Microgrid management method, device, equipment, medium and program product

Abstract

The embodiment of the invention discloses a micro-grid management method, device and equipment, a medium and a program product. Comprising the steps of obtaining to-be-processed local power grid data, performing data preprocessing on the to-be-processed local power grid data, and determining preprocessed local power grid data; inputting the preprocessed local power grid data into a pre-trained edge fine tuning model, and determining a micro-grid management strategy corresponding to the local power grid data to be processed; according to the micro-grid management strategy, each power device corresponding to the local power grid data to be processed is controlled to be adjusted; wherein the edge fine tuning model is a model obtained by performing supplementary training on a cloud power grid management model issued by a cloud based on local high-sensitivity data. According to the method, the side end data security management requirement in a cloud side end power network system architecture is fully considered, the overall management of a cloud side end micro-grid and the personalized security management of a side end are combined, and the management precision of power equipment in each side end micro-grid is improved on the basis of ensuring the security of the side end data.

Description

Technical Field

[0001] The present invention relates to the field of power network technology, and in particular to a microgrid management method, device, equipment, medium and program product. Background Art

[0002] With the substantial increase in the types of devices connected to the power network, including energy storage systems, distributed photovoltaics, charging piles and adjustable loads, these devices are not only diverse in type, but also in different locations. This leads to data heterogeneity in power network management, which poses a great challenge to power network management. Currently, the power network often uses a combination of cloud, edge and end to achieve resource management and control for different areas of the network.

[0003] However, with the rapid increase in the number of user-side devices connected to the power network, if all data is uploaded directly to the cloud platform, the cloud platform will need to analyze massive amounts of data to determine the management plan for edge devices, and it is also difficult for the cloud platform to directly manage edge devices. At the same time, different edge ends of power networks may have data with a high degree of confidentiality. If they are uniformly transmitted to the cloud platform, it will not only increase the pressure on data transmission bandwidth, but also bring high data security risks. For example, data may be stolen or tampered with during transmission, and the cloud platform may also become an attack target, resulting in data leakage. If all data cannot be transmitted to the cloud platform, it will affect the accuracy of the cloud platform's management of edge devices. The cloud platform will lack complete data support, making it difficult to accurately formulate management strategies, affecting its ability to control edge devices, and thus making it difficult to meet the cloud's control needs for the campus-level microgrid data management system (Energy Manage System, EMS) under the cloud-edge architecture of the power network. Summary of the invention

[0004] The present invention provides a microgrid management method, device, equipment, medium and program product, which fully consider the edge data security management needs in the cloud-edge power network system architecture, combine the cloud's overall management of the edge microgrid and the edge's own personalized security management, and on the basis of ensuring the security of the edge data, improve the management accuracy of the power equipment in each edge microgrid and reduce the management pressure of the operation and maintenance personnel.

[0005] In a first aspect, an embodiment of the present invention provides a microgrid management method, including:

[0006] Acquire the local power grid data to be processed, perform data preprocessing on the local power grid data to be processed, and determine the preprocessed local power grid data;

[0007] Input the pre-processed local grid data into the pre-trained edge fine-tuning model to determine the microgrid management strategy corresponding to the local grid data to be processed;

[0008] According to the microgrid management strategy, each power device corresponding to the local power grid data to be processed is controlled and adjusted;

[0009] Among them, the edge fine-tuning model is a model obtained by supplementing the training of the cloud power grid management model sent from the cloud based on local highly sensitive data.

[0010] In a second aspect, an embodiment of the present invention further provides a microgrid management device, including:

[0011] A local data acquisition module is used to acquire the local power grid data to be processed, perform data preprocessing on the local power grid data to be processed, and determine the preprocessed local power grid data;

[0012] A management strategy determination module, used to input the pre-processed local power grid data into the pre-trained edge fine-tuning model to determine the microgrid management strategy corresponding to the local power grid data to be processed;

[0013] The power equipment adjustment module is used to control the adjustment of each power equipment corresponding to the local power grid data to be processed according to the microgrid management strategy;

[0014] Among them, the edge fine-tuning model is a model obtained by supplementing the training of the cloud power grid management model sent from the cloud based on local highly sensitive data.

[0015] In a third aspect, an embodiment of the present invention further provides a microgrid management device, including:

[0016] at least one processor; and a memory communicatively coupled to the at least one processor;

[0017] The memory stores a computer program that can be executed by at least one processor, and the computer program is executed by at least one processor so that the at least one processor can implement the microgrid management method of any embodiment of the present invention.

[0018] In a fourth aspect, an embodiment of the present invention further provides a storage medium comprising computer executable instructions, which, when executed by a computer processor, are used to execute the microgrid management method of any embodiment of the present invention.

[0019] In a fifth aspect, an embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, is used to execute the microgrid management method of any embodiment of the present invention.

