Multi-source sensing integrated charging infrastructure operation and maintenance management system

Abstract

This invention discloses a multi-source sensing-integrated charging infrastructure operation and maintenance management system, relating to the field of power system operation and maintenance management. The system includes: a monitoring dataset acquisition module for acquiring monitoring datasets; a fault prediction module for inputting equipment status into an adaptation prediction unit and outputting multiple fault prediction datasets; a demand prediction module for performing demand prediction within a predetermined time period; an operation and maintenance urgency analysis module for performing operation and maintenance urgency analysis based on the fault prediction datasets and predicted charging demand; an operation and maintenance scheme optimization analysis module for analyzing operation and maintenance schemes by combining the charging station distribution topology and predetermined maintenance resources; and an operation and maintenance management module for performing operation and maintenance management based on the optimal operation and maintenance scheme. This invention solves the technical problems of low operation and maintenance efficiency and management level in existing charging infrastructure operation and maintenance management, achieving the technical effect of improving operation and maintenance efficiency and management level.

Description

Technical Field

[0001] This application relates to the field of power system operation and maintenance management, and in particular to a multi-source sensing-integrated charging infrastructure operation and maintenance management system. Background Technology

[0002] With the rapid development of the electric vehicle industry, charging infrastructure, as a crucial link supporting the popularization of electric vehicles, has expanded rapidly in scale and become increasingly widespread. However, the operation and maintenance management of charging infrastructure faces many challenges, such as frequent equipment failures, limited maintenance resources, and large fluctuations in charging demand. These problems not only affect the user's charging experience but may also threaten the safe and stable operation of the power grid. Therefore, the operation and maintenance management of charging infrastructure is particularly important. In existing charging infrastructure operation and maintenance management, a single data source monitoring method is usually adopted, which monitors the operation of charging stations only through equipment status data, ignoring the impact of environmental parameters on equipment status, leading to inaccurate fault prediction. At the same time, existing methods rely on historical charging data for simple demand forecasting, failing to fully consider the dynamic changes in future charging demand, making it difficult to effectively allocate maintenance resources.

[0003] At present, the operation and maintenance management of charging infrastructure suffers from technical problems such as low operation and maintenance efficiency and low operation and maintenance management level. Summary of the Invention

[0004] This application provides a multi-source sensing and joint operation and maintenance management system for charging infrastructure. It employs a monitoring dataset acquisition module that integrates equipment status sets and environmental parameter sets to achieve comprehensive perception and fusion of multi-source data. A fault prediction module, combined with environmental parameters, adaptively selects prediction units to perform refined analysis of equipment status and outputs fault prediction datasets corresponding one-to-one with charging piles. A demand prediction module, based on the charging dataset, dynamically predicts charging demand within a predetermined time period. An operation and maintenance urgency analysis module combines fault prediction results and demand prediction results to generate an urgency coefficient sequence. An operation and maintenance scheme optimization analysis module further combines the charging station distribution topology and predetermined maintenance resources, using optimization algorithms to determine the optimal operation and maintenance scheme. Finally, an operation and maintenance management module, based on the optimal scheme, performs intelligent operation and maintenance management of multiple charging stations within a predetermined time period. These technologies achieve the rational allocation and efficient utilization of operation and maintenance resources, resulting in improved operation and maintenance efficiency and management level.

[0005] This application provides a multi-source sensing-integrated charging infrastructure operation and maintenance management system, including:

[0006] The system includes the following modules: a monitoring dataset acquisition module for acquiring monitoring datasets from multiple charging stations within a target area, comprising a device status set and an environmental parameter set; a fault prediction module for inputting the device status set and environmental parameter set into a fault prediction channel, determining an adaptation prediction unit based on the environmental parameters, inputting the device status into the adaptation prediction unit, and outputting multiple fault prediction datasets, wherein each fault prediction data corresponds one-to-one with a charging pile; a demand prediction module for reading charging datasets from multiple charging stations within the target area, performing demand prediction within a predetermined time period based on the charging datasets, and determining multiple predicted charging demands; an operation and maintenance urgency analysis module for performing operation and maintenance urgency analysis based on the multiple fault prediction datasets and the multiple predicted charging demands, generating an urgency coefficient sequence; an operation and maintenance scheme optimization analysis module for performing operation and maintenance scheme optimization analysis based on the urgency coefficient sequence, combined with the charging station distribution topology and predetermined maintenance resources, to determine the optimal operation and maintenance scheme; and an operation and maintenance management module for performing operation and maintenance management on multiple charging stations within a predetermined time period according to the optimal operation and maintenance scheme.

[0007] In one possible implementation, a fault prediction channel is constructed, and the following processing is performed:

[0008] Configure environmental parameter thresholds, including temperature threshold, humidity threshold, and power grid load threshold; divide the temperature threshold, humidity threshold, and power grid load threshold according to a predetermined step size to obtain temperature interval set, humidity interval set, and power grid load interval set; randomly enumerate within the temperature interval set, humidity interval set, and power grid load interval set to obtain multiple environmental parameter groups; construct the fault prediction channel based on the multiple environmental parameter groups, wherein each environmental parameter group includes a temperature interval, a humidity interval, and a power grid load interval.

