An FPGA-based micro-grid power quality intelligent analysis method, system and medium

Abstract

The application relates to the technical field of micro-grid, and particularly provides a micro-grid power quality intelligent analysis method and system based on FPGA and a medium, wherein the method performs parallel processing and analysis on multi-channel synchronous sampling data through FFT transformation, RMS calculation and wavelet decomposition; a deep learning model is introduced for intelligent identification and prediction; the hardware parallel computing capacity of FPGA is used to accelerate core algorithms including deep learning inference, real-time, accurate and forward-looking monitoring, identification, classification, prediction and analysis of various power quality event disturbance types such as harmonics, interharmonics, voltage sag and transient are realized. The application significantly improves the accuracy and predictability of event detection, and meets the stringent requirements of the micro-grid under dynamic working conditions.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of micro-grid, and particularly relates to a micro-grid power quality intelligent analysis method, system and medium based on FPGA. BACKGROUND

[0002] Under the promotion of global energy structure transformation and the "double carbon" target, distributed new energy such as photovoltaic and wind power is increasingly widely used in low-voltage micro-grids. These micro-grid systems integrate various distributed power sources (such as photovoltaic panels, wind turbines), energy storage devices, power electronic converters, and various types of loads (including traditional linear loads and new non-linear dynamic loads). This integration brings significant advantages, such as improved power supply reliability and effective utilization of renewable energy. However, the complexity and dynamic characteristics of such systems also pose new challenges, particularly in terms of power quality.

[0003] Low-voltage micro-grids contain a large number of power electronic devices and frequency converters, inverters, etc., resulting in complex harmonic and interharmonic problems. The presence of interharmonics not only exacerbates electromagnetic interference within the system, but also can cause a series of special problems, such as light flickering of lighting devices, negatively affecting user power experience; at the same time, interharmonics of specific frequencies can interfere with wireless communication, power line carrier communication, and specific frequency signal control logic in industrial control systems, further affecting the stability and safety of the system. In addition, with the access of new dynamic loads such as electric vehicles, voltage transient rise, transient drop, interruption, etc. in micro-grids also become more frequent, which poses a threat to the safe operation of the power grid.

[0004] Existing monitoring and analysis technologies have obvious deficiencies in addressing the above challenges, particularly in terms of real-time response capability, disturbance identification accuracy, complex signal analysis capability, and comprehensive processing of multiple disturbance forms, which still cannot meet the requirements of high-speed dynamic characteristics of micro-grids. SUMMARY

[0005] In order to solve the problems existing in the prior art, the application provides a micro-grid power quality intelligent analysis method based on FPGA, comprising the following steps:

[0006] S1: multi-channel synchronous sampling of voltage and current data, the voltage and current data including three-phase voltage, three-phase current, neutral point to ground voltage and neutral line current, pre-processing the voltage and current data to obtain pre-processed data;

[0007] S2: the pre-processed data is processed as follows:

[0008] S2.1: performing fast Fourier transform on the pre-processed data to obtain frequency spectrum data corresponding to each channel respectively;

[0009] S2.2: Calculate the voltage RMS value sequence based on the voltage data;

[0010] S2.3: Decompose the preprocessed data using wavelet transform to extract transient features;

[0011] S3: Perform power quality analysis based on the output data in S2:

[0012] S3.1: Perform harmonic and interharmonic analysis based on the frequency spectrum data;

[0013] S3.2: Identify voltage sag, swell and interruption events based on the voltage RMS value sequence, and calculate short-term flicker and long-term flicker;

[0014] S4: Use a deep learning model to perform anomaly detection on the power quality parameters in S1-S3, and output the power quality anomaly type probability according to the anomaly type, which at least includes harmonic over-limit, voltage sudden change, fast transient, and interharmonic anomaly.

[0015] Further, the interharmonic analysis includes the following steps:

[0016] S3.1.1: Perform amplitude correction and energy correction on the frequency spectrum signals corresponding to each voltage and current data respectively;

[0017] S3.1.2: Group the spectrum lines between two harmonic frequencies with the integer multiples of the fundamental frequency as the center, calculate the root mean square value of each group, and obtain the effective value of the interharmonic group;

[0018] S3.1.3: Identify significant interharmonic components in the frequency spectrum by peak value search and frequency spectrum difference, and estimate their frequency and amplitude;

[0019] S3.1.4: Calculate the interharmonic distortion rate or the interharmonic distortion rate of a specific frequency band, and compare it with the known interharmonic frequencies that may cause equipment failure to evaluate the impact of interharmonic on the microgrid.

