Space charge detection device and method based on optical measurement

Abstract

The application relates to the technical field of power equipment testing, and particularly provides a spatial charge detection device and method based on optical measurement. The spatial charge detection device based on optical measurement comprises a sensing and collecting array which is uniformly distributed around a measured device in a circumferential direction, full-surrounding detection of spatial charge is realized, and the four optical detection technologies are cooperatively designed; through the complementary effects of Raman scattering, photoacoustic effect, LIBS and photoelectric effect, multi-dimensional detection of spatial charge is realized, and the detection limitations of single technology are compensated; the spatial charge detection method based on optical measurement adopts time-space synchronization-weight distribution-feature fusion, an adaptive weight distribution fusion algorithm based on a signal-to-noise ratio, and combines time-space synchronization calibration and Kalman filtering, so that the time-space inconsistency and random error problems of multi-source data are solved, and the precision of data fusion is improved.

Description

Technical Field

[0001] This application belongs to the field of power equipment testing technology, specifically relating to a space charge detection device and method based on optical measurement. Background Technology

[0002] As new power systems upgrade towards higher voltage and larger capacity, the insulation performance of key equipment such as high-voltage cables and transformers directly determines the operational safety of the system. The accumulation and migration of space charge within insulating materials can lead to electric field distortion, accelerate insulation aging, and induce breakdown faults. Therefore, accurate detection of its distribution characteristics has become a core requirement for equipment condition assessment.

[0003] Current space charge detection technologies are mainly divided into two categories: electrical and optical. Electrical technologies, represented by the pulse electroacoustic (PEA) method, require contact electrodes and high-voltage excitation, which can easily interfere with the original charge distribution and pose safety hazards. Optical technologies, such as Raman scattering and photoacoustic effects, have the potential for non-contact detection, but each technology has inherent drawbacks: Raman signals are easily interfered with by fluorescence, photoacoustic detection is significantly affected by environmental noise, LIBS technology has low signal strength under normal pressure, and the photoelectric effect method has difficulty distinguishing charge polarity.

[0004] In field monitoring scenarios, equipment operates in complex environments with high voltage and strong electromagnetic interference, and requires continuous operation, which traditional technologies struggle to meet. While multi-sensor fusion has been applied in other fields, a mature solution is still lacking in space charge detection: existing fusion methods fail to address the spatiotemporal synchronization and weight allocation issues of multi-source signals, and standardized computational models and testing procedures have not been established. Therefore, developing a multi-technology fusion detection device with clear testing principles, precise calculation methods, and standardized implementation steps has become crucial to overcoming industry bottlenecks. Summary of the Invention

[0005] This application addresses the limitations of existing optical measurement-based space charge detection technologies. Single optical techniques can only acquire one dimension of charge characteristics (e.g., Raman scattering for molecular correlation, photoacoustic scattering for positional information), failing to achieve simultaneous multi-parameter detection. Furthermore, LIBS technology suffers from poor quantitative accuracy due to plasma temperature fluctuations and requires complex environmental control. This application provides an optical measurement-based space charge detection device and method. The synergistic design of four optical detection technologies—Raman scattering, photoacoustic effect, LIBS, and photoelectric effect—integrates them for space charge detection, constructing a multi-physics-field synergistic detection system. This enables multi-dimensional space charge detection, overcoming the limitations of single-technology detection.

[0006] First, this application provides a space charge detection device based on optical measurement, including a laser emission module, a sensor acquisition array, a signal conditioning module, a data acquisition unit, and a signal processing unit;

[0007] The laser emitting module is connected to the sensor acquisition array via an optical fiber beam splitter, which transmits the laser emitted by the laser emitting module to the sensor acquisition array.

[0008] The sensing array consists of four types of sensors: a Raman scattering sensor, a photoacoustic sensor, a LIBS spectral sensor, and a photoelectric effect sensor. These sensors sequentially acquire four raw signals: Raman signal, photoacoustic signal, LIBS signal, and photoelectric signal.

[0009] The signal conditioning module is connected to the sensor acquisition array via a cable and is used to receive four raw signals transmitted from four types of sensors and modulate the four raw signals output by the four types of sensors respectively.

[0010] The data acquisition unit is connected to the signal modulation module via a cable and is used to synchronously acquire the four original signals modulated by the signal conditioning module.

[0011] The signal processing unit is connected to the data acquisition unit via a cable and is used to perform three-level data fusion operations on the four raw signals acquired by the data acquisition unit, and to construct a three-dimensional distribution cloud map and dynamic change curve of space charge.

[0012] In some embodiments, the laser emitting module includes a tunable laser, a collimating lens group, and an optical fiber coupler. The laser generated by the laser is collimated and then connected to an optical fiber beam splitter through the optical fiber coupler.

[0013] In some embodiments, the fiber optic beam splitter is a 1×4 fiber optic beam splitter with a splitting ratio error ≤ ±2%, which evenly distributes the laser emitted by the laser emitting module into four paths, which are then transmitted to each sensor in the sensing and acquisition array.

[0014] In some embodiments, the signal conditioning module includes a preamplifier, a low-pass filter, a high-pass filter, and a gain adjustment circuit, which amplify, filter, and reduce noise for the four original signals respectively, eliminating environmental interference signals.

[0015] In some embodiments, the signal processing unit includes an embedded processor that integrates a signal analysis module and a data fusion module. The signal analysis module performs spatiotemporal synchronization and signal feature extraction, while the data fusion module performs weight allocation and feature fusion. The signal processing unit, through the collaboration of the signal analysis module and the data fusion module, ultimately generates a three-dimensional distribution cloud map and dynamic change curve information of space charge.

[0016] In some embodiments, the system also includes a display and storage module for displaying data such as space charge distribution cloud maps and dynamic change curves in real time, and supporting local storage and USB export of historical data; it also includes a common power module for supplying power to each functional module.

[0017] Secondly, this application provides a space charge detection method based on optical measurement, implemented by the aforementioned space charge detection device based on optical measurement, specifically including the following steps:

[0018] Step 1, Equipment Deployment: Fix the sensor array around the device under test, connect the modules in the space charge detection device through optical fiber and cable, and start the power supply module to supply power to each module normally.

[0019] Step 2, parameter configuration: Set the laser parameters of the laser emission module, the signal conditioning parameters of the signal conditioning module, and the acquisition parameters of the data acquisition unit through the display module;

[0020] Step 3, signal acquisition: Start the laser emission module to output four laser beams of the same source to irradiate the insulation layer of the device under test. At the same time, start the data acquisition unit to synchronously acquire the four raw signal data output by the four sensors.

[0021] Step 4, data processing, employing a three-level data fusion operation: spatiotemporal synchronization, weight allocation, and feature fusion.

[0022] Spatiotemporal synchronization calibration uses the laser emission pulse as the time reference, timestamps the four original signals, and establishes a unified three-dimensional coordinate system to obtain four calibration signal data.

[0023] Feature extraction is performed on the four calibration signal data after spatiotemporal synchronization calibration, and the initial value of the space charge density of each sensor is calculated based on the information features obtained from the feature extraction step.

