Power equipment partial discharge monitoring and positioning system based on multi-mode sensing

Abstract

The invention relates to the technical field of partial discharge of power equipment, and particularly discloses a multi-modal sensing-based partial discharge monitoring and positioning system for power equipment, and the system comprises a multi-modal sensing unit which is configured to be arranged around to-be-monitored power equipment and is used for synchronously collecting various physical signals generated by partial discharge, the multi-mode sensing unit at least comprises an ultrahigh frequency sensing array used for receiving ultrahigh frequency electromagnetic wave signals generated by partial discharge; the ultrasonic sensing array is used for receiving ultrasonic signals generated by partial discharge; and the data acquisition and synchronization unit is connected with the multi-mode sensing unit. According to the power equipment partial discharge monitoring and positioning system based on multi-modal sensing, an integrated framework of multi-modal sensing cooperation and high-precision time synchronization is constructed, so that the problems of insufficient multi-modal cooperation, low time synchronization precision and no closed-loop optimization mechanism in existing partial discharge monitoring are effectively solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of partial discharge of power equipment, in particular to a partial discharge monitoring and positioning system for power equipment based on multi-modal sensing. BACKGROUND

[0002] Power equipment (such as gas insulated switchgear (GIS), power transformers, high-voltage switch cabinets, etc.) is the core infrastructure for ensuring the safe and stable operation of the power grid, and its insulation performance directly determines the reliability of the power grid operation. Partial discharge is the core cause of insulation degradation of power equipment: in the electric field concentration area inside the equipment (such as the inside of the bushing, the winding joint, and the contact of the disconnector), partial breakdown discharge occurs due to insulation defects (such as bubbles, suspended potential, and surface contamination). This discharge process continuously erodes the insulation medium (such as SF6 gas decomposition, transformer oil aging, and epoxy resin cracking), and if not timely monitored and located, it will eventually cause insulation breakdown, equipment explosion, and other major failures, resulting in large-scale power outages and huge economic losses. Therefore, accurate monitoring and positioning of partial discharge of power equipment is a core requirement for the operation and maintenance of the power industry.

[0003] The current mainstream partial discharge monitoring technology in the power industry mainly focuses on ultrasonic monitoring and ultra-high frequency monitoring, which are two single modal technologies. Some solutions attempt to combine temperature sensing, but due to limitations in technical architecture and algorithm design, there are still the following key technical defects, making it difficult to meet actual operation and maintenance needs:

[0004] Existing monitoring solutions generally have the problem of "single modal dependence":

[0005] For example, pure ultrasonic monitoring solutions (such as online monitoring devices based on a single ultrasonic probe) can capture mechanical vibration signals generated by partial discharge, but are easily disturbed by external environmental noise, such as switch room fan noise, operator movement vibration, and ultrasonic interference caused by motor electromagnetic radiation. Moreover, they cannot distinguish between "internal discharge" and "external environmental interference", resulting in high false positive rates.

[0006] For another example, pure ultra-high frequency monitoring solutions (such as built-in ultra-high frequency sensor systems) can detect ultra-high frequency electromagnetic waves (300MHz-3GHz) generated by discharge and have strong anti-electromagnetic interference capability, but have insufficient sensitivity to "non-corona type weak discharge" (such as suspended potential discharge), and the signal is easily shielded by the metal tank, making it impossible to cover the external area of the equipment (such as the outgoing conductor and the top of the bushing).

[0007] There are also a few solutions that attempt to combine ultrasonic and ultra-high frequency, but only for "data parallel collection", without establishing a spatiotemporal correlation logic between the two. For example, they do not determine whether the two types of signals are generated by the same discharge event, and still have the problem of "mistaking external ultrasonic interference and remote ultra-high frequency interference as joint discharge", which cannot reduce the false positive rate. SUMMARY

[0008] The present application aims to solve at least one of the technical problems in the related art. To this end, the present application aims to propose a power equipment partial discharge monitoring and positioning system based on multi-modal sensing to ensure safe and stable operation of power equipment.

[0009] To achieve the above-mentioned purpose, the first aspect of the present application proposes a power equipment partial discharge monitoring and positioning system based on multi-modal sensing, comprising:

[0010] A multi-modal sensing unit is configured to be arranged around the power equipment to be monitored, for synchronously collecting a plurality of physical signals generated by partial discharge, wherein the multi-modal sensing unit at least comprises:

[0011] A UHF sensing array is used to receive UHF electromagnetic wave signals generated by partial discharge;

[0012] An ultrasonic sensing array is used to receive ultrasonic signals generated by partial discharge;

[0013] A data acquisition and synchronization unit is connected to the multi-modal sensing unit, for high-precision synchronous sampling and time stamp marking of the UHF electromagnetic wave signals and the ultrasonic signals, to generate synchronized multi-modal data;

[0014] A central processing unit is connected to the data acquisition and synchronization unit, and is configured to:

[0015] Receive the synchronized multi-modal data;

[0016] Execute a multi-modal data fusion discrimination algorithm to identify a real joint discharge event based on the spatio-temporal correlation of the UHF electromagnetic wave signals and the ultrasonic signals;

[0017] When the joint discharge event is identified, a joint positioning algorithm is executed to calculate the discharge position coordinates of the joint discharge event in three-dimensional space.

[0018] To achieve the above-mentioned purpose, the second aspect of the present application proposes a power equipment partial discharge monitoring and positioning method based on multi-modal sensing, comprising the following steps:

[0019] First, a multi-modal sensing unit is arranged around the power equipment to be monitored, comprising a UHF sensing array and an ultrasonic sensing array, for synchronously collecting UHF electromagnetic wave signals and ultrasonic signals generated by partial discharge;

[0020] A data acquisition and synchronization unit uses a Beidou / GPS second pulse signal to realize nanosecond-level time synchronization of each sensor channel, and performs high-precision synchronous sampling and time stamp marking on the signals to generate synchronized multi-modal data;

[0021] After the central processing unit receives the data, a multi-modal data fusion discrimination algorithm is executed, and the existence of effective ultrasonic signals is searched within a preset time window based on the ultra-high frequency signal as a trigger reference, and a real joint discharge event is identified through space-time correlation;

[0022] After successful identification, a joint positioning algorithm is executed, a preliminary discharge area is calculated based on the time difference of arrival of ultra-high frequency signals, the sound speed is corrected in combination with the temperature data provided by the thermal imaging unit, and an accurate search is performed in the preliminary area using the time difference of arrival of ultrasonic signals, and finally the three-dimensional coordinates of the discharge position are calculated;

[0023] At the same time, a discharge feature vector is extracted from the multi-modal signal, and a pre-trained classification model is input to identify the discharge type;

[0024] The final result is marked on the device three-dimensional model, the type and severity are displayed, and the hierarchical alarm is triggered when the threshold is exceeded through the visualization unit.

[0025] To achieve the above purpose, the third aspect of the present application provides an electronic device, which includes a memory, a processor and a computer program stored in the memory, and the computer program is executed by the processor to realize the above-mentioned power equipment partial discharge monitoring and positioning method based on multi-modal sensing.

