An output voltage detection method and system based on an aviation power supply

Abstract

The application discloses an output voltage detection method and system based on an aviation power supply, and relates to the technical field of aviation power supply voltage detection; a differential voltage sensor is connected to the output end of the aviation power supply, a high-frequency suppression capacitor is connected in parallel, a sampling rate of 20-50 times the working frequency is set according to the type of the power supply, analog voltage signals are collected and converted, two-stage filtering is performed on the analog signals, the analog signals are amplified to 0-5V by an instrument amplifier, and then the analog signals are converted into digital voltage sequences by a 16-bit ADC; the effective value, peak value and other parameters of the alternating-current power supply, the average value, ripple coefficient and other parameters of the direct-current power supply are calculated; the characteristic parameters are compared with preset threshold values, and if the characteristic parameters exceed the threshold values, it is determined that an abnormality occurs and relevant information is recorded; the detection results are transmitted to a monitoring terminal through RS485, data is stored locally, and when an abnormality occurs, an output switch quantity is output to an alarm device. The application is adapted to aviation power supply load fluctuation and environmental temperature change, has high detection precision, safe data transmission and storage, can accurately identify an abnormality and give an early warning, and improves the reliability and efficiency of aviation power supply operation and maintenance.

Description

Technical Field

[0001] This invention relates to the field of aviation power supply voltage detection technology, and in particular to a method and system for detecting the output voltage of aviation power supplies. Background Technology

[0002] Aviation power supplies are the core power source for aircraft and airport ground support equipment. The stability of their output voltage directly determines the normal operation of avionics systems, flight control equipment, and cabin facilities, and even affects flight safety. Existing aviation power supply voltage detection technologies have significant shortcomings in anti-interference capabilities: complex interference sources exist in aviation scenarios, such as high-frequency spike interference generated by the start-up and shutdown of electrical equipment in the cabin, and 50Hz / 60Hz power frequency interference introduced by the airport ground power grid. Traditional detection equipment often uses single-stage filtering with fixed parameters, which is insufficient for suppressing high-frequency spikes and cannot dynamically adapt to drifting power frequency interference. This results in a low signal-to-noise ratio of the acquired voltage signal, leading to large deviations in the calculation of characteristic parameters such as RMS value and harmonic distortion rate—for example, the actual harmonic distortion rate of an AC 115V / 400Hz power supply is 3%, but the detected value may reach 6% due to interference, leading to misjudgment as an abnormal state. Simultaneously, the sampling rate configuration of traditional equipment is rigid, often fixed at 10kHz, without adaptation to power supply type, resulting in redundancy in DC power supply detection and loss of AC power supply information.

[0003] Existing detection methods' threshold determination mechanisms are ill-suited to the dynamic load characteristics of aviation power supplies. Aviation power supply loads fluctuate dramatically; for example, the load on avionics systems and engines surges during takeoff and decreases sharply during landing. Traditional detection methods use fixed thresholds, failing to consider the impact of load changes on voltage. When the load surges, the voltage may briefly drop to 108V (still within a safe range), but this is misjudged as abnormal because it exceeds the fixed threshold, triggering unnecessary alarms. Conversely, if the voltage slowly shifts to 121V after the load stabilizes, it may be missed because it hasn't exceeded the threshold. Furthermore, phase imbalance in multiphase aviation power supplies is a significant fault type. However, traditional detection methods often rely on single-channel independent testing, lacking a synchronous sampling mechanism. The sampling time deviation between channels can reach 10-20μs, making accurate calculation of phase differences impossible and leading to missed detections of phase imbalances. This can potentially cause overheating and damage to three-phase loads (such as air conditioning compressors).

[0004] Existing systems also have limitations in maintaining accuracy and data security: The aviation environment experiences significant temperature fluctuations, and traditional testing equipment lacks integrated temperature compensation mechanisms. For every 10°C change in temperature, testing accuracy may decrease by 0.1%FS. Over long-term use, accuracy deviations accumulate, requiring periodic manual calibration, which is costly and cannot guarantee accuracy within calibration intervals. Furthermore, aviation power supply voltage data is critical operational information, affecting equipment status and flight safety. Traditional testing data transmission often uses unencrypted RS485 communication, and storage lacks access control, posing a risk of data tampering and leakage. For example, unauthorized modification of ground power supply testing data could mask voltage anomalies, leading to misjudgments of equipment status during maintenance; data leaks could expose airport power supply layout and equipment operating patterns, posing security risks. In addition, traditional systems lack channel fault redundancy design. If a certain acquisition channel fails, system shutdown and replacement are required to resume testing, interrupting power supply monitoring and impacting operational continuity. Summary of the Invention

[0005] This invention proposes a method and system for detecting the output voltage of an aviation power supply to solve the problems mentioned in the prior art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a method for detecting the output voltage of an aviation power supply, comprising:

[0007] Signal acquisition steps: A differential voltage sensor is connected to the output of the aviation power supply. A high-frequency suppression capacitor is connected in parallel at the input of the sensor to suppress spike interference. The sampling rate is set to 20-50 times the operating frequency of the power supply. The voltage signal is continuously acquired and converted into an analog voltage signal.

[0008] Signal preprocessing steps: The analog voltage signal is filtered in two stages. The first stage is an RC low-pass filter to remove high-frequency noise, and the second stage is an adaptive notch filter to suppress power frequency interference. The filtered signal is amplified to 0-5V by an instrumentation amplifier, and then converted from analog to digital by a 16-bit ADC to obtain a digital voltage sequence.

[0009] Feature extraction steps: Calculate feature parameters for the digital voltage sequence. For AC power supplies, these include RMS value, peak value, frequency deviation, and harmonic distortion rate; for DC power supplies, these include average value, ripple coefficient, and fluctuation amount.

[0010] Anomaly detection steps: Compare the extracted feature parameters with preset thresholds. AC power thresholds include effective value deviation ±5%, frequency deviation ±10Hz, and harmonic distortion rate ≤5%. DC power thresholds include average value deviation ±10%, ripple coefficient ≤1%, and fluctuation ≤0.5V. If any feature parameter exceeds the threshold, it is determined to be an anomaly. Record the time of occurrence, duration, and corresponding feature value of the anomaly.

[0011] Data output steps: The detection results are transmitted to the aviation power monitoring terminal via digital signals and stored in the local memory. When an abnormal state is triggered, a switch signal is output to the alarm device simultaneously.

[0012] Furthermore, it also includes:

[0013] Dynamic threshold adjustment steps: The anomaly detection threshold is corrected in real time based on power load changes. First, the load type is identified by the current harmonic content, and a differentiated load influence coefficient k is configured for different loads. Then, the formula is used... Calculate the adjusted voltage RMS threshold, where U th The threshold value is the adjusted threshold, where U0 is the initial threshold, t0 is the detection start time, and t is the current time. Let U be the load current at time τ, and the integral term reflects the cumulative load change. After threshold adjustment, continuous monitoring for 3 seconds is performed. If the characteristic parameter still exceeds U... th This is then considered abnormal, avoiding misjudgments caused by instantaneous load fluctuations.

[0014] Inter-channel synchronous detection steps: Configure 2-4 independent acquisition channels, each channel is triggered to sample by a synchronous clock; after synchronous sampling, calculate the phase difference between each channel. When the phase difference is >5°, calculate the transmission delay based on the signal transmission distance between channels to generate a phase compensation value, and correct the phase data during feature extraction; at the same time, monitor the sampled values ​​of each channel in real time. If 100 consecutive sampling points of a channel exceed the normal voltage range, it is determined to be a channel fault. Automatically switch to the backup channel and record the fault channel number and fault time; calculate the cross-correlation coefficient of the digital voltage sequence of each channel.