[0020] The embodiment of the present invention provides a microgrid management method, device, equipment, medium and program product, which obtains the local power grid data to be processed, performs data preprocessing on the local power grid data to determine the preprocessed local power grid data; inputs the preprocessed local power grid data into the pre-trained edge fine-tuning model to determine the microgrid management strategy corresponding to the local power grid data to be processed; controls the various power equipment corresponding to the local power grid data to be processed to adjust according to the microgrid management strategy; wherein the edge fine-tuning model is a model obtained by supplementing the training of the cloud power grid management model sent from the cloud based on local highly sensitive data. By adopting the above technical solution, when the local power grid data to be processed is obtained, the data is first preprocessed to obtain the preprocessed local power grid data that can meet the input conditions of the edge fine-tuning model, so that when the model generates the microgrid management strategy, it will not be affected by the multi-source heterogeneity phenomenon between different devices connected to the power network. At the same time, the pre-trained edge fine-tuning model that pre-processes the local power grid data input is a model obtained by supplementing the training of the cloud power grid management model issued by the cloud and applicable to each edge node through local high-sensitive data. It fully considers the edge data security management needs in the cloud-edge power network system architecture, and combines the cloud's overall management of the edge microgrid with the edge's own personalized security management. On the basis of ensuring the security of edge data, it improves the management accuracy of power equipment in each edge microgrid and reduces the management pressure of operation and maintenance personnel.

[0021] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0023] Figure 1 A flowchart of a microgrid management method provided in Embodiment 1 of the present invention;

[0024] Figure 2 A flow chart of a microgrid management method provided in Embodiment 2 of the present invention;

[0025] Figure 3 A schematic diagram of the structure of a microgrid management device provided in Embodiment 3 of the present invention;

[0026] Figure 4A schematic diagram of the structure of a microgrid management device provided in Embodiment 4 of the present invention. DETAILED DESCRIPTION

[0027] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0028] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0029] Embodiment 1

[0030] Figure 1 A flow chart of a microgrid management method provided in Embodiment 1 of the present invention. The embodiment of the present invention can be applied to the situation of targeted management of edge microgrids with high confidentiality requirements in a cloud-edge collaborative power network. The method can be executed by a microgrid management device, which can be implemented by software and / or hardware. The microgrid management device can be configured in a microgrid management device. Optionally, the microgrid management device can be an electronic device, which can be a laptop, desktop computer, smart tablet, and edge computing device in a cloud-edge collaborative power network, etc. The embodiment of the present invention does not limit this.

[0031] like Figure 1 As shown, a microgrid management method provided by an embodiment of the present invention specifically includes the following steps:

[0032] S101, obtaining local power grid data to be processed, and performing data preprocessing on the local power grid data to be processed, and determining preprocessed local power grid data.

[0033] In this embodiment, the local power grid data to be processed can be specifically understood as data obtained from multiple power devices controlled by an edge computing device in a cloud-edge collaborative power network, and it is necessary to use it as a basis to determine the management and control strategy of each power device in the microgrid system corresponding to the edge computing device. Exemplarily, the local power grid data to be processed may include distributed power supply equipment, energy storage systems, control systems, converters, battery management systems, charging piles, smart meters, power equipment, and communication equipment, etc., which are related to the microgrid system corresponding to the edge computing device.

[0034] In this embodiment, preprocessing local power grid data can be specifically understood as processing the local data to be processed that may have problems such as data missing, anomalies, and multi-source heterogeneity. The data is correct and has a unified structure, and the data obtained can meet the input requirements of the edge fine-tuning model used for subsequent microgrid management measurement generation.

[0035] Specifically, in a cloud-edge collaborative power network, the microgrid management device located at the edge node can obtain different types of data from different power equipment in the edge node area through different data communication interfaces, and the obtained data is collectively referred to as the local power grid data to be processed. In order to reduce the data anomalies in the acquired local power grid data to be processed, and solve the problem of being unable to perform unified processing in the subsequent data processing process due to structural anomalies between different types of data, it is necessary to preprocess the local power grid data to be processed, and obtain preprocessed local power grid data that can meet the input requirements of the edge fine-tuning model for subsequent microgrid management measurement generation.

[0036] For example, the local power grid data to be processed may include data that needs to be collected in real time due to real-time changes, and may also include data that only needs to be acquired once, such as inherent characteristics of equipment. The real-time data collection process can be achieved through real-time monitoring sensors, power trading interfaces, protocol communications, etc.

[0037] Exemplarily, depending on the data structure, the local power grid data to be processed can be divided into structured data, semi-structured data and unstructured data. Among them, structured data may include the power output, voltage and current of the photovoltaic power generation system collected by the meter; real-time monitoring of power, accumulated energy, power transaction data, metering meter data, basic information of the edge node server, etc.; semi-structured data may include temperature and switch status metadata collected by sensors; fault status and log information of photovoltaic, charging piles and adjustable loads; BMS monitoring data of energy storage equipment; external environment sensor monitoring data, internal safety sensor monitoring data, smart grid communication data, etc.; unstructured data may include maintenance reports, document reports, pictures, videos, voice records, user local feedback data, etc.; the embodiment of the present invention does not limit this.

[0038] Optionally, the data preprocessing method includes at least one of the following:

[0039] Remove outliers;

[0040] Missing value processing; missing value processing includes multiple interpolation and deletion of multiple missing value records;

[0041] Duplicate record removal;

[0042] The data format is unified.

