A method and system for partial discharge detection based on diamond NV color centers

Abstract

This invention discloses a method and system for partial discharge detection based on diamond NV centers. The method involves deploying a probe integrated with diamond NV centers at a predetermined detection location on operating power equipment. A pulse sequence parameter is set through a timing control system, and a laser-microwave pulse sequence is cyclically run with the power frequency voltage phase as the time reference. Simultaneously, NV center fluorescence signals are acquired. Multiple measurements are performed by changing the timing parameters of the pulse sequence to obtain magnetic field response signals at different frequencies, achieving frequency domain detection of partial discharge signals. This enables selective measurement of high-frequency magnetic signals from partial discharges at different frequencies. This invention has the advantages of high sensitivity, adjustable frequency, and non-invasive installation, and is suitable for online monitoring and defect diagnosis of partial discharge in equipment such as power transformers, gas-insulated switchgear, and power cables.

Description

Technical Field

[0001] This invention relates to the field of online monitoring of insulation status of power equipment and quantum precision measurement technology, specifically a method and system for detecting partial discharge based on diamond NV color centers. Background Technology

[0002] Partial discharge (PD) is a major characteristic of insulation degradation in high-voltage power equipment. Online monitoring of PD is crucial for assessing the insulation health status, enabling condition-based maintenance, and preventing insulation breakdown accidents. Currently, the main technical methods for online partial discharge monitoring include ultra-high frequency (UHF) methods, ultrasonic methods, transient ground voltage methods, and high-frequency current transformer methods, all of which are widely used in engineering practice. However, all of these existing technologies have limitations to varying degrees: (i) The ultra-high frequency (UHF) method identifies discharges by detecting electromagnetic signals in the 300MHz to 3GHz frequency band emitted by partial discharges, offering strong resistance to low-frequency interference. However, its detection effectiveness heavily relies on the line-of-sight propagation path between the sensor and the discharge point. During on-site installation, the sensor must be attached to the equipment's insulating basin or the manhole of a gas-insulated switch (GIS), limiting installation location and making it difficult to achieve comprehensive coverage of all equipment in the station. Furthermore, UHF signals attenuate significantly during propagation within the equipment, leading to a sharp drop in detection sensitivity for weak discharge signals or discharge events far from the sensor's measuring point. Existing UHF sensors operate on a fixed frequency band, making it impossible to adaptively adjust the detection frequency according to the discharge type, resulting in insufficient selective identification capability for different discharge characteristics.

[0003] (II) The ultrasonic method locates and diagnoses partial discharges by detecting the acoustic signals generated. Its advantage lies in its immunity to electromagnetic interference and its ability to spatially locate the discharge source. However, ultrasonic waves attenuate significantly in gaseous, solid, and oil media. For discharges occurring inside oil-immersed transformers, the ultrasonic signal must penetrate the tank wall to be received by external sensors, resulting in a substantial signal attenuation and limited sensitivity. Furthermore, environmental noise at the substation site (such as equipment vibration and mechanical noise caused by electromagnetic forces) overlaps with the frequency band of the partial discharge ultrasonic signal, making effective differentiation through frequency domain filtering difficult and highlighting the low signal-to-noise ratio. The response frequency of ultrasonic sensors is typically in the range of 20kHz to 200kHz, limiting their ability to respond to broadband acoustic signals generated by nanosecond-level pulse discharges.

[0004] (III) The transient ground voltage method utilizes the transient ground voltage signal induced on the surface of the equipment casing by the electromagnetic waves generated by partial discharge for detection. It is simple to operate and suitable for live-line inspection of metal-enclosed equipment such as switchgear. However, this method is easily affected by grounding system impedance, equipment surface condition, and coupling from external interference sources, resulting in poor measurement repeatability and difficulty in calibrating the discharge quantity and assessing its severity. Especially in high-voltage equipment, where the casing potential distribution is complex, there is a lack of quantitative correlation between the amplitude of the transient ground voltage signal and the discharge intensity. Currently, it is mainly used for qualitative screening and lacks the ability to analyze insulation condition trends and identify defect types.

[0005] (iv) The high-frequency current transformer method couples partial discharge pulse current signals using a high-frequency current transformer connected to the equipment grounding wire or neutral point down conductor. This method offers advantages such as high sensitivity and easy installation. However, it is susceptible to interference from discharge pulses from other equipment in the grounding system, high-frequency switching interference from power electronic devices, and external electromagnetic radiation. The interference components in the field signal are complex, requiring experienced technicians to manually identify the waveform, making automated identification and continuous online monitoring difficult. Furthermore, the detection bandwidth of the high-frequency current transformer is limited by the core material and number of turns, and its high-frequency response has a cutoff frequency, resulting in distortion of fast pulse discharge signals with steep rising edges.