[0009] In a possible implementation, the fault prediction channel is constructed based on the multiple sets of environmental parameters, and the following processing is performed:

[0010] A first environmental parameter group is randomly selected from multiple environmental parameter groups, and a first charging station is randomly selected from multiple charging stations to obtain the first charging pile attribute of the first charging station. Using the first environmental parameter group as a condition constraint and the first charging pile attribute as a device constraint, a source retrieval is performed based on the Industrial Internet to obtain a first sample dataset. The first sample data includes sample device status and sample fault data, and the sample fault data includes sample fault type and sample fault node. The first sample dataset is used to train a predetermined operator to generate a first fault prediction model. Multiple fault prediction models of multiple charging stations are obtained sequentially, and a first prediction unit is constructed based on the multiple fault prediction models. Based on a binary tree, the fault prediction channel is constructed according to the mapping relationship between the first environmental parameter group and the first prediction unit.

[0011] In one possible implementation, a first fault prediction model is generated, and the following processing is performed:

[0012] Configure predetermined operators, including a digital twin model, a backpropagation neural network, and a random decision forest; divide the first sample dataset into ten equal parts and select them ten times with replacement to construct a first sample set; iterate three times to obtain three sample sets; using the sample device status as input and sample fault data as supervision, use the three sample sets to perform supervised training and cross-validation on the digital twin model, the backpropagation neural network, and the random decision forest respectively to obtain a first prediction sub-model, a second prediction sub-model, and a third prediction sub-model, wherein the prediction sub-model is identified by convergence degree; configure weight ratios based on convergence degree, and construct the first fault prediction model according to the weight ratios, the first prediction sub-model, the second prediction model, and the third prediction model.

[0013] In a possible implementation, demand forecasting is performed within a predetermined time period based on the charging dataset, and the following processing is executed:

[0014] A first charging station is randomly selected, and its first charging data is obtained, including the number of first working charging piles. The charging logs of the first charging station are retrieved under a predetermined time period to obtain a historical charging dataset, which includes the starting number of working charging piles and a limit on the number of working charging piles, and is marked with time intervals. Dynamic weights are configured based on the time intervals, and the historical charging dataset is weighted and fused according to these dynamic weights to obtain first standard charging data, including the number of first standard starting points and a first standard quantity limit. The quantity deviation between the number of first working charging piles and the number of first standard starting points is calculated, and the first standard quantity limit is compensated based on the quantity deviation to obtain a first predicted charging demand, which is then added to the plurality of predicted charging demands.

[0015] In a possible implementation, an urgency coefficient sequence is generated, and the following processing is performed:

[0016] Randomly select a first fault prediction dataset from the multiple fault prediction datasets, count the number of first fault prediction data that meet the predetermined time period, determine the first fault quantity, and calculate the first qualified quantity; obtain the first predicted charging demand corresponding to the first fault prediction dataset, and set the ratio of the first predicted charging demand to the first qualified quantity as the first emergency coefficient; obtain multiple emergency coefficients in sequence, and arrange the multiple emergency coefficients from large to small to generate the emergency coefficient sequence.

[0017] Among the possible implementation methods, an operation and maintenance solution optimization analysis is performed, and the following processing is carried out:

[0018] On the 3D simulation platform, a charging station simulation space is constructed based on the topology of the charging station distribution; the emergency coefficient sequence and the predetermined maintenance resources are rendered into the charging station simulation space to generate an optimization simulation space; within the optimization simulation space, the operation and maintenance scheme is optimized and analyzed based on the scheme evaluation function, and the optimal operation and maintenance scheme is output.

[0019] In a possible implementation, a solution evaluation function is constructed, and the following processing is performed:

[0020] The expression for the scheme evaluation function is:

[0021]

[0022] Among them, F i-1 For evaluating the adaptability of the next charging station, v1 is the weight for maintenance urgency, v2 is the weight for maintenance distance, and v3 is the weight for resource suitability. i-1 L represents the urgency factor for the next charging station. i-1 C represents the distance between the current location of the maintenance personnel and the next charging station. i-1 To determine the compatibility between operation and maintenance resources and the resource requirements of the next charging station, w1 represents the technology weight, w2 the equipment weight, w3 the material weight, and R... i-1 To ensure the compatibility of current operation and maintenance technologies with the technological requirements of the next charging station, G i-1 To ensure the compatibility between current maintenance equipment and the equipment requirements of the next charging station, H i-1 This is to assess the compatibility between current maintenance supplies and the needs of the next charging station.

[0023] The proposed multi-source sensing joint charging infrastructure operation and maintenance management system acquires monitoring datasets from multiple charging stations within a target area through a monitoring dataset acquisition module. These datasets include equipment status sets and environmental parameter sets. A fault prediction module inputs these datasets into a fault prediction channel, determines an appropriate prediction unit based on the environmental parameters, inputs the equipment status data into the prediction unit, and outputs multiple fault prediction datasets. Each fault prediction dataset corresponds one-to-one with a charging pile. A demand prediction module reads charging datasets from multiple charging stations within the target area and performs demand prediction within a predetermined time period, identifying multiple predicted charging demands. An operation and maintenance urgency analysis module performs operation and maintenance urgency analysis based on the multiple fault prediction datasets and predicted charging demands, generating an urgency coefficient sequence. An operation and maintenance scheme optimization analysis module optimizes the operation and maintenance scheme based on the urgency coefficient sequence, combined with the charging station distribution topology and predetermined maintenance resources, determining the optimal operation and maintenance scheme. Finally, an operation and maintenance management module manages the multiple charging stations within a predetermined time period according to the optimal scheme, achieving the technical effect of improving operation and maintenance efficiency and management level. Attached Figure Description

[0024] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0025] Figure 1 This is a schematic diagram of the structure of the multi-source sensing joint charging infrastructure operation and maintenance management system provided in the embodiments of this application.

