Charging pile OCPP protocol conversion adaptation method based on intelligent algorithm

Abstract

The application relates to the technical field of charging pile communication protocol, and discloses a charging pile OCPP protocol conversion and adaptation method based on an intelligent algorithm. The method collects original communication data streams of multiple brands of charging piles, and analyzes the original communication data streams to obtain basic protocol data sets containing real-time operation parameters and historical interaction records; the method obtains heterogeneous protocol instruction sets issued by a background management system for each charging pile identifier and conforming to different communication specifications. Each type of protocol field in the heterogeneous protocol instruction sets is input into a pre-trained protocol feature extraction model to generate a protocol feature vector set, and a fusion protocol feature is integrated through a dynamic weighted fusion mechanism; a corresponding standardized OCPP instruction set and an operation type identifier are generated in combination with a protocol conversion model library, and a bidirectional communication adaptation channel between the charging pile and the background system is further constructed. The method can realize communication adaptation between multiple brands of charging piles and the background system, cope with different protocol differences, and adapt to the development demand of increasing charging pile quantity and types.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of charging pile communication protocol, in particular to a charging pile OCPP protocol conversion and adaptation method based on intelligent algorithm. BACKGROUND

[0002] With the rapid development of the new energy vehicle industry, charging piles as important infrastructure, their number and variety are increasing. There are many brands of charging piles on the market, and different brands of charging piles often use their own independent communication protocols. These protocols have significant differences in data format, instruction set, and interaction logic.

[0003] The background management system needs to monitor, dispatch and manage different brands of charging piles uniformly, but due to the non-uniformity of the communication protocols followed by each charging pile, there are many obstacles in the communication between the background management system and the charging pile. For example, when the background management system issues instructions to different brands of charging piles, it needs to write separate adaptation programs for each brand's protocol, which not only increases the complexity of system development, but also greatly increases the difficulty of system maintenance.

[0004] With the continuous upgrading of charging pile technology, new communication specifications and protocols are emerging, and the original adaptation program is difficult to quickly respond to the requirements of new protocols, which limits the scalability of the system. In addition, the conversion between different protocols often relies on manual writing of conversion rules, which is prone to rule omissions or errors, affecting the accuracy and stability of communication. These problems make the information exchange between the charging pile and the background management system inefficient, restricting the large-scale development of the charging pile network and the improvement of the intelligent management level. SUMMARY

[0005] The purpose of the present application is to provide a charging pile OCPP protocol conversion and adaptation method based on intelligent algorithm to solve the problems raised in the background.

[0006] To achieve the above purpose, the present application provides a charging pile OCPP protocol conversion and adaptation method based on intelligent algorithm, which comprises:

[0007] Collecting the original communication data stream of multi-brand charging piles, analyzing the protocol characteristics of the original communication data stream to obtain a basic protocol data set, wherein the original communication data stream contains real-time running parameters and historical interaction records of the charging pile;

[0008] For each charging pile identifier in the basic protocol data set, obtain the heterogeneous protocol instruction set conforming to different communication specifications issued by the background management system for the charging pile identifier;

[0009] For each heterogeneous protocol instruction set in the obtained heterogeneous protocol instruction set, respective protocol fields within the heterogeneous protocol instruction set are input into a pre-trained protocol feature extraction model to generate a protocol feature vector set;

[0010] For each protocol feature vector set in the obtained protocol feature vector set, a dynamic weighted fusion mechanism is used to integrate feature vectors within the protocol feature vector set to generate a fused protocol feature;

[0011] For each fused protocol feature in the obtained fused protocol feature set, a corresponding standardized OCPP instruction set and its operation type identifier are generated in combination with a protocol conversion model library and the fused protocol feature;

[0012] According to the basic protocol data set, the protocol conversion model library, the generated standardized OCPP instruction set, and its operation type identifier, a bidirectional communication adaptation channel between the charging pile and the background system is constructed.

[0013] Preferably, the protocol feature analysis of the original communication data stream to obtain a basic protocol data set comprises:

[0014] Protocol syntax structure features are identified from the original communication data stream;

[0015] Based on the protocol syntax structure features, abnormal data frames are cleaned to generate purified protocol data;

[0016] Through a pre-set protocol integrity verification rule, a subset that meets the communication reliability condition is selected from the purified protocol data as the basic protocol data set.