[0015] Furthermore, in the signal preprocessing step, the adaptive notch filter uses a recursive least squares algorithm to dynamically adjust the filter coefficients. The initial filter coefficient vector is updated every 10 sampling points to ensure that the filter attenuates ≥40dB for 50Hz / 60Hz power frequency interference and ≤1dB for the power supply fundamental frequency. At the same time, a noise intensity detection step is added. By calculating the signal-to-noise ratio (SNR) of the analog voltage signal before filtering, when SNR < 30dB, the notch filter order is automatically increased from 5th to 8th order to enhance interference suppression capability; when SNR > 40dB, the filter order is decreased from 5th to 3rd order to reduce the attenuation of the fundamental frequency signal.

[0016] Furthermore, in the anomaly detection step, the continuous abnormal state is classified into three levels: Level 1 anomaly is when a single feature parameter exceeds the threshold and the deviation is ≤10%; Level 2 anomaly is when a single feature parameter deviation is >10% or two feature parameters exceed the threshold; Level 3 anomaly is when the feature parameter deviation is >20% or accompanied by voltage mutation; at the same time, a historical data comparison mechanism is added to retrieve the statistical values ​​of feature parameters for the same period in the past 7 days.

[0017] Furthermore, the data output steps also include data compression and encryption, and a new edge computing preprocessing step is added: For locally stored voltage data, a preliminary analysis is first performed by the edge computing module to extract the average, maximum, minimum, and effective values ​​every 5 seconds, storing only the statistical values ​​and the raw sampled data during abnormal periods, and then compression is performed using piecewise linear fitting; The data transmitted to the monitoring terminal is first filtered by edge computing, uploading only the statistical values, abnormal records, and prediction information; In the encryption process, the AES-128 encryption algorithm is used to process the transmitted data, and the key is distributed through an asymmetric encryption mechanism: During system initialization, the monitoring terminal generates an RSA public key and a private key, sends the public key to the detection device, the detection device generates an AES key and encrypts it with the RSA public key, and then transmits it to the monitoring terminal, which decrypts it with the private key to obtain the AES key, and the key is automatically updated every hour.

[0018] Furthermore, it also includes:

[0019] The calibration process for testing accuracy is as follows: It is performed automatically once a month, and a new temperature compensation and calibration result verification mechanism has been added: Before calibration, the current ambient temperature T is collected by the system's built-in temperature sensor, and the temperature correction value ΔU is calculated based on the preset temperature-error model; then, a standard voltage source is connected, and each standard voltage value is output sequentially. The actual detection value Udet of the testing device is collected, and the difference between the detection value and the standard value Ustd is calculated as ΔUraw=Udet-Ustd. The temperature correction value is then added to obtain the final error ΔUfinal=ΔUraw+ΔU, generating a calibration coefficient matrix, which is stored in the non-volatile memory EEPROM; During daily testing, the ambient temperature is collected in real time, and the corresponding correction value in the calibration coefficient matrix is ​​called according to the current temperature and the detection voltage value to correct the detection result; after calibration, the intermediate value of the standard voltage source is connected for verification, and the corrected detection value is collected.

[0020] A system for implementing the aforementioned output voltage detection method based on an aviation power supply includes:

[0021] Signal acquisition module: It consists of a differential voltage sensor, a high-frequency suppression capacitor, and a sampling trigger circuit. The sensor is connected in parallel to the output of the aviation power supply. The sampling trigger circuit outputs a trigger signal according to the power supply type to control the sensor to synchronously acquire voltage signals.

[0022] Preprocessing module: Includes RC low-pass filter circuit, adaptive notch filter, instrumentation amplifier, 16-bit ADC, which sequentially filters, amplifies, and converts analog signals to digital signals, and outputs digital voltage sequence;

[0023] Feature extraction module: Built on FPGA, integrating RMS value calculation unit, peak detection unit, frequency analysis unit, and harmonic analysis unit to calculate the characteristic parameters of AC / DC power supply in real time;

[0024] Anomaly detection module: Composed of MCU and threshold storage unit. The threshold storage unit pre-stores initial thresholds for different power types. The MCU compares the output of the feature extraction module with the threshold, performs anomaly judgment and classification, and generates anomaly status signal.

[0025] Data output module: includes RS485 communication interface, local memory, and switch output circuit, which transmits the detection results to the monitoring terminal and stores them to SD card. In case of abnormality, it outputs switch signals to the alarm device.

[0026] Synchronization and calibration module: includes a synchronization clock generator, standard voltage interface, and calibration coefficient storage unit, provides synchronization trigger signal for inter-channel detection, performs accuracy calibration and stores calibration coefficients;

[0027] Power supply module: It adopts a wide-voltage input switching power supply to power all modules of the system, and has built-in overcurrent protection and overvoltage protection.

[0028] Furthermore, it also includes:

[0029] Load current acquisition submodule: Integrated into the signal acquisition module, it uses a Hall current sensor connected in series in the aviation power supply output circuit to acquire the load current signal and convert it into a 0-3.3V analog signal. This signal is then converted into a digital signal by a 12-bit ADC and transmitted to the anomaly detection module. The submodule adds a current harmonic detection unit, which implements a fast Fourier transform based on FPGA to calculate the total harmonic distortion rate (THDi) of the current. It also adds a current overload prediction unit to calculate the current change slope in real time. The anomaly detection module calculates the dynamic threshold based on the current signal using a formula. The integral term reflects the cumulative load change, enabling threshold adjustment based on load changes.

[0030] Encrypted Communication Submodule: Integrated into the data output module, this submodule uses an encryption chip to receive raw detection data from the data output module, encrypts it in 128-bit groups, and transmits the encrypted data via an RS485 interface. The submodule also includes a new data integrity verification unit that calculates a CRC32 checksum for each encrypted data frame and appends it to the end of the frame. Upon receiving a data frame, the monitoring terminal first calculates the CRC32 checksum of the received data and compares it with the checksum at the end of the frame. A communication interruption reconnection mechanism is also added, monitoring the communication status via the carrier detection pin of the RS485 interface. The monitoring terminal is equipped with the same type of encryption chip and has a built-in key update unit that automatically generates a new AES key every hour.

[0031] Compared with existing technologies, the beneficial effects of this invention are:

[0032] In terms of anti-interference and detection accuracy, this invention combines two-stage filtering with adaptive notch filtering to dynamically adapt to the complex interference in aviation scenarios: high-frequency suppression capacitors and RC low-pass filters can initially filter out high-frequency spikes, while the adaptive notch filter dynamically adjusts coefficients through a recursive least squares algorithm, suppressing both fixed power frequency interference and tracking drifting power frequency interference. Even in the multi-interference environment of airport ground, it can maintain a high signal-to-noise ratio of the detection signal and avoid characteristic parameter deviations caused by interference. Simultaneously, the sampling rate is adapted according to the power supply type: the AC power supply sampling rate meets the requirement of 20-50 times the operating frequency, while the DC power supply reduces the sampling rate as needed, balancing detection integrity and efficiency. Combined with temperature compensation and periodic automatic calibration mechanisms, high-precision detection can be maintained for a long time, avoiding accuracy degradation caused by environmental temperature fluctuations and equipment aging, eliminating the need for frequent manual calibration, and reducing maintenance costs.

[0033] In terms of dynamic adaptation and reliability, the dynamic threshold adjustment mechanism of this invention can accurately match the load fluctuation characteristics of aviation power supplies: based on load current integration and load type identification, the threshold is dynamically relaxed or tightened to avoid misjudgment and missed judgment caused by sudden load increases / decreases; multi-channel synchronous sampling (deviation ≤5μs) and phase difference calculation can accurately detect phase imbalance of multi-phase power supplies. With the backup channel design, automatic switching occurs when a channel fails, avoiding detection interruption and ensuring maintenance continuity. The anomaly classification and trend prediction functions can distinguish anomalies of different severity, avoid excessive alarms for minor anomalies, and provide early warnings for parameters approaching the threshold, reserving response time for power supply adjustments and reducing the risk of fault escalation—for example, a 1-second early warning when the voltage approaches the threshold allows maintenance personnel to intervene in time and avoid triggering emergency protection.

