An edge computing terminal for aeration system health diagnosis and a deployment method thereof

Abstract

The present application relates to sewage treatment equipment technical field, disclose a kind of for aeration system health diagnosis edge computing terminal and its deployment method, terminal includes signal input module, edge computing module, man-machine interaction module and communication module, signal input module accesses pressure sensor signal, edge computing module carries out time-frequency analysis to signal, extracts characteristic parameter and carries out health diagnosis to generate diagnostic result, this module is established in learning mode Health baseline and is automatically switched to diagnostic mode after completion, man-machine interaction module exports diagnostic result and receives user operation instruction, communication module exports diagnostic result to external system;Deployment method includes that terminal is connected in parallel by high impedance input mode and accesses pressure sensor signal line, configuration channel parameter and sensor range, trigger to enter learning mode to establish health baseline, switch to diagnostic mode and output diagnostic result.The present application realizes the local real-time diagnosis and plug and play deployment of aeration system health diagnosis.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment equipment technology, specifically to an edge computing terminal for health diagnosis of aeration systems and its deployment method. Background Technology

[0002] The aeration system is the core unit of aerobic biological treatment in wastewater treatment plants, and its operating status directly affects treatment efficiency and energy consumption. The main types of aeration system failures include aerator blockage and pipeline leaks, both of which increase energy consumption and decrease treatment efficiency. Diagnostic algorithms for these failures are relatively mature; however, in practical engineering applications, the implementation of these algorithms faces the following equipment-level challenges.

[0003] First, there is a lack of dedicated on-site diagnostic equipment. Currently, there are no edge computing terminals on the market specifically designed for the health diagnosis of aeration systems. In related technologies, implementing the aforementioned diagnostic methods requires either uploading pressure signals to a host computer or the cloud for calculation, or using a general-purpose data acquisition device in conjunction with an industrial computer to run diagnostic software. The former is costly, has high latency, and relies on a network, while the latter involves dispersed equipment, complex wiring, and poor reliability in coordination.

[0004] Secondly, the core processors of existing wastewater treatment plant distributed control systems or programmable logic controllers are mostly low-frequency microcontrollers designed for industrial logic control, with a clock frequency typically below 100MHz. While these are sufficient to perform routine on / off control and PID regulation tasks, their weak floating-point arithmetic capabilities and limited memory resources make it difficult to handle computationally intensive time-frequency analysis algorithms such as real-time fast Fourier transform, power spectrum analysis, and sample entropy calculation of pressure pulsation signals. Consequently, they cannot integrate online health diagnostic functions.

[0005] Furthermore, time-frequency analysis algorithms involve significant computational demands, placing certain requirements on the processor's floating-point arithmetic capabilities and storage space. A key technical challenge in engineering implementation is how to embed these algorithms into low-cost, low-power embedded terminals while ensuring real-time performance. While existing general-purpose embedded development boards possess sufficient computing power, their interface designs are incompatible with the signal characteristics of pressure sensors, necessitating the design of additional signal conditioning circuitry.

[0006] Therefore, how to apply aeration system health diagnosis algorithms in a way that is efficient, reliable, and easy to deploy in wastewater treatment plants is a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0007] This invention provides an edge computing terminal for health diagnosis of aeration systems and its deployment method, in order to solve the technical problems in related technologies such as the lack of edge computing terminals dedicated to health diagnosis of aeration systems, the insufficient computing power of existing control systems to run frequency analysis algorithms, and the difficulty in directly deploying and running diagnostic algorithms on general embedded platforms in real time. This invention deploys the health diagnosis algorithm of aeration systems in a fixed form at the wastewater treatment plant site to achieve localized real-time diagnosis.

[0008] In a first aspect, the present invention provides an edge computing terminal for health diagnosis of an aeration system, comprising: a signal input module, the input end of which is connected to a pressure sensor to receive pressure sensor signals; the signal input module is provided with a signal conditioning circuit for processing the pressure sensor signals and outputting digital signals; an edge computing module, connected to the signal input module, for receiving digital signals; the edge computing module is provided with an embedded processor and a memory; the memory stores a time-frequency analysis diagnostic algorithm program; the embedded processor is used to execute the time-frequency analysis diagnostic algorithm program to perform time-frequency analysis on the digital signals to extract feature parameters, and perform health diagnosis based on the feature parameters to generate diagnostic results; a human-computer interaction module, connected to the edge computing module, for outputting diagnostic results to the user and receiving user operation instructions; and a communication module, connected to the edge computing module, for outputting diagnostic results to an external system; the edge computing module has a learning mode and a diagnostic mode; in the learning mode, it collects operating data to establish a health baseline for diagnostic comparison, and automatically switches to the diagnostic mode after the baseline is established, and performs health diagnosis in the diagnostic mode.

[0009] This invention solves the technical problem of incompatibility between general development board interfaces and sensor signal characteristics, which prevents direct access and deployment, by setting up a signal input module to directly connect to existing pressure sensor signals and using signal conditioning circuitry for adaptation processing. By setting up an edge computing module, the time-frequency analysis and diagnostic algorithm program is embedded in memory and executed locally by the embedded processor. The local computing power of the edge computing module directly completes time-frequency analysis and health diagnosis without relying on a host computer or cloud, solving the technical problem of insufficient computing power in existing DCS / PLCs for running time-frequency analysis algorithms. It also avoids the high cost, high latency, and network dependence issues associated with uploading pressure signals to a host computer or cloud. By setting up a human-machine interaction module and a communication module, it can both output diagnostic results on-site and receive user operation commands, and output diagnostic results to external systems, achieving on-site visualization and remote transmission of diagnostic results. The edge computing module has learning and diagnostic modes. In learning mode, it can automatically collect operating data to establish a health baseline and automatically switch to diagnostic mode after establishment, enabling the terminal to be used immediately and automatically in the field, reducing the complexity of on-site deployment and maintenance.

[0010] In one optional implementation, the signal conditioning circuit includes an anti-aliasing low-pass filter, a programmable gain amplifier, and an analog-to-digital converter (ADC). The anti-aliasing low-pass filter is used to perform anti-aliasing filtering on the pressure sensor signal, the programmable gain amplifier is used to amplify the filtered signal, and the ADC is used to convert the amplified analog signal into a digital signal. The input of the anti-aliasing low-pass filter is connected to the pressure sensor signal, and its output is connected to the input of the programmable gain amplifier. The output of the programmable gain amplifier is connected to the input of the ADC, and the ADC outputs a digital signal. In this implementation, the signal conditioning circuit uses the anti-aliasing low-pass filter to filter out high-frequency noise in the pressure sensor signal, preventing sampling aliasing and ensuring the fidelity of signal acquisition. The programmable gain amplifier matches the range of the ADC according to the sensor output amplitude, maximizing the dynamic range of the signal and avoiding signal distortion. The ADC converts the analog signal into a digital signal for processing by the edge computing module, enabling the terminal to directly adapt to the signal characteristics of existing pressure sensors without the need for additional interface conversion circuits. This achieves plug-and-play access to existing field equipment, reducing the difficulty of field modification and deployment costs.