[0017] Preferably, the dynamic weighted fusion mechanism is used to integrate the feature vectors within the protocol feature vector set to generate a fused protocol feature, comprising:

[0018] A main fusion channel based on protocol syntax tree topology analysis is constructed to extract the logical dependency relationship between protocol fields;

[0019] An auxiliary fusion channel based on communication timing characteristics is established to analyze the timestamp distribution pattern of instruction interaction;

[0020] By calculating the semantic correlation measure value of the main and auxiliary fusion channels, the fusion weight coefficient of the feature vector is dynamically adjusted.

[0021] Preferably, for each fused protocol feature in the obtained fused protocol feature set, a corresponding standardized OCPP instruction set and its operation type identifier are generated in combination with a protocol conversion model library and the fused protocol feature, comprising:

[0022] The conversion requirement coefficient of each model in the protocol conversion model library is obtained;

[0023] selecting a protocol conversion model whose conversion demand coefficient reaches an activation threshold as a core conversion engine;

[0024] inputting the fusion protocol features into the core conversion engine to output a primary OCPP instruction set and an operation type identifier thereof;

[0025] inputting the fusion protocol features into an auxiliary conversion model group in a protocol conversion model library to generate a secondary OCPP instruction set and an operation type identifier thereof;

[0026] for each instruction in the primary OCPP instruction set, performing the following verification steps: detecting whether there is an instruction with a matching operation type identifier in the secondary OCPP instruction set;

[0027] when a matching instruction is detected, counting the frequency of occurrence of the operation type in the secondary OCPP instruction set;

[0028] if the frequency exceeds a preset consensus threshold, confirming the corresponding instruction in the primary OCPP instruction set as a standardized OCPP instruction.

[0029] Preferably, the two-way communication adaptation channel between the charging pile and the background system comprises:

[0030] generating an intermediate protocol conversion instruction set and an operation type identifier thereof according to the basic protocol data set and the protocol conversion model library;

[0031] associating and mapping the operation type identifier of the intermediate protocol conversion instruction set with the operation type identifier of the standardized OCPP instruction group;

[0032] establishing an instruction conversion routing table based on the mapping result to generate configuration parameters of the two-way communication adaptation channel.

[0033] Preferably, the generation of the intermediate protocol conversion instruction set comprises:

[0034] calculating the historical calling priority of each model in the protocol conversion model library;

[0035] selecting a protocol conversion model whose historical calling priority reaches an execution standard as a target conversion model;

[0036] inputting the basic protocol data set into the target conversion model to generate an initial intermediate instruction set;

[0037] inputting the basic protocol data set into a candidate conversion model group in the protocol conversion model library to generate a reference intermediate instruction set;

[0038] For each instruction in the initial intermediate instruction set, the following verification steps are performed: verifying whether there exists an instruction in the reference intermediate instruction set with the same operation type;

[0039] When a verification exists, aggregating all reference intermediate instructions with the same operation type to form a verification instruction group;

[0040] If the size of the verification instruction group reaches a reliability threshold, confirming that the corresponding instruction in the initial intermediate instruction set is valid.

[0041] Preferably, the method further comprises:

[0042] Monitoring the real-time data throughput of the bidirectional communication adaptation channel;

[0043] When the data throughput exceeds a load threshold, starting a protocol conversion resource scheduler;

[0044] According to the output instructions of the protocol conversion resource scheduler, optimizing the concurrent processing capacity of the bidirectional communication adaptation channel.

[0045] Preferably, the starting protocol conversion resource scheduler comprises:

[0046] Parsing the operation type identifier of each instruction in the currently running standardized OCPP instruction group;

[0047] Clustering and grouping the conversion tasks based on the operation type identifier matching degree;

[0048] According to the resource occupancy rate sorting of the clustered groups, preferentially executing the conversion tasks of the high resource occupancy rate groups;

[0049] Placing the conversion tasks of the low resource occupancy rate groups into a delayed processing queue.