[0024] The signal-to-noise ratio of the calibration signal of each sensor is calculated, and the weighting coefficient is dynamically calculated based on the signal-to-noise ratio of the calibration signal of each sensor. Then, the optimal charge density estimate is obtained by performing a fusion operation based on the initial value of the space charge density of each sensor and the weighting coefficient.

[0025] Finally, based on the optimal charge density estimate, a volume data visualization algorithm is used to construct a three-dimensional space charge distribution model, and to draw distribution cloud maps and dynamic change curves.

[0026] Step 5, Result Output: The display module outputs a three-dimensional cloud map of space charge and a dynamic change curve.

[0027] In some embodiments, feature extraction is performed on the spatiotemporally synchronized calibration data, that is, extracting the information features corresponding to each calibration signal at each spatial point in a unified three-dimensional coordinate system, including: extracting the frequency shift Δ from the Raman signal. v The light intensity I is extracted using Raman spectroscopy; the amplitude A and audio frequency f are extracted from the photoacoustic signal; the characteristic peak intensity and half-peak width Δλ are extracted from the LIBS signal; the photoelectron flux density J is extracted from the photoelectric signal, and then the initial value of the space charge density of each sensor is calculated based on the information features obtained from each sensor in the feature extraction step.

[0028] In some embodiments, the weight allocation employs an adaptive weight allocation strategy, dynamically calculating the weight coefficients of each sensor based on the signal-to-noise ratio of the calibration signals of each sensor:

[0029] Where i = 1, 2, 3, 4, corresponding to the four types of sensors respectively. Let be the signal-to-noise ratio of the calibration signal of the i-th sensor, in dB, and let the weighting coefficients satisfy . .

[0030] In some embodiments, The calculation method is as follows: for the calibration signal output by the i-th sensor, extract the peak signal intensity S of the signal segment and the standard deviation of the noise intensity of the noise segment. Then, based on the core signal strength peak value S and the noise strength standard deviation... Calculate using the following formula: In the formula, SNR is the signal-to-noise ratio.

[0031] In some embodiments, the fusion operation employs a fusion algorithm combining weighted summation and Kalman filtering. Based on the initial value of the space charge density of each sensor and the weighting coefficients, the charge density of the same spatial point is weighted and fused using the following formula to obtain an estimated value of the space charge density. : ;

[0032] in, This is the initial value of the space charge density of the i-th type of sensor. Q It is the current point in the three-dimensional coordinate system. These are the weighting coefficients for sensor type i.

[0033] Then, random errors are eliminated by Kalman filtering to obtain the optimal space charge density estimate for the current space point. : ;

[0034] Wherein, K( Q Z( ) is the Kalman gain, Z( Q) represents the initial value of the space charge density detected by the sensor with the highest signal-to-noise ratio at the current spatial point.

[0035] In some embodiments, the Kalman gain K( Q The calculation process is as follows: ;

[0036] Where: p(Q|Q-1) is the prediction error covariance of the current spatial point, and H is the observation matrix. R is the transpose of the observation matrix, R is the measurement noise variance, and p is the prediction error covariance. Q|Q -1) Calculate by reverse derivation using the following formula: ;

[0037] Where, p( Q|Q ) is the current spatial point Q The posterior error covariance.

[0038] In some embodiments, in step 1, when the sensor acquisition array is fixed around the device under test, the four types of sensors in the sensor acquisition array are installed at 90° intervals around the device under test in the circumferential direction; the detection end face of all sensors in the sensor acquisition array is at the same vertical distance from the device under test; the laser incident angle α is 45°, and the sensor receiving angle β is 135°.

[0039] In some embodiments, step 3 further includes monitoring the signal-to-noise ratio (SNR) of the four raw signals and determining whether the SNR of each raw signal exceeds a threshold. If the SNR of each raw signal exceeds the threshold, the acquisition continues. If the SNR of each raw signal does not exceed the threshold, the process returns to step 2 to reset the laser parameters or the signal conditioning parameters of the corresponding signals until the SNR of all four raw signals exceeds the threshold.

[0040] Compared with the prior art, the beneficial effects that this application can achieve are:

[0041] 1. Four optical techniques—Raman scattering, photoacoustic effect, LIBS, and photoelectric effect—are integrated for space charge detection, constructing a multi-physics collaborative detection system. An adaptive weight allocation fusion algorithm based on signal-to-noise ratio is designed, significantly improving detection performance. Compared to the traditional PEA method (error ±10%) and TSDC method (error ±15%), the detection accuracy is greatly improved; spatial resolution ≤20μm, temporal resolution ≤10ms, and space charge density detection range 10... -6 C / m³-10 - 3C / m³, detection error ≤±5%.

[0042] 2. The non-contact online detection architecture enables non-contact online monitoring without the need to attach electrodes or connect to a test circuit. It can be used to perform detection while the high-voltage equipment is running normally (voltage level ≤500kV), avoiding the impact of offline detection on the operation of the power system and eliminating the interference of contact detection on charge distribution.

[0043] 3. The adaptive weight allocation fusion algorithm based on signal-to-noise ratio, combined with spatiotemporal synchronization calibration and Kalman filtering, solves the problems of spatiotemporal inconsistency and random error of multi-source data, improves the accuracy of data fusion, and significantly enhances the device's anti-interference ability against electromagnetic interference, temperature fluctuations (-20℃-60℃), and mechanical vibration (≤0.5g). The stability of test data in industrial field is ≥95%.

[0044] 4. The device features a modular design, making it easy to operate and flexible to deploy. The testing device weighs ≤15kg, supports dual power supply from mains power and lithium battery, and has a deployment time of ≤30 minutes. It can be adapted to various equipment such as high-voltage cables and transformer windings with diameters of 50mm-300mm, making it suitable for a wide range of scenarios.

[0045] 5. It provides real-time three-dimensional space charge distribution data, which can provide early warning of insulation aging and breakdown risks, improve the accuracy of equipment fault diagnosis by ≥30%, extend the equipment maintenance cycle by 20%-30%, and reduce the operation and maintenance costs of the power system. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 A schematic diagram of the space charge detection device based on optical measurement in this application is shown;

[0048] Figure 2 This application shows a schematic diagram of the arrangement between the sensor array and the device under test.

[0049] Figure 3 A flowchart of the space charge detection method in this application is shown;

[0050] Figure 4 The flowchart of the three-level data fusion operation of spatiotemporal synchronization, weight allocation, and feature fusion in this application is shown.

[0051] Figure 5 The flowchart for calculating the initial value of the space charge density of each sensor in this application is shown. Detailed Implementation

[0052] The term "comprising" in this application specification is synonymous with "including," "containing," or "characterized in," and is inclusive or open-ended, and does not exclude additional undescribed elements or method steps.

[0053] It should be noted that similar reference numerals and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, the terms "first," "second," "third," etc., are used only for distinguishing descriptions and should not be construed as indicating or implying relative importance. The technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0054] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0055] The present application will now be described in detail with reference to the accompanying drawings and specific embodiments.

[0056] Example 1

[0057] This embodiment provides a space charge detection device based on optical measurement. It includes a laser emission module, a sensor array, a signal conditioning module, a data acquisition unit, and a signal processing unit, such as... Figure 1 As shown.