[0011] In one optional implementation, the embedded processor executes the time-frequency analysis and diagnostic algorithm program using at least one of the following optimization methods: the Fast Fourier Transform employs a hybrid basis algorithm and utilizes a single-instruction multiple-data (SMD) instruction set; the edge computing module uses a sliding window to process digital signals, with the sliding step size of adjacent windows being smaller than the window length; and the sample entropy employs an approximate nearest neighbor search algorithm. This implementation significantly improves the computational efficiency of frequency domain transformation by employing a hybrid basis algorithm and utilizing a SMD instruction set to execute the Fast Fourier Transform, enabling real-time time-frequency analysis on the embedded platform. By using a sliding window to process digital signals with a sliding step size smaller than the window length, data reuse between adjacent windows is achieved, reducing the computational overhead caused by repeated calculations. By using an approximate nearest neighbor search algorithm to calculate sample entropy, the computational complexity is reduced from quadratic to linear logarithmic, significantly shortening the computation time. At least one of the above optimization methods enables the time-frequency analysis and diagnostic algorithm to run efficiently on the embedded processor, ensuring the real-time nature of health diagnosis and solving the technical problems of high computational load and high processor performance requirements in time-frequency analysis algorithms.

[0012] In one optional implementation, health diagnosis includes calculating a blockage severity index, which characterizes the severity of aerator blockage. The blockage severity index is calculated by comparing the real-time energy proportions of each frequency band with the corresponding energy proportions of each frequency band in the healthy baseline, calculating a weighted comprehensive deviation value, and then multiplying it by the amplification factor of the real-time pulsation amplitude relative to the healthy baseline. This implementation quantifies the aerator blockage severity into an intuitive blockage severity index, providing maintenance personnel with a quantitative basis for judging the severity of blockage, avoiding the subjectivity of relying on manual experience, and helping to promptly detect aerator blockage trends and carry out targeted maintenance.

[0013] In one optional implementation, the health diagnosis includes calculating a leakage severity index, which characterizes the severity of pipeline leakage. The leakage severity index is calculated by weighting the increase in real-time high-frequency energy ratio relative to the healthy baseline, the decrease in real-time pulsation amplitude relative to the healthy baseline, and the increase in real-time sample entropy relative to the healthy baseline according to preset weights. This implementation, by weighting the three dimensions—the increase in real-time high-frequency energy ratio relative to the healthy baseline, the decrease in real-time pulsation amplitude relative to the healthy baseline, and the increase in real-time sample entropy relative to the healthy baseline—into a leakage severity index, comprehensively reflects the severity of pipeline leakage from multiple dimensions. Compared to a single-parameter threshold judgment method, this multi-dimensional feature-based evaluation method effectively improves the accuracy and reliability of leakage fault diagnosis.

[0014] In one optional implementation, the terminal has a protection rating of at least IP65, an operating temperature range of -20℃ to 70℃, and a DIN35 rail mounting structure. This implementation sets the terminal's protection rating to at least IP65, providing complete dust and water spray protection, effectively resisting the high humidity and corrosive gases present in wastewater treatment plants, and preventing equipment failure due to dust and liquid intrusion. By setting the operating temperature range to -20℃ to 70℃, the terminal can operate normally in both low-temperature winter and high-temperature summer environments, adapting to extreme temperature differences in different regions. The DIN35 rail mounting structure allows the terminal to be directly mounted on the standard rails of existing control cabinets, eliminating the need for custom mounting brackets or cabinet modifications. These environmentally adaptable designs collectively ensure the long-term reliable operation of the terminal under harsh conditions in wastewater treatment plants, reducing equipment maintenance frequency and replacement costs.

[0015] In one optional implementation, the communication module supports at least one of the following protocols: Modbus RTU, Modbus TCP, OPC UA, and MQTT, and supports uploading to a remote server via 4G / 5G. In this implementation, by supporting at least one of these protocols, the communication module enables the terminal to be compatible with various existing industrial control systems and information platforms in wastewater treatment plants. Diagnostic results can be seamlessly integrated into existing distributed control systems, central control room monitoring systems, or cloud-based operation and maintenance platforms without the need for additional protocol conversion gateways. Simultaneously, by supporting 4G / 5G wireless communication for uploading to a remote server, the terminal can achieve remote data transmission in field environments lacking wired network coverage. This enables flexible output and remote monitoring of diagnostic results, reducing the complexity of system integration and implementation costs.

[0016] Secondly, the present invention provides a deployment method for an edge computing terminal for health diagnosis of an aeration system, comprising the following steps: connecting the edge computing terminal in parallel to the signal line of a pressure sensor via a high-impedance input method; configuring channel parameters and sensor range for the edge computing terminal; triggering the edge computing terminal to enter learning mode, collecting operating data to establish a health baseline; and after the health baseline is established, triggering the edge computing terminal to switch to diagnostic mode and output diagnostic results.

[0017] In the above deployment method, the edge computing terminal is connected in parallel to the signal line of the pressure sensor through a high-impedance input method. This allows the terminal to only listen to the signal without injecting any excitation into the line, and there is no need to cut off or modify the original signal line. Even if the terminal itself fails or loses power, it will not affect the signal transmission and normal operation of the original control system. After configuring the channel parameters and sensor range, the terminal is triggered to enter the learning mode, collect operating data to establish a health baseline, and automatically switch to the diagnostic mode after the health baseline is established. The entire process does not require manual calibration or parameter tuning, realizing a "plug and play" deployment method. This significantly reduces the professional skills required for on-site deployment of operation and maintenance personnel, enabling on-site personnel with a process background to independently complete the installation and commissioning of the terminal.

[0018] In one optional implementation, the input impedance of the high-impedance input method is not less than 1MΩ, and the current drawn by the edge computing terminal from the signal source does not exceed 50μA. This implementation ensures that the load error introduced by the terminal when connecting to the signal line does not exceed 0.05%, and the impact on the output of the original sensor signal is negligible, ensuring that the measurement accuracy and control function of the original control system are not affected, and guaranteeing the operational safety of the original system while acquiring diagnostic signals.

[0019] In one optional implementation, the health baseline includes a pulse amplitude benchmark, a mean pressure benchmark, an energy percentage benchmark for each frequency band, a high-frequency energy ratio benchmark, and a sample entropy benchmark. This implementation provides the terminal with a multi-dimensional and comprehensive health status reference benchmark, enabling subsequent diagnostic processes to comprehensively judge the operating status of the aeration system from multiple feature dimensions, avoiding misjudgments or omissions due to reliance on a single parameter, and improving the accuracy and reliability of fault diagnosis. Simultaneously, this health baseline is established and stored locally on the terminal, eliminating the need for external databases or cloud comparisons, allowing the terminal to independently complete diagnostic decisions on-site, ensuring that the diagnostic function continues to operate normally even in abnormal situations such as network interruptions. Attached Figure Description