[0043] Specifically, statistical methods such as Z-score and IQR can be used to remove outliers in the local power grid data to be processed; multiple possible filling values ​​are generated for each missing value through multiple interpolation, and record data containing a large number of missing values ​​in the local power grid data to be processed are deleted; hash functions or unique keys can be used to detect and delete duplicate record data in the local power grid data to be processed; based on the locally configured protocol library in the edge computing device corresponding to the local power grid data to be processed, the format of structured data, semi-structured data and unstructured data in the local power grid data to be processed is unified, and the local power grid data to be processed after the above processing is determined as preprocessed local power grid data.

[0044] Exemplarily, when preprocessing semi-structured data, taking log files as an example, log files usually contain information such as event records and system status, and the format may not be fixed. IEC 61850 is a standard protocol for smart grid communications, which includes complex data models and communication mechanisms. Use protocol parsing tools (such as the IEC 61850 parsing library) to convert communication data into a standard format. Specialized libraries such as libiec61850 can be used to parse IEC 61850 protocol data. Perform text parsing on log files and extract key fields. Use regular expressions or dedicated log parsing tools (such as Logstash and Fluentd) to extract key fields.

[0045] Exemplarily, when preprocessing unstructured data, it can be obtained through an API or directly from a file system. Unstructured data is obtained through interfaces such as documents, pictures, videos, and voices. Pictures and videos are compressed and format converted (such as Pillow and OpenCV). Text is extracted from documents and voice recognition (ASR) is performed on voices. Image recognition algorithms (such as YOLO and ResNet) are used to analyze surveillance pictures and videos. Natural language processing (NLP) algorithms are used to analyze documents and voice recordings, etc., which are not limited in the embodiments of the present invention.

[0046] S102: input the pre-processed local power grid data into the pre-trained edge fine-tuning model to determine the microgrid management strategy corresponding to the local power grid data to be processed.

[0047] Among them, the edge fine-tuning model is a model obtained by supplementing the training of the cloud power grid management model sent from the cloud based on local highly sensitive data.

[0048] In this embodiment, the edge fine-tuning model can be specifically understood as a neural network model that predicts information such as future power, power equipment status, and future resource demand in the area corresponding to the edge computing device based on the power network characteristics in the input power grid data, and generates a strategy for controlling the working status of each power equipment in the area corresponding to the edge computing device in the future.

[0049] In this embodiment, local highly sensitive data can be specifically understood as data obtained locally by the edge computing device, which is related to the working status of the local power network and has a high confidentiality level and is not suitable for transmission to the cloud. The cloud-based power grid management model can be specifically understood as a neural network model that can output corresponding power grid management strategies based on input power grid data and obtained by training the cloud-based data set of power grid data and strategies that do not contain highly sensitive data uploaded by each edge end it manages in a cloud-edge collaborative circuit network. It can be understood that the cloud-based power grid management model is completely consistent with the model architecture of the above-mentioned edge fine-tuning model. The edge fine-tuning model is a model obtained by supplementing the training of the cloud-based power grid management model based on local highly sensitive data, and adjusting different neural network layers in the model to adapt to local sensitive information.

[0050] Specifically, the preprocessed local power grid data is input into the pre-trained edge fine-tuning model, so that the edge fine-tuning model can process the preprocessed local power grid data in accordance with local highly sensitive needs, including predictions of information such as future power, power equipment status and future resource demand, and finally generate and output a strategy plan for controlling the working status of each power equipment in the area corresponding to the edge computing device in the future, and determine its output content as the microgrid management strategy corresponding to the local power grid data to be processed.

[0051] S103, controlling the power equipment corresponding to the local power grid data to be processed to make adjustments according to the microgrid management strategy.

[0052] Specifically, according to the microgrid management strategy, the operating parameters required for each power device in the microgrid area managed by the edge computing device in the cloud-edge collaborative power network in the future can be determined, and then the various operating parameters can be sent to the corresponding power equipment, so that the power equipment can quickly adjust its own operating parameters based on its own local data and control algorithm.

[0053] Optionally, a local emergency strategy for managing the microgrid area can also be configured in the edge computing device, so that the edge can respond immediately to emergencies such as equipment failures and power fluctuations without waiting for the microgrid management strategy to be generated.

[0054] The technical solution of this embodiment is to obtain the local power grid data to be processed, and perform data preprocessing on the local power grid data to be processed to determine the preprocessed local power grid data; input the preprocessed local power grid data into the pre-trained edge fine-tuning model to determine the microgrid management strategy corresponding to the local power grid data to be processed; and control the various power equipment corresponding to the local power grid data to be processed to adjust according to the microgrid management strategy; wherein the edge fine-tuning model is a model obtained by supplementing the training of the cloud power grid management model sent from the cloud based on local highly sensitive data. By adopting the above technical solution, when the local power grid data to be processed is obtained, the data is first preprocessed to obtain the preprocessed local power grid data that can meet the input conditions of the edge fine-tuning model, so that when the model generates the microgrid management strategy, it will not be affected by the multi-source heterogeneity phenomenon between different devices connected to the power network. At the same time, the pre-trained edge fine-tuning model that pre-processes the local power grid data input is a model obtained by supplementing the training of the cloud power grid management model issued by the cloud and applicable to each edge node through local high-sensitive data. It fully considers the edge data security management needs in the cloud-edge power network system architecture, and combines the cloud's overall management of the edge microgrid with the edge's own personalized security management. On the basis of ensuring the security of edge data, it improves the management accuracy of power equipment in each edge microgrid and reduces the management pressure of operation and maintenance personnel.

