Method for progress and quality collaborative management of power engineering project

Abstract

The application provides a power engineering project progress and quality collaborative management method, relates to the field of information technology, and comprises the following steps: if the power quality health state of a power distribution network is abnormal, alarm information and an intelligent diagnosis report are generated; according to the intelligent diagnosis report, the type and severity of the power quality problem are determined, corresponding treatment strategies and control instructions are generated, and the control instructions are transmitted to corresponding power quality treatment equipment through an SDN communication network; after receiving the control instructions, the power quality treatment equipment adjusts corresponding operation parameters, dynamically optimizes various power quality treatment measures for reactive power compensation and harmonic treatment, and meanwhile, power quality monitoring equipment senses the change of the operation parameters of the treatment equipment in real time, evaluates the collaborative treatment effect, forms a closed-loop feedback control, and realizes the collaborative optimization of the monitoring equipment and the treatment equipment. The application realizes intelligent monitoring, accurate diagnosis and collaborative treatment of power quality, effectively improves the power quality level of the power distribution network, and improves the use rate of power.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for collaborative management of the progress and quality of power engineering projects. Background Technology

[0002] The causes and mechanisms of power quality problems are complex and diverse, involving all aspects of the power system, including generation, transmission, distribution, and consumption. In practical engineering, the massive amounts of data collected by power quality monitoring equipment need to be transmitted in real time to the mitigation equipment so that the mitigation equipment can automatically adjust its operating parameters based on the monitoring data. However, due to the complexity and variability of the power system environment, two-way communication between monitoring and mitigation equipment faces many technical challenges. Numerous electromagnetic interference sources exist in the power system, such as power electronic equipment and electric arc furnaces, which can severely interfere with and distort the communication signals between monitoring and mitigation equipment, leading to degraded or even interrupted communication quality. Furthermore, the monitoring and mitigation equipment are distributed in different locations within the power grid, with long communication distances and significant communication delays, making real-time data transmission and control command issuance difficult. Summary of the Invention

[0003] To address the aforementioned technical problems, this invention provides a method for collaborative management of the progress and quality of power engineering projects, mainly including:

[0004] Electromagnetic interference (EMI) source characteristics of different types of EMI sources are acquired to form EMI source characteristic templates. Power quality data is collected through power quality monitoring equipment, and the collected power quality data is compressed by extracting key feature parameters through edge computing nodes to obtain compressed power quality characteristic data. The power quality characteristic data is matched and analyzed with the EMI source characteristic templates. Components in the power quality characteristic data similar to EMI source characteristics are marked as interfered power quality monitoring data. The marked interfered power quality monitoring data is processed using adaptive inter-spectral interference suppression filtering to obtain purified power quality monitoring data. Based on the purified power quality monitoring data, the power quality health status of the distribution network is determined. If the power quality health status of the distribution network is abnormal, alarm information and intelligent diagnostic reports are generated. Based on the intelligent diagnostic reports, corresponding governance strategies and control commands are generated, and the control commands are transmitted to the corresponding power quality governance equipment. After receiving the control commands, the power quality governance equipment adjusts its operating parameters accordingly. Simultaneously, the power quality monitoring equipment senses changes in the operating parameters of the power quality governance equipment in real time, evaluates the collaborative governance effect, and forms a closed-loop feedback control.

[0005] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects:

[0006] This invention discloses a method for collaborative management of progress and quality in power engineering projects. The method first establishes an electromagnetic interference source feature template. Power monitoring data is compressed in real time using edge computing nodes, and features are extracted and matched with the interference source template to identify interfered data. The interfered data is processed using adaptive inter-spectral interference suppression filtering, and key indicators are analyzed to determine the power quality health status. If the status is abnormal, a diagnostic report and mitigation strategy are generated, and control commands are sent to the mitigation equipment via an SDN network. The mitigation equipment dynamically adjusts parameters to optimize mitigation measures, and the monitoring equipment evaluates the effects in real time, forming a closed-loop feedback. This invention also utilizes virtual synchronous generator technology, enabling the mitigation equipment to simulate synchronous machine characteristics and participate in grid regulation. Two-way communication between the monitoring and mitigation equipment is achieved through a preset protocol, ensuring stable data transmission. This method realizes intelligent monitoring, accurate diagnosis, and collaborative mitigation of power quality, effectively improving the power quality level of the distribution network and increasing power utilization. Attached Figure Description

[0007] Figure 1 This is a flowchart of the method for collaborative management of progress and quality in power engineering projects according to the present invention. Detailed Implementation

[0008] To further understand the content of this invention, a detailed description of the invention is provided in conjunction with the accompanying drawings and embodiments. The specific embodiments described herein are for illustrative purposes only and are not intended to limit the invention. It should also be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0009] like Figure 1 The method for collaborative management of progress and quality in power engineering projects in this embodiment may specifically include:

[0010] S101. Obtain electromagnetic interference source characteristics of different types of electromagnetic interference sources, form electromagnetic interference source characteristic templates, and collect power quality data through power quality monitoring equipment. Extract and compress key feature parameters of the collected power quality data through edge computing nodes to obtain compressed power quality feature data.

[0011] The spectral and amplitude characteristics of different electromagnetic interference (EMI) sources are acquired to establish an EMI source feature database, which contains pre-stored EMI source feature templates. Distributed power quality monitoring equipment is used to collect power quality data in real time, including voltage, current, and harmonic parameters. If power quality parameters are received, data integrity verification is performed on the power quality parameters through the edge computing nodes. Discrete wavelet transform is applied to the verified power quality data to obtain the multi-scale decomposition feature parameters. Based on the feature parameters, the energy distribution and statistical characteristics are calculated. The support vector machine (SVM) algorithm is used to perform similarity matching between the compressed power quality feature vector and the EMI source feature templates. The feature vector is transformed to a high-dimensional space through kernel function mapping, and the distance between the feature vector and various interference source templates is calculated. The category with the smallest distance is selected as the identification result to determine the main EMI source types in the power system. Based on the identification result and the feature vector amplitude, the influence degree of the interference source is determined.

[0012] For example, an electromagnetic interference (EMI) source characteristic database is established based on the EMI sources present in the power system environment, containing the spectral and amplitude characteristics of various EMI sources. Using field measurements and historical data analysis, the electromagnetic field strength, frequency range, and waveform characteristics of different types of EMI sources are collected. The Fast Fourier Transform (FFT) algorithm is used to perform spectral analysis on the collected data to extract characteristic frequencies and amplitudes. The extracted characteristic parameters are stored in the database to form EMI source characteristic templates.

[0013] Distributed power quality monitoring equipment is used to collect power quality parameters such as voltage, current, and harmonics in real time. The raw power quality data is transmitted to edge computing nodes deployed in the substation via industrial Ethernet or fiber optic communication networks. Upon receiving the data, the edge computing nodes immediately perform data integrity verification, eliminating outliers and redundant data.

[0014] Edge computing nodes apply discrete wavelet transform to multi-scale decomposition of the verified power quality data to extract feature parameters for different frequency bands. The energy distribution and statistical characteristics of each frequency band, such as mean, variance, and kurtosis, are calculated. The feature parameter matrix is ​​then compressed using singular value decomposition, retaining the main feature vectors. Combining time-domain and frequency-domain features, a compressed power quality feature vector is generated.

[0015] A support vector machine (SVM) algorithm is used to perform similarity matching between compressed power quality feature vectors and pre-stored electromagnetic interference (EMI) source feature templates. The feature vectors are transformed into a high-dimensional space using kernel function mapping, and the distance between the feature vectors and various EMI source templates is calculated. The category with the smallest distance is selected as the identification result to determine the main EMI source types present in the current power system. Based on the identification results and feature vector amplitudes, the impact degree of the EMI source is determined. Using data from multiple monitoring points, the approximate location of the EMI source is calculated using triangulation, thus achieving EMI source localization.