[0026] Figure 2 This is a schematic diagram illustrating the process of constructing a fault prediction channel in the multi-source sensing joint charging infrastructure operation and maintenance management system provided in the embodiments of this application.

[0027] Figure labeling: Monitoring dataset acquisition module 10, fault prediction module 20, demand prediction module 30, maintenance urgency analysis module 40, maintenance solution optimization analysis module 50, maintenance management module 60. Detailed Implementation

[0028] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application.

[0029] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of this application will be provided in conjunction with the accompanying drawings. The described embodiments should not be considered as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0030] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" may be the same or different subsets of all possible embodiments and may be combined with each other without conflict. The terms "first" and "second" are used merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or modules not explicitly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only.

[0031] This application provides a multi-source sensing-integrated charging infrastructure operation and maintenance management system, such as... Figure 1 As shown, the system includes:

[0032] The monitoring dataset acquisition module 10 is used to acquire monitoring datasets from multiple charging stations within a target area. These datasets include equipment status sets and environmental parameter sets. Specifically, it determines the target area to be monitored, including all charging stations within that area, and identifies the specific equipment and environmental parameters to be monitored at each charging station. It configures and connects to various data sources, including but not limited to: sensor networks (connecting various sensors within the charging station, such as current sensors, voltage sensors, temperature sensors, humidity sensors, etc.), remote monitoring systems (accessing the charging station's remote monitoring system via a network interface to acquire real-time data), and databases (historical data is stored in the database; database connections are configured to access this data). Data is collected from data sources periodically or in real time. The collected raw data is cleaned to remove noise, duplicate data, invalid data, etc. The data is converted and standardized in terms of format and unit. The preprocessed data is divided into equipment status set and environmental parameter set. The equipment status set includes the working status (e.g., online, offline, fault) of various equipment in the charging station (such as charging piles, distribution boxes, etc.), current, voltage, power and other data reflecting the health and working conditions of the equipment. The environmental parameter set includes data on external environmental factors that may affect the operation of the charging station equipment, such as temperature, humidity, wind speed, light intensity and so on.

[0033] The fault prediction module 20 is used to input the equipment status set and environmental parameter set into the fault prediction channel, determine the appropriate prediction unit based on the environmental parameters, input the equipment status into the prediction unit, and output multiple fault prediction datasets, wherein each fault prediction data corresponds one-to-one with a charging pile. Specifically, the monitoring dataset acquisition module 10 receives the equipment status set and environmental parameter set. The environmental parameter set is analyzed in detail, including weighting, classifying, or clustering different environmental parameters to extract key environmental features. Based on the results of the environmental parameter analysis, the most suitable prediction model or algorithm for the current environmental conditions is selected as the appropriate prediction unit. The fault prediction channel is a channel including a model library or algorithm library, with each model or algorithm corresponding to a prediction unit, used to predict equipment status faults under different environmental conditions. The equipment status set data is preprocessed and formatted according to the input requirements of the prediction unit, and the preprocessed equipment status set data is input into the prediction unit. The prediction unit performs fault prediction calculations based on the input equipment status data and its internal model or algorithm, and outputs the prediction results, including the predicted fault type and the predicted fault node (i.e., which specific device or component has failed). The fault prediction results for each charging station are compiled into a dataset, which includes information such as charging station identification, predicted fault type, predicted fault time, and fault probability.

[0034] In one possible implementation, constructing a fault prediction channel includes: configuring environmental parameter thresholds, wherein the environmental parameter thresholds include a temperature threshold, a humidity threshold, and a power grid load threshold; dividing the temperature threshold, humidity threshold, and power grid load threshold according to a predetermined step size to obtain a temperature interval set, a humidity interval set, and a power grid load interval set; randomly enumerating within the temperature interval set, humidity interval set, and power grid load interval set to obtain multiple environmental parameter groups; and constructing the fault prediction channel based on the multiple environmental parameter groups, wherein each environmental parameter group includes a temperature interval, a humidity interval, and a power grid load interval.

[0035] Specifically, equipment specifications and historical operating data are collected to analyze the impact of temperature, humidity, and grid load on equipment performance. Based on the analysis results, combined with safety standards and empirical values, reasonable minimum and maximum thresholds are set for temperature, humidity, and grid load, respectively. Different predetermined step sizes are set according to the actual situation and prediction accuracy requirements of each parameter. For example, the temperature step size can be set to 1°C, the humidity step size to 5%, and the grid load step size to be set according to the sensitivity to load changes. Algorithms (such as arithmetic sequences or algorithms customized according to specific needs) are used to divide the threshold range of each parameter into multiple small intervals. An enumeration algorithm is designed to randomly select an interval from each set of intervals to form an environmental parameter group. The enumeration algorithm is repeated multiple times to generate a sufficient number of environmental parameter groups that represent possible operating environment scenarios. Based on multiple environmental parameter groups, machine learning or statistical methods are used to construct fault prediction models. These models can predict the operating status or fault probability of the equipment based on the input environmental parameter groups (i.e., temperature intervals, humidity intervals, and grid load intervals). This implementation method constructs a fault prediction channel by configuring environmental parameter thresholds, dividing interval sets, and randomly enumerating environmental parameter groups. It achieves comprehensive simulation and prediction of equipment operation under different environmental conditions, thereby improving the precision and accuracy of fault prediction channel prediction.