[0055] Embodiment 2

[0056] Figure 2A flow chart of a microgrid management method provided in Embodiment 2 of the present invention is further optimized on the basis of the above-mentioned optional technical solutions. Before obtaining the local power grid data to be processed, a local training sample set is first constructed based on the local highly sensitive data set, and the cloud power grid management model received from the cloud is supplemented with training based on the local training sample set, so as to obtain an edge fine-tuning model that meets the accuracy of local highly sensitive data processing. After the determination of the microgrid management strategy is completed, the pre-processed local power grid data is divided into the first uploaded data that needs to be encrypted, the second uploaded data with a large data volume that needs feature extraction and compression, the third uploaded data that can be uploaded without processing, and the local storage data with higher sensitivity that needs to be saved locally, according to the data characteristics of the pre-processed local power grid data including sensitivity, real-time and computational complexity. After the processing of each data that needs to be processed is completed, all the data that needs to be uploaded and the determined microgrid management strategy are uploaded to the cloud as target upload data, and the cloud data set is updated, so that the cloud can retrain or supplement the cloud power grid management model based on the microgrid management strategy that has considered the privacy content without obtaining the high privacy data of the edge, thereby improving the accuracy and adaptability of the cloud power grid management model sent from the cloud to different edges, combining the overall management of the cloud to the edge microgrid with the personalized security management of the edge itself, and on the basis of ensuring the security of the edge data, improving the management accuracy of the power equipment in each edge microgrid, and reducing the management pressure of the operation and maintenance personnel.

[0057] like Figure 2 As shown, a microgrid management method provided by an embodiment of the present invention specifically includes the following steps:

[0058] S201, receiving a cloud power grid management model sent from the cloud, and extracting multiple groups of local high-sensitivity data from the local high-sensitivity data set to construct a local training sample set.

[0059] Among them, the cloud power grid management model is trained by a cloud training sample set constructed by the cloud based on the cloud data set.

[0060] In this embodiment, the local high-sensitivity data set can be specifically understood as a set of local power grid data collected within a historical period, wherein the local power grid data collected within the historical period should include data with a sensitivity higher than a preset sensitivity threshold.

[0061] Specifically, after the edge computing device in the cloud-edge collaborative power network receives the cloud power grid management model sent from the cloud, in order to train it into a model that can be adapted to the microgrid where the edge computing device is located, multiple groups of local highly sensitive data can be extracted from the local highly sensitive data set, and the power grid data in each group of local highly sensitive data is used as a sample, and the corresponding historical microgrid management strategy is used as a label to construct a local training sample, and the set consisting of all local training samples is determined as the local training sample set.

[0062] S202: supplement the training of the cloud power grid management model through the local training sample set to determine the edge fine-tuning model.

[0063] Specifically, through the local training sample set, the cloud power grid management model is supplemented with training in the same way as its training on the cloud, and the trained cloud power grid management model is determined as the edge fine-tuning model, so that the obtained edge fine-tuning model can be suitable for processing the local power grid data to be processed with local confidentiality characteristics.

[0064] S203: Acquire the local power grid data to be processed, perform data preprocessing on the local power grid data to be processed, and determine the preprocessed local power grid data.

[0065] S204: input the pre-processed local power grid data into the pre-trained edge fine-tuning model, determine the microgrid management strategy corresponding to the local power grid data to be processed, and execute S205 and S206.

[0066] S205 , controlling the power equipment corresponding to the local power grid data to be processed to make adjustments according to the microgrid management strategy.

[0067] S206 . Divide the pre-processed local power grid data into first uploaded data, second uploaded data, third uploaded data and locally stored data according to data characteristics of the pre-processed local power grid data.

[0068] Among them, data characteristics include sensitivity, real-time and computational complexity.

[0069] Specifically, since highly sensitive data cannot be directly uploaded to the cloud for processing, it can be stored locally or uploaded to the cloud after encryption. However, data with high real-time performance often takes a lot of time to upload to the cloud and cannot guarantee real-time performance, and the required amount of data is often large, so it needs to be pre-processed on the edge before it can be uploaded to the cloud. However, due to the limitation of the computing power of the edge, data with high computational complexity often need to be uploaded to the cloud for processing. Therefore, in order to enrich the cloud data set used by the cloud for training the cloud power grid management model, after the microgrid management strategy is determined on the edge, the pre-processed local power grid data used to determine the microgrid management strategy can be divided, and the data that can be uploaded to the cloud to enrich the cloud data set can be determined. Specifically, the data characteristics corresponding to different types of data in the pre-processed local power grid data can be divided, and the first uploaded data that needs to be encrypted before being uploaded is obtained, the second uploaded data with high real-time performance and large data volume needs to be preferentially extracted and compressed on the edge to reduce the amount of uploaded data, the third uploaded data that can be directly uploaded to the cloud without processing, and the local storage data that cannot be uploaded to the cloud.