[0006] In recent years, quantum magnetic measurement technology based on diamond nitrogen-vacancy (NV) centers has made significant progress in the field of precision measurement. As a room-temperature solid-state quantum spin system, diamond NV centers possess the characteristics of laser-induced spin polarization, microwave-manipulated coherent states, and fluorescence readout spin states. They exhibit femtotes-level sensitivity and nanosecond-level time response to external magnetic fields, providing a physical basis for the direct detection of high-frequency pulsed magnetic fields from partial discharges. Existing literature (such as Wolf et al., Physical Review X, 2015, 5(4): 041001; Mizuno et al., Scientific Reports, 2020, 10(1)) has reported the application of diamond NV centers in alternating magnetic field measurement, but mainly focuses on precision measurement scenarios of continuous waves or low-frequency alternating magnetic fields such as biomagnetic fields and power frequency magnetic fields, and has not yet addressed the detection of transient pulsed magnetic fields generated by partial discharges in power equipment. Xie Yijin (Doctoral Dissertation, University of Science and Technology of China, 2020) studied magnetic measurement methods based on diamond NV color center ensembles, but no reports have been found on the engineering adaptation of pulse magnetic measurement sequences to power system power frequency phase synchronization, and tunable detection of broadband characteristics of partial discharge.

[0007] Existing magnetic measurement technology based on diamond NV centers faces the following unresolved technical challenges if directly applied to online monitoring of partial discharge: First, the pulsed magnetic field signals generated by partial discharge exhibit transient (nanosecond to microsecond) and wide-spectrum (covering several MHz to several GHz) characteristics, making it impossible for traditional continuous wave or single-frequency lock-in amplification measurement methods to effectively capture such transient signals. Second, substation sites experience strong background noise, including power frequency magnetic fields (50Hz / 60Hz) and their harmonics, as well as low-frequency magnetic disturbances caused by equipment vibration. The amplitude of this background noise far exceeds the weak high-frequency magnetic field signals radiated by partial discharge. Therefore, how to effectively capture these signals in the context of strong low-frequency noise is a significant challenge. Extracting high-frequency weak magnetic signals is a key technological bottleneck; third, the frequency characteristics of pulse magnetic fields corresponding to different discharge types (point discharge, particle discharge, floating potential discharge, air gap discharge, etc.) vary significantly. The detection frequency of existing NV color center magnetic measurement methods is fixed, making it impossible to selectively detect and identify features based on the discharge type; fourth, online monitoring requires the system to operate continuously for a long time and adapt to complex electromagnetic environments and installation conditions on site. Existing magnetic measurement devices based on NV color centers mostly adopt laboratory benchtop structures with separate laser and microwave systems, resulting in large volumes and lacking the feasibility and reliability for engineering-scale online monitoring.

[0008] To address the shortcomings of existing technologies, this invention proposes a partial discharge pulse magnetic measurement method based on diamond NV color centers and integrating an acousto-optic modulator and a microwave switch. This method achieves high-speed synchronous control of laser and microwave pulses through a timing control system. Combined with a spin echo or dynamic decoupling sequence with adjustable pulse sequence parameters, it effectively filters out low-frequency magnetic noise while selectively detecting high-frequency magnetic signals from partial discharges at different frequencies. This provides a novel quantum precision measurement solution for non-invasive, highly sensitive, and frequency-tunable online monitoring of partial discharges in power equipment. Summary of the Invention

[0009] The purpose of this invention is to provide a partial discharge detection method and system based on diamond NV centers, in order to solve the problem mentioned in the background art of the lack of a quantum precision measurement solution for non-invasive, high-sensitivity, frequency-tunable online monitoring of partial discharge in power equipment.

[0010] To achieve the above objectives, the present invention provides the following technical solution: A partial discharge detection method based on diamond NV color centers includes the following steps: S1: Deploy the probe, which integrates diamond NV color centers, at the predetermined detection location of the operating power equipment; S2: The pulse sequence parameters are set through the timing control system. The pulse sequence is used to coherently manipulate the spin state of the NV color center and detect the external magnetic field. S3: Using the power frequency voltage phase as the time reference, the laser-microwave pulse sequence is cyclically run to synchronously acquire NV color center fluorescence signals; S4: Change the timing parameters of the pulse sequence and perform multiple measurements to obtain magnetic field response signals at different frequencies, thereby realizing frequency domain detection of partial discharge signals.

[0011] As a further technical solution, the pulse sequence is a dynamic decoupling sequence, and the method for setting the pulse sequence parameters in step S2 specifically includes: N equally spaced microwave π pulses are inserted between the initial microwave π / 2 pulse and the final microwave π / 2 pulse, where N≥1; The time interval τ between adjacent microwave π pulses is adjusted by the timing control system to tune the equivalent center detection frequency f of the dynamic decoupling sequence, where f = 1 / (2τ).

[0012] As a further technical solution, the pulse sequence is a spin echo sequence, and the timing control process for cyclically running the laser-microwave pulse sequence in step S3 includes: Turn on the laser pulse and initialize the NV color center to the ground state ms=0. Turn off the laser and apply a microwave π / 2 pulse to prepare a spin coherent superposition state; After the first free evolution time, a microwave π pulse is applied to induce spin re-aggregation; After a second free evolution time, a second microwave π / 2 pulse is applied to convert the phase information into population changes; Reactivate the laser pulse and read the fluorescence signal.

