Electric energy quality data intelligent processing method based on AI large model and related equipment

Abstract

The embodiment of the invention provides an intelligent power quality data processing method based on an AI large model and related equipment, and belongs to the technical field of artificial intelligence. The method comprises the following steps: acquiring electric energy quality measurement data and an electric energy quality data processing instruction input by a target object; calling a preset intention recognition model to generate instruction semantic information based on the power quality data processing instruction; when the instruction semantic information indicates intelligent analysis of the electric energy quality measurement data, performing vectorization retrieval according to the instruction semantic information and the electric energy quality measurement data to obtain target retrieval content related to the instruction semantic information; calling a preset power quality evaluation model to generate power failure analysis data based on the power quality measurement data and the target retrieval content; and calling a preset multi-agent system to generate power quality analysis data based on the power quality measurement data and the power fault analysis data. According to the embodiment of the invention, the power quality analysis efficiency and accuracy can be improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to an intelligent processing method and related equipment for power quality data based on a large AI model. Background Technology

[0002] Renewable energy grid connection refers to connecting renewable energy sources (such as solar and wind power) to the traditional power grid, achieving interaction and coordination between renewable energy generation and the grid. Renewable energy sources such as wind and solar power are intermittent and fluctuating. Therefore, the connection of renewable energy sources may introduce new power quality problems, such as voltage flicker and harmonics. If these problems cannot be effectively detected and analyzed, they will threaten the stable operation of the power grid, thereby hindering the large-scale application of renewable energy in the power system.

[0003] Power quality analysis is a core component of power systems, involving the detection and analysis of power parameters, and is crucial for ensuring the stable operation of the power grid. Especially in the context of renewable energy grid integration, power quality analysis can monitor power parameters in real time after renewable energy grid integration, promptly identify and warn of potential power quality problems, such as voltage fluctuations, frequency deviations, and harmonic interference, to ensure the operational stability of the power grid under various conditions.

[0004] However, the power quality analysis methods in related technologies have technical problems such as relying on human experience to generate diagnostic results, lagging problem analysis, and slow equipment response, which cannot meet the requirements of modern power systems for real-time performance, accuracy, and intelligence. Summary of the Invention

[0005] The main objective of this application is to propose an intelligent power quality data processing method and related equipment based on an AI large model, aiming to improve the efficiency and accuracy of power quality analysis and meet the needs of modern power systems for real-time performance, accuracy, and intelligence.

[0006] To achieve the above objectives, a first aspect of this application proposes an intelligent power quality data processing method based on an AI large model, the method comprising: Acquire power quality measurement data and power quality data processing instructions input by the target object; The preset intent recognition model is invoked to generate instruction semantic information based on the power quality data processing instructions; When the instruction semantic information indicates that the power quality measurement data should be intelligently analyzed, a vectorized retrieval is performed based on the instruction semantic information and the power quality measurement data to obtain the target retrieval content related to the instruction semantic information; The preset power quality assessment model is invoked to generate power fault analysis data based on the power quality measurement data and the target search content; The preset multi-agent system is invoked to generate power quality analysis data based on the power quality measurement data and the power fault analysis data.

[0007] To achieve the above objectives, a second aspect of this application proposes an intelligent power quality data processing device based on an AI large model, the device comprising: The acquisition unit is used to acquire power quality measurement data and power quality data processing instructions input by the target object; The first generation unit is used to call a preset intent recognition model to generate instruction semantic information based on the power quality data processing instructions; The retrieval unit is used to perform vectorized retrieval based on the instruction semantic information and the power quality measurement data when the instruction semantic information indicates intelligent analysis of the power quality measurement data, so as to obtain target retrieval content related to the instruction semantic information; The second generation unit is used to call a preset power quality assessment model to generate power fault analysis data based on the power quality measurement data and the target search content; The third generation unit is used to call a preset multi-agent system to generate power quality analysis data based on the power quality measurement data and the power fault analysis data.

[0008] To achieve the above objectives, a third aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect.

[0009] To achieve the above objectives, a fourth aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.

[0010] This application proposes an intelligent power quality data processing method and related equipment based on an AI large-scale model. It generates instruction semantic information through an intent recognition model, accurately understanding the target object's intent and providing a clear semantic routing basis for subsequent processing. When the instruction semantic information instructs intelligent analysis of power quality measurement data, relevant knowledge is first retrieved to provide context for power quality analysis. Then, combining power quality measurement data and target retrieval content from multiple sources, power fault analysis data is generated, improving the accuracy of power quality diagnosis. Furthermore, through the collaborative integration of analysis results from various dimensions via a multi-agent system, a comprehensive and reliable power quality assessment conclusion is output, thereby achieving intelligent power quality diagnosis and improving the efficiency and accuracy of power quality analysis. Attached Figure Description

[0011] Figure 1 This is a flowchart of the intelligent power quality data processing method based on an AI large model provided in the embodiments of this application; Figure 2 This is a flowchart illustrating the process of generating the target response content provided in an embodiment of this application; Figure 3 This is another flowchart of the intelligent power quality data processing method provided in this application; Figure 4 This is a schematic diagram of the structure of the intelligent power quality data processing device based on an AI large model provided in the embodiments of this application; Figure 5 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0013] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0014] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0015] Renewable energy grid connection refers to connecting renewable energy sources (such as solar and wind power) to the traditional power grid, achieving interaction and coordination between renewable energy generation and the grid. Renewable energy sources such as wind and solar power are intermittent and fluctuating. Therefore, the connection of renewable energy sources may introduce new power quality problems, such as voltage flicker and harmonics. If these problems cannot be effectively detected and analyzed, they will threaten the stable operation of the power grid, thereby hindering the large-scale application of renewable energy in the power system.

