Cross-platform remote management scheduling method and system for charging piles

Abstract

The invention discloses a charging pile cross-platform remote management scheduling method and system, and relates to the technical field of charging pile adaptive control, and the method comprises the steps: enabling a plurality of charging pile platforms to be connected to a central scheduling server, constructing a cross-platform charging pile resource pool, and carrying out the real-time monitoring to obtain the monitoring flow of each charging pile; carrying out fault prediction, establishing a charging pile fault space, and carrying out multi-level division on a cross-platform charging pile resource pool to obtain a first cluster, a second cluster and a third cluster of charging piles; performing cross-platform scheduling compensation on the third cluster according to the first cluster and the second cluster to generate a scheduling compensation first group; a scheduling response evaluation model is introduced to perform response evaluation optimization, a second scheduling compensation group is constructed, variation joint optimization is performed according to the scheduling compensation variation capacity, and a scheduling compensation strategy is generated; and performing compensation management on the charging pile resource pool. The technical problem of low management efficiency and reliability of existing charging pile cross-platform remote scheduling is solved, and the technical effect of improving the scheduling management efficiency and reliability is achieved.

Description

Technical Field

[0001] This application relates to the field of adaptive control technology for charging piles, and in particular to a cross-platform remote management and scheduling method and system for charging piles. Background Technology

[0002] With the increasing popularity of electric vehicles, cross-platform management and efficient scheduling of charging stations have become key issues in ensuring the charging experience for electric vehicle users, improving the utilization rate of charging facilities, and achieving optimal energy allocation. Currently, the main approach to solving this problem is for each charging station platform to operate independently or achieve limited interconnection through simple protocols, employing scheduling strategies based on fixed rules or human experience for resource allocation. However, these current methods lack cross-platform resource integration and dynamic sensing capabilities, leading to delayed response to charging station faults, coarse scheduling decisions, uneven load distribution between platforms, and difficulty in adapting to fluctuations in charging demand and changes in equipment status, resulting in both resource idleness and overload.

[0003] At present, the cross-platform remote scheduling of charging piles has technical problems of low management efficiency and low reliability. Summary of the Invention

[0004] This application provides a cross-platform remote management and scheduling method and system for charging piles. It integrates data from multiple charging pile platforms to construct a unified resource pool, acquires real-time monitoring data streams of each charging pile's operating status, performs fault prediction analysis based on the monitoring data, establishes a charging pile fault space, and divides the charging piles into three clusters: high reliability, medium reliability, and those requiring compensation. The high-reliability and medium-reliability clusters are used to implement cross-platform scheduling compensation for the clusters requiring compensation, forming a preliminary scheduling scheme. The scheme is then optimized and screened using a response evaluation model, and further optimized using variability capacity analysis, ultimately generating the optimal scheduling strategy. This strategy is executed by a central scheduling server, completing the dynamic compensation management of charging pile resources across all platforms. These technical means solve the technical problems of low management efficiency and reliability in existing cross-platform remote scheduling of charging piles, achieving the technical effect of improving the efficiency and reliability of scheduling management.

[0005] This application provides a cross-platform remote management and scheduling method for charging piles, comprising: connecting multiple charging pile platforms to a central scheduling server to construct a cross-platform charging pile resource pool, and monitoring the cross-platform charging pile resource pool in real time to obtain monitoring streams for each charging pile; performing fault prediction based on the monitoring streams for each charging pile to establish a charging pile fault space, and dividing the cross-platform charging pile resource pool into multiple levels based on the charging pile fault space to obtain a first cluster, a second cluster, and a third cluster of charging piles; performing cross-platform scheduling compensation on the third cluster of charging piles based on the first cluster and the second cluster to generate a first group of scheduling compensation; introducing a scheduling response evaluation model to optimize the first group of scheduling compensation through response evaluation, constructing a second group of scheduling compensation, and performing joint optimization of the second group of scheduling compensation based on the scheduling compensation variation capacity to generate a scheduling compensation strategy; and performing compensation management on the cross-platform charging pile resource pool based on the central scheduling server and the scheduling compensation strategy.

[0006] In a possible implementation, fault prediction is performed based on the monitoring streams of each charging pile to establish a charging pile fault space, and the following processing is performed: The first charging pile monitoring stream corresponding to the first charging pile is extracted based on the monitoring streams of each charging pile, and the first charging pile fault record set corresponding to the first charging pile is retrieved; a first charging pile fault prediction model is trained based on the first charging pile fault record set, the first charging pile fault prediction model including a charging pile fault risk prediction model and a charging pile fault type prediction model; fault records of the same model of charging pile are retrieved based on the specification and model information of the first charging pile to obtain a supplementary charging pile fault record set; the first charging pile fault prediction model is reinforced and trained based on the supplementary charging pile fault record set to obtain a first charging pile fault prediction channel; the first charging pile monitoring stream is input into the first charging pile fault prediction channel to obtain a first charging pile fault prediction result, and the first charging pile fault prediction result is added to the charging pile fault space.

[0007] In a possible implementation, the cross-platform charging pile resource pool is divided into multiple levels based on the charging pile fault space to obtain a first cluster, a second cluster, and a third cluster of charging piles. The following processing is then performed: determining whether the fault risk coefficient of each charging pile in the charging pile fault space is less than the charging pile fault risk threshold, and obtaining multiple fault risk judgment results; dividing the cross-platform charging pile resource pool based on the multiple fault risk judgment results, establishing the first cluster of charging piles with a fault risk coefficient less than the charging pile fault risk threshold, and a risky charging pile cluster with a fault risk coefficient greater than or equal to the charging pile fault risk threshold; determining whether the predicted fault type of each charging pile corresponding to the risky charging pile cluster belongs to an immediately recoverable fault, and obtaining multiple fault type judgment results; dividing the risky charging pile cluster based on the multiple fault type judgment results, and obtaining the second cluster of charging piles corresponding to immediately recoverable faults, and the third cluster of charging piles corresponding to non-immediately recoverable faults.

[0008] In a possible implementation, a scheduling response evaluation model is introduced to optimize the response of the first scheduling compensation group, constructing a second scheduling compensation group, and performing the following processes: evaluating each scheduling compensation scheme within the first scheduling compensation group according to the scheduling response evaluation model to obtain multiple scheduling response evaluation results; constructing scheduling response evaluation constraints, including resource conflict risk constraints, scheduling execution risk constraints, and power grid response risk constraints; performing anomaly verification on the multiple scheduling response evaluation results according to the scheduling response evaluation constraints to obtain multiple response evaluation verification results; and optimizing and filtering the first scheduling compensation group based on the multiple response evaluation verification results to generate the second scheduling compensation group.

[0009] In a possible implementation, the scheduling compensation schemes within the first group of scheduling compensations are evaluated according to the scheduling response evaluation model to obtain multiple scheduling response evaluation results. The following processing is then performed: a first scheduling compensation scheme is extracted from the first group of scheduling compensations; based on the first and second clusters of charging piles, multiple simulated scheduling operations are performed on the third cluster of charging piles according to the first scheduling compensation scheme to obtain multiple scheduling simulation datasets; confidence fusion is performed on the multiple scheduling simulation datasets to obtain a first scheduling simulation confidence set; the scheduling response evaluation model includes a resource conflict risk evaluation model, a scheduling execution risk evaluation model, and a power grid response risk evaluation model; the first scheduling simulation confidence set is input into the scheduling response evaluation model to obtain the first scheduling response evaluation result.

[0010] In a possible implementation, the scheduling compensation second group is subjected to joint optimization based on the scheduling compensation variation capacity to generate a scheduling compensation strategy. The following processes are then performed: weight allocation is performed based on the scheduling response evaluation element set to establish a comprehensive scheduling risk analysis model, where the scheduling response evaluation element set includes resource conflict risk, scheduling execution risk, and grid response risk; based on the scheduling compensation variation capacity, the scheduling compensation second group is subjected to optimization based on the comprehensive scheduling risk analysis model to establish a compensation variation optimization first group; based on the scheduling compensation variation capacity, the compensation variation optimization first group is further optimized based on the comprehensive scheduling risk analysis model to establish a compensation variation optimization Q-th group, where Q is a positive integer greater than 1; the scheduling compensation second group, the compensation variation optimization first group, ..., the compensation variation optimization Q-th group are jointly optimized based on the comprehensive scheduling risk threshold to generate a scheduling compensation third group; the scheduling compensation third group is optimized to maximize scheduling compensation efficiency to obtain the scheduling compensation strategy.