[0020] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a hardware architecture block diagram of an edge computing terminal for health diagnosis of an aeration system according to an embodiment of the present invention; Figure 2 This is a flowchart illustrating a method for deploying an edge computing terminal for health diagnosis of an aeration system according to an embodiment of the present invention. Figure 3 This is a wiring diagram of a terminal deployment bypass access according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0022] 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, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0024] 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 technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0025] The aeration system is the core unit of aerobic biological treatment in wastewater treatment plants. Aerator blockage and pipeline leaks both lead to increased energy consumption and decreased treatment efficiency. Currently, diagnostic algorithms for these faults face the following equipment-level problems in practical engineering applications: First, there is a lack of dedicated on-site diagnostic equipment. Related technologies can only adopt cloud computing or solutions using general-purpose data acquisition devices in conjunction with industrial computers. The former is costly, has high latency, and relies on the network, while the latter involves dispersed equipment, complex wiring, and poor collaborative reliability. Second, existing DCS / PLC controllers lack sufficient computing power to handle time-frequency analysis algorithms. Third, time-frequency analysis algorithms involve large computational loads, and the interfaces of general-purpose embedded development boards are incompatible with the sensor signal characteristics, making it difficult to directly deploy and run the algorithms in real time on embedded platforms. Based on this, this invention provides an edge computing terminal for health diagnosis of aeration systems and its deployment method to solve the technical problems of lacking dedicated edge computing terminals for aeration system health diagnosis, insufficient computing power of existing control systems to run time-frequency analysis algorithms, and the difficulty in directly deploying and running diagnostic algorithms in real time on general-purpose embedded platforms.

[0026] According to an embodiment of the present invention, an edge computing terminal for health diagnosis of an aeration system is provided. Figure 1 This is a hardware architecture block diagram of an edge computing terminal for health diagnosis of an aeration system according to an embodiment of the present invention, such as... Figure 1 As shown, the edge computing terminal includes a signal input module, an edge computing module, a human-computer interaction module, and a communication module.

[0027] (1) Signal input module.

[0028] The signal input module features an input terminal and a multi-channel analog input interface. The input terminal connects to a pressure sensor to receive its signal. The multi-channel analog input interface supports 4-20mA current signals and 0-10V voltage signals, ensuring compatibility with existing pressure transmitters in wastewater treatment plants. The multi-channel analog input interface supports 8, 16, or 32 channels, allowing simultaneous connection of pressure sensors to multiple aeration branch pipes. The appropriate model can be selected based on the actual number of branch pipes.

[0029] The signal input module is equipped with a signal conditioning circuit, which is used to process the pressure sensor signal and output a digital signal.

[0030] In one alternative implementation, the signal conditioning circuitry includes an anti-aliasing low-pass filter, a programmable gain amplifier, and an analog-to-digital converter.

[0031] The anti-aliasing low-pass filter is connected to an external pressure sensor to perform anti-aliasing filtering on the pressure sensor signal, thereby filtering out high-frequency noise above the Nyquist frequency and preventing sampling aliasing. The anti-aliasing low-pass filter is a second-order Butterworth active filter with a cutoff frequency configurable to two levels: 0.4fs and 0.5fs, where fs is the sampling rate, and the stopband attenuation is not less than 40dB / decibels.

[0032] Programmable gain amplifiers are used to amplify filtered signals to match the analog-to-digital converter range to the sensor output amplitude, maximizing dynamic range. The gain of a programmable gain amplifier can be selected as 1x, 2x, 4x, or 8x.

[0033] An analog-to-digital converter (ADC) is used to convert amplified analog signals into digital signals. This ADC employs a 24-bit Σ-Δ type ADC with a maximum sampling rate of 30kSPS. The sampling rate can be configured to 100Hz, 200Hz, 500Hz, 1kHz, or 2kHz, with at least 20 effective bits and a signal-to-noise ratio (SNR) of at least 100dB. The ADC front-end incorporates a programmable digital filter to suppress power frequency (50Hz / 60Hz) interference.

[0034] The input of the anti-aliasing low-pass filter is connected to the pressure sensor signal, the output of the anti-aliasing low-pass filter is connected to the input of the programmable gain amplifier, the output of the programmable gain amplifier is connected to the input of the analog-to-digital converter, and the analog-to-digital converter outputs a digital signal.

[0035] In one optional implementation, the signal input module is further provided with a sensor excitation power supply for providing a 24V DC excitation power supply to an external two-wire or three-wire pressure transmitter, with a maximum output current of 200mA, and is equipped with overcurrent protection and short-circuit protection.

[0036] In one optional embodiment, the signal conditioning circuit further includes a signal conversion circuit, which is used to convert a 4-20mA current signal into a voltage signal through a sampling resistor, and to directly connect a 0-10V voltage signal. The sampling resistor has a resistance of 250Ω and an accuracy of 0.1%. The sampling resistor is a low-temperature drift device with a temperature drift coefficient of 25ppm / ℃ to ensure conversion accuracy over a wide temperature range.

[0037] (2) Edge computing module.

[0038] The edge computing module is connected to the signal input module to receive digital signals. The edge computing module is equipped with an embedded processor and a memory. The memory stores time-frequency analysis and diagnostic algorithm programs. The embedded processor is used to execute the time-frequency analysis and diagnostic algorithm programs to perform time-frequency analysis on the digital signals to extract feature parameters, and perform health diagnosis based on the feature parameters to generate diagnostic results.

[0039] In one alternative implementation, the embedded processor is an industrial-grade embedded processor (ARM Cortex-A series or equivalent performance) with a main frequency of not less than 800MHz, an independent floating-point arithmetic unit and digital signal processing instruction set, and hardware acceleration for time-frequency analysis algorithms such as Fast Fourier Transform.

[0040] In one optional implementation, the memory is a non-volatile memory (Flash) with a capacity of not less than 4GB, used to store the operating system, time-frequency analysis and diagnostic algorithm program, health baseline data, and historical diagnostic records.

[0041] In one alternative implementation, the edge computing module further includes working memory (DDR RAM) with a capacity of not less than 1GB, used for real-time data caching and algorithm execution.

[0042] In one alternative implementation, the edge computing module employs an embedded real-time operating system to ensure the real-time performance of data acquisition and algorithm execution. The operating system is streamlined and only retains essential components related to signal acquisition, time-frequency analysis and diagnostic algorithm execution, data storage, and communication.

[0043] In one optional implementation, to implement real-time time-frequency analysis on an embedded processor, the embedded processor includes at least one of the following optimization methods when executing the time-frequency analysis diagnostic algorithm program: Optimization Method 1: The Fast Fourier Transform (FFT) employs a hybrid basis algorithm and utilizes a Single Instruction Multiple Data (SID) instruction set. Specifically, a hybrid basis FFT algorithm with radix-4 as the primary base is used, and vectorization acceleration is achieved using the NEON SIMD instruction set of the ARM Cortex-A series processors. Taking a 1024-point FFT as an example, the computation time is <5ms.

[0044] Optimization Method Two: The edge computing module employs a sliding window to process digital signals. The sliding step size of adjacent windows is less than the window length, and adjacent window data overlaps by 50%. Data reuse and incremental update strategies reduce redundant computations. Specifically, overlapping windows (window length 10 seconds, step size 5 seconds) are used, with adjacent window data overlapping by 50%. Through data reuse and incremental update strategies, redundant computations are reduced, resulting in an overall reduction in computational load of approximately 40%.

[0045] Optimization Method 3: The sample entropy adopts an approximate nearest neighbor search algorithm instead of exhaustive matching, reducing the computational complexity of sample entropy from O(N²) to O(N log N), shortening the computation time by about 70%.