[0036] like Figure 2 As shown, in one possible implementation, constructing the fault prediction channel based on the multiple environmental parameter groups includes: randomly selecting a first environmental parameter group from the multiple environmental parameter groups, randomly selecting a first charging station from the multiple charging stations, and obtaining the first charging pile attribute of the first charging station; using the first environmental parameter group as a condition constraint and the first charging pile attribute as a device constraint, performing a source retrieval based on the Industrial Internet to obtain a first sample dataset, wherein the first sample data includes sample device status and sample fault data, and the sample fault data includes sample fault type and sample fault node; training a predetermined operator using the first sample dataset to generate a first fault prediction model; sequentially obtaining multiple fault prediction models for multiple charging stations, and constructing a first prediction unit based on the multiple fault prediction models; and constructing the fault prediction channel based on a binary tree according to the mapping relationship between the first environmental parameter group and the first prediction unit.

[0037] Specifically, a random number generator or sampling algorithm is used to select an element from the set of environmental parameter groups as the first environmental parameter group. Similarly, a charging station is randomly selected from the list of charging stations as the first charging station. The charging station management system or database is accessed to query the model, specifications, manufacturer, and other attribute information of the charging piles in the first charging station, and this attribute information is saved as the first charging pile attributes. The first environmental parameter group and the first charging pile attributes are used as query conditions and input into the data retrieval interface of the industrial internet platform. Based on the query conditions, the industrial internet platform retrieves historical data records similar to the current scenario from the database, extracts sample equipment status and sample fault data from these records, including sample fault type and sample fault node (i.e., the time point of fault occurrence), and organizes the retrieved data into the first sample dataset. A machine learning or deep learning framework is selected, and predefined operators (such as neural networks, decision trees, etc.) are configured. The first sample dataset is divided into a training set and a validation set. The operators are trained using the training set, and the parameters of the operators are adjusted through iterative optimization algorithms. The model performance is evaluated using the validation set, and necessary adjustments are made to finally generate the first fault prediction model. For each charging station, the processes of random selection, attribute acquisition, data retrieval, and model training are repeatedly executed. The resulting fault prediction models for each charging station are combined into a prediction unit, which can simultaneously predict fault conditions at multiple charging stations. A binary tree structure is designed, where each node represents an environmental parameter range. Based on the division of environmental parameter groups, the nodes of the binary tree are associated with the corresponding environmental parameter ranges. The prediction unit is used as the leaf node of the binary tree, and each leaf node corresponds to a fault prediction model for a charging station. A path is constructed from the root node to the leaf node, representing the mapping relationship from the environmental parameter group to the prediction unit. When a new environmental parameter is input, the corresponding prediction unit is found by traversing the binary tree, and the fault prediction result of that unit is output. This implementation method, by constructing a fault prediction channel based on a binary tree, realizes fault prediction and monitoring of multiple charging stations under various environmental conditions, achieving the technical effect of improving the accuracy and timeliness of fault prediction.

[0038] In one possible implementation, generating a first fault prediction model includes: configuring predetermined operators, wherein the predetermined operators include a digital twin model, a backpropagation neural network, and a random decision forest; dividing the first sample dataset into ten equal parts and selecting them ten times with replacement to construct a first sample set; iterating three times to obtain three sample sets; using the sample device status as input and sample fault data as supervision, using the three sample sets to perform supervised training and cross-validation on the digital twin model, the backpropagation neural network, and the random decision forest respectively, to obtain a first prediction sub-model, a second prediction sub-model, and a third prediction sub-model, wherein the prediction sub-model is identified by convergence; configuring weight proportions based on the convergence, and constructing the first fault prediction model according to the weight proportions, the first prediction sub-model, the second prediction model, and the third prediction model.

[0039] Specifically, digital twin models, backpropagation neural networks, and random decision forests were selected as prediction operators. Preliminary parameter settings were performed for each operator, including network structure (for neural networks), number and depth of trees (for random decision forests), and the accuracy of the physical model and simulation step size in the digital twin model. The first sample dataset (containing equipment status and fault data) was divided into ten equal parts, each with similar distribution and characteristics. One part was randomly selected with replacement from each of the ten parts, repeated ten times to form a new sample set. This sampling process was repeated three times to obtain three different sample sets, each containing different combinations of the original data. Using sample device states as input and sample fault data as supervisory information, a digital twin model, a backpropagation neural network, and a random decision forest are trained separately. Three different sample sets are used to train each operator. After each training iteration, the model's predictive performance on the remaining data is evaluated. By alternating between the test and training sets, each sample set is ensured to serve as a test set, comprehensively assessing the model's generalization ability. During training, the model's convergence is monitored, i.e., the trend of model performance changes with increasing training epochs. Models with high convergence indicate that their performance has stabilized and requires no further training. Based on the convergence of each predictive sub-model, its weight in the final predictive model is configured, with higher-convergence models receiving higher weights. Based on these weights, the first, second, and third predictive sub-models are integrated (using weighted average or other strategies) to construct the first fault prediction model. This implementation combines multiple predictive operators, leveraging their respective advantages to improve the overall prediction accuracy.

