An automatic detection transformer maintenance-free respirator and a working method thereof

Abstract

This invention discloses an automatic transformer maintenance-free breather and its operating method, belonging to the field of breather technology. It solves the problems of difficulty in early warning of power module aging and power failure in maintenance-free breathers, silicone regeneration shutdown, direct intrusion of humid air into the oil conservator leading to insulation oil oxidation and breakdown, and a significantly increased risk of transformer short circuits and explosions. The invention includes a breather body and a monitoring and control system installed within it. This invention accurately senses and diagnoses subtle signs of progressive aging of the power supply in the breather body, ensuring reliable data reporting. It also performs in-depth analysis and predicts remaining lifespan, dynamically adjusts thresholds based on risk, and generates decision commands. Finally, through an autonomous magnetorheological backup module, it provides emergency power or triggers maintenance alarms before the fault becomes critical. This completes a closed loop from latent aging identification and predictive early warning to active fault-tolerant protection without human intervention, effectively preventing breather malfunction and subsequent transformer insulation accidents caused by power failure.

Description

Technical Field

[0001] This invention relates to the field of respirator technology, and in particular to an automatic transformer detection maintenance-free respirator and its operating method. Background Technology

[0002] The transformer maintenance-free breather is an important auxiliary component of the transformer oil conservator, primarily used to prevent moisture and impurities from the air from entering the transformer, protecting the insulation performance of the transformer oil and extending the equipment's lifespan. It employs long-lasting moisture-absorbing materials and a sealed structure design, eliminating the need for frequent disassembly and desiccant replacement. It automatically filters moisture from the air, and some models can regenerate the absorbent material through their own structure or display a replacement indicator when it becomes saturated. Its outer shell is made of corrosion-resistant material, providing excellent sealing and adapting to the breathing requirements of transformers of different specifications. It effectively prevents problems such as oil deterioration and insulation degradation caused by moisture, simplifying equipment maintenance procedures and ensuring long-term stable operation of the transformer.

[0003] For example, the dual-tank maintenance-free breather device and transformer disclosed in CN215069562U have the following technical features: a housing connected to the end of a pipe communicating with the transformer's oil tank and air via a flange; a pipeline connected at one end to the flange and at the other end to a solenoid valve for switching on and off; a first silica gel tank connected to the pipeline via the solenoid valve, the first silica gel tank including a first temperature and humidity sensor configured to measure first temperature and humidity data within the first silica gel tank; a first heater chamber assembly for heating and drying the silica gel in the first silica gel tank based on the temperature and humidity data; and a second silica gel tank connected to the pipeline via the solenoid valve. The pipeline is connected to the ground. The second silicone container includes: a second temperature and humidity sensor configured to measure second temperature and humidity data within the second silicone container; a second heater chamber assembly that heats and dries the silicone in the second silicone container based on the second temperature and humidity data; a solenoid valve that switches on and off to ensure that the pipeline is always connected to one of the first and second silicone containers; an air pump located within the housing and pumping air to either the first or second silicone container; a power supply located within the housing and electrically connected to the solenoid valve, the air pump, the first and second silicone containers; and a control unit electrically connected to the power supply, the solenoid valve, the air pump, the first and second silicone containers.

[0004] In substations operating unattended for extended periods, the dual-tank maintenance-free breather for transformers, while continuously supplying dry air to the transformer oil conservator and automatically regenerating silica gel, can experience progressive aging due to the combined effects of long-term high loads, fluctuating ambient temperatures, and unstable grid voltage. This can lead to power outages and complete device malfunction. Initially, these faults are often very subtle, with signs such as power fluctuations or intermittent abnormal indicator lights. Remote monitoring may misinterpret these as ordinary grid disturbances, delaying alarms and making them difficult to identify during routine inspections or monitoring. Ultimately, this results in complete breather failure, allowing humid, unfiltered air to directly enter the oil conservator, causing rapid oxidation of the insulating oil and a sharp drop in dielectric strength. This can eventually lead to internal short circuits, overheating, or even explosions within the transformer, causing grid outages, substantial economic losses, and serious safety risks. Therefore, it is necessary to improve the transformer maintenance-free breather to include self-testing capabilities to detect and resolve problems as early as possible.

[0005] Therefore, an automatic transformer maintenance-free breather and its working method are proposed to solve or alleviate the above problems. Summary of the Invention

[0006] The purpose of this invention is to address the shortcomings of existing technologies by proposing an automatic transformer maintenance-free breather and its operating method.

[0007] To achieve the above objectives, the present invention adopts the following technical solution:

[0008] An automatic transformer detection maintenance-free respirator includes a respirator body and a monitoring and control system installed therein;

[0009] The monitoring and control system includes a magnetostrictive and photoacoustic signal capture module, an eddy current test preprocessing module, a data integrity verification module, a cloud-based piezoresistive impedance diagnostic module, an adaptive spectrum threshold alarm module, and an autonomous magnetorheological backup module connected in sequence. The detection terminals of the magnetostrictive and photoacoustic signal capture module are respectively located on the surface of the power supply coil in the respirator body, the hot zone of the power supply inside the housing, and connected to the grounding shell of the air pump. The detection terminal of the eddy current test preprocessing module is located above the circuit board of the power supply in the respirator body. The detection terminal of the eddy current test preprocessing module is connected to the expansion port of the temperature and humidity sensor located in the respirator body and inside the silicone canister. The detection terminal of the cloud-based piezoresistive impedance diagnostic module is connected to the output bus and load circuit of the power supply in the respirator body. The autonomous magnetorheological backup module is connected to the auxiliary power rail of the power supply in the respirator body.

[0010] The magnetostriction and photoacoustic signal acquisition module collects the magnetostriction and photoacoustic signals of the power supply in the respirator body and extracts physical characteristic quantities representing aging from them. The eddy current test preprocessing module collects the eddy current response signal of the power supply in the respirator body and preprocesses it to obtain impedance characteristics. The data integrity verification module is used to verify the extracted feature vectors and generate data integrity verification codes. The cloud-based piezoresistive impedance diagnosis module is used to fuse historical and real-time data for in-depth analysis to predict the remaining service life and calculate the risk score. The adaptive spectrum threshold alarm module is used to dynamically adjust the alarm threshold according to the risk score and fuse information to generate control commands.

[0011] Preferably, the monitoring and control system further includes a redundant communication module for connecting the data integrity verification module and the cloud-based piezoresistive impedance diagnostic module. The adaptive spectrum threshold alarm module is connected to the autonomous magnetorheological backup module. The redundant communication module is used to reliably transmit the verified data to the cloud via a wireless mesh network. The autonomous magnetorheological backup module is activated when a high-risk command is received and captures ambient energy to provide emergency power for the power supply in the respirator body.

[0012] Preferably, the magnetostriction and photoacoustic signal capture module includes a magnetostriction sensor unit, a photoacoustic effect detection unit, a signal conditioning and noise suppression unit, an analog-to-digital conversion unit, and a digital signal processing unit;

[0013] The magnetostrictive sensor unit includes a piezoelectric transducer, an induction coil, and an NPN bipolar transistor. The positive terminal of the piezoelectric transducer is connected to one end of the induction coil, and the other two ends of the induction coil are connected to the collector of the NPN bipolar transistor through a first current-limiting resistor. The base of the NPN bipolar transistor is connected to the ground through a base bias resistor.

[0014] The photoacoustic effect detection unit includes a photoacoustic detector and a low-noise operational amplifier. The synchronous trigger terminal of the photoacoustic detector is connected to the signal output node of the piezoelectric transducer through a parallel high-frequency filter capacitor and a reverse voltage protection diode. The cathode of the photoacoustic detector is connected to the inverting input terminal of the low-noise operational amplifier through an input resistor. The non-inverting input terminal of the low-noise operational amplifier is grounded, and its output terminal is connected to its inverting input terminal through a negative feedback resistor. The signal input node of the low-noise operational amplifier is connected in parallel with a phase compensation capacitor and a first voltage clamping diode.

[0015] The signal conditioning and noise suppression unit includes a high common-mode rejection ratio (HMR) instrumentation amplifier. The non-inverting input of the HMR instrumentation amplifier is connected to the output of a low-noise operational amplifier via a coupling resistor, and its inverting input is grounded. The output of the HMR instrumentation amplifier outputs a conditioned signal through a first low-pass filter. The differential signal input of the HMR instrumentation amplifier is connected to the output of a photoacoustic effect detection unit via an isolation transformer.

[0016] The analog-to-digital conversion unit includes an analog-to-digital converter. The positive terminal of the first analog differential input of the analog-to-digital converter is connected to the output terminal of a high common-mode rejection ratio instrumentation amplifier, and the negative terminal of its first analog differential input is grounded.