[0046] Tests showed that, under conditions of a sampling rate of 100Hz, a window length of 10 seconds, and simultaneous monitoring of 8 channels, the single-channel diagnostic calculation time was less than 50ms, and the eight-channel polling cycle was less than 400ms, meeting the requirements for real-time monitoring.

[0047] The edge computing module has a learning mode and a diagnostic mode. In the learning mode, it collects running data to establish a health baseline for diagnostic comparison. After the baseline is established, it automatically switches to the diagnostic mode to perform health diagnosis.

[0048] In one optional implementation, the health diagnosis includes calculating a blockage index, which is used to characterize the severity of aerator blockage. The blockage index is calculated by comparing the energy proportion of each frequency band in real time with the energy proportion of each frequency band in the healthy baseline and calculating the weighted comprehensive deviation value, and then multiplying it by the amplification factor of the real-time pulsation amplitude relative to the healthy baseline.

[0049] In one optional implementation, the health diagnosis includes calculating a leakage severity index, which is used to characterize the severity of pipeline leakage. The leakage severity index is calculated by weighting the increase of real-time high-frequency energy ratio relative to the health baseline, the attenuation of real-time pulsation amplitude relative to the health baseline, and the increase of real-time sample entropy relative to the health baseline according to preset weights.

[0050] In one alternative implementation, the edge computing module permanently executes the following diagnostic algorithm.

[0051] The first step is signal preprocessing. Preprocessing of the digital signal includes digital filtering, baseline drift removal, and outlier removal. Bandpass filtering is used for digital filtering, with a passband frequency ranging from 10Hz to 1000Hz.

[0052] The second step is to perform time-frequency analysis. In this embodiment, the "time-frequency analysis diagnostic algorithm program" refers to the computer program embedded in the edge computing module used to perform time-frequency analysis and related diagnostic functions. In this step, "time-frequency analysis" specifically refers to performing a Fast Fourier Transform on the preprocessed digital signal to convert the signal from the time domain to the frequency domain in order to extract frequency domain feature parameters.

[0053] The third step is feature extraction.

[0054] Feature parameters are extracted based on time-frequency analysis results. These parameters include pulsation amplitude, mean pressure, dominant frequency, high-frequency energy ratio, sample entropy, harmonic distortion rate, and low-frequency energy ratio.

[0055] The high-frequency energy ratio is defined as the percentage of energy in the 0.25fs to 0.5fs frequency band relative to the total energy in the entire frequency band (0 to 0.5fs), where fs is the sampling rate. The high-frequency energy ratio is used to detect pipeline leaks; a significant increase in the high-frequency energy ratio indicates a leak. It should be noted that when there is a leak, turbulence is generated at the leak point, and this turbulence induces high-frequency pressure pulsations. Therefore, an increase in the high-frequency energy ratio can be used as a basis for determining pipeline leaks.

[0056] The low-frequency energy ratio is defined as the percentage of energy in the 0 to 0.05 fs frequency band relative to the total energy in the entire frequency band (0 to 0.5 fs). The low-frequency energy ratio is used to distinguish between different types of blockage. When the low-frequency energy ratio increases significantly, it indicates that the surface of the aerator membrane is dominated by microbial film.

[0057] The fourth step is fault diagnosis.

[0058] Fault diagnosis is performed based on a comprehensive judgment matrix that considers pulsation amplitude, average pressure, dominant frequency offset, high-frequency energy ratio, and sample entropy, distinguishing between aerator blockage and pipeline leakage. The judgment matrix is ​​automatically generated by the terminal during the baseline learning phase based on historical data. The judgment matrix contains built-in fault discrimination rules, with the input being the direction and amplitude of change of each feature parameter relative to the healthy baseline, and the output being the fault type conclusion.

[0059] The correspondence between the direction of change of each characteristic parameter and the fault type is as follows: Increased pulsation amplitude, elevated average pressure, and negative shift in dominant frequency indicate a clogged aerator. A decrease in pulsation amplitude, a decrease in average pressure, a positive shift in the dominant frequency, and an increase in the high-frequency energy ratio indicate a pipeline leak. A decrease in high-frequency energy and sample entropy compared to normal indicates a periodic signal, suggesting blockage. Decreased sample entropy indicates enhanced signal regularity, meaning the current waveform shows high repetition with historical waveform patterns, corresponding to a blocked aerator. This is because bubble bursting remains a periodic event even when the aerator is blocked, resulting in a periodic signal pattern. An increase in high-frequency energy ratio and sample entropy indicates that the signal tends to be random, which is identified as a leakage characteristic. An increase in sample entropy indicates that the randomness of the signal is enhanced, that is, the current waveform has low repeatability with historical waveform patterns, which corresponds to the state of pipeline leakage, because turbulence noise is a random event during leakage, and the signal exhibits randomness.

[0060] The fifth step is to calculate the blockage degree index and the leakage severity index.

[0061] The system automatically calculates the clogging severity index and outputs the corresponding level. The clogging severity index is calculated by comparing the real-time energy proportion of each frequency band with the corresponding energy proportion of each frequency band in the healthy baseline, calculating the weighted comprehensive deviation value, and then multiplying it by the amplification factor of the real-time pulsation amplitude relative to the healthy baseline. The clogging severity index characterizes the severity of aerator clogging, with a value ranging from 0 to 1. The closer the clogging severity index is to 1, the more severe the clogging.

[0062] The system automatically calculates the leakage severity index and outputs the corresponding level. The leakage severity index is calculated by weighting the increase in real-time high-frequency energy ratio relative to the healthy baseline, the decrease in real-time pulsation amplitude relative to the healthy baseline, and the increase in real-time sample entropy relative to the healthy baseline according to preset weights. The leakage severity index characterizes the severity of pipeline leakage; the higher the index, the more severe the leakage.

[0063] Step 6: Subtype identification. Microbial film blockage is distinguished from inorganic scaling blockage by low-frequency energy ratio, and flange leakage, pipe cracks, and diaphragm damage are distinguished by high-frequency energy spectrum distribution morphology.

[0064] Step 7, Trend Analysis. Calculate the rate of change (first derivative) of the blockage degree index and the leakage severity index. Issue an early warning when the rate of change exceeds a preset threshold, and predict the time when the severity threshold will be reached.

[0065] In one alternative implementation, the edge computing module also includes data management functions for performing health baseline establishment, baseline updates, and data storage management.

[0066] Health Baseline Establishment: After installation or maintenance, the terminal automatically enters learning mode and continuously collects operational data for a preset duration to establish health characteristic baselines for each branch. In one example, the preset duration is 24 hours. Optionally, the terminal continuously collects operational data for 24 hours in learning mode, extracts characteristic parameters for each channel, and forms a health baseline that includes pulsation amplitude baseline, average pressure baseline, energy ratio baseline for each frequency band, high-frequency energy ratio baseline, and sample entropy baseline. This health baseline serves as a comparison basis for determining whether a system fault has occurred and assessing the severity of the fault in subsequent diagnostic modes.

