Power grid asset management system based on transformation scene

Abstract

The invention discloses a power grid asset management system based on a transformation scene. The system comprises an asset identification module, a data acquisition module, a big data analysis module and an AI decision support module. The asset identification module is used for being connected with Internet of Things equipment built in each asset; the data acquisition module is used for acquiring asset real-time data of each asset; the big data analysis module is used for performing supply chain analysis according to the supply chain data of each supply chain to obtain a corresponding supply chain analysis result; the AI decision support module is used for predicting the maintenance demand and the scrap time of each asset according to the asset real-time data to obtain the predicted maintenance demand and the predicted scrap time of each asset, and then, according to the predicted maintenance demand, the predicted scrap time and the supply chain analysis result, the supply chain analysis result is obtained; and generating a maintenance plan, a scrap replacement plan and a supply chain optimization strategy corresponding to each asset. According to the invention, the power grid assets can be automatically managed.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a power grid asset management system based on a transformation scenario. Background Art

[0002] In today's rapidly developing society, the stability and efficiency of the power grid, as the infrastructure of the national economy, directly affect the normal operation of society and people's daily life. Among them, the full life cycle management (ALM) of power grid assets is an important means to ensure the long-term reliable operation of the power grid. However, with the continuous expansion of the scale of the power grid and the increasing complexity of the structure, the traditional power grid asset management method has gradually exposed some limitations in the face of the complexity and dynamism of the modern power grid.

[0003] Key assets in the power grid, such as transformers, switchgear, cables, and distribution cabinets, are widely distributed in various regions. These assets are numerous, diverse, and widely distributed. Especially when facing the transformation of power grid enterprises, their asset management is more chaotic and duplicated, and the management difficulty is greatly increased. However, traditional asset management mainly relies on manual records and regular maintenance, manually recording asset information such as location, status, and maintenance history, and performing inspections and maintenance at specific time nodes. This management method has played an important role in the operation of power grids in the past, but in the scenario of enterprise transformation, the difficulty of asset management has greatly increased. This manual-dependent method is inefficient and prone to errors. Summary of the invention

[0004] The present invention provides a power grid asset management system based on a transformation scenario to solve the technical problem that traditional asset management relies on manual records and regular maintenance, which is inefficient and prone to errors in enterprise transformation scenarios.

[0005] In order to solve the above technical problems, an embodiment of the present invention provides a power grid asset management system based on a transformation scenario, including: an asset identification module, a data acquisition module, a big data analysis module and an AI decision support module;

[0006] The asset identification module is used to identify the identity of each asset through the device ID of the IoT device built into each asset, and connect with the IoT device built into each asset; wherein the real-time asset data includes: location information and status information of the asset;

[0007] The data acquisition module is used to obtain the real-time asset data of each asset through the connected Internet of Things device, and transmit the real-time asset data to the big data analysis module;

[0008] The big data analysis module is used to obtain the supply chain data of each supply chain, perform supply chain analysis based on the supply chain information, obtain corresponding supply chain analysis results, and transmit the obtained supply chain analysis results and the real-time asset data to the AI ​​decision support module; wherein; the supply chain data includes: supplier information, procurement information, inventory information, logistics information and product information of each supply chain; the supply chain analysis results include: the probability of occurrence of each preset risk in each supply chain;

[0009] The AI ​​decision support module is used to predict the maintenance needs and scrapping time of each asset based on the real-time data of the assets, obtain the predicted maintenance needs and predicted scrapping time of each asset, and then generate a maintenance plan, scrapping and replacement plan and supply chain optimization strategy corresponding to each asset based on the predicted maintenance needs, predicted scrapping time and the supply chain analysis results.

[0010] As a preferred solution, the asset identification module includes: a device registration and management unit, an IoT device access unit, and a data transmission unit;

[0011] The device registration and management unit is used to assign a corresponding device ID to each IoT device built into the asset, and manage each IoT device according to the device ID;

[0012] The IoT device access unit is used to identify the identity of each asset through the device ID of the IoT device built into each asset, and connect with the IoT device built into each asset according to a preset IoT access protocol;

[0013] The data transmission unit is used to transmit the real-time asset data collected by each of the IoT devices to the data collection module through a preset data transmission interface.

[0014] As a preferred solution, the data acquisition module is also used to preprocess the real-time asset data before transmitting it to the big data analysis module; wherein the preprocessing includes: data format conversion, data deduplication and data packet loss detection.

[0015] As a preferred solution, the big data analysis module includes: a data storage unit, a data cleaning unit and an analysis engine unit;

[0016] The data storage unit is used to obtain the supply chain data of each supply chain, and store the real-time asset data and the supply chain data in a preset distributed database;

[0017] The data cleaning unit is used to perform data preprocessing on the supply chain data and transmit the preprocessed supply chain data to the analysis engine unit; wherein the data preprocessing includes: data deduplication, data outlier processing, data filling and data standardization;

[0018] The analysis engine unit is used to perform supply chain analysis based on the preprocessed supply chain data, obtain corresponding supply chain analysis results, and transmit the obtained supply chain analysis results and the real-time asset data to the AI ​​decision support module.

[0019] As a preferred solution, the analysis engine unit is used to perform supply chain analysis based on the preprocessed supply chain data to obtain corresponding supply chain analysis results, including:

[0020] The analysis engine unit is used to calculate the prior probability of each preset risk in each supply chain based on the preprocessed supply chain data, and then calculate the posterior probability of each preset risk in each supply chain based on the prior probability, and use the posterior probability as the probability of each preset risk in each supply chain to obtain the corresponding supply chain analysis results.

