Partial discharge intelligent wireless sensor based on edge calculation and detection method

Abstract

The invention discloses a partial discharge intelligent wireless sensor based on edge calculation and a detection method, and the sensor employs a piezoelectric crystal to sense an ultrasonic signal excited by partial discharge, and obtains the ultrasonic signal through at least four partial discharge wireless sensors attached to a housing of a detected device. Partial discharge diagnosis is carried out according to the ultrasonic signal peak value and the discharge frequency, and when the ultrasonic signal received by any sensor meets the partial discharge diagnosis criterion, it is considered that the detected equipment has a partial discharge fault; and on the basis of all sensor coordinates and the arrival time of the main peak of the ultrasonic signal, simultaneous equations are established, and partial discharge positioning is realized by solving the equations. The device can be directly attached to a shell of electrical equipment such as a transformer for use, wiring and subsequent battery replacement maintenance are not needed, automatic networking can be achieved, on-line monitoring and positioning of internal partial discharge of the electrical equipment are achieved, and the safe operation level of the electrical equipment is improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of online monitoring of electrical equipment, and in particular relates to an intelligent wireless sensor for partial discharge based on edge computing and a detection method. Background Art

[0002] The safe and stable operation of electrical equipment is of great significance to ensuring the safety of the power system and improving the reliability of power supply. Partial discharge is an early sign of insulation aging of electrical equipment and will cause further degradation of insulation. Timely detection of partial discharge can help improve the safe operation level of electrical equipment.

[0003] Traditional partial discharge detection methods include:

[0004] CN202311848770.7 discloses a local discharge positioning method, device, terminal equipment and storage medium, including: receiving the local discharge signal of the transformer through a UHF sensor and determining the starting point of the local discharge; taking the starting point as the zero point, detecting the local discharge pulse signal within a preset time period through an ultrasonic sensor and performing time reversal to obtain an inverted pulse signal; transmitting the inverted pulse signal through an analog ultrasonic sensor set in the digital twin model of the transformer; obtaining the internal ultrasonic signal of the transformer through the analog ultrasonic sensor; generating an internal sound pressure distribution diagram based on the internal ultrasonic signal of the transformer; obtaining the maximum sound pressure point based on the internal sound pressure distribution diagram, and determining the local discharge occurrence point based on the maximum sound pressure point, wherein the digital twin model of the transformer is a digital twin model of the transformer; the analog ultrasonic sensor is a digital twin model of the ultrasonic sensor, and then, this method requires large computing resources and has a slow positioning time.

[0005] CN202311543700.0 discloses a method and probe for partial discharge detection based on ultrasonic data processing technology. The method first uses an ultrasonic probe for detection to convert the detected ultrasonic signal into an electrical signal; then the electrical signal is preprocessed including amplification and filtering; then an analog-to-digital converter is used to convert the preprocessed electrical signal into a digital signal to facilitate digital signal processing; then the digital signal is preprocessed, and then the useful signal is extracted and amplitude detection is performed; finally, data analysis is performed on the useful signal to determine whether there is partial discharge and the severity of the partial discharge, and the partial discharge spectrum and partial discharge alarm are uploaded. However, this method is easily affected by high-frequency noise on site and cannot locate partial discharge in time.

[0006] In addition, the partial discharge detection of the above inventions is based on the principle of ultrasonic detection. During the regular maintenance process, ultrasonic sensors, signal amplifiers, data acquisition cards and other equipment are carried on-site to manually detect the internal partial discharge of electrical equipment such as transformers. Many devices need to be carried, and it is difficult to detect sudden faults in time by using the method of live detection. Therefore, it is necessary to arrange sensors and detection devices on electrical equipment for a long time to form an online monitoring device. CN202311742647.7 discloses a method for locating partial discharge defects of power equipment in a blind spot of ultra-high frequency online monitoring, including: when the defect location trigger condition is met, the signal collected by the first ultra-high frequency sensor externally placed on the joint of the outlet bushing and the tank of the power equipment and the signal collected by the ultra-high frequency sensor pre-built in at the end are homologously analyzed to obtain the signal homology analysis result; according to the signal homology analysis result, the discharge defect is located in combination with the signal collected by the ultrasonic sensor externally placed on the outer surface of the equipment to obtain the partial discharge defect location result of the power equipment; the partial discharge defect location result of the power equipment is optimized by the signal collected by the second ultra-high frequency sensor externally placed on the outlet frame above the bushing, and the final partial discharge defect location result of the power equipment is obtained.