[0067] Baseline Update: Supports manual or automatic periodic updates of the health baseline. In one example, the automatic periodic update cycle is once per quarter. After triggering the baseline update, the terminal re-enters learning mode, collects operational data under the current normal operating condition, establishes a new health baseline, and automatically switches back to diagnostic mode after establishment to adapt to slow changes in characteristic parameters caused by factors such as sensor drift, changes in operating conditions, or equipment aging.

[0068] Data storage management: A circular overwrite strategy is used to manage storage space. When storage space is insufficient, the oldest stored data is automatically overwritten to ensure that storage space does not overflow.

[0069] (3) Human-computer interaction module.

[0070] The human-computer interaction module is connected to the edge computing module and is used to output diagnostic results to the user and receive user operation commands.

[0071] In one alternative implementation, the human-computer interaction module includes a touch screen, physical buttons, and status indicator lights.

[0072] The touchscreen is an industrial-grade capacitive touchscreen with a size of no less than 7 inches, used to display the following in real time: the health status of each branch pipe, including normal, warning and alarm status; blockage degree index and leakage severity index; trend curve; diagnostic conclusions and maintenance suggestions.

[0073] The physical buttons include a power button, a reset button, an confirmation button, and directional buttons, which are used for on-site operation.

[0074] The status indicator lights include a power indicator light, a running indicator light, an alarm indicator light, and a communication status indicator light, which are used to indicate the terminal's power-on status, running status, alarm status, and communication status, respectively.

[0075] (4) Communication module.

[0076] The communication module connects to the edge computing module to output diagnostic results to external systems.

[0077] In one alternative implementation, the communication module supports at least one of the Modbus RTU, Modbus TCP, OPC UA, and MQTT protocols, and supports uploading to a remote server via 4G / 5G.

[0078] In one alternative implementation, the communication module includes a wired communication interface, a wireless communication interface, and a digital output interface.

[0079] The wired communication interfaces include an Ethernet interface and an RS485 interface. The Ethernet interface uses an RJ45 connector and supports Modbus TCP and OPC UA protocols for data exchange with host computers, distributed control systems, or programmable logic controllers. The RS485 interface supports the Modbus RTU protocol and is used for data exchange with distributed control systems or programmable logic controllers.

[0080] The wireless communication interface includes a 4G / 5G module, a Wi-Fi module, and a Bluetooth module. The 4G / 5G module is used to upload diagnostic data to the cloud-based operation and maintenance platform. The Wi-Fi module is used for on-site wireless debugging. The Bluetooth module is used for near-field communication with mobile terminals.

[0081] The digital output interface is a dry contact relay output with 4 channels, used to directly output alarm signals to audible and visual alarms or central control room alarm systems.

[0082] The communication module is also used to communicate with the pressure sensor and read the real-time measurement values ​​of the pressure sensor.

[0083] (5) Storage module.

[0084] In one optional implementation, the edge computing terminal also includes a storage module. Specifically, a local data storage unit (solid-state drive or SD card, with a capacity of at least 128GB) is provided to continuously store the following data: raw pressure pulsation signals (configurable storage duration, typically 7 days), real-time characteristic parameters (1 year storage duration), and diagnostic results and alarm events (3 years storage duration). A historical data playback interface is also provided, supporting data export via USB or network interface.

[0085] (6) Power supply module.

[0086] In one alternative implementation, the edge computing terminal is further provided with a power module, which is connected to each module of the edge computing terminal and is used to supply power to each module. The power module can be integrated inside the terminal housing or set up as an independent module.

[0087] The power module uses an industrial-grade switching power supply, supporting a wide voltage input range of AC 85-265V or DC 24V to adapt to different power supply environments in wastewater treatment plants. The power module is equipped with a power filtering circuit to suppress electromagnetic interference on the power supply line, ensuring signal acquisition quality.

[0088] In one alternative implementation, the power module is also equipped with a backup battery to continuously power the terminal during a power outage for at least 30 minutes and to issue a power failure alarm during a power outage.

[0089] (7) Terminal external structure.

[0090] In one optional implementation, the edge computing terminal employs an industrial-grade protective shell with a protection rating of at least IP65 for dust and water resistance, adapting to the high humidity and corrosive gas environment of a wastewater treatment plant. The terminal is equipped with a heat dissipation structure to adapt to a wide operating temperature range of -20℃ to 70℃. The heat dissipation structure utilizes an aluminum alloy casing or active fan cooling.

[0091] In one alternative implementation, the terminal is equipped with a rail mounting structure for installation within an existing control cabinet; the terminal also has wall-mounting holes for independent wall mounting. The rail mounting structure is compatible with DIN 35mm standard rails.

[0092] In one alternative implementation, the terminal's dimensions are width × height × depth = 250mm × 180mm × 80mm.

[0093] The terminal's interface layout is as follows: The front panel features a 7-inch touchscreen display with physical buttons below it, including up, down, confirm, and back buttons. A status indicator light is located on the right side of the display. The rear panel has a DIN35 rail slot and wall-mounting holes. The bottom and / or side panels have signal input terminals, a communication interface, a power input terminal, digital output terminals, and an antenna interface. The signal input terminals are pluggable terminal blocks. The communication interfaces include an Ethernet port, an RS485 interface, a USB port, and a SIM card slot. There are two Ethernet ports, two RS485 interfaces, two USB ports, and one SIM card slot.

[0094] The edge computing terminal for health diagnosis of aeration systems provided in this embodiment solves the technical problem of incompatibility between the general development board interface and the sensor signal characteristics, which prevents direct access and deployment, by setting up a signal input module to directly connect to existing pressure sensor signals and using signal conditioning circuitry for adaptation processing. By setting up an edge computing module, the time-frequency analysis and diagnosis algorithm program is embedded in the memory and executed locally by the embedded processor. The local computing power of the edge computing module is used to directly complete time-frequency analysis and health diagnosis, without relying on a host computer or cloud. This solves the technical problem that the existing DCS / PLC has insufficient computing power to run the time-frequency analysis algorithm, and avoids the problems of high cost, large latency and network dependence caused by uploading pressure signals to a host computer or cloud. By setting up a human-machine interaction module and a communication module, it can output diagnostic results on-site and receive user operation commands, and can also output diagnostic results to external systems, realizing on-site visualization and remote transmission of diagnostic results. The edge computing module has learning mode and diagnostic mode. In learning mode, it can automatically collect operating data to establish a health baseline and automatically switch to diagnostic mode after the baseline is established. This realizes the terminal's immediate use and automated operation on-site, reducing the complexity of on-site deployment and maintenance.

[0095] Meanwhile, the terminal has complete signal acquisition, data processing, diagnostic calculation, and result output functions, and can operate independently without relying on a host computer or cloud platform. Even if the network is interrupted, the terminal can still work normally, with diagnostic data stored locally and automatically retransmitted after the network is restored, ensuring the availability and data integrity of the terminal under unstable network conditions at the wastewater treatment plant site.

[0096] According to embodiments of the present invention, the present invention also provides an embodiment of a deployment method for an edge computing terminal for health diagnosis of an aeration system. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0097] This embodiment provides a method for deploying an edge computing terminal for health diagnosis of an aeration system. Figure 2 This is a flowchart of a method for deploying an edge computing terminal for health diagnosis of an aeration system according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S201, will Figure 1 The edge computing terminal shown in the embodiment is connected in parallel to the signal line of the pressure sensor via a high-impedance input method.