[0040] The demand forecasting module 30 is used to read charging datasets from multiple charging stations within a target area, and to perform demand forecasting within a predetermined time period based on these datasets, determining multiple predicted charging demands. Specifically, it reads charging datasets from multiple charging stations within the target area from a database, file system, or real-time data source. These datasets contain historical charging records, such as charging start time, end time, charging amount, and charging pile usage. The collected charging datasets are cleaned to remove noisy data, outliers, and duplicate records. The data is formatted to ensure consistency of timestamps and correctness of data types. Features related to charging demand are extracted from the preprocessed data, including time-series data of historical charging amounts, holiday information, weather conditions (such as temperature, humidity, and rainfall), and special events (such as large-scale events and traffic control). Based on the feature selection results, the prediction model is trained using the historical charging dataset. The prediction model can be a time-series analysis model (ARIMA, LSTM, etc.), a machine learning model (regression model, decision tree, random forest, etc.), or a deep learning model. The extracted features are input into the trained model to predict charging demand within a predetermined time period. The prediction results can include the expected charging volume of each charging station in a future time period, the expected charging pile utilization rate, the maximum number of charging piles in demand, etc. Multiple predicted charging demands are output, each corresponding to a charging station and a predetermined time period.

[0041] In one possible implementation, demand forecasting within a predetermined time period based on the charging dataset includes: randomly selecting a first charging station and obtaining first charging data for the first charging station, wherein the first charging data includes the number of first working charging piles; retrieving charging logs of the first charging station under the constraint of the predetermined time period to obtain a historical charging dataset, wherein the historical charging data includes the starting number of working charging piles and a limit on the number of working charging piles, and the historical charging data is identified by time intervals; configuring dynamic weights based on time intervals, and performing weighted fusion on the historical charging dataset according to the dynamic weights to obtain first standard charging data, wherein the first standard charging data includes the number of first standard starting points and a first standard quantity limit; calculating the quantity deviation between the number of first working charging piles and the number of first standard starting points, compensating the first standard quantity limit based on the quantity deviation to obtain a first predicted charging demand, and adding it to the plurality of predicted charging demands.

[0042] Specifically, from all charging stations, a random number generator or other random sampling method is used to select one as the first charging station. The current status information of the first charging station, including the number of currently operating charging piles, is obtained in real time through the charging station management system or API interface, and is recorded as the first number of operating charging piles. Based on the predicted demand, a specific time range is set as a constraint, such as the past week or month. Based on the charging station's data recording system, all charging logs of the first charging station within the predetermined time period are retrieved. From the charging logs, the starting number of operating charging piles at each time point (i.e., the number of operating charging piles at the beginning of that time point) and the maximum number of operating charging piles (i.e., the maximum number of operating charging piles within the predetermined time period) are extracted, and the time interval of each data point is recorded. Different weights are configured for each historical charging data point based on the length of the time interval; for example, data points closer to the predicted time point have a higher weight. Using the configured dynamic weights, the historical charging dataset is weighted to obtain the average or weighted sum of the weighted starting number of operating charging piles and the maximum number of operating charging piles, which are used as the first standard starting number and the first standard number limit, respectively. The current number of working charging piles is compared with the number of standard starting points to calculate the quantity deviation. Based on the calculated deviation, the first standard quantity limit is appropriately adjusted to reflect the actual demand and potential changes of the charging stations. Specific compensation methods may include directly adding or subtracting the deviation value, or adjusting proportionally. The compensated first standard quantity limit is used as the predicted charging demand for the first charging station within a predetermined time period. This predicted charging demand is then aggregated with the predictions of other charging stations to form an overall charging demand prediction. This implementation, through dynamic weight configuration and quantity deviation compensation, better adapts to the impact of time factors on charging station demand, achieving a more accurate prediction of charging demand in the future and improving the adaptability of the prediction results.

[0043] The maintenance urgency analysis module 40 is used to perform maintenance urgency analysis based on the multiple fault prediction datasets and the multiple predicted charging demands, generating an urgency coefficient sequence. Specifically, it receives multiple fault prediction datasets from the fault prediction module 20, each dataset containing information such as the predicted fault type, predicted fault time, and fault probability for the corresponding charging pile; and receives multiple predicted charging demands from the demand prediction module 30, each demand containing information such as the expected charging volume and expected charging pile utilization rate for the corresponding charging station in a future time period. The fault prediction datasets and predicted charging demands are integrated to form a comprehensive maintenance information table for each charging station in the same time period. The table includes fields such as charging station identifier, timestamp, predicted fault information (type, time, probability), expected charging volume, and expected charging pile utilization rate. An urgency coefficient calculation model is designed, which can integrate fault prediction results and charging demand prediction results. The urgency coefficient can be calculated based on the severity and urgency of the fault, the matching degree between the predicted charging demand and the charging pile capacity, and other factors affecting maintenance urgency (such as weather conditions, special events, etc.). For example, a higher urgency factor is assigned to serious faults that have a high probability of occurring within a short period of time; if the predicted charging demand is much higher than the available capacity of charging piles, the urgency factor is increased to reflect the potential risk of charging service interruption. Based on the urgency factor calculation model, an urgency factor is generated for each charging station in each time period. The urgency factor quantifies the urgency of charging station operation and maintenance. The urgency factors of all charging stations in all time periods are arranged in chronological order to form an urgency factor sequence. This sequence visually reflects the changing trend and distribution of the urgency of charging station operation and maintenance throughout the entire target area.