[0058] The laser emission module, as the core excitation source, includes a tunable laser, a collimating lens group, and an optical fiber coupler. The output wavelength of the tunable laser is 532nm-1064nm, and the power is continuously adjustable within 0-5W. The laser generated by the laser is collimated and then connected to the optical fiber beam splitter through the optical fiber coupler, which splits it into four co-source laser beams, which are used to excite four different optical effects.

[0059] The sensing array consists of a Raman scattering sensor, a photoacoustic sensor, a LIBS spectroscopy sensor, and a photoelectric effect sensor. It is circumferentially arranged outside the insulating layer of the device under test, fixed by a bracket, and maintained at a constant distance from the surface of the insulating layer. It sequentially acquires four signals: Raman signal, photoacoustic signal, LIBS signal, and photoelectric signal. This application achieves multi-dimensional detection of space charge through the complementary effects of Raman scattering, photoacoustic, LIBS, and photoelectric effects, constructing a multi-physics-field collaborative detection system and overcoming the detection bottleneck of single technologies.

[0060] The laser emitting module is connected to the sensor array via a fiber optic beam splitter, transmitting the laser emitted by the laser emitting module to the sensor array. A 1×4 fiber optic beam splitter is used to evenly distribute the laser emitted by the laser emitting module into four paths, which are then transmitted to each sensor in the sensor array. The beam splitting ratio error is ≤±2%, ensuring the consistency of the excitation light intensity across the four detection channels.

[0061] The signal conditioning module is connected to the sensor array via a cable. It receives four signals transmitted from four different sensors and modulates the four signals output from each sensor. The signal conditioning module includes a preamplifier, a low-pass filter, a high-pass filter, and a gain adjustment circuit. These components amplify, filter, and reduce noise in the four signals output from the four sensors to eliminate environmental interference. The preamplifier has a gain of 10³-10⁻¹⁰. 6 The cutoff frequencies of the low-pass and high-pass filters are adjustable and can be set as needed. Furthermore, the signal conditioning module integrates four independent channels, allowing for simultaneous independent parameter configuration of four sensor signals, such as four cutoff frequencies and four amplifier gains.

[0062] The data acquisition unit is connected to the signal modulation module via a cable and is used to synchronously acquire the four signals modulated by the signal conditioning module. Furthermore, the data acquisition unit employs a multi-channel synchronous data acquisition card with a sampling rate ≥1GS / s and a resolution of 16 bits, supporting synchronous acquisition of four signals to ensure data time consistency.

[0063] The signal processing unit is connected to the data acquisition unit via a cable and is used to perform three-level data fusion operations on the four signals acquired by the data acquisition unit to construct a three-dimensional distribution cloud map of space charge. Furthermore, the signal processing unit has a built-in embedded processor, such as an FPGA+ARM architecture, integrating a signal analysis module and a data fusion module to realize functions such as signal feature extraction, spatiotemporal synchronization calibration, weight allocation, feature fusion, and three-dimensional reconstruction. Further, the signal analysis module (FPGA) implements high-speed parallel computing to realize signal feature extraction and spatiotemporal synchronization calibration functions, while the data fusion module (ARM) implements logical operations to realize weight allocation and feature fusion functions. Through the collaboration of the signal analysis module and the data fusion module, the signal processing unit ultimately completes the three-dimensional reconstruction of space charge, generating information such as a space charge distribution cloud map and dynamic change curves.

[0064] It also includes a display and storage module, which uses an industrial-grade touch screen to display data such as space charge distribution cloud map and dynamic change curve in real time; and has a built-in large-capacity storage chip to support local storage and USB export of historical data.

[0065] It also includes a power supply module that supports AC220V mains power or DC24V lithium battery power, with an output power of ≥50W. It has overvoltage, overcurrent and short circuit protection functions, providing stable power supply for each module of the space charge detection device and adapting to diverse power supply needs on site.

[0066] It also includes mechanical auxiliary equipment, such as brackets for fixing sensor arrays to the device under test. The brackets are made of insulating materials, such as polytetrafluoroethylene (PTFE), and their structure can be adapted to the structure of the device under test. For example, if the device under test is a circular cable, the bracket can be designed as an arc-shaped bracket, and the sensor is fixed around the circular cable through the arc-shaped bracket. The radius of the bracket can be adjusted according to the diameter of the cable under test (e.g., 50mm-300mm) to adapt to different specifications of high-voltage equipment. The device under test can also be other shapes of equipment or devices that need to detect space charge. If the device under test has a polygonal structure, a polygonal bracket can also be used.

[0067] The space charge detection device of this application optimizes the modular design of the device, reduces dependence on environmental conditions, supports dual power supply modes of mains power / lithium battery, and can be quickly deployed in field detection scenarios of various equipment such as high-voltage cables and transformers. Moreover, during the detection process, there is no need to contact the device under test or change the operating status of the device. It can achieve real-time monitoring when the high-voltage equipment is running normally, thus meeting the needs of engineering sites.

[0068] Example 2

[0069] This second embodiment provides a space charge detection method based on optical measurement, implemented by the aforementioned space charge detection device based on optical measurement. Through the interaction between a laser and the space charge within an insulating material, four different physical effects are excited. The characteristic signals generated by each effect are detected by a sensing array, indirectly obtaining information such as the distribution, density, and dynamic changes of the space charge. Specifically, it includes the following steps:

[0070] Step 1, Equipment Deployment: Fix the sensor array around the device under test, connect the various modules in the space charge detection device through optical fibers and cables, and start the power supply module to provide normal power to each module.

[0071] First, the sensor array is fixed around the device under test. The sensor array is arranged in a circumferentially uniform manner, that is, the four types of sensors in the sensor array are installed at 90° intervals around the device under test. The detection end face of all sensors in the sensor array is at the same vertical distance from the device under test. The laser incident angle α is 45° and the sensor receiving angle β is 135°.

[0072] Taking a cylindrical high-voltage cable as an example, the tested equipment consists of four types of sensors fixed circumferentially around the cable insulation layer by a bracket. Figure 2 As shown, the Raman scattering sensor is positioned at 0° of the cable's circumference, the photoacoustic sensor at 90°, the LIBS spectral sensor at 180°, and the photoelectric effect sensor at 270°. These four sensors are installed at 0°, 90°, 180°, and 270° along the cable's cross-section, forming a 360° all-encompassing detection coverage. This avoids blind spots and ensures the spatial integrity of the detection data.

[0073] The key parameters of the sensor, such as angle and distance, were also specified: d in the diagram represents the vertical distance between the sensor and the insulation layer surface, with a value of 5mm ± 1mm; the vertical distance d between the detection end face of all sensors and the cable insulation layer surface is uniformly set to 5mm to ensure consistent interaction distance between the laser and the insulation layer, improving the comparability of detection data. α is the laser incident angle, i.e., the angle between the laser beam and the normal to the insulation layer surface, set to 45°; β is the sensor receiving angle, i.e., the angle between the sensor detection direction and the laser incident direction, set to 135°; this angle combination maximizes the reception of effective scattered and effect signals while reducing ambient stray light interference.

[0074] Then, the laser emitting module, fiber optic beam splitter, sensor acquisition array, signal conditioning module, data acquisition unit, signal processing unit, and power supply module are connected in sequence. After checking that the link is complete, the power supply module is started.