[0016] In establishing the electromagnetic interference (EMI) source characteristic database, a combination of on-site measurement and historical data analysis was adopted. Taking EMI caused by switching operations as an example, data was collected at multiple locations within the substation using a portable electromagnetic field strength meter, with a measurement frequency range of 1 kHz to 1 MHz and a sampling rate of 10 MS / s. The collected time-domain data was analyzed using a Fast Fourier Transform (FFT) algorithm to extract characteristic frequencies and amplitudes. The results showed that EMI generated by switching operations was mainly concentrated in the 10 kHz to 100 kHz frequency band, with a peak amplitude of approximately 80 dB μV / m. These characteristic parameters were stored in the database to form a characteristic template for EMI sources caused by switching operations.

[0017] The power quality monitoring equipment employs high-precision voltage and current transformers with a sampling rate of 25.6kHz and a resolution of 16 bits. The raw power quality data is transmitted to edge computing nodes via industrial Ethernet at a transmission rate of 100Mbps. Upon receiving the data, the edge computing nodes use a checksum algorithm to verify data integrity, eliminating outliers and redundant data. For example, if a voltage surge exceeding 30% of the rated value is detected, the data is marked as abnormal and discarded.

[0018] The verified power quality data was subjected to multi-scale decomposition using a 5-level discrete wavelet transform, with Daubechies 4 wavelets as the basis functions. Energy distribution and statistical characteristics, such as mean, variance, and kurtosis, were calculated for each frequency band. The feature parameter matrix was then compressed using singular value decomposition, retaining the first 10 principal feature vectors. Combining time-domain and frequency-domain features, a compressed 50-dimensional power quality feature vector was generated.

[0019] A support vector machine algorithm with radial basis function kernels is used to perform similarity matching between the compressed power quality feature vector and pre-stored electromagnetic interference source feature templates. The kernel function parameter γ is set to 0.1, and the penalty factor C is set to 10. The distance between the feature vector and each type of interference source template is calculated, and the category with the smallest distance is selected as the identification result. If it is identified as harmonic interference and the total harmonic distortion rate exceeds 5%, it is judged as a serious impact. Using data from three monitoring points, the approximate location of the interference source is calculated using triangulation, with the error range controlled within 10 meters, thus achieving the localization of the interference source.

[0020] S102. The step of matching and analyzing the power quality characteristic data with the electromagnetic interference source characteristic template, and marking the components in the power quality characteristic data that are similar to the characteristics of the electromagnetic interference source as interfered power quality monitoring data. Specifically, the compressed power quality characteristic data is transmitted to the data analysis center; the data analysis center matches and analyzes the received power quality characteristic data with the pre-established electromagnetic interference source characteristic template to determine whether there are components in the power quality characteristic data that are similar to the characteristics of the electromagnetic interference source; if similar components are found, the monitoring data is determined to be subject to electromagnetic interference and is marked as interfered power quality monitoring data.

[0021] The system receives compressed power quality characteristic data sent by the smart meter, which is transmitted to the data analysis center via TCP / IP protocol. Based on the compressed power quality characteristic data, a Huffman decoding algorithm is used to decompress the data, obtaining the original power quality characteristic data. A pre-established electromagnetic interference source feature template library is acquired, and a cosine similarity algorithm is used to calculate the similarity between the original power quality characteristic data and each sample in the feature template library. If the similarity exceeds a preset threshold, the original power quality characteristic data is determined to be subject to electromagnetic interference, and it is marked as interfered power quality monitoring data using a bitmap labeling method. For the interfered power quality monitoring data, a decision tree algorithm is used to identify the type of interference source, and an interference severity index is calculated based on the similarity value and interference characteristic parameters, classifying the interference severity into three levels: mild, moderate, and severe.

[0022] For example, compressed power quality feature data is transmitted to the data analysis center via TCP / IP protocol, and a data checksum algorithm is used to ensure the reliability and integrity of the data transmission. At the data receiving end, a Huffman decoding algorithm is used to decompress the data and restore the original power quality feature data. The decompressed data is then verified for integrity using a cyclic redundancy check algorithm to eliminate any erroneous data that may have been generated during transmission. The data analysis center uses a pre-established electromagnetic interference source feature template library to perform feature extraction and matching analysis on the received power quality feature data. A cosine similarity algorithm is used to calculate the similarity between the power quality feature data and each sample in the feature template library. The cosine similarity calculation formula is: cos(θ)=(A·B) / (||A||·||B||), where A and B represent the power quality feature data vector and the feature template vector, respectively, and ||A|| and ||B|| represent the Euclidean norm of the vectors. Based on the similarity calculation results, a similarity threshold is set to determine whether there are components in the power quality feature data similar to electromagnetic interference source features. Thresholds were determined through statistical analysis of historical data. Specifically, the similarity distribution between known interfered samples and normal samples was calculated, and the intersection of the two distributions was selected as the initial threshold. If the similarity exceeded the preset threshold, the monitoring data was determined to be subject to electromagnetic interference, and it was marked as interfered power quality monitoring data using a bitmap labeling method. Further analysis was performed on the marked interfered data, employing a decision tree algorithm to identify the type of interference source. A decision tree model was constructed based on parameters such as the spectral characteristics, amplitude changes, and duration of the power quality characteristic data. By traversing the decision tree nodes, the type of interference source was determined, such as harmonic interference, transient interference, or flicker interference. Based on the similarity value and interference characteristic parameters, an interference severity index was calculated, classifying the interference severity into three levels: mild, moderate, and severe. The analysis results were added to the labeled data. During data transmission, the compressed power quality characteristic data was transmitted to the data analysis center at a rate of 1 Mbps using the TCP / IP protocol. The data packet size was set to 1460 bytes, and flow control was implemented using a sliding window mechanism, with the window size dynamically adjusted; the initial value was set to 65535 bytes. Data verification is performed using a 32-bit Cyclic Redundancy Check (CRC-32) algorithm, with the checksum stored in the packet header. Upon receiving data, the receiving end first performs a CRC-32 check; if the check fails, a retransmission is requested. After successful check, the data is decompressed using a Huffman decoding algorithm at a decompression rate of 10MB / s. The decompressed data is then processed using an MD5 algorithm to generate a 128-bit checksum, which is compared with the MD5 value of the original data to ensure data integrity. The data analysis center uses a pre-established electromagnetic interference source feature template library containing feature templates for 100 common electromagnetic interference sources, each represented by a 50-dimensional feature vector. Cosine similarity algorithm is used to perform matching analysis on the received power quality feature data.During the calculation process, both power quality feature data and template data are normalized to unit vectors to eliminate the influence of amplitude differences. For each feature template, its cosine similarity with the power quality data is calculated, with a value range of [-1, 1]. The calculation speed is optimized to process 1000 samples per second. The similarity threshold is determined using historical data statistical analysis. 10,000 known samples are used, of which 5,000 are normal samples and 5,000 are interfered samples. The similarity distribution between these samples and the feature templates is calculated, and probability density function curves are plotted. The intersection point of the two distributions is found and used as the initial threshold, with an initial value set to 0.85. The system automatically updates the threshold after processing 1000 new samples to adapt to changes in data distribution. If the calculated similarity exceeds the current threshold, the monitoring data is determined to be subject to electromagnetic interference. Bitmap labeling is used to mark it as interfered data, with each bit representing the type of interference: 1 indicates the presence of that type of interference, and 0 indicates its absence. The data marked as interfered is further analyzed, and a decision tree algorithm is used to identify the type of interference source. The decision tree model contains 15 feature nodes with a maximum depth of 5, using the Gini coefficient as the node splitting criterion. Features include harmonic content, voltage fluctuation amplitude, and interference duration. The average traversal time for the decision tree is 0.5 milliseconds. Based on the decision tree output, interference sources are categorized into three main types: harmonic interference, transient interference, and flicker interference. When calculating the interference severity index, three factors are considered: similarity value, interference amplitude, and duration, using a weighted average method with weights of 0.4, 0.3, and 0.3, respectively. The index range of 0-1 is divided into three levels: 0-0.3 for mild, 0.3-0.7 for moderate, and 0.7-1 for severe. The analysis results are added to the labeled data in JSON format, including interference type, severity, and key parameters.

