and deploy on site, and suitable for mobile inspection scenarios.

[0073] 2. Achieving ultra-high sensitivity detection of weak magnetic fields: Utilizing PZT / Metglas composite magnetoelectric-piezoelectric materials, the magnetoelectric coupling coefficient reaches 31.5 mV / cm·Oe, and the magnetic noise spectral density is as low as 0.9 × 10⁻⁶. -12 With a sensitivity of T / √Hz, reaching the pT level, it can effectively detect weak magnetic field signals after attenuation over long distances. This is achieved through mechanical resonance design and a Q value > 10. 4 The acoustic cavity achieves strain amplification, and the equivalent magnetic sensitivity in the resonant state is increased by two orders of magnitude.

[0074] 3. Achieving high-precision spatial orientation and multipath suppression: Employing a hybrid L-shaped coprime sparse array architecture, combined with an improved MUSIC algorithm and chaotic genetic algorithm optimization, a two-dimensional spatial resolution of 0.5° and a direction-finding accuracy of 0.8° (at a 10dB signal-to-noise ratio) are achieved, with a multipath suppression rate exceeding 90%. This effectively solves the problem of decreased measurement accuracy caused by multipath effects and environmental interference in complex electromagnetic environments.

[0075] 4. Superior anti-interference capability: Utilizing a 180° inverse-phase modulation differential anti-interference sensing structure, it can suppress over 99% of common-mode electromagnetic noise and mechanical interference, while simultaneously doubling the sensitivity of odd-order harmonic response. Combined with a physically constrained embedded current inversion algorithm, it ensures that measurement results conform to physical laws such as Ampere's circuital law, further enhancing measurement reliability in complex environments.

[0076] 5. Achieve high-precision temperature compensation over a wide temperature range: Integrating a wireless passive SAW temperature sensor with an intelligent decoupling algorithm, it achieves ±1% temperature compensation accuracy at a distance of 3m, with a temperature measurement range covering -40℃ to 120℃, effectively solving the impact of temperature drift of magnetoelectric materials on the accuracy of long-distance measurements.

[0077] 6. Multi-voltage level adaptation: Through a gradient array configuration design, it can adapt to the long-distance current measurement needs of various voltage levels from 10kV to 500kV. Specifically, the 35kV and 10kV voltage level devices achieve a measurement accuracy of level 2.5 or higher at safe distances under single-parameter laboratory current conditions, with a measurement range covering 10A~1000A; the 500kV, 220kV, and 110kV voltage level devices achieve level 3 or higher, with a measurement range covering 50A~2000A. The frequency response range of each voltage level device covers 50Hz fundamental to 1250Hz (25th harmonic), meeting the requirements of power quality monitoring.

[0078] 7. Intelligent and Self-Diagnostic Capabilities: The equipment features self-diagnostic functions for operational faults, automatically identifying abnormal nodes in the sensor array and generating diagnostic reports. Fault detection response time is no more than 100ms. Built-in overvoltage and overcurrent protection circuits dynamically monitor voltage and current parameters, automatically triggering circuit breaker protection or current limiting when exceeding safety thresholds to prevent damage to core components. The device's total power consumption does not exceed 2W (Bluetooth communication mode) and 3.5W in 4G communication mode. A single charge of the 10000mAh lithium battery provides over 8 hours of battery life (Bluetooth mode) and over 6 hours (4G mode).

[0079] Not all possible combinations of the technical features in the above embodiments are described; however, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0080] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the device and its core ideas. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A magnetoelectric-piezoelectric effect sensing array current measuring device, characterized in that, include: The sensor array, array signal processing module, temperature compensation module, and current inversion module are connected in sequence. The sensing array includes multiple sensing units based on the magnetoelectric-piezoelectric effect, and the sensing array is a hybrid L-shaped coprime sparse array; The sensor array is used to acquire the magnetic field signal of the conductor under test at multiple times and convert it into multiple voltage signals. The array signal processing module is used to perform signal conditioning, analog-to-digital conversion and digital signal processing on multiple voltage signals at multiple times to obtain magnetic field spatial distribution data. Digital signal processing is implemented using an improved multiple signal classification algorithm; The temperature compensation module is used to obtain the ambient temperature of the conductor under test, perform temperature compensation on the magnetic field spatial distribution data, and obtain the temperature-compensated magnetic field spatial distribution data. The current inversion module is used to determine the current value of the conductor under test based on the temperature-compensated magnetic field spatial distribution data using the built-in current inversion model.

2. The magnetoelectric-piezoelectric effect sensor array current measuring device according to claim 1, characterized in that, The sensing unit comprises an upper protective layer, a magnetostrictive layer, an intermediate adhesive layer, a piezoelectric layer, and a lower protective layer, which are stacked sequentially from top to bottom. The magnetostrictive layer is used to generate mechanical strain under the action of an alternating magnetic field when current flows through the conductor under test and an alternating magnetic field is generated. The intermediate adhesive layer is used to transfer the mechanical strain generated by the magnetostrictive layer to the piezoelectric layer; The piezoelectric layer is used to generate a polarization voltage under the received mechanical strain, and outputs a voltage signal that is proportional to the magnetic field strength of the alternating magnetic field.