[0050] Preferably, the optimizing the concurrent processing capacity of the bidirectional communication adaptation channel comprises:

[0051] Establishing a state value evaluation function of the protocol conversion model;

[0052] When the state value evaluation function is lower than an active threshold, releasing the computing resources occupied by the corresponding protocol conversion model;

[0053] According to the historical conversion task load law, predicting the resource demand peak value;

[0054] Reserving a computing resource buffer for high-priority protocol conversion models.

[0055] Preferably, the releasing the computing resources occupied by the corresponding protocol conversion model comprises:

[0056] Identifying the shared weight modules between protocol conversion models;

[0057] Eliminate the repetition weight module by the communication feature fingerprint comparison algorithm;

[0058] Only unload the memory space corresponding to the non-shared weight module;

[0059] When the idle duration of the shared weight module exceeds the recovery threshold, trigger the distributed memory recovery mechanism.

[0060] Compared with the prior art, the beneficial effects of the present application are:

[0061] By collecting the original communication data stream of multiple brand charging piles and performing protocol feature analysis, the protocol characteristics of different charging piles can be comprehensively mastered, providing sufficient basic information for subsequent protocol conversion. The heterogeneous protocol instruction set issued by the background management system for different charging pile identifiers enables the conversion process to closely integrate with actual management needs, ensuring that the converted instructions match the actual application scenario.

[0062] Using a pre-trained protocol feature extraction model to process various protocol fields in the heterogeneous protocol instruction set generates a protocol feature vector set, which can accurately capture key information in the protocol fields, reduce information loss, and improve the accuracy of feature representation. Through a dynamic weighted fusion mechanism, the feature vectors are integrated to generate a fusion protocol feature, which can be reasonably weighted according to the importance of different features, making the fused feature better reflect the overall characteristics of the protocol and enhance the representativeness of the feature.

[0063] Combining the protocol conversion model library and the fusion protocol feature generates corresponding standardized OCPP instruction sets and their operation type identifiers, effectively converting heterogeneous protocols into standardized protocols, eliminating differences between different protocols, and enabling charging piles of different brands to communicate with the background system using a unified protocol. The constructed bidirectional communication adaptation channel breaks down the communication barriers between charging piles and the background system, enabling efficient and smooth information exchange between the two, facilitating the background system to manage and schedule multiple brand charging piles, and improving the collaborative working ability of the entire charging pile network.

[0064] This method does not require manual writing of a large number of conversion rules, reducing the dependence on human experience, reducing errors caused by human factors, and improving the efficiency and flexibility of protocol conversion, enabling quick adaptation to new communication protocols and specifications, enhancing the adaptability and scalability of the system. BRIEF DESCRIPTION OF DRAWINGS

[0065] Figure 1 The working principle diagram of the charging pile OCPP protocol conversion adaptation method based on intelligent algorithm described in the present application;

[0066] Figure 2A flow chart for generating fusion protocol features for a dynamic weighted fusion mechanism;

[0067] Figure 3 A flow chart for generating and verifying a standardized OCPP instruction set;

[0068] Figure 4 A flow chart for generating and verifying an intermediate protocol conversion instruction set;

[0069] Figure 5 A flow chart for a concurrent processing capability optimization strategy. DETAILED DESCRIPTION

[0070] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of the present application.

[0071] Please refer to Figure 1 The present application provides a charging pile OCPP protocol conversion adaptation method based on intelligent algorithm, which comprises:

[0072] Collecting original communication data streams of multi-brand charging piles, performing protocol feature analysis on the original communication data streams to obtain a basic protocol data set, wherein the original communication data streams contain real-time running parameters and historical interaction records of the charging piles; for each charging pile identifier in the basic protocol data set, obtaining a heterogeneous protocol instruction set issued by a background management system for the charging pile identifier, which conforms to different communication specifications; for each heterogeneous protocol instruction set in the obtained heterogeneous protocol instruction group, inputting each type of protocol field in the heterogeneous protocol instruction set into a pre-trained protocol feature extraction model to generate a protocol feature vector set; for each protocol feature vector set in the obtained protocol feature vector group, integrating the feature vectors in the protocol feature vector set by using a dynamic weighted fusion mechanism to generate a fusion protocol feature; for each fusion protocol feature in the obtained fusion protocol feature set, generating a corresponding standardized OCPP instruction set and its operation type identifier in combination with a protocol conversion model library and the fusion protocol feature; and constructing a bidirectional communication adaptation channel between the charging pile and the background system according to the basic protocol data set, the protocol conversion model library, the generated standardized OCPP instruction group and its operation type identifier.

