Rural distributed energy management system

Abstract

The invention discloses a rural distributed energy management system. The system comprises a sensing layer, an edge layer, a communication layer and a platform layer, the sensing layer is used for collecting real-time multi-source data, and the multi-source data comprises any multiple of electric parameter data, meteorological data or environmental data; the edge layer is used for carrying out preprocessing, anomaly detection and autonomous control on the multi-source data; the communication layer is used for guaranteeing stable transmission of data to be transmitted among the layers, and the communication layer comprises at least two communication modes; the communication mode comprises LoRa, NB-IoT (Narrow Band Internet of Things) or 4G / 5G; and the platform layer is used for performing predictive maintenance and global energy scheduling according to the multi-source data and a preset algorithm. According to the embodiment of the invention, effectiveness and stability can be improved. The method can be widely applied to the technical field of energy management.

Description

Technical Field

[0001] This application relates to the field of energy management technology, and in particular to a management system for rural distributed energy. Background Technology

[0002] With the transformation and upgrading of rural energy structures and the widespread application of distributed energy, rural power systems are gradually showing characteristics of diversification, decentralization, and a high proportion of new energy integration. However, due to the small scale of rural power grids, weak infrastructure, and complex communication conditions, existing energy systems still face prominent problems in some aspects, such as the effectiveness and stability of fault early warning and proactive prevention. Summary of the Invention

[0003] The main objective of this application is to propose a management system for rural distributed energy, aiming to improve its effectiveness and stability.

[0004] To achieve the above objectives, this application proposes a management system for rural distributed energy, comprising a sensing layer, an edge layer, a communication layer, and a platform layer; wherein, The sensing layer is used to collect real-time multi-source data, which includes any combination of electrical parameter data, meteorological data, or environmental data. The edge layer is used for preprocessing, anomaly detection, and autonomous control of the multi-source data; The communication layer is used to ensure stable transmission of data between layers, and the communication layer includes at least two communication methods; the communication methods include LoRa, NB-IoT or 4G / 5G. The platform layer is used for predictive maintenance and global energy scheduling based on the multi-source data and preset algorithms.

[0005] In some embodiments, the preprocessing, anomaly detection, and autonomous control of the multi-source data include: Redundant data is removed from the multi-source data; Anomaly fluctuations are detected in the preprocessed multi-source data using a time series and threshold model. If the anomaly fluctuations do not exceed a preset fluctuation range, the system automatically corrects the data and maintains autonomous operation. If the anomaly fluctuations exceed the preset fluctuation range, an alarm is triggered, and the data of the anomaly fluctuations is sent to the platform layer for further analysis.

[0006] In some embodiments, the process of removing redundant data from the multi-source data includes: Based on preset normal ranges and rates of change, obvious abnormal and noisy data are removed; And / or, median filtering and moving average algorithms are used to smooth the multi-source data, eliminating instantaneous interference and noise data.

[0007] In some embodiments, the abnormal fluctuation detection of preprocessed multi-source data using a time series and threshold model includes: Based on the preprocessed multi-source data, a normal distribution model of key parameters is established using a sliding window, and the moving average and standard deviation are calculated. When the collected data points deviate from the threshold range of the mean or standard deviation, they are identified as abnormal fluctuations.

[0008] In some embodiments, the communication layer includes a primary transmission link and a backup transmission link, and ensuring stable transmission of data to be transmitted between layers includes: The primary and backup transmission links are determined based on the performance parameters of various communication methods. These performance parameters include any one or more of the following: anti-interference capability, coverage, bandwidth, or energy consumption. The data level is determined based on the data to be transmitted. If the data level is level one, multipath transmission is performed using the primary transmission link and the backup transmission link. If the data level is level two, the primary transmission link is used for transmission. If the delay of the primary transmission link exceeds the preset delay or terminal, the system automatically switches to the backup transmission link for transmission.

[0009] In some embodiments, the predictive maintenance and global energy scheduling based on the multi-source data and a preset algorithm includes: Feature extraction and pattern recognition are performed on the multi-source data, and machine learning models are used to predict the probability of equipment failure. By combining scheduling algorithms with load forecasting, the coordinated operation of distributed energy resources and energy storage systems can be achieved.