[0007] However, the wiring (power supply and data transmission) and maintenance costs of online monitoring devices based on wired sensors are relatively high, which restricts the large-scale promotion of this technology. Summary of the invention

[0008] In order to solve the shortcomings of the prior art, the present invention provides a partial discharge intelligent wireless sensor and detection method based on edge computing. The sensor uses a piezoelectric crystal to sense the ultrasonic signal excited by partial discharge, and integrates a sensitive unit, a signal conditioning circuit, an intelligent processing module and a storage unit, a wireless communication unit, a clock circuit, an energy acquisition unit, a power management module, etc. to form a small sensor, which can be directly attached to the casing of electrical equipment such as a transformer, without the need for wiring and subsequent battery replacement maintenance, and can automatically form a network to realize online monitoring of partial discharge inside electrical equipment, thereby improving the safe operation level of electrical equipment.

[0009] The present invention adopts the following technical solution.

[0010] The first aspect of the present invention proposes a partial discharge intelligent sensor based on edge computing, including a sensor housing, an ultrasonic sensitive unit, a signal conditioning circuit, an intelligent processing module and a wireless communication unit; specifically:

[0011] The signal conditioning circuit, intelligent processing module and wireless communication unit are all placed in the sensor housing;

[0012] A groove is provided in the sensor housing for embedding the ultrasonic sensitive unit. One end of the ultrasonic sensitive unit extends out of the sensor housing. A spring is provided at one end of the ultrasonic sensitive unit located in the housing. A magnetic ring is arranged on the inner side of the housing around the ultrasonic sensitive unit. Under the action of the magnetic ring and the spring, the sensor is tightly adsorbed onto the housing of the detected device.

[0013] After the ultrasonic sensing unit detects the discharge ultrasonic signal, it is amplified by the signal conditioning circuit and then input to the intelligent processing module;

[0014] The intelligent processing module performs partial discharge diagnosis based on the ultrasonic signal peak value and the number of discharges. When it is determined that the equipment to be inspected has a partial discharge fault, the wireless communication unit receives the coordinate signals of other intelligent sensors installed on the equipment to be inspected and the main peak arrival time signal of the ultrasonic signal; based on the sensor coordinates and the main peak arrival time of the ultrasonic signal, the simultaneous equations are used to realize partial discharge positioning.

[0015] Preferably, the clock synchronization module is connected to the intelligent processing module and the storage unit, and is used to align the time when multiple sensors are networked;

[0016] The energy collection unit is a small solar panel, which is fixedly mounted on a panel bracket and absorbs solar energy from the environment to provide energy for the sensor.

[0017] The power management module is connected to the power access ends of all modules including the sensitive unit, signal conditioning circuit, intelligent processing module and storage unit, wireless communication unit, clock synchronization module and energy acquisition unit, including the solar panel power management circuit and energy storage battery. The energy storage battery is used to store the energy absorbed by the solar panel to power the various functional modules of the sensor.

[0018] Preferably, the ultrasonic sensitive unit is a piezoelectric crystal, the resonance frequency of the piezoelectric crystal is 20-200 kHz, and the peak sensitivity is ≥80 dB.

[0019] Preferably, the signal conditioning circuit comprises a filter, which is a bandpass filter with a bandpass frequency of 20-200 kHz, and the amplification factor of the amplifier circuit is 100 times.

[0020] The second aspect of the present invention provides a partial discharge detection method based on edge computing using the partial discharge smart sensor according to the first aspect of the present invention, specifically comprising:

[0021] Step 1: Obtain ultrasonic signals through at least four partial discharge wireless sensors attached to the housing of the device to be tested, and extract ultrasonic signals at each sensor end;

[0022] Step 2: At each sensor end, partial discharge diagnosis is performed based on the extracted ultrasonic signal, the ultrasonic signal peak value and the number of discharges. When the ultrasonic signal received by any sensor meets the partial discharge diagnosis criterion, it is considered that the detected equipment may have a partial discharge fault, and then the process goes to step 3, otherwise it returns to step 1 to continue ultrasonic signal extraction;

[0023] Step 3: At any sensor end, based on the coordinates of all sensors and the arrival time of the main peak of the ultrasonic signal, establish the time equation for the fault point to reach each sensor, and the simultaneous equations to achieve partial discharge positioning by solving the equations.

[0024] Preferably, in step 1, short-time energy and short-time frequency are selected as characteristic parameters of the ultrasonic signal to extract the endpoints of the arrival time, so as to obtain the main peak arrival time and peak-to-peak value of the ultrasonic signal, wherein the short-time energy and short-time frequency are respectively the energy and frequency of each frame after the ultrasonic signal is framed; and the endpoints are the positions where the effective signal in the ultrasonic signal starts and ends.

[0025] Preferably, in step 1, the signal is first divided into frames according to a set length, and the short-time energy and short-time frequency are extracted from each frame; when the short-time energy of a frame signal is greater than the set energy threshold TH1, the center point of the signal frame is recorded as point A, and the center frequency of the frame corresponding to point A is calculated. When the center frequency is also greater than the set frequency threshold TH2, it is considered that the point is the main peak arrival time of the partial discharge ultrasonic signal.

[0026] Search for the peak-to-peak value of the signal in the period 2ms after this point, and record it as the peak-to-peak value of the main peak of the ultrasonic signal.