[0016] Power quality analysis is a core component of power systems, involving the detection and analysis of power parameters, and is crucial for ensuring the stable operation of the power grid. Especially in the context of renewable energy grid integration, power quality analysis can monitor power parameters in real time after renewable energy grid integration, promptly identify and warn of potential power quality problems, such as voltage fluctuations, frequency deviations, and harmonic interference, to ensure the operational stability of the power grid under various conditions.

[0017] However, the power quality analysis methods and equipment in related technologies have many shortcomings. Specifically, the power parameter acquisition equipment in these technologies (such as IEC61000-4-30 Class A standard instruments) can only record some power parameters (such as voltage or current) simultaneously, and cannot achieve comprehensive acquisition of the power parameters required after the grid connection of new energy sources, such as the simultaneous acquisition of multi-dimensional indicators like harmonics, flicker, and three-phase imbalance. This leads to incomplete power quality assessment results, and subsequent analysis relies on human experience to generate diagnostic reports, which takes a long time (usually several hours to several days), failing to meet the real-time decision-making needs of modern power systems. In addition, the power quality analysis instruments in these technologies can only provide basic power parameter measurement functions. Power fault diagnosis and power quality assessment rely entirely on human experience, resulting in low accuracy in power quality analysis. For example, when a voltage drop or harmonic distortion is detected, the power quality analysis methods in these technologies cannot automatically generate corresponding power quality diagnostic reports or solutions. Professionals are required to manually analyze historical data and case databases to provide diagnostic reports or solutions, which is inefficient and inaccurate.

[0018] The power quality analysis equipment in related technologies lacks intelligent early warning mechanisms, making it unable to detect abnormal states of the power system in real time. For example, it cannot detect voltage interruptions or frequency deviations in a timely manner, thus failing to provide prompt solutions. When anomalies are detected, staff must interrupt their work processes and seek assistance from external experts, leading to delays in handling power system anomalies and impacting system reliability. Furthermore, the power quality analysis equipment in these technologies cannot provide real-time problem-solving or operational guidance. When staff encounter technical difficulties such as incorrect parameter settings or untraceable faults, they must consult manuals or seek remote assistance from power experts, resulting in low problem-solving efficiency. Finally, the power quality analysis equipment in these technologies lacks intelligent control functions (such as voice control or gesture control). In complex operating conditions (such as high-noise or high-altitude environments), staff must manually operate the interface of the power quality analysis equipment, which limits multi-task parallel processing, such as simultaneously detecting data and setting power parameters, leading to low work efficiency and increased operational error rates. In summary, the power quality analysis methods in these technologies cannot meet the real-time, accuracy, and intelligent requirements of modern power systems.

[0019] Based on this, embodiments of this application provide a method and related equipment for intelligent processing of power quality data based on AI large models, aiming to solve the above problems and improve the accuracy, real-time performance and intelligence of power quality analysis in modern power systems.

[0020] The intelligent power quality data processing method and related equipment based on AI large model provided in this application are specifically described through the following embodiments. First, the intelligent power quality data processing method based on AI large model in this application embodiment is described.

[0021] The intelligent power quality data processing method based on an AI large-scale model provided in this application relates to the field of artificial intelligence technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the intelligent power quality data processing method based on an AI large-scale model, but is not limited to the above forms.

[0022] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0023] Figure 1 This is an optional flowchart of the intelligent power quality data processing method based on an AI large model provided in the embodiments of this application. Figure 1 The method may include, but is not limited to, steps S101 to S105.

[0024] Step S101: Obtain power quality measurement data and power quality data processing instructions input by the target object; Step S102: Invoke the preset intent recognition model to generate instruction semantic information based on power quality data processing instructions; Step S103: When the instruction semantic information indicates that the power quality measurement data should be intelligently analyzed, a vectorized retrieval is performed based on the instruction semantic information and the power quality measurement data to obtain the target retrieval content related to the instruction semantic information. Step S104: Call the preset power quality assessment model to generate power fault analysis data based on power quality measurement data and target search content; Step S105: Call the preset multi-agent system to generate power quality analysis data based on power quality measurement data and power fault analysis data.

[0025] Steps S101 to S105, as illustrated in this embodiment, generate instruction semantic information through an intent recognition model to accurately understand the target object's intent and provide a clear semantic routing basis for subsequent processing. When the instruction semantic information instructs intelligent analysis of power quality measurement data, relevant knowledge is first retrieved to provide context for power quality analysis. Then, power fault analysis data is generated by combining power quality measurement data and multi-source information such as target retrieval content, improving the accuracy of power quality diagnosis. Furthermore, the analysis results from various dimensions are collaboratively integrated through a multi-agent system to output comprehensive and reliable power quality assessment conclusions, thereby realizing intelligent power quality diagnosis and improving the efficiency and accuracy of power quality analysis.