[0010] In some embodiments, feature extraction from the multi-source data includes: Based on the timestamps of the multi-source data, a time-period characteristic curve of load and power generation output is constructed. The key events of the time period characteristic curve are marked and feature vectors are extracted; the feature vectors include the time of occurrence, duration, or fluctuation amplitude.

[0011] In some embodiments, pattern recognition is performed in the following manner: Based on clustering algorithms, the operating status of the multi-source data is divided into several classic modes, including normal, light overload overvoltage, or heavy overload low voltage. By associating the classic patterns with related factors, we obtain the association patterns.

[0012] In some embodiments, predicting the probability of device failure using a machine learning model includes: The extracted feature vectors and the identified patterns are input into the trained machine learning model to obtain the probability of equipment failure.

[0013] In some embodiments, the system includes several types of sensing sensors, which are used to collect the multi-source data. The sensing sensors integrate a GNSS timing module and are supplemented by an NTP network time synchronization mechanism to form a dual time synchronization system, so that the timestamp error of the collected multi-source data is controlled within a preset time range.

[0014] The embodiments of this application include at least the following beneficial effects: The management system for rural distributed energy in the embodiments of this application includes a sensing layer, an edge layer, a communication layer, and a platform layer; the sensing layer is used to collect real-time multi-source data, which includes any combination of electrical parameter data, meteorological data, or environmental data, thereby improving the effectiveness of the data; the edge layer is used to preprocess, detect anomalies, and perform autonomous control on the multi-source data, thereby achieving multi-dimensional monitoring, anomaly identification, and risk prediction of the energy system; the communication layer is used to ensure stable transmission of data to be transmitted between layers, and includes at least two communication methods, including LoRa, NB-IoT, or 4G / 5G, thereby increasing the stability of data transmission; the platform layer is used to perform predictive maintenance and global energy scheduling based on multi-source data and preset algorithms, thereby solving the coordination problem of the entire chain of "sensing-edge-communication-platform" and improving reliability and real-time performance. Attached Figure Description

[0015] Figure 1 This is a structural block diagram of a management system for rural distributed energy provided in an embodiment of this application; Figure 2 This is a data processing flowchart of a management system for rural distributed energy provided in an embodiment of this application. Detailed Implementation

[0016] 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 of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0017] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0018] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0019] 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.

[0020] See Figure 1 This application proposes a management system for rural distributed energy, comprising a sensing layer, an edge layer, a communication layer, and a platform layer; wherein, The perception layer is used to collect real-time multi-source data, which includes any combination of electrical parameter data, meteorological data, or environmental data. Edge layer, used for preprocessing multi-source data, anomaly detection, and autonomous control; The communication layer is used to ensure stable transmission of data between layers. The communication layer includes at least two communication methods, including LoRa, NB-IoT, or 4G / 5G. The platform layer is used for predictive maintenance and global energy scheduling based on multi-source data and preset algorithms.

[0021] The management system for rural distributed energy collects electrical parameters (such as voltage, current, active power, reactive power, frequency, etc.) and meteorological environmental data (such as temperature, humidity, light intensity, wind speed, etc.) or environmental data from multiple sources at the sensing layer, performs local autonomy and data filtering at the edge layer, provides remote interconnection at the communication layer, and completes centralized management at the cloud platform layer, so as to achieve multi-level division of labor and collaboration.

[0022] While performing data cleaning and feature extraction, the edge layer supports outlier removal and rapid pattern recognition. In the event of communication interruption, it performs local adjustments to energy storage units and load equipment according to a preset autonomous control strategy, thereby maintaining the continuity and stability of the system.

[0023] The communication layer ensures stable transmission of critical data even in rural areas with weak signal by setting up primary and backup links and switching between multiple standards; among them, LoRa or NB-IoT are used for low-power wide-coverage transmission, and 4G / 5G are used for high-bandwidth low-latency transmission.