[0027] Preferably, the energy threshold TH1 and the frequency threshold TH2 are determined in the following manner:

[0028] When the sensor is installed on the equipment housing, it automatically collects background noise signals in the absence of partial discharge on site;

[0029] Every 2 hours, extract the background noise signal of 1s in length, calculate the short-time energy and center frequency, and extract 20 segments of signals in total;

[0030] Calculate the 25% quantile Q1 and 75% quantile Q3 of all short-time energies and center frequencies; calculate the interquartile range IQR: IQR = Q3-Q1; consider values ​​lower than Q1-1.5×IQR or higher than Q3+1.5×IQR as outliers and exclude them;

[0031] After excluding outliers:

[0032] The energy threshold TH1 is 3 times the short-term energy average value of multiple background noise signals with a length of 1s;

[0033] The frequency threshold TH2 takes the average value of the center frequencies of multiple background noise signals with a length of 1 s.

[0034] Preferably, in step 2, the diagnostic criterion for partial discharge is:

[0035] The combination of the peak-to-peak value of the main peak of the ultrasonic signal and the number of high-frequency pulses is graded and judged as follows:

[0036] 1) When the peak-to-peak value of the main peak is greater than the discharge threshold TH3, and the number of high-frequency pulses is greater than the set number threshold N, it is determined to be a high-probability partial discharge;

[0037] 2) When the peak-to-peak value of the main peak is greater than TH3, but the number of high-frequency pulses is less than the threshold value N, or when the peak-to-peak value of the main peak is less than TH3, but the number of high-frequency pulses is greater than the threshold value N, it is determined to be a medium-likelihood partial discharge;

[0038] 3) When the peak-to-peak value of the main peak is greater than TH1 but less than or equal to TH3, and the number of high-frequency pulses is greater than or equal to 0.5×N, it is judged as a low-probability partial discharge.

[0039] Preferably, the discharge threshold TH3 is twice the energy threshold TH1;

[0040] The basis for judging the high-frequency pulse is that the central frequency of the signal is greater than the frequency threshold TH2; the time period T is set to be a 3-minute time period from the time when the first pulse discharge occurs, and the number threshold N=6.

[0041] Preferably, in step 3, the partial discharge fault point (x0, y0, z0) is located according to the following simultaneous equations:

[0042]

[0043] In the formula, (x1, y1, z1), (x2, y2, z2), (x n ,y n ,z n ) are the coordinates of the first, second, and nth sensors, respectively, τ 12 , τ 1n are the time delays of the ultrasonic signal reaching the 2nd to nth sensors relative to the time it reaches the 1st sensor, (x0, y0, z0) are the coordinates of the partial discharge fault point to be determined, v is the velocity of the ultrasonic signal, and T is the time it takes for the ultrasonic signal to reach the first sensor.

[0044] Preferably, after the local discharge is located in step 3, the location information is sent through a wireless communication unit for early warning.

[0045] The beneficial effects of the present invention are that, compared with the prior art, the edge computing-based local discharge intelligent wireless sensor and detection method proposed by the present invention: (1) the sensor integrates a sensitive unit, a signal conditioning circuit, an intelligent processing module and a storage unit, a wireless communication unit, a clock circuit, an energy acquisition unit, a power management module, etc., thereby realizing the miniaturization and high integration of the sensor; (2) the sensor contains a power management module and the energy acquisition unit is used to power the sensor; it can be used by directly attaching it to the surface of the equipment, and does not require wiring and subsequent maintenance when it is placed on the electrical equipment for long-term online monitoring, which greatly reduces the cost; and the fixed structure based on the combination of magnetic rings and springs is stable and firm; (3) the wireless communication module is used to perform short-range networking between multiple sensors, which can realize automatic networking and positioning of local discharge fault points, improve the convenience of deploying local discharge online monitoring devices, and use the clock synchronization module to align time, which is conducive to avoiding the problem of detection time difference and improving positioning accuracy; (4) the wireless communication module can also send early warning information to a long distance, and can provide timely feedback when a sudden local discharge fault occurs, which is conducive to ensuring the safe and stable operation of power transformers and power systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] Figure 1 It is a schematic diagram of the local discharge intelligent wireless sensor based on edge computing in the present invention;

[0047] Figure 2 It is a schematic diagram of the fixing structure of the sensor in the present invention;

[0048] Figure 3 It is a schematic diagram of the signal conditioning circuit in the present invention;

[0049] Figure 4 This is a flow chart of the partial discharge detection method based on intelligent wireless sensors in the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical scheme and advantages of the present invention clearer, the technical scheme of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. The embodiments described in this application are only part of the embodiments of the present invention, not all of them. Based on the spirit of the present invention, other embodiments obtained by ordinary technicians in this field without creative work are all within the scope of protection of the present invention.