[0020] Preferably, the method further includes three-phase imbalance analysis, specifically including:

[0021] S2.5: Perform symmetric component decomposition on the preprocessed data to extract symmetric components, including the amplitude and phase of positive sequence, negative sequence and zero sequence;

[0022] S3.5: Calculate the voltage imbalance and current imbalance according to the symmetric components.

[0023] Based on the above scheme, the method further includes a triggering mechanism for power quality events:

[0024] Level 1 trigger: Monitor instantaneous voltage or current values. When the instantaneous voltage or current value changes beyond a preset threshold, trigger an interrupt signal to initiate an abnormal response and capture transient disturbances.

[0025] Second-level trigger: Based on power quality parameters, when the logical combination conditions configured by the user are met, the event log is triggered.

[0026] Furthermore, the power quality event triggering mechanism also includes a third-level trigger: based on the anomaly type probability output by the CNN model, when the probability of predicting a certain anomaly pattern increases significantly, and the first or second-level triggering conditions are not met, the following operations are performed:

[0027] Adjust the threshold parameters of the first or second level trigger conditions to improve the sensitivity of capturing this type of abnormal pattern;

[0028] CNN identifies patterns as suspicious or high-risk, triggering data logging for subsequent analysis or CNN model training.

[0029] Generate early warning information to indicate potential power quality problems or equipment malfunctions.

[0030] Preferably, the method further includes S4: adaptive analysis based on power quality parameters, specifically including:

[0031] S4.1: By monitoring power quality parameters, assess the current operating condition of the microgrid. The power quality parameters include at least spectral entropy, RMS rate of change, load information, and harmonics / interharmonics.

[0032] S4.2: Determine the type of disturbance:

[0033] If transient features are detected, analyze the wavelet data and start high-speed waveform recording;

[0034] If harmonic or interharmonic anomalies are detected, increase the FFT analysis frequency and the harmonic and interharmonic analysis depth.

[0035] If the operating conditions are stable, perform routine testing;

[0036] S4.3: Adjust the parameters of the triggering mechanism according to the operating conditions and disturbance type.

[0037] Based on the same inventive concept, this invention also provides an FPGA-based intelligent power quality analysis system for microgrids, using the aforementioned intelligent power quality analysis method for microgrids, including:

[0038] The data acquisition module is used to collect current and voltage data of the microgrid.

[0039] The preprocessing module is used to receive and process data from the data acquisition module;

[0040] The power quality analysis module is used to analyze and calculate the output data of the preprocessing module. It includes at least an FFT analysis unit, an RMS calculation unit, a wavelet transform feature extraction unit, and a neural network anomaly detection unit. The FFT analysis unit, RMS calculation unit, wavelet transform feature extraction unit, and neural network anomaly detection unit process in parallel.

[0041] The storage module is used to store the data from the data acquisition module, the preprocessing module, and the power quality analysis module.

[0042] Furthermore, the data acquisition module employs an 8-channel synchronous sampling ADC to acquire three-phase voltage, three-phase current, one neutral line current, and one neutral point-to-ground voltage. The front end of each channel is configured with an analog filter. The preprocessing module receives the ADC sampling data output by the data acquisition module and performs digital filtering, data format conversion, extraction, noise reduction, and normalization on the ADC sampling data before outputting preprocessed data. The preprocessed data is then input into the power quality analysis module.

[0043] Based on the above scheme, the FFT analysis unit processes the 8-channel preprocessed data. The FFT analysis unit includes a high-precision mode and a fast update mode. The high-precision mode outputs high-resolution spectrum data, and the fast update mode outputs the spectrum amplitude and phase data corresponding to each channel data.

[0044] Preferably, the power quality analysis module also includes:

[0045] The harmonic analysis unit acquires the high-resolution spectrum data and obtains harmonic characteristics, which include at least the total harmonic distortion rate and harmonic parameters of each channel.

[0046] The interharmonic analysis unit acquires the high-resolution spectrum data, performs interharmonic grouping and interharmonic parameter calculation, and obtains interharmonic characteristics.

[0047] The voltage fluctuation and flicker calculation unit calculates short-time flicker and long-time flicker based on the voltage RMS value sequence obtained from the RMS calculation unit.

[0048] Furthermore, the feature matrix is ​​input into the neural network anomaly detection unit, the feature matrix including at least the interharmonic feature, harmonic feature and time-domain feature, and outputs the probability of multiple anomaly types, the anomaly types including at least harmonic over-limit, voltage sudden change, fast transient and interharmonic anomaly; the neural network anomaly detection unit is trained using historical power quality parameters and corresponding anomaly types.