[0024] The cloud-based management platform utilizes big data analytics and artificial intelligence algorithms to fuse historical operational data with real-time sensing data, identifying energy supply and demand imbalances, equipment failures, and environmental anomalies. It also improves system operating efficiency through predictive maintenance. Furthermore, the cloud-based management platform supports global energy scheduling and optimization, dynamically adjusting energy allocation based on distributed generation units, energy storage systems, and load characteristics. It also ensures the reliability of data transmission and policy execution through security authentication and access control mechanisms.

[0025] The management system for rural distributed energy achieves comprehensive perception and efficient management of the system's operational status by constructing a layered sensing architecture. This architecture combines multi-source sensor data acquisition, autonomous edge computing processing, redundant multi-standard communication transmission, and cloud-based big data analysis and optimized scheduling. Through multi-layered collaboration and intelligent control, the reliability, real-time performance, and operational efficiency of rural distributed energy systems are effectively improved, providing technical support for the safe, stable, and sustainable development of rural energy. Compared to existing single-layer sensing or centralized control methods, this invention constructs a layered architecture encompassing a sensing layer, edge layer, communication layer, and cloud management layer, forming an integrated end-edge-cloud multi-layered collaborative system.

[0026] Through the aforementioned layered collaborative mechanism, this invention provides a low-cost, scalable, and low-maintenance intelligent sensing and scheduling solution for rural distributed energy systems, while ensuring data acquisition integrity ≥99%, timestamp error ≤±200ms, end-to-end control execution latency ≤200ms, annual terminal failure rate ≤1%, and equipment cost ≤400CNY / node. This significantly improves the system's real-time performance, reliability, and replicability. Furthermore, this invention incorporates a "zero-trust" security concept in its system design, deploying multi-level access control and encryption mechanisms at the edge, cloud, and terminal layers to ensure data trustworthiness and controllable access. Simultaneously, a multi-level online and preventative maintenance mechanism is constructed, enabling fault visualization, early detection of potential hazards, and rapid response through equipment status monitoring and predictive analysis. Through modular design and a unified interface standard, this method facilitates the subsequent integration and expansion of various energy types such as photovoltaic, wind, and biomass energy, meeting the needs of long-term, low-cost operation and maintenance and sustainable development of rural energy systems.

[0027] In some embodiments, preprocessing, anomaly detection, and autonomous control of multi-source data include: Redundant data is removed from multi-source data; Anomaly fluctuations are detected in preprocessed multi-source data using time series and threshold models. If the abnormal fluctuations do not exceed the preset fluctuation range, the system automatically corrects the fluctuations and maintains autonomous operation. If the abnormal fluctuations exceed the preset fluctuation range, an alarm is triggered and the abnormal fluctuation data is sent to the platform layer for further analysis.

[0028] Lightweight filtering effectively removes redundant and noisy data, while compression algorithms normalize the data, reducing upload bandwidth pressure. Real-time monitoring of key operating curves is achieved through time-series modeling and threshold judgment algorithms. For power fluctuations and equipment status anomalies, the system employs a K-means clustering diagnostic algorithm to identify and classify normal and abnormal states, distinguishing potential early equipment failures or gradual degradation. If abnormal fluctuations do not exceed a set threshold, the system automatically adjusts strategies (such as load switching) using its built-in Python / Lua rule engine, automatically correcting and compensating to maintain autonomous operation. If the anomaly exceeds the threshold, a local audible and visual alarm is immediately triggered, and the original data, corrected data, and event tags are uploaded to the cloud management layer via an encrypted link for subsequent in-depth diagnostics and source analysis using machine learning models. Simultaneously, preset emergency control commands are executed (such as circuit breaker tripping or activating backup compensation devices).

[0029] In some embodiments, redundant data removal from multi-source data includes: Based on preset normal ranges and rates of change, obvious abnormal and noisy data are removed; And / or, use median filtering and moving average algorithms to smooth multi-source data and remove instantaneous interference and noise data.

[0030] First, based on a set reasonable range and rate of change, flash value judgment is performed to directly eliminate obvious abnormal and noisy data. Then, algorithms such as median filtering and moving average are used to smooth the running curves of multi-source data to suppress transient interference. The processed data is aggregated at the edge layer, such as calculating the average, maximum, and minimum values ​​within 5 minutes, which significantly reduces the amount of uplink data and bandwidth pressure.