[0075] Step 2, parameter configuration: Set the laser parameters of the laser emission module, the signal conditioning parameters of the signal conditioning module, and the acquisition parameters of the data acquisition unit through the display module.

[0076] The laser parameters of the laser emission module include: output wavelength, pulse frequency, and power. The output wavelength is selected based on the characteristics of the insulating material. For example, if the insulating material is polyvinyl chloride, the output wavelength is 532nm; if the insulating material is cross-linked polyethylene, the output wavelength is 1064nm. The pulse frequency is adjustable from 1kHz to 10kHz, and the power is adjustable from 1W to 5W.

[0077] The signal conditioning parameters of the signal conditioning module include: preamplifier gain and filter cutoff frequency. The preamplifier gain ranges from 10³ to 10⁻¹⁰. 6 The appropriate cutoff frequency can be selected based on signal strength. The filter cutoff frequency is set according to the characteristics of the four sensors: 100MHz for Raman signals; 10kHz for photoacoustic signals; 500MHz for LIBS signals; and 1MHz for photoelectric signals.

[0078] The data acquisition unit's acquisition parameters include: sampling rate, acquisition duration, and storage interval. The sampling rate, i.e., the sampling frequency, is set to 1 GS / s. The acquisition duration for a single acquisition can be selected within the range of 1 s to 10 min, and the storage interval can be set within the range of 10 ms to 1 s.

[0079] Step 3, Signal Acquisition: Start the laser emission module to output four laser beams of the same origin to irradiate the insulation layer of the device under test. At the same time, start the data acquisition unit to synchronously acquire the four signals output by the four sensors.

[0080] When simultaneously acquiring four raw signals from four different sensors, the signal-to-noise ratio of each raw signal is also considered. Real-time monitoring is performed, where i = 1, 2, 3, 4, corresponding to four different sensors. Let be the signal-to-noise ratio (SNR) of the original signal from the i-th sensor; and determine the SNR of each original signal. Does it exceed the threshold? If the signal-to-noise ratio is... If the signal-to-noise ratio (SNR) exceeds the threshold, continue acquisition. If the SNR does not exceed the threshold, return to step 2 to reset the laser parameters or the corresponding signal conditioning parameters until the SNR of all four signals exceeds the threshold. For example, if the SNR threshold is 20dB, then... Is it ≥20dB? If <20dB, return to step 2 to adjust the amplifier gain or laser power; if ≥20dB, continue data acquisition. The values ​​of 'i' correspond to the angles at which the four types of sensors are installed on the cross-section of the device under test. When i=1, it corresponds to the sensor installed at 0° on the device under test; when i=2, 3, and 4, they correspond to the sensors installed at 90°, 180°, and 270° respectively. The installation angles and categories of each sensor are fixed and matched to ensure consistency between signal acquisition and parameter configuration.

[0081] in, The calculation method is as follows: for the raw signals output by the i types of sensors, extract the peak value S of the signal characteristic intensity of the signal segment and the standard deviation of the noise intensity of the noise segment. Then, based on the peak signal strength S and the standard deviation of noise intensity... Calculate using the following formula: (Formula 1), where SNR is the signal-to-noise ratio.

[0082] Furthermore, the peak signal strength S extracted here does not calculate the strength of every signal feature; each sensor only needs to calculate it once for its core signal feature that can characterize the signal-to-noise ratio. The specific logic of the core signal characteristics for each type of sensor is as follows: The core signal characteristic for a Raman sensor is light intensity. The core signal characteristic of a photoacoustic sensor is the amplitude A of the sound wave, while the core signal characteristic of a LIBS sensor is the intensity of the characteristic peak. The core signal characteristic of a photoelectric sensor is the photoelectron flux density J.

[0083] Taking a photoacoustic sensor as an example, the process is as follows: The photoacoustic sensor acquires signals and extracts the core signal characteristic amplitude A (the sound frequency f is an auxiliary parameter, not involved in charge density calculation, nor used for SNR); from the time series of amplitude A, the peak value S of the signal segment is extracted, that is, the maximum value of A; from the signal-free segment, such as the silent period before laser emission, the standard deviation of the noise intensity is calculated. ;

[0084] Substituting into the formula, the signal-to-noise ratio of the original signal output by the photoacoustic sensor is calculated as follows: Similarly, the signal-to-noise ratio of the raw signals from other sensors can be calculated.

[0085] In step 3, when acquiring the four raw signals, the signal-to-noise ratio (SNR) is calculated and judged in real time. This determines whether the raw signals currently detected by the four sensors are valid. If the SNR of a certain signal is below standard, it is an invalid signal and needs to be directly returned to the parameter configuration stage for adjustment. When the SNR meets the standard, it is a valid signal and can enter the subsequent data processing flow. By calculating and judging the SNR, invalid data from the four raw signals acquired by the four sensors is avoided from entering the subsequent process, ensuring the effectiveness and accuracy of space charge detection.

[0086] Step 4, data processing: First, a three-level data fusion operation is adopted, consisting of spatiotemporal synchronization, weight allocation, and feature fusion. This involves sequentially performing spatiotemporal synchronization calibration, feature extraction, weight allocation, and fusion operation to obtain the optimal space charge density estimate. Then, based on the optimal space charge density estimate, a volumetric data visualization algorithm is used to construct a three-dimensional space charge distribution model, drawing distribution cloud maps and dynamic change curves. The data processing flow is as follows: Figure 4 As shown.

[0087] Step 41, Spatiotemporal Synchronization Calibration: This includes time synchronization and spatial synchronization. The raw data acquired from the sensor array is transmitted to the signal processing unit. Using the laser emission pulse as the time reference, the four signals are timestamped to achieve time synchronization. Based on the circumferential position of the sensors, the data is transformed to a unified r-θ-z three-dimensional coordinate system to achieve spatial synchronization. Through spatiotemporal synchronization calibration, four calibration signal data are obtained in the same three-dimensional coordinate system, eliminating the time delay and spatial point deviation of the four raw signal data.

[0088] Furthermore, the r-θ-z three-dimensional coordinate system is a cylindrical coordinate system used to locate spatial points within the insulation layer of the device under test (such as a high-voltage cable). The definitions and units of each parameter are as follows: r is the radial distance, in millimeters (mm) or meters (m), representing the distance from a spatial point to the central axis of the device; θ is the circumferential angle, in degrees (°) or radians (rad), representing the circumferential position of a spatial point on the cross-section of the device (corresponding to the sensor installation angles of 0° / 90° / 180° / 270°); Z is the axial length, in millimeters (mm) or meters (m), representing the position of a spatial point along the length of the device under test. The space charge density ρ is the "attribute value" corresponding to each spatial point Q(r,θ,z)). Through spatiotemporal synchronous calibration, it is mapped to a unified r-θ-z three-dimensional coordinate system to obtain the space charge density after fusion of the four calibration signals corresponding to each spatial point Q(r,θ,z).

[0089] The space charge density ρ is the charge density at each space point. Q The "attribute value" corresponding to (r, θ, z) is: we first determine a specific spatial point within the insulation layer using the r-θ-z three-dimensional coordinate system.Q (For example, r=5mm, θ=0°, z=10), and then through detection and data fusion, the charge density ρ at this location is obtained. The final three-dimensional distribution cloud map uses "color depth / brightness" to represent the magnitude of the charge density (for example, the redder the color, the higher the charge density), and the r-θ-z three-dimensional coordinate system is the "spatial coordinate frame" of the spatial charge density cloud map.