[0011] In a possible implementation, based on the scheduling compensation variation capacity, the second scheduling compensation group is subjected to variation optimization according to the scheduling comprehensive risk analysis model to establish a first compensation variation optimization group, and the following processes are performed: comprehensive risk calculation is performed on the second scheduling compensation group according to the scheduling comprehensive risk analysis model to obtain a scheduling comprehensive risk set; variation capacity is allocated to the second scheduling compensation group based on the scheduling compensation variation capacity and the scheduling comprehensive risk set to obtain a variation capacity allocation set; variation adjustment is performed on the second scheduling compensation group according to the variation capacity allocation set to generate a first compensation variation group; and response evaluation optimization is performed on the first compensation variation group according to the scheduling response evaluation model to generate the first compensation variation optimization group.

[0012] In a possible implementation, the cross-platform charging pile resource pool is monitored in real time to obtain the monitoring stream of each charging pile, and the following processing is performed: each charging pile in the cross-platform charging pile resource pool is monitored in real time to obtain multiple monitoring datasets; the multiple monitoring datasets are monitored in real time to generate the monitoring stream of each charging pile.

[0013] In a possible implementation, the following process is performed: fault operation and maintenance of the cross-platform charging pile resource pool is carried out according to the charging pile fault space.

[0014] This application also provides a cross-platform remote management and scheduling system for charging piles, comprising: a cross-platform charging pile resource pool construction module, used to connect multiple charging pile platforms to a central scheduling server, construct a cross-platform charging pile resource pool, and monitor the cross-platform charging pile resource pool in real time to obtain monitoring streams for each charging pile; a multi-level partitioning module for the resource pool, used to perform fault prediction based on the monitoring streams for each charging pile, establish a charging pile fault space, and partition the cross-platform charging pile resource pool into a first cluster, a second cluster, and a third cluster of charging piles based on the charging pile fault space; a cross-platform scheduling compensation module, used to perform cross-platform scheduling compensation on the third cluster of charging piles based on the first cluster and the second cluster, generating a first group of scheduling compensation; a mutation joint optimization module, used to introduce a scheduling response evaluation model to perform response evaluation optimization on the first group of scheduling compensation, construct a second group of scheduling compensation, and perform mutation joint optimization on the second group of scheduling compensation based on the scheduling compensation mutation capacity, generating a scheduling compensation strategy; and a compensation management module, used to perform compensation management on the cross-platform charging pile resource pool based on the central scheduling server and the scheduling compensation strategy.

[0015] This application proposes a cross-platform remote management and scheduling method and system for charging piles. First, multiple charging pile platforms are connected to a central scheduling server to construct a cross-platform charging pile resource pool. The pool is monitored in real time to obtain the monitoring streams for each charging pile. Next, fault prediction is performed based on these monitoring streams to establish a charging pile fault space. The cross-platform charging pile resource pool is then divided into three clusters based on these fault spaces: a first cluster, a second cluster, and a third cluster. Cross-platform scheduling compensation is then performed on the third cluster based on the first and second clusters to generate a first scheduling compensation group. A scheduling response evaluation model is then introduced to optimize the first scheduling compensation group, constructing a second scheduling compensation group. Furthermore, the second scheduling compensation group undergoes joint optimization based on its variation capacity to generate a scheduling compensation strategy. Finally, the central scheduling server manages the cross-platform charging pile resource pool according to the scheduling compensation strategy, thereby improving the efficiency and reliability of scheduling management. Attached Figure Description

[0016] 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.

[0017] Figure 1 This is a flowchart illustrating a cross-platform remote management and scheduling method for charging piles, as provided in an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of the structure of a cross-platform remote management and scheduling system for charging piles provided in an embodiment of this application.

[0019] Figure labeling: Cross-platform charging pile resource pool construction module 10, resource pool multi-level partitioning module 20, cross-platform scheduling and compensation module 30, mutation joint optimization module 40, compensation management module 50. Detailed Implementation

[0020] The above description is merely 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, specific embodiments of this application are given below.

[0021] 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.

[0022] In the following description, references to "some embodiments" describe a subset of all possible embodiments. However, it is understood that "some embodiments" can be the same or different subsets of all possible embodiments and can 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.

[0023] This application provides a cross-platform remote management and scheduling method for charging piles, such as... Figure 1 As shown, the method includes: Step S100: Connect multiple charging pile platforms to the central scheduling server to build a cross-platform charging pile resource pool, and monitor the cross-platform charging pile resource pool in real time to obtain the monitoring stream of each charging pile.

[0024] Specifically, a charging pile platform refers to a system that manages charging piles, provided by the charging pile manufacturer or operating company, used to control the operation of charging piles and record charging data. The central dispatch server is a server used for centralized management and dispatching of charging pile resources, receiving charging pile data, analyzing data, and generating dispatch instructions. Multiple charging pile platforms are connected to the central dispatch server using network communication protocols (such as TCP / IP, MQTT, etc.). Each charging pile platform sends real-time data (such as charging pile status, power level, usage, etc.) to the central dispatch server through its communication interface. On the central dispatch server, a charging pile resource pool database is created using a database management system (such as MySQL, MongoDB, etc.). Information on the connected charging piles (including charging pile ID, location, platform, technical parameters, etc.) is stored in the database, forming a cross-platform charging pile resource pool. This cross-platform charging pile resource pool contains charging pile information from different charging pile platforms for unified management and dispatch. A data acquisition module (such as a Linux-based daemon or Windows service program) runs on the central dispatch server, periodically (e.g., every second or every minute) acquiring real-time data from the charging pile platforms. The data acquisition module receives monitoring data from charging piles through polling or subscription and stores it as a monitoring stream. The monitoring stream can be a time-series data structure that records the status information of the charging piles at different points in time.

[0025] For example, multiple charging pile platforms (such as Platform A, Platform B, and Platform C) send charging pile data to a central dispatch server via the MQTT protocol. The data acquisition module on the central dispatch server retrieves real-time data from each platform every 10 seconds, including information such as the charging pile's battery level, usage status (idle, charging, faulty, etc.), and charging power. This data is stored in a charging pile resource pool table in a MySQL database, forming a cross-platform charging pile resource pool. Simultaneously, the data acquisition module records each acquired data stream as a monitoring stream; for example, charging pile A1 is "idle" with 80% battery at 10:00:00, and "charging" with 75% battery at 10:00:10.

[0026] In one possible implementation, the cross-platform charging pile resource pool is monitored in real time to obtain monitoring streams for each charging pile. Step S100 further includes step S110, which involves real-time monitoring of each charging pile within the cross-platform charging pile resource pool to obtain multiple monitoring datasets. Specifically, a data acquisition module is deployed on a central dispatch server. This module establishes a connection with each charging pile platform via network communication protocols (such as TCP / IP, MQTT, etc.). Each charging pile platform is responsible for collecting real-time data (including charging pile status, power, charging power, fault information, etc.) of the charging piles it manages. The real-time data obtained from each charging pile platform is stored as an independent monitoring dataset. Each monitoring dataset can be a time-series data structure, recording the status information of a single charging pile at different points in time. For example, database tables or file systems can be used to store these datasets. The collected data is formatted to ensure data consistency and readability. For example, the data is converted into a unified JSON or CSV format for easier subsequent processing.

[0027] For example, the central dispatch server establishes connections with charging pile platforms A, B, and C via the MQTT protocol. Platform A manages charging piles A1 and A2, platform B manages charging piles B1 and B2, and platform C manages charging piles C1 and C2. The data acquisition module retrieves real-time data from each platform every 10 seconds. This data is stored as independent monitoring datasets, such as: Monitoring Dataset 1: Real-time data of charging pile A1 (status, battery level, charging power, etc.); Monitoring Dataset 2: Real-time data of charging pile A2; Monitoring Dataset 3: Real-time data of charging pile B1; Monitoring Dataset 4: Real-time data of charging pile B2; Monitoring Dataset 5: Real-time data of charging pile C1; Monitoring Dataset 6: Real-time data of charging pile C2.