[0031] In some embodiments, abnormal fluctuation detection is performed on preprocessed multi-source data using a time series and threshold model, including: Based on the preprocessed multi-source data, a normal distribution model of key parameters is established using a sliding window, and the moving average and standard deviation are calculated. When the collected data points deviate from the threshold range of the mean or standard deviation, they are identified as abnormal fluctuations.

[0032] First, the system uses sliding window technology to establish a short-term normal distribution model of key parameters (such as voltage, current, and power) and calculates their moving average and standard deviation in real time. When the collected data points fluctuate abnormally, such as deviating from the mean or exceeding 3 times the standard deviation, they are judged as abnormal fluctuations.

[0033] Referring to Table 1, through the introduction of the aforementioned edge layer mechanism, the system can not only achieve efficient and low-latency data processing in rural distributed energy scenarios, but also enhance its proactive defense and rapid response capabilities against abnormal states, ultimately supporting the platform layer's optimized scheduling and multi-layer collaborative management.

[0034] Table 1

[0035] In some embodiments, the communication layer includes a primary transmission link and a backup transmission link to ensure stable transmission of data between layers, including: The primary and backup transmission links are determined based on the performance parameters of various communication methods. These performance parameters include any one or more of the following: anti-interference capability, coverage, bandwidth, or energy consumption. The data level is determined based on the data to be transmitted. If the data level is Level 1, multipath transmission is performed using the primary transmission link and the backup transmission link. If the data level is level 2, the primary transmission link is used for transmission. If the delay of the primary transmission link exceeds the preset delay or the terminal, the system will automatically switch to the backup transmission link for transmission.

[0036] Redundancy switching and reliability assurance mechanisms at the communication layer are crucial for ensuring the stable operation of rural distributed energy systems in complex electromagnetic and geographical environments. Specifically, at least two communication methods are pre-configured within the same system, such as LoRa and NB-IoT or 4G / 5G complementing each other, enabling automatic switching between primary and backup links. When the primary link experiences high latency, signal attenuation, or link interruption leading to a decline in transmission quality, the system can switch to the backup link within milliseconds, ensuring the continuity and real-time nature of data transmission. To ensure reliable transmission of critical control commands and high-value operational data in the complex rural environment, a multi-path parallel transmission and intelligent switching mechanism is introduced at the communication layer. This mechanism first replicates data packets at the edge gateway and then redundantly transmits them through dual primary and backup channels (e.g., "fiber / 4G as the primary link + LoRa as the backup link"), leveraging the physical differences between different communication standards to significantly improve data delivery rates. Specifically, the link decision engine integrated on the platform side evaluates the anti-interference capability, coverage, bandwidth, and energy consumption of each channel in real time based on characteristics such as rural terrain undulations, vegetation obstruction, and wireless interference. During transmission, the system continuously monitors the signal quality (RSSI / SNR) and end-to-end latency of the main link. If the performance falls below a set threshold, it can seamlessly switch to a backup link within 30 seconds. Simultaneously, for critical data, a dual-mode concurrent strategy of "NB-IoT and LoRaWAN" is employed to ensure that at least one signal can reach the platform. This dynamic routing mechanism based on environmental characteristics constitutes a highly reliable redundant communication network, effectively addressing single-link interruptions and ensuring the continuity and real-time nature of service data, thereby significantly improving data delivery rate and packet loss resilience.

[0037] Meanwhile, the platform comprehensively weighs the anti-interference capabilities, coverage, bandwidth requirements, and energy consumption levels of different communication methods based on the characteristics of the complex rural environment (such as undulating terrain, vegetation obstruction, and diverse sources of wireless interference), thus forming a scenario-specific communication adaptation strategy. For example, LoRa is prioritized in long-distance, low-power scenarios; 4G / 5G is used in real-time control scenarios requiring high bandwidth and high speed; and NB-IoT is used in scenarios with dense nodes but moderate traffic, to achieve the best balance between performance and cost. Through this design, the communication layer can maintain highly reliable data transmission in rural environments with frequent interference and severe signal attenuation, meeting the real-time and stability requirements of distributed energy systems and providing reliable basic data support for subsequent cloud-based big data analysis and scheduling optimization.