[0098] In one alternative implementation, the input impedance of the high-impedance input method is not less than 1MΩ, and the current drawn by the edge computing terminal from the signal source does not exceed 50μA.

[0099] Among them, the high impedance input method refers to the terminal input presenting a high equivalent impedance, so that the current drawn from the signal source is extremely small, thereby avoiding substantial impact on the original signal.

[0100] Parallel connection refers to the connection of the terminal with the sensor and the original control system (such as PLC or DCS) in parallel. That is, the terminal is not connected in series with the signal loop, but only bridging the two ends of the signal line to obtain the signal. The significance of doing this is that the terminal can obtain the pressure signal required for diagnosis without changing the electrical parameters of the original signal line, and without affecting the normal operation and gas supply control function of the original control system. Moreover, the terminal operates independently, and even if it fails, it will not cause the original control system to fail, thus ensuring the reliability and safety of the system.

[0101] Figure 3 Wiring diagram for deploying bypass access to the terminal. (Example) Figure 3 As shown, a pressure sensor is installed on the aeration branch pipe. The pressure sensor is connected to the original control system (such as a PLC or DCS) via a signal line to transmit pressure signals to the original control system. The original control system reads the pressure signals and controls the air supply. The edge computing terminal is connected in parallel to this signal line through parallel connection points (T-joints or terminal blocks). Specifically, the terminal's signal input module is connected to the signal line and the common terminal through a high-impedance differential amplifier (input impedance ≥ 1MΩ). It only obtains signals from the signal line, does not change the original electrical parameters of the signal line, and does not introduce signals into the signal loop.

[0102] like Figure 3 As shown, there is no series connection between the edge computing terminal and the original control system of the present invention. Both are connected to the signal line of the pressure sensor. However, the access of the terminal will not affect the signal reception of the original control system. Even if the terminal fails, it will not cause the signal line to be interrupted, and the original control system can still operate normally.

[0103] Step S202: Configure channel parameters and sensor range to the edge computing terminal.

[0104] Among them, the channel parameters refer to the configuration information of the signal input channel corresponding to the aeration branch pipe monitored by the terminal, including the channel number, the corresponding branch pipe number, the sampling rate, etc.

[0105] Sensor range refers to the measurement range of the connected pressure sensor, including the lower limit and the upper limit of the range. It is used to convert the current signal (such as 4-20mA) or voltage signal (such as 0-10V) output by the sensor into the corresponding pressure physical quantity value.

[0106] Maintenance personnel input the above configuration information through the terminal's human-machine interface (such as a touchscreen). The terminal initializes the parameters of the signal input module according to the configuration information, enabling the terminal to correctly identify and parse the sensor signals connected to each channel, ensuring the accuracy of subsequent signal acquisition and diagnostic calculations. After configuration, the terminal can continuously acquire pressure signals from each channel according to the set sampling rate in subsequent steps.

[0107] During the above deployment process, the terminal achieves bypass access and plug-and-play configuration of the signal lines to be deployed in the following ways.

[0108] For bypass access, the terminal input adopts a high-impedance input method. Specifically, the terminal is connected in parallel to a 4-20mA current loop or voltage signal line through a high-impedance differential amplifier. For the 4-20mA loop, the terminal input samples the voltage difference generated by the loop current across the sampling resistor through the high-impedance differential amplifier, without changing the total loop impedance. For the 0-10V voltage signal, the terminal input uses a JFET input stage operational amplifier with an input impedance of not less than 10MΩ and an input bias current of less than 10pA. When connected in parallel with high-impedance input, the current drawn by the terminal from the signal source is less than 50μA, and the introduced load error is less than 0.05%, which has a negligible impact on the original sensor's signal output. The terminal input also has transient suppression diodes and electrostatic discharge protection circuits to prevent surges and electrostatic discharge from damaging the terminal; the terminal input also has a common-mode filter to suppress common-mode interference introduced by long-line transmission.

[0109] For plug-and-play configuration, the terminal comes pre-installed with a time-frequency analysis and diagnostic algorithm. During initial installation, maintenance personnel configure the following parameters via the terminal's human-machine interface (e.g., touchscreen): number of channels, sensor range, signal type (4-20mA or 0-10V), branch number, and sampling rate. The sampling rate can be configured to 100Hz, 200Hz, 500Hz, or 1000Hz. 100Hz is suitable for routine monitoring, balancing power consumption and accuracy; 200Hz or 500Hz are suitable for scenarios requiring higher frequency resolution, such as precise leak diagnosis; and 1000Hz is suitable for accurate analysis of high-frequency leak characteristics. After configuration, the terminal automatically enters learning mode to establish a health baseline, and automatically switches to diagnostic mode after learning is complete.

[0110] Step S203: Trigger the edge computing terminal to enter learning mode and collect running data to establish a health baseline.

[0111] In one alternative implementation, the health baseline includes a pulsation amplitude benchmark, a mean pressure benchmark, a frequency band energy ratio benchmark, a high-frequency energy ratio benchmark, and a sample entropy benchmark.

[0112] This step involves using external operations (such as performing a confirmation operation or sending a specific command on the terminal's human-computer interaction interface) to bring the terminal from standby or configuration state into learning mode.

[0113] Learning mode is a working state of an edge computing terminal. In this state, the terminal does not perform health diagnosis, but continuously collects pressure signals from each channel and extracts feature parameters to establish a comparison benchmark for subsequent diagnosis.

[0114] Operational data refers to the pressure pulsation signals collected by the terminal under normal operating conditions of the aeration system and the characteristic parameters extracted after time-frequency analysis. The system is required to be in a fault-free normal operating state during the collection period.

[0115] A health baseline is a benchmark dataset that characterizes the normal levels of various characteristic parameters in the healthy state of an aeration system. The terminal continuously collects operational data for a sufficient duration in learning mode, and after processing, forms the health baseline for each channel. This baseline serves as a comparison basis for determining whether a system malfunction has occurred and assessing the severity of the malfunction in subsequent diagnostic modes.

[0116] The learning mode setting enables the terminal to adapt to different working conditions and different field environments, avoiding the problem of mismatch between factory preset parameters and actual working conditions, and realizing the terminal's plug-and-play capability and rapid on-site deployment.

[0117] Step S204: After the health baseline is established, the edge computing terminal is triggered to switch to diagnostic mode and output the diagnostic results.

[0118] After the learning completion conditions are met, this step switches the terminal from learning mode to diagnostic mode through external operation (such as confirmation on the terminal's human-computer interaction interface). Diagnostic mode is a working state of the edge computing terminal. In this state, the terminal uses a health baseline as a comparison benchmark to perform time-frequency analysis and feature extraction on the real-time collected pressure signals, and performs health diagnosis based on the degree of deviation of the feature parameters from the baseline, outputting the diagnostic results.

[0119] Outputting diagnostic results refers to the terminal displaying diagnostic conclusions to the user through the human-machine interaction module and / or sending them to external systems (such as distributed control systems, central control room monitoring systems, or cloud-based operation and maintenance platforms) through the communication module to achieve fault alarms, data recording, and remote monitoring.