[0034] Regarding data security and operational efficiency, this invention optimizes end-to-end data management through edge computing and encryption technology: after initial data analysis, edge computing stores only key statistical values ​​and abnormal raw data, combined with piecewise linear fitting compression, significantly reducing storage and transmission bandwidth usage; the combination of AES-128 encryption and RSA key distribution ensures the security of data transmission, storage, and export, preventing tampering and leakage, and meeting aviation information security requirements. The multi-interface design for data output (RS485, switch input) is adaptable to different monitoring terminals (cabin monitoring screen, ground operation and maintenance platform), and the linkage between abnormal alarms and protection signals can quickly trigger backup power or protection circuits, reducing the scope of fault impact. The overall system's wide-voltage power supply and overcurrent / overvoltage protection adapt to power fluctuations in aviation scenarios, ensuring the stable operation of the detection system itself, further improving the safety and efficiency of aviation power supply operation and maintenance. Attached Figure Description

[0035] Figure 1 This is a schematic block diagram of an output voltage detection system based on an aviation power supply proposed in this invention;

[0036] Figure 2 This is a schematic diagram of an output voltage detection method based on an aviation power supply proposed in this invention;

[0037] Figure 3 This is a line graph showing the comparison of the false judgment rates of fixed and dynamic thresholds under different load currents for the output voltage detection method based on aviation power supply proposed in this invention.

[0038] Figure 4 This is a line graph showing the comparison of detection errors with and without temperature compensation at different temperatures for the output voltage detection method based on aviation power supply proposed in this invention. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] In the description of this invention, it should be understood that the terms "center," "longitudinal," "lateral," "length," "width," "thickness," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," "clockwise," and "counterclockwise," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not 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.

[0041] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of the stated features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified. Furthermore, the terms "installed," "connected," and "linked" should be interpreted broadly; for example, they may refer to a fixed connection, a detachable connection, or an integral connection; they may refer to a mechanical connection or an electrical connection; they may refer to a direct connection or an indirect connection through an intermediate medium; and they may 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. The invention will now be described in further detail with reference to the accompanying drawings.

[0042] Reference Figures 1 to 4A method for detecting the output voltage of an aviation power supply, comprising:

[0043] Signal acquisition steps: A differential voltage sensor is connected to the output of the aviation power supply. A high-frequency suppression capacitor (capacitance 100-500pF) is connected in parallel at the sensor input to suppress spike interference. The sampling rate is set to 20-50 times the power supply operating frequency (sampling rate 8-20kHz when AC power supply is 400Hz, sampling rate 1-5kHz when DC power supply is 28V). The voltage signal is continuously acquired and converted into an analog voltage signal (range ±150V, accuracy ±0.1%FS).

[0044] Signal preprocessing steps: The analog voltage signal is filtered in two stages. The first stage is an RC low-pass filter (cutoff frequency 1kHz) to remove high-frequency noise, and the second stage is an adaptive notch filter (center frequency 50Hz / 60Hz) to suppress power frequency interference. The filtered signal is amplified to 0-5V by an instrumentation amplifier (gain 10-100 times, common-mode rejection ratio ≥120dB), and then converted to digital voltage sequence by a 16-bit ADC.

[0045] Feature extraction steps: Calculate feature parameters for the digital voltage sequence. For AC power, these include RMS value (calculated based on sampled values ​​within half a cycle), peak value (maximum and minimum values ​​within each cycle), frequency deviation (difference between the actual frequency and 400Hz), and harmonic distortion rate (ratio of total harmonic content to fundamental frequency). For DC power, these include average value (arithmetic mean of 100 consecutive sample points), ripple coefficient (ratio of RMS ripple to average value), and fluctuation (difference between maximum and minimum voltage within 5 seconds).

[0046] Anomaly detection steps: Compare the extracted feature parameters with preset thresholds. AC power supply thresholds include RMS deviation ±5% (115V power supply corresponds to 109.25-120.75V), frequency deviation ±10Hz, and harmonic distortion rate ≤5%; DC power supply thresholds include average deviation ±10% (28V power supply corresponds to 25.2-30.8V), ripple coefficient ≤1%, and fluctuation ≤0.5V. If any feature parameter exceeds the threshold, it is judged as an anomaly, and the time of occurrence, duration, and corresponding feature value of the anomaly are recorded.

[0047] Data output steps: The detection results (normal / abnormal status, real-time voltage value, characteristic parameters) are transmitted to the aviation power monitoring terminal in the form of digital signals (RS485 protocol, baud rate 9600bps) and stored in the local memory (a data file is generated every 5 minutes in CSV format). When an abnormal status is triggered, a switch signal (active high level) is output to the alarm device.

[0048] This invention also includes:

[0049] Dynamic threshold adjustment steps: This step corrects the anomaly detection threshold in real time based on power load changes. First, the load type is identified by the current harmonic content (THDi > 8% for inductive loads, phase lead of fundamental > 10° for capacitive loads, THDi < 3% and phase deviation < 5° for resistive loads). Differentiated load influence coefficients k are then configured for different loads (k = 0.005 for inductive loads, k = 0.003 for capacitive loads, k = 0.002 for resistive loads). Then, the adjustment is performed using the formula... Calculate the adjusted voltage RMS threshold, where U th U0 is the adjusted threshold (in V), U0 is the initial threshold (in V), t0 is the detection start time (in seconds), and t is the current time (in seconds). The load current at time τ (in A) is represented by the integral term, which reflects the cumulative load change. After threshold adjustment, monitoring continues for 3 seconds. If the characteristic parameter still exceeds U... th This is considered abnormal to avoid misjudgment caused by instantaneous load fluctuations; when the load current stabilizes within ±2A within 30 seconds, U th It automatically recovers to the initial threshold to ensure detection accuracy under normal conditions.

[0050] Multi-channel synchronous detection step: This step configures 2-4 independent acquisition channels (corresponding to different phases or different output ports of the aviation power supply). Each channel is triggered for sampling by a synchronous clock (accuracy ±1μs), ensuring that the sampling time deviation is ≤5μs. After synchronous sampling, the phase difference between each channel is calculated (based on zero-crossing detection, the AC power supply detects 2 zero-crossings per cycle, and the phase difference is calculated by converting the zero-crossing time difference between adjacent channels). When the phase difference is >5°, the transmission delay (approximately 1.67ns) is calculated based on the signal transmission distance between channels (e.g., the distance between channel 1 and channel 2 is 0.5m, and the signal propagation speed is 3×10^8m / s). A phase compensation value is generated (the phase compensation amount corresponding to the delay = 360° × delay time × power supply frequency), which is corrected during feature extraction. Phase data; simultaneously monitor the sampling values ​​of each channel in real time. If 100 consecutive sampling points of a channel exceed the normal voltage range (e.g., the sampling value of the AC 115V power supply channel is continuously <100V or >130V), it is determined to be a channel fault, and automatically switches to the backup channel (one backup channel is pre-configured), and records the fault channel number and the time of the fault; calculate the cross-correlation coefficient of the digital voltage sequence of each channel (reflecting the synchronicity of voltage changes between channels). When the correlation coefficient between the phase voltage channels of the AC power supply is <0.95, it is determined to be a phase imbalance abnormality, and the phase difference of each channel is recorded. When the data is output, the "phase imbalance" status and the specific phase deviation value are marked. It is suitable for comprehensive detection of multi-phase aviation power supplies (e.g., three-phase 115V / 400Hz).