[0026] In step S101 of some embodiments, power quality measurement data is acquired by a power quality analysis device deployed on the end side. This power quality analysis device employs a synchronous 8-channel 16-bit analog-to-digital (ADC) architecture, acquiring power parameters in real time at a sampling rate of 100kHz to obtain power quality measurement data. This data may include multi-dimensional parameters such as voltage, current, frequency, harmonics (51st harmonic analysis, with total harmonic distortion (THD) between -0.2% and 0.2%), flicker, and three-phase imbalance. The power quality analysis device provided in this application achieves a voltage acquisition accuracy of ±0.05% and a current acquisition accuracy of ±0.1%. The acquired power quality measurement data can then be digitally filtered using a 128th-order finite impulse response (FIR) filter to ensure data quality. The signal-to-noise ratio of the digital filter is greater than 80dB to ensure that the time delay deviation between multi-channel acquired data does not exceed 1 microsecond. Thus, the embodiments of this application can solve the technical problem of incomplete power parameter acquisition in related technologies by improving hardware accuracy and adopting real-time algorithms to achieve synchronous recording of the acquired power parameters.

[0027] Furthermore, this application embodiment can also acquire power quality data processing instructions input by the target object, where the target object can be a power system maintenance personnel or staff member, and the power quality data processing instructions can be voice instructions or text instructions. The power quality analysis device can integrate a text input module and a voice recognition module to collect the voice or text instructions input by the target object. For example, the voice instructions can be instructions such as "start measuring voltage" or "start line detection," and then the power quality data processing instructions can be semantically recognized, and the corresponding power quality data processing tasks can be executed according to the recognized semantic results.

[0028] Power quality data processing instructions and power quality measurement data together constitute the input data for subsequent calls to the Artificial Intelligence (AI) large model to perform power quality data analysis tasks. This improves the real-time and comprehensiveness of power parameter acquisition. Combined with multimodal input instructions, it provides high-precision and complete data for subsequent intelligent analysis tasks of the model.

[0029] In step S102 of some embodiments, a preset intent recognition model can be invoked to perform semantic parsing on the power quality data processing instructions. When the input power quality data processing instruction is a voice instruction, the voice instruction can be converted to text before semantic recognition. The intent recognition model can adopt a two-level classification architecture: first, the power quality data processing instructions are classified at a coarse-grained level to obtain the semantic type (i.e., intent type); then, fine-grained intent parsing is performed to obtain the specific instruction semantic information. The instruction semantic type can include: control operation type, product usage type, other practical power quality scenario problems, and power quality-related knowledge type problems. Control operation commands are used to control the power quality analysis equipment to perform corresponding operations, such as: starting voltage measurement, opening the flicker interface, stopping recording, and detecting the line. Product usage commands are used to ask questions about how to use the product (e.g., the power quality analysis equipment), such as: how to measure harmonics, how to record data, and how to set parameters. Other practical power quality scenario questions are used to address various problems encountered in actual power quality analysis scenarios, such as: how to detect and locate voltage sag sources in the distribution network, and how to assess the impact of renewable energy grid connection on power quality at the point of common coupling. Power quality related knowledge questions refer to questions raised by the target audience regarding the basic theories and technical standards in the field of power quality, such as: please explain the definition of THD, and why renewable energy grid connection causes power quality problems. It should be noted that the above examples do not constitute a limitation of this application.

[0030] Instruction semantic information includes specific operational content and operational objects for different instruction semantic types. For example, for control operation instructions, the parsed instruction semantic information may include operation task codes, operation parameters, and operational objects. The intent recognition model provided in this application achieves a problem classification accuracy of 99.2% and an intent recognition F1 score of 97.5%. This method of recognizing instruction semantics through a model improves the accuracy of semantic recognition of power quality data processing instructions, thereby improving the accuracy of subsequent power quality data analysis.

[0031] In some embodiments, the method provided in this application further includes the following steps: When the instruction semantic information indicates that the power parameter acquisition equipment should be controlled, the instruction semantic information is parsed to obtain the equipment operation control information; Based on the equipment operation control information, the target power parameter acquisition unit is determined from multiple power parameter acquisition units, and the target power parameter acquisition unit is called to perform power parameter detection to obtain the target power parameters.

[0032] In this embodiment, when the instruction semantic information indicates control of the power parameter acquisition device, the instruction semantic information is parsed to obtain structured device operation control information. The power parameter acquisition device can be the aforementioned power quality analysis device, comprising multiple power parameter acquisition units. Each power parameter acquisition unit can be used to acquire a specific type of power parameter or power parameters at a specific circuit location, such as a voltage acquisition unit, a current acquisition unit, and a harmonic analysis unit. Different power parameter acquisition units can correspond to different ADC channels or signal processing modules in hardware. The device operation control information can be used to drive the power parameter acquisition device to acquire power parameters. The device operation control information can include specific task execution instruction codes, acquisition objects, operation content (e.g., measurement, detection, or recording operations), and operation parameters (e.g., sampling frequency).

[0033] Then, based on the equipment operation control information, the target power parameter acquisition unit can be determined from multiple power parameter acquisition units. Specifically, it can be selected based on the task execution instruction code, operation parameters, acquisition object, and circuit location information in the equipment operation control information. For example, when the equipment operation control information indicates that the B-phase voltage is to be measured, the unit for acquiring the B-phase voltage can be accurately located from multiple power parameter acquisition units. Furthermore, the target power parameter acquisition unit can be invoked to perform power parameter detection to obtain the target power parameters. After the power parameter acquisition indicated by the instruction semantic information is completed, the complete target power parameters are output.

[0034] For example, parsing the semantic information of the instruction yields a task execution instruction code of 1001, where the object to be collected is voltage, and the operation content is measurement. Further, the instruction code can be sent to a power parameter acquisition device to control the device to call the voltage measurement module to perform voltage measurement and output the measured real-time voltage value and the corresponding acquisition time to the target object. It is understood that the above examples do not constitute a limitation on the embodiments of this application.