[0013] As a further technical solution, in step S3, the laser-microwave pulse sequence cycle uses the power frequency voltage phase as the synchronization reference, triggers multiple measurement sequences in each power frequency cycle, and precisely controls the opening and closing of the acousto-optic modulator and microwave switch through the timing control system to achieve nanosecond-level synchronous control of laser pulses and microwave pulses.

[0014] As a further technical solution, the method following step S4 also includes: S5: Convert the acquired fluorescence signals from multiple frequency channels into a spectral response feature vector of the high-frequency magnetic field of partial discharge; S6: Input the spectral response feature vector into the pre-trained neural network model; S7: The neural network model outputs the defect type of partial discharge and its confidence level to achieve online diagnosis of insulation defects.

[0015] As a further technical solution, the neural network model is trained using a partial discharge spectral feature library of different defect types as training samples. The defect types include, but are not limited to, tip discharge, particle discharge, floating potential discharge, and air gap discharge.

[0016] A partial discharge detection system based on diamond NV centers for implementing the method, characterized in that it comprises: The laser excitation and modulation unit is used to generate and modulate laser pulses to perform spin initialization and fluorescence readout of the NV color center; The microwave pulse control unit is used to generate and output microwave pulses in a set sequence to coherently control the spin state of the NV color center. The magnetic field biasing unit is used to generate a constant bias magnetic field in the region where the diamond is located, so as to cause the ground state of the NV color center to undergo Zeeman splitting and select the working energy level; The fluorescence readout and acquisition unit is used to collect the fluorescence generated by the NV color center and convert it into an electrical signal; The timing control system is used to receive the power frequency phase synchronization signal and control the timing synchronization between the laser excitation and modulation unit, the microwave pulse control unit, and the fluorescence readout and acquisition unit.

[0017] As a further technical solution, the laser excitation and control unit includes a 532nm laser and an acousto-optic modulator. The acousto-optic modulator is driven by a digital pulse signal provided by the timing control system to achieve rapid switching and power control of the laser pulse. The microwave pulse control unit includes an arbitrary waveform generator, a microwave source, a microwave switch, and a microwave radiation structure. The microwave switch is triggered by the timing control system to gate and shape the microwave pulse.

[0018] As a further technical solution, the probe is a fully solid-state miniaturized structure, which internally encapsulates a diamond single crystal, the magnetic field bias unit, and the microwave radiation structure. The probe is connected to an external host via a cable. The external host integrates the laser excitation and control unit, the microwave pulse control unit, the fluorescence readout and acquisition unit, and the timing control system.

[0019] As a further technical solution, the timing control system adjusts the time interval τ between adjacent microwave π pulses in the dynamic decoupling sequence to change the equivalent center detection frequency of the sequence, so as to adapt to the characteristic frequency components corresponding to different partial discharge types and achieve selective detection of high-frequency magnetic signals of partial discharge.

[0020] Compared with the prior art, the beneficial effects of the present invention are: 1. High sensitivity for detecting high-frequency weak magnetic signals. By effectively filtering out static and low-frequency magnetic noise through spin echo sequences and extending the spin coherence time, combined with the resonant enhancement response of the dynamic decoupling sequence to the alternating magnetic field at a specific frequency, weak pulsed magnetic field signals generated by partial discharge can be extracted in environments with strong electromagnetic interference. 2. Flexible and adjustable detection frequency. By adjusting the time interval between adjacent π pulses in the dynamic decoupling sequence through the timing control system, the detection frequency of the equivalent center of the sequence can be changed. It can cover the typical spectrum range of partial discharge without changing the hardware and adapt to the characteristic frequencies corresponding to different discharge types and propagation paths. 3. Non-invasive online continuous monitoring. The probe has a fully solid-state miniaturized structure, which is attached to the surface of the equipment shell without electrical contact. With continuous pulse cyclic acquisition synchronized with power frequency phase, it can operate for a long time without power interruption. 4. Capable of defect type identification. Partial discharge spectral response characteristics are obtained through multi-frequency sweep measurements, and the discharge type is automatically identified by combining this with a neural network model, providing a quantitative basis for insulation diagnosis. Attached Figure Description

[0021] Figure 1 This is a flowchart of a partial discharge detection method based on diamond NV color centers according to the present invention; Figure 2 This is a system architecture diagram of a partial discharge detection system based on diamond NV color centers according to the present invention; Figure 3 This is a schematic diagram of the spin echo sequence and dynamic decoupling sequence in the embodiment; Figure 4 This is a schematic diagram of the laser excitation and control unit of a partial discharge detection system based on diamond NV color centers according to the present invention; Figure 5 This is a schematic diagram of the microwave pulse control unit of a partial discharge detection system based on diamond NV color centers according to the present invention. Detailed Implementation