[0051] In this invention, the adaptive notch filter uses a recursive least squares algorithm to dynamically adjust the filter coefficients in the signal preprocessing step. Specifically, an initial filter coefficient vector (length 5-10, based on a preset power supply fundamental frequency of 400Hz) is set, and the coefficients are updated every 10 sampling points to ensure that the filter attenuates ≥40dB for 50Hz / 60Hz power frequency interference and ≤1dB for the power supply fundamental frequency (400Hz). Simultaneously, a noise intensity detection step is added. By calculating the signal-to-noise ratio (SNR) of the analog voltage signal before filtering (SNR=20lg(signal RMS / noise RMS)), when SNR<30dB (strong interference environment, such as near airport ground power supply), the notch filter order is automatically increased from 5th to 8th order to enhance interference suppression capability and ensure interference attenuation ≥45dB; when SNR>40dB (weak interference environment, such as near airport ground power supply), the notch filter order is automatically increased from 5th to 8th order to enhance interference suppression capability and ensure interference attenuation ≥45dB. The filter order is reduced from 5th to 3rd order to reduce the attenuation of the fundamental signal (fundamental attenuation ≤ 0.5dB). The interference suppression ratio is calculated every 100ms (interference suppression ratio = interference amplitude before filtering / interference amplitude after filtering). If the suppression ratio < 20dB (i.e., attenuation < 26dB), the filter coefficient vector is reinitialized and the update frequency is increased (updated every 5 sampling points) until the suppression ratio ≥ 20dB. When a power frequency interference frequency drift is detected (e.g., from 50Hz to 51Hz or 49Hz), the drift amount is identified by a frequency tracking algorithm (based on fast Fourier transform, the interference frequency is calculated every 20ms). The filter center frequency is synchronously adjusted within 100ms to ensure effective suppression of the drifted power frequency interference. This is especially suitable for scenarios with complex power frequency interference, such as airport ground power supply and cabin hybrid power supply.

[0052] In this invention, during the anomaly detection step, persistent abnormal states (exceeding 5 seconds) are classified into three levels: Level 1 anomaly is when a single characteristic parameter exceeds the threshold and the deviation is ≤10% (e.g., an AC 115V power supply with an effective voltage of 122V, exceeding the upper limit of 120.75V but ≤126.5V (115V×110%)); Level 2 anomaly is when a single characteristic parameter deviation is >10% or two or more characteristic parameters exceed the threshold (e.g., an AC power supply with an effective voltage of 127V and a harmonic distortion rate of 6%); Level 3 anomaly is when a characteristic parameter deviation is >20% or accompanied by voltage... Sudden change (a change of ≥10V within 1 second, such as a sudden drop of 28V DC power to 18V); A new historical data comparison mechanism is added to retrieve the statistical values ​​(average, maximum, and minimum) of characteristic parameters from the same time period over the past 7 days (e.g., if the current detection time is 10:00-10:05, retrieve the voltage data from 10:00-10:05 of the previous 7 days). If the current abnormal characteristic parameter is within ±12% of the historical statistical value (e.g., the maximum effective value for the same period in history was 121V, and the current value is 122V), and there are no other abnormal characteristic parameters, then the abnormality level is adjusted. The level is lowered (e.g., from Level 1 to "Watchlist," only a notification is displayed on the monitoring terminal, without triggering audible or visual alarms); if the current abnormal parameter deviates from historical statistical values ​​by more than 15% (e.g., the historical maximum effective value was 121V, the current value is 125V), the original abnormal level is maintained; in addition, an abnormal trend prediction function is added, which calculates the slope of characteristic parameter changes in real time (e.g., the change in effective voltage value per second). When the parameter approaches the threshold and the slope is greater than 0.5V / s (e.g., AC effective value is 120V, threshold is 120.75V, slope is 0.6V / s), the abnormal level is determined. One second in advance, a "threshold approach warning" is issued on the monitoring terminal, and the parameter value at the predicted time is recorded to reserve response time for power supply adjustment. Different levels of anomalies correspond to different output strategies. Level 1 anomalies are only displayed in green text on the monitoring terminal. Level 2 anomalies trigger an audible and visual alarm (sound intensity 80dB, red light flashing frequency 2Hz). Level 3 anomalies, in addition to audible and visual alarms, simultaneously output a protection signal (continuous high level for 2 seconds, triggering the backup protection circuit of the aviation power supply to prepare to cut off the fault output) and send an emergency fault message to the ground operation and maintenance terminal.

[0053] In this invention, the data output step also includes data compression and encryption processing, and a new edge computing preprocessing stage is added: For locally stored voltage data, preliminary analysis is first performed through the edge computing module to extract key statistical values ​​(average, maximum, minimum, and effective values) every 5 seconds. Only the statistical values ​​and the original sampled data during abnormal periods (such as the original data 10 seconds before and after the abnormality) are stored. Then, piecewise linear fitting compression is used (10 lines are fitted to 1000 original sample points, with a fitting error ≤0.5%). After compression, the data volume is reduced to 10% of the original, reducing local storage usage. The data transmitted to the monitoring terminal is first filtered by edge computing, and only key statistical values, abnormal records, and prediction information are uploaded, reducing remote transmission bandwidth usage (such as 5MB of data transmitted after the original 100MB per hour is transmitted). In the encryption processing stage, the AES-128 encryption algorithm is used to process the transmitted data, and the key is distributed through an asymmetric encryption (RSA-2048) mechanism: During system initialization, the monitoring terminal generates an RSA public key and a private key, and distributes the public key to the monitoring terminal. The testing device generates an AES key, encrypts it with an RSA public key, and then transmits it to the monitoring terminal. The monitoring terminal decrypts the AES key using its private key to obtain the AES key. The key is automatically updated every hour, maintaining the current encrypted link during the update process to avoid transmission interruption. Data transmission adopts a fragmentation mechanism, dividing files exceeding 1024 bytes into 1024-byte fragments. Each fragment is independently encrypted and given a fragment number (e.g., fragment 1, fragment 2). The monitoring terminal reassembles the fragments according to their numbers upon receiving them. If a fragment is detected as missing (e.g., fragment 3 is not received), it requests the testing device to retransmit the corresponding fragment. Locally stored data is encrypted using the AES-128 algorithm. The encryption key is independent of the transmission key, and the storage key is embedded in the system security chip (e.g., the built-in security unit of STM32H7) to prevent unauthorized reading. When exporting data (e.g., via USB interface), the operator's authorization password (a 6-digit numeric password; three consecutive incorrect passwords lock the export function for 30 minutes) must be verified to ensure the security of aviation power supply voltage data throughout the entire process of transmission, storage, and export, preventing data tampering or leakage.

[0054] This invention also includes:

[0055] Accuracy calibration steps: Automatically executed monthly, with added temperature compensation and calibration result verification mechanisms: Before calibration, the system's built-in temperature sensor (DS18B20, measurement range -40℃ to 85℃, accuracy ±0.5℃) collects the current ambient temperature T. Based on a preset temperature-error model (e.g., for every 10℃ increase in temperature, the detection error increases by 0.01%FS; for every 10℃ decrease in temperature, the error increases by 0.008%FS), the temperature correction value ΔU is calculated (ΔU = detection value × error coefficient × (T - 25℃) / 10℃, with 25℃ as the base). The standard temperature is then connected; subsequently, a standard voltage source (accuracy ±0.01%, supporting DC 25V, 28V, 31V, AC 100V, 115V, 130V / 400Hz) is connected, and the standard voltage values ​​are output sequentially. The actual detection value Udet of the detection device is collected, and the difference between the detection value and the standard value Ustd is calculated as ΔUraw=Udet-Ustd. Then, the temperature correction value is added to obtain the final error ΔUfinal=ΔUraw+ΔU, generating a calibration coefficient matrix (containing correction values ​​for different voltage points and different temperature ranges, such as 28V). The correction value for 25℃ is -0.02V, and the correction value for 28V / 35℃ is -0.022V, which is stored in non-volatile memory (EEPROM). During routine testing, the ambient temperature is collected in real time, and the corresponding correction value in the calibration coefficient matrix is ​​called according to the current temperature and the detection voltage value to correct the detection results, thereby improving the detection accuracy from ±0.1%FS to ±0.05%FS. After calibration, the intermediate value of the standard voltage source (such as DC 28V, AC 115V) is connected for verification, and the corrected detection value is collected. If the error is >0. If the error is 0.05%FS (e.g., the standard value of DC 28V, the corrected detection value is 28.014V, and the error of 0.05%FS corresponds to 0.014V; if the error is 0.015V, it will exceed the limit), the calibration process will be re-executed until the verification is successful. In addition, when the system detects that the ambient temperature changes by more than 15°C within 1 hour (e.g., the temperature changes suddenly due to the start-up and shutdown of the air conditioning in the cabin), it will automatically trigger temporary calibration (only calibrating the currently commonly used voltage point, such as DC 28V or AC 115V) to ensure that the detection accuracy still meets the requirements when the temperature changes drastically, and to maintain high-precision detection capability over a long period of time.