[0026] Compared with the prior art, the beneficial effects of the present application are:

[0027] The power equipment partial discharge monitoring and positioning system based on multi-modal sensing of the embodiment of the present application effectively solves the problems of insufficient multi-modal cooperation, low time synchronization accuracy and lack of closed-loop optimization mechanism in existing partial discharge monitoring by constructing a multi-modal sensing cooperation + high-precision time synchronization integrated architecture; wherein the multi-modal sensing array combined with the fusion discrimination algorithm of space-time correlation can accurately distinguish between real discharge inside the device and external environmental interference, and avoid false positives and false negatives;

[0028] The nanosecond-level synchronization realized by the UWB double-sided bidirectional ranging ensures the effective correlation of different modal signals, laying a foundation for multi-source fusion; the adaptive dynamic compensation mechanism combines with the device structure and real-time environmental parameters to correct the signal propagation characteristics, improving the accuracy of discharge positioning, and finally realizing accurate monitoring, comprehensive positioning, intelligent early warning and efficient operation and maintenance of power equipment partial discharge, and ensuring the safe and stable operation of the power grid. BRIEF DESCRIPTION OF DRAWINGS

[0029] The disclosure of the present application will be described with reference to the accompanying drawings. It should be understood that the drawings are only for illustrative purposes, and are not intended to limit the scope of protection of the present application. In the drawings, the same reference numerals are used to refer to the same components. Among them:

[0030] Figure 1 is a schematic diagram of a framework structure of a power equipment partial discharge monitoring and positioning system based on multi-modal sensing provided by the present application;

[0031] Figure 2 is a schematic diagram of alignment of UHF and AE signals under nanosecond-level synchronization in a power equipment partial discharge monitoring and positioning system based on multi-modal sensing provided by the present application (a) and (b);

[0032] Figure 3 is a schematic diagram of signal energy threshold determination beyond and in a power equipment partial discharge monitoring and positioning system based on multi-modal sensing provided by the present application (c) and (d);

[0033] Figure 4 is a schematic diagram of UHF coarse positioning area superimposed AE precise positioning result (e) and positioning error distribution (f) in a power equipment partial discharge monitoring and positioning system based on multi-modal sensing provided by the present application;

[0034] Figure 5 is a curve change diagram of sound velocity vs. temperature in a power equipment partial discharge monitoring and positioning system based on multi-modal sensing provided by the present application;

[0035] Figure 6 is a schematic diagram of positioning error change with temperature in a power equipment partial discharge monitoring and positioning system based on multi-modal sensing provided by the present application;

[0036] Figure 7 is a schematic diagram of a three-dimensional model of a calibration discharge position in a power equipment partial discharge monitoring and positioning system based on multi-modal sensing provided by the present application;

[0037] Figure 8 is a schematic diagram of pre-warning level color mapping in a power equipment partial discharge monitoring and positioning system based on multi-modal sensing provided by the present application;

[0038] Figure 9 is a schematic diagram of a flow of a power equipment partial discharge monitoring and positioning method based on multi-modal sensing provided by the present application;

[0039] Figure 10 is a schematic diagram of a structure of an electronic device provided by the present application. DETAILED DESCRIPTION

[0040] Embodiments of the present application are described below in detail, examples of which are shown in the drawings, wherein the same or similar reference signs represent the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary and are intended to explain the present application, and cannot be understood as a limitation of the present application.

[0041] A multi-modal sensor-based power equipment partial discharge monitoring and positioning system is described below with reference to the drawings. Embodiment I

[0042] Figure 1 A multi-modal sensor-based power equipment partial discharge monitoring and positioning system is described below with reference to the drawings.

[0043] As Figure 1 shown, in order to realize accurate monitoring and positioning of partial discharge of large power equipment, the deployment and composition of the three core units of the system in specific implementation are as follows:

[0044] 1. A multi-modal sensor unit is configured to be arranged around the power equipment to be monitored, for synchronously collecting various physical signals generated by partial discharge, and the multi-modal sensor unit at least includes:

[0045] a UHF sensor array for receiving UHF electromagnetic wave signals generated by partial discharge;

[0046] an ultrasonic sensor array for receiving ultrasonic signals generated by partial discharge.

[0047] Among them, the UHF sensor array: at least 4 built-in or external UHF sensors are used. For oil-immersed transformers, the sensors can be installed through the valve (such as the oil drain valve) reserved on the tank wall, so that the antenna part can effectively receive the UHF electromagnetic wave signals generated by the partial discharge inside the transformer, and the frequency range is usually 300MHz-3GHz. The 4 sensors are distributed in a non-coplanar and non-collinear manner in space to form a three-dimensional sensing array, which is the geometric basis for three-dimensional positioning.

[0048] The ultrasonic (AE) sensor array: also uses at least 4 piezoelectric ultrasonic sensors, which are tightly attached to the outer surface of the transformer tank wall by magnetic adsorption or adhesion. The arrangement of the sensors should fully consider the tank structure, uniformly cover the key areas such as windings, bushings and tap changers, and avoid obstacles as much as possible. They are used to receive the signals of the stress waves, i.e. ultrasonic signals, generated by partial discharge and transmitted to the outer wall after propagating in the tank structure, and the frequency range is usually 20kHz-300kHz.

[0049] 2、Data acquisition and synchronization unit, connected with multi-modal sensing unit, used for high-precision synchronous sampling and time stamping of UHF electromagnetic wave signal and ultrasonic signal, to generate synchronized multi-modal data. This unit is the bridge connecting front-end sensing and back-end processing, which is composed of multiple parallel acquisition channels and a unified synchronization clock source.

[0050] Among them, the importance of high-precision synchronization here is: to ensure the accuracy of the subsequent TDOA (Time Difference of Arrival) positioning algorithm, time synchronization is the technical cornerstone of the system. In this embodiment, the Beidou / GPS time service module is adopted, and the second pulse signal (PPS) output by the module is transmitted to each acquisition channel. The pulse front edge accuracy of the PPS signal can reach nanosecond level, and based on this, each channel can realize strict time alignment and high-precision time stamping.

[0051] Synchronous sampling: under the trigger and calibration of the PPS signal, the analog-to-digital converters (ADCs) of all channels sample the UHF and AE signals in a completely synchronized manner. The sampling rate of the UHF signal is not less than 2GSa / s to ensure the true restoration of the pulse waveform; the sampling rate of the AE signal is not less than 5MSa / s.

[0052] As Figure 2 The figure shows the alignment of UHF and AE signals under nanosecond-level synchronization. In figure (a), the blue solid line represents the UHF signal actually collected, which contains background noise. The red dashed line outlines the shape of the ideal discharge pulse submerged in noise, which has a nanosecond-level (~10ns) steep rising edge and a very narrow pulse width. The black dashed line accurately marks the arrival time of the UHF pulse, =2.0μs. In figure (b), the green solid line represents the AE signal collected, which also contains significant noise, simulating mechanical noise sources such as oil pump vibration and fan rotation. The magenta dashed line represents the ideal AE signal, which is a decaying oscillation wave with a frequency of about 150kHz. The black dashed line again marks the UHF trigger time . The red dashed line marks the arrival time of the AE signal ≈62.0μs. The yellow highlighted area in the figure represents the preset time window . The window starts at , and its width (set to ~70μs in the figure) is (maximum size of the device / speed of sound), ensuring that the AE signal from the farthest sensor can be captured.

[0053] 3. A central processing unit, connected with the data acquisition and synchronization unit, which is usually composed of high-performance embedded industrial computers or servers, with high-speed data bus and powerful processors (CPU / GPU / FPGA) inside, capable of real-time analysis and calculation of massive synchronous multi-modal data from the data acquisition unit

[0054] The central processing unit is configured to:

[0055] receive the synchronous multi-modal data;

[0056] execute a multi-modal data fusion discrimination algorithm, based on the spatio-temporal correlation of the ultra-high frequency electromagnetic wave signals and the ultrasonic signals, to identify a real joint discharge event;

[0057] When the joint discharge event is identified, execute a joint positioning algorithm to calculate the discharge position coordinates of the joint discharge event in three-dimensional space.

[0058] The workflow of the system aims to solve the two technical pain points of noise interference and low positioning accuracy that exist in traditional single-mode monitoring methods. The core principle is to use the physical correlation that local discharge inevitably produces both electromagnetic and acoustic signals, and through a two-step strategy of discrimination first and positioning later, to accurately lock the real discharge signal and position it from complex background noise. The specific steps include:

[0059] Step 1: Multi-modal data fusion discrimination based on spatio-temporal correlation, which aims to explore whether the signal is a real local discharge.