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

[0023] Please see Figure 1 The present invention provides a partial discharge detection method based on diamond NV centers, comprising the following steps: S1: Probe Deployment Steps. Deploy the probe, integrated with a diamond NV color center, at the designated testing location on the operating power equipment. Specifically, the testing personnel select the measurement point on the outer casing surface near the high-voltage conductor or a weak point in the insulation, based on the type of equipment under test (e.g., power transformer, gas-insulated switchgear, or power cable) and the characteristics of the insulation structure. The probe adopts a fully solid-state miniaturized packaging structure and is fixed to the equipment casing surface by magnetic mounting, clamps, or adhesive bonding. The distance between the diamond single crystal inside the probe and the inner wall of the equipment casing is controlled within 5mm to ensure that the pulsed magnetic field signal generated by partial discharge can be effectively coupled to the NV color center. After the probe is deployed, connect the probe to the external host via a shielded cable.

[0024] S2: Pulse Sequence Parameter Setting Steps. The pulse sequence parameters are set through the timing control system. This pulse sequence is used for coherent manipulation of the NV color center spin state and detection of the external magnetic field. Based on the on-site testing requirements and the historical characteristics of partial discharge in the device under test, the operator selects the pulse sequence type—spin echo sequence or dynamic decoupling sequence—through the host computer software. The timing control system automatically configures parameters such as laser pulse width, microwave pulse width, pulse interval, and number of cycles according to the selected sequence type. If the dynamic decoupling sequence is selected, the number N of microwave π pulses inserted between the initial and final π / 2 pulses, as well as the time interval τ between adjacent π pulses, must also be set to determine the equivalent center detection frequency of the sequence.

[0025] S3: Synchronous Acquisition Steps. Using the power frequency voltage phase as the time reference, a laser-microwave pulse sequence is cyclically run to synchronously acquire the fluorescence signal of the NV color center. The timing control system receives the power frequency phase synchronization signal (50Hz or 60Hz) provided by the voltage transformer of the device under test or the station's synchronous clock source, using the power frequency cycle as the basic time frame. Within each power frequency cycle, the timing control system triggers multiple measurement sequences according to a preset delay time. Each measurement sequence includes a complete laser initialization, microwave manipulation, and fluorescence readout process. During the laser pulse activation period, the acousto-optic modulator chops the 532nm continuous laser into laser pulses with adjustable width (typically 1μs~10μs), irradiating the diamond NV color center to complete spin initialization. The microwave switch is activated during the microwave manipulation window, applying microwave pulses to the microwave radiation structure inside the probe to achieve coherent manipulation of the NV color center spin state. The fluorescence signal is acquired by a high-speed photodetector during the laser readout window, and stored in a buffer after analog-to-digital conversion.

[0026] S4: Frequency Domain Scanning Measurement Steps. Multiple measurements are performed by changing the timing parameters of the pulse sequence to obtain magnetic field response signals at different frequencies, thus achieving frequency domain detection of the partial discharge signal. For the same measurement point, the timing control system sequentially changes the timing parameters of the pulse sequence according to a preset frequency sweep parameter table—when using a dynamic decoupling sequence, the time interval τ between adjacent π pulses is gradually adjusted so that the equivalent center detection frequency f=1 / (2τ) traverses the typical spectrum range of partial discharge (e.g., from 1MHz to 1GHz) according to a set step size. Fluorescence data is continuously collected for a preset duration (e.g., 100 power frequency cycles) at each frequency point, and the magnetic field response amplitude of that frequency channel is recorded after phase-locked amplification or cumulative averaging. After completing the scan of all frequency points, the partial discharge frequency domain response curve of that measurement point is obtained.

[0027] Please see Figure 3 When the pulse sequence is a dynamic decoupling sequence, the method for setting the pulse sequence parameters in step S2 is as follows: N equally spaced microwave π pulses (labeled π) are inserted between the initial microwave π / 2 pulse (labeled π / 2) and the final microwave π / 2 pulse (labeled π / 2), where N ≥ 1. The time interval between adjacent microwave π pulses is denoted as τ, and the equivalent center detection frequency f of the sequence is determined by the formula f = 1 / (2τ). By adjusting the value of τ through the timing control system, the detection frequency can be continuously tuned. When τ matches the half-cycle of a certain high-frequency component of the partial discharge, this frequency component is coherently accumulated and enhanced under the action of multiple π pulses in the sequence, while noise components deviating from this frequency are effectively suppressed, thereby achieving frequency-selective detection. Figure 3 The timing structure of the spin echo sequence is also shown for comparison. The spin echo sequence consists of a π / 2 pulse, a π pulse, and a final π / 2 pulse, and is suitable for wideband low-frequency noise filtering scenarios.