[0073] Embodiment 1: Please refer to Figure 2In the protocol feature analysis of the original communication data stream to obtain the basic protocol data set, the protocol syntax structure features are identified from the original communication data stream, which includes the positioning of data frame start symbol, end symbol and field separator, and the extraction of data type (such as numerical type, character type, Boolean type), length constraint and value range of each field. By scanning the original communication data stream frame by frame, comparing with the preset syntax rule template, the differences of different brands of charging pile protocols in field arrangement order, check digit calculation method, etc. are determined. Based on the identified protocol syntax structure features, the abnormal data frames are cleaned, and the specific operations include eliminating the data frames that do not conform to the syntax rules, such as the data frames whose field length exceeds the specified range and the data frames whose data type does not match the definition; for the data frames that have some missing fields but can be completed by adjacent valid data frames or historical data, interpolation method or pattern matching method is used for repair to generate purified protocol data. Through the preset protocol integrity verification rule, the subset that meets the communication reliability condition is selected from the purified protocol data as the basic protocol data set, and the protocol integrity verification rule includes data frame sequence continuity check, i.e. verifying whether the sequence number of the data frame in the instruction interaction process is continuous, without skipping or repeating; it includes key field integrity check, i.e. ensuring that the fields related to the charging pile running state, instruction execution result and other key information are not missing; it includes check digit verification, i.e. by recalculating the check digit of the data frame and comparing with the original check digit, confirming that the data has not been tampered or damaged in the transmission process, only the data that passes all verification rules is included in the basic protocol data set.

[0074] When the dynamic weighted fusion mechanism is used to integrate the feature vectors in the protocol feature vector set to generate the fused protocol features, a main fusion channel based on protocol syntax tree topology analysis is constructed. The protocol instructions are parsed into a syntax tree structure, in which the root node represents the entire instruction and the child nodes represent the protocol fields at different levels. By traversing the node relationship of the syntax tree, the logical dependency relationships between the protocol fields are extracted, such as the fact that the "charging start instruction" field must depend on the existence of the "charging pile number" field and the "user identity authentication" field to be effective, and the fact that the "charging power adjustment" field has a numerical correlation relationship with the "current charging current" field. An auxiliary fusion channel based on communication timing characteristics is established, and information such as the sending timestamp, receiving timestamp, and response timestamp of each instruction is collected. The timestamp distribution pattern of instruction interaction is analyzed, including the sending interval law of instructions of the same type, the response delay distribution between different instructions, the instruction frequency difference between peak and valley periods, and other timing characteristics. The semantic correlation measure value of the main and auxiliary fusion channels is calculated. This value is obtained by quantifying the degree of association between the logical dependency relationships and the timing patterns. For example, when two protocol fields have a strong logical dependency relationship and often appear together in timing, their semantic correlation measure value is high; otherwise, it is low. The fusion weight coefficient of the feature vector is dynamically adjusted according to the semantic correlation measure value. Feature vectors with high semantic correlation measure values are given higher weights and have a greater influence in the integration process. Finally, all feature vectors are weighted and summed according to the adjusted weights to generate the fused protocol features. These fused protocol features not only contain the logical association information between the protocol fields, but also integrate the timing characteristics of instruction interaction, and can more comprehensively reflect the essential features of the protocol.