[0060] The physical basis of this step is that when local discharge occurs, the UHF electromagnetic wave propagates at nearly the speed of light ( ), while the ultrasonic wave propagates much slower in insulating oil ( ). Therefore, for the same discharge event, the UHF sensor will receive the signal almost instantaneously, while the AE sensor will have a significant delay proportional to the propagation distance. Random electromagnetic interference (such as mobile phone signals) or mechanical vibration noise (such as accessory vibration) do not have this strict "electricity first and sound later" and reasonable time delay characteristics.

[0061] The algorithm implementation process of this step is:

[0062] UHF signal preliminary trigger: the central processing unit monitors the data stream of all UHF channels in real time. When the signal amplitude of any UHF channel exceeds the preset noise baseline threshold, the system records the arrival time of the pulse and marks it as a candidate discharge event. The high sensitivity of UHF signals ensures that weak real discharges are not missed;

[0063] Establish dynamic time window: starting from , the system will open a preset time window . The width of the window is calculated according to the maximum geometric size of the device under test and the speed of sound , and the formula can be expressed as: , which represents the longest time required for an ultrasonic wave to propagate from any point in the device to the farthest sensor;

[0064] AE signal correlation confirmation: the system then checks whether there are valid ultrasonic signals in this time window, i.e. signals with energy exceeding their own noise threshold.

[0065] Then make event discrimination decision:

[0066] Real event: if within , the system successfully detects valid signals from multiple (e.g. at least 3) AE sensors, the central processing unit determines that this "candidate discharge event" is a "real joint discharge event", which means that the electromagnetic and acoustic phenomena are cross-verified.

[0067] Noise removal: if no valid AE signal is detected within , it is determined that the initial UHF pulse is external electromagnetic interference and is removed from the data set; otherwise, if there is only AE signal without UHF trigger, it is determined as mechanical vibration noise.

[0068] Second step: joint positioning based on multi-modal data cooperation, aiming to solve the problem of positioning accuracy.

[0069] The algorithm implementation process of this step is:

[0070] UHF coarse positioning: the system first uses the arrival time stamp of a group of UHF signals associated with the event. By calculating the time difference of arrival (TDOA) between different sensors and solving a set of hyperbolic equations, an initial estimated area of the discharge location can be obtained. Due to the complex propagation path of UHF signals inside the device, there may be some errors in the positioning result, but it is sufficient to define a reliable search range;

[0071] AE precise positioning: next, the system uses the arrival time stamp of a group of AE signals associated with the event. Similarly, the TDOA algorithm is used, but the key difference is that the search algorithm for solving the sound source position is constrained within the initial estimated area obtained in the first step. This electromagnetic-guided-acoustic strategy has two advantages:

[0072] First, it eliminates the multi-solution problem. When using AE signals for global positioning, multiple pseudo-solutions often appear due to sensor layout and time measurement errors. However, by limiting the search range, these pseudo-solutions can be effectively excluded, and a unique true solution can be found.

[0073] Second, it improves the computational efficiency. Instead of searching the entire three-dimensional space of the device, the convergence speed of the positioning algorithm is greatly accelerated.

[0074] Finally, the accurate positioning result of AE, i.e., the precise coordinates (x, y, z) of the joint discharge event in the three-dimensional model of the device, is calculated and output.

[0075] As Figure 4 The UHF coarse positioning area is superimposed with the AE accurate positioning result (e) and the positioning error distribution (f). In the left graph (e), the blue circular area represents the coarse positioning result based on UHF signals, with a radius of about 1.2 m, reflecting the ability of UHF methods to lock the approximate area of partial discharge, but with limited accuracy. The red scattered points represent the accurate positioning results calculated by the AE algorithm, distributed near the true discharge source, indicating that AE positioning achieves secondary optimization within the UHF area. The yellow pentagram marks the true discharge source coordinates (2.5 m, 3.5 m), which can be visually seen that the AE positioning result is significantly closer to the true position.

[0076] In the right graph (f), the error distribution histogram of AE positioning results and true position is shown. Most of the positioning result errors are concentrated in the 0.1-0.3 m interval, with an average error of about 0.18 m and a maximum error of not more than 0.5 m. This graph directly proves that the joint algorithm process of UHF coarse positioning + AE accurate positioning can balance range locking and accuracy optimization.

[0077] Compared with existing single modality monitoring technology, this system uses the inherent correlation between different physical signals to establish a powerful logical filter. Any single source noise that cannot cause both electromagnetic and acoustic phenomena (whether external electromagnetic interference or internal mechanical vibration) will be effectively eliminated, thus fundamentally solving the false alarm and missed alarm problems. Each discharge event output by the system has been cross-verified, and the result is highly reliable; and the joint positioning algorithm creatively adopts a two-step strategy of UHF coarse positioning + AE accurate positioning. It is not a simple weighted average of the two positioning results, but uses the global arrival characteristics of UHF signals to provide prior information for AE positioning, and then uses the advantage of relatively clear AE signal propagation path for accurate solution. This cooperative mechanism overcomes the inherent defects of single modality positioning, can improve the positioning accuracy from meters to centimeters, and realizes accurate tracing of the fault point.

[0078] In addition to this, the system can not only answer the question of whether there is discharge and where the discharge is, but also provides rich information for subsequent fault diagnosis through the synchronous acquisition of multi-modal data. Since the electromagnetic and acoustic waveforms of the discharge are captured simultaneously, the type (corona, surface, air gap) and severity of the partial discharge can be more accurately determined by analyzing the internal relationship between the characteristics of these waveforms, such as the steepness of the UHF pulse and the proportion of the AE signal energy, which is difficult to achieve with any single information source.

[0079] By way of example, the above-mentioned multi-modal sensing unit further comprises:

[0080] One or more long-wave infrared thermal imaging sensing units are added. The unit is installed in a position that can overlook the key area of the power equipment to be monitored (such as a large oil-immersed transformer), ensuring that its field of view can completely cover the main heat dissipation surface of the transformer tank.

[0081] The thermal imaging sensing unit periodically (e.g., once every minute) captures the infrared radiation of the device surface and converts it into a two-dimensional temperature distribution map This map is transmitted in the form of a digital matrix to the central processing unit in real time. The central processing unit then registers the real-time temperature map with the pre-established three-dimensional digital model of the power equipment, thereby obtaining real-time temperature data of any point on the device surface.

[0082] In addition, the central processing unit is further configured to compensate for the propagation speed of the ultrasonic wave signal using the temperature distribution data. This is because the accuracy of the ultrasonic positioning algorithm fundamentally depends on the accuracy of the formula distance = speed x time.

[0083] In traditional acoustic positioning technology, the propagation speed of ultrasonic waves in insulating media (such as transformer oil) is usually assumed to be a fixed constant. However, this is a serious idealized assumption that deviates from reality. Physical principles show that the speed of sound in a liquid medium is closely related to the temperature and density of the medium. For large power transformers, the internal oil temperature distribution is extremely uneven during operation, with the top oil temperature being tens of degrees Celsius different from the bottom oil temperature. Ignoring this temperature gradient and using a fixed average sound speed for calculation will inevitably result in a huge positioning error, with an error amplitude of tens of centimeters or even more, making the positioning result lose its practical guiding significance. This is a major technical bottleneck that has long plagued high-precision acoustic positioning in the field.