[0028] When the pulse sequence is a spin echo sequence, the timing control process for cyclically running the laser-microwave pulse sequence in step S3 is as follows: First, the timing control system sends an activation signal to the acousto-optic modulator, and a 532nm laser pulse irradiates the diamond NV color center for a duration of 2μs~5μs, polarizing the electron spin of the NV color center from an arbitrary initial state to the ground state ms=0, thus completing the spin initialization. Secondly, the acousto-optic modulator is turned off to cut off the laser, and the timing control system sends a trigger signal to the microwave switch. The microwave pulse manipulation unit generates a microwave π / 2 pulse (pulse width approximately 50ns~200ns), which acts on the NV color center, preparing its spin state from the ms=0 state to a superposition state (|0 ms=0 ns ... +|-1 ) / √2, forming a coherent superposition; Then, the microwave switch is turned off, and the system enters the first free evolution stage, which lasts for τ_free (typically 0.5μs~10μs). During this stage, the pulsed magnetic field generated by the partial discharge couples with the spin of the NV color center, accumulating a phase increment in the spin phase that is proportional to the magnetic field amplitude. Subsequently, the microwave switch is turned on again, and a microwave π pulse (with a pulse width approximately twice that of a π / 2 pulse) is applied to achieve spin state reversal, thereby compensating for the phase drift caused by low-frequency noise in subsequent evolution (i.e., the spin echo effect). After a second free evolution stage with a time equal to that of the first free evolution, the microwave switch is turned on and a second microwave π / 2 pulse is applied, which converts the spin phase information into population differences in different spin states. Finally, the acousto-optic modulator restarts the laser pulse, inducing fluorescence from the NV color center. The fluorescence intensity is proportional to the population of the ms=0 state, thus converting the magnetic field signal into a readable optical signal. Through this sequence, static and slowly varying low-frequency magnetic noise is effectively filtered out, retaining only transient or high-frequency magnetic field signals that match the sequence timing window.

[0029] Please see Figure 2 In a preferred embodiment, in step S3, the laser-microwave pulse sequence is cyclically synchronized with the power frequency voltage phase as the synchronization reference. The timing control system receives an externally input power frequency synchronization signal (such as a substation synchronization clock or a voltage transformer secondary side signal), and triggers multiple measurement sequences within each power frequency cycle according to a preset phase delay (such as 0°, 45°, 90°, etc.). The timing control system internally employs a high-precision field-programmable gate array (FPGA) as the core timing unit, with a timing resolution better than 1ns. This allows for precise control of the relative delay between the acousto-optic modulator drive signal and the microwave switch trigger signal, ensuring that the timing error between the laser pulse and the microwave pulse is less than 5ns. The rise / fall time of the acousto-optic modulator is approximately 10ns~50ns, and the switching time of the microwave switch is approximately 2ns~5ns. Together, they achieve nanosecond-level precision laser-microwave pulse synchronization control, meeting the requirements for detecting nanosecond-level partial discharge pulse signals.

[0030] Step S4 is followed by a defect type intelligent identification step, as detailed below: S5: Construction of Spectral Feature Vector. The fluorescence signals from multiple frequency channels acquired in step S4 are preprocessed by denoising and normalization to convert them into spectral response feature vectors of the high-frequency magnetic field of partial discharge. Specifically, assuming that the sweep frequency measurement obtains fluorescence modulation amplitude data [A1, A2, …, AM] for M frequency channels, with corresponding detection frequencies [f1, f2, …, fM], these data are combined into an M-dimensional feature vector X = [A1, A2, …, AM]^T. To further improve the feature discrimination, the feature vector can also be normalized, that is, the amplitude of each channel is divided by the maximum amplitude of all channels at the measurement point, so that the value range of each element of the feature vector is normalized to the interval [0,1].

[0031] S6: Neural Network Inference. The spectral response feature vector X is input into a pre-trained neural network model. This neural network model is a multilayer perceptron (MLP) or a one-dimensional convolutional neural network (1D-CNN), with the number of input layer nodes equal to the number of frequency channels M, the number of output layer nodes equal to the number of defect type categories K, 2-3 hidden layers, the ReLU activation function, and the Softmax activation function in the output layer to output the probability distribution of each category.

[0032] S7: Defect Type Output. After forward computation, the neural network model outputs the defect type of partial discharge and its confidence level (probability value), enabling online diagnosis of insulation defects. The system uses the category with the highest confidence level as the diagnostic result and displays the confidence percentage of that result. If the highest confidence level is lower than a preset threshold (e.g., 0.7), a "Not Identified" message is output, suggesting manual review.

[0033] Further, the training process of the neural network model described in step S6 is as follows: Beforehand, under laboratory or field testing conditions, a large amount of partial discharge frequency domain response data is collected for four typical defect types: tip discharge, particle discharge, floating potential discharge, and air gap discharge. A spectral feature sample library labeled with defect type tags is constructed. Each sample contains fluorescence modulation amplitude vectors of M frequency channels and their corresponding defect type labels. The sample library is divided into a training set, a validation set, and a test set in a 7:1.5:1.5 ratio. The neural network is trained using the cross-entropy loss function and the Adam optimizer until the accuracy of the validation set converges. After training, the network weights and bias parameters are stored in the system memory for use during the online monitoring phase.