[0049] Furthermore, the system also includes an adaptive analysis module, which includes an operating condition assessment model that uses power quality parameters to assess the operating conditions of the microgrid and dynamically adjusts the power quality analysis parameters based on the operating condition assessment results.

[0050] On the other hand, a computer-readable storage medium is also provided having a computer program thereon, the computer program comprising:

[0051] The processor-executable instructions, when executed by the processing unit of the on-chip system platform, perform the steps of the microgrid power quality intelligent analysis method as described above;

[0052] The configuration data of the FPGA, and / or the configuration data of the FPGA, when the configuration data is loaded into the programmable logic cell of the field-programmable gate array (FPGA), performs the steps of the method described above.

[0053] Compared with the prior art, the present invention has the following beneficial effects:

[0054] 1. It achieves accurate measurement and classification of various power quality events such as harmonics, interharmonics, flicker, and transient events. It also supports CNN anomaly prediction. Combined with power quality parameters, it can identify, classify, and predict multiple types of disturbances, significantly improving the accuracy of anomaly detection.

[0055] 2. The heterogeneous architecture of FPGA and multi-core ARM processor and the optimized hardware and software co-design significantly improve the real-time performance and data processing capabilities of power quality analysis, and can meet the millisecond-level monitoring requirements under the dynamic operating conditions of microgrids.

[0056] 3. The three-level intelligent triggering mechanism combines rapid hardware response, flexible software logic, and learning and prediction capabilities, which can more effectively capture various power quality events and dynamically optimize the triggering strategy to reduce missed alarms and false alarms.

[0057] 4. An adaptive analysis strategy based on parameters such as spectral entropy and THD enables the system to dynamically adjust the analysis algorithm and its parameters according to the actual power grid operating conditions, thereby improving the relevance and efficiency of the analysis. Attached Figure Description

[0058] Figure 1 This is a flowchart of the power quality analysis method of this application;

[0059] Figure 2 This is a flowchart of the interharmonic analysis in this application;

[0060] Figure 3 This is a flowchart of the adaptive analysis for this application;

[0061] Figure 4This is a structural diagram of the power quality analysis system of this application;

[0062] Figure 5 This is a block diagram of the system hardware architecture of this application;

[0063] Figure 6 This is a block diagram of the system software architecture of this application. Detailed Implementation

[0064] The invention will be further described below with reference to specific embodiments.

[0065] Example 1

[0066] like Figure 1 As shown, this application provides an intelligent power quality analysis method for microgrids based on FPGA, which follows international standards for the definition and measurement of relevant parameters. The method includes the following steps:

[0067] Step S1: Use an 8-channel synchronous sampling ADC to collect voltage and current data. The voltage and current data include three-phase voltage (Ua, Ub, Uc), three-phase current (Ia, Ib, Ic), neutral point to ground voltage (Vng), and neutral line current (In). Each analog input channel is equipped with a filter at the front end to prevent high-frequency signal aliasing.

[0068] Preferably, a Butterworth anti-aliasing filter with a cutoff frequency of 120 kHz is selected;

[0069] To improve signal quality and facilitate subsequent processing, the voltage and current data undergo filtering, format conversion, extraction, denoising, and normalization preprocessing after initial filtering. Specifically, to suppress out-of-band noise, digital filtering is performed, such as using a configurable 32nd-order FIR filter. To facilitate processing, the data format is standardized, such as converting 16-bit integer data to fixed-point or floating-point format. The sampling rate is reduced as needed to save subsequent resources. Random noise is removed using moving average or wavelet thresholding. Normalization is performed to unify the data scale of each channel, scaling the data to a single interval.

[0070] The preprocessed voltage and current data are obtained after the above preprocessing.

[0071] Step S2, the preprocessed voltage and current data output from S1 are processed as follows:

[0072] Step S2.1: Perform FFT transformation on the preprocessed voltage and current data to obtain the spectrum data corresponding to each channel;

[0073] Step S2.2: Calculate the root mean square (RMS) value sequence of the power frequency cycle / half cycle of voltage and current data for each channel in parallel, and extract the RMS rate of change, which can provide voltage and current fluctuation information at different time resolutions.

[0074] Step S2.3: The preprocessed voltage and current data are decomposed using a 4-layer Daubechies-4 mother wavelet to extract wavelet features such as energy and significance coefficients at each scale, which are used to characterize fast transient events.