[0090] Step 42, Feature Extraction and Initial Value Calculation: Information features are extracted from the four calibration signal data after spatiotemporal synchronization calibration. The frequency shift Δν and light intensity are extracted from the Raman signal. ; Extracting amplitude A and audio frequency f from photoacoustic signals; Extracting characteristic peak intensity from LIBS signals And half-peak width Δλ; photoelectron flux density J extracted from photoelectric signals.

[0091] Among them, the frequency shift Δν of the Raman signal reflects the space charge density and light intensity. The uniformity of density distribution is verified to assist in calculating the signal-to-noise ratio; the amplitude A of the photoacoustic signal reflects the space charge density, the sound frequency f reflects the spatial location of the space charge, and the sound wave propagation time is related to the depth of the space charge; the characteristic peak intensity of the LIBS signal... The half-width at half-maximum (WHM) Δλ reflects the space charge density, indicating the dynamic migration rate of the space charge. Plasma lifetime is related to charge migration. The photoelectron flux density J of the photoelectric signal directly reflects the charge polarity and density, and the amount of photoelectrons released is positively correlated with the charge polarity. By extracting multi-dimensional features, the limitations of single technologies in obtaining only "one-dimensional information" are avoided. For example, traditional Raman spectroscopy cannot distinguish charge polarity, and photoacoustic methods are difficult to accurately quantify density. This provides data support for the comprehensive detection of "space charge density, polarity, three-dimensional position, and dynamic changes."

[0092] Then, based on the information features obtained from each sensor in the feature extraction step, the initial value of the space charge density for each sensor is calculated: Initial value of the space charge density of the Raman sensor. Initial value of space charge density of photoacoustic sensor Initial value of space charge density of LIBS sensor Initial value of space charge density of photoelectric sensor ,like Figure 5 As shown.

[0093] (1) Initial value of space charge density of Raman sensor based on Raman scattering effect Calculations are performed. When a laser irradiates an insulating material, photons undergo inelastic collisions with the material molecules, producing Raman scattered light. When space charge exists within the material, the charge alters the polarization state and vibrational energy levels of the molecules, causing a frequency shift in the Raman scattered light. The frequency shift Δν is related to the space charge density. It is directly proportional. Therefore, based on the scattered light signal captured by the Raman scattering sensor, the frequency shift Δν and light intensity can be analyzed. It can calculate the space charge density distribution. Its core relationship is:

[0094] (Formula 2);

[0095] in, This is a proportionality coefficient, determined by the properties of the insulating material, and its unit is cm. -1 ·m² / C, This represents the inherent frequency shift when there is no charge, i.e., the Raman characteristic peak shift (baseline value) of the insulating material molecules when there is no space charge, expressed in cm. -1 The initial value of the space charge density of the Raman sensor can be calculated according to (Formula 2). :

[0096] Right now (Formula 3).

[0097] proportionality coefficient The calculation process is as follows: A standard calibration sample with a known space charge density is used, such as two identical insulating material blocks, and then two known space charge values ​​are pre-injected into each block. , The Raman frequency shifts of the two samples were measured respectively. , Substitute into the formula:

[0098] (Formula 4);

[0099] (Formula 5);

[0100] Solve by simultaneously solving equations (4) and (5):

[0101] (Formula 6).

[0102] Example: If .

[0103]

[0104] The calculation process is as follows: A standard insulating sample without charge, such as pure XLPE or oil paper, is placed in the detection device. The laser parameters are set as follows: wavelength λ = 1064 nm, power P = 2 W, integration time t = 100 ms. Raman spectra are acquired, and Gaussian fitting is performed on the characteristic peaks. For example, the CH vibration peak of XLPE is located at 2850 cm⁻¹. -1 At that time, that is :

[0105] (Formula 7);

[0106] in, Let A be the ordinate (Raman intensity) of the Raman spectrum as a function of the abscissa (laser frequency), where A is the peak amplitude. Where σ is the peak frequency and σ is the full width at half maximum (FWHM). Record As the reference frequency when there is no charge ,but , The incident laser frequency, expressed in Hz or .

[0107] (2) Initial value of space charge density of photoacoustic sensor based on photoacoustic effect Calculations are performed. A laser pulse irradiates the insulating layer; the material absorbs the laser energy and undergoes thermal expansion. Simultaneously, the space charge undergoes a minute displacement under the laser radiation pressure. Both factors together excite acoustic wave vibrations, and the acoustic wave amplitude A is related to the space charge density. Proportional. A photoacoustic sensor (microphone array) captures this sound wave signal, and by analyzing the amplitude A and frequency characteristics, the spatial distribution of the charge is calculated. The core relationship is:

[0108] (Formula 8);

[0109] in, The sound-to-electric conversion coefficient is expressed in Pa·m. 3 / (C·W), This represents the laser pulse power, measured in W.

[0110] laser pulse power Calculation method: Directly measure the laser output power using a power meter. If it is a pulsed laser, the duty cycle must be considered in the calculation. (Formula 9);

[0111] in, The average power is the reading from the power meter; η is the duty cycle. For example, if the pulse frequency f = 5 kHz and the pulse width τ = 10 ns, then η = f·τ = 5 × 10 -5 .

[0112] Sound-to-electric conversion coefficient Calculation process: using known space charge density Standard samples, at laser power Measuring the amplitude of sound waves .

[0113] Calculated according to (Formula 8) value: (Formula 10).

[0114] Space charge density of photoacoustic sensors Calculation process: The output voltage signal U of the photoacoustic sensor is converted into the acoustic wave amplitude A through the calibration curve (A = U / G, where G is the sensor gain, in Pa / V).

[0115] Substituting into (Equation 8) and performing inversion calculations yields the initial value of the space charge density:

[0116] (Formula 11).

[0117] (3) The space charge density of the LIBS sensor based on the LIBS effect Initial values ​​are calculated. A high-energy laser pulse is focused onto the surface of an insulating material, causing localized ionization and the formation of plasma. Space charge affects the degree of ionization and lifetime of the plasma, leading to changes in the intensity I and width Δλ of the characteristic peaks in the plasma emission spectrum. A LIBS spectral sensor acquires the spectral signal, and by analyzing the characteristic peak parameters, the space charge release process is assessed. The core relationship is:

[0118] (Formula 12);

[0119] in, The spectral response coefficient is expressed in counts·m. 4 / (C·J 2 ), The energy density of the laser pulse is expressed in J / m².

[0120] Laser pulse energy density The calculation is as follows: First, the laser pulse energy E is measured by an energy meter (in J), and the focused spot area S (in m²) is calculated. Then, the result is obtained by formula (13):

[0121] (Formula 13); where the focused spot area S is calculated as follows:

[0122] (Formula 14); where d is the beam diameter, measured by a beam analyzer, in meters.

[0123] Spectral response coefficient Calculation process: For a known space charge density Standard samples, at energy density Measurement of characteristic peak intensity The spectrometer's counting unit is "counts".