[0023] S103. The power quality monitoring data marked as interfered is processed using adaptive inter-spectral interference suppression filtering to obtain purified power quality monitoring data. The power quality health status of the distribution network is then determined based on the purified power quality monitoring data. Specifically, the power quality monitoring data marked as interfered is processed using adaptive inter-spectral interference suppression filtering, including adaptively adjusting filter parameters according to the spectral characteristics of the electromagnetic interference source, selectively filtering out interfered frequency bands, and outputting purified power quality monitoring data. The purified power quality monitoring data is then transmitted to the distribution automation master station system, and key indicators of voltage sag and harmonic distortion are analyzed to determine the power quality health status of the distribution network.

[0024] Spectral analysis is performed on the interfered power quality monitoring data, and the power spectral density of the monitoring data is calculated using the Fast Fourier Transform algorithm. Based on the power spectral density, the energy distribution characteristics of the interfered frequency band are identified, and the center frequency and bandwidth information of the interfered frequency band are obtained. The normalized minimum mean square error algorithm is used to dynamically adjust the adaptive filter parameters, where the adaptive filter parameters include filter coefficients. If the filter coefficients match the characteristics of the interfered frequency band, the purified power quality monitoring data is output. The purified power quality monitoring data is transmitted to the distribution automation master station system via a transport layer security protocol. The purified data received by the distribution automation master station system is decrypted and its integrity is verified. If the verification passes, the purified data is stored in a real-time database. The power quality data stored in the real-time database is retrieved, and spectral analysis is performed on the power quality data. Key indicator parameters are extracted from the spectral analysis results, including the duration and amplitude of voltage sags and harmonic content. A power quality health index is calculated based on the key indicator parameters to determine the power quality health status of the distribution network.

[0025] For example, based on the spectral characteristics of the electromagnetic interference source, a Fast Fourier Transform (FFT) algorithm is used to perform spectral analysis on the interfered power quality monitoring data, calculate the power spectral density of the signal, identify the energy distribution characteristics of the interfered frequency band, and obtain the center frequency and bandwidth information of the interference frequency band. The frequency resolution is set to 1 Hz, the number of sampling points to 8192, and the Hanning window function is used to reduce spectral leakage. The interference frequency is located using a peak detection algorithm, and the detection threshold is set to three times the average power spectral density. The Normalized Minimum Mean Square Error (NLMS) algorithm is used to dynamically adjust the adaptive filter parameters, initializing the filter coefficients as a zero vector with a step size μ of 0.1. The filter coefficients are continuously optimized through iterative calculations to match the frequency response characteristics of the filter with the characteristics of the interference frequency band. The iteration stopping condition is set to a mean square error less than a preset threshold of 10^-6 or reaching the maximum number of iterations of 1000. This achieves precise suppression of the interference signal and outputs purified power quality monitoring data. Transport Layer Security (TLS) version 1.3 is used to transmit the purified data to the distribution automation master station system to ensure data transmission security. In the distribution automation master station system, the received cleaned data undergoes TLS decryption and integrity verification. After successful verification, the data is stored in a real-time database. The database uses the time-series database InfluxDB, with a data retention policy of 30 days and a compressed storage format to reduce storage space usage. The stored power quality data is analyzed using a Fast Fourier Transform (FFT) algorithm to extract key parameters such as the duration and amplitude of voltage sags and harmonic content. Voltage sags are judged when their amplitude is below 90% of the rated voltage and their duration is between 0.5 and 30 cycles. Harmonic content calculation considers harmonics from the 2nd to the 50th order and uses Total Harmonic Distortion (THD) as the metric. A weighted average method is used to calculate the power quality health index, with the following weights: 40% for voltage sag, 40% for harmonic distortion, and 20% for voltage deviation. A health index threshold is set: greater than 90 indicates healthy, 70-90 indicates sub-healthy, and less than 70 indicates unhealthy, thus determining the power quality health status of the distribution network. In practical applications, power quality monitoring equipment collects data at a sampling rate of 10kHz, generating 10,000 data points per second. The Fast Fourier Transform algorithm uses 8192 points for calculation, achieving a frequency resolution of 1.22Hz. For a 50Hz power frequency system, it can accurately analyze up to the 100th harmonic. The Hanning window function has a main lobe width of 2.67 frequency intervals, effectively reducing spectral leakage. The peak detection algorithm searches for interference frequencies in the 5-2500Hz range; typical switching power supply interference at around 20kHz is accurately identified, with a power spectral density of -40dB / Hz, far exceeding the background noise level of -60dB / Hz. The normalized minimum mean square error algorithm uses an initial step size of 0.1, which gradually decreases during iteration, stopping when the mean square error drops to 10^-6 or reaches 837 iterations, with an average convergence time of 0.5 seconds.The total harmonic distortion (THD) of the filtered signal was reduced from 4.7% to 1.2%, meeting the national power quality standards. The transport layer security protocol uses the ECDHE_RSA_WITH_AES_256_GCM_SHA384 encryption suite, and key exchange employs the elliptic curve Diffie-Hellman algorithm for forward security. In the InfluxDB database, voltage data points are stored at 50ms intervals, achieving a compression ratio of 10:1; 30 days of data occupy approximately 500MB of storage space. Voltage sag analysis detected an event lasting 22 cycles with an amplitude drop to 82% of the rated voltage. Harmonic analysis showed that the 5th harmonic had the highest content, at 2.8% of the fundamental frequency. The power quality health index was calculated to be 85 points, indicating a sub-healthy state, primarily affected by harmonic distortion. An alarm message was automatically generated, recommending checking nonlinear loads and considering the installation of passive filters.

[0026] S104. If the power quality health status of the distribution network is abnormal, an alarm message and an intelligent diagnostic report are generated. Based on the intelligent diagnostic report, corresponding governance strategies and control commands are generated, and the control commands are transmitted to the corresponding power quality governance equipment. Specifically, based on the intelligent diagnostic report, the type and severity of the power quality problem are determined, and corresponding governance strategies and control commands are generated; the control commands are transmitted to the corresponding power quality governance equipment through the SDN (Software Defined Network) communication network.

[0027] The system acquires power quality health status indicators for the distribution network and uses a comprehensive evaluation method to determine whether power quality is abnormal. If the health index falls below a preset threshold, an alarm mechanism is triggered. Based on the alarm mechanism, a random forest algorithm is used to analyze the power quality anomaly data to determine the specific type of power quality problem. For each specific type of power quality problem, a corresponding governance strategy is retrieved from a preset governance strategy library, and control commands are generated. A software-defined network controller (SDN) encapsulates the control commands into standardized network data packets, which are then processed using the P4 programmable protocol. The power quality governance device receiving the standardized network data packets performs authentication and command decryption. If authentication is successful, the control operation corresponding to the control command is executed. The power quality governance device feeds back the operation results to the SDN, which updates the network status based on the operation results.