[0084] How to use temperature data for compensation? The present invention proposes an innovative compensation process of electromagnetic pre-positioning guidance and thermal imaging data correction, whose working principle is as follows:

[0085] Step 1: Preliminary positioning and region definition. After the system confirms a real combined discharge event through spatio-temporal correlation, it first calculates an initial estimated region of the discharge using the time difference of arrival (TDOA) of signals from the UHF sensor array ;

[0086] Step 2: Extracting local equivalent temperature. The central processing unit queries and extracts the real-time surface temperature value corresponding to the initial estimated region in the device's three-dimensional model based on its location. Through a heat conduction model, the local equivalent temperature inside the region can be estimated ;

[0087] Step 3: Calculating the corrected sound speed. The central processing unit calls a pre-set physical model or lookup table to calculate the corrected local sound speed of the region based on the local equivalent temperature . This model is usually a temperature-dependent function, which can be represented by a linear approximation formula:

[0088] where , and are known constants related to the characteristics of the insulating medium;

[0089] Step 4: Performing high-precision acoustic positioning. Finally, the system substitutes this accurately compensated local sound speed into the ultrasonic TDOA positioning algorithm to solve the final discharge position coordinates.

[0090] The core contribution of the above steps lies in solving the key problem of sound speed uncertainty from a physical level. By introducing thermal imaging as a new dimension of information, the positioning model is transformed from a static, idealized model to a dynamic, accurate model that can adapt to the real-time operating state of the device. This enables the positioning error to be steadily improved from the traditional decimeter level to the centimeter level, achieving a qualitative leap in positioning accuracy.

[0091] As Figure 5 shows the sound speed vs. temperature curve, showing the fitting of the theoretical formula and the measured data. Figure 5 The blue solid line is the theoretical formula of sound speed changing with temperature (about 331.4 + 0.6T m / s), the red dots are the measured sound speed data, and the green dashed line is the linear fitting result based on the measured data. From the curve, it can be seen that the measured data and the theoretical formula have basically the same trend, but there is a slight deviation above 40°C, with an average deviation of about ±2 m / s. This result shows that the sound speed-temperature fitting method based on measured data can optimize the theoretical model, making the sound speed estimation closer to the actual working condition.​

[0092] In some embodiments of the present application, the definition of spatio-temporal correlation is:

[0093] In the preset time window , both the ultra-high frequency sensor array and the ultrasonic sensor array detect signals exceeding the respective preset energy thresholds and ;

[0094] wherein the preset time window : the width of this window is not an arbitrary value, but is scientifically set based on physical principles. This ensures that the window size is sufficient to cover the longest time consumption of ultrasonic waves propagating from any point in the device to any sensor.

[0095] Energy threshold and : these two thresholds are set above the background noise level of the respective channel, for preliminary judgment of the effectiveness of the signal.

[0096] Since the high-voltage device site is an extremely complex electromagnetic and mechanical noise environment, including electromagnetic noise sources: such as radio broadcasting, mobile phone communication, switch operation, etc., which will generate a large amount of pulse interference, which is extremely easy to be mis-captured as discharge signals by sensitive UHF sensors; such as mechanical noise sources: such as oil pump vibration, fan rotation, accessory loosening, etc., which will generate stress waves, which will be mis-captured by AE sensors. If only the signal of a single sensor is judged as partial discharge, it will lead to a very high false alarm rate, making the monitoring system useless. How to accurately extract the real discharge signal from the vast amount of noise signals is the primary problem that all monitoring systems must solve.

[0097] By defining the spatio-temporal correlation, the system can perform event discrimination. The discrimination algorithm of this embodiment ingeniously uses the physical nature that partial discharge must simultaneously produce electromagnetic waves (near light speed) and acoustic waves (slow speed), and designs a "electromagnetic trigger, acoustic confirmation" logic criterion:

[0098] Step one (trigger): the algorithm takes the UHF signal as the "sentinel". When any UHF sensor detects a pulse signal with energy exceeding , record its arrival time , and take this as the benchmark to immediately open a confirmation window with a time length of ;

[0099] Step two (confirmation): the algorithm checks the data of all ultrasonic sensors within this confirmation window ;

[0100] Step 3 (decision): If at least one (or a pre-set number) of the ultrasonic sensors detects a signal with energy exceeding within this window, the system determines that the two signals originate from the same physical event, possessing spatiotemporal correlation, and thus confirms it as a real combined discharge event.

[0101] If no valid ultrasonic signal is detected when the confirmation window is closed, the system determines that the initial UHF signal is a random electromagnetic interference and discards it.

[0102] In summary, this definition establishes a strong logical constraint: a random external electromagnetic interference cannot spontaneously generate a matching ultrasonic signal within the appropriate time delay inside the device; conversely, a random mechanical vibration cannot spontaneously generate a precisely advanced ultra-high frequency electromagnetic pulse. Therefore, this spatiotemporal correlation criterion can fundamentally filter out most irrelevant noise, reducing the false positive rate to a very low level, ensuring the high reliability of the monitoring results. And by removing a large amount of invalid noise data at the very front end of data processing, the computational burden of subsequent positioning and analysis algorithms is greatly reduced, ensuring the real-time response capability of the entire system.

[0103] As Figure 3 exceeds and signal energy threshold determination diagrams. The blue curve (UHF signal energy) and green curve (AE signal energy) in figures (c) and (d) represent the partial discharge signal energy detected by the ultra-high frequency (UHF) and acoustic emission (AE) sensors, respectively. The UHF signal energy ranges from about 20 dB to 100 dB, with background noise maintained at around 30 dB, while partial discharge events produce peak signals up to 65-89 dB. The energy range of AE signals is slightly lower, reflecting the attenuation characteristics of sound wave propagation. The red dashed lines in the figure ( = 50 dB, = 45 dB) represent the set signal energy thresholds. Through these thresholds, normal background noise and real discharge events can be effectively distinguished, avoiding false positives in traditional methods.

[0104] Furthermore, Figure 3 the red solid dots in the figure identify signal points exceeding the threshold, which correspond to partial discharge events occurring at different time points (e.g. 1.5 seconds, 3.2 seconds, etc.). Through bimodal signal processing, the figure clearly shows the change of the signal on the time axis and whether it exceeds the set threshold. Under the multi-modal analysis of UHF signals and AE signals, the accuracy and reliability of event recognition can be effectively improved, thereby greatly reducing the false positive rate.

[0105] In some embodiments of the present application, in order to achieve zero false alarm event discrimination from strong noise background, the central processing unit in the system executes a multi-modal data fusion discrimination algorithm, which uses multi-physical dimension information cross verification to ensure that only real partial discharge events can pass through, thereby fundamentally solving the false alarm problem, and the specific steps include:

[0106] Step one: preliminary detection of ultra-high frequency electromagnetic wave signals to identify candidate discharge pulses;

[0107] Technical details: 1. Signal preprocessing: the original UHF signal received from the data acquisition and synchronization unit is not used directly, but first filtered by digital filtering (e.g. band-pass filtering) to filter out noise outside the frequency band, and denoised by wavelet transform and other techniques to improve the signal-to-noise ratio SNR;

[0108] 2. Dynamic threshold setting: the system does not use a fixed amplitude threshold, but implements an adaptive threshold algorithm. The central processing unit periodically analyzes the background noise level during the signal-free period and dynamically sets a trigger threshold higher than the noise baseline (e.g. 5 times higher than the root mean square value);

[0109] 3. Candidate event generation: when the amplitude of the pre-processed UHF signal pulse continuously exceeds this dynamic threshold, the system identifies the starting point of the pulse as a candidate discharge pulse. At the same time, the system accurately records the time stamp of the pulse front arriving at the sensor and extracts its waveform data.