[0035] Thus, this embodiment controls the power quality analysis equipment based on the semantic recognition results of power quality data processing instructions, improving the intelligence level of human-computer interaction. This allows the target object to complete the power parameter acquisition task through simple and intuitive operation, thereby improving the efficiency of power quality detection. Simultaneously, by accurately locating and calling specific power parameter acquisition units, the accuracy of parameter acquisition is improved.

[0036] In some embodiments, the method provided in this application further includes the following steps: When the instruction semantic information indicates that power quality related knowledge should be obtained, power quality related knowledge is retrieved in the corresponding preset knowledge base according to the instruction semantic information to obtain power quality related knowledge. Get the preset prompt word template; Based on the prompt word template, a preset large language model is invoked to generate the target response content based on power quality correlation knowledge and instruction semantic information.

[0037] In this embodiment, when the instruction semantic information indicates the retrieval of power quality-related knowledge, the intent type is knowledge query. Power quality-related knowledge can be retrieved from the corresponding preset knowledge base based on the instruction semantic information. The preset knowledge base is a pre-constructed power quality professional knowledge base, containing a power quality industry knowledge graph and a fault case library. The knowledge graph covers multiple standards and contains millions of entity nodes, while the fault case library records multiple fault cases and various fault modes from historical data. Power quality-related knowledge refers to knowledge fragments retrieved from the knowledge base through vectorized retrieval, which may include concept definitions, standard clauses, and technical principles. For example, when the instruction semantic information instructs the target object to inquire about the national standard limit for voltage sag, a search can be performed in the knowledge base based on this instruction semantic information, returning specific clauses regarding voltage sag limits and their technical explanations.

[0038] It is understandable that regardless of whether the power quality data processing command input by the target object is a text command or a voice command, it can be converted into corresponding text data and then vectorized for retrieval. This application embodiment uses retrieval enhancement technology to retrieve power quality-related knowledge. Specifically, the acge_text_embedding vector model converts the text data corresponding to the command semantic information into 1536-dimensional text vector data. Then, based on the converted vector data, a search is performed in the pgvector vector library (capable of large-scale retrieval in a short time) to find the vector with the highest similarity to the text vector data corresponding to the command semantic information, and the target retrieval content corresponding to that vector is returned.

[0039] During a search, the search can be routed to the corresponding knowledge base based on the semantic information of the instruction. For example, the search path can be determined based on the identified intent type. If the intent type is related to product usage, the search will proceed to the knowledge base corresponding to the product manual or operating guide. If the intent type is related to real-world power quality issues, the search will proceed to the fault case library to find similar historical solutions. If the intent type is related to power quality knowledge, the search will proceed to the industry knowledge graph library to find theoretical knowledge and standard provisions. Thus, this embodiment of the application enables multi-database routing, improving the recall rate of the search content.

[0040] Then, the retrieved target search content can be integrated and optimized using a pre-defined large language model, and the integrated and optimized results can be returned to the target audience. The large language model refers to a professional language model that has been fine-tuned with knowledge of the power quality domain, possessing the ability to understand power quality professional knowledge and reason. The target response content refers to the accurate and complete response content generated by the large language model under the guidance of prompt word templates, integrating retrieved power quality-related knowledge and the semantics of the original question.

[0041] Specifically, preset prompt word templates can be obtained. These templates may include role settings, output formats, and task requirements, ensuring the accuracy and standardization of the content generated by the large language model. Further, based on the prompt word templates, target retrieval content, and instruction semantic information, the large language model is invoked to generate target response content using power quality association knowledge and instruction semantic information.

[0042] This application's embodiments construct a power quality knowledge question-answering mechanism, solving the technical problem of low efficiency caused by manual knowledge retrieval in related technologies. By combining retrieval enhancement technology with a large language model, the accuracy and readability of the response content can be ensured, improving the efficiency of power quality knowledge acquisition.

[0043] In some embodiments, please refer to Figure 2 , Figure 2 This is a flowchart illustrating the process of generating the target response content provided in this application. Specifically, the application layer of the power quality analysis device can receive power quality data processing instructions (e.g., asking a question) input by the target object. If it is a voice instruction, it will first be converted to text. Then, the semantic type of the voice instruction can be determined through an intent recognition model. Figure 2 As shown, semantic types can include control operation commands (such as start measurement) and knowledge-based question-and-answer commands. Knowledge-based question-and-answer commands include product usage questions, power quality-related questions, timeliness issues, assistant identity-related questions, and general questions. Different types of commands can generate corresponding domain-specific responses by calling the corresponding AI large-scale model, or be directly sent to the control system to control equipment for parameter acquisition and other operations. Each AI large-scale model can be used to solve problems in a specific domain or dimension and generate corresponding small-scale model analysis conclusions. Specifically, AI large-scale models can include equipment fault diagnosis models, arc detection models, equipment health assessment models, and power image recognition models, etc. It is understood that the above examples do not limit the embodiments of this application.