[0028] Step S120 involves real-time monitoring of the multiple monitoring datasets to generate monitoring streams for each charging pile. Specifically, multiple independent monitoring datasets are merged to generate a comprehensive monitoring stream. The monitoring stream can be a continuous time series data, recording the status changes of each charging pile at different points in time. For example, data stream processing frameworks (such as Apache Kafka, Apache Flink, etc.) can be used to achieve real-time data fusion. During the generation of the monitoring stream, the data is cleaned and preprocessed to remove noisy data (such as outliers, duplicate data, etc.), fill in missing data (such as through interpolation methods), and extract key features (such as fault characteristics, usage frequency, etc.). The monitoring stream is ensured to be updated in real time, reflecting the latest status of the charging piles. For example, the real-time performance of the monitoring stream is guaranteed by setting the data update time interval (such as per second or per minute).

[0029] For example, Apache Kafka can be used as the data stream processing framework to merge the six monitoring datasets mentioned above. The generated monitoring stream can be a continuous time series data, recording the status changes of each charging station. For example: Time point 1: Charging station A1 is in "idle" with 80% battery level; Time point 2: Charging station A1 is in "charging" with 75% battery level; Time point 3: Charging station B1 is in "fault" with 50% battery level; Time point 4: Charging station C2 is in "idle" with 90% battery level; and so on.

[0030] This implementation ensures data integrity and accuracy by storing real-time data from each charging station as an independent monitoring dataset. Independent storage of each dataset facilitates subsequent data processing and analysis. By generating a comprehensive monitoring stream, the system can gain a more complete understanding of the charging station's operational status, leading to more rational resource allocation and scheduling. For example, based on data from the monitoring stream, priority can be given to compensating charging stations in good condition and concentrated locations, thereby improving resource utilization.

[0031] Step S200: Based on the monitoring streams of each charging pile, perform fault prediction, establish a charging pile fault space, and divide the cross-platform charging pile resource pool into multiple levels according to the charging pile fault space to obtain the first cluster of charging piles, the second cluster of charging piles, and the third cluster of charging piles.

[0032] Specifically, machine learning algorithms (such as Support Vector Machines (SVM) and neural networks) are used to analyze the monitoring stream data of each charging pile. Labeled fault data (such as historical fault cases) is used to train the machine learning model, enabling it to identify charging pile fault patterns. After model training, real-time monitoring stream data is input into the model for fault prediction, forecasting the probability and type of future charging pile faults. Based on the fault prediction results, a charging pile fault space is established. The charging pile fault space is a data structure, such as a multidimensional array or database table, that records information related to charging pile faults, including the fault probability, fault type (such as hardware fault, software fault, communication fault, etc.), and fault time for each charging pile. Clustering algorithms (such as K-means and DBSCAN) are used to perform multi-level partitioning of the cross-platform charging pile resource pool. Based on factors such as the charging pile's fault probability, location, and usage frequency, the charging piles are divided into different clusters for targeted scheduling and management. For example, the first cluster of charging piles can be charging piles with a low probability of failure and concentrated locations; the second cluster of charging piles can be charging piles with a medium probability of failure and more dispersed locations; and the third cluster of charging piles can be charging piles with a high probability of failure or special problems.

[0033] For example, a neural network can be used to predict faults based on monitoring data from each charging station. This monitoring data includes parameters such as voltage, current, and temperature of the charging stations. This data is input into the neural network model, which, based on the trained parameters, outputs the fault probability for each charging station. For example, the fault probability of charging station A1 is 0.1, that of charging station B2 is 0.5, and that of charging station C3 is 0.8. Based on these fault probabilities, a fault space for the charging stations is established, and the K-means algorithm is used to divide the charging stations into three clusters. The first cluster includes charging stations with a fault probability below 0.2, such as charging station A1; the second cluster includes charging stations with a fault probability between 0.2 and 0.6, such as charging station B2; and the third cluster includes charging stations with a fault probability above 0.6, such as charging station C3.

[0034] In one possible implementation, fault prediction is performed based on the monitoring streams of each charging pile to establish a charging pile fault space. Step S200 further includes step S210, extracting the first charging pile monitoring stream corresponding to the first charging pile based on the monitoring streams of each charging pile, and retrieving the first charging pile fault record set corresponding to the first charging pile. Specifically, the monitoring stream of any charging pile (the first charging pile) is extracted from the comprehensive charging pile monitoring stream using a data filtering algorithm, for example, filtering the corresponding data from the monitoring stream based on the charging pile ID. The historical fault record set of the first charging pile is retrieved from the database, and these records include information such as fault time, fault type, and fault cause.

[0035] For example, assuming the monitoring stream contains data from multiple charging piles, the monitoring stream for charging pile A1 (the first charging pile monitoring stream) is extracted using a filtering algorithm. Simultaneously, the historical fault record set for charging pile A1 is retrieved from the database. For example: Fault record 1: Time 2024-05-10, Fault type "Hardware fault", Fault cause "Charging module damage"; Fault record 2: Time 2024-06-15, Fault type "Communication fault", Fault cause "Network connection interruption".

[0036] Step S220: Based on the first charging pile fault record set, train a first charging pile fault prediction model. This model includes a charging pile fault risk prediction model and a charging pile fault type prediction model. Specifically, machine learning algorithms (such as logistic regression, support vector machine, neural network, etc.) are used to train the historical fault record set of the first charging pile. The charging pile fault risk prediction model is used to predict the probability of a charging pile malfunctioning within a future period; the charging pile fault type prediction model is used to predict the possible fault types. Fault-related features, such as voltage fluctuations, current anomalies, and temperature changes, are extracted from the monitoring stream and used as input to the model.

[0037] For example, a neural network is used to train a set of historical fault records for charging pile A1, with features including monitoring data such as voltage, current, and temperature. After training, two models are obtained: a charging pile fault risk prediction model, which outputs the probability (e.g., 0.1) of charging pile A1 failing within the next 24 hours; and a charging pile fault type prediction model, which outputs the possible fault types (e.g., "hardware fault").

[0038] Step S230: Based on the specifications and model information of the first charging pile, retrieve fault records of charging piles of the same model to obtain a supplementary fault record set. Specifically, based on the specifications and model information of the first charging pile (such as brand, model, etc.), retrieve historical fault records of charging piles of the same model from the database. These records serve as supplementary data to enhance the training effect of the model.

[0039] For example, assuming that the model of charging pile A1 is "Model X", historical fault records of charging piles of the same model (such as charging piles A2 and A3) can be retrieved from the database. For example, the fault record of charging pile A2 is: time 2024-07-01, fault type is "hardware fault", and fault cause is "charging module damage"; the fault record of charging pile A3 is: time 2024-08-05, fault type is "communication fault", and fault cause is "network connection interruption".

[0040] Step S240: The first charging pile fault prediction model is reinforced and trained based on the supplementary fault record set of the charging pile to obtain the first charging pile fault prediction channel. Specifically, the supplementary fault record set of the same model of charging pile is merged with the historical fault record set of the first charging pile, and the first charging pile fault prediction model is retrained. By increasing the sample size, the accuracy and generalization ability of the model are improved. Methods such as cross-validation are used to optimize the model to ensure its performance in different scenarios.

[0041] For example, by merging the fault records of charging piles A2 and A3 with the historical fault records of charging pile A1, the fault prediction model for the first charging pile can be retrained. The optimized model can more accurately predict the fault risk and fault type of charging pile A1.

[0042] Step S250: Input the first charging pile monitoring stream into the first charging pile fault prediction channel to obtain the first charging pile fault prediction result, and add the first charging pile fault prediction result to the charging pile fault space. Specifically, input the real-time monitoring stream of the first charging pile into the optimized fault prediction model to obtain fault prediction results, including fault risk coefficient and predicted fault type. Store the fault prediction result in the charging pile fault space for subsequent scheduling and management.

[0043] This approach increases the diversity and quantity of training data by combining historical fault records of charging piles with fault records of charging piles of the same model, thereby improving the accuracy and generalization ability of the fault prediction model.