[0038] In some embodiments, predictive maintenance and global energy scheduling based on multi-source data and a preset algorithm include: Feature extraction and pattern recognition are performed on multi-source data, and machine learning models are used to predict the probability of equipment failure. By combining scheduling algorithms with load forecasting, the coordinated operation of distributed energy resources and energy storage systems can be achieved.

[0039] The cloud-based management platform's data analysis and scheduling optimization not only undertakes the core task of centralized processing but also enhances the system's intelligence level through multi-dimensional algorithm models. In terms of operation scheduling, the cloud platform integrates advanced scheduling optimization algorithms, combining short-term and medium-to-long-term load forecasts provided by edge-side LSTM models to perform global optimization and coordinated control of heterogeneous energy sources such as distributed photovoltaic, wind, energy storage, and small hydropower. This scheduling mechanism first uses a precise digital twin model established at the platform layer to conduct all-weather, multi-period simulations of the "source-grid-load-storage" system, aiming to optimize the overall system's operational economy and energy utilization efficiency to generate a scheduling plan. Specifically, when renewable energy output fluctuates, the platform quickly issues commands to energy storage PCS (Power Conversion System) or flexible loads to achieve real-time power smoothing and balancing. During peak load periods, it prioritizes the joint power supply of distributed energy and energy storage units, executing precise load reduction. Furthermore, under platform commands, it can combine edge autonomy strategies (such as automatically switching to backup power when SOC is too low) to achieve optimized control through "cloud-edge collaboration." This approach enables the system to play a crucial role in peak load reduction, valley filling, and smoothing renewable energy fluctuations.

[0040] In some embodiments, feature extraction from multi-source data includes: Based on the timestamps of multi-source data, a time-period characteristic curve of load and power generation output is constructed. Mark key events on the time-cycle characteristic curve and extract feature vectors; the feature vectors include the time of occurrence, duration, or fluctuation amplitude.

[0041] In one specific embodiment, based on accurate GNSS / NTP timestamps, daily / weekly periodic characteristic curves of load and power generation output are constructed to identify the "peak-valley-flat" time period patterns; finally, for events such as voltage over-limit and frequency deviation initially marked by the edge-side K-means clustering diagnostic algorithm, feature vectors such as their occurrence time, duration, and fluctuation amplitude are extracted.

[0042] In some embodiments, pattern recognition is performed in the following manner: Based on clustering algorithms, the operating status of multi-source data is divided into several classic modes, including normal, light overload overvoltage, or heavy overload low voltage. By associating classic patterns with related factors, we obtain related patterns.

[0043] In one specific embodiment, the platform utilizes the aforementioned features for in-depth analysis. Through the K-means clustering algorithm, it classifies the operating states in the multi-dimensional feature space into typical patterns such as "normal," "light load overvoltage," and "heavy load low voltage," achieving automatic classification and profiling of operating conditions. Correlation analysis identifies the coupling relationships between different patterns and meteorological conditions and user behavior. For example, it establishes a correlation model of "photovoltaic output - irradiance - cloud cover changes," or an energy consumption pattern of "tea processing - specific time period - sudden load increase."

[0044] In some embodiments, using a machine learning model to predict the probability of device failure includes: The extracted feature vectors and the identified patterns are input into the trained machine learning model to obtain the probability of equipment failure.

[0045] In one specific embodiment, these extracted features and identified patterns serve as key inputs to drive the LSTM model to make more accurate short-term load and renewable energy output predictions, and to establish equipment failure probability models. This enables a deep understanding of the system's operating status, predictive maintenance and accurate health assessment of key equipment, and closed-loop management from global optimization to local execution, thereby avoiding the impact of potential failures on system stability.