[0028] For example, based on the power quality health status indicators of the distribution network, a comprehensive evaluation method is used to determine whether any anomalies exist. If the health index is below a preset threshold of 70 points, an alarm mechanism is triggered, and an intelligent diagnostic report is generated. The alarm information includes structured data on the time, type, and severity of the anomaly. The intelligent diagnostic report includes three parts: problem description, cause analysis, and impact assessment. It analyzes the power quality anomaly data using a random forest algorithm to identify the specific type of power quality problem, such as harmonic pollution, voltage fluctuations, and three-phase imbalance. Based on the deviation of each indicator, the severity of the problem is quantitatively assessed and classified into three levels: minor, moderate, and severe. Based on the intelligent diagnostic results, a preset governance strategy library is invoked to automatically generate corresponding governance strategies and control instructions for different types and severity of power quality problems. The governance strategy library is constructed by combining expert knowledge and historical data, and the strategies are updated periodically using an incremental learning algorithm. For harmonic pollution problems, active filters are selected and their parameters are adjusted; for voltage fluctuations, reactive power compensation devices are switched on and off; for three-phase imbalance, automatic load regulators are used for balancing. The generated control commands include specific details such as device selection, parameter settings, and operation timing. A software-defined network controller encapsulates these commands into standardized network data packets. The P4 programmable protocol is used to process the data packets, enabling flexible routing strategies. The control commands are encrypted using the AES-256 encryption algorithm, and digital signature technology ensures their integrity and non-repudiation. The encrypted control commands are accurately transmitted to the corresponding power quality management devices via an SDN switch, achieving precise power quality management of the distribution network. Upon receiving the control commands, the power quality management devices first perform authentication and command decryption. After successful authentication, they execute the corresponding control operations, such as adjusting the switching capacity of reactive power compensation devices, modifying the harmonic compensation parameters of active power filters, or changing the tap level of voltage regulators. After execution, the management devices feed back the operation results to the SDN controller, forming a closed-loop control. The SDN controller updates the network status based on the feedback results and records the execution status in the power quality management database for subsequent strategy optimization and performance evaluation. In practical applications, the power quality health status indicators of the distribution network are collected every 5 minutes through a real-time monitoring system. When the health index drops to 68 points, below the preset threshold of 70 points, the system immediately triggers an alarm mechanism. The alarm information records the anomaly occurrence time as 10:15:30 on September 20, 2024, the anomaly type as harmonic pollution, and the anomaly severity as moderate. The random forest algorithm uses 100 decision trees, with 20 feature dimensions including total harmonic distortion (THD) of voltage, total harmonic distortion (THD) of current, and the content of each harmonic. The algorithm analysis results show that the 5th harmonic content reaches 6.2%, exceeding the national standard limit of 4%.The intelligent diagnostic report indicated that the problem might be caused by newly installed frequency converter equipment nearby, with an estimated impact area of ​​users in an industrial area within a 3-kilometer radius. Based on the diagnostic results, the governance strategy library automatically matched a governance solution, selecting a 100kvar active power filter for harmonic suppression. Control commands were generated, including parameters such as device ID, target harmonic order, and compensation capacity. The commands were encapsulated using the P4 protocol, encrypted using the AES-256 algorithm, with a key length of 256 bits. The SDN controller adopted the Ryu framework, configured with a shortest path first algorithm for routing. The control commands passed through a 3-hop SDN switch, with transmission latency controlled within 10ms. After receiving the commands, the active power filter completed parameter adjustments within 50ms, reducing the 5th harmonic content to 3.8%. The entire process, from anomaly detection to governance completion, took less than 1 second, achieving rapid response and precise governance of power quality issues. The execution results were fed back to the control center via the SDN network; the updated health index rose to 76 points, and the alarm status was automatically cleared.

[0029] S105. After receiving control commands, the power quality management equipment adjusts its operating parameters accordingly and dynamically optimizes various power quality management measures such as reactive power compensation and harmonic control. At the same time, the power quality monitoring equipment senses changes in the operating parameters of the management equipment in real time, evaluates the collaborative management effect, forms a closed-loop feedback control, and realizes the collaborative optimization of the monitoring equipment and the management equipment.

[0030] The system receives control commands carrying the unique identifier of the power quality monitoring equipment, which are issued by the power quality monitoring system of the power quality monitoring equipment. Based on the control commands, key parameter information is obtained through parsing by an embedded processor. Voltage and current waveform data are acquired using a high-speed data acquisition unit, with the sampling rate of the high-speed data acquisition unit set to a preset value. The waveform data is processed using a fast Fourier transform algorithm to obtain harmonic content and power factor. Based on the harmonic content and power factor, the reactive power compensation capacity and harmonic mitigation parameters are calculated using the recursive least squares method. If the reactive power compensation capacity exceeds a preset threshold, an improved fuzzy control algorithm is triggered. The improved fuzzy control algorithm uses voltage deviation rate, total harmonic distortion rate (THD), and power factor as input variables. Based on these input variables, an adjustment amount for the control equipment is determined using a pre-established fuzzy rule base. The adjustment amount is then defuzzified using the centroid method to obtain specific adjustment instructions. These instructions are used to control the reactive power compensation device, active filter, and static var generator to adjust their operating parameters. The voltage compliance rate, THD compliance rate, and average power factor are then obtained. Finally, a weighted average calculation is performed on these parameters based on preset weights to obtain the control effect evaluation result.

[0031] For example, after receiving control commands, the power quality management equipment parses the command content and extracts key parameter information through an embedded processor. Based on a preset equipment type and parameter mapping table, the commands are converted into specific equipment operation instructions. For the coordination of multiple management devices, a priority ranking mechanism is adopted, dynamically allocating management tasks according to the severity of the power quality problem and the management capacity of the equipment. Reactive power compensation devices, active power filters, and static var generators adjust their corresponding operating parameters according to their respective operation instructions. The power quality monitoring equipment acquires voltage and current waveform data in real time through a high-speed data acquisition unit, with a sampling rate set to 25.6kHz to capture harmonics up to the 50th order. The acquired raw data is filtered to remove high-frequency noise, and then a fast Fourier transform algorithm is used to calculate key indicators such as harmonic content and power factor. To process large amounts of real-time data, a sliding window method is used, performing calculations every 10 cycles, and the results are compressed and stored in a time-series database, retaining detailed data for the most recent 24 hours and statistical data for 30 days. Based on the real-time data provided by the monitoring equipment, the optimal reactive power compensation capacity and harmonic management parameters are dynamically calculated using a recursive least squares method. This algorithm rapidly adapts to changes in the power grid state by continuously updating regression coefficients. The objective function includes voltage deviation, total harmonic distortion (THD), and power factor, with weights set to 0.4, 0.4, and 0.2, respectively. The optimization results are used to adjust the operating parameters of the governance equipment in real time, such as the switching capacity of reactive power compensation devices, the harmonic compensation current of active power filters, and the output reactive power of static var generators. An improved fuzzy control algorithm is employed to achieve collaborative optimization between monitoring and governance equipment. Input variables include voltage deviation, THD, and power factor, each divided into three fuzzy sets: low, medium, and high. Output variables are the adjustment amounts of the governance equipment, divided into five fuzzy sets: significantly reduced, slightly reduced, maintained, slightly increased, and significantly increased. Fuzzy rules are designed based on expert experience, such as "if the voltage deviation is high and the power factor is low, then reactive power compensation should be significantly increased." Defuzzification is performed using the centroid method to obtain specific adjustment instructions. Simultaneously, a collaborative control mechanism is introduced, considering the complementarity of multiple governance devices, such as the coordinated operation of active power filters and reactive power compensation devices, to improve the overall governance effect. The effectiveness of power quality management is evaluated using comprehensive indicators, including voltage compliance rate, total harmonic distortion (THD) compliance rate, and average power factor, calculated through a weighted average, with weights set according to local power grid characteristics. In practical applications, after receiving control commands, the embedded processor of the power quality management system operates at a clock frequency of 100MHz, completing command parsing within 5ms. A device type and parameter mapping table is stored in a 256KB EEPROM, containing 500 mapping relationships. The priority ranking mechanism sets the weights for voltage deviation, harmonic distortion, and power factor to 0.4, 0.4, and 0.2, respectively. For example, when the THD reaches 8%, the active filter receives the highest priority.The monitoring equipment uses a 24-bit ADC, generating 5120 data points over 10 cycles at a sampling rate of 25.6kHz. A 120th-order FIR filter with a cutoff frequency of 6.4kHz is used as the digital filter. The FFT algorithm completes harmonic calculations within 50ms, accurate to the 50th harmonic. The time-series database uses InfluxDB, achieving a compression ratio of 10:1, with detailed 24-hour data occupying approximately 500MB of storage space. The forgetting factor for the recursive least squares method is set to 0.98, and the regression coefficients are updated every 100ms. The fuzzy sets for the input variables of the fuzzy control algorithm are: voltage deviation rate ([-5%, -2%], [-3%, 3%], and [2%, 5%]), total harmonic distortion rate ([0, 2%], [1%, 5%], and [4%, 8%]), and power factor ([0.85, 0.9], [0.88, 0.95], and [0.93, 1]). The adjustment ranges for the output variables are divided into five intervals: [-10%, -5%], [-7%, -2%], [-3%, 3%], [2%, 7%], and [5%, 10%]. In the collaborative control mechanism, when the harmonic distortion rate exceeds 5%, the active filter is activated first, while the switching capacity of the reactive power compensation device is reduced to avoid harmonic amplification. In the evaluation of the treatment effect, the weights for voltage compliance rate, total harmonic distortion rate compliance rate, and average power factor are 0.4, 0.4, and 0.2, respectively. A comprehensive index score greater than 90 is considered excellent, 80-90 is good, 70-80 is acceptable, and scores below 70 require further optimization.