[0075] The above-mentioned parallel processing and analysis of multi-channel synchronous sampling data is achieved through FFT transformation, RMS calculation and wavelet decomposition, realizing real-time and high-speed processing and analysis of the acquired data.

[0076] Step S2.4: Based on the preprocessed voltage and current data, calculate the active power and reactive power, which are used for load status monitoring.

[0077] Step S3: Perform power quality analysis based on the output data from S2.

[0078] Step S3.1: Extract the amplitude and phase information of each harmonic based on the spectrum data, identify the fundamental wave and harmonic components based on the amplitude and phase information, and analyze the harmonics and interharmonics;

[0079] Based on amplitude and phase information, the compliance of each parameter is evaluated, the total harmonic distortion (THD) of each channel is calculated, THD represents the ratio of the effective value of all harmonic components to the fundamental frequency; the harmonic content of each order is calculated to determine whether a certain harmonic exceeds the standard; and the amplitude is used to determine whether it exceeds the maximum allowable value.

[0080] Furthermore, historical data on THD and key subharmonics are used to analyze trends and provide early warnings.

[0081] like Figure 2 As shown, interharmonic analysis includes the following steps:

[0082] Step S3.1.1: In order to compensate for the influence of the window function, amplitude correction and energy correction are performed on the 8-channel spectrum data corresponding to each voltage and current.

[0083] Step S3.1.2: Using integer multiples of the fundamental frequency as the center, group the spectral lines between the two harmonic frequencies, calculate the root mean square (RMS) value of each group, and obtain the effective value of the interharmonic group / subgroup.

[0084] Step S3.1.3: Identify significant interharmonic components in the spectral data through peak search and spectral difference, and estimate their frequency and amplitude.

[0085] Step S3.1.4: Calculate the interharmonic distortion rate (TIHD) of each frequency band or the interharmonic distortion rate of a specific frequency band, compare it with the known interharmonic frequencies that may cause equipment failure, and evaluate the impact of interharmonics on the microgrid.

[0086] Step S3.2: Identify voltage sag, voltage swell, and voltage interruption events by detecting changes in the voltage RMS value, and extract a first feature parameter, which includes at least: occurrence time, duration, depth / height, and missing voltage; the first feature parameter is used to identify and classify voltage transient events.

[0087] Furthermore, by combining the low-frequency voltage interharmonic information provided by interharmonic analysis, we can help determine the cause of flicker or assess the severity caused by interharmonics.

[0088] Step S3.3: Perform digital filtering on the voltage RMS value sequence to extract the voltage envelope, and calculate the short-time flicker (P_st) and long-time flicker (P_lt).

[0089] When the calculated P_st value exceeds the preset threshold, an alarm event is triggered and relevant data is recorded for subsequent analysis and use.

[0090] Step S3.4: Based on the wavelet features extracted in S2.3, identify fast transient events and extract second feature parameters. The second feature parameters include at least the peak value, duration, and oscillation frequency. The second feature parameters are used to accurately classify and analyze transient events.

[0091] Voltage sags, sags, and interruptions are identified by detecting changes in the voltage RMS value, while wavelet feature analysis is performed to identify fast transient events and to perform preliminary classification of fast transient events.

[0092] Furthermore, to detect the power quality imbalance of the microgrid, the method also includes three-phase imbalance analysis, specifically including:

[0093] Step S2.5: Perform symmetric component decomposition on the preprocessed voltage and current data, and extract the symmetric components, which include the amplitude and phase of positive sequence, negative sequence and zero sequence.

[0094] Step S3.5: Calculate the voltage unbalance (VUF) and current unbalance (CUF) based on the symmetrical components.

[0095] If the voltage imbalance is detected to be continuously exceeding the standard, a preliminary load adjustment strategy or energy storage unit compensation suggestion will be generated based on the microgrid's operating status and configuration information.

[0096] It should be noted that the detection and calculation of various power quality parameters comply with international standards.

[0097] Furthermore, this invention uses a deep learning model to intelligently analyze power quality parameters. During microgrid system operation, the collected or calculated power quality parameters are integrated into a feature matrix and input into the model for learning and analysis. In this embodiment, a lightweight neural network (CNN) can be used to detect anomalies in power quality parameters and output the type of power quality anomaly and its corresponding probability. These anomaly types include harmonic exceedances, voltage surges, rapid transients, interharmonic anomalies, and unknown anomalies. In addition, the model is trained using supervised learning, i.e., using historical power quality parameters and corresponding anomaly types as training data. This invention can also employ other models such as RNNs, LSTMs, and support vector machines.