[0021] As a preferred solution, the AI ​​decision support module includes: an AI model unit and a decision engine unit;

[0022] The AI ​​model unit is used to predict the maintenance demand and scrapping time of each asset according to the real-time data of the assets, obtain the predicted maintenance demand and predicted scrapping time of each asset, and transmit the predicted maintenance demand and predicted scrapping time to the decision engine unit; wherein the predicted maintenance demand includes: the next predicted maintenance time and predicted maintenance type of each asset;

[0023] The decision engine unit is used to obtain a maintenance plan, a scrapping and replacement plan for each asset and an optimal supply chain corresponding to each asset based on the predicted maintenance time, predicted maintenance type, predicted scrapping time and the supply chain analysis result.

[0024] As a preferred solution, the AI ​​model unit is used to predict the maintenance requirements and scrapping time of each asset according to the real-time data of the asset, and obtain the predicted maintenance requirements and predicted scrapping time of each asset, including:

[0025] The AI ​​model unit is used to input the real-time data of the assets into a preset LSTM network so that the LSTM network captures the time dependency in the asset status data and outputs the corresponding implicit state, and inputs the implicit state into a preset random forest model so that the random forest model predicts the maintenance requirements and scrapping time of each asset based on the implicit state, and outputs the predicted maintenance requirements and predicted scrapping time of each asset.

[0026] As a preferred solution, the power grid asset management system based on the transformation scenario further includes: a data management and visualization module and a real-time monitoring and alarm module;

[0027] The data management and visualization module is used to store and manage the real-time asset data of each asset, and to visualize the real-time asset data and the supply chain analysis results;

[0028] The real-time monitoring and alarm module is used to monitor the location information and status information of each asset based on the real-time data of the asset, and to issue an alarm when it is detected that an asset is in an abnormal state.

[0029] As a preferred solution, the data management and visualization module includes: a cloud database management unit, a visualization unit and a report generation unit;

[0030] The cloud database management unit is used to store and manage the real-time asset data of each asset;

[0031] The visualization unit is used to visualize the real-time asset data of each asset and the supply chain analysis results;

[0032] The report generation unit is used to generate a corresponding asset management report based on the real-time asset data, the supply chain analysis results and a preset report template.

[0033] As a preferred solution, the real-time monitoring and alarm module includes: an online monitoring unit, an abnormality detection unit and an alarm notification unit;

[0034] The online monitoring unit is used to monitor the current status and current location of each asset based on the acquired real-time asset data;

[0035] The abnormality detection unit is used to monitor whether each asset is in an abnormal state according to the current state and current position of each asset;

[0036] The alarm notification unit is used to issue an alarm when it is detected that an asset is in an abnormal state.

[0037] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0038] The present invention provides a power grid asset management system based on a transformation scenario, comprising an asset identification module, a data acquisition module, a big data analysis module and an AI decision support module; the asset identification module is used to identify the identity of each asset through the device ID of the Internet of Things device built into each asset, and connect with the Internet of Things device built into each asset; wherein the asset real-time data includes: the location information and status information of the asset; the data acquisition module is used to obtain the real-time asset data of each asset through the connected Internet of Things device, and transmit the real-time asset data to the big data analysis module; the big data analysis module is used to obtain the supply chain data of each supply chain, and perform supply chain analysis based on the supply chain information, Obtain the corresponding supply chain analysis results, and transmit the obtained supply chain analysis results and the real-time data of the assets to the AI ​​decision support module; wherein; the supply chain data includes: supplier information, procurement information, inventory information, logistics information and product information of each supply chain; the supply chain analysis results include: the probability of occurrence of each preset risk in each supply chain; the AI ​​decision support module is used to predict the maintenance requirements and scrapping time of each asset according to the real-time data of the assets, obtain the predicted maintenance requirements and predicted scrapping time of each asset, and then generate the corresponding maintenance plan, scrapping replacement plan and supply chain optimization strategy for each asset according to the predicted maintenance requirements, predicted scrapping time and the supply chain analysis results.

[0039] The power grid asset management system based on the transformation scenario of the present invention integrates an asset identification module, a data acquisition module, a big data analysis module and an AI decision support module, which can automatically manage power grid assets, solve the problems of low efficiency and frequent errors in traditional manual management methods, and significantly improve the management efficiency and accuracy of power grid assets. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a structural diagram of a power grid asset management system based on a transformation scenario provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solutions and advantages of this application clearer, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0042] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by technicians in the technical field to which this application belongs; the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned figure descriptions and any variations thereof are intended to cover non-exclusive inclusions.

[0043] In the description of the embodiments of the present application, the technical terms "first", "second", etc. are only used to distinguish different objects, and cannot be understood as indicating or implying relative importance or implicitly indicating the number, specific order or primary and secondary relationship of the indicated technical features. In the description of the embodiments of the present application, the meaning of "multiple" is more than two, unless otherwise clearly and specifically defined.

[0044] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0045] In the description of the embodiments of the present application, the term "and / or" is only a description of the association relationship of the associated objects, indicating that there may be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0046] In the description of the embodiments of the present application, the term "multiple" refers to more than two (including two). Similarly, "multiple groups" refers to more than two groups (including two groups), and "multiple pieces" refers to more than two pieces (including two pieces).

[0047] In the description of the embodiments of the present application, unless otherwise clearly specified and limited, technical terms such as "installed", "connected", "connected", "fixed" and the like should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to the specific circumstances.