[0032] The monitoring equipment collects power quality indicators in real time during the power quality management process, including voltage RMS value, current RMS value, voltage distortion rate, current distortion rate, voltage imbalance, current imbalance, harmonic content, harmonic current amplitude, power factor, and reactive power. The pre-set management effect evaluation function is used to evaluate the collaborative management effect. If the management effect does not meet expectations, the evaluation result is used as a feedback signal to form a closed-loop feedback control.

[0033] Raw voltage and current signal data collected by power quality monitoring equipment are acquired. After median filtering to remove outliers, the raw data is smoothed using a moving average method to obtain preprocessed data. A fast Fourier transform is performed on the preprocessed data to obtain the spectral analysis results of the voltage and current signals. Voltage distortion rate, current distortion rate, voltage imbalance, and current imbalance are calculated based on the spectral analysis results. A weighted sum of the voltage distortion rate, current distortion rate, voltage imbalance, and current imbalance is obtained using a preset governance effect evaluation function to obtain a comprehensive evaluation score. If the comprehensive evaluation score is lower than a preset threshold, the governance effect is determined to be unsatisfactory. Based on the deviation and rate of change of the comprehensive evaluation score from the target value, a fuzzy PID control algorithm is used to adjust the PID (Proportional-Integral-Derivative) control parameters. The fuzzy PID control algorithm includes fuzzification of input variables, execution of fuzzy rule inference, and defuzzification using the centroid method. Control commands for the power quality management equipment are generated based on the adjusted PID parameters; the control commands are then sent to the power quality management equipment via industrial Ethernet, which uses a star topology and Modbus TCP communication protocol.

[0034] For example, using power quality monitoring equipment, voltage and current signals are acquired in real time using synchronous sampling technology. The sampling frequency is set to 10kHz, the sampling accuracy is 24 bits, and a data frame is formed every 10 power cycles. Median filtering is applied to the acquired raw data to remove outliers, followed by smoothing using a moving average method. The effective voltage and current values ​​are calculated using the root mean square algorithm, with a calculation period of 200ms. A fast Fourier transform algorithm is used to perform spectral analysis on the preprocessed voltage and current signals, with a Hanning window selected as the window function. Voltage distortion rate, current distortion rate, voltage imbalance, and current imbalance are calculated. For harmonic content, the amplitude and phase of the 2nd to 50th harmonics are calculated, and the fundamental frequency is determined using the zero-crossing detection method. The total harmonic distortion (THD) is used to characterize the degree of harmonic pollution. Simultaneously, power factor and reactive power are calculated using vector operations, with the calculation period synchronized with the effective value calculation. Based on a preset governance effect evaluation function, the calculated power quality indicators are weighted and summed. The evaluation function employs fuzzy comprehensive evaluation, standardizing each indicator to the [0,1] interval. Weight allocation uses the analytic hierarchy process (AHP), with weights for voltage deviation, total harmonic distortion (THD), current imbalance, and power factor at 0.3, 0.3, 0.2, and 0.2, respectively. A comprehensive evaluation score is calculated, ranging from 0 to 100. If the score is below a preset threshold of 80, the treatment effect is deemed unsatisfactory, triggering an optimized control process. The evaluation results are used as feedback signals to the closed-loop controller, employing a fuzzy PID control algorithm. Input variables are the deviation *e* and the rate of change *ec* between the evaluation score and the target value; output variables are the adjustment amounts of PID parameters Kp, Ki, and Kd. The input variables are fuzzified and divided into 7 fuzzy subsets. Forty-nine fuzzy rules are formulated based on expert experience. Defuzzification is performed using the centroid method to obtain the adjustment amounts of the PID parameters. Based on the optimized PID parameters, control commands for each treatment device are generated, such as the switching capacity of the reactive power compensation device and the compensation current command for the active filter. Control commands are transmitted to each power quality management device via industrial Ethernet. The network topology adopts a star structure, and the communication protocol uses Modbus TCP to ensure the real-time performance and reliability of command transmission. In practical applications, the power quality monitoring equipment uses a 24-bit ADC for data acquisition, with a sampling frequency of 10kHz, forming a data frame containing 100 sampling points every 200ms. Outliers are removed using a 5-point median filter, followed by smoothing using a 20-point moving average. The calculated effective voltage value is 220.5V, and the effective current value is 45.2A. FFT analysis uses 2048 points with a Hanning window coefficient of 0.5. The calculated total harmonic distortion (THD) rate is 2.8% for voltage and 4.5% for current, with a voltage imbalance of 1.2% and a current imbalance of 2.1%. The 5th harmonic content is the highest, at 1.5% for voltage and 3.2% for current.The calculated power factor is 0.92, and the reactive power is 8.6 kvar. In the fuzzy comprehensive evaluation, the membership degrees of voltage deviation, harmonic distortion rate, unbalance, and power factor are 0.95, 0.85, 0.92, and 0.88, respectively. The eigenvalue of the weight matrix obtained by the analytic hierarchy process is 4.15, and the consistency ratio CR is 0.05, which is less than 0.1, satisfying the consistency requirement. The comprehensive evaluation score is 87 points, which is higher than the threshold of 80 points, so no optimization control is required. In the fuzzy PID control, the deviation e and the deviation change rate ec are both divided into seven fuzzy subsets: NB, NM, NS, ZO, PS, PM, and PB. Forty-nine fuzzy rules are used, such as "IFeisNBandecisNBTHENKpisPB,KiisNB,KdisPS". After defuzzification, the PID parameter adjustments are obtained as ΔKp=0.15, ΔKi=-0.02, and ΔKd=0.05. Control commands are transmitted via the Modbus TCP protocol, with a frame format of "01030000000AC5CD", where 01 is the slave address, 03 is the function code (read holding register), 0000 is the start address, 000A is the number of registers, and C5CD is the CRC checksum. The average network transmission delay is 15ms, and the packet loss rate is less than 0.1%, meeting the requirements for real-time control.