[0110] Step two: based on the arrival time of each candidate discharge pulse, search for whether there is valid ultrasonic signal collected by the ultrasonic sensor array within the preset time window ;

[0111] Technical details: 1. Scientific setting of time window: the setting of the preset time window is the key to this algorithm, which is based on clear physical principles rather than empirical guesses. This window represents the maximum theoretical time required for ultrasonic waves to propagate from any point within the device to the farthest sensor, ensuring that we do not miss any potentially associated acoustic signals;

[0112] 2. Definition of valid ultrasonic signal: within this time window, the system will simultaneously search all ultrasonic sensor data. A valid ultrasonic signal must meet two conditions: (a) its signal energy exceeds the background noise threshold of its channel; (b) its appearance is not isolated, usually requiring detection on at least N (e.g. N >= 3) ultrasonic sensors to rule out local mechanical interference on a single sensor.

[0113] Step three: If exists, confirm the corresponding UHF signal and ultrasonic signal of the group as a joint discharge event;

[0114] Technical details: 1. Association confirmation: If the system successfully retrieves the valid ultrasonic signal that meets the conditions within the time window of step two, the authenticity of the candidate discharge pulse is cross-verified in the acoustic dimension.

[0115] 2. Data packaging and transmission: The central processing unit will formally confirm this event as a joint discharge event at this time. It will package the UHF signal waveform associated with the event, all related AE signal waveforms, and their respective accurate time stamps into a complete data set, and transmit it to the subsequent joint positioning algorithm module for processing.

[0116] 3. Noise elimination: If no valid ultrasonic signal is detected within the time window After turning off, the system determines that the initial UHF pulse is random electromagnetic interference (such as space radiation), and decides to discard it without entering the subsequent processing flow.

[0117] During the execution of the multi-modal data fusion discrimination algorithm by the central processing unit described above, there are two challenges, including:

[0118] Challenge one: the problem of missed reports caused by signal attenuation

[0119] In large equipment, a real but weak energy partial discharge may produce a severely attenuated ultrasonic signal during transmission, resulting in a signal below the detection threshold when it reaches the sensor. At this time, although the UHF signal is captured, it cannot be confirmed by the valid AE signal, and the algorithm may misjudge it as noise, resulting in a missed report.

[0120] For example, in this embodiment, a hierarchical confirmation mechanism is introduced to further optimize step three, which specifically includes:

[0121] High-confidence event: UHF signal is strong, and multiple AE sensors detect clear signals. Directly confirmed as a joint discharge event;

[0122] Medium-confidence event (suspected event): UHF signal is strong, but only 1 or 2 AE sensors detect weak signals. The system does not discard it immediately, but marks it as a suspected discharge event and performs timing correlation analysis, i.e., observes whether such events repeatedly occur near the location within a short period of time. If repeated, the confidence is increased.

[0123] The above flexible processing strategy not only maintains the immunity of the system to strong noise, but also greatly improves the detection sensitivity of early weak discharge, solving the problem of missed reports caused by rigid criteria.

[0124] Challenge 2: Correlation confusion problem under dense or continuous background noise

[0125] In some severe stages of fault development, multiple partial discharges may occur in a very short time (millisecond level) in succession; or, there is a persistent, pulsed mechanical noise source in the equipment. This will cause multiple UHF pulses and multiple AE signals to occur in the same time window, resulting in false correlation.

[0126] For example, the present embodiment can adopt a more advanced matching strategy, which specifically includes:

[0127] ① Preliminary positioning guidance: for the first UHF pulse in the window, the system can first use the time difference of its arrival at different UHF sensors to quickly estimate a rough discharge area;

[0128] ② Arrival time prediction and matching: based on this rough area, the system can predict when the ultrasonic signal should arrive at each AE sensor;

[0129] ③ Best match search: then, the system will find a set of AE signals that best match the "predicted arrival time sequence" among the multiple AE signals actually detected in the window.

[0130] This refined correlation method based on prediction first and then matching can accurately find the unique multi-modal signal set corresponding to each discharge event even in a dense signal situation, greatly improving the algorithm's analysis capability and accuracy under complex working conditions, and pointing to high-value information of the real physical event every time, laying a solid and reliable data foundation for subsequent accurate positioning and state assessment.

[0131] In some embodiments of the present application, the joint positioning algorithm includes:

[0132] First, based on the time difference of arrival of signals received by each sensor in the very high frequency sensor array , a preliminary discharge location area is calculated

[0133] Then, based on the time difference of arrival of signals received by each sensor in the ultrasonic sensor array , and combined with the sound speed corrected by the temperature data provided by the thermal imaging sensing unit , an accurate search is performed within the preliminary discharge location area to obtain the final discharge location coordinates .

[0134] In the above, the initial discharge location area ​This isn't the final result, but rather a preliminary search space calculated from the UHF sensor array signals. Once a joint discharge event is confirmed, the time difference of arrival (TDOA) algorithm is used to quickly determine the approximate area of ​​the discharge, leveraging the near-light-speed propagation of UHF signals and their minimal influence on the medium. This crucial step significantly reduces the computational complexity and ambiguity of subsequent acoustic positioning. It's like using a radar to initially locate the approximate target area, avoiding aimless searches and effectively eliminating ghost solutions, a common occurrence in acoustic positioning.

[0135] The speed of sound calculation based on temperature compensation , is to solve the biggest error source of traditional acoustic positioning, that is, the sound speed It is considered a fixed constant. However, in actual large-scale power equipment, the temperature distribution of the internal insulating medium (such as transformer oil) is extremely uneven, resulting in significant differences in sound velocity at different locations. Ignoring the influence of temperature is the fundamental reason for the poor positioning accuracy of traditional methods.

[0136] Therefore, the thermal imaging sensor unit is introduced to provide a dynamic and accurate physical parameter for acoustic positioning. It no longer relies on an inaccurate average sound speed, but is based on Real-time temperature of the area , calculate a local sound speed that fully matches the current working conditions through this physical formula .

[0137] Precise positioning: Get the precise local sound speed After that, the system will Within the area, the arrival time difference data of the ultrasonic AE sensor array is used for the final iterative search or analytical solution to obtain the centimeter-level discharge position coordinates.

[0138] For example, the corrected sound speed It can be determined by the following formula:

[0139] in, is the corrected speed of sound, The propagation speed of ultrasonic waves in insulating media at 0 degrees Celsius is is the Celsius temperature of the discharge location area measured by the thermal imaging sensor unit.

[0140] As an example, first set the following device parameters:

[0141] Equipment to be monitored: 500kV gas-insulated switchgear (GIS), tank length 15 meters, bushing height 8 meters, rated current 3150A;

[0142] UHF sensor array: 4 sensors (2 built-in + 2 external), frequency band 300MHz-3GHz, sensitivity ≤1pC, sampling rate 2GSa / s;

[0143] Ultrasonic sensor array: 4 sensors, frequency 20-300kHz, sampling rate 5MSa / s, amplitude measurement range 0-80dBμV, strong magnetic adsorption installation;

[0144] Thermal imaging sensor unit: resolution 640x512, temperature measurement range -40℃-125℃, accuracy ±0.5℃, sampling period 1 minute;

[0145] Data acquisition and synchronization unit: Beidou PPS synchronization, timestamp accuracy 1ns, data transmission rate 100Mbps.

[0146] Then the partial discharge positioning calculation is performed:

[0147] Step 1: Multimodal data acquisition

[0148] UHF sensor: 1 candidate discharge pulse is detected, arrival time , amplitude 6pC;

[0149] Ultrasonic sensor: In the window (14:30:00.123456789s~14:30:00.132256789s), 4 sensors all detect valid signals, among which sensor 1 arrives at time , amplitude 45dBμV;

[0150] Thermal imaging unit: The temperature of the candidate discharge area (near the GIS sleeve) is collected .

[0151] Step 2: Data synchronization and discrimination

[0152] Synchronization verification: UHF and ultrasonic signal timestamp deviation ≤1ns, meeting the synchronization requirements;

[0153] Spacetime correlation decision: UHF signal exceeds threshold (6pC>5pC), AE signal exceeds threshold (45dBμV>30dBμV), and AE arrives within , confirming it as a joint discharge event.