[0034] Please see Figure 2 The present invention also provides a partial discharge detection system based on diamond NV centers for implementing the above method. The system includes: a laser excitation and control unit, a microwave pulse manipulation unit, a magnetic field biasing unit, a fluorescence readout and acquisition unit, a timing control system, and a probe. Wherein: The laser excitation and modulation unit is used to generate and modulate laser pulses for spin initialization and fluorescence readout of the NV color centers. For example... Figure 4 As shown, this unit consists of a 532nm continuous-wave laser (output power adjustable from 0 to 2W), an acousto-optic modulator (AOM), an optical fiber coupler, and a spatial optical path. The continuous-wave laser output from the 532nm laser is collimated and then incident on the acousto-optic modulator. The acousto-optic modulator is driven by a digital pulse signal provided by a timing control system. By controlling the driving power and pulse width, the laser pulses are rapidly switched (switching time < 50ns) and the power is adjusted. The first-order diffracted light output from the acousto-optic modulator is coupled through an optical fiber or spatial optical path and enters the probe, illuminating the diamond single crystal surface. The spot diameter is approximately 50μm to 200μm.

[0035] The microwave pulse manipulation unit generates and outputs microwave pulses in a set sequence to coherently manipulate the spin states of the NV color center. For example... Figure 5 As shown, this unit consists of an arbitrary waveform generator (AWG), a microwave signal source, a microwave switch, a power amplifier, and a microwave radiation structure. The AWG generates an intermediate frequency pulse envelope waveform (frequency range DC~500MHz) triggered by a timing control system. The microwave signal source generates a local oscillator microwave carrier signal (frequency range 2.6GHz~3.0GHz). Both are up-converted by a mixer to produce the target microwave pulse frequency. The microwave switch (switching time <5ns) is controlled by the timing control system trigger signal, performing nanosecond-level precision gating and shaping of the microwave pulse to ensure that the microwave pulse is output only within a specified control window. The power amplifier amplifies the microwave pulse power to +30dBm~+37dBm, which is then fed into the microwave radiation structure inside the probe via a coaxial cable. The microwave radiation structure can be in the form of a microstrip line, coplanar waveguide, or ring resonator, mounted near the diamond single crystal surface with a spacing controlled within 0.5mm to generate a uniform and sufficiently strong microwave magnetic field to drive the spin flipping of the NV color center.

[0036] The magnetic field biasing unit is used to generate a constant bias magnetic field in the region where the diamond is located. This unit is implemented using a small neodymium iron boron permanent magnet or a miniature electromagnetic coil, installed near the diamond single crystal inside the probe. The permanent magnet generates a static, uniform magnetic field of approximately 100 Gauss to 500 Gauss in the diamond region, causing Zeeman splitting of the ground state ms=±1 energy level of the NV color center (splitting distance approximately 2.8 MHz / Gauss). By adjusting the position and orientation of the permanent magnet, a pair of transition energy levels (ms=0) is selected. ms=-1 or ms=0 The resonant frequency (ms=+1) is used as the operating frequency for microwave control and magnetic field sensing, and its corresponding frequency is usually set in the range of 2.87GHz±Δ.

[0037] The fluorescence readout and acquisition unit is used to collect the fluorescence generated by the excited NV color center and convert it into an electrical signal. This unit consists of an objective lens, a dichroic mirror, a filter, and a high-speed photodetector. The red fluorescence (center wavelength approximately 637 nm) generated by the diamond NV color center under 532 nm laser excitation is collected by the objective lens inside the probe. After reflection by the dichroic mirror, the residual 532 nm excitation light is filtered out by a long-pass filter (cutoff wavelength > 600 nm). Finally, it is converted into an electrical signal by a high-speed photodetector (bandwidth > 10 MHz, such as an avalanche photodiode or photomultiplier tube). The electrical signal is pre-amplified and then sent to a high-speed acquisition card (sampling rate > 100 MS / s) for analog-to-digital conversion, or sent to a lock-in amplifier for synchronous demodulation to extract the weak fluorescence changes that are time-dependent on the pulse sequence.

[0038] The timing control system, as the core synchronization unit of the entire system, receives the power frequency phase synchronization signal and controls the timing synchronization between the aforementioned units. Based on an FPGA architecture, the timing control system integrates a high-precision clock source (frequency stability better than 1ppm), a multi-channel digital delay generator (delay resolution better than 1ns), and a programmable pulse sequence generator. The timing control system receives the power frequency phase synchronization signal (50Hz / 60Hz) from the voltage transformer of the device under test, uses its zero-crossing point as a reference for phase-locked loop frequency multiplication, and generates the master clock for the entire measurement sequence. Based on this, the system's internal counter outputs the following three synchronization signals according to a preset time parameter table: a laser pulse sequence driving the acousto-optic modulator, a microwave gating signal triggering the microwave switch, and a sample-and-hold signal triggering the high-speed acquisition card. The relative delay between the three signals can be adjusted online via the host computer software to adapt to the timing requirements of different pulse sequences.