[0098] Example 2

[0099] Based on Example 1, this embodiment adds a triggering mechanism for power quality events to ensure timely response and effective data capture for various power quality events, including:

[0100] (1) First-level triggering: monitor the instantaneous value of voltage or current. When the instantaneous value of voltage or current changes beyond the preset threshold, an interrupt signal is immediately generated to start the abnormal response mechanism and capture transient disturbances, such as impact, rapid overvoltage, and the initial stage of voltage drop.

[0101] On the one hand, the interrupt signal is marked and responded to after it is generated. On the other hand, the data before and after the trigger moment is automatically latched into the buffer to provide high-accuracy raw waveforms for subsequent analysis.

[0102] The first-level trigger is a direct hardware trigger, which directly monitors the instantaneous values ​​of line voltage and current through an 8-channel analog comparator. In this embodiment, when the instantaneous voltage value exceeds 120% of the rated voltage peak value or falls below 80% of the rated voltage peak value, a hardware interrupt signal is immediately generated. This trigger operation has the shortest response time, enabling rapid response to power quality issues.

[0103] (2) Second-level triggering is based on power quality parameters, including at least RMS sequence value, THD, voltage flicker value, transient characteristic energy (energy distribution under different wavelet scales in wavelet transform), possible interharmonic disturbances and interharmonic group amplitude. When these parameters meet the user-preset logical combination conditions (such as THD exceeding the limit, P_st exceeding the limit, transient energy increase, interharmonic amplitude exceeding the standard, etc.), the event record is triggered.

[0104] The second-level trigger is a software logic trigger, which combines and configures the calculated power quality parameters using logical conditions and employs priority queue scheduling to ensure that the event judgment and triggering delay is controlled to be less than 50µs, and records the detailed parameters of the power quality event.

[0105] (3) The third level of triggering is to predict the probability of anomalies through the CNN model. Since the training data of the CNN model contains rich harmonics, interharmonics and various combined disturbance features, its prediction can reflect potential early faults or complex disturbance risks that are not easily captured by fixed thresholds.

[0106] Based on the anomaly type probabilities output by the CNN model, when the probability of predicting a certain anomaly pattern increases significantly, and the first or second level triggering conditions are not met, the following operations can be performed:

[0107] Adjust the threshold parameters of the first or second level trigger conditions to improve the sensitivity of capturing this type of abnormal pattern;

[0108] At this point, the CNN considers the pattern suspicious or high-risk, triggering data recording for subsequent analysis or CNN model training.

[0109] Generate early warning information to alert maintenance personnel to potential power quality problems or specific equipment malfunctions.

[0110] The third-level trigger improves the predictability and accuracy of power quality analysis and reduces false alarms. At the same time, by updating the CNN model and trigger logic, it can continuously adapt to new anomaly patterns and evolution characteristics that emerge in microgrids.

[0111] By working together through the three triggering mechanisms mentioned above, it is possible to ensure rapid response to power quality events, accurately capture power quality events, and achieve anomaly warnings and data recording, thereby realizing comprehensive and efficient power quality event monitoring and data management.

[0112] Example 3

[0113] Based on Example 2, this embodiment, in order to improve the accuracy of power quality analysis, further includes step S4: performing adaptive analysis based on power quality parameters, such as... Figure 3 As shown, it specifically includes:

[0114] Step S4.1: By monitoring power quality parameters, assess the current operating condition of the microgrid. The operating conditions include stable, slight fluctuation, severe disturbance, and suspected fault. The power quality parameters include at least spectral entropy, RMS rate of change, load information, and harmonics / interharmonics.

[0115] Spectral entropy is an indicator that quantifies the uniformity or complexity of a signal's power spectrum distribution. A low spectral entropy value indicates that the signal energy is mainly concentrated on a few frequency components (such as a pure fundamental signal), while a high spectral entropy value indicates that the signal energy is more dispersed across multiple frequency components (such as a distorted signal containing abundant harmonics, interharmonics, or noise). Therefore, by monitoring changes in spectral entropy, the overall power quality status or the occurrence of disturbances can be effectively reflected, providing an important basis for operational condition assessment.

[0116] Step S4.2, make adjustments based on the operating condition assessment results:

[0117] If transient features are detected, wavelet data is analyzed first, and high-speed waveform recording is initiated.