[0017] Preferably, the eddy current test preprocessing module includes an eddy current excitation and detection unit, a signal isolation and amplification unit, a digital signal processing core unit, and a data buffer output unit;

[0018] The eddy current excitation and detection unit includes a bridge balance circuit composed of an excitation coil and a detection coil, a function signal generator, and a PNP bipolar transistor. The output terminal of the function signal generator is connected to the common connection point of the excitation coil and the detection coil. The output amplitude adjustment terminal of the function signal generator is set by a first adjustable potentiometer. The emitter of the PNP bipolar transistor is energized, its collector is connected to the output terminal of the function signal generator, and its base is biased and controlled by a base drive resistor.

[0019] The signal isolation and amplification unit includes a programmable gain amplifier and a signal isolation transformer. The positive terminal of the differential input of the programmable gain amplifier is connected to the signal output node of the bridge balanced circuit, and its negative terminal is grounded. Its gain is set through its parallel digital control interface. The primary winding side of the signal isolation transformer is connected to the probe signal output terminal of the eddy current excitation and detection unit. A second voltage clamping diode is connected in series on the primary winding side of the signal isolation transformer, and an impedance matching resistor and a first buffer transistor are connected in parallel on the secondary winding side. The base of the first buffer transistor is connected to one end of the secondary winding of the signal isolation transformer, the collector of the first buffer transistor is connected to the positive terminal of the differential input of the programmable gain amplifier, and the emitter of the first buffer transistor is grounded.

[0020] The digital signal processing core unit includes a microprocessor, whose analog signal acquisition port is connected to the single-ended output of a programmable gain amplifier. The first set of general-purpose input and output terminals of the microprocessor is connected to the clock signal input terminal and the data latch signal input terminal of the data buffer output unit through pull-down resistors.

[0021] The data buffer output unit includes a serial shift register. The serial data input terminal of the serial shift register is connected to the data bus of the microprocessor, and its parallel data output terminal is used to connect to the data integrity verification module.

[0022] Preferably, the data integrity verification module includes a Hamming code verification and processing unit, a secure hash operation unit, and a non-volatile storage unit;

[0023] The Hamming code verification and processing unit includes a field-programmable gate array (FPGA), and its front end is connected to a signal input buffer circuit composed of a low-offset operational amplifier. The non-inverting input terminal of the low-offset operational amplifier receives the verification data signal from the eddy current test preprocessing module, its inverting input terminal is grounded, and its output terminal is connected to the general input / output port group of the FPGA through a second low-pass filter.

[0024] The secure hashing unit includes a hashing algorithm chip and a first precision reference voltage source. Its power input terminal is connected to an NPN current-enhancing transistor. The base of the NPN current-enhancing transistor is connected to the cathode of the first precision reference voltage source. The anode of the first precision reference voltage source is grounded, and its cathode is connected to a second current-limiting resistor. A tantalum decoupling capacitor is connected in parallel with the second current-limiting resistor. The specific low-voltage complementary metal-oxide-semiconductor level output terminal group of the field-programmable gate array is respectively connected to the data input terminal, clock input terminal, and enable control terminal of the hashing algorithm chip.

[0025] The non-volatile memory cell includes an electrically erasable programmable read-only memory chip, whose serial data line and serial clock line are powered through a first pull-up resistor, and whose write protection control terminal is connected to a general-purpose input / output terminal of a field-programmable gate array. The data line and clock line generated by the hash algorithm chip are connected to the corresponding serial interface terminal of the electrically erasable programmable read-only memory chip.

[0026] Preferably, the redundant communication module includes a wireless transceiver unit, a mesh network control and processing unit, a radio frequency front-end amplification and matching unit, and an antenna and switching unit;

[0027] The wireless transceiver unit includes an integrated radio frequency transceiver chip, whose serial peripheral interface's host input / slave output terminal, host output / slave input terminal, serial clock terminal, and chip select terminal are connected to the corresponding host interface terminal of the mesh network control and processing unit.

[0028] The radio frequency front-end amplification and matching unit includes a radio frequency power amplifier. The signal input terminal of the radio frequency power amplifier is connected to the radio frequency output terminal of the integrated radio frequency transceiver chip through a π-type impedance matching network. The power supply terminal of the radio frequency power amplifier is connected to the power supply through a third current-limiting resistor. A high-frequency decoupling capacitor is connected in parallel with the third current-limiting resistor. An electrostatic discharge protection diode is connected to the output terminal of the radio frequency power amplifier.

[0029] The mesh network control and processing unit includes a mesh network coordinator, whose radio frequency signal output terminal is connected to the signal input terminal of the antenna and the switching unit through a first coaxial connector;

[0030] The antenna and switching unit includes a helical antenna, a balun, and an antenna path switching circuit. The unbalanced port of the balun is connected to the RF signal output of the mesh network coordinator, and its balanced port is connected to the feed center point of the helical antenna. The antenna path switching circuit includes a PIN diode and a PNP switching transistor. The anode of the PIN diode is connected to the RF signal, the cathode of the PIN diode is energized through a first bias resistor, the emitter of the PNP switching transistor is energized, the collector of the PNP switching transistor is connected to the cathode of the PIN diode, and the base of the PNP switching transistor is connected to the general-purpose input / output terminal of the mesh network coordinator.

[0031] Preferably, the cloud-based piezoelectric impedance diagnostic module includes an excitation signal generation and conditioning unit, a cloud-based signal acquisition and analysis unit, and a remote spectrum database unit;

[0032] The excitation signal generation and conditioning unit includes a direct digital frequency synthesis function generator, a signal amplifier, and a digital programmable gain amplifier. The synchronization frame control terminal of the direct digital frequency synthesis function generator is used to receive cloud control commands, and its waveform signal output terminal is connected to the in-phase input terminal of the signal amplifier. The output terminal of the signal amplifier is connected to a virtual test probe, which is connected to the power supply output bus and load circuit in the respirator body. The gain amplitude of the signal amplifier is set through the parallel digital gain control interface of the digital programmable gain amplifier.

[0033] The cloud-based signal acquisition and analysis unit includes a high-precision instrumentation amplifier and a cloud-based field-programmable gate array (FPGA) server. The non-inverting input of the high-precision instrumentation amplifier is connected to the return signal of the virtual test probe, its inverting input is grounded, and its output outputs a differentially amplified signal through a high-frequency filter capacitor. The analog input port group of the cloud-based FPGA server is connected to the high-frequency filter capacitor.

[0034] The remote spectrum database unit includes a network physical layer transceiver chip, whose transmit data line group and receive data line group are respectively connected to the output data bus and input data bus of the cloud field programmable gate array server.

[0035] Preferably, the adaptive spectral threshold alarm module includes a spectral threshold calculation and control unit, an audible and visual alarm driving and feedback unit, and a probabilistic logic decision unit;

[0036] The spectral threshold calculation and control unit includes a microcontroller and an operational amplifier. The analog voltage output pin of the digital-to-analog converter in the microcontroller is connected to the non-inverting input of the operational amplifier through an integral filter circuit. The output of the operational amplifier is connected to the signal input of the power amplifier of the audible and visual alarm drive and feedback unit.

[0037] The sound and light alarm drive and feedback unit includes a power amplifier, a speaker, and a current-enhancing transistor. The output signal of the operational amplifier is connected to the non-inverting input of the power amplifier as a frequency control input. The output of the power amplifier is connected to one end of the speaker and supplies power to the speaker through the collector and emitter paths of the current-enhancing transistor. A first freewheeling diode is connected in parallel across the two ends of the speaker. The operating status feedback signal of the speaker is acquired through the first general-purpose input / output pin of the microcontroller.

[0038] The probabilistic logic decision unit includes a programmable logic chip, whose data input pin, clock input pin, and enable control pin are respectively connected to the corresponding communication interface pins of the microcontroller, and whose parallel data output pin group is connected to the external interrupt request pin of the microcontroller. The feedback signal collected by the first general-purpose input / output pin of the microcontroller is transmitted to the data input pin of the programmable logic chip.

[0039] Preferably, the autonomous magnetorheological backup module includes an energy conversion and storage unit, a power switching and control unit, and a self-diagnosis and oscillation feedback unit;

[0040] The energy conversion and storage unit includes a magnetorheological damper, an adjustable linear regulator, a ferrite inductor energy storage device, and a rectifier bridge. The first end of the electromagnetic coil of the magnetorheological damper is connected to the voltage output terminal of the adjustable linear regulator. The voltage adjustment terminal of the adjustable linear regulator outputs current through a fourth current-limiting resistor, and its voltage input terminal is energized through a P-channel MOSFET. A second freewheeling diode is connected in parallel across the two ends of the electromagnetic coil of the magnetorheological damper. The electrical signal output terminal of the magnetorheological damper is connected to the AC input terminal of the rectifier bridge through a pulse transformer. The DC positive output terminal of the rectifier bridge is connected to an energy storage capacitor through an energy storage inductor.