[0044] For knowledge-based question-and-answer commands, the data needed to generate the response can be retrieved from the knowledge base first, thus entering the vector retrieval process. Relevant content can be retrieved from the instrumentation industry knowledge base. Specifically, the question data of the target object can be semantically vectorized first; for voice question data, it can be converted into text data before semantic vectorization. Furthermore, the Euclidean distance algorithm can be used to find the text blocks with the highest similarity in the vector library. The vector library is generated based on the instrumentation industry knowledge base, which contains numerous source files in formats such as PDF, HTML, or TXT. These include instrumentation product manuals, industry standards, industry patent documents, published industry papers, industry journal articles, industry academic reports, and industry question-and-answer data. An unstructured loader can be used to load these source files, and then the original text data can be extracted. Further, a text splitter can be used to cut long texts into appropriately sized text chunks for subsequent processing and retrieval. The text blocks can then be vectorized using an embedding model. Furthermore, the vectorized text blocks are stored in a vector library and mapped to the original text data, thus generating a vector library for vector retrieval.

[0045] Following the steps outlined above, after finding the content corresponding to the most similar vector in the vector library, these contents can be used for summary retrieval and comprehensive sorting of the retrieval content. Finally, the most relevant information fragments are selected as the retrieval recall content.

[0046] Furthermore, the user's (i.e., the target audience's) question and the retrieved content can be populated into a preset prompt word template. The integrated prompt words and the small model's analysis conclusions form a complete context. Then, the integrated prompt words and the small model's analysis conclusions can be input into a large language model. The large language model outputs the target response content and delivers it to the target audience at the application layer.

[0047] In step S103 of some embodiments, when the instruction semantic information indicates that intelligent analysis of power quality measurement data is required, vectorized retrieval can be performed first based on the instruction semantic information and power quality measurement data. It is understood that the knowledge retrieval described above is a retrieval operation performed by the target object as long as it obtains the answer to the relevant question. However, the embodiments of this application require diagnosing power quality, generating a diagnostic report and providing a solution based on power quality measurement data. Before this, relevant knowledge needs to be retrieved to provide data support for generating the diagnostic report and solution.

[0048] Specifically, the acge_text_embedding vector model can be used to convert the current power quality measurement data and command semantic information into vector representations, then perform similarity matching in the pgvector vector database, and finally return the target retrieval content corresponding to the vector with the highest similarity.

[0049] In step S104 of some embodiments, the power fault analysis data is an intermediate diagnostic result generated by fusing and analyzing real-time collected power quality measurement data and target retrieval content through a power quality assessment model. The power fault analysis data may include a feature parameter set, fault warning information, problem location data, and preliminary diagnostic results. The feature parameter set can be obtained by feature extraction from power quality measurement data and target retrieval content, and may include the following core feature parameter values: voltage imbalance, frequency deviation, harmonic distortion rate, voltage sag amplitude, and flicker index, etc. These parameters are output in numerical form (JSON format) for subsequent analysis. The fault warning information is an abnormal state information of the power system determined based on the feature parameters, and may include a warning identifier, potential fault type (e.g., voltage sag and frequency fluctuation), probability score corresponding to the fault type (e.g., voltage sag prediction accuracy of 96.7%, probability score of 0.967), and severity level (high / medium / low risk), etc. The problem location data may include the specific location of the power fault (e.g., grid node ID) and timestamp. The preliminary diagnostic results include a preliminary cause analysis corresponding to the abnormal state of the power system.

[0050] In some embodiments, a preset power quality assessment model is invoked to generate power fault analysis data based on power quality measurement data and target search content, including the following steps: The power quality measurement data and the target search content are fused to obtain the target fused data. Multi-scale feature extraction is performed on the target fusion data to obtain power quality feature data; The system calls a preset power quality assessment model to generate power fault analysis data based on power quality characteristic data and target search content.

[0051] In this embodiment, the power quality measurement data and the target search content can first be normalized and denoised. Denoising can be achieved through sliding window filtering. Then, the power quality measurement data and the target search content can be fused to obtain target fused data, which can be achieved through timestamp alignment and feature association mapping. Further, multi-scale feature extraction is performed on the target fused data to obtain power quality feature data. Multi-scale feature extraction refers to analyzing power quality data simultaneously from different time scales and frequency dimensions.

[0052] In some embodiments, multi-scale feature extraction is performed on the target fusion data to obtain power quality feature data, including the following steps: Continuous wavelet transform is performed on the target fusion data to obtain time-frequency joint domain feature information; The target fusion data is subjected to a fast Fourier transform to obtain frequency domain feature information; Power quality characteristic data are obtained by performing feature calculations based on the joint time-frequency domain feature information and the frequency domain feature information.

[0053] In this embodiment, a continuous wavelet transform model can be invoked to perform continuous wavelet transform on the target fused data. Specifically, sampling is performed in 0.5ms time windows, with 1024 sampling points acquired in each time window. Time-series signals such as voltage and current are decomposed into different frequency sub-bands (e.g., high-frequency transients and low-frequency steady states) to capture the time-frequency characteristics of transients. For example, when analyzing voltage sag events, continuous wavelet transform can accurately identify the start time, duration, and frequency component changes of the voltage drop, outputting a coefficient matrix containing three-dimensional information of time, frequency, and amplitude, i.e., time-frequency joint domain feature information.

[0054] Then, a Fast Fourier Transform (FFT) can be performed on the target fused data to obtain frequency domain feature information. This frequency domain feature information can include each harmonic component, specifically the amplitude and phase of each harmonic. Specifically, steady-state power quality measurement data can be analyzed using a frequency resolution of 0.1%, meaning that a minute frequency change of 0.05 Hz can be detected at a fundamental frequency of 50 Hz.