[0044] In one possible implementation, the cross-platform charging pile resource pool is divided into multiple levels according to the charging pile fault space to obtain a first cluster, a second cluster, and a third cluster of charging piles. Step S200 further includes step S260, determining whether the fault risk coefficient of each charging pile in the charging pile fault space is less than the charging pile fault risk threshold, and obtaining multiple fault risk judgment results. Specifically, a charging pile fault risk threshold (e.g., 0.5) is set to distinguish the fault risk level of charging piles. The charging pile fault space is traversed, and the fault risk coefficient of each charging pile is judged. If the fault risk coefficient is less than the threshold, it is marked as low risk; otherwise, it is marked as high risk.

[0045] Step S270: Based on the multiple fault risk assessment results, the cross-platform charging pile resource pool is divided, and a first cluster of charging piles with fault risk coefficients less than the charging pile fault risk threshold and a risk charging pile cluster with fault risk coefficients greater than or equal to the charging pile fault risk threshold are established. Specifically, based on the fault risk assessment results, the charging piles are divided into two clusters: the first cluster is a set of charging piles with fault risk coefficients less than the threshold, representing low-risk charging piles; the risk charging pile cluster is a set of charging piles with fault risk coefficients greater than or equal to the threshold, representing high-risk charging piles.

[0046] Step S280: Determine whether the predicted fault type of each charging pile corresponding to the risky charging pile cluster belongs to the immediately recoverable fault type, and obtain multiple fault type judgment results. Specifically, define the "immediately recoverable fault" type (fault types that can be quickly recovered through simple operations, such as communication faults, which can be repaired by restarting) and the "non-immediately recoverable fault" type (fault types that require a longer time to repair, such as hardware faults, which require repair or replacement of parts). Traverse the risky charging pile cluster and classify and judge according to the predicted fault type of each charging pile.

[0047] Step S290: Based on the multiple fault type judgment results, the risk charging pile clusters are divided to obtain a second cluster of charging piles corresponding to immediately recoverable faults and a third cluster of charging piles corresponding to non-immediately recoverable faults. Specifically, based on the fault type judgment results, the risk charging pile clusters are further divided into two clusters: the second cluster of charging piles is the set of charging piles whose predicted fault type is immediately recoverable fault; the third cluster of charging piles is the set of charging piles whose predicted fault type is non-immediately recoverable fault.

[0048] This implementation uses a multi-layered approach, subdividing the charging pile resource pool into low-risk charging piles (first cluster), charging piles with quickly recoverable faults (second cluster), and charging piles requiring longer repair times (third cluster). This refined classification makes resource management more targeted, enabling different scheduling strategies to be adopted based on the different risks and fault types of charging piles.

[0049] In one possible implementation, the method further includes: performing fault operation and maintenance on the cross-platform charging pile resource pool based on the charging pile fault space.

[0050] Specifically, utilizing the established charging pile fault space, charging piles are categorized into different fault types—high-risk, medium-risk, and low-risk—based on fault prediction results (such as fault risk coefficients and predicted fault types). Charging piles are then prioritized according to their fault risk coefficients and predicted fault types. High-risk charging piles are processed first, followed by medium-risk ones, and low-risk ones are processed last. Specific fault handling strategies are formulated based on fault type and priority. For example, for hardware faults, maintenance personnel are dispatched for on-site inspection and repair; for communication faults, remote restarts or network configuration adjustments are used; and for software faults, remote software updates or device restarts are employed. Fault handling strategies are executed, and the process is monitored in real time. A central scheduling server records the progress and results of fault handling to ensure timely resolution. The results of fault handling are recorded in a database and fed back into the charging pile fault space to update the fault prediction model. This feedback mechanism continuously optimizes fault prediction and handling strategies.

[0051] This approach, through fault priority sorting and classification, can quickly identify and handle faults in high-risk charging piles, reduce the impact of faults on users, and improve overall fault handling efficiency.

[0052] Step S300: Perform cross-platform scheduling compensation on the third cluster of charging piles based on the first cluster and the second cluster of charging piles to generate the first scheduling compensation group.

[0053] Specifically, cross-platform scheduling compensation refers to the process of relocating charging piles from one cluster to another based on the characteristics of different clusters within the charging pile resource pool, in order to compensate for insufficient or faulty charging pile resources in the target cluster. The first group of scheduling compensation refers to multiple scheduling compensation schemes formed after the scheduling compensation process. Each scheduling compensation scheme includes a set of scheduled charging piles used to compensate for the charging piles in the target cluster.

[0054] Specifically, a scheduling compensation strategy is formulated based on the characteristics of the first and second clusters of charging piles (such as low failure probability, concentrated location, and high usage frequency). For example, charging piles in the first cluster are prioritized for compensation because of their low failure probability. If the number of charging piles in the first cluster is insufficient to meet the compensation needs, suitable charging piles are selected from the second cluster to supplement them. The scheduling module of the central scheduling server allocates charging piles from the first and second clusters to the third cluster for compensation according to the formulated scheduling compensation strategy. The scheduling module can use optimization algorithms (such as linear programming and genetic algorithms) to initially determine multiple scheduling compensation schemes, which constitute the first scheduling compensation group. Each scheduling compensation scheme in the first scheduling compensation group includes a set of charging piles that, after scheduling, can effectively compensate the charging piles in the third cluster.

[0055] For example, if 5 charging piles in the third cluster need compensation, 3 charging piles in the first cluster are available, and 4 charging piles in the second cluster are available, the scheduling compensation strategy prioritizes compensating the 3 charging piles in the first cluster. If the compensation capacity of these 3 charging piles is insufficient, then 2 charging piles are selected from the second cluster to supplement them. Ultimately, these 5 charging piles (3 from the first cluster and 2 from the second cluster) form one of the scheduling compensation schemes in the first scheduling compensation group, used to compensate the charging piles in the third cluster.

[0056] Step S400: Introduce a scheduling response evaluation model to perform response evaluation optimization on the first group of scheduling compensation, construct a second group of scheduling compensation, and perform joint optimization of the second group of scheduling compensation based on the scheduling compensation mutation capacity to generate a scheduling compensation strategy.

[0057] Specifically, a scheduling response evaluation model is constructed to assess the effectiveness of scheduling compensation. This model quantifies and evaluates multiple scheduling compensation schemes in the first group of scheduling compensation schemes based on multiple indicators (such as compensation time, compensation cost, and resource utilization). For example, a weighted scoring method is used to score each indicator, and then a comprehensive score is calculated to evaluate the performance of each scheduling compensation scheme in the first group of scheduling compensation schemes.

[0058] Based on the scoring results of the scheduling response evaluation model, the first group of scheduling compensation schemes is optimized and adjusted to improve scheduling performance. For example, if the compensation time of a certain scheduling compensation scheme is too long, its scheduling scheme can be adjusted to select a closer charging pile for compensation; if the compensation cost is too high, charging pile resources can be reallocated to reduce the compensation cost. The optimized first group of scheduling compensation schemes forms the second group of scheduling compensation schemes. The second group of scheduling compensation schemes, guided by the scheduling response evaluation model, has been optimized and adjusted to have better scheduling performance.

[0059] Mutation operations (such as randomly adjusting the scheduling order of charging piles or changing the allocation ratio of charging piles) are introduced to perform joint optimization on the second group of scheduling compensation. Mutation operations are used to increase the diversity of scheduling schemes and avoid getting trapped in local optima. Through multiple mutations and evaluations, the optimal scheduling compensation strategy is finally generated.

[0060] In one possible implementation, a scheduling response evaluation model is introduced to optimize the response of the first group of scheduling compensation schemes, thereby constructing a second group of scheduling compensation schemes. Step S400 further includes step S410, whereby each scheduling compensation scheme within the first group of scheduling compensation schemes is evaluated according to the scheduling response evaluation model to obtain multiple scheduling response evaluation results. Specifically, a scheduling response evaluation model is constructed, which quantitatively evaluates the scheduling compensation schemes based on multiple indicators. Weights are assigned to each evaluation indicator, and each scheduling compensation scheme in the first group of scheduling compensation schemes is input into the scheduling response evaluation model to calculate a comprehensive score for each scheme.