[0041] The power switching and control unit includes a switching controller field-effect transistor and a timer integrated circuit. The source of the switching controller field-effect transistor is connected to an energy storage capacitor, and the drain of the switching controller field-effect transistor is connected to the auxiliary power rail of the power supply in the respirator body. The pulse signal output terminal of the timer integrated circuit is connected to the gate of the switching controller field-effect transistor through a fifth current resistor.

[0042] The self-diagnosis and oscillation feedback unit includes a self-diagnosis oscillator and a feedback operational amplifier. The signal output terminal of the self-diagnosis oscillator is connected to the non-inverting input terminal of the feedback operational amplifier, and the output terminal of the feedback operational amplifier is connected to the gate of the field-effect transistor of the switch controller through a second buffer transistor.

[0043] The present invention also provides a method for operating an automatic transformer maintenance-free breather, based on the automatic transformer maintenance-free breather described above, comprising the following steps:

[0044] S1. Multi-source data synchronization and feature extraction steps: The magnetostriction and photoacoustic signals of the power supply in the respirator body are collected through the magnetostriction and photoacoustic signal capture module. The eddy current response signal of the power supply in the respirator body is collected through the eddy current test preprocessing module. The ambient temperature and humidity signals are collected simultaneously. The collected signals are time-aligned and resampled. The dimensionless physical feature vectors that are robust to the environment and related to the aging of the power supply in the respirator body are extracted.

[0045] S2. Local health assessment and preprocessing steps: In the eddy current test preprocessing module and data integrity verification module, the feature vector is preliminarily processed and verified. Based on the pre-established health benchmark model, the deviation of the current state from the health benchmark is calculated, and preliminary health status indicators and data integrity verification codes are generated.

[0046] S3. Reliable data transmission steps: The verified health status indicators and related characteristic data are reliably transmitted to the cloud via a wireless mesh network through redundant communication modules;

[0047] S4. Cloud-based in-depth diagnosis and prediction steps: In the cloud-based piezoresistive diagnosis module, historical data and real-time uploaded data are integrated to perform piezoresistive spectrum matching and in-depth analysis. Based on the time series prediction model, the probability distribution of the remaining service life of the power supply in the respirator body and the real-time risk score are calculated.

[0048] S5. Dynamic Decision-Making and Alarm Generation Steps: In the adaptive spectrum threshold alarm module, the local alarm threshold is dynamically adjusted based on the risk score and prediction results sent from the cloud. Based on fuzzy inference rules, local and cloud information are integrated to generate alarm or control commands of different priorities.

[0049] S6. Command Execution and Autonomous Backup Steps: When a high-risk or power failure command is generated, the autonomous magnetorheological backup module is activated to provide emergency power to the power supply in the respirator body and monitor the execution feedback to form a safety closed-loop control.

[0050] S7. Model self-update and system self-check steps: During normal system operation, use data increments from healthy periods to update the local and cloud-based health baseline models; continuously monitor the self-diagnostic signals of each functional module.

[0051] The present invention has the following beneficial effects:

[0052] This invention enables precise sensing and early diagnosis of subtle signs of progressive aging in the power supply of the respirator, ensuring reliable data reporting. It also performs in-depth analysis and predicts remaining lifespan, dynamically adjusts thresholds based on risk, and generates decision instructions. Finally, through an autonomous magnetorheological backup module, it provides emergency power or triggers maintenance alarms before the fault becomes critical. This completes a closed loop from latent aging identification and predictive early warning to active fault-tolerant protection without human intervention, effectively preventing respirator malfunction and subsequent transformer insulation accidents caused by power supply failure. Attached Figure Description

[0053] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 This is a schematic diagram of the structure of the present invention;

[0055] Figure 2 This is a structural block diagram of the monitoring and control system in this invention;

[0056] Figure 3 This is a flowchart of the present invention.

[0057] 1. Respirator body; 101. Housing; 102. Air pump; 103. Power supply; 104. Silicone canister; 201. Magnetostriction and photoacoustic signal capture module; 202. Eddy current test preprocessing module; 203. Data integrity verification module; 204. Redundant communication module; 205. Cloud-based piezoresistive impedance diagnosis module; 206. Adaptive spectrum threshold alarm module; 207. Autonomous magnetorheological backup module. Detailed Implementation

[0058] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0059] Therefore, the following detailed description of the embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the invention without inventive effort are within the scope of protection of the invention.

[0060] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0061] In the description of this invention, it should be understood that the terms "center," "upper," "lower," "left," "right," "vertical," "horizontal," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of this invention is in use, or the orientation or positional relationship commonly understood by those skilled in the art. They are only used to facilitate the description of this invention and to simplify the description, and are not intended to indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0062] Furthermore, the terms "first," "second," and "third" are used only to distinguish descriptions and should not be interpreted as indicating or implying relative importance.

[0063] In the description of this invention, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0064] An automatic transformer maintenance-free breather, such as Figure 1 and Figure 2As shown, the device includes a respirator body 1 and a monitoring and control system disposed therein. The monitoring and control system includes a magnetostriction and photoacoustic signal acquisition module 201, an eddy current test preprocessing module 202, a data integrity verification module 203, a redundant communication module 204, a cloud-based piezoresistive impedance diagnostic module 205, an adaptive spectrum threshold alarm module 206, and an autonomous magnetorheological backup module 207, which are connected in sequence. The detection terminals of the magnetostriction and photoacoustic signal acquisition module 201 are respectively located on the coil surface of the power supply 103 in the respirator body 1 and inside the housing 101. The hot zone of 03 is connected to the grounding housing of the air pump 102. The detection end of the eddy current test preprocessing module 202 is set above the circuit board of the power supply 103 in the respirator body 1. The detection end of the eddy current test preprocessing module 202 is connected to the expansion port of the temperature and humidity sensor in the respirator body 1 and located in the silicone canister 104. The detection end of the cloud piezoresistive impedance diagnostic module 205 is connected to the output bus and load circuit of the power supply 103 in the respirator body 1. The autonomous magnetorheological backup module 207 is connected to the auxiliary power rail of the power supply 103 in the respirator body 1.

[0065] Among them, the magnetostriction and photoacoustic signal acquisition module 201 collects the magnetostriction and photoacoustic signals of the power supply 103 in the respirator body 1 and extracts the physical characteristic quantities characterizing aging from them. Furthermore, the magnetostriction and photoacoustic signal acquisition module 201 includes a magnetostriction sensor unit, a photoacoustic effect detection unit, a signal conditioning and noise suppression unit, an analog-to-digital conversion unit, and a digital signal processing unit.

[0066] The magnetostrictive sensor unit includes a piezoelectric transducer, an induction coil, and an NPN bipolar transistor. The positive terminal of the piezoelectric transducer is connected to one end of the induction coil, and the other two ends of the induction coil are connected to the collector of the NPN bipolar transistor through a first current-limiting resistor. The base of the NPN bipolar transistor is connected to the ground through a base bias resistor.

[0067] The photoacoustic effect detection unit includes a photoacoustic detector and a low-noise operational amplifier. The synchronous trigger terminal of the photoacoustic detector is connected to the signal output node of the piezoelectric transducer through a parallel high-frequency filter capacitor and a reverse voltage protection diode. The cathode of the photoacoustic detector is connected to the inverting input terminal of the low-noise operational amplifier through an input resistor. The non-inverting input terminal of the low-noise operational amplifier is grounded, and its output terminal is connected to its inverting input terminal through a negative feedback resistor. The signal input node of the low-noise operational amplifier is connected in parallel with a phase compensation capacitor and a first voltage clamping diode.

[0068] The signal conditioning and noise suppression unit includes a high common-mode rejection ratio (CMRR) instrumentation amplifier. The non-inverting input of the CMRR instrumentation amplifier is connected to the output of a low-noise operational amplifier through a coupling resistor, and its inverting input is grounded. The output of the CMRR instrumentation amplifier outputs the conditioned signal through a first low-pass filter. The differential signal input of the CMRR instrumentation amplifier is connected to the output of the photoacoustic effect detection unit through an isolation transformer.

[0069] The analog-to-digital conversion unit includes an analog-to-digital converter. The positive terminal of the first analog differential input of the analog-to-digital converter is connected to the output terminal of a high common-mode rejection ratio instrumentation amplifier, and the negative terminal of its first analog differential input is grounded.

[0070] Among them, the eddy current test preprocessing module 202 collects the eddy current response signal of the power supply 103 in the respirator body 1 and preprocesses it to obtain impedance characteristics. Furthermore, the eddy current test preprocessing module 202 includes an eddy current excitation and detection unit, a signal isolation and amplification unit, a digital signal processing core unit, and a data buffer output unit.

[0071] The eddy current excitation and detection unit includes a bridge balancing circuit composed of an excitation coil and a detection coil, a function signal generator, and a PNP bipolar transistor. The output terminal of the function signal generator is connected to the common connection point of the excitation coil and the detection coil. The output amplitude adjustment terminal of the function signal generator is set by a first adjustable potentiometer. The emitter of the PNP bipolar transistor is energized, its collector is connected to the output terminal of the function signal generator, and its base is biased and controlled by a base drive resistor.