[0046] like Figure 2As shown, the platform first cleans and aligns the heterogeneous data from multiple sources uploaded by the sensing layer and edge layer, including electrical parameters (U / I / P / Q / Hz), environmental meteorological data (irradiance, wind speed, temperature and humidity), and equipment safety status (smoke detection, water immersion). In the feature extraction stage, the platform first calculates the moving average, standard deviation, and number of short-term over-limits for key indicators from the power data to quantify operational stability. Then, based on accurate GNSS / NTP timestamps, it constructs load and power generation cycle characteristic curves, extracts feature vectors, and standardizes the features. In the pattern recognition stage, the platform uses the above features for in-depth analysis. Through the K-means clustering algorithm, it classifies the operating status in the multi-dimensional feature space into typical patterns such as "normal," "light load overvoltage," and "heavy load low voltage," achieving automatic classification and profiling of operating conditions. Correlation analysis identifies the coupling relationship between different patterns and meteorological conditions and user behavior, such as establishing a correlation model of "photovoltaic output - irradiance - cloud layer change," or an energy consumption pattern of "tea processing - specific time period - sudden load increase." Ultimately, these extracted features and identified patterns serve as key inputs to drive the LSTM model for more accurate short-term load and renewable energy output predictions, and to establish equipment failure probability models. This enables a deep understanding of the system's operational status, predictive maintenance and accurate health assessment of critical equipment, and closed-loop management from global optimization to local execution, thereby avoiding the impact of potential failures on system stability. In terms of operation scheduling, the cloud platform integrates advanced scheduling optimization algorithms, combined with short-term and medium-to-long-term load prediction results provided by the edge-side LSTM model, to perform global optimization and coordinated control of heterogeneous energy sources such as distributed photovoltaic, wind, energy storage, and small hydropower. This scheduling mechanism first uses a precise digital twin model established at the platform layer to conduct all-weather, multi-period simulations of the "source-grid-load-storage" system, generating a scheduling plan with the goal of optimizing the overall system's operational economy and energy utilization efficiency. Specifically, when renewable energy output fluctuates, the platform quickly issues commands to energy storage PCS (Power Conversion System) or flexible loads to achieve real-time power smoothing and balancing. During peak load periods, it prioritizes the joint power supply of distributed energy resources and energy storage units, executing precise load reduction. Furthermore, under platform commands, it can combine edge autonomy strategies (such as automatically switching to backup power when SOC is too low) to achieve optimized "cloud-edge collaborative" control. This approach enables the system to play a crucial role in peak load reduction, off-peak load filling, and smoothing renewable energy fluctuations.

[0047] All data adheres to standardized MQTT 5.0 or HTTPS protocols, is encapsulated in a unified JSON or Protobuf format, and uploaded to the platform via a communication module with breakpoint resume capability. For abnormal data or control commands issued by the platform, the system executes local autonomous logic based on its built-in Python / Lua rule engine, realizing a complete closed loop of "data preprocessing - anomaly diagnosis - edge autonomy - reliable upload," ensuring the real-time performance, reliability, and efficiency of the entire data processing process.

[0048] In terms of scalability, the platform adopts a modular and service-oriented architecture, which can flexibly support the addition, deletion, and updating of different functional modules, and has the ability to seamlessly connect with photovoltaic, wind, biomass energy, and future new energy types. At the same time, the platform ensures the compatibility and scalability of multi-energy collaboration through a unified data interface and standardized protocols, laying the foundation for subsequent system iterations and regional large-scale promotion.

[0049] In some embodiments, the system includes several types of sensing sensors for collecting multi-source data. The sensing sensors integrate GNSS timing modules and are supplemented by NTP network time synchronization mechanisms to form a dual time synchronization system, so that the timestamp error of the collected multi-source data is controlled within a preset time range.