[0120] The diagnostic mode is the normal operating state of the terminal. After learning is completed, it will run in this mode for a long time to achieve continuous real-time monitoring of the health status of the aeration system.

[0121] The deployment method of the edge computing terminal for health diagnosis of aeration systems provided in this embodiment connects the edge computing terminal in parallel to the signal line of the pressure sensor through a high-impedance input method. This allows the terminal to only listen to the signal without injecting any excitation into the line, and there is no need to cut off or modify the original signal line. Even if the terminal itself fails or loses power, it will not affect the signal transmission and normal operation of the original control system. After configuring the channel parameters and sensor range, the terminal is triggered to enter the learning mode, collect operating data to establish a health baseline, and automatically switch to the diagnostic mode after the health baseline is established. The entire process does not require manual calibration or parameter tuning, realizing a "plug and play" deployment method. This significantly reduces the professional skills required for on-site deployment of operation and maintenance personnel, enabling on-site personnel with a process background to independently complete the installation and commissioning of the terminal.

[0122] In this embodiment, the workflow of the edge computing terminal after deployment includes the following steps.

[0123] Step 1: Power-on self-test and initialization.

[0124] After the terminal is powered on, it performs a hardware self-test. The hardware self-test includes ADC communication testing, storage read / write testing, network connectivity testing, and status checks of each module. If the self-test passes, the terminal enters the working state. If the self-test fails, the touchscreen displays fault information, and the status indicator lights flash as an alarm.

[0125] Step 2: Learning the health baseline.

[0126] Upon initial installation or manual triggering of baseline learning, the terminal enters learning mode, continuously collecting normal operation data and automatically calculating the baseline characteristics of each channel. The data collection period for normal operation must be no less than 24 hours, during which the aeration system must be in a fault-free, normal operating condition. The baseline characteristics include pulsation amplitude baseline, average pressure baseline, energy percentage baseline for each frequency band, high-frequency energy ratio baseline, and sample entropy baseline. After learning is complete, the terminal automatically switches to diagnostic mode.

[0127] Step 3: Real-time data acquisition and diagnosis.

[0128] The terminal continuously acquires pressure signals from each channel at a preset sampling rate and performs the following calculations in units of a sliding window: Signal preprocessing includes digital filtering and baseline removal. Time-frequency transformation, which uses Fast Fourier Transform; Feature extraction includes pulsation amplitude, average pressure, high-frequency energy ratio, sample entropy, harmonic distortion rate, and low-frequency energy ratio. Fault determination is based on the combined changes in pulsation amplitude, average pressure, main frequency offset, high-frequency energy ratio, and sample entropy, and outputs whether it is blocked, leaking, or normal according to the built-in judgment matrix. Severity calculation: Automatically calculate the blockage degree index and the leakage severity index, and output the corresponding level as normal, slight, moderate or severe. The blockage degree index reflects the comprehensive deviation of the energy proportion of each frequency band relative to the healthy baseline, and the leakage severity index reflects the comprehensive degree of high-frequency energy increase, pulsation amplitude attenuation and sample entropy increase. Subtype identification distinguishes between microbial film blockage and inorganic scaling blockage by low-frequency energy ratio, and distinguishes between flange leakage, pipeline cracks and membrane damage by high-frequency energy spectrum distribution morphology.

[0129] In one optional implementation, the preset sampling rate is 100Hz. The sliding window has a window length of 0.5 seconds and a sliding step size of 0.25 seconds.

[0130] Step 4: Output and Display of Results.

[0131] Real-time diagnostic results are output in the following ways: The touchscreen display shows the health status of each branch pipe, the values ​​and trend curves of the blockage degree index and the air leakage severity index, the text description of the diagnostic conclusions and maintenance suggestions. The health status of each branch pipe is marked with green, yellow or red to correspond to normal, warning and alarm respectively. Digital output; alarm signals are output via relay dry contacts. The communication interface outputs diagnostic data to the host system via Modbus TCP, Modbus RTU, OPC UA, or MQTT protocols. Upload to the cloud: Diagnostic data is uploaded to the cloud-based operation and maintenance platform via 4G / 5G.

[0132] Step 5: Data storage and trend analysis.

[0133] The terminal continuously stores raw pressure pulsation signals, real-time characteristic parameters, diagnostic results, and alarm events. The storage duration of the raw pressure pulsation signals is configured. The terminal automatically calculates the rate of change of the blockage degree index and the leakage severity index daily. When the rate of change exceeds a preset threshold, an early warning is issued, and the time when the severity threshold is predicted is provided.

[0134] Step Six: Remote Operation and Maintenance and Firmware Upgrade.

[0135] The terminal supports remote access via a network interface, allowing maintenance personnel to remotely view diagnostic data, export historical records, and modify configuration parameters. Remote access requires authentication configuration. The terminal supports remote firmware upgrades, downloading upgrade packages via Ethernet or 4G / 5G networks and automatically completing verification and installation.

[0136] The present invention has the following advantages compared with the prior art.

[0137] First, an edge computing terminal hardware architecture specifically designed for health diagnosis of aeration systems. This invention is specifically designed for the fault mechanisms (clogging and leakage) and application environments (high humidity and corrosive gases in wastewater treatment plants) of aeration systems. It is an embedded hardware platform that integrates multi-channel signal input (compatible with 4-20mA / 0-10V), signal conditioning (anti-aliasing filtering, programmable gain amplification, high-precision ADC), edge computing (embedded processor + FPU + DSP instruction set), local storage, multi-protocol communication (Modbus / OPC UA / MQTT), and human-computer interaction (touchscreen + buttons + indicator lights). It is dedicated to diagnosing clogging and leakage in aeration systems, filling a product gap in this niche market.

[0138] Secondly, the deployment method bypasses the original control system. The terminal connects in parallel to the existing pressure sensor signal line via a high-impedance input method, only reading the signal without injecting any excitation into the line or modifying the original signal value. Even if the terminal itself fails, it will not affect the normal operation of the original control system and the gas supply control function. The initial installation only requires configuring the channel parameters and sensor range. The terminal automatically completes health baseline learning and diagnostic mode switching. Ordinary maintenance personnel can complete the deployment of one terminal within 15 minutes, achieving a convenient plug-and-play deployment method.

[0139] Third, the embedded implementation method of solidified time-frequency analysis and diagnostic algorithms. This invention solidifies time-frequency analysis algorithms such as Fast Fourier Transform and sample entropy into resident algorithms of embedded terminals. Through the hybrid basis FFT algorithm and optimization measures such as single instruction multiple data stream instruction set acceleration, sliding window data reuse (adjacent window sliding step size is less than window length), and sample entropy approximate nearest neighbor search, it achieves real-time operation on the embedded platform, outputting the blockage degree index, leakage severity index, and fault type identification results. Maintenance personnel do not need to understand the algorithm details, but only need to view the diagnostic conclusions and maintenance suggestions, realizing algorithm solidification and professional and reliable diagnostic functions.

[0140] Fourth, the automatic health baseline learning and self-configuration method. After new installation or maintenance, the terminal automatically enters learning mode, continuously collecting 24 hours of normal operation data, and automatically calculates the characteristic parameter baselines for each channel (including pulsation amplitude baseline, average pressure baseline, energy ratio baseline for each frequency band, high-frequency energy ratio baseline, and sample entropy baseline). Upon completion, it automatically switches to diagnostic mode. Maintenance personnel only need to configure channel parameters and sensor ranges; manual calibration or parameter tuning is unnecessary, achieving automatic terminal deployment.