[0072] The signal isolation and amplification unit includes a programmable gain amplifier and a signal isolation transformer. The positive terminal of the differential input of the programmable gain amplifier is connected to the signal output node of the bridge balancing circuit, and its negative terminal is grounded. Its gain is set through its parallel digital control interface. The primary winding side of the signal isolation transformer is connected to the probe signal output terminal of the eddy current excitation and detection unit. A second voltage clamping diode is connected in series on the primary winding side of the signal isolation transformer, and an impedance matching resistor and a first buffer transistor are connected in parallel on the secondary winding side. The base of the first buffer transistor is connected to one end of the secondary winding of the signal isolation transformer, the collector of the first buffer transistor is connected to the positive terminal of the differential input of the programmable gain amplifier, and the emitter of the first buffer transistor is grounded.

[0073] The core unit of digital signal processing includes a microprocessor, whose analog signal acquisition port is connected to the single-ended output of a programmable gain amplifier. The first set of general-purpose input and output terminals of the microprocessor is connected to the clock signal input terminal and the data latch signal input terminal of the data buffer output unit through pull-down resistors.

[0074] The data buffer output unit includes a serial shift register. The serial data input terminal of the serial shift register is connected to the data bus of the microprocessor, and its parallel data output terminal is used to connect to the data integrity verification module 203.

[0075] The data integrity verification module 203 is used to verify the extracted feature vector and generate a data integrity verification code. Furthermore, the data integrity verification module 203 includes a Hamming code verification and processing unit, a secure hash operation unit, and a non-volatile storage unit.

[0076] The Hamming code verification and processing unit includes a field-programmable gate array (FPGA), and its front end is connected to a signal input buffer circuit composed of a low-offset operational amplifier. The non-inverting input of the low-offset operational amplifier receives the verification data signal from the eddy current test preprocessing module 202, its inverting input is grounded, and its output is connected to the general input / output port group of the FPGA through a second low-pass filter.

[0077] The secure hashing unit includes a hashing algorithm chip and a first precision reference voltage source. Its power supply 103 input terminal is powered through an NPN current-enhancing transistor. The base of the NPN current-enhancing transistor is connected to the cathode of the first precision reference voltage source. The anode of the first precision reference voltage source is grounded, and its cathode is powered through a second current-limiting resistor. A tantalum decoupling capacitor is connected in parallel with the second current-limiting resistor. A specific low-voltage complementary metal-oxide-semiconductor level output group of the field-programmable gate array is respectively connected to the data input terminal, clock input terminal, and enable control terminal of the hashing algorithm chip.

[0078] The non-volatile memory cell includes an electrically erasable programmable read-only memory chip, whose serial data line and serial clock line are powered through a first pull-up resistor, and whose write protection control terminal is connected to a general-purpose input / output terminal of a field-programmable gate array. The data line and clock line generated by the hash algorithm chip are connected to the corresponding serial interface terminal of the electrically erasable programmable read-only memory chip.

[0079] The redundant communication module 204 is used to reliably transmit the verified data to the cloud through the wireless Mesh network. Furthermore, the redundant communication module 204 includes a wireless transceiver unit, a mesh network control and processing unit, an RF front-end amplification and matching unit, and an antenna and switching unit.

[0080] The wireless transceiver unit includes an integrated radio frequency transceiver chip, whose serial peripheral interface's host input / slave output, host output / slave input, serial clock, and chip select terminals are connected to the corresponding host interface terminals of the mesh network control and processing unit.

[0081] The RF front-end amplification and matching unit includes an RF power amplifier. The signal input terminal of the RF power amplifier is connected to the RF output terminal of the integrated RF transceiver chip through a π-type impedance matching network. The power supply terminal 103 of the RF power amplifier is connected to the power supply through a third current-limiting resistor. A high-frequency decoupling capacitor is connected in parallel with the third current-limiting resistor. An electrostatic discharge protection diode is connected to the output terminal of the RF power amplifier.

[0082] The mesh network control and processing unit includes a mesh network coordinator, whose radio frequency signal output is connected to the signal input of the antenna and the switching unit via a first coaxial connector.

[0083] The antenna and switching unit includes a helical antenna, a balun, and an antenna path switching circuit. The unbalanced port of the balun is connected to the RF signal output of the mesh network coordinator, and its balanced port is connected to the feed center point of the helical antenna. The antenna path switching circuit includes a PIN diode and a PNP switching transistor. The anode of the PIN diode is connected to the RF signal, and the cathode of the PIN diode is energized through a first bias resistor. The emitter of the PNP switching transistor is energized, the collector of the PNP switching transistor is connected to the cathode of the PIN diode, and the base of the PNP switching transistor is connected to the general-purpose input / output terminal of the mesh network coordinator.

[0084] Among them, the cloud-based piezoresistive impedance diagnostic module 205 is used to integrate historical and real-time data for in-depth analysis to predict the remaining service life and calculate the risk score. Furthermore, the cloud-based piezoresistive impedance diagnostic module 205 includes an excitation signal generation and conditioning unit, a cloud-based signal acquisition and analysis unit, and a remote spectrum database unit.

[0085] The excitation signal generation and conditioning unit includes a direct digital frequency synthesis function generator, a signal amplifier, and a digital programmable gain amplifier. The synchronization frame control terminal of the direct digital frequency synthesis function generator is used to receive cloud control commands. Its waveform signal output terminal is connected to the in-phase input terminal of the signal amplifier. The output terminal of the signal amplifier is connected to a virtual test probe. The virtual test probe is connected to the output bus and load circuit of the power supply 103 in the respirator body 1. The gain amplitude of the signal amplifier is set through the parallel digital gain control interface of the digital programmable gain amplifier.

[0086] The cloud-based signal acquisition and analysis unit includes a high-precision instrumentation amplifier and a cloud-based field-programmable gate array (FPGA) server. The non-inverting input of the high-precision instrumentation amplifier is connected to the return signal of the virtual test probe, its inverting input is grounded, and its output outputs the differentially amplified signal through a high-frequency filter capacitor. The analog input port group of the cloud-based FPGA server is connected to the high-frequency filter capacitor.

[0087] The remote spectrum database unit includes a network physical layer transceiver chip, whose transmit data line group and receive data line group are respectively connected to the output data bus and input data bus of the cloud field programmable gate array server.

[0088] The adaptive spectrum threshold alarm module 206 is used to dynamically adjust the alarm threshold according to the risk score and fuse information to generate control commands. Furthermore, the adaptive spectrum threshold alarm module 206 includes a spectrum threshold calculation and control unit, an audible and visual alarm drive and feedback unit, and a probabilistic logic decision unit.

[0089] The spectrum threshold calculation and control unit includes a microcontroller and an operational amplifier. The analog voltage output pin of the digital-to-analog converter in the microcontroller is connected to the non-inverting input of the operational amplifier through an integral filter circuit. The output of the operational amplifier is connected to the signal input of the power amplifier of the audible and visual alarm drive and feedback unit.

[0090] The sound and light alarm drive and feedback unit includes a power amplifier, a speaker, and a current-enhancing transistor. The output signal of the operational amplifier is connected to the non-inverting input of the power amplifier as a frequency control input. The output of the power amplifier is connected to one end of the speaker and supplies power to the speaker through the collector and emitter path of the current-enhancing transistor. A first freewheeling diode is connected in parallel across the two ends of the speaker. The operating status feedback signal of the speaker is acquired through the first general-purpose input / output pin of the microcontroller.

[0091] The probabilistic logic decision unit includes a programmable logic chip, whose data input pin, clock input pin, and enable control pin are respectively connected to the corresponding communication interface pins of the microcontroller, and whose parallel data output pin group is connected to the external interrupt request pin of the microcontroller. The feedback signal collected by the first general-purpose input / output pin of the microcontroller is transmitted to the data input pin of the programmable logic chip.

[0092] Among them, the autonomous magnetorheological backup module 207 is activated when a high-risk command is received and captures ambient energy to provide emergency power to the power supply 103 in the respirator body 1. Furthermore, the autonomous magnetorheological backup module 207 includes an energy conversion and storage unit, a power switching and control unit, and a self-diagnosis and oscillation feedback unit.

[0093] The energy conversion and storage unit includes a magnetorheological damper, an adjustable linear regulator, a ferrite inductor energy storage device, and a rectifier bridge. The first end of the electromagnetic coil of the magnetorheological damper is connected to the voltage output terminal of the adjustable linear regulator. The voltage adjustment terminal of the adjustable linear regulator outputs current through a fourth current-limiting resistor, and its voltage input terminal is energized through a P-channel MOSFET. A second freewheeling diode is connected in parallel across the electromagnetic coil of the magnetorheological damper. The electrical signal output terminal of the magnetorheological damper is connected to the AC input terminal of the rectifier bridge through a pulse transformer. The DC positive output terminal of the rectifier bridge is connected to an energy storage capacitor through an energy storage inductor.