[0075] Example 2: see Figure 3In the protocol feature analysis of the original communication data stream to obtain the basic protocol data set, the protocol syntax structure features are identified from the original communication data stream, which requires frame-by-frame analysis of the communication data of the multi-brand charging pile, extraction of key syntax elements such as start markers, end markers and field separators in the data frame, determination of the data type of each field, such as numerical fields that may be used to represent charging voltage, current and other parameters, character fields that may be used to identify charging pile number, fault code and other information, and the length limit and position sequence of each field in the data frame. Based on these identified protocol syntax structure features, abnormal data frames are cleaned up. Specifically, for those data frames whose field length exceeds the specified range, whose data type does not match the definition, or which lack key separators, they are determined as abnormal and removed; for data frames in which some fields are missing but the reasonable values can be inferred from the rules of the preceding and following valid data frames, a filling method based on historical data distribution is used for repair, and the purified protocol data is generated after such processing. Through the preset protocol integrity verification rule, a subset that meets the communication reliability condition is selected from the purified protocol data as the basic protocol data set. The protocol integrity verification rule includes a continuity check of the data frame sequence, i.e. verifying whether the sequence number of the data frame is sequentially increased and has no repetition according to the sending order, to ensure the continuity of the instruction interaction process; includes a completeness check of the key information field, such as the instructions involving charging start, stop, fault alarm and other core operations must contain complete parameter information; includes a data integrity check, which is performed by recalculating the check value of the data frame and comparing it with the original check value to confirm that the data has not been error or tampered in the transmission process. Only the data that passes all these verifications will be included in the basic protocol data set.

[0076] When using a dynamic weighted fusion mechanism to integrate feature vectors within the protocol feature vector set to generate fused protocol features, a main fusion channel based on protocol syntax tree topology analysis is constructed. First, each protocol instruction is converted into a syntax tree structure, where the root node represents the entire instruction, and each child node represents a different level of protocol field. By analyzing the parent-child and sibling relationships between nodes in the syntax tree, the logical dependencies between protocol fields are extracted. For example, the value of the "charging mode selection" field directly affects the effective range of the "charging current limit" field, and the status of the "identity authentication result" field determines the validity of the "charging start permission" field. An auxiliary fusion channel based on communication timing characteristics is established, collecting timestamp information such as the sending time, receiving time, and response time of each instruction during the interaction process. The distribution patterns of these timestamps are analyzed, including the frequency of occurrence of the same type of instruction within a unit of time, the distribution of time intervals between different instructions, and the fluctuation range of response time. For example, during peak electricity consumption periods, the sending frequency of charging request instructions increases significantly, while fault instructions are usually accompanied by shorter response times. The semantic relevance metric of the primary and secondary fusion channels is calculated. This metric is derived by comparing the co-occurrence of fields in logical dependencies with the temporal correlation of fields in temporal features. For example, when two fields have a strong logical dependency and frequently appear simultaneously in temporal sequence, their semantic relevance metric is higher. The fusion weight coefficients of the feature vectors are dynamically adjusted based on the semantic relevance metric, assigning greater weight to feature vectors with high semantic relevance metric values, allowing them to play a greater role in the integration process. Finally, all feature vectors are combined according to the adjusted weights to generate a fusion protocol feature. This fusion protocol feature integrates the logical correlation and temporal features of the protocol fields, providing a more comprehensive reflection of the overall characteristics of the protocol.

[0077] Example 3: See Figure 4 When generating the intermediate protocol conversion instruction set, the historical call priority of each model in the protocol conversion model library is calculated. This priority is determined by combining multiple historical performance indicators of the model. Specifically, the number of past calls for each model is collected, the number of successfully completed conversion tasks is counted, the time consumed for each conversion is recorded, and the system resources used during runtime, including memory usage and processor utilization, are monitored. Using this data, a specific algorithm is employed to calculate the historical call priority of each model, thus reflecting the model's overall performance in practical applications. Protocol conversion models whose historical call priority meets the execution criteria are selected as target conversion models. The execution criteria are set based on the system's overall requirements for conversion efficiency and stability; only models that demonstrate high success rates, fast processing speeds, and reasonable resource usage in historical calls are selected as target conversion models.