[0154] Step 3: Joint positioning calculation

[0155] Preliminary positioning (UHF): Calculate the TDOA of sensors A (built-in, coordinates (2,3,4)m) and B (external, coordinates (10,3,4)m) , combined with the electromagnetic wave speed , and obtain the initial region : “X∈[8,10]m, Y∈[2,4]m, Z∈[3,5]m” (error ≤ 1 meter);

[0156] Speed ​​of sound correction: hour, ;

[0157] Precise positioning: Calculate the TDOA of sensor 1 (coordinates (2,3,4)m) and sensor 2 (coordinates (5,3,4)m) , combined with , solve for the discharge coordinates The positioning error is ±0.08 m, that is, the discharge is located 8.5 m inside the GIS casing.

[0158] Figure 4 shows a schematic diagram (e) of the UHF coarse positioning area superimposed on the AE precise positioning results, as well as a distribution diagram (f) of the positioning error. In the left figure (e), the blue circular area represents the coarse positioning result based on the UHF signal, with a radius of approximately 1.2m. This shows that the UHF method can locate the approximate area of ​​the partial discharge, but with limited accuracy. The red scattered points represent the precise positioning results calculated by the AE algorithm, which are distributed near the actual discharge source, indicating that AE positioning achieves secondary optimization within the UHF area. The yellow five-pointed star marks the coordinates of the actual discharge source (2.5m, 3.5m). It can be seen that the AE positioning result is significantly closer to the actual location.

[0159] Figure (f) on the right shows a histogram of the error distribution between the AE positioning results and the true position. Most positioning results show errors concentrated in the 0.1–0.3m range, with an average error of approximately 0.18m and a maximum error of no more than 0.5m. This figure visually demonstrates that the combined UHF coarse positioning + AE fine positioning algorithm can achieve both range lock and accuracy optimization.

[0160] And as Figure 6 Shows how the positioning error varies with temperature. Figure 6 The red curve represents the positioning error without temperature compensation. The error increases significantly with increasing temperature, reaching approximately 3.5 m at 70°C. The green curve shows the results after implementing the proposed temperature compensation algorithm, with the error remaining stable within the 0.2–0.4 m range. This comparison fully demonstrates the effectiveness of the temperature compensation mechanism, resolving the positioning deviation caused by temperature variations in traditional methods and significantly improving the reliability and accuracy of discharge source location.

[0161] In the above content, since the precise search is based only on the time difference of arrival (TDOA) of the AE signal, the following limitations are not considered:

[0162] 1. Multipath effect: The ultrasonic waves are reflected on the surface of the metal tank body and insulating parts of the equipment, generating "reflected waves" and "direct waves" superimposed, resulting in a TDOA calculation deviation of ≥15μs;

[0163] 2. Fuzzy boundary: The initial discharge area (UHF coarse positioning) The boundary is a "fuzzy range", and the search domain is not narrowed by combining the equipment structure (such as the diameter of the GIS tank body and the aperture of the bushing), resulting in slow convergence (time consumption >500ms) and "pseudo-solution" of the search algorithm (such as the least squares method). For example, the position corresponding to the reflected wave is misjudged as the real discharge point, and the actual positioning accuracy is reduced to more than 0.2 meters, and the real-time performance cannot meet the online monitoring requirements.

[0164] Based on the above limitations, the following steps are adopted in this embodiment to further improve the accuracy and efficiency, specifically including:

[0165] First step: Direct wave separation: Based on the "amplitude attenuation characteristics" and "phase consistency" of AE signals, the direct wave, i.e. the direct propagation signal generated by the real discharge, is extracted from the superimposed signal, and the reflected wave (interference signal) is removed, ensuring that the TDOA calculation is based on effective signals;

[0166] Second step: Dynamic search domain optimization: According to the three-dimensional structure of the equipment, such as the inner wall coordinates of the GIS tank body and the aperture range of the bushing, the boundary of the initial area is corrected to the "equipment physical constraint boundary" (such as excluding the area where discharge cannot exist inside the metal tank body), and the search range is reduced;

[0167] Third step: Improved particle swarm algorithm: Taking "minimum direct wave TDOA error" as the objective function, the real discharge position is quickly optimized in the dynamic search domain, taking into account accuracy and efficiency.

[0168] In some embodiments of the present application, the central processing unit is further configured to:

[0169] First, extract the discharge feature vector from the multi-modal signals corresponding to the joint discharge event. The original multi-modal signal is a complex time-domain waveform, containing a large amount of redundant information and noise. Directly inputting the original signal into the model not only has a huge amount of calculation, but also makes it difficult for the model to quickly converge and learn effective rules. Therefore, it is necessary to first extract the identity information that can most significantly distinguish different discharge types, i.e. to construct a highly condensed and representative discharge feature vector. The feature combination selected in this embodiment is based on a deep understanding of the discharge physical mechanism, taking into account the characteristics of both electromagnetic waves and sound waves.

[0170] By way of example, the discharge feature vector includes:

[0171] UHF signal features:

[0172] Pulse Amplitude: This is the peak voltage of the UHF signal pulse waveform. It is directly related to the energy intensity released in a single discharge event. Generally, internal gas gap discharge has a relatively large pulse amplitude due to the short path and concentrated energy; while the amplitude of corona discharge is usually small and distributed discretely.

[0173] Pulse Rise Time: Defined as the time required for the pulse to rise from 10% to 90% of its peak value. This parameter reflects the speed of discharge development. For example, the discharge initiated by sharp electrodes (such as corona) develops extremely fast, with a rise time usually in the nanosecond (ns) level, very steep. Surface discharge, due to its development along the surface of the insulator, has a relatively tortuous path, and its rise time may be slightly longer.

[0174] Pulse Width: Usually refers to the time duration when the pulse amplitude drops to 50% (half-height width). It represents the duration of a single discharge pulse. Different types of discharge have different durations of plasma processes, which are directly reflected in the pulse width.

[0175] Ultrasonic signal features:

[0176] Energy: The energy of the ultrasonic signal is usually obtained by calculating the square sum (or integral) of the signal within a certain time window. It reflects the intensity of the sound pressure generated by the instantaneous expansion of the gas in the discharge process. Internal gas discharge, being enclosed by a metal shell, is more easily coupled to the sensor by sound wave energy, showing a higher energy value. Surface discharge is next, and corona discharge produces the weakest sound wave.

[0177] Duration: Refers to the time elapsed from the start to the decay of the ultrasonic signal to the background noise level. This parameter is related to the continuity of the discharge process. For example, surface discharge may be accompanied by multiple consecutive small discharges, forming a long-duration sound wave event envelope.

[0178] Dominant Frequency: By performing Fourier transform (FFT) or other spectral analysis methods on the ultrasonic signal, the frequency point with the highest energy is found. Different types of discharge have different spectral characteristics of their sound wave signals. For example, internal gas gap discharge produces sound waves with frequencies usually concentrated in the lower frequency band (such as tens to hundreds of kHz), while the dominant frequency of surface discharge may be wider or higher.

[0179] Then, the discharge feature vector is input into a pre-trained discharge type classification model to identify the type of partial discharge, including corona discharge, surface discharge, or internal gas gap discharge.

[0180] Among them, in order to effectively classify the discharge feature vector, a variety of mature machine learning or deep learning models can be used. For example:

[0181] Support Vector Machine (SVM): It is effective for small sample and high dimension classification problems, and has strong generalization ability; Random Forest: It is composed of multiple decision trees and has good robustness and is not easy to overfit; Multilayer Perceptron (MLP) or other neural network models: It can learn complex nonlinear relationships between features, and has higher classification accuracy when the data volume is sufficient.