[0048] Embodiment 1

[0049] Please refer to Figure 1, is a schematic diagram of a structure of a power grid asset management system based on a transformation scenario provided by an embodiment of the present invention, including: an asset identification module, a data acquisition module, a big data analysis module and an AI decision support module;

[0050] The asset identification module is used to identify the identity of each asset through the device ID of the IoT device built into each asset, and connect with the IoT device built into each asset; wherein the real-time asset data includes: location information and status information of the asset;

[0051] The data acquisition module is used to obtain the real-time asset data of each asset through the connected Internet of Things device, and transmit the real-time asset data to the big data analysis module;

[0052] The big data analysis module is used to obtain the supply chain data of each supply chain, perform supply chain analysis based on the supply chain information, obtain corresponding supply chain analysis results, and transmit the obtained supply chain analysis results and the real-time asset data to the AI ​​decision support module; wherein; the supply chain data includes: supplier information, procurement information, inventory information, logistics information and product information of each supply chain; the supply chain analysis results include: the probability of occurrence of each preset risk in each supply chain;

[0053] The AI ​​decision support module is used to predict the maintenance needs and scrapping time of each asset based on the real-time data of the assets, obtain the predicted maintenance needs and predicted scrapping time of each asset, and then generate a maintenance plan, scrapping and replacement plan and supply chain optimization strategy corresponding to each asset based on the predicted maintenance needs, predicted scrapping time and the supply chain analysis results.

[0054] Preferably, the power grid asset management system based on the transformation scenario also includes: a data management and visualization module and a real-time monitoring and alarm module; the data management and visualization module is used to store and manage the real-time asset data of each asset, and visualize the real-time asset data and the supply chain analysis results; the real-time monitoring and alarm module is used to monitor the location information and status information of each asset based on the real-time asset data, and to alarm when it is detected that an asset is in an abnormal state.

[0055] Specifically, the present invention combines AI technology to perform full life cycle management of power grid assets. The power grid asset management system based on the transformation scenario includes an asset identification module, a data acquisition module, a big data analysis module, an AI decision support module, a data management and visualization module, and a real-time monitoring and alarm module;

[0056] The asset identification module identifies the identity of each asset based on the device ID of the IoT device built into each asset, and connects with the IoT device built into each asset;

[0057] The data acquisition module is used to collect real-time data of assets, including asset location, asset status and environmental information, through connected IoT devices;

[0058] The big data analysis module is used to analyze the supply chain information of assets and generate insights for optimizing the asset supply chain;

[0059] The AI ​​decision support module stores, cleans and analyzes real-time asset data to predict the maintenance needs and scrapping time of each asset;

[0060] The data management and visualization module is used to manage and visualize the collected data, supporting real-time query and report generation;

[0061] The real-time monitoring and alarm module monitors the status of connected assets and triggers an alarm when an abnormal situation occurs.

[0062] The present invention combines AI to manage the entire life cycle of power grid assets, and comprehensively solves the problems of inefficiency and error-proneness caused by the reliance on manual methods for power grid asset management. The system integrates multiple modules, among which the asset identification module automatically identifies assets, reducing errors in manual records; the data acquisition module obtains the location information, status and environmental data of assets in real time to ensure the timeliness and accuracy of the data; the big data analysis module conducts in-depth analysis of the asset supply chain and provides optimization solutions; the AI ​​decision support module accurately predicts maintenance needs and scrapping time by storing, cleaning and analyzing real-time data, thereby reducing failure risks and maintenance costs; the data management and visualization module effectively manages and displays data, facilitating real-time query and report generation; the real-time monitoring and alarm module continuously monitors the status of assets, and immediately triggers an alarm when an abnormality occurs, ensuring that problems are responded to in a timely manner. Through the integration of these functions, the system significantly improves the automation and intelligence level of power grid asset management, and solves the problems of low efficiency and frequent errors in traditional methods.

[0063] Preferably, the asset identification module includes: a device registration and management unit, an Internet of Things device access unit and a data transmission unit; the device registration and management unit is used to assign a corresponding device ID to each built-in Internet of Things device of each asset, and manage each of the Internet of Things devices according to the device ID; the Internet of Things device access unit is used to identify the identity of each asset through the device ID of the Internet of Things device built into each asset, and connect with the Internet of Things device built into each asset according to a preset Internet of Things access protocol; the data transmission unit is used to transmit the real-time asset data collected by each of the Internet of Things devices to the data acquisition module through a preset data transmission interface.

[0064] Specifically, the asset identification module includes: an IoT device access unit, a device registration and management unit, and a data transmission interface;

[0065] (1) IoT device access unit: Assets are connected to the Internet through built-in or attached IoT modules to achieve real-time data transmission and location updates; each asset is identified and located online through a GPS module, and the IoT device access unit continuously monitors the status and location of the asset and sends the data to the data acquisition module in real time; the IoT device access unit adopts IoT access protocols such as MQTT, CoAP, HTTP, etc. to ensure stable and low-latency data transmission between assets and management systems.

[0066] Among them, the real-time asset data collected by IoT devices include:

[0067] (a) Asset location data:

[0068] GPS coordinates: Obtain the longitude and latitude of the asset through the GPS module and update the location of the asset in real time;

[0069] Location information update time: record the timestamp of the latest location data to ensure the timeliness of the data.

[0070] (b) Asset status data:

[0071] Operation status: For example, the asset's online, offline, faulty, normal operation, and other status information.