[0094] The power switching and control unit includes a switching controller field-effect transistor and a timer integrated circuit. The source of the switching controller field-effect transistor is connected to an energy storage capacitor, and the drain of the switching controller field-effect transistor is connected to the auxiliary power supply 103 rail of the power supply 103 in the respirator body 1. The pulse signal output terminal of the timer integrated circuit is connected to the gate of the switching controller field-effect transistor through a fifth current resistor.

[0095] The self-diagnostic and oscillation feedback unit includes a self-diagnostic oscillator and a feedback operational amplifier. The signal output terminal of the self-diagnostic oscillator is connected to the non-inverting input terminal of the feedback operational amplifier, and the output terminal of the feedback operational amplifier is connected to the gate of the field-effect transistor of the switch controller through a second buffer transistor.

[0096] To address the issue of gradual aging and eventual functional failure of the transformer breather power supply module 103 in long-term unattended substations due to continuous operation, environmental stress, and power grid disturbances, the monitoring and control system of the breather body 1 begins by capturing physical changes in the internal materials and electrical state of the power supply 103 during operation. The magnetostrictive and photoacoustic signal capture module 201 undertakes this task. The magnetostrictive sensor unit is attached to the surface of the power supply 103. When these magnetic materials undergo microscopic magnetic domain movement or mechanical deformation under long-term power supply heating and current stress, it will induce minute changes in the surface charge, thereby inducing a microvolt-level voltage signal in the induction coil connected in series.

[0097] At the same time, the photoacoustic detector emits a beam of light modulated by a specific frequency, which shines on the hot area of ​​the power supply 103 inside the housing 101. The periodic hot spots generated inside these components due to increased dielectric loss or increased contact resistance will cause extremely slight thermoelastic expansion at the location of the hot area, which will excite ultrasonic waves. The high-sensitivity electret microphone inside the photoacoustic detector converts it into an electrical signal.

[0098] These two weak analog signals, originating from different physical effects, then enter a preprocessing channel composed of discrete and integrated circuits. The magnetostrictive signal first flows into a common-emitter amplifier circuit based on an NPN bipolar transistor. The base of the transistor obtains a suitable bias voltage through a high-precision metal film resistor, which initially amplifies the signal by tens of times.

[0099] The photoacoustic signal is fed into a non-inverting amplifier circuit composed of a low-noise operational amplifier. A negative feedback resistor network is connected between the inverting input and output of the op-amp to set the precise gain, while its non-inverting input receives the signal through a DC blocking capacitor.

[0100] Afterward, the two signals are combined and enter the instrumentation amplifier, which can greatly suppress common-mode interference from power supply 103 or surrounding equipment. The amplified signal then passes through a low-pass filter to filter out the high-frequency noise generated by the switching power supply 103. The purified analog signal is acquired by the analog-to-digital converter, which converts the voltage value into digital code at a rate of thousands of times per second.

[0101] These digital streams are then fed into a field-programmable gate array (FPGA), where thousands of programmable logic units are configured as digital filters and signal processors to perform real-time calculations on the data and extract characteristic values ​​representing the signal envelope shape, harmonic content, and impulse response time.

[0102] In parallel with the above modules is the eddy current test preprocessing module 202. This module applies an alternating magnetic field to the surface of the circuit board of the power supply 103 through the excitation coil driven by the function generator. The oscillator and waveform synthesis circuit integrated inside the function generator generate a frequency-adjustable sine wave. This waveform drives the coil after passing through a current buffer stage composed of a PNP transistor.

[0103] In the adjacent detection coil, the distribution of the eddy current field is changed due to micro-cracks in the copper foil of the circuit board caused by thermal fatigue or separation between capacitor electrode layers, thereby inducing a differential voltage. This differential voltage is potential isolated by an isolation transformer and then sent to a programmable gain amplifier. The gain of the amplifier is set by several resistors connected to its logic control pins and can be automatically adjusted according to the signal strength.

[0104] The amplified signal is analyzed by a microprocessor using a fast Fourier transform. The arithmetic logic unit and a dedicated multiplier inside the microprocessor work together to calculate the real and imaginary parts of the impedance, thereby obtaining the impedance spectrum.

[0105] All feature data acquired from different sensor paths are then fed into the data integrity verification module 203, which first uses a programmable logic device to implement Hamming code verification and generate a check bit for each data word.

[0106] Meanwhile, the hash algorithm chip performs mathematical digest calculations on the data block. The chip contains shift registers and nonlinear logic functions to generate a unique fingerprint code.

[0107] The processed data frames are stored in an electrically erasable programmable read-only memory chip. After ensuring data reliability, the redundant communication module 204 starts working, and the integrated RF transceiver chip reads the data. The data is modulated onto the carrier using Gaussian frequency shift keying. The modulated RF signal passes through a π-type impedance matching network and enters an RF power amplifier to enhance the transmission strength. The multi-stage field-effect transistors inside the RF power amplifier amplify the signal. The entire wireless network is coordinated by a mesh network coordinator, which continuously monitors the channel and selects the optimal relay path based on the received signal strength and packet error rate.

[0108] Finally, the signal is radiated out through the helical antenna. After the data arrives at the remote cloud-based piezoresistive impedance diagnostic module 205, the virtual instrument in the cloud server simulates the function of the actual impedance analyzer through software algorithms. It compares the uploaded impedance spectrum data with a huge historical aging database, which adopts a distributed storage architecture.

[0109] The comparison process employs a particle filtering algorithm, which uses a large number of randomly sampled particles to simulate the possible evolution trajectory of the health of power supply 103, thereby estimating the probability density function of the remaining lifetime and extracting a conservative prediction value from it.

[0110] Combining the current degradation stage information obtained from cluster analysis, a quantitative comprehensive risk score is calculated. This risk score is sent back to the adaptive spectrum threshold alarm module 206 of the local device via the downlink. After receiving the score, the microcontroller in the module converts the digital score into an analog voltage through its integrated digital-to-analog converter. This voltage is then used by a voltage divider network to generate a reference level.

[0111] This reference level is fed into the non-inverting input of the operational amplifier, while the inverting input is connected to a signal from the local health index calculation circuit.

[0112] The health index is obtained by calculating the regularized Mahalanobis distance between all currently extracted feature values ​​and the health benchmark model stored in electrically erasable programmable read-only memory. This distance calculation involves inverting the benchmark covariance matrix and adding a small constant to ensure numerical stability.

[0113] When the health index signal falls below the reference level dynamically set by the risk score, the output state of the operational amplifier flips. This flip signal triggers a programmable logic device, which is pre-written with probabilistic inference rules based on Bayes' theorem to perform final confirmation and priority classification of the alarm.

[0114] Once the highest risk level is confirmed, the instruction is transmitted to the autonomous magnetorheological backup module 207. The adjustable linear regulator in this module starts to work, and its adjustment terminal output voltage drives the electromagnetic coil of the magnetorheological damper to generate a magnetic field, changing the viscosity of the magnetic fluid in the magnetorheological damper to efficiently capture the mechanical energy of the transformer vibration. The generated AC power is converted into DC power by the rectifier bridge, and then filtered and stored by the energy storage inductor and energy storage capacitor.

[0115] When a voltage drop in power supply 103 is detected, the level changes, driving the gate of the power MOSFET to turn it on, releasing the energy in the energy storage capacitor to the auxiliary power supply 103 rail.

[0116] This ensures that early signs of aging can be identified during the long incubation period that cannot be detected by routine manual inspections and simple remote monitoring, and that countermeasures can be taken before the function is completely lost. This maintains the ability of the respirator to continuously dry the air, isolates the path of moisture invading the oil conservator, protects the chemical and electrical properties of the transformer insulating oil, prevents catastrophic consequences such as short circuits, overheating, or even explosions caused by insulation deterioration, and ensures continuous power supply to the power grid.

[0117] This invention also provides a method for operating an automatic transformer maintenance-free breather, based on the above-mentioned automatic transformer maintenance-free breather, such as... Figure 3 As shown, it includes the following steps:

[0118] S1. Multi-source data synchronization and feature extraction steps: The magnetostriction and photoacoustic signals of the power supply 103 in the respirator body 1 are collected by the magnetostriction and photoacoustic signal capture module 201. The eddy current response signal of the power supply 103 in the respirator body 1 is collected by the eddy current test preprocessing module 202. The ambient temperature and humidity signals are collected simultaneously. The collected signals are time aligned and resampled. The dimensionless physical feature vectors that are robust to the environment and related to the aging of the power supply 103 in the respirator body 1 are extracted.