[0070] Optionally, according to data characteristics of the pre-processed local power grid data, the pre-processed local power grid data is divided into first uploaded data, second uploaded data, third uploaded data and locally stored data, including:

[0071] Determine the data in the pre-processed local power grid data whose sensitivity is higher than a preset sensitivity threshold and whose real-time performance is lower than a preset real-time performance threshold as the first uploaded data;

[0072] Determine the data in the pre-processed local power grid data, which has a sensitivity lower than a preset sensitivity threshold, a real-time performance higher than a preset real-time performance threshold, and a computational complexity higher than a preset complexity threshold, as the second uploaded data;

[0073] Determine the data in the pre-processed local power grid data whose sensitivity is lower than a preset sensitivity threshold, whose real-time performance is lower than a preset real-time performance threshold, and whose computational complexity is higher than a preset complexity threshold as the third uploaded data;

[0074] The data other than the first uploaded data, the second uploaded data and the third uploaded data in the pre-processed local power grid data are determined as the locally stored data.

[0075] In this embodiment, the preset sensitivity threshold can be specifically understood as a threshold that is pre-set according to actual conditions to distinguish the confidentiality level of the power grid data collected by the edge, and then determine whether the power grid data can be uploaded to the cloud or uploaded to the cloud after encryption. The preset real-time threshold can be specifically understood as a threshold that is pre-set according to actual conditions to distinguish the timeliness of processing the power grid data collected by the edge, and then determine whether the power grid data needs to be processed at the edge without being uploaded to the cloud. The preset complexity threshold can be specifically understood as a threshold that is pre-set according to actual conditions to distinguish whether the computing power of the edge can process the collected power grid data.

[0076] Specifically, according to the different data types of data acquired by the edge computing device, it has corresponding data characteristics, and the data characteristics of different types of data can be pre-set according to actual conditions. Since the pre-processed local power grid data is only the data obtained after data pre-processing for the local power grid data to be processed, it does not affect the data characteristics of the local power grid data to be processed itself, so the pre-processed local power grid data can be divided according to the data characteristics of different data types in the pre-processed local power grid data. The data with a sensitivity higher than the preset sensitivity threshold and a real-time lower than the preset real-time threshold, that is, the data that does not need to be processed by itself at the edge and needs to be encrypted before uploading is determined as the first uploaded data. The data with a sensitivity lower than the preset sensitivity threshold, a real-time higher than the preset real-time threshold, and a computational complexity higher than the preset complexity threshold, that is, although the real-time is high, the data cannot be completed at the edge due to the complexity of the data operation, and the confidentiality is not high and does not need to be encrypted is determined as the second uploaded data. The data with a sensitivity lower than the preset sensitivity threshold, a real-time lower than the preset real-time threshold, and a computational complexity higher than the preset complexity threshold, that is, the data that does not need to be pre-processed due to real-time and sensitivity factors, but only cannot be completed at the edge due to the complexity of the data operation is determined as the third uploaded data. The data in the pre-processed local power grid data other than the above data that needs to be uploaded to the cloud is determined as the locally stored data.

[0077] Exemplarily, the following table shows the data upload requirements of some local power grid data of different data types:

[0078]

[0079] As shown in the above table, "○" indicates that the characteristics of the data type are to be determined, which may or may not be present, and the processing measures under different characteristics are different; "√" indicates that the data type has the data characteristics; "×" indicates that the data type does not have the data characteristics. The data characteristics and local processing and cloud processing methods in the above table can be understood as some specific examples of dividing the first uploaded data, the second uploaded data, the third uploaded data and the locally stored data in the embodiment of the present invention. In actual use, the processing method of pre-processing local power grid data of different data types can be determined in the manner shown in the above table, or any other division method using data characteristics can be used. The embodiment of the present invention does not limit this. In essence, the pre-processed local power grid data is divided into the first uploaded data that needs to be encrypted and uploaded to the cloud for processing, the second uploaded data that needs to be feature extracted and compressed and transmitted to the cloud for processing, the third uploaded data that can be directly uploaded to the cloud without processing, and the locally stored data that needs to be retained locally for processing.

[0080] S207: Encrypt the first uploaded data and determine it as the first data to be uploaded.

[0081] Specifically, the first uploaded data is encrypted using an encryption method agreed upon with the cloud, and a result obtained by the processing is determined as the first data to be uploaded.

[0082] Exemplarily, the encryption processing method may include a public-private key pair, a lightweight terminal cross-domain authentication protocol, end-to-end encryption, and VPN encrypted transmission, etc., which is not limited in the embodiments of the present invention.

[0083] S208: sort, extract features and compress the second uploaded data to determine it as the second data to be uploaded.

[0084] Optionally, sorting the second uploaded data includes:

[0085] selecting an initial pivot in the second uploaded data and initializing a first pointer and a second pointer;

[0086] Scanning the second uploaded data based on the positions of the first pointer and the second pointer, and adjusting the positions of the second uploaded data, the first pointer and the second pointer;

[0087] determining a position where the first pointer and the second pointer meet as a final pivot, and dividing the adjusted second uploaded data into two sub-arrays according to the final pivot position;

[0088] Recursively sort each subarray.