[0061] Step S420: Construct scheduling response evaluation constraints, which include resource conflict risk constraints, scheduling execution risk constraints, and grid response risk constraints. Specifically, scheduling response evaluation constraints are used to verify whether the scheduling compensation scheme meets specific conditions, including resource conflict risk constraints, scheduling execution risk constraints, and grid response risk constraints. The resource conflict risk constraints ensure that the scheduling compensation scheme does not lead to charging pile resource conflicts; the scheduling execution risk constraints ensure that no technical problems (such as communication interruptions or equipment failures) occur during the execution of the scheduling compensation scheme; and the grid response risk constraints ensure that the scheduling compensation scheme does not cause excessive load on the grid. For example, checking whether the scheduling scheme involves multiple users simultaneously reserving the same charging pile, checking whether the scheduling scheme considers the current state of the charging pile (such as whether it is in a fault state), and checking whether the scheduling scheme avoids large-scale charging during peak grid hours (such as 17:00-19:00).

[0062] Step S430: Perform anomaly verification on the multiple scheduling response evaluation results according to the scheduling response evaluation constraints to obtain multiple response evaluation verification results. Specifically, verify the evaluation result of each scheduling compensation scheme to check whether it meets the above three constraints. If a scheme does not meet any constraint, it is marked as abnormal and an anomaly label is added. The verification results are divided into two categories: normal and abnormal.

[0063] Step S440: Based on the multiple response evaluation and verification results, optimize and filter the first group of scheduling compensation to generate the second group of scheduling compensation. Specifically, select scheduling compensation schemes marked as normal from the first group of scheduling compensation, and form the second group of scheduling compensation with the selected normal schemes.

[0064] This approach introduces a scheduling response evaluation model and constraints to comprehensively evaluate and verify scheduling compensation schemes, ensuring that the selected schemes meet the requirements in terms of resource allocation, execution risk, and grid response, thereby improving the reliability and feasibility of scheduling schemes.

[0065] In one possible implementation, the scheduling compensation schemes within the first group of scheduling compensations are evaluated according to the scheduling response evaluation model to obtain multiple scheduling response evaluation results. Step S410 further includes step S411, extracting a first scheduling compensation scheme from the first group of scheduling compensations. Specifically, a specific scheduling compensation scheme is selected from the first group of scheduling compensations as the first scheduling compensation scheme. The selection method can be random selection or based on a certain priority rule (such as prioritizing the scheme with the highest resource utilization).

[0066] For example, suppose there are three scheduling compensation schemes in the first group of charging piles: Scheme 1: Prioritize the use of charging piles in the first group for compensation; Scheme 2: Prioritize the use of charging piles in the second group for compensation; Scheme 3: Use a mix of charging piles from the first and second groups for compensation. Scheme 1 is selected as the first scheduling compensation scheme.

[0067] Step S412: Based on the first cluster and the second cluster of charging piles, the third cluster of charging piles is subjected to multiple simulated scheduling operations according to the first scheduling compensation scheme to obtain multiple scheduling simulation datasets. Specifically, the first scheduling compensation scheme is simulated multiple times using a simulation environment (such as simulation software or a virtual platform) to evaluate the performance of the scheme under different scenarios. The results of each simulated scheduling are recorded, including scheduling time, resource utilization, grid load, and other information, forming multiple scheduling simulation datasets.

[0068] Step S413: Confidence fusion is performed on the multiple scheduling simulation datasets to obtain a first scheduling simulation confidence set. Specifically, statistical analysis is conducted on the multiple scheduling simulation datasets to extract confidence intervals or confidence values ​​for key indicators (such as scheduling time, resource utilization, and grid load). Confidence fusion can be achieved using statistical methods (such as mean, standard deviation, and confidence interval calculation). The fused result generates the first scheduling simulation confidence set for subsequent evaluation.

[0069] For example, confidence fusion is performed on three scheduling simulation datasets: the mean scheduling time is 11 minutes, the standard deviation is 1 minute, and the confidence interval is [10, 12]; the mean resource utilization rate is 77.67%, the standard deviation is 2.51%, and the confidence interval is [75.16%, 80.18%]; and the mean grid load is 21%, the standard deviation is 1%, and the confidence interval is [20%, 22%]. The resulting first scheduling simulation confidence set is: scheduling time: [10, 12]; resource utilization rate: [75.16%, 80.18%]; grid load: [20%, 22%.

[0070] Step S414: The dispatch response evaluation model includes a resource conflict risk evaluation model, a dispatch execution risk evaluation model, and a power grid response risk evaluation model. The first dispatch simulation confidence set is input into the dispatch response evaluation model to obtain the first dispatch response evaluation result. Specifically, the first dispatch simulation confidence set is input into the dispatch response evaluation model, which includes three sub-models. The resource conflict risk evaluation model is used to assess whether the dispatch scheme will lead to resource conflicts (such as repeated allocation of charging piles); the dispatch execution risk evaluation model is used to assess the technical problems that the dispatch scheme may encounter during execution (such as communication interruption, equipment failure); and the power grid response risk evaluation model is used to assess the impact of the dispatch scheme on the power grid load to ensure that it will not put excessive pressure on the power grid. Each sub-model outputs a risk coefficient, representing the risk level of the dispatch scheme in the corresponding aspect.

[0071] For example, if the first scheduling simulation confidence set is input into the scheduling response evaluation model, the resource conflict risk evaluation model outputs a resource conflict risk coefficient of 0.1 (low risk); the scheduling execution risk evaluation model outputs a scheduling execution risk coefficient of 0.2 (medium risk); and the power grid response risk evaluation model outputs a power grid response risk coefficient of 0.3 (medium risk). Therefore, the evaluation results for the first scheduling response are: resource conflict risk coefficient 0.1, scheduling execution risk coefficient 0.2, and power grid response risk coefficient 0.3.

[0072] This implementation method, through multiple simulated scheduling and confidence fusion, enables the system to comprehensively evaluate the performance of scheduling schemes under different scenarios, ensuring that the selected schemes have low risks in terms of resource conflicts, execution risks, and grid response, thus avoiding scheduling failures caused by resource conflicts, execution risks, or excessive grid load, thereby optimizing scheduling decisions.

[0073] In one possible implementation, the second group of scheduling compensation is subjected to joint optimization based on the scheduling compensation variation capacity to generate a scheduling compensation strategy. Step S400 further includes step S450, which involves weight allocation based on the scheduling response evaluation element set to establish a comprehensive scheduling risk analysis model. The scheduling response evaluation element set includes resource conflict risk, scheduling execution risk, and grid response risk. Specifically, weights are allocated based on the scheduling response evaluation element set (resource conflict risk, scheduling execution risk, and grid response risk). The weights can be adjusted according to actual needs and experience. For example, resource conflict risk weight: 0.4, scheduling execution risk weight: 0.3, grid response risk weight: 0.3. A comprehensive scheduling risk analysis model is established, which integrates the above three risk indicators to calculate the comprehensive risk value of each scheduling compensation scheme. The formula for calculating the comprehensive risk value can be expressed as: Comprehensive risk value = w1 × resource conflict risk coefficient + w2 × scheduling execution risk coefficient + w3 × grid response risk coefficient, where w1, w2, and w3 are the weights of resource conflict risk, scheduling execution risk, and grid response risk, respectively.

[0074] Step S460: Based on the scheduling compensation mutation capacity, perform mutation optimization on the second group of scheduling compensation according to the scheduling comprehensive risk analysis model to establish a first group for compensation mutation optimization. Specifically, the scheduling compensation mutation capacity is the range or degree to which mutation operations are allowed on the scheduling compensation scheme. Based on the scheduling compensation mutation capacity, perform mutation operations on each scheme in the second group of scheduling compensation, such as randomly adjusting the allocation order of charging piles or changing the allocation ratio of charging piles, to increase the diversity of schemes and avoid getting trapped in local optima. Evaluate the mutated schemes and select schemes that meet the preset risk conditions to enter the first group for compensation mutation optimization.

[0075] Step S470: Based on the scheduling compensation mutation capacity, continue mutation optimization on the first compensation mutation optimization group according to the scheduling comprehensive risk analysis model, and establish the Qth compensation mutation optimization group, where Q is a positive integer greater than 1. Specifically, continue mutation operations on the schemes in the first compensation mutation optimization group and evaluate them according to the scheduling comprehensive risk analysis model. Repeat this process Q times, generating a new compensation mutation optimization group each time, until the predetermined number of iterations is reached.