[0050] Sensors include, but are not limited to, low-cost electrical parameter meters, environmental and meteorological sensors, and security and status sensors. Electrical parameter acquisition devices such as voltage, current, power, and frequency are deployed on the user's energy consumption side and the distributed power supply side. Environmental sensors such as temperature and humidity, solar irradiance, wind speed and direction, and air pressure are also deployed to achieve comprehensive perception of power operation and the external environment. Each node adopts a modular plug-and-play design, facilitating low-cost expansion and maintenance. All acquisition nodes integrate a GNSS timing module, supplemented by an NTP network time synchronization mechanism, forming a dual time synchronization system to ensure that the timestamp error of the acquired data is controlled within ±200ms. The nodes incorporate lightweight data verification and anomaly detection algorithms to mark packet loss, transmission delay, and data drift in real time. Combined with a 15-minute rolling statistical mechanism, the data integrity rate is dynamically calculated, ensuring ≥99%. Through SoC integrated design, sensing, communication, and processing modules are integrated onto the same circuit board, keeping the average cost below 400 CNY per node. Furthermore, PCB ruggedization, sealing protection, and a wide temperature range (-40℃~+85℃) enhance the reliability of the equipment in harsh rural environments, achieving a node MTBF ≥50,000 hours. At the edge layer, edge computing gateways are deployed at key locations in rural distribution areas or microgrids. These gateways possess multi-channel data aggregation, preprocessing, and protocol conversion capabilities, supporting multiple standard interfaces such as Modbus, DL / T645, and IEC61850. When a communication link interruption or disconnection from the upper-level cloud platform is detected, the edge gateway automatically switches to local control mode, utilizing a built-in rule base and historical data to perform autonomous control, such as load shedding, energy storage scheduling, and inverter power limiting, ensuring safe system operation.

[0051] Running a lightweight machine learning model specifically optimized for rural scenarios within the edge gateway is the core of achieving intelligent sensing and rapid autonomy. The gateway strictly follows the models and strategies issued by the platform layer, running an LSTM short-term prediction model to accurately predict the distribution network load curve, providing a decision-making basis for real-time scheduling of power sources, grid, load, and storage. Simultaneously, the gateway utilizes a K-means clustering diagnostic algorithm to rapidly cluster and recognize patterns in collected operational data such as voltage and frequency, enabling millisecond-level identification and early warning of abnormal states such as voltage exceeding limits and frequency deviations. This "prediction-diagnosis" closed loop is entirely completed at the edge. Combined with a built-in Python / Lua rule engine, it can autonomously execute preset control strategies such as "automatic power cut-off for overload" or "automatic switching to backup power when energy storage SOC is too low" when disconnected from the platform, effectively improving the power supply resilience and autonomy of the vulnerable rural power grid. Employing edge encryption chips and local key management, the system achieves multi-layered encrypted transmission from the data acquisition end to the cloud and supports whitelist / blacklist device access policies. At the cloud layer, it completes unified data storage, algorithm analysis, and energy scheduling. The platform provides standardized APIs to support integration with scheduling centers and third-party energy management platforms, and uses distributed identity authentication and blockchain notarization technology to ensure the credibility and traceability of data sharing.

[0052] In terms of security, the platform employs a multi-level access control and distributed encryption mechanism, combined with a zero-trust architecture, to dynamically verify and control access throughout the entire process of data collection, transmission, storage, and retrieval, ensuring data trustworthiness and privacy security in complex rural power environments. For data processing, the system adopts a modular and service-oriented architecture, providing high flexibility for the access and processing of heterogeneous energy data such as photovoltaic, wind, and biomass power. Data originates from intelligent fusion terminals deployed on-site. Their acquisition modules aggregate multi-energy stream data through multiple interfaces (RS485 / CAN / 4-20mA) and wireless networks (ZigBee / LoRa), performing plug-and-play identification and initial filtering. In the edge computing module, the system efficiently processes the raw data.

[0053] In summary, the intelligent sensing and multi-layer collaborative management method for rural distributed energy systems proposed in this invention constructs a secure and reliable architecture spanning the entire "end-edge-cloud" chain. At the sensing layer, modular sensing units and distributed deployment achieve high-completeness acquisition of power and environmental parameters, while GNSS / NTP dual time synchronization technology ensures timing accuracy. At the edge layer, autonomous control and rapid response mechanisms are introduced to support independent operation and millisecond-level adjustment in communication anomaly scenarios. At the platform layer, multi-source data cross-validation and intelligent algorithms are combined to achieve global scheduling optimization and early warning of operational anomalies.

[0054] Meanwhile, this invention incorporates a zero-trust security strategy throughout the entire chain, dynamically authenticating and controlling access during data acquisition, transmission, and storage to ensure information reliability. At the operation and maintenance level, a multi-level online monitoring and preventative maintenance mechanism is established, utilizing terminal operation logs and health assessments to achieve visualized management of node status and early intervention for potential faults. Through these measures, the sensing layer equipment can achieve high reliability, low operation and maintenance costs, and long-term stable operation in complex rural power environments, providing solid support for the safe, economical, and sustainable development of rural energy systems.