[0039] The probe is a fully solid-state miniaturized structure, internally encapsulating a diamond single crystal, the magnetic field biasing unit, and the microwave radiation structure. The diamond single crystal is synthesized by electron-grade chemical vapor deposition (CVD), with dimensions of approximately 2mm × 2mm × 0.5mm and an internal NV color center concentration of approximately 10^14 / cm^3 to 10^16 / cm^3. The probe housing is made of non-magnetic stainless steel or polyetheretherketone (PEEK) to avoid interference with the measured magnetic field. The probe connects to an external host via a multi-core shielded cable, which includes a laser fiber, a microwave coaxial cable, a power supply line, and a fluorescence signal line. The external host integrates the laser and acousto-optic modulator driver in the laser excitation and control unit, the microwave source and switch in the microwave pulse control unit, the detector and acquisition card in the fluorescence readout and acquisition unit, and the timing control system, forming an integrated portable measurement host.

[0040] Please continue reading. Figure 2 The signal flow and data flow of the entire system are as follows: Timing control system ( Figure 2The laser excitation and control unit (marked "5") receives the power frequency synchronization signal (marked "6") and sends synchronization trigger signals to the laser excitation and control unit (marked "1") and the microwave pulse control unit (marked "2"), respectively. The laser pulse generated by the laser excitation and control unit is transmitted via optical fiber to the probe (marked "9"), irradiating the diamond NV color center; the microwave pulse generated by the microwave pulse control unit is transmitted via coaxial cable to the microwave radiation structure inside the probe, controlling the spin state of the NV color center. The magnetic field bias unit (… Figure 2 The unit marked "3" provides a constant bias magnetic field in the diamond region. The fluorescence generated by the NV color center is collected by the fluorescence readout and acquisition unit (marked "4") and converted into an electrical signal. After acquisition, the electrical signal is sent to the data analysis module (marked "7"), and finally output and saved by the result display and storage module (marked "8").

[0041] In a preferred embodiment, the timing control system adjusts the time interval τ between adjacent microwave π pulses in the dynamic decoupling sequence to change the equivalent center detection frequency of the sequence, thereby adapting to the characteristic frequency components corresponding to different partial discharge types and achieving selective detection of high-frequency magnetic signals of partial discharges. Specifically, for tip discharges (whose pulse rise edge is steep and high-frequency components can reach hundreds of MHz), τ is set to a smaller value (e.g., 2ns~5ns) to focus the detection frequency in the high-frequency band; for particulate discharges (whose pulse duration is relatively long and characteristic frequencies are concentrated in the range of several MHz to tens of MHz), τ is set to a larger value (e.g., 50ns~200ns) to shift the detection frequency down to the mid-to-low frequency band. Through this frequency tuning mechanism, the same hardware system can achieve optimal detection matching for different discharge types without any hardware replacement.

[0042] Specific application examples

[0043] The following is a specific application example of online partial discharge monitoring on a gas-insulated sealed switch (GIS) device.

[0044] The probe, integrated with a diamond NV color center, is fixed to the surface of the GIS housing near the detection port of the busbar insulator. An ultrasonic coupling agent is applied to the bottom of the probe to enhance mechanical coupling with the housing surface. The probe is connected to an external host via a shielded cable. After the host is powered on, the system automatically executes a self-test program to confirm that the laser output power, microwave source frequency, detector bias voltage, and timing control system clock status are all normal.

[0045] The operator sets the measurement parameters via the host touchscreen: selects the dynamic decoupling sequence, sets the number of π pulses N=16, the sweep frequency range to 10MHz~500MHz, the sweep step size to 10MHz, a total of 49 frequency channels, and acquires 200 power frequency cycles (approximately 4 seconds, 50Hz power frequency) at each frequency point, with a total measurement time of approximately 196 seconds. The timing control system uses the power frequency synchronization signal provided by the substation as a reference and automatically cycles through steps S1 to S4 according to the preset sweep frequency parameter table.

[0046] After measurement, the system automatically inputs the fluorescence modulation amplitude data of 49 frequency channels into the built-in neural network model (pre-trained with spectral feature libraries for four defect types: tip discharge, particle discharge, floating potential discharge, and air gap discharge). The neural network output shows: "Floating potential discharge, confidence level 92.3%". Based on this diagnostic result, maintenance personnel arranged a power outage for repair. Upon disassembly, a floating potential component (an unsecured shield) was found inside the GIS, consistent with the diagnostic result, verifying the effectiveness of this method.

[0047] The names of the steps, units, and modules in this specific embodiment do not constitute a limitation on the scope of protection of this invention. Those skilled in the art can make adaptive adjustments to the parameters according to the specific equipment type and site conditions in practical applications, and such adjustments all fall within the scope of protection of this invention.