[0124] Substitute into the formula: (Formula 15);

[0125] Finally, the space charge density of the LIBS sensor is derived according to (Equation 11). Initial value:

[0126] (Formula 16)

[0127] (4) Initial value of space charge density of photoelectric sensor based on photoelectric effect Calculations are performed. Photoelectric effect: When a weak auxiliary electric field (Eauxiliary ≤ 1 kV / m, which does not affect equipment operation) is applied to the surface of the insulating layer, and laser irradiation is performed, the space charge absorbs photon energy and undergoes a transition, releasing photoelectrons. The number of photoelectrons N is related to the space charge density. They are directly proportional. Photoelectric effect sensors detect the intensity of photoelectron flow, thereby indirectly obtaining information about charge accumulation. The core relationship is:

[0128] (Formula 17)

[0129] in, Photovoltaic conversion factor, in photons -1 ·m 3 , Laser photon flux, measured in photons / s. The calculation process is as follows:

[0130] (Formula 18);

[0131] Where h is Planck's constant, in J·s, with a value of 6.626 × 10⁻⁶. -34 J·s, where c is the speed of light in a vacuum, and c = 3 × 10⁻⁶. 8 m / s. λ is the laser wavelength, in meters (m). The laser frequency is expressed in Hz; h This represents the energy of a single laser photon, measured in J.

[0132] Due to laser frequency And the laser wavelength λ satisfies (Formula 19); where c is the speed of light in a vacuum, therefore... Substitute ,get (Equation 20).

[0133] Photoelectric conversion coefficient Calculation process: For a known space charge density Standard samples, at photon flux Measuring the number of photoelectrons It can be counted using a photoelectric sensor. Substitute into formula (Formula 15) for calculation: (Formula 21).

[0134] The initial value of the space charge density of the photoelectric sensor can be calculated according to (Formula 17). for: (Equation 22).

[0135] Step 43, Weight Allocation Step: The weight allocation adopts an adaptive weight allocation strategy, based on the signal-to-noise ratio of the calibration signals of each sensor. Weighting coefficients for each sensor Dynamic calculations are performed, where i = 1, 2, 3, 4, corresponding to the four types of sensors respectively:

[0136] (Formula 23);

[0137] in, The signal-to-noise ratio (SNR) of the calibration signal for the i-th sensor is expressed in dB, with weighting coefficients. satisfy .

[0138] The signal-to-noise ratio (SNR) of the calibration signals from the four sensors is also calculated according to (Formula 1). The only difference is the SNR of the calibration signals from each sensor. During the calculation, the core signal characteristics representing the signal-to-noise ratio of each sensor are extracted in step 42 from the four calibration signal data after spatiotemporal synchronization calibration. Based on this, the signal-to-noise ratio of the calibration signals of each sensor is calculated. It is also mapped to a spatial point Q in a unified three-dimensional coordinate system. The weight coefficients of each sensor obtained are also mapped to a spatial point Q in a unified three-dimensional coordinate system. Finally, it ensures that the subsequent weighted fusion operation is performed for each spatial point Q, calling the signal-to-noise ratio of the corresponding sensor to calculate the weight coefficient, realizing the charge density fusion at that point, eliminating the temporal and spatial deviation of the detection results of each sensor, and improving the accuracy of the detection results.

[0139] Step 44: The fusion operation employs a fusion algorithm combining weighted summation and Kalman filtering. First, the space charge of each sensor is calculated based on the information features obtained from the feature extraction step. Then, the space charge density of the same spatial point is fused and calculated using (Formula 23) to obtain the spatial point... QSpace charge density estimate at [location] :

[0140] (Formula 24);

[0141] Where i = 1, 2, 3, 4 are fixed to correspond to the Raman sensor, photoacoustic sensor, LIBS sensor, and photoelectric sensor, respectively. This is the initial value of the space charge density of the i-th type of sensor. It is the weighting coefficient of the i-th sensor.

[0142] In a specific embodiment, (Formula 24) expands to (Formula 25):

[0143] When i=1, it is a Raman sensor. When i=2, it is a photoacoustic sensor. When i=3, it is a LIBS sensor. When i=4, it is a photoelectric sensor. According to (Formula 1), the signal-to-noise ratio of the Raman sensor calibration signal is: The signal-to-noise ratio of the photoacoustic sensor calibration signal is The signal-to-noise ratio of the LIBS sensor calibration signal is The signal-to-noise ratio of the photoelectric sensor calibration signal is Then, according to (Formula 23), the weighting coefficient of the Raman sensor is obtained. Weighting coefficients of photoacoustic sensors Weighting coefficients of LIBS sensors Weighting coefficients of photoelectric sensors Substituting the initial values ​​of the space charge density of each sensor and the weighting coefficients into (Equation 24) yields the corresponding spatial points. Q Estimated space charge density after fusion:

[0144] (Formula 25); where, Q It is the current spatial point in the three-dimensional coordinate system.

[0145] Then, random errors are eliminated by Kalman filtering to obtain the optimal space charge density estimate:

[0146] (Formula 26);

[0147] Wherein, K( Q Z( ) represents the Kalman gain. Q The value is the initial measured value of the space charge density detected by a certain sensor at the current spatial point. The single source data with the highest SNR among the four sensors is selected.

[0148] "Measured value" refers to the initial value of space charge density directly detected by the sensor with the highest signal-to-noise ratio at the current spatial point. Specifically: in multi-sensor fusion scenarios, "Z ( Q ")" represents the initial values ​​of the charge density from the four sensors at the current spatial point. In this process, the initial value with the highest signal-to-noise ratio and the strongest reliability is selected as the "measured value." For example, if the Raman sensor has the highest SNR, then... .

[0149] Kalman gain K( Q The calculation process is as follows:

[0150] (Formula 27);

[0151] Where: p( Q|Q -1) represents the prediction error covariance of the current spatial point, and H is the observation matrix. The transpose of the observation matrix, R is the measurement noise variance; the prediction error covariance p( Q|Q -1) Calculate by reverse derivation using the following formula: ;

[0152] Here, H is used to establish the mapping relationship between the "system state quantity" and the "actual observation value". In the space charge detection scenario of this application: the "state quantity" of the system is the space charge density ρ (i.e. the target quantity that we need to estimate); the "actual observation value" is Z(Q); since the "state quantity" and the "observation value" are the same physical quantity (both are charge density), the value of H is 1 (a unitless scalar).

[0153] Where p( Q|Q ) is the current spatial point Q The posterior error covariance represents the magnitude of the error in the estimated charge density of spatial point Q after Kalman filtering optimization. The smaller the value, the more reliable the estimation result.

[0154] p( Q|Q The initial spatial points (e.g., ...) represent the recursive computational complexity of the Kalman filter. The specific computational process is as follows: First, initial spatial points (e.g., ...) are given... Q The prediction error covariance p(1|0) of (1,1,1) is assigned an initial value, which is usually set according to experience, for example, p(1|0)=0.1, with the unit consistent with the square of the charge density; calculate the Kalman gain K(1) at this point;

[0155] Then, p(1|1) = (1 - K(1)·H)·p(1|0) (Formula 29);

[0156] p( of subsequent spatial point Q) Q|Q -1) Directly use the p( of the previous point)Q -1| Q -1) Assignment, i.e.:

[0157] p( Q|Q -1)= p( Q -1|Q-1)+Q sys Q sys The system noise variance is empirically taken as 10. -9 (C / m) 3 ) 2 Then recursively calculate K( Q ) and p( Q|Q (Unit: C / m) 3 .