[0044] In one possible implementation, generating the emergency coefficient sequence includes: randomly selecting a first fault prediction dataset from the plurality of fault prediction datasets, counting the number of first fault prediction data that meet the predetermined time period, determining the first fault quantity, and calculating the first qualified quantity; obtaining the first predicted charging demand corresponding to the first fault prediction dataset, and setting the ratio of the first predicted charging demand to the first qualified quantity as the first emergency coefficient; sequentially obtaining a plurality of emergency coefficients, and arranging the plurality of emergency coefficients from largest to smallest to generate the emergency coefficient sequence.

[0045] Specifically, all fault prediction datasets are listed, and a dataset is selected as the first fault prediction dataset using a random number generator or similar mechanism. Each fault prediction entry in the first fault prediction dataset is iterated through, and it is checked whether the predicted fault time of each entry falls within a predetermined time period. The number of faults falling within the predetermined time period is counted to obtain the first fault count. The first fault count is then subtracted from the total number of charging piles at the first charging station corresponding to the first fault prediction dataset to obtain the first qualified count, representing the number of charging piles available within the predetermined time period. The charging demand prediction data associated with the first fault prediction dataset is found, and the total predicted charging demand is extracted as the first predicted charging demand. The ratio of the first predicted charging demand to the first qualified count is calculated, and this ratio is set as the first urgency coefficient. The above steps are repeated for each remaining fault prediction dataset, and all calculated urgency coefficients are collected and sorted according to their magnitude to generate a sorted urgency coefficient sequence. This implementation method obtains the urgency coefficient through basic mathematical operations, which is simple and easy to implement, achieving the technical effect of improving the efficiency of urgency coefficient calculation.

[0046] The operation and maintenance (O&M) scheme optimization and analysis module 50 is used to perform O&M scheme optimization analysis based on the urgency coefficient sequence, combined with the charging station distribution topology and predetermined maintenance resources, to determine the optimal O&M scheme. Specifically, it receives the urgency coefficient sequence from the O&M urgency analysis module 40, understands and analyzes the O&M urgency of each charging station in different time periods, identifies charging stations and time periods with high urgency coefficients, and focuses these areas on O&M scheme optimization. It acquires the geographical location information and distribution of charging stations, constructs a charging station distribution topology map, and analyzes the distance, coverage, and interdependencies between charging stations to consider inter-regional synergy when optimizing the O&M scheme. It statistically analyzes and statistically analyzes predetermined maintenance resources, including maintenance personnel, maintenance vehicles, spare parts, etc., and evaluates the availability, allocation capability, and response time of maintenance resources to ensure timely and effective fault repair and equipment maintenance during the O&M process. Based on the emergency coefficient sequence, charging station distribution topology, and scheduled maintenance resources, multiple possible operation and maintenance plans are formulated. Each operation and maintenance plan is simulated and evaluated to assess factors such as feasibility, cost-effectiveness, and impact on business. Optimization algorithms (such as genetic algorithms and simulated annealing algorithms) are used to compare and select multiple operation and maintenance plans to determine the optimal operation and maintenance plan.

[0047] In one possible implementation, the operation and maintenance scheme optimization analysis includes: constructing a charging station simulation space based on the charging station distribution topology on a three-dimensional simulation platform; rendering the emergency coefficient sequence and the predetermined maintenance resources into the charging station simulation space to generate an optimization simulation space; and performing operation and maintenance scheme optimization analysis based on the scheme evaluation function within the optimization simulation space to output the optimal operation and maintenance scheme.

[0048] Specifically, the technology stack and toolset of the 3D simulation platform are determined. Data such as the geographical location, scale, and type of charging stations are collected. Using 3D modeling software or platform functions, 3D models are constructed based on the topological distribution of the charging stations. These models are then imported into the 3D simulation platform, and the relative positions and interaction logic between the models are set. Data rendering functions are developed or integrated into the 3D simulation platform. An emergency coefficient sequence is used as input data, and a visualization method (such as color depth, icon size, etc.) is designed to represent the urgency level of different charging stations. Scheduled maintenance resources are also added to the charging station simulation space in a visual manner, representing resource availability and distribution. The rendering effect is dynamically adjusted based on the emergency coefficient and resource distribution to generate an optimization simulation space. A scheme evaluation function is defined, which is a function constructed based on multiple evaluation indicators to evaluate the merits of different operation and maintenance schemes. Within the optimization simulation space, algorithms (such as genetic algorithms and particle swarm optimization algorithms) are used to generate and simulate different operation and maintenance schemes. The scheme evaluation function is applied to each scheme to calculate its score or evaluation indicator. Based on the score or evaluation indicator, the operation and maintenance schemes are ranked or filtered, and the optimal operation and maintenance scheme is output. The optimal operation and maintenance scheme is the one with the highest score that satisfies specific constraints. This implementation method constructs a charging station simulation space through a 3D simulation platform and presents the emergency coefficient sequence and scheduled maintenance resources in a visual manner, enabling operation and maintenance personnel to intuitively understand the current operation and maintenance needs and resource distribution, thereby achieving the technical effect of more easily discovering potential problems and optimization points.