[0076] Step S480: Based on the comprehensive scheduling risk threshold, the second group of scheduling compensation, the first group of compensation variation optimization, ..., the Qth group of compensation variation optimization are jointly optimized to generate the third group of scheduling compensation. Specifically, all schemes in the second group of scheduling compensation, the first group of compensation variation optimization, ..., the Qth group of compensation variation optimization are comprehensively evaluated, and schemes with a comprehensive risk value less than or equal to the comprehensive scheduling risk threshold (the upper limit of the comprehensive risk value used to filter scheduling compensation schemes) are selected to form the third group of scheduling compensation. The comprehensive scheduling risk threshold can be set according to actual needs, for example, 0.2.

[0077] Step S490 involves optimizing the scheduling compensation efficiency of the third group of scheduling compensation schemes to obtain the scheduling compensation strategy. Specifically, the schemes in the third group of scheduling compensation schemes are further screened with the goal of maximizing scheduling compensation efficiency. Scheduling compensation efficiency can be measured by indicators such as compensation time and resource utilization. For example, the scheme with the shortest compensation time and the highest resource utilization is selected as the final scheduling compensation strategy.

[0078] This approach, through multiple rounds of variational and joint optimization, comprehensively assesses the overall risk of dispatch compensation schemes, ensuring that the selected scheme has low risk in terms of resource conflicts, execution risks, and grid response. By maximizing dispatch compensation efficiency, it selects the scheme with the shortest compensation time and the highest resource utilization, thereby improving overall dispatch efficiency.

[0079] In one possible implementation, based on the scheduling compensation variation capacity, variation optimization is performed on the second group of scheduling compensation according to the scheduling comprehensive risk analysis model to establish a first group of compensation variation optimization. Step S460 further includes step S461, performing comprehensive risk calculation on the second group of scheduling compensation according to the scheduling comprehensive risk analysis model to obtain a scheduling comprehensive risk set. Specifically, the scheduling comprehensive risk analysis model is used to calculate the comprehensive risk of each scheme in the second group of scheduling compensation. The comprehensive risk value is obtained by weighted summation of resource conflict risk, scheduling execution risk, and grid response risk. The comprehensive risk value of each scheme is stored in the scheduling comprehensive risk set for variation capacity allocation.

[0080] Step S462: Based on the scheduling compensation mutation capacity, allocate mutation capacity to the second group of scheduling compensation according to the scheduling comprehensive risk set to obtain a mutation capacity allocation set. Specifically, allocate mutation capacity to each scheme according to the comprehensive risk value in the scheduling comprehensive risk set. The mutation capacity may include the number of mutation schemes and can be dynamically adjusted according to the level of the comprehensive risk value. For example, allocate more mutation capacity to schemes with higher comprehensive risk values ​​to reduce risk through mutation. Store the mutation capacity of each scheme in the mutation capacity allocation set.

[0081] Step S463: Adjust the scheduling compensation second group according to the mutation capacity allocation set to generate a compensation mutation first group. Specifically, perform mutation operations on each scheme according to the mutation capacity allocation set. Mutation operations may include adjusting the allocation order of charging piles, changing the allocation ratio of charging piles, etc. The magnitude and number of mutations are determined according to the allocated mutation capacity. The mutated schemes are then combined into the compensation mutation first group.

[0082] Step S464: Optimize the response of the first compensation variation group according to the scheduling response evaluation model to generate the first compensation variation optimization group. Specifically, the scheduling response evaluation model is used to evaluate each scheme in the first compensation variation group, calculating the resource conflict risk coefficient, scheduling execution risk coefficient, and grid response risk coefficient for each scheme. The resource conflict risk coefficient, scheduling execution risk coefficient, and grid response risk coefficient for each scheme are verified to ensure that each risk coefficient meets preset constraints. For example: resource conflict risk coefficient ≤ 0.1, scheduling execution risk coefficient ≤ 0.2, and grid response risk coefficient ≤ 0.3. If any risk coefficient of a scheme does not meet the constraints, the scheme is marked as abnormal and an abnormal label is added; otherwise, it is marked as normal. The schemes marked as normal form the first compensation variation optimization group.

[0083] This implementation, through mutation capacity allocation, allows for fine-grained mutation adjustments based on the risk level of each scheme, rather than a one-size-fits-all approach. This makes mutation operations more targeted and more effectively reduces scheme risk. Individual risk verification allows for the selection of schemes that fully meet the constraints from multiple scheduling compensation options, preventing scheduling failures caused by excessively high risk coefficients and thus optimizing scheduling decisions.

[0084] Step S500: Based on the central scheduling server, perform compensation management on the cross-platform charging pile resource pool according to the scheduling compensation strategy.

[0085] Specifically, based on the generated scheduling compensation strategy, the scheduling module of the central scheduling server generates specific scheduling instructions to control the compensation operation of the charging piles. These instructions include parameters such as charging pile allocation information (e.g., assigning charging pile A to charging pile B for compensation), compensation time, and compensation power. The scheduling instructions are sent to the corresponding charging pile platform via network communication protocols (e.g., TCP / IP, MQTT). Upon receiving the instructions, the charging pile platform controls the charging piles to perform compensation operations according to the instructions. For example, it might control charging pile A to start charging charging pile B, or adjust the charging power of the charging pile. The central scheduling server monitors the compensation operation process of the charging piles in real time and records the compensation results (e.g., compensation completion time, compensation power). If any abnormalities occur during the compensation process (e.g., charging pile malfunction, communication interruption), the central scheduling server promptly adjusts the scheduling strategy and reallocates charging pile resources to ensure the smooth completion of the compensation operation.

[0086] For example, according to the scheduling compensation strategy, the central scheduling server generates a scheduling instruction to allocate charging pile A1 to charging pile B1 for compensation, with a compensation power of 30kW and a compensation time of 1 hour. The scheduling instruction is sent to charging pile A platform and charging pile B platform via the MQTT protocol. After receiving the instruction, charging pile A platform controls charging pile A1 to start charging charging pile B1. The central scheduling server monitors the charging process in real time, recording the charging amount and charging time. If charging pile A1 malfunctions during charging, the central scheduling server will regenerate the scheduling instruction to allocate charging pile A2 to charging pile B1 for compensation, ensuring the successful completion of the compensation operation.

[0087] This application's embodiments utilize data from multiple charging pile platforms to construct a unified resource pool. Real-time monitoring acquires the operational status data stream of each charging pile, and fault prediction analysis is performed based on the monitoring data to establish a charging pile fault space. Accordingly, the charging piles are divided into three clusters: high reliability, medium reliability, and those requiring compensation. Resources from the high and medium reliability clusters are used to implement cross-platform scheduling compensation for the clusters requiring compensation, forming a preliminary scheduling scheme. This scheme is then optimized and screened using a response evaluation model, followed by secondary optimization using variability capacity analysis, ultimately generating the optimal scheduling strategy. This strategy is executed by the central scheduling server, completing the dynamic compensation management of charging pile resources across the entire platform. These technical means solve the technical problems of low management efficiency and reliability in existing cross-platform remote scheduling of charging piles, achieving the technical effect of improving the efficiency and reliability of scheduling management.

[0088] In the above text, refer to Figure 1 A cross-platform remote management and scheduling method for charging piles according to an embodiment of the present invention is described in detail. Next, reference will be made to... Figure 2 This invention describes a cross-platform remote management and scheduling system for charging piles according to an embodiment of the present invention.

[0089] A cross-platform remote management and scheduling system for charging piles according to an embodiment of the present invention addresses the technical problems of low management efficiency and reliability in existing cross-platform remote scheduling of charging piles, thereby improving the efficiency and reliability of scheduling management. The cross-platform remote management and scheduling system for charging piles includes: a cross-platform charging pile resource pool construction module 10, a multi-level resource pool partitioning module 20, a cross-platform scheduling compensation module 30, a variation joint optimization module 40, and a compensation management module 50.