[0089] For example, the first record in the second uploaded data may be used as the initial pivot, denoted as P i; Initialize the first pointer low and the second pointer high, low points to the starting position in the second uploaded data, and high points to the end position in the second uploaded data; Scan the second uploaded data based on the low pointer and the high pointer, and in the process of scanning from left to right, start from high and scan to the left to find the first keyword less than P i The record is moved to the position pointed to by low; in the process of scanning from right to left, start from low and scan to the right to find the first keyword greater than P i , move the record to the position pointed to by high; swap the two records pointed to by low and high, and then adjust the positions of the first pointer and the second pointer. Assume that the record pointed to by low is less than or equal to P i , then move the low pointer to the right; assuming that the record pointed to by high is greater than or equal to P i , then move the high pointer to the left; repeat the above steps until low and high meet, and determine the position as the final pivot. Then divide the adjusted second upload data into two sub-arrays according to the final pivot position. The left sub-array contains all keywords less than P. i The right subarray contains all records with keywords greater than P i , and then all sub-arrays can be recursively sorted to obtain the data required for feature extraction.

[0090] For example, the basis for division using the final pivot position can be timestamp or key field, etc. Taking timestamp as an example, the two sub-arrays divided can be the load power, battery status, photovoltaic power and other power generation power, grid electricity price, battery capacity, grid power consumption and other data at time t, which can be used as data set A t,n Indicates. t,n Represents the charging and discharging action data set taken by the nth energy storage power station or system at time t.

[0091] It is understandable that S207 and S208 may be executed simultaneously or in any order. The embodiment of the present invention is written by taking the sequential execution as an example.

[0092] S209 , taking the first data to be uploaded, the second data to be uploaded, the third uploaded data and the microgrid management strategy as target uploaded data.

[0093] S210, uploading the target uploaded data to the cloud to update the cloud data set in the cloud, and allowing the cloud to retrain the cloud power grid management model or supplement the training of the cloud power grid management model based on the updated cloud data set.

[0094] Optionally, the target upload data can be uploaded to the cloud at a certain period or as required, and during the upload process, an efficient communication protocol (such as MQTT or CoAP, etc.) can be used to achieve low-latency and low-bandwidth data transmission.

[0095] It can be understood that after the cloud retrains the cloud power grid management model or supplements the training of the cloud power grid management model based on the updated cloud data set, the cloud will send the newly trained cloud power grid management model to each edge it manages, so that each edge can execute steps S201-S210 as mentioned above, and the embodiments of the present invention will not be repeated here.

[0096] The technical solution of this embodiment is to first construct a local training sample set based on the local high-sensitivity data set before obtaining the local power grid data to be processed, and to supplement the training of the cloud power grid management model received from the cloud based on the local training sample set, so as to obtain an edge fine-tuning model that meets the accuracy of local high-sensitivity data processing. After the determination of the microgrid management strategy is completed, the pre-processed local power grid data is divided into the first uploaded data that needs to be encrypted, the second uploaded data with a large data volume that needs feature extraction and compression, the third uploaded data that can be uploaded without processing, and the local storage data with higher sensitivity that needs to be saved locally, according to the data characteristics of the pre-processed local power grid data including sensitivity, real-time and computational complexity. After the processing of each data that needs to be processed is completed, all the data that needs to be uploaded and the determined microgrid management strategy are uploaded to the cloud as target upload data, and the cloud data set is updated, so that the cloud can retrain or supplement the cloud power grid management model based on the microgrid management strategy that has considered the privacy content without obtaining the high privacy data of the edge, thereby improving the accuracy and adaptability of the cloud power grid management model sent from the cloud to different edges, combining the overall management of the cloud to the edge microgrid with the personalized security management of the edge itself, and on the basis of ensuring the security of the edge data, improving the management accuracy of the power equipment in each edge microgrid, and reducing the management pressure of the operation and maintenance personnel.

[0097] Embodiment 3

[0098] Figure 3 A schematic diagram of the structure of a microgrid management device provided in Embodiment 3 of the present invention is shown in FIG. Figure 3 As shown, the microgrid management device includes a local data acquisition module 31 , a management strategy determination module 32 and a power equipment adjustment module 33 .

[0099] Among them, the local data acquisition module 31 is used to obtain the local power grid data to be processed, and perform data preprocessing on the local power grid data to be processed to determine the preprocessed local power grid data; the management strategy determination module 32 is used to input the preprocessed local power grid data into the pre-trained edge fine-tuning model to determine the microgrid management strategy corresponding to the local power grid data to be processed; the power equipment adjustment module 33 is used to control the various power equipment corresponding to the local power grid data to be processed according to the microgrid management strategy; wherein the edge fine-tuning model is a model obtained by supplementing the training of the cloud power grid management model sent from the cloud based on local high-sensitivity data.

[0100] The technical solution of the embodiment of the present invention, when obtaining the local power grid data to be processed, first preprocesses the data to obtain the preprocessed local power grid data that can meet the input conditions of the edge fine-tuning model, so that the model will not be affected by the multi-source heterogeneity phenomenon between different devices connected to the power network when generating the microgrid management strategy. At the same time, the pre-trained edge fine-tuning model input by the pre-processed local power grid data is a model obtained after supplementing the training of the cloud power grid management model issued by the cloud and applicable to each edge node through local high-sensitive data. It fully considers the edge data security management requirements in the cloud-edge power network system architecture, combines the overall management of the cloud to the edge microgrid with the personalized security management of the edge itself, and on the basis of ensuring the security of the edge data, improves the management accuracy of the power equipment in each edge microgrid and reduces the management pressure of the operation and maintenance personnel.