[0051] Embodiment 1 of the present invention discloses a local discharge intelligent wireless sensor based on edge computing such as Figure 1 As shown, it includes a sensitive unit, a signal conditioning circuit, an intelligent processing module and a storage unit, a wireless communication unit, a clock synchronization module, an energy acquisition unit, and a power management module.

[0052] The signal conditioning circuit, intelligent processing module and wireless communication unit are all placed in the sensor housing;

[0053] like Figure 2 As shown, a groove is provided in the sensor housing for embedding the ultrasonic sensitive unit, one end of the ultrasonic sensitive unit extends out of the sensor housing, a spring is provided at one end of the ultrasonic sensitive unit located in the housing, and a magnetic ring is arranged on the inner side of the housing around the ultrasonic sensitive unit. Under the action of the magnetic ring and the spring, the sensor is tightly adsorbed onto the housing of the detected device;

[0054] After the ultrasonic sensing unit detects the discharge ultrasonic signal, it is amplified by the signal conditioning circuit and then input to the intelligent processing module;

[0055] The intelligent processing module performs partial discharge diagnosis based on the ultrasonic signal peak value and the number of discharges. When it is determined that the equipment to be inspected has a partial discharge fault, the wireless communication unit receives the coordinate signals of other intelligent sensors installed on the equipment to be inspected and the main peak arrival time signal of the ultrasonic signal; based on the sensor coordinates and the main peak arrival time of the ultrasonic signal, the simultaneous equations are used to realize partial discharge positioning.

[0056] Preferably, the clock synchronization module is connected to the intelligent processing module and the storage unit, and is used to align the time when multiple sensors are networked;

[0057] The energy collection unit is a small solar panel, which is fixedly mounted on a panel bracket and absorbs solar energy from the environment to provide energy for the sensor.

[0058] The power management module is connected to the power access ends of all modules including the sensitive unit, signal conditioning circuit, intelligent processing module and storage unit, wireless communication unit, clock synchronization module and energy acquisition unit, including the solar panel power management circuit and energy storage battery. The energy storage battery is used to store the energy absorbed by the solar panel to power the various functional modules of the sensor.

[0059] Preferably, the ultrasonic sensitive unit is a piezoelectric crystal, the resonant frequency of the piezoelectric crystal is 20-200kHz, and the peak sensitivity is ≥80dB. Specifically, the piezoelectric crystal used in this embodiment is an R15α piezoelectric sensor, the resonant frequency of the piezoelectric sensor is 75kHz (Ref V / (m / s)), and the peak sensitivity is 80dB (Ref V / (m / s)).

[0060] Preferably, the magnetic ring is a strong NdFeB magnetic ring with an outer diameter of 50 mm, an inner diameter of 35 mm and a thickness of 10 mm.

[0061] Preferably, the signal conditioning circuit comprises a filter, which is a bandpass filter with a bandpass frequency of 20-200 kHz, and the amplification factor of the amplifier circuit is 100 times.

[0062] The signal conditioning circuit includes two-stage amplifier circuits. The first stage amplification factor is: 1+R7 / R8=1+9k / 1k=10; the second stage amplification factor is: 1+R10 / R11=1+9k / 1k=10, and the total amplification factor is 10*10=100, that is, the amplification factor is 100 times.

[0063] Specifically, the signal conditioning circuit includes a filter, specifically: a bandpass filter is composed of two-stage high-pass filters and two-stage low-pass filters, and the transfer functions of the high-pass and low-pass filters are both Butterworth.

[0064] The high pass filter cutoff frequency is

[0065] The low-pass filter cutoff frequency is

[0066] Therefore, the total passband frequency is 20-200kHz.

[0067] Preferably, the signal conditioning circuit is as follows Figure 3 As shown, specifically:

[0068] The signal conditioning circuit includes a first operational amplifier U1A, a second operational amplifier U1B, a third operational amplifier U2A, a fourth operational amplifier U2B, a first resistor R1, a second resistor R2, a third resistor R3, a fourth resistor R4, a fifth resistor R5, a sixth resistor R6, a seventh resistor R7, an eighth resistor R8, a ninth resistor R9, a tenth resistor R10, an eleventh resistor R11, a first capacitor C1, a second capacitor C2, a third capacitor C3, and a fourth capacitor C4:

[0069] R9 is connected to the positive input terminal of the third operational amplifier U2A, one end of the eighth resistor R8 and one end of the seventh resistor R7 are connected to the negative input terminal of the third operational amplifier U2A, and the output terminal of the third operational amplifier U2A and the other end of the seventh resistor are connected to the positive input terminal of the fourth operational amplifier U2B;

[0070] One end of the eleventh resistor R11 and one end of the tenth resistor R10 are connected to the negative input end of the fourth operational amplifier U2B, and the output end of the fourth operational amplifier U2B and the other end of the tenth resistor R10 are connected to one end of the fourth capacitor C4;