[0072] Health status: Environmental parameters such as temperature, humidity, vibration, pressure, etc. of the asset, to monitor whether the working environment of the equipment is within the normal range.

[0073] Power consumption: Monitor the energy consumption of assets in real time to determine the efficiency of equipment.

[0074] (2) Device registration and management unit: Each asset is registered and managed through a unique IoT device ID; the details are as follows:

[0075] Unique ID assignment: Each IoT device is registered with a unique device ID when it is connected to the system. These IDs are usually generated based on MAC addresses, serial numbers, or dedicated device identifiers to ensure the uniqueness and traceability of each asset.

[0076] Device management platform: The system includes a device management platform for centralized management of all registered devices. The management platform provides functions such as device status monitoring, device information update, and exception handling to ensure that each asset can be effectively tracked and managed throughout its life cycle. Exception handling here refers to handling exceptions that can be directly judged by the device. For example, when the temperature exceeds the temperature threshold, it is judged as an exception, and the exception is reported to remind relevant personnel to handle it.

[0077] (3) Data transmission interface: assets upload data to the data management and visualization module through the Internet of Things;

[0078] The data transmission interface adopts encryption technologies such as AES and SSL / TLS to ensure that the data of assets is not tampered with or leaked during the transmission process.

[0079] Data format and protocol conversion: The interface supports multiple data formats, such as JSON, XML, CSV, etc., and can perform protocol conversion as needed so that the data can be recognized and processed by other modules in the system. For example, the data collected by the sensor is converted from the original format to a standard format that can be parsed by the system.

[0080] Data redundancy and fault-tolerance mechanism: The interface design takes data redundancy and fault-tolerance mechanism into consideration to ensure that in the event of network fluctuations or device failures, data can continue to be transmitted through caching or retransmission mechanisms to avoid data loss.

[0081] Preferably, the data acquisition module is also used to preprocess the real-time asset data before transmitting the real-time asset data to the big data analysis module; wherein the preprocessing includes: data format conversion, data deduplication and data packet loss detection.

[0082] Specifically, the data acquisition module transmits the collected data to the big data analysis module through the MQTT or HTTP protocol, and stores the data in a database based on the big data analysis module for call by the AI ​​decision support module, data management and visualization module, and real-time monitoring and alarm module.

[0083] Before the data is transmitted to the big data analysis module, the data acquisition module performs preliminary processing on the data, including data format conversion, deduplication, packet loss detection, etc., to ensure the integrity and consistency of the data; and to facilitate subsequent analysis and processing, the data acquisition module will add tags to the real-time data of each asset, including asset ID, timestamp, location, etc. These tags can help the system quickly identify and associate data from different assets.

[0084] Preferably, the big data analysis module includes: a data storage unit, a data cleaning unit and an analysis engine unit; the data storage unit is used to obtain the supply chain data of each supply chain, and store the asset real-time data and the supply chain data in a preset distributed database; the data cleaning unit is used to perform data preprocessing on the supply chain data, and transmit the preprocessed supply chain data to the analysis engine unit; wherein the data preprocessing includes: data deduplication, data outlier processing, data filling and data standardization; the analysis engine unit is used to perform supply chain analysis based on the preprocessed supply chain data to obtain corresponding supply chain analysis results, and transmit the obtained supply chain analysis results and the asset real-time data to the AI ​​decision support module.

[0085] Preferably, the analysis engine unit is used to perform supply chain analysis based on the preprocessed supply chain data to obtain corresponding supply chain analysis results, including: the analysis engine unit is used to calculate the prior probability of occurrence of each preset risk in each supply chain based on the preprocessed supply chain data, and then calculate the posterior probability of occurrence of each preset risk in each supply chain based on the prior probability, and use the posterior probability as the probability of occurrence of each preset risk in each supply chain to obtain the corresponding supply chain analysis results.

[0086] Specifically, the big data analysis module includes: a distributed database, a data cleaning unit and an analysis engine unit;

[0087] (1) Distributed database: used to store real-time asset data uploaded from IoT devices and supply chain data obtained over the Internet;

[0088] The supply chain data includes:

[0089] (a) Supplier information:

[0090] Supplier Name: Supplier's name and basic information;

[0091] Contact information: supplier’s contact address, phone number and email address;

[0092] Reputation rating: data based on historical performance, such as delivery time, quality control, etc.

[0093] (b) Procurement data:

[0094] Purchase Orders: Details of each purchase order, including order number, product category, quantity, unit price and total price;

[0095] Purchase time: record the generation time and delivery time of the purchase order;

[0096] Payment status: payment information, including the status of paid, pending payment, and overdue payment;

[0097] (c) Inventory data:

[0098] Inventory levels: real-time inventory quantities, including current inventory, inventory in transit, and safety stock levels;

[0099] Inventory turnover rate: a metric used to evaluate the efficiency of inventory management;

[0100] Expired or unsalable product information: Monitor which products are approaching expiration or unsalable, thereby optimizing inventory management;

[0101] (d) Logistics information:

[0102] Transportation status: including transportation mode, transportation company, real-time location and status updates during transportation;

[0103] Delivery time: record the actual delivery time compared with the expected delivery time;

[0104] Transportation cost: the cost of each transportation route;.

[0105] (e) Product information:

[0106] Product Specifications: Detailed specifications and features of each product;

[0107] Product Lifecycle: Information on managing a product throughout its lifecycle, from introduction to retirement.

[0108] Database architecture: The architecture combines a distributed file system (such as HDFS) and a distributed database (such as HBase and Cassandra) to ensure that the system can be expanded to handle large-scale data sets and provide high availability and fault tolerance.