[0048] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A partial discharge detection method based on diamond NV color centers, characterized in that, Includes the following steps: S1: Deploy the probe, which integrates diamond NV color centers, at the predetermined detection location of the operating power equipment; S2: The pulse sequence parameters are set through the timing control system. The pulse sequence is used to coherently manipulate the spin state of the NV color center and detect the external magnetic field. S3: Using the power frequency voltage phase as the time reference, the laser-microwave pulse sequence is cyclically run to synchronously acquire NV color center fluorescence signals; S4: Change the timing parameters of the pulse sequence and perform multiple measurements to obtain magnetic field response signals at different frequencies, thereby realizing frequency domain detection of partial discharge signals.

2. The partial discharge detection method based on diamond NV centers according to claim 1, characterized in that, The pulse sequence is a dynamic decoupling sequence, and the method for setting the pulse sequence parameters in step S2 specifically includes: N equally spaced microwave π pulses are inserted between the initial microwave π / 2 pulse and the final microwave π / 2 pulse, where N≥1; The time interval τ between adjacent microwave π pulses is adjusted by the timing control system to tune the equivalent center detection frequency f of the dynamic decoupling sequence, where f = 1 / (2τ).

3. The partial discharge detection method based on diamond NV color centers according to claim 1, characterized in that, The pulse sequence is a spin echo sequence, and the timing control process for cyclically running the laser-microwave pulse sequence in step S3 includes: Turn on the laser pulse and initialize the NV color center to the ground state ms=0. Turn off the laser and apply a microwave π / 2 pulse to prepare a spin coherent superposition state; After the first free evolution time, a microwave π pulse is applied to induce spin re-aggregation; After a second free evolution time, a second microwave π / 2 pulse is applied to convert the phase information into population changes; Reactivate the laser pulse and read the fluorescence signal.

4. The partial discharge detection method based on diamond NV color centers according to claim 1, characterized in that, In step S3, the laser-microwave pulse sequence cycle uses the power frequency voltage phase as the synchronization reference, triggers multiple measurement sequences within each power frequency cycle, and precisely controls the opening and closing of the acousto-optic modulator and microwave switch through the timing control system to achieve nanosecond-level synchronous control of laser pulses and microwave pulses.

5. The partial discharge detection method based on diamond NV color centers according to claim 1, characterized in that, The process following step S4 also includes: S5: Convert the acquired fluorescence signals from multiple frequency channels into a spectral response feature vector of the high-frequency magnetic field of partial discharge; S6: Input the spectral response feature vector into a pre-trained neural network model; S7: The neural network model outputs the defect type of partial discharge and its confidence level to achieve online diagnosis of insulation defects.

6. The partial discharge detection method based on diamond NV color centers according to claim 5, characterized in that, The neural network model is trained using a partial discharge spectral feature library of different defect types as training samples. The defect types include, but are not limited to, tip discharge, particle discharge, floating potential discharge, and air gap discharge.

7. A partial discharge detection system based on diamond NV centers for implementing the method of any one of claims 1 to 6, characterized in that, include: The laser excitation and modulation unit is used to generate and modulate laser pulses to perform spin initialization and fluorescence readout of the NV color center; The microwave pulse control unit is used to generate and output microwave pulses in a set sequence to coherently control the spin state of the NV color center. The magnetic field biasing unit is used to generate a constant bias magnetic field in the region where the diamond is located, so as to cause the ground state of the NV color center to undergo Zeeman splitting and select the working energy level; The fluorescence readout and acquisition unit is used to collect the fluorescence generated by the NV color center and convert it into an electrical signal; The timing control system is used to receive the power frequency phase synchronization signal and control the timing synchronization between the laser excitation and modulation unit, the microwave pulse control unit, and the fluorescence readout and acquisition unit.

8. The partial discharge detection system based on diamond NV color centers according to claim 7, characterized in that, The laser excitation and control unit includes a 532nm laser and an acousto-optic modulator. The acousto-optic modulator is driven by a digital pulse signal provided by the timing control system to achieve rapid switching and power control of the laser pulse. The microwave pulse control unit includes an arbitrary waveform generator, a microwave source, a microwave switch, and a microwave radiation structure. The microwave switch is triggered by the timing control system to gate and shape the microwave pulse.

9. The partial discharge detection system based on diamond NV color centers according to claim 7, characterized in that, The probe is a fully solid-state miniaturized structure, which internally encapsulates a diamond single crystal, the magnetic field bias unit, and the microwave radiation structure. The probe is connected to an external host via a cable. The external host integrates the laser excitation and control unit, the microwave pulse control unit, the fluorescence readout and acquisition unit, and the timing control system.

10. The partial discharge detection system based on diamond NV centers according to claim 7, characterized in that, The timing control system adjusts the time interval τ between adjacent microwave π pulses in the dynamic decoupling sequence to change the equivalent center detection frequency of the sequence, thereby adapting to the characteristic frequency components corresponding to different partial discharge types and achieving selective detection of high-frequency magnetic signals of partial discharge.

Images