[0071] The other end of the fourth capacitor C4 and one end of the fifth resistor R5 are connected to one end of the first capacitor C1;

[0072] The other end of the first capacitor C1 and one end of the sixth resistor R6 are connected to the positive input end of the first operational amplifier U1A, one end of the third resistor R3 is connected to the negative input end of the first operational amplifier U1A, and the output end of the first operational amplifier U1A, the other end of the third resistor R3, and the other end of the fifth resistor R5 are connected to one end of the first resistor R1;

[0073] The other end of the first resistor R1, one end of the third capacitor C3 and one end of the second resistor R2 are connected;

[0074] The other end of the second resistor R6 and one end of the second capacitor C2 are connected to the positive input end of the second operational amplifier U1B, one end of the fourth resistor R4 is connected to the negative input end of the second operational amplifier U1B, and the output end of the second operational amplifier U1B and the other end of the fourth resistor R4 are connected to the other end of the third capacitor C3 as the output end of the signal conditioning circuit;

[0075] The first operational amplifier U1A, the second operational amplifier U1B, the third operational amplifier U2A, and the fourth operational amplifier U2B are all powered by dual power supplies, with the positive power rail connected to 5V VCC and the negative power rail connected to -5V VEE;

[0076] The other ends of the sixth resistor R6 , the eighth resistor R8 , the eleventh resistor R11 and the second capacitor C2 are grounded; the other end of the ninth resistor R9 is connected to the input signal.

[0077] Preferably, the first operational amplifier U1A, the second operational amplifier U1B, the third operational amplifier U2A, and the fourth operational amplifier U2B are all NE5532AI operational amplifiers; the resistance of the first resistor R1 is 5.6kΩ, the resistance of the second resistor R2 is 5.6kΩ, the resistance of the third resistor R3 is 10kΩ, the resistance of the fourth resistor R4 is 10kΩ, the resistance of the fifth resistor R5 is 5.6kΩ, the resistance of the sixth resistor R6 is 11.5kΩ, the resistance of the seventh resistor R7 is 9kΩ, the resistance of the eighth resistor R8 is 1kΩ, the resistance of the ninth resistor R9 is 7.5kΩ, the resistance of the tenth resistor R10 is 9kΩ, the resistance of the eleventh resistor R11 is 1kΩ, the capacitance of the first capacitor C1 is 1nF, the capacitance of the second capacitor C2 is 100pF, the capacitance of the third capacitor C3 is 200pF, and the capacitance of the fourth capacitor C4 is 1nF.

[0078] Preferably, the intelligent processing module and storage unit are Nano 33BLE Rev2, whose processor is nRF52840, and the storage unit includes 256KB static random access memory, 1MB flash memory, and integrated u- NINA-B306 Bluetooth chip.

[0079] Preferably, the wireless communication module includes Bluetooth and LoRa modules, Bluetooth is used for short-range networking between multiple sensors, and the LoRa module is used to send warning information to long-distance data terminals.

[0080] Preferably, the Bluetooth chip used is a Bluetooth chip integrated in the intelligent processing module and the storage unit, and its communication distance is 10 meters. The LoRa communication module is a HC-14 patch type LoRa communication module, and its communication distance is 3000 meters.

[0081] Preferably, the clock synchronization module is a SKG09DT Beidou timing module, which has a positioning accuracy of 2.5m and a timing accuracy of 10ns.

[0082] Preferably, the energy acquisition unit is a Weixue 5.5V, 6W monocrystalline silicon solar panel.

[0083] Preferably, the power management module is a Weixue 5V, 3A solar power module.

[0084] Preferably, in order to reduce the volume and weight, a 1000mAh lithium battery is used in this embodiment. The solar power module has an MPPT power point tracking function, an onboard overcharge and over-discharge protection function, and supports a maximum stable output of 5V and 3A.

[0085] like Figure 4 As shown, embodiment 1 of the present invention proposes a partial discharge detection method based on edge computing using the partial discharge intelligent sensor described in embodiment 1 of the present invention. This embodiment is aimed at partial discharge monitoring of power transformers, and specifically includes:

[0086] Step 1: Obtain ultrasonic signals through at least four partial discharge wireless sensors attached to the housing of the device to be tested, and extract ultrasonic signals at each sensor end;

[0087] Preferably, in step 1, short-time energy and short-time frequency are selected as characteristic parameters of the ultrasonic signal to extract the endpoints of the arrival time, so as to obtain the main peak arrival time and peak-to-peak value of the ultrasonic signal, wherein the short-time energy and short-time frequency are respectively the energy and frequency of each frame after the ultrasonic signal is framed; and the endpoints are the positions where the effective signal in the ultrasonic signal starts and ends.