[0055] Furthermore, power quality characteristic data can be obtained by performing feature calculations based on time-frequency joint domain feature information and frequency domain feature information. Power quality characteristic data is a set of feature parameters, including voltage imbalance, frequency deviation, and harmonic distortion rate. Feature calculations are performed based on the output results of wavelet transform and Fourier transform to obtain the specific values ​​of the aforementioned feature parameters. For example, the duration of voltage sag can be calculated based on the time-frequency coefficients obtained from wavelet transform, and the total harmonic distortion rate can be calculated based on the spectral data obtained from Fourier transform. Another example is the voltage imbalance formula: |(V_max-V_min) / V_avg|×100%, where V_max refers to the maximum voltage value among the three-phase voltages (phase A, phase B, and phase C), V_min refers to the minimum voltage value among the three-phase voltages, and V_avg refers to the arithmetic mean of the three-phase voltages. Finally, power quality characteristic data containing specific values ​​of various features is output, and the output frequency can be set to be once every 0.5ms. The embodiments of this application deeply integrate real-time power quality measurement data with domain expertise, thereby improving the accuracy and timeliness of fault diagnosis.

[0056] In some embodiments, the power quality assessment model includes a time-series prediction network, which calls a preset power quality assessment model to generate power fault analysis data based on power quality characteristic data and target search content, including the following steps: The time-series prediction network is invoked to generate power fault trend prediction data based on power quality characteristic data and target search content; Power fault analysis data is generated based on power fault trend prediction data and power quality characteristic data.

[0057] In this embodiment, a power quality assessment model can be invoked to generate power fault analysis data based on power quality feature data and target search content. The power quality assessment model includes a time-series prediction network based on the Transformer-XL architecture. This network can perform trend prediction based on time feature sequences, specifically generating power fault trend prediction data based on power quality feature data and target search content. For example, when a voltage imbalance exceeding a 2% threshold is detected, the power quality assessment model can predict the probability of equipment failure within the next 10 seconds and provide a corresponding warning.

[0058] Specifically, the time-series prediction network includes a self-attention layer, which calculates the relationships between different features (such as the co-variation pattern of voltage imbalance and frequency deviation) and outputs a fused feature vector. The network can also perform location encoding on power quality feature data, providing timestamp information. This power quality feature data is a time-series sequence, including real-time feature values ​​of the aforementioned feature parameters, with a sampling interval of 0.5 ms. The target retrieval content can include historical fault patterns (e.g., voltage drop events from the past week) to provide contextual reference. The time-series prediction network can capture the temporal dependencies between different feature values ​​using this data, avoiding prediction biases caused by sequence length limitations in traditional models.

[0059] The time-series prediction network also includes a fully connected layer, which is used for regression prediction. This fully connected layer outputs power fault trend prediction data for a future time window (e.g., the next 10 seconds). The power fault trend prediction data refers to the predicted trend of power quality status changes over a future period, specifically including the probability of fault occurrence (e.g., voltage drop risk value), feature evolution curves (e.g., predicted frequency deviation), predicted probability distributions (e.g., a 96.7% probability of a voltage drop occurring within 5 seconds), and trend reports (which can be text or structured data). The trend report may include the predicted event type, time point, confidence interval, and feature change graph. For example, the time-series prediction network receives the current and historical voltage imbalance time series, combines it with fault development patterns under similar operating conditions, analyzes long-term dependencies through its segment-level recurrence mechanism and self-attention layer, outputs a 96.7% probability of a voltage drop occurring within the next 10 seconds, and provides the evolution curves of the feature parameters.

[0060] Furthermore, power fault analysis data can be generated based on power fault trend prediction data and power quality characteristic data, and the power fault analysis data is the generated structured diagnostic results.

[0061] This application embodiment mines feature data sequences through a time-series prediction network to predict the development trend of power faults in advance, providing a time window for proactive intervention and improving the early warning capability of the power system.

[0062] In step S105 of some embodiments, a pre-defined multi-agent system can be invoked to generate final power quality analysis data based on power quality measurement data and power fault analysis data. The multi-agent system employs a group-based dense and hybrid expert model architecture, with multiple agents working collaboratively, each responsible for analysis tasks in a specific domain. For example, a voltage characteristic analysis agent processes voltage-related parameters, and a frequency analysis agent handles frequency deviation issues. These agents reach a unified diagnostic conclusion through information sharing and voting mechanisms. The output power quality analysis data includes a complete diagnostic report, remediation recommendations, and preventative measures; for example, voltage imbalance leading to equipment risks suggests adjusting load distribution.

[0063] In some embodiments, the multi-agent system includes multiple agents, and the process of calling a preset multi-agent system to generate power quality analysis data based on power quality measurement data and power fault analysis data includes the following steps: Multiple intelligent agents are invoked to generate corresponding power quality analysis sub-results based on power quality measurement data and power fault analysis data, respectively. The power quality analysis results are weighted and fused to obtain power quality analysis data.

[0064] In this embodiment, an agent refers to multiple AI analysis modules employing a hybrid expert model architecture. Each agent focuses on fault analysis tasks in a specific domain. Each agent can generate corresponding power quality analysis sub-results based on power quality measurement data and power fault analysis data. These sub-results represent local diagnostic conclusions formed by each agent from its professional perspective. For example, a voltage analysis agent, based on voltage measurement data and related fault analysis data, analyzes events such as voltage dips and swells, generating a power quality analysis sub-result of "voltage dip amplitude 30%, mainly affecting precision instruments." Similarly, a harmonic analysis agent, based on current measurement data and harmonic characteristics, generates a power quality analysis sub-result of "5th harmonic exceeding standard, suspected to be caused by inverter load."