[0158] The three-level data fusion operation proposed in this application, which combines spatiotemporal synchronization, weight allocation, and feature fusion, dynamically allocates weights based on the signal-to-noise ratio and combines Kalman filtering to eliminate random errors, thereby improving the accuracy and stability of the detection results.

[0159] Step 45, based on Construct a three-dimensional space charge distribution model and plot a three-dimensional space charge distribution cloud map and dynamic change curve.

[0160] based on Construct a three-dimensional space charge distribution model and plot a three-dimensional space charge distribution cloud map: that is, for the three-dimensional coordinates constructed in step 41, take all... Q The information features corresponding to the points are then calculated, and all of them are obtained. Q The optimal spatial density distribution estimate corresponding to a point is obtained based on the corresponding values ​​of all points in the three-dimensional coordinate system. Draw a three-dimensional cloud map of space charge.

[0161] The space charge dynamic change curve is a two-dimensional curve. Its core principle is to "select a characteristic dimension to show the trend of charge density change over time / space." There are two common ways to determine this value:

[0162] (1) Fixed spatial dimension that changes over time; this is the most common method of value selection. A fixed spatial point is selected. Q For example, at the radial midpoint r=5mm and circumferential 0° of the cable insulation layer, the optimal space charge density estimate in this area is continuously monitored. With "time" as the horizontal axis, "space charge density" Plot a curve with the vertical axis to show the dynamic change of space charge density at the same location over time (e.g., the process of charge accumulation / dissipation).

[0163] (2) Fixed time, varying with spatial dimension. A fixed point in time is selected, and along a certain spatial dimension, such as the radial direction r of the cable, the charge density at different locations is extracted; using "a certain dimension in the three-dimensional coordinate system" as the abscissa, such as the radial distance r, and "spatial charge density" as the metric. Plot a curve with the vertical axis as the ordinate to show the differences in charge density distribution at different spatial points at the same time.

[0164] Step 5: Result output. The display module displays the three-dimensional distribution cloud map and dynamic change curve of space charge in real time, and stores the processing results to the local storage module. It supports USB export or network transmission to the background management system. Furthermore, it can also display the maximum and average values ​​of space charge density.

[0165] After outputting the results, determine if any of the following conditions for continuing detection are met: ① The preset monitoring time / detection area has not been reached; ② The equipment is operating normally and there is a need for data supplementation; ③ A user's instruction to continue detection has been received. If the conditions are met, return to step 3 to reacquire the signal and continue executing steps 4 and 5; if the conditions are not met, end the space charge detection.

[0166] The space charge detection device based on optical measurement in this application features a modular design, weighs ≤15kg, supports dual power supply of mains power / lithium battery, has a deployment time of ≤30 minutes, and can be adapted to various devices such as high-voltage cables and transformer windings with diameters of 50mm-300mm, making it suitable for a wide range of scenarios.

[0167] The space charge detection method based on optical measurement in this application, the adaptive weight allocation fusion algorithm based on signal-to-noise ratio, and the combination of spatiotemporal synchronization calibration and Kalman filtering solve the problems of spatiotemporal inconsistency and random error of multi-source data, significantly improve the accuracy of data fusion, and significantly enhance the device's anti-interference ability against electromagnetic interference, temperature fluctuations (-20℃-60℃), and mechanical vibration (≤0.5g). The stability of test data in industrial field is ≥95%.

[0168] The space charge detection device and method based on optical measurement of this application can provide real-time three-dimensional distribution data of space charge, which can provide early warning of insulation aging and breakdown risks, improve the accuracy of equipment fault diagnosis by ≥30%, extend the equipment maintenance cycle by 20%-30%, and reduce the operation and maintenance costs of power systems.

[0169] Example 3

[0170] Taking the space charge detection of a 110kV cross-linked polyethylene (XLPE) high-voltage cable as an example:

[0171] Test object: 110kV XLPE high-voltage cable, insulation layer thickness 10mm, outer diameter 120mm.

[0172] Detection device: The space charge detection device described in this application.

[0173] Parameter settings: Select a 1064nm wavelength laser for the laser emission module, with a power of 3W and a pulse frequency of 5kHz; The sensor array is set to... Figure 2 The setup was adjusted with the bracket radius set to 60mm and the sensor distance from the insulation layer set to 5mm; the signal conditioning module amplifier gain was set to 10. 4 The filter cutoff frequency is configured according to each sensor type; the data acquisition unit has a sampling rate of 1GS / s and an acquisition time of 5min.

[0174] Auxiliary equipment: Laptop (for data export and analysis), AC220V mains power supply interface.

[0175] Testing process: After the device is deployed, online testing is carried out under normal cable conditions. The signal-to-noise ratio is stable at 25dB-30dB, which meets the testing requirements.

[0176] Test results: A three-dimensional distribution cloud map of the space charge inside the cable insulation layer was successfully obtained, revealing the maximum and average space charge density at a distance from the outer insulation layer.

[0177] Verification results: Compared with the offline PEA method test results, the accuracy of the results detected by the space detection device and detection method in this application is significantly improved.

[0178] Application value: Based on the test results, maintenance personnel took timely measures to strengthen local insulation, avoiding the risk of insulation breakdown caused by further charge accumulation and ensuring the safe operation of the cable.

[0179] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A space charge detection device based on optical measurement, characterized in that, It includes a laser emission module, a sensor acquisition array, a signal conditioning module, a data acquisition unit, and a signal processing unit; The laser emitting module is connected to the sensor acquisition array via an optical fiber beam splitter, which transmits the laser emitted by the laser emitting module to the sensor acquisition array. The sensing array consists of four types of sensors: a Raman scattering sensor, a photoacoustic sensor, a LIBS spectral sensor, and a photoelectric effect sensor. These sensors sequentially acquire four raw signals: Raman signal, photoacoustic signal, LIBS signal, and photoelectric signal. The signal conditioning module is connected to the sensor acquisition array via a cable and is used to receive four raw signals transmitted from four types of sensors and modulate the four raw signals output by the four types of sensors respectively. The data acquisition unit is connected to the signal modulation module via a cable and is used to synchronously acquire the four original signals modulated by the signal conditioning module. The signal processing unit is connected to the data acquisition unit via a cable. It is used to perform a three-level data fusion operation of spatiotemporal synchronization, weight allocation, and feature fusion on the four raw signals acquired by the data acquisition unit, and to construct a three-dimensional distribution cloud map and dynamic change curve of space charge. The three-level data fusion operation process of spatiotemporal synchronization-weight allocation-feature fusion includes: spatiotemporal synchronization calibration uses the laser emission pulse as the time reference, timestamps the four original signals, and establishes a unified three-dimensional coordinate system to obtain four calibration signal data. Feature extraction is performed on the four calibration signal data after spatiotemporal synchronization calibration, and the initial value of the space charge density of each sensor is calculated based on the information features obtained from the feature extraction step. The signal-to-noise ratio of the calibration signal of each sensor is calculated, and the weighting coefficient is dynamically calculated based on the signal-to-noise ratio of the calibration signal of each sensor. Then, the optimal charge density estimate is obtained by performing a fusion operation based on the initial value of the space charge density of each sensor and the weighting coefficient. Finally, based on the optimal charge density estimate, a volume data visualization algorithm is used to construct a three-dimensional space charge distribution model, and to draw distribution cloud maps and dynamic change curves.