[0109] Data partitioning and indexing: Improve query speed and data access efficiency by partitioning data and creating indexes. Partitions can be based on dimensions such as time, geographic location, or asset type, while indexes can speed up the retrieval of specific data.

[0110] Partitioning strategy:

[0111] Time partitioning: partitioning data based on timestamps. A common practice is to create partitions by day, week, or month. This is particularly effective for time series data (such as sensor data) and can quickly locate data within a specific time range;

[0112] Geographical location partitioning: Data is partitioned based on geographic location (such as country, province, or city). This is very effective for application scenarios that require regional analysis;

[0113] Asset type partitioning: Partitioning by asset type or category facilitates independent management and analysis of data for different types of assets.

[0114] Index Type:

[0115] Primary key index: ensures the uniqueness of each record, usually the primary key of the data table;

[0116] Secondary index: Create an index on non-primary key fields to speed up queries on specific fields. For example, you can create a secondary index on asset status or supplier name;

[0117] Full-text indexing: Full-text indexing of text fields (such as descriptions or comments) to support fuzzy queries and full-text searches;

[0118] Index implementation:

[0119] When creating a table, you can specify the index through a specific DDL (Data Definition Language) statement;

[0120] Choose an appropriate index type, taking into account the frequency of queries and the frequency of data updates to balance query performance and write performance.

[0121] Data storage format: Data storage uses columnar storage formats (such as Parquet and ORC). These formats are optimized for big data analysis, can improve query efficiency and compression ratio, and reduce storage space usage.

[0122] (2) a data cleaning unit, used to pre-process supply chain related data;

[0123] Data deduplication: Eliminate duplicate data to ensure that each record is unique and avoid affecting analysis results due to duplicate data.

[0124] Outlier processing: Detect and process outliers through statistical methods (such as Z-score, IQR) to ensure the robustness of the data set.

[0125] Data filling: For missing data, interpolation, mean filling or prediction filling methods based on machine learning models are used to ensure data integrity.

[0126] Data standardization and normalization: Standardize data so that data from different sources can be compared and analyzed under the same standard, such as unit conversion, data format unification, etc.

[0127] (3) an analysis engine unit, which performs supply chain analysis based on the preprocessed supply chain related data; the supply chain analysis adopts a risk assessment model of a Bayesian network and a Markov chain Monte Carlo (MCMC) method, and is specifically implemented as follows:

[0128] A Bayesian network is a directed acyclic graph (DAG) that is used to represent conditional dependencies between random variables. The network consists of nodes (representing random variables) and directed edges (representing dependencies);

[0129] Nodes: X1,X2,…,X n Represents various risk factors or events in the supply chain, for example: X1 represents the risk of supplier bankruptcy; X2 represents the risk of transportation delay; X3 represents the risk of market volatility;

[0130] Directed edge: X i →X j Represents variable X i For variable X j impacts, such as shipping delays that may affect inventory shortages;

[0131] Conditional probability table: For each node X i , given its parent node (denoted as Pa(X i )) Under the condition, define X i The conditional probability distribution P(X i |Pa(X i ));

[0132] Use Bayesian networks for reasoning and calculate posterior probabilities to assess the risk level of each risk factor;

[0133] Posterior probability calculation:

[0134] Given an observed evidence variable E = e, calculate a risk variable X i The posterior probability P(X i |E=e);

[0135] Calculate using Bayes' formula:

[0136]

[0137] Where: P(X i ) represents the prior probability, that is, the probability of X before any evidence is given. i Initial evaluation of P(E=e|Xi ) represents the likelihood acceptance book, that is, assuming X i The probability of observing evidence E = e under all possible circumstances is obtained by calculating the marginal likelihood of all possible X. i Summing gives:

[0138]

[0139] Markov Chain Monte Carlo (MCMC) Method:

[0140] MCMC methods are used to approximate complex posterior distributions of Bayesian networks, especially in high dimensions and complex dependencies. Commonly used MCMC algorithms include Gibbs Sampling and Random Walk Monte Carlo (Metropolis-Hastings);

[0141] Gibbs sampling algorithm: By gradually updating the value of each random variable, a series of sampling values ​​are generated to approximate the target distribution. The specific steps are as follows:

[0142] Initialization: For each random variable X i Choose an initial value

[0143] Iterative update: For each X i , given the current values ​​of other variables, sample new values

[0144]

[0145] Repeat the iterative update steps until enough samples are generated to approximate the posterior distribution;

[0146] Metropolis-Hastings algorithm: Generate samples that conform to the target distribution by accepting or rejecting the proposed state transition. The specific steps are as follows:

[0147] Initialization: Select initialization state X (0) ;

[0148] Generate proposal state: From the proposal distribution Q(X ′ |X (t) ) to sample a new state X ′ ;

[0149] Accept or reject the proposed state: calculate the acceptance probability α:

[0150]

[0151] Randomly select a value u, and accept the new state X if u < α (t+1) =X ′ ; Otherwise keep the current state X (t+1) =X (t) ;

[0152] Repeatedly generate the proposal state and the acceptance or rejection proposal state to generate enough samples;

[0153] Posterior probability distribution: The calculated posterior probability of each risk factor is used to assess the likelihood of each risk in the supply chain.

[0154] Use charts (such as heat maps and risk curves) to display the distribution and interrelationships of various risk factors to facilitate intuitive understanding and decision-making.