[0056] This invention also discloses a system for detecting the output voltage of an aviation power supply, comprising:

[0057] The signal acquisition module consists of a differential voltage sensor (model LV25-P, input ±150V, output 0-5V, linearity ±0.1%), a high-frequency suppression capacitor (100pF ceramic capacitor, withstand voltage 500V), and a sampling trigger circuit (composed of a comparator LM311, trigger level 0.5V). The sensor is connected in parallel to the output of the aviation power supply. The sampling trigger circuit outputs a trigger signal according to the power supply type (AC / DC) (AC power supply trigger frequency 8-20kHz, DC power supply trigger frequency 1-5kHz) to control the sensor to synchronously acquire voltage signals.

[0058] The preprocessing module includes an RC low-pass filter circuit (1kΩ metal film resistor, accuracy ±1%; 160nF monolithic capacitor, accuracy ±5%, cutoff frequency 1kHz), an adaptive notch filter (composed of an operational amplifier OP07, gain-bandwidth product 1MHz, center frequency adjustable range 45-65Hz), an instrumentation amplifier (AD620, common-mode rejection ratio 120dB, gain adjustable range 10-100 times), and a 16-bit ADC (ADS8344, conversion rate 100kSPS, integral nonlinearity ±0.5LSB), which sequentially filters, amplifies, and converts the analog signal to digital, outputting a digital voltage sequence.

[0059] Feature extraction module: Built on FPGA (model EP4CE10, 10K logic units, 100MHz main frequency), integrating an effective value calculation unit (implementing half-cycle integration operation, integration step size 1μs), a peak detection unit (comparing continuous sampled values ​​to find extreme values, response time ≤1μs), a frequency analysis unit (based on zero-crossing detection to calculate the cycle, detecting 2 zero-crossings per cycle, frequency resolution 0.1Hz), and a harmonic analysis unit (fast Fourier transform processing, 1024 points, transformation time ≤1ms), to calculate the characteristic parameters of AC / DC power supply in real time;

[0060] Anomaly detection module: Composed of MCU (STM32H743, 400MHz, 2MB Flash) and threshold storage unit (EEPROM, 128KB). The threshold storage unit pre-stores initial thresholds for different power supply types (AC 115V / 400Hz, DC 28V). The MCU compares the output of the feature extraction module with the threshold, performs anomaly judgment and classification, and generates anomaly status signals (including time, duration, and level).

[0061] Data output module: includes RS485 communication interface (MAX485 chip, communication distance 1200m, supports 9600bps baud rate), local memory (16GB SD card, supports SPI interface, read / write speed 10MB / s), and switch output circuit (relay output, capacity 250V / 1A, response time ≤10ms), which transmits the detection results to the monitoring terminal, stores them to the SD card, and outputs switch signals to the alarm device when there is an abnormality;

[0062] Synchronization and Calibration Module: Includes a synchronization clock generator (DS3231, real-time clock accuracy ±2ppm, temperature compensation range -40℃ to 85℃), a standard voltage interface (BNC interface, impedance 50Ω, supports external standard source connection), and a calibration coefficient storage unit (EEPROM, capacity 128KB), providing synchronous trigger signals for multi-channel detection, performing accuracy calibration, and storing calibration coefficients; Power Supply Module: Employs a wide-input (9-36V) switching power supply (module LM2596, output 5V / 3A, 3.3V / 2A, efficiency ≥85%) to power all modules of the system. It features built-in overcurrent protection (current limit 5A, response time ≤100μs) and overvoltage protection (voltage limit 6V, response time ≤50μs) to ensure stable system operation during aviation power fluctuations.

[0063] This invention also includes:

[0064] Load current acquisition submodule: Integrated into the signal acquisition module, this submodule uses a Hall current sensor (ACS758, range ±50A, accuracy ±1%, response time ≤4μs) connected in series in the aviation power supply output circuit to acquire the load current signal and convert it into a 0-3.3V analog signal. This signal is then converted into a digital signal by a 12-bit ADC (ADS7822, conversion rate 50kSPS, integral nonlinearity ±1LSB) and transmitted to the anomaly detection module. The submodule also includes a new current harmonic detection unit, which implements a fast Fourier transform (512 points) based on an FPGA to calculate the total harmonic distortion rate THDi (THDi=√(I2²+I3²+…+In²) / I1, where I1 is the fundamental current and I2 to In are the harmonic currents). When THDi > 8%, it is determined to be a load anomaly (e.g., damage to internal load components leading to increased harmonics), and a current anomaly signal is simultaneously output to the anomaly detection module. The normal detection module and the anomaly detection module correlate current anomalies with voltage anomalies (e.g., increased voltage ripple and THDi > 8%, indicating a power supply anomaly caused by the load). Simultaneously, an overload prediction unit is added to calculate the current change slope in real time (slope = ΔI / Δt, where ΔI is the current difference between adjacent 10ms, and Δt = 10ms). When the slope > 2A / s and the current > 80% of the rated current (e.g., the rated output current of an aviation power supply is 30A, and the current is 25A), an overload warning signal is issued to the data output module 500ms in advance. The data output module synchronously transmits the warning information to the monitoring terminal, forming a linked protection mechanism with the voltage anomaly detection. The anomaly detection module calculates a dynamic threshold based on the current signal using a formula. The integral term reflects the cumulative load change, enabling threshold adjustment based on load changes and improving the system's adaptability to complex load scenarios (e.g., mixed power supply of inductive and resistive loads in the cabin).

[0065] Encrypted Communication Submodule: Integrated into the data output module, this submodule uses an encryption chip (AT88SC0104 dedicated AES128 chip, encryption rate ≥1Mbps, supports AES-128 / 256). It receives raw detection data (key statistical values, anomaly records, and predictive information) from the data output module, groups it into 128-bit blocks, and performs encryption operations (10 rounds of byte replacement, row shifting, column mixing, and round key addition). The encrypted data is transmitted via an RS485 interface. The submodule also includes a data integrity verification unit. For each encrypted data frame (including frame header, data segment, and frame trailer), it calculates a CRC32 checksum (polynomial 0xEDB88320) and adds it to the frame trailer. After receiving a data frame, the monitoring terminal first calculates the CRC32 checksum of the received data and compares it with the frame trailer checksum. If they do not match, a retransmission request is sent to the submodule, which then re-encrypts and retransmits the corresponding data. According to the frame, error-free data transmission is ensured; at the same time, a communication interruption reconnection mechanism is added. The communication status is monitored through the carrier detection pin (CD pin) of the RS485 interface. If no carrier is detected for 100ms (determined as a communication interruption), a reconnection attempt is made every 100ms (sending a reconnection request message). After successful reconnection, the submodule retransmits the critical data during the interruption (such as abnormal records and statistical values ​​5 seconds before and after the interruption) to prevent data loss; the encryption chip's key is stored in the built-in security fuse area (non-erasable, only readable by chip instructions) to prevent the key from being illegally tampered with or extracted; the monitoring terminal is equipped with the same type of encryption chip and has a built-in key update unit. A new AES key is automatically generated every hour, encrypted with RSA-2048 and sent to the submodule. The submodule decrypts and updates the key to ensure the long-term security of the encrypted link and meet the strict requirements of avionics systems for data transmission security and integrity.

[0066] A specific implementation of an output voltage detection method and system based on aviation power supply:

[0067] Example 1: Airport Ground Three-Phase 115V / 400Hz Aviation Power Supply Testing Scenario (Application Scenario: Ground power supply for the boarding bridge of Terminal 3 of an international airport, providing ground power for narrow-body aircraft such as Boeing 737 and Airbus A320. The power supply outputs three-phase 115V / 400Hz, with a rated load current of 30A. Daily monitoring is required for the effective voltage value, phase difference, and harmonic distortion rate. It also needs to cope with ground power grid frequency interference (50Hz) and the sudden load increase during aircraft startup (current rises from 10A to 25A within 30 seconds). The system is deployed in the boarding bridge power supply control cabinet and connects to the airport operation and maintenance cloud platform.)