[0101] Optionally, the data preprocessing method includes at least one of the following:

[0102] Remove outliers;

[0103] Missing value processing; missing value processing includes multiple interpolation and deletion of multiple missing value records;

[0104] Duplicate record removal;

[0105] The data format is unified.

[0106] Optionally, the microgrid management device further includes: a model fine-tuning module, which is used to:

[0107] Before obtaining the local power grid data to be processed, a cloud power grid management model sent from the cloud is received, and multiple groups of local high-sensitivity data are extracted from the local high-sensitivity data set to construct a local training sample set;

[0108] Supplementary training of the cloud-based power grid management model is performed through the local training sample set to determine the edge fine-tuning model;

[0109] Among them, the cloud power grid management model is trained by a cloud training sample set constructed by the cloud based on the cloud data set.

[0110] Optionally, the microgrid management device further includes: a data uploading module, which is used to:

[0111] After determining the microgrid management strategy corresponding to the local power grid data to be processed, dividing the pre-processed local power grid data into first uploaded data, second uploaded data, third uploaded data and locally stored data according to data characteristics of the pre-processed local power grid data;

[0112] Encrypting the first uploaded data and determining it as the first data to be uploaded;

[0113] sorting, extracting features and compressing the second uploaded data to determine it as the second data to be uploaded;

[0114] The first data to be uploaded, the second data to be uploaded, the third data to be uploaded and the microgrid management strategy are taken as target uploaded data;

[0115] Uploading the target uploaded data to the cloud to update the cloud data set in the cloud, and allowing the cloud to retrain the cloud power grid management model or supplement the training of the cloud power grid management model based on the updated cloud data set;

[0116] Among them, data characteristics include sensitivity, real-time and computational complexity.

[0117] Optionally, according to data characteristics of the pre-processed local power grid data, the pre-processed local power grid data is divided into first uploaded data, second uploaded data, third uploaded data and locally stored data, including:

[0118] Determine the data in the pre-processed local power grid data whose sensitivity is higher than a preset sensitivity threshold and whose real-time performance is lower than a preset real-time performance threshold as the first uploaded data;

[0119] Determine the data in the pre-processed local power grid data, which has a sensitivity lower than a preset sensitivity threshold, a real-time performance higher than a preset real-time performance threshold, and a computational complexity higher than a preset complexity threshold, as the second uploaded data;

[0120] Determine the data in the pre-processed local power grid data whose sensitivity is lower than a preset sensitivity threshold, whose real-time performance is lower than a preset real-time performance threshold, and whose computational complexity is higher than a preset complexity threshold as the third uploaded data;

[0121] The data other than the first uploaded data, the second uploaded data and the third uploaded data in the pre-processed local power grid data are determined as the locally stored data.

[0122] Optionally, sorting the second uploaded data includes:

[0123] selecting an initial pivot in the second uploaded data and initializing a first pointer and a second pointer;

[0124] Scanning the second uploaded data based on the positions of the first pointer and the second pointer, and adjusting the positions of the second uploaded data, the first pointer and the second pointer;

[0125] determining a position where the first pointer and the second pointer meet as a final pivot, and dividing the adjusted second uploaded data into two sub-arrays according to the final pivot position;

[0126] Recursively sort each subarray.

[0127] The microgrid management device provided in the embodiment of the present invention can execute the microgrid management method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0128] Embodiment 4

[0129] Figure 4 A schematic diagram of the structure of a microgrid management device provided for Embodiment 4 of the present invention. The microgrid management device 40 may be intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The microgrid management device 40 may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0130] like Figure 4 As shown, the microgrid management device 40 includes at least one processor 41, and a memory connected to the at least one processor 41, such as a read-only memory (ROM) 42, a random access memory (RAM) 43, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 41 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 42 or the computer program loaded from the storage unit 48 to the random access memory (RAM) 43. In RAM 43, various programs and data required for the operation of the microgrid management device 40 can also be stored. The processor 41, ROM 42 and RAM 43 are connected to each other through a bus 44. An input / output (I / O) interface 45 is also connected to the bus 44.

[0131] Multiple components in the microgrid management device 40 are connected to the I / O interface 45, including: an input unit 46, such as a keyboard, a mouse, etc.; an output unit 47, such as various types of displays, speakers, etc.; a storage unit 48, such as a disk, an optical disk, etc.; and a communication unit 49, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 49 allows the microgrid management device 40 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0132] The processor 41 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 41 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The processor 41 executes the various methods and processes described above, such as a microgrid management method.

[0133] In some embodiments, the microgrid management method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 48. In some embodiments, part or all of the computer program may be loaded and / or installed on the microgrid management device 40 via the ROM 42 and / or the communication unit 49. When the computer program is loaded into the RAM 43 and executed by the processor 41, one or more steps of the microgrid management method described above may be performed. Alternatively, in other embodiments, the processor 41 may be configured to execute the microgrid management method in any other suitable manner (e.g., by means of firmware).