[0118] If harmonic or interharmonic anomalies are detected, increase the FFT analysis frequency and the analysis depth of harmonics and interharmonics; if persistent interharmonics are detected, switch to a higher resolution FFT configuration or increase the execution frequency of interharmonics.

[0119] If the operating conditions are found to be stable, then routine testing will be emphasized.

[0120] Step S4.3: Adjust the FFT analysis window or RMS calculation window according to the perturbation characteristics, and configure the window length according to the requirements to balance resolution and response speed.

[0121] Step S4.4: Adjust the parameters of the triggering mechanism, such as the threshold values ​​of THD and P_st, according to the operating conditions and disturbance type.

[0122] Example 4

[0123] Based on the intelligent power quality analysis method for microgrids described in the above embodiments, this embodiment provides an intelligent power quality analysis system for microgrids based on FPGA, such as... Figure 4 As shown, it includes:

[0124] The data acquisition module uses an 8-channel synchronous sampling ADC to acquire three-phase voltage, three-phase current, one neutral line current and one neutral point-to-ground voltage. The front end of each channel is equipped with an analog filter.

[0125] The preprocessing module receives ADC sampling data output by the data acquisition module, performs digital filtering, data format conversion, extraction, noise reduction and normalization on the ADC sampling data, and outputs preprocessed data, which is then input into the power quality analysis module.

[0126] The power quality analysis module is used to analyze and calculate the output data of the preprocessing module. It includes at least an FFT analysis unit, an RMS calculation unit, a wavelet transform feature extraction unit, and a neural network anomaly detection unit. The FFT analysis unit processes the 8-channel preprocessed data and includes a high-precision mode and a fast update mode. The high-precision mode uses a long data window to output high-frequency spectrum data for accurate harmonic and interharmonic analysis. The fast update mode uses a short data window to achieve millisecond-level fast spectrum updates for dynamic monitoring and rapid event response, outputting the spectrum amplitude and phase data corresponding to each channel.

[0127] The RMS computing unit performs parallel calculations of the root mean square (RMS) value sequences of voltage and current at power frequency or half-cycle for each channel; the Daubechies-4 (db4) mother wavelet is used to achieve 4-level wavelet packet decomposition of the signal and extract fast transient features.

[0128] The feature matrix is ​​input into the neural network anomaly detection unit. The feature matrix includes at least the interharmonic feature, harmonic feature, and time-domain feature. The probability of multiple anomaly types is output. The anomaly types include at least harmonic over-limit, voltage sudden change, fast transient, and interharmonic anomaly. The neural network anomaly detection unit is trained using historical power quality parameters and corresponding anomaly types.

[0129] The storage module includes a high-speed cache (DDR4) and non-volatile memory (Flash) for storing data from the data acquisition module, preprocessing module, and power quality analysis module.

[0130] Furthermore, the power quality analysis module also includes: a harmonic analysis unit, which acquires the high-resolution spectrum data and obtains harmonic characteristics, wherein the harmonic characteristics include at least the total harmonic distortion rate and harmonic parameters of each channel;

[0131] The interharmonic analysis unit acquires the high-resolution spectrum data, performs interharmonic grouping and interharmonic parameter calculation, and obtains interharmonic characteristics.

[0132] The voltage fluctuation and flicker calculation unit calculates short-time flicker and long-time flicker based on the voltage RMS value sequence obtained from the RMS calculation unit.

[0133] The symmetrical component calculation unit is used to calculate voltage unbalance and current unbalance.

[0134] Furthermore, the power quality analysis system also includes an adaptive analysis module, which includes an operating condition assessment model that uses power quality parameters to assess the operating conditions of the microgrid and dynamically adjusts the quality analysis parameters based on the operating condition assessment results.

[0135] likeFigure 5 and Figure 6 As shown, the power quality analysis system and method in this application are based on a collaborative computing architecture of a multi-core ARM processor (PS end) and an FPGA (PL end). The power quality analysis is achieved through task division between the PL end and the PS end. The PS end is responsible for performing complex system management, advanced algorithm processing, AI inference control and human-computer interaction tasks, while the PL end undertakes high-speed data acquisition, parallel signal preprocessing, real-time spectrum analysis and key hardware acceleration tasks, thus forming an efficient heterogeneous computing architecture.

[0136] Specifically, the PL side mainly implements the following functions: high-speed ADC interface control, 8-channel synchronous sampling data buffer, real-time RMS value calculation, wavelet transform and FFT analysis, dual-mode spectrum analysis (high-precision mode and fast update mode), CNN inference accelerator, and a first-level hardware triggering mechanism; it processes multiple computing units in parallel, and its core goal is to provide low-latency, high-throughput data stream processing capabilities and provide high-quality basic feature data for the PS side.