[0088] Preferably, in step 1, the signal is first divided into frames according to a set length, and the short-time energy and short-time frequency are extracted from each frame; when the short-time energy of a frame signal is greater than the set energy threshold TH1, the center point of the signal frame is recorded as point A, and the center frequency of the frame corresponding to point A is calculated. When the center frequency is also greater than the set frequency threshold TH2, it is considered that the point is the main peak arrival time of the partial discharge ultrasonic signal.

[0089] Search for the peak-to-peak value of the signal in the period 2ms after this point, and record it as the peak-to-peak value of the main peak of the ultrasonic signal.

[0090] Specifically, in this embodiment, the frame length F is selected to be 0.02 ms, and the frame shift SH is selected to be 0.01 ms.

[0091] It should be noted that the calculation of short-time energy is as follows:

[0092]

[0093] Where x[n+m] is the signal value of the mth sampling point in the nth frame, w[m] is the value of the window function, M is the number of sampling points per frame, and m represents the mth sampling point. The window function includes rectangular window, Hanning window, and Hamming window, which are used to reduce edge effects.

[0094] The short-time frequency is calculated as follows:

[0095]

[0096] Where k is the frequency index, corresponding to components of different frequencies.

[0097] Preferably, the energy threshold TH1 and the frequency threshold TH2 are determined in the following manner:

[0098] When the sensor is installed on the equipment housing, it automatically collects background noise signals in the absence of partial discharge on site;

[0099] Every 2 hours, extract the background noise signal of 1s in length, calculate the short-time energy and center frequency, and extract 20 segments of signals in total;

[0100] Calculate the 25% quantile Q1 and 75% quantile Q3 of all short-time energies and center frequencies; calculate the interquartile range IQR: IQR = Q3-Q1; consider values ​​lower than Q1-1.5×IQR or higher than Q3+1.5×IQR as outliers and exclude them;

[0101] After excluding outliers:

[0102] The energy threshold TH1 is 3 times the short-term energy average value of multiple background noise signals with a length of 1s;

[0103] The frequency threshold TH2 takes the average value of the center frequencies of multiple background noise signals with a length of 1 s.

[0104] Specifically, the number of multi-segment background noise signals set in this embodiment is 5.

[0105] Step 2: At each sensor end, partial discharge diagnosis is performed based on the extracted ultrasonic signal, the ultrasonic signal peak value and the number of discharges. When the ultrasonic signal received by any sensor meets the partial discharge diagnosis criterion, it is considered that the detected equipment may have a partial discharge fault, and then the process goes to step 3, otherwise it returns to step 1 to continue ultrasonic signal extraction;

[0106] Preferably, in step 2, the partial discharge diagnosis criterion is:

[0107] The classification is performed by combining the peak-to-peak value of the main peak of the ultrasonic signal and the number of high-frequency pulses. Specifically:

[0108] 1) When the peak-to-peak value of the main peak is greater than the discharge threshold TH3, and the number of high-frequency pulses is greater than the number threshold N, it is determined to be a high-probability partial discharge;

[0109] 2) When the peak-to-peak value of the main peak is greater than TH3, but the number of high-frequency pulses is less than the threshold value N, or when the peak-to-peak value of the main peak is less than TH3, but the number of high-frequency pulses is greater than the threshold value N, it is determined to be a medium-likelihood partial discharge.

[0110] 3) When the peak-to-peak value of the main peak is greater than TH1 but less than or equal to TH3, and the number of high-frequency pulses is greater than or equal to 0.5×N, it is judged as a low-probability partial discharge.

[0111] Preferably, in this embodiment, the discharge threshold TH3 is taken as twice the energy threshold TH1; the judgment basis of the high-frequency pulse is that the center frequency of the signal segment is greater than the frequency threshold TH2; the time period T is set to a 3-minute time period from the time when the first pulse discharge occurs, and the number threshold N=6.

[0112] Step 3: At any sensor end, based on the coordinates of all sensors and the arrival time of the main peak of the ultrasonic signal, establish the time equation for the fault point to reach each sensor, and the simultaneous equations to achieve partial discharge positioning by solving the equations.

[0113] Preferably, the coordinates of each sensor and the main peak arrival time are combined to form a set of simultaneous equations. In this embodiment, four sensors are selected, which are located on four sides of the transformer, and the simultaneous equations are as follows:

[0114]

[0115] In the formula, (x1, y1, z1), (x2, y2, z2), (x n ,y n ,z n ) are the coordinates of the first, second, and nth sensors, respectively, τ 12 , τ 1nare the time delays of the ultrasonic signal reaching the 2nd to nth sensors relative to the time it reaches the 1st sensor, (x0, y0, z0) are the coordinates of the partial discharge fault point to be determined, v is the velocity of the ultrasonic signal, and T is the time it takes for the ultrasonic signal to reach the first sensor.

[0116] Preferably, after the local discharge is located in step 3, the location information is sent through a wireless communication unit for early warning.