[0141] Fifth, the terminal integrates multiple functions and provides visualized multi-dimensional diagnostic results. It combines signal acquisition, data processing, fault diagnosis, result display, data storage, remote communication, and alarm output into one unit, requiring no additional equipment. A single terminal can simultaneously monitor up to 32 aeration branch pipes. The terminal uses a touchscreen to graphically display the health status of each branch pipe (green, yellow, or red indicators correspond to normal, warning, and alarm, respectively), the blockage index, and the leakage severity index, along with historical trend curves, textual descriptions of diagnostic conclusions, and maintenance recommendations. Maintenance personnel can intuitively obtain diagnostic information without needing to interpret raw data.

[0142] Sixth, industrial-grade design, adaptable to harsh environments. The terminal uses industrial-grade components and an IP65 protective shell, and operates in a wide temperature range (-20℃~70℃), adapting to the high humidity, corrosive gases, and large temperature fluctuations of the wastewater treatment plant environment, ensuring long-term reliable operation of the terminal under the harsh conditions of the wastewater treatment plant.

[0143] Seventh, open interconnectivity and predictive maintenance capabilities. The terminal supports at least one industrial communication protocol among Modbus RTU, Modbus TCP, OPC UA, and MQTT, allowing for easy integration of diagnostic data into existing distributed control systems, central control room monitoring systems, or cloud-based maintenance platforms. The terminal continuously calculates the rate of change of the blockage degree index and the leakage severity index, issuing an early warning when the rate exceeds a preset threshold and predicting the time to reach the severe threshold, thus upgrading from "reactive maintenance" to "predictive maintenance."

[0144] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0145] The following is a detailed reference. Figure 4 This diagram illustrates a structural schematic suitable for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 401, which can perform various appropriate actions and processes according to a program stored in read-only memory (ROM) 402 or a program loaded from memory 408 into random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the electronic device. The processor 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0146] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4 Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0147] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 409, or installed from a memory 408, or installed from a ROM 402. When the computer program is executed by the processor 401, it performs the functions defined in the deployment method of the edge computing terminal for health diagnosis of an aeration system according to embodiments of the present invention.

[0148] Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0149] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the deployment method of the edge computing terminal for health diagnosis of an aeration system shown in the above embodiments is implemented.

[0150] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0151] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and all such modifications and variations fall within the scope defined by the appended claims.

Claims

1. An edge computing terminal for health diagnosis of an aeration system, characterized in that, include: The signal input module has an input terminal for connecting to a pressure sensor to receive the pressure sensor signal. The signal input module is equipped with a signal conditioning circuit for processing the pressure sensor signal and outputting a digital signal. An edge computing module, connected to the signal input module, is used to receive the digital signal. The edge computing module is equipped with an embedded processor and a memory. The memory stores a time-frequency analysis and diagnostic algorithm program. The embedded processor is used to execute the time-frequency analysis and diagnostic algorithm program to perform time-frequency analysis on the digital signal to extract feature parameters, and perform health diagnosis based on the feature parameters to generate a diagnostic result. The human-computer interaction module is connected to the edge computing module and is used to output the diagnostic results to the user and receive user operation instructions. A communication module, connected to the edge computing module, is used to output the diagnostic results to an external system; The edge computing module has a learning mode and a diagnostic mode. In the learning mode, it collects running data to establish a health baseline for diagnostic comparison, and automatically switches to the diagnostic mode after the baseline is established to perform the health diagnosis.

2. The edge computing terminal for health diagnosis of an aeration system according to claim 1, characterized in that, The signal conditioning circuit includes an anti-aliasing low-pass filter, a programmable gain amplifier, and an analog-to-digital converter; the anti-aliasing low-pass filter is used to perform anti-aliasing filtering on the pressure sensor signal, the programmable gain amplifier is used to amplify the filtered signal, and the analog-to-digital converter is used to convert the amplified analog signal into a digital signal. The input terminal of the anti-aliasing low-pass filter is connected to the pressure sensor signal, and the output terminal is connected to the input terminal of the programmable gain amplifier. The output terminal of the programmable gain amplifier is connected to the input terminal of the analog-to-digital converter, and the analog-to-digital converter outputs the digital signal.

3. The edge computing terminal for health diagnosis of an aeration system according to claim 1, characterized in that, When the embedded processor executes the time-frequency analysis and diagnostic algorithm program, it includes at least one of the following optimization methods: The Fast Fourier Transform employs a mixed-basis algorithm and utilizes a single-instruction multiple-data instruction set; The edge computing module uses a sliding window to process the digital signal, and the sliding step size of adjacent windows is smaller than the window length. The sample entropy is calculated using an approximate nearest neighbor search algorithm.

4. The edge computing terminal for health diagnosis of an aeration system according to claim 1, characterized in that, The health diagnosis includes calculating a blockage index, which is used to characterize the severity of aerator blockage; The congestion index is calculated by comparing the energy proportion of each frequency band in real time with the energy proportion of each frequency band in the healthy baseline and calculating the weighted comprehensive deviation value, and then multiplying it by the amplification factor of the real-time pulsation amplitude relative to the healthy baseline.

5. The edge computing terminal for health diagnosis of an aeration system according to claim 1, characterized in that, The health diagnosis includes calculating a leakage severity index, which is used to characterize the severity of pipeline leakage; The air leakage severity index is calculated by weighting the increase of real-time high-frequency energy ratio relative to the healthy baseline, the decrease of real-time pulsation amplitude relative to the healthy baseline, and the increase of real-time sample entropy relative to the healthy baseline according to preset weights.

6. The edge computing terminal for health diagnosis of an aeration system according to claim 1, characterized in that, The terminal has a protection rating of no less than IP65, an operating temperature range of -20℃ to 70℃, and is equipped with a DIN35 rail mounting structure.

7. The edge computing terminal for health diagnosis of an aeration system according to claim 1, characterized in that, The communication module supports at least one of the following protocols: Modbus RTU, Modbus TCP, OPC UA, and MQTT, and supports uploading to a remote server via 4G / 5G.

8. A method for deploying an edge computing terminal for health diagnosis of an aeration system, wherein the edge computing terminal is the terminal described in claim 1, characterized in that, Includes the following steps: The edge computing terminal is connected in parallel to the signal line of the pressure sensor via a high-impedance input method. Configure channel parameters and sensor ranges for the edge computing terminal; The edge computing terminal is triggered to enter learning mode to collect operational data and establish a health baseline. Once the health baseline is established, the edge computing terminal is triggered to switch to diagnostic mode and output the diagnostic results.

9. The method according to claim 8, characterized in that, The input impedance of the high-impedance input method is not less than 1MΩ, and the current drawn by the edge computing terminal from the signal source does not exceed 50μA.

10. The method according to claim 8, characterized in that, The health baseline includes the pulsation amplitude baseline, the mean pressure baseline, the energy ratio of each frequency band baseline, the high-frequency energy ratio baseline, and the sample entropy baseline.

Images