[0090] The cross-platform charging pile resource pool construction module 10 is used to connect multiple charging pile platforms to the central dispatch server, construct a cross-platform charging pile resource pool, and monitor the cross-platform charging pile resource pool in real time to obtain the monitoring flow of each charging pile; the resource pool multi-level partitioning module 20 is used to perform fault prediction based on the monitoring flow of each charging pile, establish a charging pile fault space, and partition the cross-platform charging pile resource pool into a first cluster, a second cluster, and a third cluster of charging piles based on the charging pile fault space; the cross-platform scheduling compensation module 30 is used to perform cross-platform scheduling compensation on the third cluster of charging piles based on the first cluster and the second cluster, generating a first group of scheduling compensation; the mutation joint optimization module 40 is used to introduce a scheduling response evaluation model to perform response evaluation optimization on the first group of scheduling compensation, construct a second group of scheduling compensation, and perform mutation joint optimization on the second group of scheduling compensation based on the scheduling compensation mutation capacity to generate a scheduling compensation strategy; the compensation management module 50 is used to perform compensation management on the cross-platform charging pile resource pool based on the central dispatch server and the scheduling compensation strategy.

[0091] The specific configuration of the resource pool multi-level partitioning module 20 will be described in detail below. As mentioned above, based on the monitoring streams of each charging pile, fault prediction is performed to establish a charging pile fault space. The resource pool multi-level partitioning module 20 may further include: a data extraction unit for extracting the first charging pile monitoring stream corresponding to the first charging pile based on the monitoring streams of each charging pile, and retrieving the first charging pile fault record set corresponding to the first charging pile; a first charging pile fault prediction model training unit for training a first charging pile fault prediction model based on the first charging pile fault record set, wherein the first charging pile fault prediction model includes a charging pile fault risk prediction model and a charging pile fault type prediction model; a same-model charging pile fault record retrieval unit for retrieving the same-model charging pile fault records based on the specification and model information of the first charging pile, and obtaining a supplementary charging pile fault record set; a reinforcement training unit for reinforcing the first charging pile fault prediction model based on the supplementary charging pile fault record set, and obtaining a first charging pile fault prediction channel; and a charging pile fault prediction unit for inputting the first charging pile monitoring stream into the first charging pile fault prediction channel, obtaining the first charging pile fault prediction result, and adding the first charging pile fault prediction result to the charging pile fault space.

[0092] The cross-platform charging pile resource pool is divided into three clusters based on the charging pile fault space to obtain a first cluster, a second cluster, and a third cluster of charging piles. The multi-level resource pool division module 20 may further include: a fault risk judgment unit for judging whether the fault risk coefficient of each charging pile in the charging pile fault space is less than the charging pile fault risk threshold, obtaining multiple fault risk judgment results; a charging pile cluster division unit for dividing the cross-platform charging pile resource pool based on the multiple fault risk judgment results, establishing a first cluster of charging piles with a fault risk coefficient less than the charging pile fault risk threshold, and a risky charging pile cluster with a fault risk coefficient greater than or equal to the charging pile fault risk threshold; an immediate recovery fault judgment unit for judging whether the predicted fault type of each charging pile corresponding to the risky charging pile cluster belongs to an immediately recoverable fault, obtaining multiple fault type judgment results; and a risky charging pile cluster division unit for dividing the risky charging pile cluster based on the multiple fault type judgment results, obtaining a second cluster of charging piles corresponding to immediately recoverable faults, and a third cluster of charging piles corresponding to non-immediately recoverable faults.

[0093] The specific configuration of the variation joint optimization module 40 will be described in detail below. As mentioned above, a scheduling response evaluation model is introduced to evaluate and optimize the first group of scheduling compensation, and a second group of scheduling compensation is constructed. The variation joint optimization module 40 may further include: a scheme evaluation unit for evaluating each scheduling compensation scheme in the first group of scheduling compensation according to the scheduling response evaluation model, and obtaining multiple scheduling response evaluation results; an evaluation constraint construction unit for constructing scheduling response evaluation constraints, which include resource conflict risk constraints, scheduling execution risk constraints, and power grid response risk constraints; an anomaly verification unit for performing anomaly verification on the multiple scheduling response evaluation results according to the scheduling response evaluation constraints, and obtaining multiple response evaluation verification results; and an optimization screening unit for optimizing and screening the first group of scheduling compensation according to the multiple response evaluation verification results, and generating the second group of scheduling compensation.

[0094] Specifically, the scheduling compensation schemes within the first group of scheduling compensation are evaluated according to the scheduling response evaluation model to obtain multiple scheduling response evaluation results. The scheme evaluation unit may further include: a first scheduling compensation scheme extraction subunit for extracting a first scheduling compensation scheme based on the first group of scheduling compensation; a simulation scheduling subunit for performing multiple simulated scheduling operations on the third group of charging piles based on the first and second clusters of charging piles and according to the first scheduling compensation scheme to obtain multiple scheduling simulation datasets; a confidence fusion subunit for performing confidence fusion on the multiple scheduling simulation datasets to obtain a first scheduling simulation confidence set; and a scheduling response evaluation subunit for inputting the first scheduling simulation confidence set into the scheduling response evaluation model, which includes a resource conflict risk evaluation model, a scheduling execution risk evaluation model, and a power grid response risk evaluation model, to obtain a first scheduling response evaluation result.

[0095] The process involves performing joint optimization of the second group of scheduling compensation based on the scheduling compensation variation capacity to generate a scheduling compensation strategy. The joint optimization module 40 may further include: a scheduling comprehensive risk analysis model establishment unit, used to allocate weights based on a scheduling response evaluation element set to establish a scheduling comprehensive risk analysis model, where the scheduling response evaluation element set includes resource conflict risk, scheduling execution risk, and grid response risk; a variation optimization unit, used to perform variation optimization on the second group of scheduling compensation based on the scheduling compensation variation capacity and the scheduling comprehensive risk analysis model to establish a compensation variation optimization first group; an iterative variation optimization unit, used to continue performing variation optimization on the first group of compensation variation optimization based on the scheduling compensation variation capacity and the scheduling comprehensive risk analysis model to establish a compensation variation optimization Q-th group, where Q is a positive integer greater than 1; a joint optimization unit, used to perform joint optimization on the second group of scheduling compensation, the first group of compensation variation optimization, ..., the Q-th group of compensation variation optimization based on a scheduling comprehensive risk threshold to generate a scheduling compensation third group; and a scheduling compensation efficiency maximization optimization unit, used to perform scheduling compensation efficiency maximization optimization on the third group of scheduling compensation to obtain the scheduling compensation strategy.

[0096] Specifically, based on the scheduling compensation variation capacity, the second group of scheduling compensation is optimized for variation according to the scheduling comprehensive risk analysis model to establish a first group of compensation variation optimization. The variation optimization unit may further include: a comprehensive risk calculation subunit for performing comprehensive risk calculation on the second group of scheduling compensation according to the scheduling comprehensive risk analysis model to obtain a scheduling comprehensive risk set; a variation capacity allocation subunit for allocating variation capacity to the second group of scheduling compensation based on the scheduling compensation variation capacity and the scheduling comprehensive risk set to obtain a variation capacity allocation set; a variation adjustment subunit for adjusting variation of the second group of scheduling compensation according to the variation capacity allocation set to generate a first group of compensation variation; and a response evaluation optimization subunit for performing response evaluation optimization on the first group of compensation variation according to the scheduling response evaluation model to generate the first group of compensation variation optimization.

[0097] The specific configuration of the cross-platform charging pile resource pool construction module 10 will be described in detail below. As mentioned above, to monitor the cross-platform charging pile resource pool in real time and obtain monitoring streams for each charging pile, the cross-platform charging pile resource pool construction module 10 may further include: a first real-time monitoring unit for real-time monitoring of each charging pile in the cross-platform charging pile resource pool to obtain multiple monitoring datasets; and a second real-time monitoring unit for real-time monitoring of the multiple monitoring datasets to generate the monitoring streams for each charging pile.

[0098] The system may further include: a fault operation and maintenance module for performing fault operation and maintenance on the cross-platform charging pile resource pool based on the charging pile fault space.

[0099] The charging pile cross-platform remote management and scheduling system provided in this embodiment of the invention can execute the charging pile cross-platform remote management and scheduling method provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0100] 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.