[0137] The PS (Power Supply) architecture runs an asymmetric multiprocessing (AMP) system. Some cores run an RTOS to ensure hard real-time communication and event response with the FPGA, while the remaining cores run a Linux operating system. This system is responsible for executing IEC standard-compliant harmonic / interharmonic post-processing algorithms, flicker calculation, spectral entropy analysis, deep learning model management, adaptive analysis strategy decision-making, three-level intelligent triggering mechanism coordination, network communication protocol stack, data storage, and a web interface. Furthermore, the PS uses a dynamic configuration mechanism to adjust parameters and switch modes for the PL (Power Processor) functional modules, achieving optimal utilization of system resources and intelligent closed-loop control.

[0138] Furthermore, the power quality analysis method according to the present invention can be recorded in a computer-readable recording medium. Specifically, according to the present invention, when the program is executed on an embedded processor (including a multi-core ARM processor running Linux and RTOS) of an MPSoC platform, or when the configuration code is loaded into a programmable logic unit of an FPGA, the system is enabled to perform the steps of the microgrid power quality intelligent analysis method as described above.

[0139] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, program segment, or portion of code containing at least one executable instruction for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0140] In general, various exemplary embodiments of the present invention can be implemented in hardware or dedicated circuitry, software, firmware, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device. When aspects of embodiments of the present invention are illustrated or described as block diagrams, flowcharts, or represented using certain other images, it will be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as non-limiting examples in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or certain combinations thereof.

[0141] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0142] While the specific embodiments of the present invention have been described above, they are not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A FPGA-based micro-grid power quality intelligent analysis method, characterized in that, Comprise: S1: multi-channel synchronous sampling voltage, current data, the voltage, current data includes three-phase voltage, three-phase current, neutral point to ground voltage and neutral line current, the voltage, current data is pretreated to obtain pretreated data; S2: the pretreated data is processed as follows: S2.1: the fast Fourier transform is carried out on the pretreated data to obtain the frequency spectrum data corresponding to each channel respectively; S2.2: the voltage RMS value sequence is calculated based on the voltage data; S2.3: the pretreated data is decomposed by wavelet transform to extract transient characteristics; S3: based on the output data in S2, the power quality analysis is carried out: S3.1: based on the frequency spectrum data, the harmonic and interharmonic analysis is carried out; S3.2: based on the voltage RMS value sequence, voltage sag, swell and interruption events are identified, and short-time flicker and long-time flicker are calculated; S4: the power quality parameters in S1-S3 are integrated into a feature matrix to input a deep learning model for anomaly detection, and the power quality anomaly type and probability are output, the anomaly type at least includes harmonic overrun, voltage sudden change, fast transient, interharmonic anomaly; The method further comprises a triggering mechanism for power quality events: First level trigger: monitor the voltage or current instantaneous value, when the voltage or current instantaneous value changes beyond the preset threshold, trigger an interruption signal, which is used to start the abnormal response and capture the transient disturbance; Second level trigger: based on the power quality parameters, when the logic combination condition configured by the user is met, trigger event recording; Third level trigger: based on the anomaly type probability output by the CNN model, when the probability of predicting a certain anomaly mode significantly increases, and the first or second level trigger condition is not met, the following operations are performed: Adjust the threshold parameters of the first or second level trigger condition to improve the capture sensitivity of this type of anomaly mode; At this time, the CNN considers that the mode is suspicious or has high risk, triggers data recording for subsequent analysis or CNN model training; Generate a warning information to prompt the existence of potential power quality problems or equipment abnormalities. 2.The FPGA-based micro-grid power quality intelligent analysis method of claim 1, wherein, The interharmonic analysis comprises the following steps: S3.1.1: the amplitude correction and energy correction are carried out on the frequency spectrum signals corresponding to each voltage, current data respectively; S3.1.2: taking the integer multiple of the fundamental frequency as the center, the frequency spectrum lines between two harmonic frequencies are grouped, the root mean square values of each group are calculated to obtain the effective value of the interharmonic group; S3.1.3: through peak value search and frequency spectrum difference, the significant interharmonic components in the frequency spectrum are identified, and their frequency and amplitude are estimated; S3.1.4: the interharmonic distortion rate or the interharmonic distortion rate of a specific frequency band is calculated, which is compared with the known interharmonic frequency causing equipment failure to evaluate the influence of interharmonic on the microgrid. 3.The FPGA-based micro-grid power quality intelligent analysis method of claim 1, wherein, The method further comprises three-phase imbalance analysis, specifically comprising: The pretreated data is subjected to symmetrical component decomposition to extract symmetrical components, the symmetrical components include the amplitude and phase of positive sequence, negative sequence and zero sequence; According to the symmetrical components, the voltage unbalance degree and the current unbalance degree are calculated. 4.The FPGA-based micro-grid power quality intelligent analysis method of claim 1, wherein, The method further comprises adaptive analysis based on power quality parameters, specifically comprising: S4.1: evaluate the operation condition of the current micro-grid by monitoring power quality parameters, the power quality parameters at least including spectral entropy, RMS rate of change, load information and harmonic / interharmonic; S4.2: make disturbance type judgment: if transient characteristics are detected, analyze wavelet data and start high-speed recording; if harmonic or interharmonic anomaly is detected, increase FFT analysis frequency, harmonic and interharmonic analysis depth; if the condition is stable, make regular detection; S4.3: adjust the parameters of the trigger mechanism according to the condition and the disturbance type.