2. The space charge detection device based on optical measurement according to claim 1, characterized in that, The laser emitting module includes a tunable laser, a collimating lens group, and an optical fiber coupler. The laser generated by the laser is collimated and then connected to an optical fiber beam splitter through the optical fiber coupler.

3. The space charge detection device based on optical measurement according to claim 1, characterized in that, The fiber optic beam splitter uses a 1×4 fiber optic beam splitter with a splitting ratio error of ≤±2%, which evenly distributes the laser emitted by the laser emission module into four paths, which are then transmitted to the sensors in the sensing and acquisition array.

4. The space charge detection device based on optical measurement according to claim 1, characterized in that, The signal conditioning module includes a preamplifier, a low-pass filter, a high-pass filter, and a gain adjustment circuit, which amplify, filter, and reduce noise for the four original signals to eliminate environmental interference signals.

5. The space charge detection device based on optical measurement according to claim 1, characterized in that, The signal processing unit includes an embedded processor, which integrates a signal analysis module and a data fusion module. The signal analysis module realizes spatiotemporal synchronization and signal feature extraction functions, while the data fusion module is used to realize weight allocation and feature fusion functions. Through the collaboration of the signal module and the data fusion module, the signal processing unit ultimately generates a three-dimensional distribution cloud map of space charge and dynamic change curve information.

6. The space charge detection device based on optical measurement according to claim 1, characterized in that, It also includes a display and storage module for real-time display of the three-dimensional distribution cloud map and dynamic change curve information of space charge, and supports local storage and USB export of historical data; it also includes a common power module for powering the various functional modules.

7. A space charge detection method based on optical measurement, characterized in that, The space charge detection device based on optical measurement as described in any one of claims 1-6 comprises the following steps: Step 1, Equipment Deployment: Fix the sensor array around the device under test, connect the modules in the space charge detection device through optical fiber and cable, and start the power supply module to supply power to each module normally. Step 2, parameter configuration: Set the laser parameters of the laser emission module, the signal conditioning parameters of the signal conditioning module, and the acquisition parameters of the data acquisition unit through the display module; Step 3, signal acquisition: Start the laser emission module to output four laser beams of the same source to irradiate the insulation layer of the device under test. At the same time, start the data acquisition unit to synchronously acquire the four raw signal data output by the four sensors. Step 4, data processing, employing a three-level data fusion operation: spatiotemporal synchronization, weight allocation, and feature fusion. Spatiotemporal synchronization calibration uses the laser emission pulse as the time reference, timestamps the four original signals, and establishes a unified three-dimensional coordinate system to obtain four calibration signal data. Feature extraction is performed on the four calibration signal data after spatiotemporal synchronization calibration, and the initial value of the space charge density of each sensor is calculated based on the information features obtained from the feature extraction step. The signal-to-noise ratio of the calibration signal of each sensor is calculated, and the weighting coefficient is dynamically calculated based on the signal-to-noise ratio of the calibration signal of each sensor. Then, the optimal charge density estimate is obtained by performing a fusion operation based on the initial value of the space charge density of each sensor and the weighting coefficient. Finally, based on the optimal charge density estimate, a volume data visualization algorithm is used to construct a three-dimensional space charge distribution model, and to draw distribution cloud maps and dynamic change curves. Step 5, Result Output: The display module outputs a three-dimensional cloud map of space charge and a dynamic change curve.

8. The space charge detection method based on optical measurement according to claim 7, characterized in that, Feature extraction is performed on the spatiotemporally synchronized calibration data, that is, extracting the information features corresponding to each calibration signal at each spatial point in a unified three-dimensional coordinate system, including: extracting the frequency shift Δ from the Raman signal. v and light intensity I 拉曼 ; Extracting amplitude A and audio frequency f from photoacoustic signals; Extracting characteristic peak intensity from LIBS signals And the half-peak width Δλ; the photoelectron flux density J is extracted from the photoelectric signal, and then the initial value of the space charge density of each sensor is calculated based on the information features obtained from the feature extraction step.

9. The space charge detection method based on optical measurement according to claim 7, characterized in that, The weight allocation adopts an adaptive weight allocation strategy, which assigns weight coefficients to each sensor based on the signal-to-noise ratio of the calibration signals of each sensor. Perform dynamic calculations: Where i = 1, 2, 3, 4, corresponding to the four types of sensors respectively. The signal-to-noise ratio of the calibration signal for the i-th sensor is expressed in dB, and the weighting coefficient is... satisfy .

10. The space charge detection method based on optical measurement according to claim 9, characterized in that, The calculation method is as follows: for the calibration signal output by the i-th type of sensor, extract the peak value of the core signal intensity S of the signal segment and the standard deviation of the noise intensity of the noise segment. Then, based on the peak signal strength S and the standard deviation of noise intensity... Calculate using the following formula: ; In the formula, SNR is the signal-to-noise ratio.

11. The space charge detection method based on optical measurement according to claim 7, characterized in that, The fusion operation employs a fusion algorithm combining weighted summation and Kalman filtering. Based on the initial value of the space charge density of each sensor and the weighting coefficients, the charge density at the same spatial point is weighted and fused using the following formula to obtain the estimated value of the fused space charge density. : ; in, This is the initial value of the space charge density of the i-th type of sensor. Q It is the current point in the three-dimensional coordinate system. These are the weighting coefficients for the i-th type of sensor; Then, random errors are eliminated by Kalman filtering to obtain the optimal space charge density estimate for the current space point: ; where K( Q Z( ) is the Kalman gain, Z( Q ) represents the initial value of the space charge density detected by the sensor with the highest signal-to-noise ratio at the current spatial point.

12. The space charge detection method based on optical measurement according to claim 11, characterized in that, Kalman gain K( Q The calculation process is as follows: ; Where: p( Q|Q -1) represents the prediction error covariance of the current spatial point, and H is the observation matrix. R is the transpose of the observation matrix, R is the measurement noise variance, and p is the prediction error covariance. Q|Q -1) Calculate by reverse derivation using the following formula: ;where p( Q|Q ) is the current spatial point Q The posterior error covariance.

13. The space charge detection method based on optical measurement according to claim 7, characterized in that, In step 1, when the sensor array is fixed around the device under test, the four types of sensors in the sensor array are installed at 90° intervals around the device under test in the circumferential direction; the detection end face of all sensors in the sensor array is at the same vertical distance from the device under test; the laser incident angle α is 45° and the sensor receiving angle β is 135°.

14. The space charge detection method based on optical measurement according to claim 7, characterized in that, Step 3 also includes monitoring the signal-to-noise ratio (SNR) of the four raw signals and determining whether the SNR of each raw signal exceeds the threshold. If the SNR of each raw signal exceeds the threshold, the acquisition continues. If the SNR of each raw signal does not exceed the threshold, the process returns to step 2 to reset the laser parameters or the signal conditioning parameters of the corresponding signals until the SNR of all four raw signals exceeds the threshold.

Images