[0068] I. System Module Deployment and Parameter Configuration Implementation

[0069] Signal acquisition module deployment: Three differential voltage sensors (model LV25-P, input ±150V, output 0-5V, linearity ±0.1%) are used, connected in parallel to the output terminals of phases A, B, and C of the three-phase power supply. A 100pF high-frequency suppression ceramic capacitor (withstand voltage 500V) is connected in series at the input terminal of each sensor to suppress the 20-30kHz spike interference during aircraft startup. The sampling trigger circuit outputs a synchronous trigger signal from the FPGA (EP4CE10) with a frequency set to 20kHz (50 times that of the 400Hz power supply) to ensure that the three-phase sampling time deviation is ≤5μs. One spare sensor (of the same model) is configured and combined with A through a relay switching circuit. The spare trigger signal is synchronized with the main channel.

[0070] Preprocessing module configuration: The RC low-pass filter circuit uses a 1kΩ metal film resistor (accuracy ±1%) and a 160nF monolithic capacitor (accuracy ±5%), with a cutoff frequency of 1kHz, to filter out high-frequency noise conducted from the power supply line; the adaptive notch filter is composed of an operational amplifier OP07, with an initial center frequency of 50Hz, a feedback resistor of 100kΩ, and a capacitor of 0.1μF. The filter coefficient is dynamically adjusted through a recursive least squares algorithm and updated every 10 sampling points (0.5ms); the instrumentation amplifier AD620 is set to a gain of 20 times to amplify the filtered ±2.5V signal to 0-5V; the 16-bit ADCADS8344 conversion rate is set to 100kSPS, with a single conversion time of 10μs. The three-phase data is collected in the order of A→B→C, with 1024 sampling points collected for each phase to form a digital voltage sequence.

[0071] Feature extraction module implementation: The FPGA's built-in RMS calculation unit performs half-cycle (1.25ms) integration on each of the 1024 sampling points for each phase, using the following formula: (N=500, corresponding to half a cycle of sampling), calculate the effective value of phase A as 114.8V, phase B as 115.2V, and phase C as 115.0V; the peak detection unit compares the sampling values ​​of each cycle to obtain the peak values ​​of phase A as 163.5V, phase B as 164.1V, and phase C as 163.8V; the frequency analysis unit calculates the actual frequency as 400.2Hz with a deviation of +0.2Hz by detecting zero crossings (2 zero crossings per cycle); the harmonic analysis unit performs a 1024-point FFT, and the total harmonic distortion (THD) is 2.3% (fundamental amplitude 115V, 3rd harmonic 3.2V, 5th harmonic 1.8V).

[0072] Anomaly detection module configuration: Threshold storage unit (EEPROM) pre-stores the initial threshold of three-phase power supply: effective value 109.25-120.75V, phase difference ≤5°, THD≤5%; MCU (STM32H743) is connected to the load current acquisition submodule (Hall sensor ACS758, range ±50A) to acquire the A-phase current signal, convert it into a 0-3.3V digital signal, and calculate the current THDi=7.5% (determined as inductive load, k=0.005).

[0073] Data output module deployment: RS485 communication interface (MAX485) baud rate set to 9600bps, connected to the airport operation and maintenance cloud platform, transmission cycle 5 seconds / time; local storage uses 16GB SD card, files named in the format "20240610-1030-A320", storing the effective voltage value, phase difference, and current value every 5 seconds; switch output circuit (relay) connected to the power control cabinet alarm light (red) and backup power switching signal terminal, outputting a high level to trigger in case of abnormality.

[0074] Synchronization and calibration module configuration: The DS3231 synchronization clock generator has an accuracy of ±2ppm and is calibrated with the airport NTP server once a day; the standard voltage interface (BNC) automatically connects to an external standard source (FLUKE5520A, accuracy ±0.01%) at 9:00 on the 1st of each month, outputs a 115V / 400Hz standard voltage, collects and detects a value of 114.98V, calculates an error of -0.02V, generates a calibration coefficient matrix (115V corresponds to a correction value of +0.02V), and stores it in EEPROM.

[0075] II. Implementation of Key Steps and Application of Formulas

[0076] Dynamic threshold adjustment implementation: When the aircraft starts (10:30), the current rises from 10A to 25A (within 30 seconds), and the MCU adjusts according to the formula. Calculate the A-phase threshold: U0 = 120.75V, k = 0.005, t0 = 10:30:00, t = 10:30:30. , then U th =120.75×(1+0.005×525 / 100)=120.75×1.02625≈123.92V; At this time, the voltage of phase A drops to 108.5V (original threshold 109.25V), but does not exceed the lower limit of the adjusted threshold 109.25×(1-0.005×525 / 100)≈106.48V, which is judged to be normal to avoid misjudgment.

[0077] Multi-channel phase difference detection implementation: FPGA calculates the zero-crossing time of the three phases: Phase A zero-crossing time is 10:30:30.000000s, Phase B is 10:30:30.000694s, and Phase C is 10:30:30.001389s; Phase difference calculation: The time difference between Phase B and Phase A is 0.694ms, corresponding to a phase difference of 360° × 0.694ms × 400Hz ≈ 99.94° (standard 120°, deviation 20.06°). Phase compensation value is called: the distance between phases A and B is 0.3m, the signal propagation delay is 0.3m / 3×10^8m / s=1ns, the compensation phase is 360°×1ns×400Hz≈0.000144°, and the corrected phase difference is ≈20.06°. Since the correlation coefficient is 0.92<0.95, it is determined to be an abnormal phase imbalance. Record "phase B lags phase A by 20.06°" and trigger a level 2 alarm (red light flashes at 2Hz, sound at 80dB). Adaptive notch filtering implementation: When the amplitude of 50Hz power frequency interference from the ground grid reaches 1.5V (superimposed on the 400Hz fundamental frequency), with an SNR of 28dB < 30dB, the notch filter order is automatically increased from 5th to 8th, the feedback resistor is adjusted to 150kΩ, and the capacitor is 0.08μF. The interference suppression ratio is calculated every 100ms: 1.5V before filtering / 0.07V after filtering ≈ 21.4 > 20dB, which meets the requirements. When the power grid interference drifts to 51Hz, the frequency tracking algorithm identifies the drift within 20ms, and the center frequency is synchronously adjusted to 51Hz, maintaining the suppression ratio above 20dB, and the THD decreases from 4.8% during interference to 2.3%.

[0078] III. Verification of Detection Results and Data Comparison

[0079] After running continuously for one week, the system detected three phase imbalances (all B-phase lagging) and two load surges (current rose to 28A), with no false positives or false negatives. Traditional detection equipment (single channel, fixed threshold) made five false positives (caused by load surges) and two false negatives (phase difference of 18° not detected) during the same period.

[0080] Table 1: Comparison of the effects of traditional equipment and this system for ground power detection in Example 1

[0081] Evaluation indicators Traditional single-channel testing equipment This system performs multi-channel fusion detection. Voltage RMS detection error ±0.5%FS ±0.05%FS Phase imbalance detection sensitivity It can only detect deviations greater than 30°. Detectable deviation >5° Power frequency interference suppression ratio 15dB 21dB False positive rate due to sudden load increase 35% (5 / 14 times) 0% (0 / 14 times) Data upload delay 200ms 50ms

[0082] Table 1 shows that traditional equipment suffers from low phase detection sensitivity due to its single channel and lack of synchronous sampling, and its fixed threshold is prone to misjudgment. This system, through multi-channel synchronization, dynamic thresholding, and adaptive filtering, significantly reduces detection errors, resulting in more sensitive phase detection and stronger interference suppression capabilities, making it fully adaptable to the complex operating conditions of ground power supplies. For example, the voltage RMS detection error has been reduced from ±0.5%FS to ±0.05%FS, ensuring that maintenance personnel accurately grasp the power supply output status; the 50ms data latency meets the real-time monitoring requirements of the airport cloud platform, avoiding delays in maintenance response.