[0055] 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.

[0056] 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.

[0057] 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.

[0058] 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.

[0059] 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.

[0060] 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.

[0061] 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.

[0062] 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.

[0063] 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.

[0064] 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. A management system for rural distributed energy, characterized by, The system comprises a perception layer, an edge layer, a communication layer and a platform layer; wherein, The perception layer is configured to collect real-time multi-source data, the multi-source data comprising any one or more of electrical parameter data, meteorological data or environmental data; The edge layer is configured to pre-process, detect abnormalities and autonomously control the multi-source data; The communication layer is configured to ensure stable transmission of data to be transmitted between layers, the communication layer comprising at least two types of communication modes; the communication modes comprising LoRa, NB-IoT or 4G / 5G; The platform layer is configured to perform predictive maintenance and global energy scheduling based on the multi-source data and a preset algorithm.

2. The system of claim 1, wherein, The pre-processing, abnormality detection and autonomous control of the multi-source data comprises: removing redundant data from the multi-source data; detecting abnormal fluctuations in the pre-processed multi-source data using a time series and threshold model; if the abnormal fluctuations do not exceed a preset fluctuation range, automatically correcting and maintaining autonomous operation; if the abnormal fluctuations exceed the preset fluctuation range, triggering an alarm and sending the data of the abnormal fluctuations to the platform layer for further analysis by the platform layer.

3. The system of claim 2, wherein, The removing of redundant data from the multi-source data comprises: removing obvious abnormal and noise data based on a preset normal range and rate of change; and / or, using a median filter and a moving average algorithm to smooth the multi-source data and remove transient interference and noise data.

4. The system of claim 2, wherein, The detecting of abnormal fluctuations in the pre-processed multi-source data using a time series and threshold model comprises: based on the pre-processed multi-source data, using a sliding window to establish a normal distribution model of key parameters and calculating a moving average and a standard deviation; when a data point collected deviates from the threshold range of the average value or the standard deviation, it is determined to be an abnormal fluctuation.

5. The system of claim 1, wherein, The communication layer comprises a main transmission link and a backup transmission link, and the stable transmission of data to be transmitted between layers comprises: determining the main transmission link and the backup transmission link according to the index parameters of various types of communication modes, the index parameters comprising any one or more of anti-interference ability, coverage, bandwidth or energy consumption; determining the data level according to the data to be transmitted; if the data level is level one, using the main transmission link and the backup transmission link for multi-path transmission; if the data level is level two, using the main transmission link for transmission; if the delay of the main transmission link exceeds a preset delay or terminal, automatically switching to the backup transmission link for transmission.

6. The system of claim 1, wherein, The predictive maintenance and global energy scheduling based on the multi-source data and a preset algorithm comprises: extracting features and identifying patterns from the multi-source data, and predicting the probability of equipment failure using a machine learning model; coordinating the operation of distributed energy and energy storage systems through a scheduling algorithm combined with load forecasting.

7. The system of claim 6, wherein, The feature extraction from the multi-source data comprises: based on the timestamp of the multi-source data, constructing a time period characteristic curve of load and power generation output; labeling key events of the time period characteristic curve and extracting feature vectors; the feature vectors comprising time of occurrence, duration or fluctuation amplitude.

8. The system of claim 6, wherein, The mode recognition is performed in the following manner: Based on a clustering algorithm, the operating state of the multi-source data is divided into multiple classic modes, and the classic modes include normal, light load overvoltage or heavy load low voltage; The classic mode is associated with the associated factors to obtain the associated mode.

9. The system of claim 6, wherein, The device fault probability is predicted by using the machine learning model, which includes: The extracted feature vector and the recognized mode are input into the trained machine learning model to obtain the device fault probability.

10. The system of claim 1, wherein, The system includes a plurality of sensing sensors, and a plurality of sensing sensors are used to collect the multi-source data. The sensing sensor integrates a GNSS timing module, and is assisted by an NTP network time correction mechanism to form a dual time synchronization system, so that the time stamp error of the collected multi-source data is controlled within a preset time range.

Images