[0117] It should be noted that the equations need to be solved to locate the partial discharge. In this embodiment, a particle swarm optimization algorithm is used to search for possible solutions of the equations to locate the partial discharge, and the location information of the partial discharge is sent through the wireless sensor module.

[0118] The specific process of solving the equation is as follows:

[0119] 1) Initialization:

[0120] The dimension of each particle is defined as 4 dimensions (corresponding to the unknown numbers x0, y0, z0, T).

[0121] Initialize the position and velocity of the particles. The positions of the particles are randomly distributed in the solution space, and the solution space is taken as the length, width, and height of the internal space of the transformer to be tested.

[0122] 2) Define the objective function:

[0123] Define the objective function based on the system of equations. Subtract the left side from the right side of each equation and calculate the sum of the squares of the residuals:

[0124] Error=((x0-x1) 2 +(y0-y1) 2 +(z0-z1) 2 -v 2 T 2 ) 2

[0125] +((x0-x2) 2 +(y0-y2) 2 +(z0-z2) 2 -v 2 (T+τ 12 ) 2 ) 2

[0126] +((x0-x3) 2 +(y0-y3) 2 +(z0-z3) 2 -v 2 (T+τ 13 ) 2 ) 2

[0127] +((x0-x4) 2 +(y0-y4) 2 +(z0-z4) 2 -v 2 (T+τ 14 ) 2 ) 2

[0128] The smaller the value of the error function of the calculation result is, the closer the position of the particle is to the solution of the equation group and the location of the partial discharge fault point.

[0129] 3) Iteratively update particles:

[0130] Update the particle's velocity and position. The velocity is updated based on the current particle's position, local optimal position, and global optimal position:

[0131] v i (t+1)=wv i (t)+c1r1(p best -x i (t))+c2r2(g best -x i (t))

[0132] Where: v i (t+1) is the velocity of the i-th particle at the next moment, x i (t) is the current particle position, p best is the best historical position of the particle, g best is the global optimal position, w is the inertia weight, c1 and c2 are acceleration factors, r1 and r2 are random numbers, and then the particle position is updated according to the updated velocity:

[0133] x i (t+1)=x i (t)+v i (t+1)

[0134] When the error of the global optimal solution is less than the set threshold or the number of iterations reaches the maximum value, the iteration stops.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the relevant field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents, and any modifications or equivalent replacements that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A partial discharge intelligent sensor based on edge computing, comprising a sensor housing, an ultrasonic sensing unit, a signal conditioning circuit, an intelligent processing module and a wireless communication unit; characterized in that: The signal conditioning circuit, intelligent processing module and wireless communication unit are all placed in the sensor housing; A groove is provided in the sensor housing for embedding the ultrasonic sensitive unit. One end of the ultrasonic sensitive unit extends out of the sensor housing. A spring is provided at one end of the ultrasonic sensitive unit located in the housing. A magnetic ring is arranged on the inner side of the housing around the ultrasonic sensitive unit. Under the action of the magnetic ring and the spring, the sensor is tightly adsorbed onto the housing of the detected device. After the ultrasonic sensing unit detects the discharge ultrasonic signal, it is amplified by the signal conditioning circuit and then input to the intelligent processing module; The intelligent processing module performs partial discharge diagnosis based on the ultrasonic signal peak value and the number of discharges. When it is determined that the equipment to be inspected has a partial discharge fault, the wireless communication unit receives the coordinate signals of other intelligent sensors installed on the equipment to be inspected and the main peak arrival time signal of the ultrasonic signal; based on the sensor coordinates and the main peak arrival time of the ultrasonic signal, the simultaneous equations are used to realize partial discharge positioning.

2. The partial discharge intelligent sensor according to claim 1 further comprises an energy acquisition unit, a clock synchronization module and a power management module, characterized in that: A clock synchronization module, connected to the intelligent processing module and the storage unit, is used to align the time when multiple sensors are networked; The energy collection unit is a small solar panel, which is fixedly mounted on a panel bracket and absorbs solar energy from the environment to provide energy for the sensor. The power management module is connected to the power access ends of all modules including the sensitive unit, signal conditioning circuit, intelligent processing module and storage unit, wireless communication unit, clock synchronization module and energy acquisition unit, including the solar panel power management circuit and energy storage battery. The energy storage battery is used to store the energy absorbed by the solar panel to power the various functional modules of the sensor.

3. According to claim 1, a partial discharge intelligent wireless sensor based on edge computing is characterized in that: The ultrasonic sensitive unit is a piezoelectric crystal, the resonance frequency of the piezoelectric crystal is 20-200kHz, and the peak sensitivity is ≥80dB.

4. According to the edge computing-based partial discharge intelligent wireless sensor of claim 1, it is characterized by: The signal conditioning circuit comprises a filter, which is a bandpass filter with a bandpass frequency of 20-200kHz, and the amplification factor of the amplifier circuit is 100 times.