[0182] In actual operation, the central processing unit performs the following operations:

[0183] 1. receiving a real-time multi-modal signal of a combined discharge event from a sensor;

[0184] 2. calculating a six-dimensional discharge feature vector of the event in real time according to the foregoing method ;

[0185] 3. inputting the into a pre-trained classification model solidified in the system as input;

[0186] 4. The model outputs a classification result, such as "surface discharge", according to the decision boundary learned by the model, and can also output a confidence score.

[0187] In some embodiments of the present application, the data acquisition and synchronization unit uses the second pulse signal (PPS) of the global positioning system (GPS) or Beidou satellite navigation system (BDS) to realize nanosecond-level time synchronization of each sensing channel.

[0188] In some embodiments of the present application, the system further comprises a visualization and alarm unit, which is the user interaction core and operation and maintenance decision outlet of the system, and its core function is to convert the abstract data (such as three-dimensional coordinates, discharge type code, and warning value) output by the central processing unit into intuitive and operable operation and maintenance information, and at the same time ensure timely response to faults through a hierarchical alarm mechanism.

[0189] The visualization and alarm unit is configured to:

[0190] 1. mark the calculated discharge position coordinates on the three-dimensional digital model of the power equipment .

[0191] A red dynamic flashing icon (such as a solid circle with a diameter of 5mm, with a flashing frequency of 2 times per second) can be used to mark the discharge position, and the icon color changes with the warning level (emergency level I is red, important level II is orange, general level III is yellow, and attention level IV is blue), which intuitively distinguishes the emergency level of the fault;

[0192] At the same time, mark the key information beside the icon: discharge coordinates (accurate to 0.01 m, such as X:8.50 m, Y:3.20 m, Z:4.10 m), positioning error range (such as ±0.08 m), positioning time (accurate to seconds, such as 2024-08-2514:30:00);

[0193] Support interactive query: click the icon with the mouse, and a "positioning basis details" (such as the sensor number involved in the positioning, the signal arrival time difference of each sensor, and the corrected sound speed value) will pop up, making it easy for operation and maintenance personnel to trace the reliability of the positioning process.

[0194] For example, Figure 7 The figure shows the calibration of the discharge position in the cube model. The gray-white semi-transparent cube in the figure represents the three-dimensional geometric model of the power equipment, and the red five-point star mark point is the partial discharge source position (coordinate example: X=0.6, Y=0.4, Z=0.7, relative to the normalized size of the equipment). This figure directly shows that the positioning result can be accurately placed near the true discharge position in three-dimensional space. The positioning error of this method in the three-dimensional equipment model is controlled within 0.2 m, effectively improving the credibility of the discharge source calibration.

[0195] 2. Display the discharge type and discharge severity assessment results.

[0196] (1) Discharge type display content:

[0197] Core results: discharge type, such as "internal air gap discharge of bushing", "outgoing line corona discharge", and "insulation surface discharge";

[0198] Confidence: the output confidence of the discharge type classification model, which is calculated based on the matching degree of the discharge feature vector and the model training sample, such as the cosine similarity of 6-dimensional features such as ultra-high frequency pulse amplitude and ultrasonic main frequency.

[0199] (2) Discharge severity assessment result display content:

[0200] Based on the above calculation logic of early warning value W = position weight + signal strength weight + trend weight , the following quantitative information is displayed:

[0201] Early warning level: I level (emergency, W≥80 points), II level (important, 60≤W<80 points), III level (general, 40≤W<60 points), IV level (attention, W<40 points);

[0202] Each weight score: such as "position weight =40 points (key area inside the bushing), signal strength weight = 25 points (UHF amplitude 6 pC, AE amplitude 45 dBμV), trend weight = 22 points (discharge frequency increased by 15% within 10 minutes), total warning value W = 87 points (level I)”;

[0203] Trend curve: shows the “discharge frequency-time” and “signal amplitude-time” curves within the last 1 hour (e.g., one data point every 5 minutes), intuitively presenting the discharge development trend, for example, “discharge frequency increased from 5 times / minute to 18 times / minute, showing an accelerating upward trend”.

[0204] 3. When the monitoring index exceeds the preset warning threshold, an alarm is triggered.

[0205] Among them, the monitoring index type: select 3 types of core indicators directly related to discharge hazards, set the threshold in combination with “GB / T11022-2020 High Voltage AC Switching Device and Control Device Standard Common Technical Requirements” and device operation and maintenance experience, for example, as shown in the following table:

[0206] Monitoring indicators Threshold setting logic Examples Discharge frequency Based on the device historical normal data (such as normal operation ≤5 times / minute), the threshold is set to 3 times the normal maximum value Threshold = 15 times / minute Early warning value According to the early warning level, the first level alarm threshold = 80 points, the second level = 60 points First level alarm threshold = 80 points Key area discharge duration If the discharge in the key area (such as the sleeve, winding) lasts more than 30 minutes, an alarm should be triggered to avoid continuous damage to the insulation Threshold = 30 minutes

[0207] Threshold calibration mechanism: support dynamic adjustment of threshold based on device operation time, environmental conditions (such as humidity, temperature). For example, GIS devices running for more than 10 years, the insulation performance decreases, the discharge frequency threshold can be lowered to 12 times / minute, to ensure that the alarm sensitivity of old devices is higher.

[0208] Alarm level and method: level I alarm (emergency): on-site sound and light alarm (continuous buzzing of beeper beside the device, constant red warning light) + remote multi-channel notification (SMS to maintenance personnel, pop-up window on maintenance platform, telephone alarm to dispatch center), notification interval 5 minutes (until confirmation of receipt);

[0209] Level II alarm (important): on-site yellow warning light flashing (1 time / second) + remote SMS + platform pop-up window, notification interval 15 minutes;

[0210] Level III / IV alarm (general / concern): only the maintenance platform displays alarm information, no on-site sound and light to avoid interfering with normal maintenance.

[0211] Alarm confirmation and closed loop: after the alarm is triggered, the maintenance personnel need to “confirm receipt” and “feedback processing results” (such as “scheduled for maintenance” “false alarm (environmental interference)”) in the system, the system automatically records the “alarm triggering-confirmation-processing-closed loop” whole process log (save time ≥1 year), meeting the power industry operation and maintenance traceability requirements.

[0212] For example, Figure 8A warning level color map is shown. Given the warning level for different areas, the color mapping (green / yellow / red) visually represents the risk status. Green areas in the diagram indicate normal operation, yellow areas indicate minor anomalies or cautionary conditions, and red areas indicate alarm conditions with severe discharge risks. This diagram demonstrates that by overlaying warning levels on each area, operations and maintenance personnel can quickly determine the overall health of the equipment and high-risk locations. For example, the red area near (X=7, Y=3) in the diagram indicates that priority attention is required. Example 2

[0213] like Figure 9 As shown, corresponding to the above system embodiment, the present invention also proposes a method for monitoring and locating partial discharge of power equipment based on multimodal sensing, comprising the following steps:

[0214] First, multimodal sensing units, including ultra-high frequency (UHF) and ultrasonic sensor arrays, are deployed around the power equipment to be monitored. These sensors are used to synchronously collect UHF electromagnetic wave signals and ultrasonic signals generated by partial discharge.

[0215] The data acquisition and synchronization unit uses Beidou / GPS second pulse signals to achieve nanosecond-level time synchronization of each sensor channel, and performs high-precision synchronous sampling and time stamping on the signals to generate synchronized multimodal data;

[0216] After receiving the data, the central processing unit executes a multimodal data fusion and identification algorithm. Using the ultra-high frequency signal as a trigger, it searches for valid ultrasonic signals within a preset time window and identifies true joint discharge events through spatiotemporal correlation.