[0134] Optionally, an embodiment of the present invention further provides a computer program product, including a computer program, which, when executed by a processor, implements the microgrid management method provided by any embodiment of the present invention.

[0135] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0136] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0137] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0138] To provide interaction with a user, the systems and techniques described herein may be implemented on a microgrid management device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to a user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which a user can provide input to the microgrid management device. Other types of devices may also be used to provide interaction with a user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0139] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0140] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0141] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0142] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A microgrid management method, characterized in that: include: Acquire the local power grid data to be processed, and perform data preprocessing on the local power grid data to be processed to determine the preprocessed local power grid data; Inputting the pre-processed local power grid data into a pre-trained edge fine-tuning model, and determining a microgrid management strategy corresponding to the local power grid data to be processed; Controlling the power equipment corresponding to the local power grid data to be processed to adjust according to the microgrid management strategy; Among them, the edge fine-tuning model is a model obtained by supplementing the training of the cloud power grid management model sent from the cloud based on local highly sensitive data.

2. The microgrid management method according to claim 1, characterized in that: Before obtaining the local power grid data to be processed, the method further includes: Receive the cloud power grid management model sent from the cloud, and extract multiple groups of local high-sensitivity data from the local high-sensitivity data set to construct a local training sample set; Supplementary training is performed on the cloud power grid management model through the local training sample set to determine an edge fine-tuning model; The cloud power grid management model is trained by a cloud training sample set constructed by the cloud based on a cloud data set.

3. The microgrid management method according to claim 1, characterized in that: After determining the microgrid management strategy corresponding to the local power grid data to be processed, the method further includes: According to data characteristics of the pre-processed local power grid data, the pre-processed local power grid data is divided into first uploaded data, second uploaded data, third uploaded data and locally stored data; Encrypting the first uploaded data and determining it as first data to be uploaded; performing data sorting, feature extraction and compression on the second uploaded data to determine it as the second data to be uploaded; Taking the first data to be uploaded, the second data to be uploaded, the third uploaded data and the microgrid management strategy as target uploaded data; Uploading the target uploaded data to the cloud to update the cloud data set of the cloud, and allowing the cloud to retrain the cloud power grid management model or supplement the training of the cloud power grid management model according to the updated cloud data set; The data characteristics include sensitivity, real-time and computational complexity.

4. The microgrid management method according to claim 3, characterized in that: The method of dividing the pre-processed local power grid data into first uploaded data, second uploaded data, third uploaded data and locally stored data according to data characteristics of the pre-processed local power grid data comprises: Determine the data in the pre-processed local power grid data whose sensitivity is higher than a preset sensitivity threshold and whose real-time performance is lower than a preset real-time performance threshold as the first uploaded data; Determine the data in the pre-processed local power grid data whose sensitivity is lower than a preset sensitivity threshold, whose real-time performance is higher than a preset real-time performance threshold, and whose computational complexity is higher than a preset complexity threshold as the second uploaded data; Determine the data in the pre-processed local power grid data whose sensitivity is lower than a preset sensitivity threshold, whose real-time performance is lower than a preset real-time performance threshold, and whose computational complexity is higher than a preset complexity threshold as the third uploaded data; The data other than the first uploaded data, the second uploaded data and the third uploaded data in the pre-processed local power grid data are determined as locally stored data.

5. The microgrid management method according to claim 3, characterized in that: The step of sorting the second uploaded data includes: selecting an initial pivot in the second uploaded data and initializing a first pointer and a second pointer; Scanning the second uploaded data based on the positions of the first pointer and the second pointer, and adjusting the positions of the second uploaded data, the first pointer, and the second pointer; determining a position where the first pointer and the second pointer meet as a final pivot, and dividing the adjusted second uploaded data into two sub-arrays according to the final pivot position; Recursively sort each of the sub-arrays.

6. The microgrid management method according to any one of claims 1 to 5, characterized in that: The data preprocessing method includes at least one of the following: Remove outliers; Missing value processing; wherein, the missing value processing includes multiple interpolation and deletion of multiple missing value records; Deletion of duplicate records; The data format is unified.

7. A microgrid management device, characterized in that: include: A local data acquisition module is used to acquire the local power grid data to be processed, and perform data preprocessing on the local power grid data to be processed to determine the preprocessed local power grid data; A management strategy determination module, used for inputting the pre-processed local power grid data into a pre-trained edge fine-tuning model to determine a microgrid management strategy corresponding to the local power grid data to be processed; A power equipment adjustment module, used for controlling the power equipment corresponding to the local power grid data to be processed to adjust according to the microgrid management strategy; Among them, the edge fine-tuning model is a model obtained by supplementing the training of the cloud power grid management model sent from the cloud based on local highly sensitive data.

8. A microgrid management device, characterized in that: include: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the microgrid management method according to any one of claims 1 to 6.

9. A storage medium containing computer executable instructions, characterized in that: The computer executable instructions are used to execute the microgrid management method as described in any one of claims 1 to 6 when executed by a computer processor.

10. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the microgrid management method according to any one of claims 1 to 6.

Images