[0119] More specifically, extract the dimensionless physical feature vectors that are environmentally robust and related to the aging of the power supply module 103, including:

[0120] The magnetostriction signal acquired by the magnetostriction and photoacoustic signal acquisition module 201 is subjected to Hilbert transform to obtain its envelope, and the total harmonic distortion of the envelope signal spectrum is calculated as the first characteristic quantity characterizing the nonlinear aging of the magnetic core.

[0121] For the photoacoustic signal acquired by the magnetostriction and photoacoustic signal acquisition module 201, the first-order normalized spectral moment of its impulse response is calculated as the second characteristic quantity characterizing the thermal relaxation properties of local hot spots.

[0122] The impedance spectrum acquired by the eddy current test preprocessing module 202 is used to calculate the normalized Euclidean distance between the current trajectory center and the healthy reference center, which is used as the third characteristic quantity characterizing the change of conductor eddy current loss.

[0123] Principal component analysis was used to orthogonally decompose the synchronously collected environmental temperature and humidity data with the first, second, and third feature quantities to obtain the final feature vector decoupled from the environmental factors.

[0124] S2. Local health assessment and preprocessing steps: In the eddy current test preprocessing module 202 and the data integrity verification module 203, the feature vector is preliminarily processed and verified. Based on the pre-established health benchmark model, the deviation of the current state from the health benchmark is calculated, and preliminary health status indicators and data integrity verification codes are generated.

[0125] More specifically, it calculates the deviation of the current state from the health baseline, including:

[0126] Using the pre-stored health baseline mean vector and covariance matrix in the storage unit of the data integrity verification module 203, the regularized Mahalanobis distance of the current feature vector is calculated. The calculation method of the regularized Mahalanobis distance is as follows: first, calculate the difference vector between the current feature vector and the health baseline mean vector; then calculate the product of the difference vector and a weighted matrix, and then perform an inner product with the transpose of the difference vector; where the weighted matrix is ​​the inverse matrix of the sum of the health baseline covariance matrix and an identity matrix adjusted by the regularization parameter.

[0127] S3. Reliable data transmission steps: The verified health status indicators and related characteristic data are reliably transmitted to the cloud via a wireless mesh network through the redundant communication module 204;

[0128] More specifically, reliable transmission via wireless mesh networks includes:

[0129] The mesh network coordinator in the Sub-GHz redundant communication module 204 dynamically selects the transmission path based on real-time link quality indicators; Hamming encoding is performed on the data payload to be transmitted, and cyclic redundancy check codes are added to form data frames; data transmission is carried out in the Sub-GHz band using carrier sense multi-point access and collision avoidance mechanisms until cloud confirmation is received or the maximum number of retransmissions is reached.

[0130] S4. Cloud-based in-depth diagnosis and prediction steps: In the cloud-based piezoresistive diagnostic module 205, historical data and real-time uploaded data are integrated to perform piezoresistive spectrum matching and in-depth analysis. Based on the time series prediction model, the remaining service life probability distribution and real-time risk score of the power supply 103 in the respirator body 1 are calculated.

[0131] More specifically, the remaining lifespan probability distribution and real-time risk score of power supply module 103 are calculated, including:

[0132] The historical sequence of health status indicators is input into a particle filter-based prediction model to simulate a large number of possible future degradation trajectories; the moment when these trajectories first cross the preset failure threshold is statistically analyzed to form a probability distribution of the remaining service life, and the low quantile is taken as a conservative estimate.

[0133] The calculation of the real-time risk score is a linear combination of the S-shaped function mapping value of the conservative estimate of the remaining useful life and the weighted value of the current degradation severity; wherein, the current degradation severity is calculated by the confidence level of the state prototype obtained by cluster analysis and its preset weight.

[0134] S5. Dynamic decision-making and alarm generation steps: In the adaptive spectrum threshold alarm module 206, the local alarm threshold is dynamically adjusted according to the risk score and prediction results sent from the cloud. Based on fuzzy inference rules, local and cloud information are integrated to generate alarm or control commands with different priorities.

[0135] More specifically, dynamically adjusting local alarm thresholds includes:

[0136] The adaptive spectrum threshold alarm module 206 receives the real-time risk score sent from the cloud; it subtracts the product of the risk score and a positive coefficient from the preset baseline warning threshold to obtain the dynamic warning line; it subtracts the product of the risk score and another positive coefficient from the preset baseline alarm threshold to obtain the dynamic alarm line; when the locally calculated health status index is lower than the dynamic warning line or the dynamic alarm line, the corresponding level of alarm is triggered.

[0137] S6. Command execution and autonomous backup steps: When a high-risk or power supply 103 failure command is generated, the autonomous magnetorheological backup module 207 is activated to provide emergency power to the power supply 103 in the respirator body 1 and to monitor the execution feedback, forming a safety closed-loop control.

[0138] More specifically, the startup of the autonomous magnetorheological backup module 207 includes:

[0139] The high-risk command issued by the adaptive spectrum threshold alarm module 206 triggers the switch controller in the autonomous magnetorheological backup module 207.

[0140] The switch controller uses pulse width modulation to control the magnetorheological damper to capture vibration energy from the environment and charges the backup power supply 103 bus through rectification, filtering and energy storage circuits.

[0141] Meanwhile, the output voltage of the backup power supply 103 is monitored by the self-diagnosis and oscillation feedback unit to form a closed-loop control and ensure voltage stability;

[0142] S7. Model self-updating and system self-checking steps: During normal system operation, update the local and cloud-based health baseline models using incremental data from healthy periods; continuously monitor the self-diagnostic signals of each functional module;

[0143] More specifically, incremental updates to the local and cloud-based health benchmark models include:

[0144] When the system is determined to be in a healthy state and the environment is stable, an exponentially weighted moving average algorithm is used to iteratively update the mean vector and covariance matrix of the health benchmark in local storage and cloud database with a certain learning rate using the current feature vector data.

[0145] To address the problem caused by the progressive aging of the power supply module 103 of the breather in long-term unattended substations, this working method is based on the automatic detection of the maintenance-free breather of the transformer, and performs multi-source data synchronization and feature extraction steps. The magnetostrictive and photoacoustic signal capture module 201 is responsible for collecting the unique physical signals generated by the power supply module 103 during operation, while the eddy current test preprocessing module 202 obtains the electrical response characteristics of its internal conductors.

[0146] This raw information is processed and transformed into a set of digital features that can resist environmental interference and directly reflect the aging process.

[0147] Next, in the local health assessment and preprocessing steps, the data integrity verification module 203 verifies and encapsulates these feature quantities to ensure the reliability of subsequent analysis, and calculates the real-time health status deviation based on the pre-stored health benchmark model to form a preliminary quantitative assessment index.

[0148] The next step is reliable data transmission. The redundant communication module 204 uses a self-organizing wireless mesh network to continuously and stably send local status indicators and feature data to the remote computing center.

[0149] In the cloud-based deep diagnostic and prediction step, the cloud-based piezoelectric impedance diagnostic module 205 gathers historical and real-time data, simulates the degradation trajectory of the power supply module 103 through a complex algorithm model, accurately calculates the probabilistic distribution of its remaining service life, and generates an objective real-time risk score.

[0150] Once this score is issued, the dynamic decision-making and alarm generation steps are initiated. The adaptive spectrum threshold alarm module 206 receives the score and dynamically adjusts the trigger thresholds for local warnings and alarms based on its value, so that the sensitivity of the alarm to risk can change synchronously with the actual threat level. At the same time, the module integrates information from multiple sources to form clear instructions.

[0151] When the command determines that the risk is extremely high, the command execution and autonomous backup steps are immediately activated; the autonomous magnetorheological backup module 207 starts working, drawing mechanical energy from the surrounding environment and converting it into electrical energy, thereby providing temporary emergency power to the critical circuits in the event of a failure of the main power supply 103, ensuring that the monitoring and protection functions are not interrupted.

[0152] Throughout the process, there are also model self-updating and system self-checking steps. This method uses confirmed health data to continuously fine-tune its health baseline model to adapt to the normal drift of the equipment, and continuously monitors the operating status of each functional module to ensure that the entire monitoring system is always reliable.

[0153] It effectively prevented the loss of breather function caused by the latent failure of power supply 103, avoided the deterioration of insulating oil due to moisture, and ultimately prevented the chain reaction of serious electrical faults in the transformer, ensuring the safety and economy of power grid operation.