[0217] After successful recognition, a joint positioning algorithm is executed. First, the preliminary discharge area is calculated based on the arrival time difference of the UHF signal. Then, the sound speed is corrected based on the temperature data provided by the thermal imaging unit. The arrival time difference of the ultrasonic signal is used to conduct a precise search within the preliminary area. Finally, the three-dimensional coordinates of the discharge location are calculated.

[0218] At the same time, discharge feature vectors are extracted from multimodal signals and input into a pre-trained classification model to identify discharge types;

[0219] The final result is marked with the location, display type and severity on the three-dimensional model of the equipment through the visualization unit, and a graded alarm is triggered when the threshold is exceeded. Example 3

[0220] Corresponding to the above embodiment, the present invention further provides an electronic device.

[0221] like Figure 10As shown is a structural schematic diagram of an electronic device in the present application, the electronic device 200 comprises a processor 201 and a memory 203. Wherein, the processor 201 and the memory 203 are connected, such as connected through a bus 202. Optionally, the electronic device 200 can further comprise a transceiver 204. It needs to be explained that the transceiver 204 is not limited to one in the actual application, and the structure of the electronic device 200 does not constitute a limitation to the embodiments of the present application.

[0222] The processor 201 can be a CPU, a general processor, a DSP, an ASIC, an FPGA or other programmable logic device, a transistor logic device, a hardware component or any combination thereof. It can realize or execute various exemplary logic blocks, modules and circuits described in combination with the disclosure. The processor 201 can also be a combination of computing functions, such as one or more microprocessor combinations, combinations of DSP and microprocessor, etc.

[0223] The bus 202 can include a channel for transmitting information between the above-mentioned components. The bus 202 can be a PCI bus or an EISA bus, etc. The bus 202 can be divided into an address bus, a data bus, a control bus, etc. For the convenience of representation, Figure 10 In the figure, only one thick line is used to represent, but it does not mean that there is only one bus or only one type of bus.

[0224] The memory 203 is used to store the computer program corresponding to the power equipment partial discharge monitoring and positioning method based on multi-modal sensing of the above-mentioned embodiments of the present application, which is controlled and executed by the processor 201. The processor 201 is used to execute the computer program stored in the memory 203 to realize the content shown in the foregoing method embodiments.

[0225] Wherein, the electronic device 200 includes but is not limited to: notebook computers, PAD (tablet computers) and other mobile terminals, and fixed terminals such as desktop computers and the like. Figure 10 The electronic device 200 shown is only an example, which should not bring any limitation to the function and use range of the embodiments of the present application.

[0226] It is to be appreciated that the above description and the examples that follow are intended to be illustrative only and that changes can be made to the description and examples without departing from the scope of the application. Note also that the use of particular brand names in the description is solely for illustration and should not be construed as an endorsement of such brands.

[0227] It should be understood that aspects of the application can be implemented in hardware, software, firmware or combinations thereof. In the embodiments described above, various steps or methods can be implemented, in part, or in whole, in software and / or firmware that are stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, and in another embodiment, any of the following techniques, which are well known in the art, can be used to implement the functions of the application: a discrete logic circuit having logic gates for implementing logic functions upon data signals, an application specific integrated circuit having appropriate combinational logic gates, a programmable gate array (PGA), a field programmable gate array (FPGA), and / or the like.

[0228] In the description of the present application, the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" are intended to mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present application. The appearances of the above terms in various places in the specification are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0229] Furthermore, the terms "first", "second", etc. are used only for descriptive purposes and do not connote or imply relative importance or a quantity of the indicated technical features. Thus, a feature defined with "first", "second", etc. can include at least one of the features implicitly or explicitly. In the description of the present application, the meaning of "a plurality" is at least two, for example, two, three, etc., unless otherwise specifically defined.

[0230] Although the embodiments of the present application have been shown and described above, it is understood that the above-described embodiments are exemplary and are not to be construed as limiting the present application, and that changes, modifications, substitutions and variations can be made by those skilled in the art without departing from the scope of the present application.

Claims

1. A partial discharge monitoring and positioning system for power equipment based on multimodal sensing, characterized in that: include: The multimodal sensing unit is configured to be arranged around the power equipment to be monitored and is used to synchronously collect multiple physical signals generated by partial discharge. The multimodal sensing unit includes at least: UHF sensor array, used to receive UHF electromagnetic wave signals generated by partial discharge; An ultrasonic sensor array, used for receiving ultrasonic signals generated by partial discharge; a data acquisition and synchronization unit, connected to the multimodal sensing unit, for performing high-precision synchronous sampling and time-stamping on the ultra-high frequency electromagnetic wave signal and the ultrasonic signal to generate synchronized multimodal data; The central processing unit is connected to the data acquisition and synchronization unit and is configured to: receiving the synchronized multimodal data; executing a multimodal data fusion identification algorithm to identify a true joint discharge event based on the spatiotemporal correlation between the ultra-high frequency electromagnetic wave signal and the ultrasonic signal; After the joint discharge event is identified, a joint positioning algorithm is executed to calculate the discharge position coordinates of the joint discharge event in three-dimensional space.

2. The system according to claim 1, wherein: The multimodal sensing unit further comprises: Thermal imaging sensor unit, used to obtain temperature distribution data on the surface of the power equipment to be monitored; The central processing unit is further configured to compensate for a propagation speed of the ultrasonic signal using the temperature distribution data.

3. The system according to claim 1, wherein: The definition of the spatiotemporal correlation is: In the preset time window The UHF sensor array and the ultrasonic sensor array both detect energy exceeding their respective preset energy thresholds. and signal.

4. The system according to claim 1, wherein: The specific steps of the central processing unit executing the multimodal data fusion identification algorithm include: Step 1: Perform preliminary detection on the UHF electromagnetic wave signal to identify candidate discharge pulses; Step 2: Based on the arrival time of each candidate discharge pulse, Searching whether there is a valid ultrasonic signal collected by the ultrasonic sensor array; Step 3: If so, the corresponding UHF signal and ultrasonic signal of this group are confirmed as a joint discharge event.

5. The system according to claim 1 or 2, characterized in that The joint positioning algorithm includes: Based on the arrival time difference of the signal received by each sensor in the UHF sensor array , calculate a preliminary discharge location area ; Based on the arrival time difference of the signal received by each sensor in the ultrasonic sensor array , and combined with the sound velocity corrected by the temperature data provided by the thermal imaging sensor unit , in the initial discharge position area Perform precise search within to obtain the final discharge position coordinates .

6. The system according to claim 5, characterized in that The corrected sound speed Determined by the following formula: in, is the corrected speed of sound, The propagation speed of ultrasonic waves in insulating media at 0 degrees Celsius is is the Celsius temperature of the discharge position area measured by the thermal imaging sensor unit.

7. The system according to claim 1, wherein: The central processing unit is further configured to: extracting a discharge feature vector from the multimodal signal corresponding to the joint discharge event; The discharge feature vector is input into a pre-trained discharge type classification model to identify the type of partial discharge, which includes corona discharge, surface discharge or internal air gap discharge.

8. The system according to claim 7, characterized in that The discharge characteristic vector includes: The pulse amplitude, pulse rise time, and pulse width of UHF signals, as well as the energy, duration, and main frequency of ultrasonic signals.

9. The system according to claim 1, wherein: The data acquisition and synchronization unit uses the second pulse signal of the global positioning system or the Beidou satellite navigation system to achieve nanosecond-level time synchronization of each sensor channel.

10. The system according to claim 1, wherein: The system also includes a visualization and alarm unit configured to: Mark the calculated discharge position coordinates on the 3D digital model of the power equipment ; Display discharge type and discharge severity assessment results; When the monitoring indicator exceeds the preset alarm threshold, an alarm is triggered.

Images