[0035] By utilizing virtual synchronous generators, reactive power compensation devices and active power filters are controlled as virtual devices with synchronous generator characteristics. By sensing grid frequency and voltage parameters in real time, the inertial response and damping characteristics of synchronous generators are simulated to participate in reactive power regulation and harmonic suppression of the grid. The operating status and governance effect of virtual synchronous generators are evaluated in real time through monitoring equipment, and the control parameters and strategies of virtual synchronous generators are optimized.

[0036] The system employs a phasor measurement unit to acquire grid frequency and voltage parameters. Based on these parameters, a frequency estimate is determined using zero-crossing detection and phase-locked loop (PLL) technology. Using the frequency estimate and voltage parameters, the virtual rotor angular velocity and power angle are calculated using a pre-defined second-order oscillation equation model in the virtual synchronous generator controller. Control commands for reactive power and active power are generated based on these virtual rotor angular velocity and power angle. These control commands are then sent to the reactive power compensation device and the active power filter. The reactive power compensation device adjusts the number and timing of capacitor switching according to the control commands, while the active power filter adjusts the amplitude and phase of the output current. The output voltage, current, and power data of the virtual synchronous generator are acquired through monitoring equipment. If the output voltage, current, and power data deviate from preset thresholds, the virtual inertia and damping coefficients are optimized online using a recursive least squares method.

[0037] For example, a phasor measurement unit is used to collect grid frequency and voltage parameters in real time. Frequency is estimated using zero-crossing detection combined with phase-locked loop (PLL) technology, and voltage phasor information is calculated using a discrete Fourier transform algorithm. The sampling frequency is set to 10kHz, the data frame length is 1000 points, and the PLL loop filter bandwidth is 5Hz. The collected data is transmitted to a virtual synchronous generator controller via industrial Ethernet, with a transmission delay controlled within 10ms. Based on the received frequency and voltage parameters, the virtual synchronous generator controller calculates the virtual rotor angular velocity and power angle using a preset second-order oscillation equation model. The virtual inertia constant in the model is initially set to 5s, the damping coefficient to 20, and the rated power to 1MW. Based on the calculation results, control commands for reactive and active power are generated, including power setpoints, response speed, and steady-state error limits. For frequency abrupt changes, the maximum angular acceleration limit of the virtual rotor is set to 0.1. For voltage fluctuations, the maximum voltage regulation rate is set to 0.1 pu / s. The generated control commands are sent to the reactive power compensation device and the active power filter respectively. The reactive power compensation device adjusts the number and timing of capacitor banks switching according to the commands, using a predictive control algorithm to switch them 0.5 cycles in advance to reduce operational impact. The active power filter adjusts the amplitude and phase of the output current according to the commands, with a response time of less than 5 ms. The coordinated control strategy of the two adopts a hierarchical structure: the reactive power compensation device is responsible for basic reactive power regulation, and the active power filter is responsible for rapid dynamic compensation and harmonic suppression to avoid mutual interference. Real-time monitoring equipment collects the output voltage, current, and power data of the virtual synchronous generator at a sampling rate of 20 kHz. Wavelet transform algorithm is used to perform time-frequency analysis on the collected data, using the db4 wavelet with a decomposition level of 5. The operating status and mitigation effect of the virtual synchronous generator are evaluated, including transient response time, steady-state error, and harmonic suppression rate. The recursive least squares method was used to optimize control parameters such as virtual inertia and damping coefficient online, with a forgetting factor set to 0.98 and an initial covariance matrix 100 times the identity matrix. The optimization results were used to adjust the virtual synchronous generator model parameters in real time via an adaptive controller, with an adjustment period of 100ms and a single adjustment amplitude limited to within ±10% of the original value. In practical applications, a 24-bit ADC was used for data acquisition in the phasor measurement unit, with a sampling frequency of 10kHz. Zero-crossing detection employed a linear interpolation algorithm, and a second-order Butterworth filter was used for the phase-locked loop. When the grid frequency abruptly changed from 50Hz to 49.5Hz, the frequency estimation error converged to within ±0.01Hz within 10ms. After receiving the frequency change information, the virtual synchronous generator controller calculated from the second-order oscillation equation model that the virtual rotor angular velocity decreased from 314.16rad / s to 311.02rad / s, with a power angle change of -0.157rad. The generated control commands require the reactive power compensation device to add 20 kvar of capacitive reactive power within 100 ms, and the active power filter to provide 10 kW of active power support within 5 ms. The reactive power compensation device adopts a three-phase bridge inverter topology, with an IGBT switching frequency of 10 kHz and a space vector PWM control strategy. The active power filter uses instantaneous reactive power theory for harmonic detection, with a response time of 3 ms, effectively suppressing harmonics from the 2nd to the 13th order. In coordinated control, the adjustment speed of the reactive power compensation device is limited to 100 kvar / s, and the dynamic response range of the active power filter is ±20% of its rated capacity. The data collected by the monitoring equipment is transformed by wavelet transform to obtain the energy distribution of five frequency bands, where d1 (5-10 kHz) reflects switching ripple, and d3 (1.25-2.5 kHz) reflects higher harmonics. The recursive least squares method updates the parameter estimate every 100ms. When insufficient system damping is detected, the virtual damping coefficient is adjusted from 20 to 22.5, so that the power angle oscillation decays to less than 10% of the initial amplitude within 200ms.Throughout the process, the steady-state voltage deviation was controlled within ±0.5%, and the total harmonic distortion rate was reduced from 4.5% to 1.8%, meeting the power quality standards.

[0038] According to the pre-set collaborative governance strategy, control commands are transmitted from the monitoring equipment to each governance device through the preset communication protocol and interface standard, using the specified communication data frame format and data transmission rate, to carry out bidirectional communication between the monitoring equipment and the governance equipment. If a communication failure or abnormality occurs, a communication failure diagnosis method is used to locate and handle the fault, ensuring stable data transmission between the monitoring equipment and the governance equipment.

[0039] The system acquires control data and generates a standard data frame containing device address, function code, data length, and verification information. The standard data frame supports a preset industrial communication protocol. It receives the standard data frame and performs a CRC (Cyclic Redundancy Check) verification. If the CRC verification passes, the system parses the standard data frame, extracts control commands, and executes corresponding governance operations based on the control commands. A heartbeat mechanism is established, with an initial heartbeat interval set. Network load data is acquired, and the heartbeat interval is adjusted based on the network load data. If a preset number of heartbeat packets are not received consecutively, a communication fault diagnosis program is triggered. The communication fault diagnosis program is executed, using ping (Packet Internet Groper) testing and port status checks to determine the fault type. If a network link fault is identified, the system switches to a pre-configured backup communication channel; if a device fault is identified, the fault information is recorded. The system acquires device communication demand data and network status data, and executes a dynamic bandwidth allocation algorithm based on these data. The bandwidth allocation is adjusted based on the execution result of the dynamic bandwidth allocation algorithm.