[0083] Example 2: Aircraft Cabin 28V DC Aviation Power Supply Testing Scenario (Application Scenario: A Boeing 787 passenger aircraft of a certain airline uses a DC power supply system to power cabin lighting and avionics control modules. It outputs 28V DC with a rated load current of 20A. Cabin temperature fluctuates from -10℃ to 40℃. Monitoring of average voltage, ripple coefficient, and fluctuation is required to address sudden temperature changes caused by air conditioning start-up and shutdown (from 25℃ to 35℃ within 1 hour) and load fluctuations during avionics module switching (current increases from 8A to 15A). The system is integrated into the cabin power control box and interfaces with the aircraft cockpit monitoring screen and ground maintenance terminal.)

[0084] I. System Module Deployment and Parameter Configuration Implementation

[0085] Signal acquisition module deployment: One differential voltage sensor (LV25-P, input 0-50V, output 0-5V) is used, connected in parallel between the positive and negative terminals of the DC power supply. A 220pF high-frequency suppression capacitor is connected in parallel at the sensor input terminal (to suppress the 15kHz spike during avionics module switching). The sampling trigger frequency is set to 5kHz (the DC power supply does not require a high sampling rate). The trigger signal is output by the MCU (STM32H743) and synchronously connected to a temperature sensor (DS18B20, -40℃ to 85℃). The ambient temperature is collected every 100ms.

[0086] Preprocessing module configuration: The RC low-pass filter circuit uses a 1kΩ resistor and a 160nF capacitor (cutoff frequency 1kHz) to filter out high-frequency components in the ripple; the center frequency of the adaptive notch filter is set to 60Hz (to accommodate a small amount of power frequency interference in the cabin), and the gain-bandwidth product is 1MHz; the AD620 instrumentation amplifier is set to a gain of 10 times to amplify the 0-2.8V signal to 0-28V, corresponding to an output of 0-5V; the 16-bit ADCADS8344 has a conversion rate of 50kSPS and collects 1000 sampling points to form a digital voltage sequence.

[0087] Feature extraction module implementation: FPGA calculates the average voltage value = 27.9V (arithmetic average of 1000 sampling points), the effective value of ripple = 0.21V (calculation of the effective value of ripple component), and the ripple coefficient = 0.21 / 27.9≈0.75%; fluctuation amount = maximum 28.1V - minimum 27.7V within 5 seconds = 0.4V; the temperature sensor collects the current cabin temperature of 32℃ and transmits it to the synchronization and calibration module.

[0088] Anomaly detection module configuration: Threshold storage unit pre-stores DC power supply thresholds: average value 25.2-30.8V, ripple coefficient ≤1%, fluctuation ≤0.5V; MCU connected to load current acquisition submodule (ACS758, range ±30A), acquisition current 12A, THDi=2.5% (judged as resistive load, k=0.002); enable anomaly trend prediction, calculate voltage average change slope = 0.05V / s (currently 27.9V, threshold upper limit 30.8V), no risk of approaching the threshold.

[0089] Data output module deployment: RS485 interface connects to cockpit monitoring screen (baud rate 9600bps), local storage uses 16GB SD card, data is stored in "flight date-flight number-time", and one compressed file is generated every 5 minutes (original 10MB / file, compressed 1MB / file); the encrypted communication submodule uses AT88SC0104 chip, AES-128 encrypted data transmission, the key is updated hourly from the ground maintenance terminal via aircraft satellite link, and data transmission is maintained without interruption during the update process.

[0090] II. Implementation of Key Steps and Application of Formulas

[0091] Temperature compensation and calibration implementation: Automatic calibration is performed on the 1st of each month, connected to a standard DC power supply (FLUKE5520A, output 25V, 28V, 31V). The current ambient temperature is 32℃. The temperature correction value ΔU is calculated as: ΔU = detected value × 0.001 × (32-25) / 10 = detected value × 0.0007. When calibrating the 28V standard value, the original detected value is 27.98V, ΔU = 27.98 × 0.0007 ≈ 0.0196V, and the final error is 27.98 + 0.0196 - 28 ≈ -0.0004V. A calibration coefficient matrix is ​​generated (correction value +0.02V corresponding to 28V / 32℃). During routine testing, if the temperature suddenly rises to 35℃, the correction value +0.022V is applied. The detected value of 27.9V is corrected to 27.922V, with an error < 0.05%FS.

[0092] Data compression and encryption implementation: The edge computing module extracts key statistical values ​​(average value, ripple coefficient) every 5 seconds, and only stores the raw data during abnormal periods (such as fluctuations of 0.45V close to the threshold); piecewise linear fitting compresses 1000 sampling points into 10 straight lines with a fitting error of 0.3%; during encrypted transmission, data frames are grouped into 128-bit blocks and a CRC32 checksum is added (e.g., data segment 0x12345678, checksum 0x87654321). The cockpit monitoring screen receives the data, verifies its consistency, and then decrypts it, ensuring no data loss or tampering.

[0093] III. Verification of Detection Results and Data Comparison

[0094] During 10 flights (8 hours each), the system detected two instances of ripple coefficient exceeding the upper limit (0.95%, less than 1%) and one instance of sudden temperature change (10°C increase in 1 hour), both of which were accurately recorded and alarmed. In contrast, traditional equipment had a detection error exceeding 0.3%FS due to temperature changes during the same period, and missed one ripple anomaly.

[0095] Table 2: Comparison of the effects of traditional equipment and this system for detecting DC power supply in the cabin in Example 2

[0096] Evaluation indicators Traditional equipment without temperature compensation This system includes temperature compensation and encryption. Temperature change detection error ±0.3%FS (at 35℃) ±0.05%FS (at 35℃) Data compression rate Uncompressed (100%) 10%（1 / 10） Data encryption security Unencrypted (transmitted in plaintext) AES-128 encryption (tamper-proof) Ripple anomaly missed detection rate 20% (1 / 5 times) 0% (0 / 5 times) Key update stability Keyless mechanism 100% update success rate (10 attempts)

[0097] Table 2 shows that traditional equipment lacks temperature compensation, resulting in large detection errors during sudden temperature changes, and lacks data security mechanisms. This system maintains high accuracy through temperature compensation, reduces storage footprint through data compression (adapting to limited cabin storage space), and ensures data security through AES-128 encryption (compliant with aviation information security standards). For example, at 35℃, the detection error is reduced from ±0.3%FS to ±0.05%FS, ensuring accurate monitoring of the avionics module's power supply voltage; a 10% compression rate allows a 16GB SD card to store 1600 hours of data, meeting the long-term recording needs of intercontinental flights; and 100% key update success rate avoids the risk of data leakage due to encrypted link interruptions.