[0065] Then, the power quality analysis sub-results can be weighted and fused to obtain power quality analysis data. The weights used in the weighting and fusion process are determined based on the confidence level and historical accuracy of each agent's expertise. Power quality analysis sub-results generated by multiple agents are fused using a weighted average or voting mechanism to obtain the final power quality analysis data. The power quality analysis data refers to the final diagnostic report and solutions generated after fusion. For example, the weight of the power quality analysis sub-result corresponding to the voltage analysis agent is 0.6, the weight of the power quality analysis sub-result corresponding to the harmonic analysis agent is 0.3, and the weight of the power quality analysis sub-result corresponding to the frequency analysis agent is 0.1. Based on these weights, the fault type, fault cause analysis, and remediation suggestions of each power quality analysis sub-result are comprehensively calculated, ultimately generating a unified diagnostic conclusion and solution.

[0066] This application's embodiments construct an intelligent diagnostic mechanism based on multi-expert collaborative decision-making. Multiple specialized agents conduct in-depth analysis of complex power quality issues from different dimensions, avoiding the limitations of single-model analysis. A weighted fusion mechanism integrates professional judgments from various fields, fully leveraging the expertise of each agent, while the weight allocation reflects the varying importance of different pieces of evidence. Thus, this application's embodiments enhance the diagnostic capabilities of power systems in handling complex faults. When facing complex power quality analysis events involving multiple coupled factors, it can provide more comprehensive and accurate analysis results, thereby improving the reliability of diagnostic conclusions and solutions.

[0067] In some embodiments, the power quality analysis data includes a power quality diagnostic report and a corresponding power quality management plan. After the power quality analysis data is generated by calling a preset multi-agent system based on power quality measurement data and power fault analysis data, the following steps are also included: Push power quality diagnostic reports and power quality management solutions to the target audience; The system provides voice broadcasts of power quality diagnostic reports and power quality management solutions.

[0068] In this embodiment, the power quality diagnostic report refers to a comprehensive assessment document generated based on multi-agent system analysis, which may include current status assessment, problem location, severity analysis, and root cause of the fault. The power quality management solution refers to the specific solutions provided for the power fault problems identified in the diagnostic report, which may include equipment modification plans, parameter adjustment suggestions, and preventive measures. For example, when detecting a harmonic exceedance problem, a diagnostic report containing "5th harmonic content reaches 8.7%, exceeding the national standard limit" can be pushed to the target object, along with a solution "It is recommended to install a 5th harmonic filter and adjust the capacitor bank configuration." This content can be pushed through various channels such as device displays, mobile terminals, or web interfaces.

[0069] This application embodiment can integrate a speech synthesis module to broadcast power quality diagnostic reports and power quality management solutions to the target audience via voice announcements. For example, upon detecting a voltage sag, in addition to pushing a text report, it will also send the following voice prompt to the target audience: "Warning, a voltage sag in phase A has been detected, with an amplitude of 30% and a duration of 200 milliseconds. It is recommended to check the starting circuit of large motors." Voice announcements can adapt to noisy industrial environments or work scenarios requiring hands-free operation, ensuring that critical information is conveyed to on-site personnel in a timely and accurate manner.

[0070] This application's embodiments ensure information integrity through multimodal output, improving information transmission efficiency and user experience. Text push ensures information accuracy and traceability, facilitating subsequent analysis and archiving; voice broadcast provides an instant and convenient way to obtain information, ensuring personnel safety and operational efficiency under complex operating conditions. It effectively solves the problems of delayed report generation and limited information transmission in related power quality analysis methods, enabling on-site personnel to quickly obtain diagnostic results and take timely remedial measures, thereby improving the real-time nature and effectiveness of power quality management.

[0071] In some embodiments, please refer to Figure 3 , Figure 3This is another flowchart of the intelligent power quality data processing method provided in this application. In this method, the end-side detection equipment, such as the aforementioned power quality analysis equipment, is deployed at the power system site and can be used for human-computer interaction and data acquisition. The instrumentation AI big data model is the core AI analysis engine deployed in the cloud, used to receive data collected by the end-side detection equipment and process deeper professional issues in the instrumentation field. The power data analysis platform is a software platform for overall management, responsible for receiving, synthesizing, storing, and displaying the results generated in the cloud. Specifically, the end-side detection equipment 301 can receive a voice command from the target object instructing it to "generate a report." Then, the end-side detection equipment 302 queries historical or real-time power quality measurement data in its local database and sends the generated power quality measurement data to the power data analysis platform 301 and the instrumentation AI big data model 303. The instrumentation AI model 303 generates a preliminary diagnostic conclusion on power quality based on power quality measurement data and sends it to the power data analysis platform 301. The power data analysis platform 301 generates a power quality analysis conclusion based on the preliminary diagnostic conclusion and the power quality measurement data, and sends it to the end-side detection device 302. The end-side detection device 302 receives the power quality analysis conclusion and transmits it to the instrumentation AI model 303. The instrumentation AI model 303 further analyzes, processes, and packages the power quality analysis conclusion, and sends the processed results back to the power data analysis platform 301. The power data analysis platform 301 generates a data analysis image based on the processed results, and then generates a power quality analysis report based on the data analysis image, power quality measurement data, and processed results. Finally, the power data analysis platform 301 sends the generated power quality analysis report to the end-side detection device 302, which stores and displays the power quality analysis report.