[0049] In one possible implementation, the scheme evaluation function is constructed, including:

[0050] The expression for the scheme evaluation function is:

[0051]

[0052] Among them, F i-1 For evaluating the adaptability of the next charging station, v1 is the weight for maintenance urgency, v2 is the weight for maintenance distance, and v3 is the weight for resource suitability. i-1 L represents the urgency factor for the next charging station. i-1 C represents the distance between the current location of the maintenance personnel and the next charging station. i-1 To determine the compatibility between operation and maintenance resources and the resource requirements of the next charging station, w1 represents the technology weight, w2 the equipment weight, w3 the material weight, and R... i-1 To ensure the compatibility of current operation and maintenance technologies with the technological requirements of the next charging station, G i-1 To ensure the compatibility between current maintenance equipment and the equipment requirements of the next charging station, H i-1This refers to the compatibility between current maintenance resources and the resource requirements of the next charging station. Specifically, the scheme evaluation function consists of three parts: urgency coefficient, maintenance distance, and resource compatibility. The urgency coefficient reflects the urgency of the maintenance needs of the next charging station within a predetermined time period. The maintenance distance is used to assess maintenance efficiency; a shorter maintenance distance means higher efficiency and lower costs. Resource compatibility assesses the degree of matching between current maintenance resources (technology, equipment, and materials) and the resource requirements of the next charging station. Weighting coefficients in the expression are used to balance the importance of different factors in the scheme evaluation. This implementation comprehensively evaluates the merits of maintenance schemes by integrating multiple factors such as urgency, maintenance distance, and resource compatibility, avoiding the limitations of optimizing a single indicator and achieving the technical effect of improving the scientific and rational nature of decision-making.

[0053] The operation and maintenance management module 60 is used to perform operation and maintenance management on multiple charging stations within a predetermined time period according to the optimal operation and maintenance plan. Specifically, the operation and maintenance management module 60 receives the optimal operation and maintenance plan from the operation and maintenance plan optimization and analysis module 50, interprets the optimal operation and maintenance plan, and clarifies the specific operation and maintenance tasks for each charging station within the predetermined time period, including inspection, repair, and preventive maintenance. Based on the nature, urgency, and resource availability of the tasks, the operation and maintenance tasks are assigned to the corresponding operation and maintenance teams or individuals, and clear task completion time and quality requirements are set. The operation and maintenance teams or individuals perform the corresponding operation and maintenance work according to the assigned tasks, such as inspection, fault diagnosis and repair, and preventive maintenance. During the operation and maintenance process, the operating status and maintenance progress of the charging stations are monitored in real time, and the operation and maintenance plan is adjusted as necessary based on the actual operation and maintenance situation and feedback. This application employs a monitoring dataset acquisition module that integrates equipment status sets and environmental parameter sets to achieve comprehensive perception and fusion of multi-source data. A fault prediction module, combined with environmental parameters, adaptively selects prediction units to perform refined analysis of equipment status, outputting fault prediction datasets corresponding one-to-one with charging piles. A demand prediction module, based on the charging dataset, dynamically predicts charging demand within a predetermined time period. An operation and maintenance urgency analysis module combines fault prediction results and demand prediction results to generate an urgency coefficient sequence. An operation and maintenance scheme optimization analysis module further combines the charging station distribution topology and predetermined maintenance resources, using optimization algorithms to determine the optimal operation and maintenance scheme. Finally, an operation and maintenance management module, based on the optimal scheme, performs intelligent operation and maintenance management of multiple charging stations within a predetermined time period. These technical means achieve reasonable allocation and efficient utilization of operation and maintenance resources, resulting in improved operation and maintenance efficiency and management level.

[0054] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0055] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application. In some cases, the actions or steps described in this application can be performed in a different order than that shown in the embodiments and still achieve the desired results. Furthermore, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