3. The magnetoelectric-piezoelectric effect sensor array current measuring device according to claim 1, characterized in that, The hybrid L-type coprime sparse array includes a triangular electromagnetic vector sensor located at the origin, a first scalar quantum array arranged along the x-axis, and a second scalar quantum array arranged along the y-axis; both the first and second scalar quantum arrays are composed of multiple sensing units. Each sensing unit is arranged along the x-axis and y-axis with a spacing between sensing units that are coprime integers, for the purpose of synchronously acquiring the multiple voltage signals in space.

4. The magnetoelectric-piezoelectric effect sensor array current measuring device according to claim 1, characterized in that, The array signal processing module includes a signal conditioning circuit, an analog-to-digital conversion circuit, and a digital signal processing unit connected in sequence. The signal conditioning circuit is used to condition the multiple voltage signals to obtain conditioned analog signals; the signal conditioning includes amplification and filtering. The analog-to-digital conversion circuit is used to perform analog-to-digital conversion on multiple conditioned analog signals to obtain multiple digital signals. The digital signal processing unit is used to process multiple digital signals at multiple times using an improved multi-signal classification algorithm to obtain the magnetic field spatial distribution data.

5. The magnetoelectric-piezoelectric effect sensor array current measuring device according to claim 1, characterized in that, An improved multi-signal classification algorithm is used to process multiple digital signals at multiple time points to obtain the spatial distribution data of the magnetic field, including: Construct an array signal model based on multiple digital signals at multiple times; Based on the array signal model, construct the array covariance matrix; The array covariance matrix is ​​decomposed into eigenvalues ​​to obtain the signal subspace and the noise subspace. Tensor decomposition is performed on the array covariance matrix to reconstruct the virtual comatrix covariance matrix; By utilizing the orthogonality between the noise subspace and the steering vector of the sensing array, a spatial spectrum function is constructed based on the virtual covariance matrix, and a spectral peak search is performed on the spatial spectrum function to determine the estimated value of the direction of arrival of the tested conductor. Based on the coprime array configuration of the sensor array, the direction of arrival (DOA) estimate is de-ambigued to eliminate false spectral peaks and obtain an unambiguous DOA estimate. Based on the direction of arrival estimation results, the spatial distribution data of the magnetic field is generated.

6. The magnetoelectric-piezoelectric effect sensor array current measuring device according to claim 1, characterized in that, The temperature compensation module includes a surface acoustic wave temperature sensor, an RF read / write unit, and a temperature-magnetic field decoupling unit; the surface acoustic wave temperature sensor and the RF read / write unit are wirelessly connected, and the RF read / write unit is electrically connected to the temperature-magnetic field decoupling algorithm unit. The radio frequency read / write unit is used to transmit a radio frequency query signal to the surface acoustic wave temperature sensor and receive the reflected echo signal generated by the surface acoustic wave temperature sensor according to the radio frequency query signal. The temperature-magnetic field decoupling unit has a built-in temperature-magnetic field decoupling algorithm, which is used to determine the ambient temperature of the conductor under test based on the reflected echo signal, and to perform temperature compensation on the magnetic field spatial distribution data according to the ambient temperature, so as to obtain the temperature-compensated magnetic field spatial distribution data.

7. The magnetoelectric-piezoelectric effect sensor array current measuring device according to claim 1, characterized in that, The process of determining the current inversion model includes: Obtain the training dataset; the training dataset includes sample magnetic field spatial distribution data and corresponding true current values; A deep learning inversion network is constructed; the deep learning inversion network includes three convolutional neural networks and two fully connected layers connected in sequence; each convolutional neural network is constructed based on a 3×3 filter, a batch normalization layer and a ReLU activation function; The deep learning inversion network is trained on the training dataset using a loss function that incorporates physical constraints to obtain the current inversion model; the physical constraints include Ampere's circuital law constraints, current amplitude limits, and L2 regularization terms.

8. The magnetoelectric-piezoelectric effect sensor array current measuring device according to claim 1, characterized in that, The sensor array is provided with at least one differential anti-interference structure; The differential anti-interference structure includes two sensing units and a differential operation circuit; the two sensing units are symmetrically arranged along a direction perpendicular to the conductor being measured, and both sensing units are connected to the differential operation circuit. The differential operation circuit is used to subtract the voltage signals output by the two sensing units to obtain a differential voltage signal, and output the differential voltage signal as one voltage signal of the sensing array.

9. The magnetoelectric-piezoelectric effect sensor array current measuring device according to claim 1, characterized in that, The magnetoelectric-piezoelectric effect sensing array current measurement device further includes: a main control and communication module; the main control and communication module includes a microcontroller and a wireless communication unit connected in sequence, the microcontroller is connected to the current inversion module, and the wireless communication unit is wirelessly connected to a remote external device; The microcontroller is used to receive the current value; The wireless communication unit is used to send the current value to an external device.

10. The magnetoelectric-piezoelectric effect sensing array current measuring device according to claim 9, characterized in that, The magnetoelectric-piezoelectric effect sensing array current measurement device further includes: a protective housing; The sensor array, the array signal processing module, the temperature compensation module, the current inversion module, and the main control and communication module are all encapsulated within the protective housing; the protective housing has an IP65 or higher protection rating.

Images