[0078] The base protocol data set is input into the target conversion model, which has built-in parsing rules and conversion logic for multiple common protocol formats, and can perform preliminary conversion processing on various instructions in the base protocol data set. During the conversion process, the model will identify the protocol type identifier in the data, call the corresponding conversion module according to the identifier, perform format adjustment, unit conversion, and semantic mapping operations on the parameters in the instructions, and convert the original protocol instructions into an intermediate format to generate an initial intermediate instruction set. At the same time, the base protocol data set is input into the candidate conversion model group in the protocol conversion model library, which contains a batch of models optimized for specific brand charging pile protocols or special application scenarios. These models may have unique advantages when handling some edge cases or complex protocol structures. Each model in the candidate conversion model group will independently convert the base protocol data and output corresponding intermediate instructions, which together form a reference intermediate instruction set.

[0079] For each instruction in the initial intermediate instruction set, it is necessary to verify whether there is an instruction with the same operation type in the reference intermediate instruction set. The operation type is determined by the function code or key field in the instruction, such as charging start, parameter query, fault reporting, etc. When there is an instruction with the same operation type, all these instructions with the same operation type are aggregated together to form a verification instruction group. The instructions in the verification instruction group may differ in parameter details or format, but the core operation function is consistent. If the size of the verification instruction group reaches the reliability threshold, which is set according to the system's requirements for instruction conversion accuracy, and is usually related to the size of the candidate conversion model group, when a sufficient number of candidate models convert instructions with the same operation type, it indicates that the corresponding instruction in the initial intermediate instruction set has consistency and universality in conversion logic, and the instruction is confirmed to be valid. Through this process, each instruction in the initial intermediate instruction set is checked one by one, and finally an intermediate protocol conversion instruction set that has passed validity verification is formed.

[0080] In calculating the historical call priority, the following formula can be used:

[0081]

[0082] wherein, represents the historical call priority, represents the number of successful conversions of the model in history, represents the total number of calls of the model in history, represents the average conversion speed of the model (unit: instructions / second), represents the average memory usage of the model (unit: MB), represents the average processor occupancy rate of the model (unit: %), , , are the weight coefficients of each index respectively, and .

[0083] Example 4: Monitor the real-time data throughput of the bidirectional communication adaptation channel, the real-time data throughput is realized by continuously collecting the number of instructions transmitted per unit time, the total number of data frame bytes in the channel. The collection process covers the input port and output port of the channel, respectively records the original protocol instruction flow entering the channel and the standardized OCPP instruction flow output, at the same time marks the transmission time point of each instruction, forms a continuous throughput monitoring sequence. When the data throughput exceeds the load threshold, the load threshold is set according to the hardware processing capacity of the channel, the upper limit of the protocol conversion model calculation and the historical peak flow, at this time the protocol conversion resource scheduler is started, and the conversion task in the channel is dynamically managed.

[0084] Analyze the operation type identifier of each instruction in the currently running standardized OCPP instruction group, the operation type identifier is determined by the function code field in the instruction header, for example, function code "0x01" represents charging parameter query, "0x02" represents charging start control, "0x03" represents fault state reporting, etc. Each operation type identifier corresponds to an instruction structure and processing logic, for example, parameter query type instruction contains a list of parameter numbers to be queried, control type instruction contains specific operation parameter values. Based on the matching degree of operation type identifier, the conversion task is clustered and grouped, the matching degree is calculated by comparing the consistency of function code field and the correlation degree of instruction parameters, the conversion tasks with the same function code and high parameter correlation degree are classified into the same group, for example, all parameter query instructions for the same charging pile are clustered into a group, and start control instructions for different charging piles are clustered into another group.

[0085] According to the resource occupation rate sorting of clustering and grouping, the resource occupation rate is determined by counting the CPU processing time, memory occupation and protocol conversion model calling times required by each group of tasks, for example, fault state reporting type task needs to process a large number of state code mapping, so the resource occupation rate is usually higher than that of ordinary parameter query task. The conversion task of high resource occupation rate group is preferentially executed, and more computing cores and memory space are allocated to it to shorten the waiting time of the task in the processing queue; the conversion task of low resource occupation rate group is put into the delay processing queue, the delay processing queue adopts the first-in first-out scheduling strategy, and the tasks are executed in turn after the high resource occupation rate group tasks are processed.