[0154] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An automatic detection transformer maintenance-free respirator, characterized by, Includes the respirator body (1) and a monitoring and control system installed therein; The monitoring and control system includes a magnetostrictive and photoacoustic signal capture module (201), an eddy current test preprocessing module (202), a data integrity verification module (203), a cloud-based piezoresistive impedance diagnosis module (205), an adaptive spectrum threshold alarm module (206), and an autonomous magnetorheological backup module (207) connected in sequence. The detection end of the magnetostrictive and photoacoustic signal capture module (201) is respectively set on the coil surface of the power supply (103) in the respirator body (1), the hot zone of the power supply (103) in the housing (101), and connected to the grounding shell of the air pump (102). The detection end of the eddy current test preprocessing module (202) is located above the circuit board of the power supply (103) in the respirator body (1). The detection end of the eddy current test preprocessing module (202) is connected to the expansion port of the temperature and humidity sensor in the respirator body (1) and located in the silicone canister (104). The detection end of the cloud-based piezoresistive impedance diagnostic module (205) is connected to the output bus and load circuit of the power supply (103) in the respirator body (1). The autonomous magnetorheological backup module (207) is connected to the auxiliary power rail of the power supply (103) in the respirator body (1). The magnetostriction and photoacoustic signal acquisition module (201) acquires the magnetostriction and photoacoustic signals of the power supply (103) in the respirator body (1) and extracts physical characteristic quantities characterizing aging from them. It performs a Hilbert transform on the acquired magnetostriction signal to obtain its envelope and calculates the total harmonic distortion of the envelope signal spectrum as the first characteristic quantity characterizing the nonlinear aging of the magnetic core. It calculates the first-order normalized spectral moment of the impulse response of the acquired photoacoustic signal as the second characteristic quantity characterizing the thermal relaxation characteristics of local overheating points. The eddy current test preprocessing module (202) acquires the eddy current response signal of the power supply (103) in the respirator body (1) and preprocesses it to obtain impedance characteristics. It calculates the impedance spectrum of the acquired signal. The normalized Euclidean distance between the current trajectory center and the health benchmark center is used as the third characteristic quantity to characterize the change of conductor eddy current loss. The principal component analysis method is used to orthogonally decompose the synchronously collected environmental temperature and humidity data with the first, second and third characteristic quantities to obtain the final feature vector decoupled from the environmental factors. The data integrity verification module (203) is used to verify the extracted feature vector to generate a data integrity verification code. The cloud-based piezoelectric impedance diagnosis module (205) is used to fuse historical and real-time data for in-depth analysis to predict the remaining service life and calculate the risk score. The adaptive spectrum threshold alarm module (206) is used to dynamically adjust the alarm threshold according to the risk score and fuse information to generate control commands.

2. The self-detecting, maintenance-free breathing apparatus for a transformer according to claim 1, wherein The monitoring and control system also includes a redundant communication module (204) for connecting the data integrity verification module (203) and the cloud-based piezoresistive impedance diagnostic module (205). The adaptive spectrum threshold alarm module (206) is connected to the autonomous magnetorheological backup module (207). The redundant communication module (204) is used to reliably transmit the verified data to the cloud via a wireless mesh network. The autonomous magnetorheological backup module (207) is activated when a high-risk command is received and captures ambient energy to provide emergency power to the power supply (103) in the respirator body (1).

3. The self-sufficient, maintenance-free breathing apparatus for a transformer according to claim 2, characterized in that The magnetostriction and photoacoustic signal capture module (201) includes a magnetostriction sensor unit, a photoacoustic effect detection unit, a signal conditioning and noise suppression unit, an analog-to-digital conversion unit, and a digital signal processing unit; The magnetostrictive sensor unit includes a piezoelectric transducer, an induction coil, and an NPN bipolar transistor. The positive terminal of the piezoelectric transducer is connected to one end of the induction coil, and the other two ends of the induction coil are connected to the collector of the NPN bipolar transistor through a first current-limiting resistor. The base of the NPN bipolar transistor is connected to the ground through a base bias resistor. The photoacoustic effect detection unit includes a photoacoustic detector and a low-noise operational amplifier. The synchronous trigger terminal of the photoacoustic detector is connected to the signal output node of the piezoelectric transducer through a parallel high-frequency filter capacitor and a reverse voltage protection diode. The cathode of the photoacoustic detector is connected to the inverting input terminal of the low-noise operational amplifier through an input resistor. The non-inverting input terminal of the low-noise operational amplifier is grounded, and its output terminal is connected to its inverting input terminal through a negative feedback resistor. The signal input node of the low-noise operational amplifier is connected in parallel with a phase compensation capacitor and a first voltage clamping diode. The signal conditioning and noise suppression unit includes a high common-mode rejection ratio (HMR) instrumentation amplifier. The non-inverting input of the HMR instrumentation amplifier is connected to the output of a low-noise operational amplifier via a coupling resistor, and its inverting input is grounded. The output of the HMR instrumentation amplifier outputs a conditioned signal through a first low-pass filter. The differential signal input of the HMR instrumentation amplifier is connected to the output of a photoacoustic effect detection unit via an isolation transformer. The analog-to-digital conversion unit includes an analog-to-digital converter. The positive terminal of the first analog differential input of the analog-to-digital converter is connected to the output terminal of a high common-mode rejection ratio instrumentation amplifier, and the negative terminal of its first analog differential input is grounded.

4. The automatic detection transformer maintenance-free respirator according to claim 2, characterized in that, The eddy current test preprocessing module (202) includes an eddy current excitation and detection unit, a signal isolation and amplification unit, a digital signal processing core unit, and a data buffer output unit. The eddy current excitation and detection unit includes a bridge balance circuit composed of an excitation coil and a detection coil, a function signal generator, and a PNP bipolar transistor. The output terminal of the function signal generator is connected to the common connection point of the excitation coil and the detection coil. The output amplitude adjustment terminal of the function signal generator is set by a first adjustable potentiometer. The emitter of the PNP bipolar transistor is energized, its collector is connected to the output terminal of the function signal generator, and its base is biased and controlled by a base drive resistor. The signal isolation and amplification unit includes a programmable gain amplifier and a signal isolation transformer. The positive terminal of the differential input of the programmable gain amplifier is connected to the signal output node of the bridge balanced circuit, and its negative terminal is grounded. Its gain is set through its parallel digital control interface. The primary winding side of the signal isolation transformer is connected to the probe signal output terminal of the eddy current excitation and detection unit. A second voltage clamping diode is connected in series on the primary winding side of the signal isolation transformer, and an impedance matching resistor and a first buffer transistor are connected in parallel on the secondary winding side. The base of the first buffer transistor is connected to one end of the secondary winding of the signal isolation transformer, the collector of the first buffer transistor is connected to the positive terminal of the differential input of the programmable gain amplifier, and the emitter of the first buffer transistor is grounded. The digital signal processing core unit includes a microprocessor, whose analog signal acquisition port is connected to the single-ended output of a programmable gain amplifier. The first set of general-purpose input and output terminals of the microprocessor is connected to the clock signal input terminal and the data latch signal input terminal of the data buffer output unit through pull-down resistors. The data buffer output unit includes a serial shift register. The serial data input terminal of the serial shift register is connected to the data bus of the microprocessor, and its parallel data output terminal is used to connect to the data integrity verification module (203).

5. The automatic detection transformer maintenance-free respirator according to claim 2, characterized in that, The data integrity verification module (203) includes a Hamming code verification and processing unit, a secure hash operation unit, and a non-volatile storage unit; The Hamming code verification and processing unit includes a field-programmable gate array (FPGA), and its front end is connected to a signal input buffer circuit composed of a low offset operational amplifier. The non-inverting input terminal of the low offset operational amplifier receives the verification data signal from the eddy current test preprocessing module (202), its inverting input terminal is grounded, and its output terminal is connected to the general input / output port group of the FPGA through a second low-pass filter. The secure hashing unit includes a hashing algorithm chip and a first precision reference voltage source. Its power input terminal is connected to an NPN current-enhancing transistor. The base of the NPN current-enhancing transistor is connected to the cathode of the first precision reference voltage source. The anode of the first precision reference voltage source is grounded, and its cathode is connected to a second current-limiting resistor. A tantalum decoupling capacitor is connected in parallel with the second current-limiting resistor. The specific low-voltage complementary metal-oxide-semiconductor level output terminal group of the field-programmable gate array is respectively connected to the data input terminal, clock input terminal, and enable control terminal of the hashing algorithm chip. The non-volatile memory cell includes an electrically erasable programmable read-only memory chip, whose serial data line and serial clock line are powered through a first pull-up resistor, and whose write protection control terminal is connected to a general-purpose input / output terminal of a field-programmable gate array. The data line and clock line generated by the hash algorithm chip are connected to the corresponding serial interface terminal of the electrically erasable programmable read-only memory chip.