[0040] For example, based on a preset collaborative governance strategy, the monitoring device generates a standard data frame containing the device address, function code, data length, and verification information, supporting multiple industrial communication protocols such as Modbus-TCP, Profinet, and EtherCAT. Control commands are encapsulated in the data field and sent to each governance device via industrial Ethernet at a rate of 100Mbps. The transmitted data is encrypted using the TLS 1.3 encryption protocol, and authentication is performed using X.509 certificates to ensure communication security. Upon receiving the data frame, the governance device performs CRC verification and data parsing, extracts the control commands, and executes the corresponding governance operations. Execution results and device status information are prioritized; high-priority information (such as fault alarms) is sent immediately, while low-priority information (such as normal status) is sent in batches periodically. The processed information is packaged into a return data frame and fed back to the monitoring device through the same communication channel. A heartbeat mechanism is established between the monitoring and governance devices, with an initial heartbeat interval set to 2 seconds, adaptively adjusted according to network load, ranging from 1 to 5 seconds. If no heartbeat packet is received for three consecutive times, a communication fault diagnosis procedure is triggered. The diagnostic program uses a depth-first search algorithm to analyze the network topology and locate faulty nodes. Simultaneously, it initiates communication quality monitoring, recording metrics such as packet loss rate and latency. When these metrics exceed preset thresholds, it automatically adjusts communication parameters or routing strategies. The communication fault diagnosis program determines the fault type and location through step-by-step ping tests and port status checks. If it's a network link failure, it automatically switches to a pre-configured backup communication channel, such as a 4G wireless network. If it's a device failure, the fault information is recorded in a log file, and an alarm message is sent to the higher-level control system. Simultaneously, a dynamic bandwidth allocation algorithm is activated, adjusting bandwidth allocation in real time based on the communication needs of each device and network conditions to optimize overall communication efficiency. For devices that are chronically faulty, the system automatically adjusts the network topology to bypass the faulty node, ensuring normal communication for other devices. In practical applications, the monitoring equipment uses a multi-protocol communication module, supporting Modbus-TCP, Profinet, and EtherCAT protocols. Taking Modbus-TCP as an example, a data frame contains a 1-byte device address, a 1-byte function code, a 2-byte start address, a 2-byte register count, and a 2-byte CRC checksum. Control commands are encapsulated in data fields with a maximum length of 252 bytes. TLS 1.3 encryption uses the ECDHE-ECDSA-AES256-GCM-SHA384 cipher suite, and the X.509 certificate uses a 2048-bit RSA key. The industrial Ethernet operates in full-duplex mode, achieving an actual communication rate of 97Mbps. After receiving data, the CRC check takes no more than 0.5ms. Priority sorting uses a heap sort algorithm, setting alarm information priority to 1 and status information priority to 2, with a processing time complexity of O(nlogn).The heartbeat mechanism initially uses a 2-second interval, increasing to 3 seconds when network load exceeds 80% and decreasing to 1.5 seconds when it falls below 20%. A depth-first search algorithm is used for network topology analysis, with an average search depth of 4 layers and a time consumption of approximately 10ms. Communication quality monitoring records the average packet loss rate and round-trip latency within 100ms; parameter adjustments are triggered when the packet loss rate exceeds 1% or the latency exceeds 50ms. The ping test timeout is set to 100ms; three consecutive timeouts are considered a fault. The backup 4G network uses LTE Cat4 technology with a maximum downlink speed of 150Mbps. The dynamic bandwidth allocation algorithm is based on a weighted fair queue with an allocation granularity of 64kbps, adjusted every 500ms. Automatic network topology adjustment uses Dijkstra's algorithm to calculate the shortest path, with a complexity of O(ElogV), where E is the number of edges and V is the number of vertices. The average fault response time of the entire system is controlled within 1 second, and communication reliability reaches 99.999%.

[0041] Obviously, those skilled in the art can make various modifications and variations to the embodiments of this application without departing from the spirit and scope of the embodiments of this application. Therefore, if these modifications and variations to the embodiments of this application fall within the scope of the claims of this application and their equivalents, this application also intends to include these modifications and variations.

Claims

1. A method for collaborative management of schedule and quality in power engineering projects, characterized in that, The method includes: The electromagnetic interference source characteristics of different types of electromagnetic interference sources are obtained to form electromagnetic interference source characteristic templates. Power quality data is collected through power quality monitoring equipment. The collected power quality data is then extracted and compressed through edge computing nodes to obtain compressed power quality characteristic data. The power quality characteristic data is matched and analyzed with the electromagnetic interference source characteristic template. The components in the power quality characteristic data that are similar to the electromagnetic interference source characteristics are marked as power quality monitoring data that are interfered with. The power quality monitoring data marked as being disturbed is processed by an adaptive inter-spectral interference suppression filter to obtain purified power quality monitoring data, and the power quality health status of the distribution network is judged based on the purified power quality monitoring data. If the power quality health status of the distribution network is abnormal, alarm information and intelligent diagnostic reports are generated. Based on the intelligent diagnostic reports, corresponding governance strategies and control instructions are generated, and the control instructions are transmitted to the corresponding power quality governance equipment. After receiving control commands, the power quality management equipment adjusts its operating parameters accordingly. At the same time, the power quality monitoring equipment senses changes in the operating parameters of the power quality management equipment in real time, evaluates the collaborative management effect, and forms a closed-loop feedback control.

2. The method according to claim 1, characterized in that, The process involves acquiring electromagnetic interference source characteristics of different types of electromagnetic interference sources to form electromagnetic interference source characteristic templates, and collecting power quality data through power quality monitoring equipment. The collected power quality data is then processed by edge computing nodes to extract and compress key feature parameters, resulting in compressed power quality feature data, including: The spectral and amplitude characteristics of different electromagnetic interference sources are obtained, and an electromagnetic interference source feature database is established. The electromagnetic interference source feature database contains pre-stored electromagnetic interference source feature templates. Distributed power quality monitoring equipment is used to collect power quality data in real time, including voltage parameters, current parameters, and harmonic parameters. Power quality parameters are verified for data integrity through edge computing nodes; Discrete wavelet transform is applied to the verified power quality data to obtain the feature parameter matrix of multi-scale decomposition. The feature parameter matrix is ​​subjected to dimensionality reduction and compression to generate a compressed power quality feature vector.

3. The method according to claim 1, characterized in that, The step of matching and analyzing power quality characteristic data with the electromagnetic interference source characteristic template, and marking components in the power quality characteristic data similar to electromagnetic interference source characteristics as interfered power quality monitoring data, includes: Edge computing nodes transmit compressed power quality characteristic data to the data analysis center via TCP / IP protocol; The data analysis center decompresses the compressed power quality characteristic data to obtain the original power quality characteristic data. The cosine similarity algorithm is used to calculate the similarity between the original power quality feature data and each sample in the electromagnetic interference source feature template; If the similarity exceeds a preset threshold, the original power quality characteristic data is determined to be subject to electromagnetic interference, and it is marked as interfered power quality monitoring data using a bitmap marking method.

4. The method according to claim 1, characterized in that, The process involves applying adaptive inter-spectral interference suppression filtering to the power quality monitoring data marked as interfered, resulting in purified power quality monitoring data. Based on this purified data, the power quality health status of the distribution network is determined, including: Spectral analysis was performed on the power quality monitoring data affected by interference, and the power spectral density of the monitoring data was calculated using the fast Fourier transform algorithm. Based on the power spectral density, the energy distribution characteristics of the interfered frequency band are identified, and the center frequency and bandwidth information of the interfered frequency band are obtained. The normalized minimum mean square error algorithm is used to dynamically adjust the parameters of the adaptive filter, wherein the adaptive filter parameters include the filter coefficients. If the filter coefficients match the characteristics of the interference frequency band, then the purified power quality monitoring data will be output. The purified power quality monitoring data is transmitted to the distribution automation master station system via a transport layer security protocol. The purification data received by the power distribution automation master station system is decrypted and its integrity is verified. If the verification is successful, the purification data is stored in the real-time database. Obtain power quality data stored in the real-time database and perform spectrum analysis on the power quality data; Extract key indicator parameters from the spectrum analysis results; The power quality health index is calculated based on the key indicator parameters to determine the power quality health status of the distribution network.