[0155] Preferably, the AI ​​decision support module includes: an AI model unit and a decision engine unit; the AI ​​model unit is used to predict the maintenance requirements and scrapping time of each asset based on the real-time data of the assets, obtain the predicted maintenance requirements and predicted scrapping time of each asset, and transmit the predicted maintenance requirements and predicted scrapping time to the decision engine unit; wherein the predicted maintenance requirements include: the next predicted maintenance time and predicted maintenance type of each asset; the decision engine unit is used to obtain the maintenance plan, scrapping and replacement plan of each asset and the optimal supply chain corresponding to each asset based on the predicted maintenance time, predicted maintenance type, predicted scrapping time and the supply chain analysis results.

[0156] Preferably, the AI ​​model unit is used to predict the maintenance requirements and scrapping time of each asset based on the real-time data of the assets, and obtain the predicted maintenance requirements and predicted scrapping time of each asset, including: the AI ​​model unit is used to input the real-time data of the assets into a preset LSTM network, so that the LSTM network captures the time dependency in the asset status data and outputs the corresponding implicit state, and inputs the implicit state into a preset random forest model, so that the random forest model predicts the maintenance requirements and scrapping time of each asset based on the implicit state, and outputs the predicted maintenance requirements and predicted scrapping time of each asset.

[0157] Specifically, the AI ​​decision support module includes: an AI model unit and a decision engine unit;

[0158] (1) AI model unit: used to analyze asset status data and predict maintenance needs and scrapping time;

[0159] The AI ​​model combines the long short-term memory network (LSTM) based on time series analysis with the random forest (RandomForest), which specifically includes the following parts:

[0160] Input Layer: Input asset historical status data (such as temperature, vibration, working hours, etc.), real-time data (such as current status, location, environmental information, etc.) and other external factors (such as weather, usage conditions, market demand, etc.).

[0161] LSTM Network:

[0162] Purpose: To process time series data, capture the time dependencies in asset status data, and generate hidden states.

[0163] Structure: The LSTM network contains multiple LSTM units, each of which consists of an input gate, a forget gate, an output gate, and a memory unit, and is used to process time series data.

[0164] Output: Hidden state h t , used as input for subsequent random forests;

[0165] Random Forest:

[0166] Purpose: To predict asset maintenance requirements and scrapping time based on the implicit states generated by LSTM and other non-time series data. The implicit states are the intermediate outputs generated by the LSTM network when processing time series data. They reflect the model's understanding and memory of the input data at each time step, and can capture the time dependency and dynamic changes in the data; the non-time series data refers to static or other forms of data relative to time series data. These data do not change over time, but provide a supplementary source of information, such as market demand data for related products, industry trends, etc.

[0167] Structure: The random forest model consists of multiple decision trees, each of which generates a prediction based on the input features. The final result is obtained by voting or averaging.

[0168] Output Layer:

[0169] Maintenance Needs Forecasting: Predict when an asset will need its next maintenance and the type of maintenance required.

[0170] Scrap time prediction: predict the remaining useful life (RUL) of an asset.

[0171] The loss function of the model is as follows:

[0172] Maintenance demand prediction loss, predict the next maintenance time point, in order to measure the difference between the prediction time and the actual maintenance time point:

[0173]

[0174] Where: represents the i-th maintenance time point predicted by the model; t i Indicates the actual maintenance time point; N indicates the number of samples;

[0175] Prediction of scrap time loss:

[0176]

[0177] Where: RUL represents the remaining useful life of the j-th asset predicted by the model; j represents the actual remaining useful life; M represents the number of samples;

[0178] Total loss function:

[0179] L total =α·L maintenance +β·L RUL ;

[0180] Where: α and β represent weight coefficients.

[0181] (2) Decision engine unit: Based on the predicted maintenance needs, scrapping time and supply chain analysis results, it automatically generates supply chain optimization strategies to optimize the full life cycle management of assets.

[0182] Maintenance strategy generation: Based on the predicted maintenance needs, the decision engine generates a detailed maintenance plan, including time, required resources, personnel arrangements, etc.

[0183] Assume the maintenance cost is C m , if maintenance is performed at time t, the optimization objective is:

[0184] Minimize∑ t C m (t)×P m (t);

[0185] Where: P m (t) represents the probability of performing maintenance at time t;

[0186] Scrapping and replacement strategy generation: Based on the scrapping time forecast, determine when assets need to be replaced and generate a replacement plan.

[0187] Assume that the cost after scrapping is C r , then the optimization objective is:

[0188] Minimize∑ t C r (t)×Pr (t);

[0189] Where: P r (t) represents the probability that the asset needs to be scrapped at time t;

[0190] Supply chain optimization strategy generation: Combined with the results of the analysis engine unit of the big data analysis module, the supply chain management strategy is optimized, and the best supply chain for each asset is selected to ensure the timely supply of required resources during maintenance and scrapping.

[0191] Output:

[0192] Maintenance plan: detailed maintenance schedule and resource arrangements.

[0193] Scrapping and replacement planning: replacement timing and resource allocation.

[0194] Supply chain optimization strategy: optimization suggestions for the supply chain, including resource allocation, supplier selection, etc.

[0195] Preferably, the data management and visualization module includes: a cloud database management unit, a visualization unit and a report generation unit; the cloud database management unit is used to store and manage the real-time asset data of each asset; the visualization unit is used to visualize the real-time asset data of each asset and the supply chain analysis results; the report generation unit is used to generate a corresponding asset management report based on the real-time asset data, the supply chain analysis results and a preset report template.