[0101] 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 cross-platform remote management and scheduling method for charging piles, characterized in that, The method includes: Multiple charging pile platforms are connected to a central dispatch server to build a cross-platform charging pile resource pool, and the cross-platform charging pile resource pool is monitored in real time to obtain the monitoring stream of each charging pile. Based on the monitoring streams of each charging pile, fault prediction is performed, a charging pile fault space is established, and the cross-platform charging pile resource pool is divided into multiple levels according to the charging pile fault space to obtain the first cluster, the second cluster, and the third cluster of charging piles. Based on the first cluster and the second cluster of charging piles, cross-platform scheduling compensation is performed on the third cluster of charging piles to generate the first group of scheduling compensation. A scheduling response evaluation model is introduced to evaluate and optimize the first group of scheduling compensations, a second group of scheduling compensations is constructed, and the second group of scheduling compensations is jointly optimized based on the variation capacity of scheduling compensations to generate a scheduling compensation strategy. Based on the central scheduling server, the cross-platform charging pile resource pool is compensated and managed according to the scheduling compensation strategy.

2. The cross-platform remote management and scheduling method for charging piles as described in claim 1, characterized in that, Based on the monitoring streams of each charging pile, fault prediction is performed, and a charging pile fault space is established, including: Extract the first charging pile monitoring stream corresponding to the first charging pile based on the monitoring stream of each charging pile, and retrieve the first charging pile fault record set corresponding to the first charging pile. Based on the first charging pile fault record set, a first charging pile fault prediction model is trained. The first charging pile fault prediction model includes a charging pile fault risk prediction model and a charging pile fault type prediction model. Based on the specifications and model information of the first charging pile, a fault record retrieval of the same model of charging pile is performed to obtain a supplementary set of charging pile fault records. The first charging pile fault prediction model is reinforced and trained based on the charging pile fault supplementary record set to obtain the first charging pile fault prediction channel. The first charging pile monitoring stream is input into the first charging pile fault prediction channel to obtain the first charging pile fault prediction result, and the first charging pile fault prediction result is added to the charging pile fault space.

3. The cross-platform remote management and scheduling method for charging piles as described in claim 1, characterized in that, The cross-platform charging pile resource pool is divided into three levels based on the charging pile fault space, resulting in a first cluster, a second cluster, and a third cluster of charging piles, including: Determine whether the fault risk coefficient of each charging pile in the charging pile fault space is less than the charging pile fault risk threshold, and obtain multiple fault risk judgment results. Based on the multiple fault risk assessment results, the cross-platform charging pile resource pool is divided, and a first cluster of charging piles with a fault risk coefficient less than the fault risk threshold of the charging pile is established, as well as a risky charging pile cluster with a fault risk coefficient greater than or equal to the fault risk threshold of the charging pile. Determine whether the predicted fault type of each charging pile corresponding to the risky charging pile cluster belongs to the immediate recovery fault, and obtain multiple fault type judgment results; Based on the results of the multiple fault type judgments, the risk charging pile clusters are divided to obtain a second cluster of charging piles corresponding to immediately recoverable faults and a third cluster of charging piles corresponding to non-immediately recoverable faults.

4. The cross-platform remote management and scheduling method for charging piles as described in claim 1, characterized in that, A scheduling response evaluation model is introduced to evaluate and optimize the first group of scheduling compensations, thereby constructing a second group of scheduling compensations, including: The scheduling compensation schemes within the first group of scheduling compensation are evaluated based on the scheduling response evaluation model to obtain multiple scheduling response evaluation results; Construct scheduling response evaluation constraints, which include resource conflict risk constraints, scheduling execution risk constraints, and power grid response risk constraints. Based on the scheduling response evaluation constraints, anomaly verification is performed on the multiple scheduling response evaluation results to obtain multiple response evaluation verification results. Based on the multiple response evaluation and verification results, the first group of scheduling compensation is optimized and selected to generate the second group of scheduling compensation.

5. The cross-platform remote management and scheduling method for charging piles as described in claim 4, characterized in that, The scheduling compensation schemes within the first group of scheduling compensation are evaluated based on the aforementioned scheduling response evaluation model, resulting in multiple scheduling response evaluation results, including: The first scheduling compensation scheme is extracted based on the first group of scheduling compensation; Based on the first cluster of charging piles and the second cluster of charging piles, the third cluster of charging piles is simulated and scheduled multiple times according to the first scheduling compensation scheme to obtain multiple scheduling simulation datasets. The first scheduling simulation confidence set is obtained by performing confidence fusion on the multiple scheduling simulation datasets. The scheduling response evaluation model includes a resource conflict risk evaluation model, a scheduling execution risk evaluation model, and a power grid response risk evaluation model. The first scheduling simulation confidence set is input into the scheduling response evaluation model to obtain the first scheduling response evaluation result.

6. The cross-platform remote management and scheduling method for charging piles as described in claim 1, characterized in that, Based on the scheduling compensation mutation capacity, the second group of scheduling compensation is subjected to joint mutation optimization to generate a scheduling compensation strategy, including: A comprehensive scheduling risk analysis model is established by weighting the scheduling response evaluation element set, which includes resource conflict risk, scheduling execution risk, and power grid response risk. Based on the scheduling compensation variation capacity, the scheduling compensation second group is optimized for variation according to the scheduling comprehensive risk analysis model to establish the compensation variation optimization first group. Based on the scheduling compensation mutation capacity, according to the scheduling comprehensive risk analysis model, the first group of compensation mutation optimization is further optimized to establish the Qth group of compensation mutation optimization, where Q is a positive integer greater than 1. Based on the comprehensive scheduling risk threshold, the second group of scheduling compensation, the first group of compensation variation optimization, ... the Qth group of compensation variation optimization are jointly optimized to generate the third group of scheduling compensation. The scheduling compensation strategy is obtained by optimizing the scheduling compensation efficiency of the third group of scheduling compensation.

7. The cross-platform remote management and scheduling method for charging piles as described in claim 6, characterized in that, Based on the aforementioned scheduling compensation mutation capacity, and according to the aforementioned scheduling comprehensive risk analysis model, mutation optimization is performed on the second group of scheduling compensation to establish a first group of compensation mutation optimization, including: Based on the comprehensive risk analysis model of scheduling, the comprehensive risk of the second group of scheduling compensation is calculated to obtain the comprehensive risk set of scheduling. Based on the scheduling compensation mutation capacity, the second group of scheduling compensation is allocated mutation capacity according to the scheduling comprehensive risk set to obtain a mutation capacity allocation set. Based on the mutation capacity allocation set, the scheduling compensation second group is mutated and adjusted to generate the compensation mutation first group; The first group of compensation mutations is optimized by evaluating the response of the scheduling response evaluation model to generate the first group of compensation mutation optimization.

8. The cross-platform remote management and scheduling method for charging piles as described in claim 1, characterized in that, Real-time monitoring of the cross-platform charging pile resource pool to obtain monitoring streams for each charging pile, including: Each charging pile in the cross-platform charging pile resource pool is monitored in real time to obtain multiple monitoring datasets; The multiple monitoring datasets are monitored in real time to generate monitoring streams for each charging pile.

9. The cross-platform remote management and scheduling method for charging piles as described in claim 1, characterized in that, The cross-platform charging pile resource pool is operated and maintained based on the charging pile fault space.

10. A cross-platform remote management and scheduling system for charging piles, characterized in that, The system is used to implement the cross-platform remote management and scheduling method for charging piles as described in any one of claims 1-9, and the system includes: The cross-platform charging pile resource pool construction module is used to connect multiple charging pile platforms to the central scheduling server, construct a cross-platform charging pile resource pool, and monitor the cross-platform charging pile resource pool in real time to obtain the monitoring stream of each charging pile. The resource pool multi-level partitioning module is used to predict faults based on the monitoring streams of each charging pile, establish a charging pile fault space, and partition the cross-platform charging pile resource pool into a multi-level partition based on the charging pile fault space to obtain a first cluster of charging piles, a second cluster of charging piles, and a third cluster of charging piles. The cross-platform scheduling compensation module is used to perform cross-platform scheduling compensation on the third cluster of charging piles based on the first cluster of charging piles and the second cluster of charging piles, and generate the first group of scheduling compensation. The mutation joint optimization module is used to introduce a scheduling response evaluation model to perform response evaluation optimization on the first group of scheduling compensation, construct a second group of scheduling compensation, and perform mutation joint optimization on the second group of scheduling compensation according to the scheduling compensation mutation capacity to generate a scheduling compensation strategy. The compensation management module is used to perform compensation management on the cross-platform charging pile resource pool based on the central scheduling server and the scheduling compensation strategy.

Images