Claims

1. A multi-source perception joint charging infrastructure operation and maintenance management system, characterized in that, The method comprises the following steps: a monitoring data set acquisition module is configured to acquire monitoring data sets of a plurality of charging stations in a target area, wherein the monitoring data sets comprise a device state set and an environmental parameter set; a fault prediction module is configured to input the device state set and the environmental parameter set into a fault prediction channel, determine an adaptive prediction unit based on the environmental parameters, input the device state into the adaptive prediction unit, and output a plurality of fault prediction data sets, wherein the fault prediction data and the charging piles are one-to-one corresponding; a demand prediction module is configured to read charging data sets of a plurality of charging stations in a target area, perform demand prediction within a predetermined time period based on the charging data sets, and determine a plurality of predicted charging demands; an operation and maintenance urgency analysis module is configured to perform operation and maintenance urgency analysis based on the plurality of fault prediction data sets and the plurality of predicted charging demands, and generate an urgency coefficient sequence; an operation and maintenance scheme optimization analysis module is configured to perform operation and maintenance scheme optimization analysis based on the urgency coefficient sequence, in combination with charging station distribution topology and predetermined maintenance resources, and determine an optimal operation and maintenance scheme; an operation and maintenance management module is configured to perform operation and maintenance management of a plurality of charging stations within a predetermined time period according to the optimal operation and maintenance scheme; wherein the demand prediction within a predetermined time period based on the charging data set comprises: randomly selecting a first charging station and acquiring first charging data of the first charging station, wherein the first charging data comprises a first number of working charging piles; retrieving charging logs of the first charging station with the predetermined time period as a constraint, and acquiring a historical charging data set, wherein the historical charging data comprises a starting number of working charging piles and a limit value of the number of working charging piles, and the historical charging data is marked with a time interval; configuring a dynamic weight based on the time interval, weighting and fusing the historical charging data set according to the dynamic weight to obtain first standard charging data, wherein the first standard charging data comprises a first standard starting number and a first standard limit value; calculating the number deviation of the first number of working charging piles and the first standard starting number, compensating the first standard limit value based on the number deviation to obtain a first predicted charging demand, and adding it to the plurality of predicted charging demands. 2.The multi-source perception joint charging infrastructure operation and management system of claim 1, wherein, The fault prediction channel is constructed, comprising: configuring environmental parameter thresholds, wherein the environmental parameter thresholds comprise temperature thresholds, humidity thresholds, and power grid load thresholds; dividing the temperature thresholds, humidity thresholds, and power grid load thresholds according to a predetermined step size to obtain a temperature interval set, a humidity interval set, and a power grid load interval set; randomly enumerating in the temperature interval set, humidity interval set, and power grid load interval set to obtain a plurality of environmental parameter groups, and constructing the fault prediction channel based on the plurality of environmental parameter groups, wherein each environmental parameter group comprises a temperature interval, a humidity interval, and a power grid load interval.

3. The multi-source perception joint charging infrastructure operation and management system according to claim 2, wherein, The fault prediction channel is constructed based on the plurality of environmental parameter groups, comprising: randomly selecting a first environmental parameter group from the plurality of environmental parameter groups and a first charging station from the plurality of charging stations, and acquiring first charging pile attributes of the first charging station; The first sample data set is used to train a predetermined operator to generate a first fault prediction model. A plurality of fault prediction models of a plurality of charging stations are sequentially obtained, and a first prediction unit is constructed based on the plurality of fault prediction models. Based on a binary tree, a mapping relationship between the first environment parameter group and the first prediction unit is constructed to build the fault prediction channel. Generating a first fault prediction model includes:

4. The multi-source perception joint charging infrastructure operation and management system according to claim 3, characterized in that, Configuring a predetermined operator, wherein the predetermined operator includes a digital twin model, a BP neural network, and a random decision forest; The first sample data set is divided into ten equal parts, and ten times are selected with replacement to construct a first sample set, which is iteratively selected three times to obtain three sample sets. Using the three sample sets, the digital twin model, the BP neural network, and the random decision forest are supervised trained and cross-validated with the sample device state as input and the sample fault data as supervision to obtain a first prediction sub-model, a second prediction sub-model, and a third prediction sub-model, wherein the prediction sub-models are identified by convergence degree. Based on the convergence degree, a weight ratio is configured, and the first fault prediction model is constructed based on the weight ratio, the first prediction sub-model, the second prediction sub-model, and the third prediction sub-model. Generating an emergency coefficient sequence includes: 5.The multi-source perception jointed charging infrastructure operation management system of claim 1, wherein, Randomly selecting a first fault prediction data set from the plurality of fault prediction data sets, counting the number of first fault prediction data that meets the predetermined time period, determining the first fault number, and calculating the first qualified number; The first prediction charging demand corresponding to the first fault prediction data set is obtained, and the ratio of the first prediction charging demand to the first qualified number is set as the first emergency coefficient; A plurality of emergency coefficients are sequentially obtained and arranged in descending order to generate the emergency coefficient sequence. Optimization analysis of operation and maintenance scheme includes: 6.The multi-source perception jointed charging infrastructure operation management system of claim 1, wherein, On a three-dimensional simulation platform, a charging station simulation space is constructed according to the charging station distribution topology; The emergency coefficient sequence and the predetermined maintenance resource are rendered to the charging station simulation space to generate an optimization simulation space; In the optimization simulation space, operation and maintenance scheme optimization analysis is performed based on a scheme evaluation function, and the optimal operation and maintenance scheme is output. Constructing a scheme evaluation function includes:

7. The multi-source perception jointed charging infrastructure operation and management system according to claim 6, characterized in that, ​ The expression of the scheme evaluation function is: wherein, is the evaluation fitness of the next charging station, is the operation and maintenance urgency weight, is the operation and maintenance distance weight, is the resource adaptation weight, is the emergency coefficient of the next charging station, is the interval distance between the current position of the operation and maintenance personnel and the next charging station, is the adaptation degree of the operation and maintenance resource and the resource demand of the next charging station, is the technology weight, is the equipment weight, is the material weight, is the adaptation degree of the current operation and maintenance technology and the technology demand of the next charging station, is the adaptation degree of the current operation and maintenance equipment and the equipment demand of the next charging station, is the adaptation degree of the current operation and maintenance material and the material demand of the next charging station.

Images