[0196] Specifically, the data management and visualization module includes: a cloud database management unit, a visualization unit, and a report generation unit;

[0197] (1) Cloud database management unit:

[0198] The cloud database management unit is used to store and manage asset data collected from IoT devices. It uses SQL and NoSQL databases to meet the storage requirements of different types of data.

[0199] SQL database: suitable for storing structured data, such as basic information of assets, time series data, etc. Commonly used SQL databases include MySQL, PostgreSQL, etc.

[0200] Data table design: The data table design should include fields such as asset ID, location, status, and collection time to facilitate subsequent query and analysis.

[0201] Data indexing: To improve query efficiency, you need to create indexes for commonly used query conditions (such as asset ID and time range).

[0202] NoSQL database: Suitable for storing unstructured data, such as device logs, pictures, video data, etc. Commonly used NoSQL databases include MongoDB, Cassandra, etc.

[0203] Data model design: NoSQL databases use a flexible data model design and can optimize the storage of different types of asset data, such as using document-based databases to store data in JSON format.

[0204] Data backup and recovery: The cloud database management unit should be configured with a regular data backup mechanism to ensure data security and reliability. At the same time, it should support rapid data recovery in the event of data damage or loss.

[0205] Data security and permission management: Through user role and permission control, ensure that only authorized users can access and operate specific data. In addition, the cloud database management unit must support data encryption transmission and storage to ensure the security of data during transmission and storage.

[0206] (2) Visualization unit:

[0207] The visualization unit displays key information such as the real-time status, location, supply chain risks, etc. of assets in the form of charts and dashboards. It mainly includes the following functions:

[0208] Dashboard: Used to display key asset indicators in real time, such as asset health status, location distribution, supply chain risks, etc. The dashboard can be configured with different views to meet the needs of different users.

[0209] Real-time update: The dashboard should support real-time update of data, so that users can check the latest asset status at any time.

[0210] Interactive functions: The dashboard should support user interactive operations, such as clicking on an asset in the chart to view its detailed information or historical data.

[0211] Chart generation: The visualization unit should support the generation of multiple chart types, such as line charts, bar charts, pie charts, heat maps, etc., to display data of different dimensions.

[0212] Dynamic filtering: Users can use the dynamic filtering function to customize the display range and conditions of the chart and flexibly view the required data.

[0213] Data aggregation and analysis: The visualization unit should provide data aggregation and analysis functions, such as asset status changes within a time range, health status of assets in an area, etc.

[0214] Map visualization: For asset location data, the visualization unit should support map visualization based on geographic information system (GIS).

[0215] Asset distribution map: Displays the geographical distribution of assets through a map, combining visual elements such as color and size to indicate the status, health and other information of the assets.

[0216] Track playback: For mobile assets, you can view their historical location tracks to help users analyze asset usage and movement patterns.

[0217] (3) Report generation unit:

[0218] The report generation unit is responsible for generating various periodic reports to help users regularly understand the status and management of assets. This unit should have the following functions:

[0219] Predefined report module: The report generation unit predefines a variety of report templates, such as asset health report, maintenance record report, supply chain risk analysis report, etc. Users can select the appropriate template to generate reports according to their needs.

[0220] Report template configuration: Users can customize the fields and display methods in the report template according to specific needs to meet different report requirements.

[0221] Report generation and export: Supports the generation of reports in PDF, Excel and other formats, making it convenient for users to save, share and print.

[0222] Automated report generation: The report generation unit should support automated report generation. Users can preset the time and frequency of report generation, such as daily, weekly, and monthly report generation. The system automatically completes report generation and sends it to the specified mailbox or storage location.

[0223] Custom fields and filters: When generating reports, users can flexibly configure custom fields and filters to generate data reports that meet specific conditions.

[0224] Report history record and query: The report generation unit should save the historical report records, and users can query and download historical reports at any time.

[0225] Preferably, the real-time monitoring and alarm module includes: an online monitoring unit, an abnormality detection unit and an alarm notification unit; the online monitoring unit is used to monitor the current status and current position of each asset based on the acquired real-time data of the assets; the abnormality detection unit is used to monitor whether each asset is in an abnormal state based on the current status and current position of each asset; the alarm notification unit is used to alarm when it is detected that an asset is in an abnormal state.

[0226] Specifically, the real-time monitoring and alarm module includes: an online monitoring unit, an abnormality detection unit and an alarm notification unit;

[0227] (1) Online monitoring units monitor asset location and status in real time based on IoT data;

[0228] (2) The anomaly detection unit automatically detects and issues penalty alarms based on the thresholds set by the AI ​​model; the thresholds can be set based on the mean and standard deviation of historical data, for example:

[0229] Upper threshold = mean + k × standard deviation

[0230] Lower threshold = mean - k × standard deviation

[0231] Wherein, k is a constant (such as 2 or 3) used to control the sensitivity of the threshold.

[0232] (3) The alarm notification unit is based on the detection results of the abnormality detection unit. If an alarm is triggered, it notifies relevant personnel to handle the problem.