[0072] Please see Figure 4 This application also provides an intelligent power quality data processing device based on an AI large model, which can realize the above-mentioned intelligent power quality data processing method based on an AI large model. The device includes: The acquisition unit 401 is used to acquire power quality measurement data and power quality data processing instructions input by the target object; The first generation unit 402 is used to call a preset intent recognition model to generate instruction semantic information based on power quality data processing instructions; The retrieval unit 403 is used to perform vectorized retrieval based on the instruction semantic information and the power quality measurement data when the instruction semantic information indicates that intelligent analysis of power quality measurement data is to be performed, so as to obtain target retrieval content related to the instruction semantic information. The second generation unit 404 is used to call a preset power quality assessment model to generate power fault analysis data based on power quality measurement data and target search content; The third generation unit 405 is used to call a preset multi-agent system to generate power quality analysis data based on power quality measurement data and power fault analysis data.

[0073] The specific implementation of the AI-based large model-based intelligent power quality data processing device is basically the same as the specific implementation of the AI-based large model-based intelligent power quality data processing method described above, and will not be repeated here.

[0074] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned intelligent power quality data processing method based on an AI large model. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0075] Please see Figure 5 , Figure 5 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 501 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 502 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 502 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 502 and is called and executed by the processor 501 to implement the intelligent power quality data processing method based on an AI large model according to the embodiments of this application. The input / output interface 503 is used to implement information input and output; The communication interface 504 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 505 transmits information between various components of the device (e.g., processor 501, memory 502, input / output interface 503, and communication interface 504); The processor 501, memory 502, input / output interface 503, and communication interface 504 are connected to each other within the device via bus 505.

[0076] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent power quality data processing method based on an AI large model.

[0077] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0078] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0079] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0080] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0081] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0082] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0083] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0084] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0085] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0086] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0087] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0088] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. An AI large model-based power quality data intelligent processing method, characterized in that, The method comprises: acquiring power quality measurement data and power quality data processing instructions input by a target object; calling a preset intention recognition model to generate instruction semantic information based on the power quality data processing instructions; when the instruction semantic information indicates intelligent analysis of the power quality measurement data, performing vectorization retrieval according to the instruction semantic information and the power quality measurement data to obtain target retrieval content related to the instruction semantic information; calling a preset power quality evaluation model to generate power failure analysis data based on the power quality measurement data and the target retrieval content; calling a preset multi-agent system to generate power quality analysis data based on the power quality measurement data and the power failure analysis data.

2. The method of claim 1, wherein, The calling of the preset power quality evaluation model to generate power failure analysis data based on the power quality measurement data and the target retrieval content comprises: performing data fusion on the power quality measurement data and the target retrieval content to obtain target fusion data; performing multi-scale feature extraction on the target fusion data to obtain power quality feature data; calling a preset power quality evaluation model to generate power failure analysis data based on the power quality feature data and the target retrieval content.

3. The method of claim 2, wherein, The power quality evaluation model comprises a time series prediction network, and the calling of the preset power quality evaluation model to generate power failure analysis data based on the power quality feature data and the target retrieval content comprises: calling the time series prediction network to generate power failure trend prediction data based on the power quality feature data and the target retrieval content; generating power failure analysis data based on the power failure trend prediction data and the power quality feature data.

4. The method of claim 2, wherein, The multi-scale feature extraction on the target fusion data to obtain power quality feature data comprises: performing continuous wavelet transform on the target fusion data to obtain time-frequency joint domain feature information; performing fast Fourier transform on the target fusion data to obtain frequency domain feature information; performing feature calculation based on the time-frequency joint domain feature information and the frequency domain feature information to obtain power quality feature data.

5. The method of claim 1, wherein, The multi-agent system comprises a plurality of agents, and the calling of the preset multi-agent system to generate power quality analysis data based on the power quality measurement data and the power failure analysis data comprises: calling the plurality of agents to generate corresponding power quality analysis sub-results based on the power quality measurement data and the power failure analysis data; performing weighted fusion on the power quality analysis sub-results to obtain power quality analysis data.

6. The method of claim 1, wherein, The method further comprises: when the instruction semantic information indicates control of a power parameter acquisition device, analyzing the instruction semantic information to obtain device operation control information, the power parameter acquisition device comprising a plurality of power parameter acquisition units; determining a target power parameter acquisition unit in the plurality of power parameter acquisition units based on the device operation control information, and calling the target power parameter acquisition unit to perform power parameter detection to obtain target power parameters.

7. The method of claim 1, wherein, The method further comprises: When the instruction semantic information indicates obtaining power quality correlation knowledge, power quality correlation knowledge retrieval is performed in a corresponding preset knowledge base according to the instruction semantic information, and power quality correlation knowledge is obtained; A preset prompt word template is obtained; According to the prompt word template, a preset large language model is called to generate target reply content based on the power quality correlation knowledge and the instruction semantic information.

8. The method of claim 1, wherein, The power quality analysis data includes power quality diagnosis reports and corresponding power quality treatment schemes, and after the preset multi-agent system is called to generate power quality analysis data based on the power quality measurement data and the power failure analysis data, the method further includes: The power quality diagnosis reports and the power quality treatment schemes are pushed to the target object; The power quality diagnosis reports and the power quality treatment schemes are audibly broadcasted.

9. An electronic device, comprising: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the AI large model-based power quality data intelligent processing method of any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the AI large model-based power quality data intelligent processing method of any one of claims 1 to 8.

Images