5. An FPGA-based micro-grid power quality intelligent analysis system, characterized in that, The system is configured to perform the method of claim 1, comprising: a data acquisition module for acquiring current and voltage data of the micro-grid; a preprocessing module for receiving and processing data of the data acquisition module; a power quality analysis module for analyzing and calculating the output data of the preprocessing module, at least including an FFT analysis unit, an RMS calculation unit, a wavelet transform feature extraction unit and a neural network anomaly detection unit, the FFT analysis unit, the RMS calculation unit, the wavelet transform feature extraction unit and the neural network anomaly detection unit being processed in parallel; a storage module for storing data of the data acquisition module, the preprocessing module and the power quality analysis module. 6.The FPGA-based micro-grid power quality intelligent analysis system of claim 5, wherein, The data acquisition module uses an 8-channel synchronous sampling ADC to acquire three-phase voltage, three-phase current, one neutral line current and one neutral point to ground voltage, and the front end of the channel is configured with an analog filter; The preprocessing module receives ADC sampling data output by the data acquisition module, and outputs preprocessed data after digital filtering, data format conversion, decimation, denoising and normalization of the ADC sampling data, and the preprocessed data is input into the power quality analysis module. 7.The FPGA-based micro-grid power quality intelligent analysis system of claim 6, wherein, The FFT analysis unit processes 8-channel preprocessed data, and the FFT analysis unit includes a high-precision mode and a fast update mode, the high-precision mode outputs high-resolution spectral data, and the fast update mode outputs spectral amplitude and phase data corresponding to each channel data. 8.The FPGA-based micro-grid power quality intelligent analysis system of claim 7, wherein, The power quality analysis module further comprises: a harmonic analysis unit that obtains the high-resolution spectral data to obtain harmonic characteristics, the harmonic characteristics at least including total harmonic distortion rate of each channel and harmonic parameters of each order; an interharmonic analysis unit that obtains the high-resolution spectral data to perform interharmonic grouping and interharmonic parameter calculation to obtain interharmonic characteristics; a voltage fluctuation and flicker calculation unit that calculates short-time flicker and long-time flicker according to the voltage RMS value sequence obtained by the RMS calculation unit. 9.The FPGA-based micro-grid power quality intelligent analysis system of claim 8, wherein, The neural network anomaly detection unit inputs a feature matrix, the feature matrix at least including the interharmonic characteristics, the harmonic characteristics and the time domain characteristics, outputs probabilities of multiple abnormal types, the abnormal types at least including harmonic overrun, voltage sudden change, fast transient and interharmonic anomaly; the neural network anomaly detection unit is trained using historical power quality parameters and corresponding abnormal types. 10.The FPGA-based micro-grid power quality intelligent analysis system of claim 8, wherein, It further comprises an adaptive analysis module, the adaptive analysis module comprises a condition evaluation model, uses power quality parameters to evaluate the operation condition of the micro-grid; according to the condition evaluation result, dynamically adjust the quality analysis parameters.

11. A computer-readable storage medium storing a computer program, characterized in that, The computer program comprises: instructions executable by a processor, which when executed by a processing unit of a system-on-chip platform, perform the steps of the microgrid power quality intelligent analysis method according to any one of claims 1 to 4; and / or configuration data for an FPGA, which when loaded into programmable logic cells of a field programmable gate array (FPGA), perform the steps of the method according to any one of claims 1 to 4.

Images