5. The method according to claim 1, characterized in that, If the power quality health status of the distribution network is abnormal, an alarm message and an intelligent diagnostic report are generated. Based on the intelligent diagnostic report, corresponding governance strategies and control commands are generated, and the control commands are transmitted to the corresponding power quality governance equipment, including: The system acquires the power quality health status indicators of the distribution network, uses a comprehensive evaluation method to determine whether the power quality is abnormal, and triggers an alarm mechanism if the health index is lower than the preset threshold. Based on the alarm mechanism, the random forest algorithm is used to analyze the power quality anomaly data to obtain the specific types of power quality problems; For each specific type of power quality problem, a corresponding governance strategy is retrieved from a preset governance strategy library, and control instructions are generated. The control commands are encapsulated into standardized network data packets by a software-defined network controller, and the data packets are processed using the P4 programmable protocol. The standardized network data packets are transmitted to the power quality management equipment.

6. The method according to claim 1, characterized in that, After receiving the control command, the power quality management equipment adjusts its operating parameters accordingly, including: The power quality management equipment obtains key parameter information by parsing the control commands through an embedded processor. Voltage and current waveform data are acquired using a high-speed data acquisition unit, and the sampling rate of the high-speed data acquisition unit is set to a preset value. Waveform data is processed using the Fast Fourier Transform algorithm to obtain harmonic content and power factor; Based on harmonic content and power factor, the reactive power compensation capacity and harmonic mitigation parameters are calculated using the recursive least squares method. If the reactive power compensation capacity exceeds the preset threshold, the improved fuzzy control algorithm will be triggered. The improved fuzzy control algorithm's input variables include voltage deviation rate, total harmonic distortion rate, and power factor. Based on the input variables, the adjustment amount of the treatment equipment is determined through a pre-established fuzzy rule base; The adjustment amount is defuzzified using the center of gravity method to obtain specific adjustment instructions; The operating parameters of the reactive power compensation device, active filter, and static var generator are adjusted according to the adjustment instructions. Obtain the voltage compliance rate, total harmonic distortion rate compliance rate, and average power factor; The voltage compliance rate, total harmonic distortion rate compliance rate, and average power factor are weighted and averaged according to preset weights to obtain the evaluation results of the treatment effect.

7. The method according to claim 6, characterized in that, The power quality monitoring equipment senses changes in the operating parameters of the power quality management equipment in real time, evaluates the collaborative management effect, and forms a closed-loop feedback control. It also includes: real-time collection of power quality indicators during the power quality management process, such as voltage RMS, current RMS, voltage distortion rate, current distortion rate, voltage imbalance, current imbalance, harmonic content, harmonic current amplitude, power factor, and reactive power. A preset management effect evaluation function is used to evaluate the collaborative management effect. If the management effect does not meet expectations, the evaluation result is used as a feedback signal to form a closed-loop feedback control. A virtual synchronous generator is used to control the reactive power compensation device and active power filter management equipment. Virtual devices with synchronous generator characteristics participate in reactive power regulation and harmonic suppression of the power grid by sensing grid frequency and voltage parameters in real time, simulating the inertial response and damping characteristics of a synchronous generator. The monitoring equipment evaluates the operating status and governance effect of the virtual synchronous generator in real time, and optimizes the control parameters and strategies of the virtual synchronous generator. According to the pre-set collaborative governance strategy, control commands are transmitted from the monitoring equipment to each governance device through preset communication protocols and interface standards, using the specified communication data frame format and data transmission rate, to conduct bidirectional communication between the monitoring equipment and the governance devices. If a communication failure or abnormality occurs, a communication fault diagnosis method is used to locate and handle the fault.

8. The method according to claim 7, characterized in that, The system collects power quality indicators in real time during the power quality management process, including RMS voltage, RMS current, voltage distortion rate, current distortion rate, voltage imbalance, current imbalance, harmonic content, harmonic current amplitude, power factor, and reactive power, using monitoring equipment. A preset management effect evaluation function is used to assess the collaborative management effect. If the management effect does not meet expectations, the evaluation result is used as a feedback signal to form a closed-loop feedback control, including: The system acquires raw voltage and current signal data collected by power quality monitoring equipment. After median filtering to remove outliers, the raw data is smoothed using a moving average method to obtain preprocessed data. A fast Fourier transform is performed on the preprocessed data to obtain the spectral analysis results of the voltage and current signals. Voltage distortion rate, current distortion rate, voltage imbalance, and current imbalance are calculated based on the spectral analysis results. A weighted sum of these parameters is obtained using a preset governance effect evaluation function to obtain a comprehensive evaluation score. If the comprehensive evaluation score is lower than a preset threshold, the governance effect is deemed unsatisfactory. Based on the deviation and rate of change of the comprehensive evaluation score from the target value, a fuzzy PID control algorithm is used to adjust the proportional-integral-derivative (PI-DE) control parameters. The fuzzy PID control algorithm includes fuzzification of the input variables, execution of fuzzy rule inference, and defuzzification using the centroid method. Control commands for the power quality governance equipment are generated based on the adjusted PI-DE control parameters. The control commands are then sent to the power quality governance equipment via an industrial Ethernet network.

9. The method according to claim 7, characterized in that, The method utilizes a virtual synchronous generator to control reactive power compensation devices and active power filters as virtual devices with synchronous generator characteristics. By sensing grid frequency and voltage parameters in real time, it simulates the inertial response and damping characteristics of a synchronous generator to participate in reactive power regulation and harmonic suppression of the grid. Monitoring equipment is used to evaluate the operating status and governance effect of the virtual synchronous generator in real time, and to optimize the control parameters and strategies of the virtual synchronous generator, including: The system employs a phasor measurement unit to acquire grid frequency and voltage parameters. Based on these parameters, a frequency estimate is determined using zero-crossing detection and phase-locked loop (PLL) technology. Using the frequency estimate and voltage parameters, the virtual rotor angular velocity and power angle are calculated using a pre-defined second-order oscillation equation model in the virtual synchronous generator controller. Control commands for reactive power and active power are generated based on these virtual rotor angular velocity and power angle. These control commands are then sent to the reactive power compensation device and the active power filter. The reactive power compensation device adjusts the number and timing of capacitor switching according to the control commands, while the active power filter adjusts the amplitude and phase of the output current. The output voltage, current, and power data of the virtual synchronous generator are acquired through monitoring equipment. If the output voltage, current, and power data deviate from preset thresholds, the virtual inertia and damping coefficients are optimized online using a recursive least squares method.

10. The method according to claim 7, characterized in that, According to a pre-set collaborative governance strategy, control commands are transmitted from monitoring devices to various governance devices through a preset communication protocol and interface standard, using a specified communication data frame format and data transmission rate, enabling bidirectional communication between the monitoring devices and the governance devices. If a communication failure or anomaly occurs, a communication fault diagnosis method is used to locate and handle the fault, including: The system acquires control data and generates a standard data frame containing device address, function code, data length, and verification information. The standard data frame supports a preset industrial communication protocol. A cyclic redundancy check (CR) is performed on the standard data frame. If the CR passes, the standard data frame is parsed, control commands are extracted, and corresponding management operations are executed based on the control commands. A heartbeat mechanism is established, an initial heartbeat interval is set, network load data is acquired, and the heartbeat interval is adjusted based on the network load data. If a preset number of heartbeat packets are not received consecutively, a communication fault diagnosis program is triggered. The communication fault diagnosis program is executed, and the fault type is determined through Internet packet explorer testing and port status checks. If a network link fault is identified, the system switches to a pre-configured backup communication channel; if a device fault is identified, the fault information is recorded. Device communication requirement data and network status data are acquired, and a dynamic bandwidth allocation algorithm is executed based on these data. The bandwidth allocation is adjusted based on the execution result of the dynamic bandwidth allocation algorithm.

Images