[0098] The above are merely preferred embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for detecting the output voltage of an aviation power supply, characterized in that, include: Signal acquisition steps: A differential voltage sensor is connected to the output of the aviation power supply. A high-frequency suppression capacitor is connected in parallel at the input of the sensor to suppress spike interference. The sampling rate is set to 20-50 times the operating frequency of the power supply. The voltage signal is continuously acquired and converted into an analog voltage signal. Signal preprocessing steps: The analog voltage signal is filtered in two stages. The first stage is an RC low-pass filter to remove high-frequency noise, and the second stage is an adaptive notch filter to suppress power frequency interference. The filtered signal is amplified to 0-5V by an instrumentation amplifier, and then converted from analog to digital by a 16-bit ADC to obtain a digital voltage sequence. Feature extraction steps: Calculate feature parameters for the digital voltage sequence. For AC power supplies, these include RMS value, peak value, frequency deviation, and harmonic distortion rate; for DC power supplies, these include average value, ripple coefficient, and fluctuation amount. Anomaly detection steps: Compare the extracted feature parameters with preset thresholds. AC power thresholds include effective value deviation ±5%, frequency deviation ±10Hz, and harmonic distortion rate ≤5%. DC power thresholds include average value deviation ±10%, ripple coefficient ≤1%, and fluctuation ≤0.5V. If any feature parameter exceeds the threshold, it is determined to be an anomaly. Record the time of occurrence, duration, and corresponding feature value of the anomaly. Data output steps: The detection results are transmitted to the aviation power monitoring terminal via digital signals and stored in the local memory. When an abnormal state is triggered, a switch signal is synchronously output to the alarm device. Dynamic threshold adjustment steps: The anomaly detection threshold is corrected in real time based on power load changes. First, the load type is identified by the current harmonic content, and a differentiated load influence coefficient k is configured for different loads. Then, the formula is used... Calculate the adjusted voltage RMS threshold, where U th The threshold value is the adjusted threshold, where U0 is the initial threshold, t0 is the start time of detection, and t is the current time. Let U be the load current at time τ, and the integral term reflects the cumulative load change. After threshold adjustment, continuous monitoring for 3 seconds is performed. If the characteristic parameter still exceeds U... th This is then considered abnormal, to avoid misjudgment caused by instantaneous load fluctuations; The calibration process for testing accuracy is performed automatically once a month, and a new temperature compensation and calibration result verification mechanism has been added: Before calibration, the current ambient temperature T is collected by the system's built-in temperature sensor, and the temperature correction value ΔU is calculated based on the preset temperature-error model; then, a standard voltage source is connected, and each standard voltage value is output in sequence, the actual detection value Udet of the detection device is collected, the difference between the detection value and the standard value Ustd is calculated as ΔUraw=Udet-Ustd, and then the temperature correction value is added to obtain the final error ΔUfinal=ΔUraw+ΔU, generating a calibration coefficient matrix and storing it in the non-volatile memory EEPROM; During routine testing, the ambient temperature is collected in real time. Based on the current temperature and the test voltage value, the corresponding correction value in the calibration coefficient matrix is ​​called to correct the test results. After calibration, the intermediate value of the standard voltage source is connected for verification, and the corrected test value is collected. In the signal preprocessing step, the adaptive notch filter uses a recursive least squares algorithm to dynamically adjust the filter coefficients. The initial filter coefficient vector is updated every 10 sampling points to ensure that the filter attenuates ≥40dB for 50Hz / 60Hz power frequency interference and ≤1dB for the power fundamental frequency. At the same time, a noise intensity detection step is added. By calculating the signal-to-noise ratio (SNR) of the analog voltage signal before filtering, when SNR < 30dB, the notch filter order is automatically increased from 5th to 8th order to enhance interference suppression capability; when SNR > 40dB, the filter order is decreased from 5th to 3rd order to reduce the attenuation of the fundamental frequency signal.

2. The method for detecting the output voltage of an aviation power supply according to claim 1, characterized in that, Also includes: Inter-channel synchronization detection steps: Configure 2-4 independent acquisition channels, and each channel is triggered to sample via a synchronization clock; After synchronous sampling, the phase difference between each channel is calculated. When the phase difference is greater than 5°, the transmission delay is calculated based on the signal transmission distance between channels to generate a phase compensation value, which is used to correct the phase data during feature extraction. At the same time, the sampling values ​​of each channel are monitored in real time. If 100 consecutive sampling points of a channel exceed the normal voltage range, it is determined to be a channel fault. The channel is automatically switched to a backup channel and the fault channel number and fault time are recorded. The cross-correlation coefficient is calculated for the digital voltage sequence of each channel.

3. The method for detecting the output voltage of an aviation power supply according to claim 1, characterized in that, In the anomaly detection step, persistent abnormal states are classified into three levels: Level 1 anomaly is when a single feature parameter exceeds the threshold and the deviation is ≤10%; Level 2 anomaly is when the deviation of a single feature parameter is >10% or two feature parameters exceed the threshold; Level 3 anomaly is when the deviation of a feature parameter is >20% or accompanied by voltage mutation; at the same time, a historical data comparison mechanism is added to retrieve the statistical values ​​of feature parameters for the same period in the past 7 days.

4. The method for detecting the output voltage of an aviation power supply according to claim 1, characterized in that, The data output process also includes data compression and encryption, and adds an edge computing preprocessing step: For locally stored voltage data, the edge computing module performs preliminary analysis to extract the average, maximum, minimum, and effective values ​​every 5 seconds, storing only the statistical values ​​and raw sampled data during abnormal periods, and then compresses it using piecewise linear fitting; The data transmitted to the monitoring terminal is first filtered by edge computing, uploading only the statistical values, abnormal records, and prediction information; In the encryption process, the AES-128 encryption algorithm is used to process the transmitted data, and the key is distributed through an asymmetric encryption mechanism: During system initialization, the monitoring terminal generates an RSA public key and a private key, sends the public key to the detection device, the detection device generates an AES key and encrypts it with the RSA public key, then transmits it to the monitoring terminal, the monitoring terminal decrypts it with the private key to obtain the AES key, and the key is automatically updated every hour.

5. A system for implementing the output voltage detection method based on an aviation power supply as described in any one of claims 1-4, characterized in that, include: Signal acquisition module: It consists of a differential voltage sensor, a high-frequency suppression capacitor, and a sampling trigger circuit. The sensor is connected in parallel to the output of the aviation power supply. The sampling trigger circuit outputs a trigger signal according to the power supply type to control the sensor to synchronously acquire voltage signals. Preprocessing module: Includes RC low-pass filter circuit, adaptive notch filter, instrumentation amplifier, 16-bit ADC, which sequentially filters, amplifies, and converts analog signals to digital signals, and outputs digital voltage sequence; Feature extraction module: Built on FPGA, integrating RMS value calculation unit, peak detection unit, frequency analysis unit, and harmonic analysis unit to calculate the characteristic parameters of AC / DC power supply in real time; Anomaly detection module: Composed of MCU and threshold storage unit. The threshold storage unit pre-stores initial thresholds for different power types. The MCU compares the output of the feature extraction module with the threshold, performs anomaly judgment and classification, and generates anomaly status signal. Data output module: includes RS485 communication interface, local memory, and switch output circuit, which transmits the detection results to the monitoring terminal and stores them to SD card. In case of abnormality, it outputs switch signals to the alarm device. Synchronization and calibration module: includes a synchronization clock generator, standard voltage interface, and calibration coefficient storage unit, provides synchronization trigger signal for inter-channel detection, performs accuracy calibration and stores calibration coefficients; Power supply module: It adopts a wide-voltage input switching power supply to power all modules of the system, and has built-in overcurrent protection and overvoltage protection.

6. The system for detecting the output voltage of an aviation power supply according to claim 5, characterized in that, Also includes: Load current acquisition submodule: Integrated into the signal acquisition module, it uses a Hall current sensor connected in series in the aviation power supply output circuit to acquire the load current signal and convert it into a 0-3.3V analog signal. After being converted into a digital signal by a 12-bit ADC, it is transmitted to the anomaly detection module. The submodule adds a current harmonic detection unit, which implements fast Fourier transform based on FPGA to calculate the total harmonic distortion rate THDi of the current. Meanwhile, an overload prediction unit is added to calculate the slope of current change in real time; the anomaly detection module calculates the dynamic threshold based on the current signal using a formula, with the integral term reflecting the cumulative load change, thus realizing threshold adjustment based on load changes.

7. The system for detecting the output voltage of an aviation power supply according to claim 5, characterized in that, Also includes: Encrypted Communication Submodule: Integrated into the data output module, this submodule uses an encryption chip to receive raw detection data from the data output module, encrypts it in 128-bit groups, and transmits the encrypted data via an RS485 interface. The submodule also includes a new data integrity verification unit that calculates a CRC32 checksum for each encrypted data frame and appends it to the end of the frame. Upon receiving a data frame, the monitoring terminal first calculates the CRC32 checksum of the received data and compares it with the checksum at the end of the frame. A communication interruption reconnection mechanism is also added, monitoring the communication status via the carrier detection pin of the RS485 interface. The monitoring terminal is equipped with the same type of encryption chip and has a built-in key update unit that automatically generates a new AES key every hour.

Images