5. A partial discharge detection method based on edge computing using the partial discharge intelligent sensor according to any one of claims 1 to 4, characterized in that: The method comprises: Step 1: Obtain ultrasonic signals through at least four partial discharge wireless sensors attached to the housing of the device to be tested, and extract ultrasonic signals at each sensor end; Step 2: At each sensor end, partial discharge diagnosis is performed based on the extracted ultrasonic signal, the ultrasonic signal peak value and the number of discharges. When the ultrasonic signal received by any sensor meets the partial discharge diagnosis criterion, it is considered that the detected equipment may have a partial discharge fault, and then the process goes to step 3, otherwise it returns to step 1 to continue ultrasonic signal extraction; Step 3: At any sensor end, based on the coordinates of all sensors and the arrival time of the main peak of the ultrasonic signal, establish the time equation for the fault point to reach each sensor, and the simultaneous equations to achieve partial discharge positioning by solving the equations.

6. The partial discharge detection method according to claim 5, characterized in that: In step 1, short-time energy and short-time frequency are selected as characteristic parameters of the ultrasonic signal to extract the endpoints of the arrival time, and the arrival time and peak-to-peak value of the main peak of the ultrasonic signal are obtained. The short-time energy and short-time frequency are the energy and frequency of each frame after the ultrasonic signal is framed; the endpoints are the positions where the effective signal in the ultrasonic signal starts and ends.

7. The partial discharge detection method according to claim 5 or 6, characterized in that: In step 1, the signal is first divided into frames according to the set length, and the short-time energy and short-time frequency are extracted from each frame; when the short-time energy of a frame signal is greater than the set energy threshold TH1, the center point of the signal frame is recorded as point A, and the center frequency of the frame corresponding to point A is calculated. When the center frequency is also greater than the set frequency threshold TH2, it is considered that this point is the main peak arrival time of the local discharge ultrasonic signal. Search for the peak-to-peak value of the signal in the period 2ms after this point, and record it as the peak-to-peak value of the main peak of the ultrasonic signal.

8. The partial discharge detection method according to claim 7, characterized in that: The energy threshold TH1 and the frequency threshold TH2 are determined in the following manner: When the sensor is installed on the equipment housing, it automatically collects background noise signals in the absence of partial discharge on site; Every 2 hours, extract the background noise signal of 1s in length, calculate the short-time energy and center frequency, and extract 20 segments of signals in total; Calculate the 25% quantile Q1 and 75% quantile Q3 of all short-time energies and center frequencies; calculate the interquartile range IQR: IQR = Q3-Q1; consider values ​​lower than Q1-1.5×IQR or higher than Q3+1.5×IQR as outliers and exclude them; After excluding outliers: The energy threshold TH1 is 3 times the short-term energy average value of multiple background noise signals with a length of 1s; The frequency threshold TH2 takes the average value of the center frequencies of multiple background noise signals with a length of 1 s.

9. The partial discharge detection method according to claim 8, characterized in that: In step 2, the diagnostic criteria for partial discharge are: The classification is performed by combining the peak-to-peak value of the main peak of the ultrasonic signal and the number of high-frequency pulses, specifically: 1) When the peak-to-peak value of the main peak is greater than the discharge threshold TH3, and the number of high-frequency pulses is greater than the set number threshold N, it is determined to be a high-probability partial discharge; 2) When the peak-to-peak value of the main peak is greater than TH3, but the number of high-frequency pulses is less than the threshold value N, or when the peak-to-peak value of the main peak is less than TH3, but the number of high-frequency pulses is greater than the threshold value N, it is determined to be a medium-likelihood partial discharge; 3) When the peak-to-peak value of the main peak is greater than TH1 but less than or equal to TH3, and the number of high-frequency pulses is greater than or equal to 0.5×N, it is judged as a low-probability partial discharge.

10. The partial discharge detection method according to claim 9, characterized in that: The discharge threshold TH3 is twice the energy threshold TH1; The judgment basis of the high-frequency pulse is that the central frequency of the signal is greater than the frequency threshold TH2; the time period T is set to a 3-minute time period from the judgment of the first pulse discharge, and the number threshold N=6.

11. The partial discharge detection method according to claim 5, characterized in that: In step 3, the partial discharge fault point (x0, y0, z0) is located according to the following simultaneous equations: In the formula, (x1, y1, z1), (x2, y2, z2), (x n ,y n ,z n ) are the coordinates of the first, second, and nth sensors, respectively, τ 12 , τ 1n are the time delays of the ultrasonic signal reaching the 2nd to nth sensors relative to the time it reaches the 1st sensor, (x0, y0, z0) are the coordinates of the partial discharge fault point to be determined, v is the velocity of the ultrasonic signal, and T is the time it takes for the ultrasonic signal to reach the first sensor.

12. The partial discharge detection method according to claim 1, characterized in that: In step 3, after the partial discharge is located, the location information is sent through the wireless communication unit for early warning.

Images