6. The automatic detection transformer maintenance-free respirator according to claim 2, characterized in that, The redundant communication module (204) includes a wireless transceiver unit, a mesh network control and processing unit, an RF front-end amplification and matching unit, and an antenna and switching unit. The wireless transceiver unit includes an integrated radio frequency transceiver chip, whose serial peripheral interface's host input / slave output terminal, host output / slave input terminal, serial clock terminal, and chip select terminal are connected to the corresponding host interface terminal of the mesh network control and processing unit. The radio frequency front-end amplification and matching unit includes a radio frequency power amplifier. The signal input terminal of the radio frequency power amplifier is connected to the radio frequency output terminal of the integrated radio frequency transceiver chip through a π-type impedance matching network. The power supply terminal of the radio frequency power amplifier is connected to the power supply through a third current-limiting resistor. A high-frequency decoupling capacitor is connected in parallel with the third current-limiting resistor. An electrostatic discharge protection diode is connected to the output terminal of the radio frequency power amplifier. The mesh network control and processing unit includes a mesh network coordinator, whose radio frequency signal output terminal is connected to the signal input terminal of the antenna and the switching unit through a first coaxial connector; The antenna and switching unit includes a helical antenna, a balun, and an antenna path switching circuit. The unbalanced port of the balun is connected to the RF signal output of the mesh network coordinator, and its balanced port is connected to the feed center point of the helical antenna. The antenna path switching circuit includes a PIN diode and a PNP switching transistor. The anode of the PIN diode is connected to the RF signal, the cathode of the PIN diode is energized through a first bias resistor, the emitter of the PNP switching transistor is energized, the collector of the PNP switching transistor is connected to the cathode of the PIN diode, and the base of the PNP switching transistor is connected to the general-purpose input / output terminal of the mesh network coordinator.

7. The automatic detection transformer maintenance-free respirator according to claim 2, characterized in that, The cloud-based piezoresistive impedance diagnostic module (205) includes an excitation signal generation and conditioning unit, a cloud-based signal acquisition and analysis unit, and a remote spectrum database unit; The excitation signal generation and conditioning unit includes a direct digital frequency synthesis function generator, a signal amplifier, and a digital programmable gain amplifier. The synchronization frame control terminal of the direct digital frequency synthesis function generator is used to receive cloud control commands. Its waveform signal output terminal is connected to the in-phase input terminal of the signal amplifier. The output terminal of the signal amplifier is connected to a virtual test probe. The virtual test probe is connected to the output bus and load circuit of the power supply (103) in the respirator body (1). The gain amplitude of the signal amplifier is set through the parallel digital gain control interface of the digital programmable gain amplifier. The cloud-based signal acquisition and analysis unit includes a high-precision instrumentation amplifier and a cloud-based field-programmable gate array (FPGA) server. The non-inverting input of the high-precision instrumentation amplifier is connected to the return signal of the virtual test probe, its inverting input is grounded, and its output outputs a differentially amplified signal through a high-frequency filter capacitor. The analog input port group of the cloud-based FPGA server is connected to the high-frequency filter capacitor. The remote spectrum database unit includes a network physical layer transceiver chip, whose transmit data line group and receive data line group are respectively connected to the output data bus and input data bus of the cloud field programmable gate array server.

8. The automatic detection transformer maintenance-free respirator according to claim 2, characterized in that, The adaptive spectral threshold alarm module (206) includes a spectral threshold calculation and control unit, an audible and visual alarm driving and feedback unit, and a probabilistic logic decision unit; The spectral threshold calculation and control unit includes a microcontroller and an operational amplifier. The analog voltage output pin of the digital-to-analog converter in the microcontroller is connected to the non-inverting input of the operational amplifier through an integral filter circuit. The output of the operational amplifier is connected to the signal input of the power amplifier of the audible and visual alarm drive and feedback unit. The sound and light alarm drive and feedback unit includes a power amplifier, a speaker, and a current-enhancing transistor. The output signal of the operational amplifier is connected to the non-inverting input of the power amplifier as a frequency control input. The output of the power amplifier is connected to one end of the speaker and supplies power to the speaker through the collector and emitter paths of the current-enhancing transistor. A first freewheeling diode is connected in parallel across the two ends of the speaker. The operating status feedback signal of the speaker is acquired through the first general-purpose input / output pin of the microcontroller. The probabilistic logic decision unit includes a programmable logic chip, whose data input pin, clock input pin, and enable control pin are respectively connected to the corresponding communication interface pins of the microcontroller, and whose parallel data output pin group is connected to the external interrupt request pin of the microcontroller. The feedback signal collected by the first general-purpose input / output pin of the microcontroller is transmitted to the data input pin of the programmable logic chip.

9. The automatic detection transformer maintenance-free respirator according to claim 2, characterized in that, The autonomous magnetorheological backup module (207) includes an energy conversion and storage unit, a power switching and control unit, and a self-diagnosis and oscillation feedback unit; The energy conversion and storage unit includes a magnetorheological damper, an adjustable linear regulator, a ferrite inductor energy storage device, and a rectifier bridge. The first end of the electromagnetic coil of the magnetorheological damper is connected to the voltage output terminal of the adjustable linear regulator. The voltage adjustment terminal of the adjustable linear regulator outputs current through a fourth current-limiting resistor, and its voltage input terminal is energized through a P-channel MOSFET. A second freewheeling diode is connected in parallel across the two ends of the electromagnetic coil of the magnetorheological damper. The electrical signal output terminal of the magnetorheological damper is connected to the AC input terminal of the rectifier bridge through a pulse transformer. The DC positive output terminal of the rectifier bridge is connected to an energy storage capacitor through an energy storage inductor. The power switching and control unit includes a switch controller field-effect transistor and a timer integrated circuit. The source of the switch controller field-effect transistor is connected to an energy storage capacitor, and the drain of the switch controller field-effect transistor is connected to the auxiliary power rail of the power supply (103) in the respirator body (1). The pulse signal output terminal of the timer integrated circuit is connected to the gate of the switch controller field-effect transistor through a fifth current resistor. The self-diagnosis and oscillation feedback unit includes a self-diagnosis oscillator and a feedback operational amplifier. The signal output terminal of the self-diagnosis oscillator is connected to the non-inverting input terminal of the feedback operational amplifier, and the output terminal of the feedback operational amplifier is connected to the gate of the field-effect transistor of the switch controller through a second buffer transistor.

10. A method for operating an automatic transformer maintenance-free breather, based on the automatic transformer maintenance-free breather as described in any one of claims 2-9, characterized in that, Includes the following steps: S1. Multi-source data synchronization and feature extraction steps: The magnetostriction signal and photoacoustic signal of the power supply (103) in the respirator body (1) are collected by the magnetostriction and photoacoustic signal acquisition module (201). The Hilbert transform is performed on the collected magnetostriction signal to obtain its envelope, and the total harmonic distortion of the envelope signal spectrum is calculated as the first characteristic quantity characterizing the nonlinear aging of the magnetic core. The first normalized spectral moment of the impulse response of the collected photoacoustic signal is calculated as the second characteristic quantity characterizing the thermal relaxation characteristics of the local overheating point. The power supply (103) of the respirator body (1) is collected by the eddy current test preprocessing module (202). The eddy current response signal of the source (103) is used to calculate the normalized Euclidean distance between the current trajectory center and the health reference center of the collected impedance spectrum. This distance is used as the third characteristic quantity to characterize the change of conductor eddy current loss. At the same time, the ambient temperature and humidity signals are collected. The collected signals are time-aligned and resampled. The principal component analysis method is used to orthogonally decompose the synchronously collected ambient temperature and humidity data with the first, second and third characteristic quantities to obtain the final feature vector decoupled from the environmental factors. The dimensionless physical feature vector that is robust to the environment and related to the aging of the power supply (103) in the respirator body (1) is extracted. S2. Local health assessment and preprocessing steps: In the eddy current test preprocessing module (202) and data integrity verification module (203), the feature vector is preliminarily processed and verified. Based on the pre-established health benchmark model, the deviation of the current state from the health benchmark is calculated, and preliminary health status indicators and data integrity verification codes are generated. S3. Reliable data transmission steps: The verified health status indicators and related characteristic data are reliably transmitted to the cloud via a wireless mesh network through the redundant communication module (204); S4. Cloud-based in-depth diagnosis and prediction steps: In the cloud-based piezoresistive diagnosis module (205), historical data and real-time uploaded data are integrated to perform piezoresistive spectrum matching and in-depth analysis. Based on the time series prediction model, the remaining service life probability distribution and real-time risk score of the power supply (103) in the respirator body (1) are calculated. S5. Dynamic decision-making and alarm generation steps: In the adaptive spectrum threshold alarm module (206), the local alarm threshold is dynamically adjusted according to the risk score and prediction results sent from the cloud. Based on the fuzzy reasoning rules, local and cloud information are integrated to generate alarm or control instructions with different priorities. S6. Command execution and autonomous backup steps: When a high-risk or power supply (103) failure command is generated, the autonomous magnetorheological backup module (207) is activated to provide emergency power to the power supply (103) in the respirator body (1) and monitor the execution feedback to form a safety closed-loop control. S7. Model self-update and system self-check steps: During normal system operation, use data increments from healthy periods to update the local and cloud-based health baseline models; continuously monitor the self-diagnostic signals of each functional module.

Images