[0233] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A power grid asset management system based on a transformation scenario, characterized in that: include: Asset identification module, data collection module, big data analysis module and AI decision support module; The asset identification module is used to identify the identity of each asset through the device ID of the IoT device built into each asset, and connect with the IoT device built into each asset; wherein the real-time asset data includes: location information and status information of the asset; The data acquisition module is used to obtain the real-time asset data of each asset through the connected Internet of Things device, and transmit the real-time asset data to the big data analysis module; The big data analysis module is used to obtain the supply chain data of each supply chain, perform supply chain analysis based on the supply chain information, obtain corresponding supply chain analysis results, and transmit the obtained supply chain analysis results and the real-time asset data to the AI ​​decision support module; wherein; the supply chain data includes: supplier information, procurement information, inventory information, logistics information and product information of each supply chain; the supply chain analysis results include: the probability of occurrence of each preset risk in each supply chain; The AI ​​decision support module is used to predict the maintenance needs and scrapping time of each asset based on the real-time data of the assets, obtain the predicted maintenance needs and predicted scrapping time of each asset, and then generate a maintenance plan, scrapping and replacement plan and supply chain optimization strategy corresponding to each asset based on the predicted maintenance needs, predicted scrapping time and the supply chain analysis results.

2. The power grid asset management system based on the transformation scenario according to claim 1, characterized in that: The asset identification module includes: a device registration and management unit, an IoT device access unit, and a data transmission unit; The device registration and management unit is used to assign a corresponding device ID to each IoT device built into the asset, and manage each IoT device according to the device ID; The IoT device access unit is used to identify the identity of each asset through the device ID of the IoT device built into each asset, and connect with the IoT device built into each asset according to a preset IoT access protocol; The data transmission unit is used to transmit the real-time asset data collected by each of the IoT devices to the data collection module through a preset data transmission interface.

3. The power grid asset management system based on the transformation scenario according to claim 1, characterized in that: The data acquisition module is also used to pre-process the real-time asset data before transmitting it to the big data analysis module; wherein the pre-processing includes: data format conversion, data deduplication and data packet loss detection.

4. The power grid asset management system based on the transformation scenario according to claim 1, characterized in that: The big data analysis module includes: a data storage unit, a data cleaning unit and an analysis engine unit; The data storage unit is used to obtain the supply chain data of each supply chain, and store the real-time asset data and the supply chain data in a preset distributed database; The data cleaning unit is used to perform data preprocessing on the supply chain data and transmit the preprocessed supply chain data to the analysis engine unit; wherein the data preprocessing includes: data deduplication, data outlier processing, data filling and data standardization; The analysis engine unit is used to perform supply chain analysis based on the preprocessed supply chain data, obtain corresponding supply chain analysis results, and transmit the obtained supply chain analysis results and the real-time asset data to the AI ​​decision support module.

5. The power grid asset management system based on the transformation scenario according to claim 4, characterized in that: The analysis engine unit is used to perform supply chain analysis based on the preprocessed supply chain data to obtain corresponding supply chain analysis results, including: The analysis engine unit is used to calculate the prior probability of each preset risk in each supply chain based on the preprocessed supply chain data, and then calculate the posterior probability of each preset risk in each supply chain based on the prior probability, and use the posterior probability as the probability of each preset risk in each supply chain to obtain the corresponding supply chain analysis results.

6. The power grid asset management system based on the transformation scenario according to claim 1, characterized in that: The AI ​​decision support module includes: an AI model unit and a decision engine unit; The AI ​​model unit is used to predict the maintenance demand and scrapping time of each asset according to the real-time data of the assets, obtain the predicted maintenance demand and predicted scrapping time of each asset, and transmit the predicted maintenance demand and predicted scrapping time to the decision engine unit; wherein the predicted maintenance demand includes: the next predicted maintenance time and predicted maintenance type of each asset; The decision engine unit is used to obtain a maintenance plan, a scrapping and replacement plan for each asset and an optimal supply chain corresponding to each asset based on the predicted maintenance time, predicted maintenance type, predicted scrapping time and the supply chain analysis result.

7. The power grid asset management system based on the transformation scenario according to claim 6, characterized in that: The AI ​​model unit is used to predict the maintenance requirements and scrapping time of each asset according to the real-time data of the asset, and obtain the predicted maintenance requirements and predicted scrapping time of each asset, including: The AI ​​model unit is used to input the real-time data of the assets into a preset LSTM network so that the LSTM network captures the time dependency in the asset status data and outputs the corresponding implicit state, and inputs the implicit state into a preset random forest model so that the random forest model predicts the maintenance requirements and scrapping time of each asset based on the implicit state, and outputs the predicted maintenance requirements and predicted scrapping time of each asset.

8. The power grid asset management system based on the transformation scenario according to claim 1, characterized in that: Also includes: Data management and visualization module and real-time monitoring and alarm module; The data management and visualization module is used to store and manage the real-time asset data of each asset, and to visualize the real-time asset data and the supply chain analysis results; The real-time monitoring and alarm module is used to monitor the location information and status information of each asset based on the real-time data of the asset, and to issue an alarm when it is detected that an asset is in an abnormal state.

9. The power grid asset management system based on the transformation scenario according to claim 8, characterized in that: The data management and visualization module includes: a cloud database management unit, a visualization unit and a report generation unit; The cloud database management unit is used to store and manage the real-time asset data of each asset; The visualization unit is used to visualize the real-time asset data of each asset and the supply chain analysis results; The report generation unit is used to generate a corresponding asset management report based on the real-time asset data, the supply chain analysis results and a preset report template.

10. The power grid asset management system based on the transformation scenario according to claim 8, characterized in that: Real-time monitoring and alarm module, including: online monitoring unit, abnormality detection unit and alarm notification unit; The online monitoring unit is used to monitor the current status and current location of each asset based on the acquired real-time asset data; The abnormality detection unit is used to monitor whether each asset is in an abnormal state according to the current state and current position of each asset; The alarm notification unit is used to issue an alarm when it is detected that an asset is in an abnormal state.

Images