[0086] According to the output instructions of the protocol conversion resource scheduler, the concurrent processing capability of the bidirectional communication adaptation channel is optimized, specifically including dynamically adjusting the size of the thread pool in the channel, adding dedicated processing threads for high-priority task groups; adjusting the calling order of the protocol conversion model, giving priority to high-resource-occupancy groups to use the model instances with better performance; setting the threshold of the task buffer queue, and temporarily enabling the standby computing resources for processing when the length of the low-priority task queue exceeds the threshold.

[0087]

[0088] Through the above process, the protocol conversion resource scheduler can intelligently schedule according to the type and resource demand of the task when the channel load is too high, ensuring efficient processing of critical conversion tasks, while balancing the overall utilization of system resources, and realizing dynamic optimization of the concurrent processing capability of the bidirectional communication adaptation channel.

[0089] Embodiment 5: refer to Figure 5 When optimizing the concurrent processing capability of the bidirectional communication adaptation channel, a state value evaluation function of the protocol conversion model is established, which comprehensively considers the current conversion efficiency of the model, i.e. the number of protocol conversion tasks completed per unit time; the utilization of system resources by the model, including memory occupation ratio and processor usage; and the length of the task queue associated with the model, i.e. the number of tasks waiting to be processed by the model. Through the comprehensive calculation of these indicators, a quantitative evaluation of the current running state of the model is formed. When the state value evaluation function is lower than the active threshold, it indicates that the running efficiency of the protocol conversion model in the current system environment is low, and its contribution to the overall conversion process is limited. At this time, the computing resources occupied by the model are released, including recovering the processor computing time allocated to the model, emptying the memory space occupied by the model during running, and releasing the occupation of the specific protocol conversion interface.

[0090] According to the historical conversion task load law, the resource demand peak value is predicted, and the historical conversion task load law is obtained by analyzing the number of different types of protocol conversion tasks and resource consumption curves in different time periods in the past period, for example, the number of charging-related instruction conversion tasks during the morning and evening peak periods on weekdays is usually significantly higher than that during other periods, and the corresponding resource demand also increases accordingly. Based on these laws, the time series analysis method is used to infer the maximum value of resource demand that may occur within a certain period in the future, i.e. the resource demand peak value.

[0091] The computing resource buffer is reserved for high-priority protocol conversion models, which are determined according to the importance of the instruction types they process, for example, models responsible for charging safety parameter conversion and emergency fault instruction processing are set as high priority. The computing resource buffer includes specially allocated processor cores, independent memory blocks and preferential network bandwidth, ensuring that these high-priority models can continuously obtain sufficient computing support without being affected by other tasks during peak resource demand periods.

[0092] When releasing the computing resources occupied by the corresponding protocol conversion model, first identify the shared weight modules between the protocol conversion models, which are basic algorithm components commonly relied on by multiple models when performing conversion tasks, such as general field parsing modules and data format conversion functions. Through a communication feature fingerprint comparison algorithm, the weight modules of each model are feature extracted to generate unique feature fingerprints, and repeated weight modules are identified by comparing fingerprints, and one instance is retained and the other repeated modules are eliminated. Only the memory space corresponding to the non-shared weight modules is unloaded, which are unique to a specific model and will not affect the normal operation of other models. When the idle duration of the shared weight module exceeds the recovery threshold, the idle duration is calculated by recording the difference between the last call time and the current time, and the recovery threshold is dynamically adjusted according to the tightness of the system memory resources. At this time, the distributed memory recovery mechanism is triggered to release the shared weight module from the memory of all nodes, and the recovered memory resources are unified into the system resource pool for other modules to apply.

[0093] It should be noted that, in the present document, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0094] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A method for converting and adapting OCPP protocol of charging piles based on intelligent algorithm, characterized in that, The method comprises the following steps: Collecting original communication data streams of multi-brand charging piles, performing protocol feature analysis on the original communication data streams to obtain a basic protocol data set, wherein the original communication data streams contain real-time operation parameters and historical interaction records of the charging piles; For each charging pile identifier in the basic protocol data set, obtaining a heterogeneous protocol instruction set issued by a background management system for the charging pile identifier, which conforms to different communication specifications; For each heterogeneous protocol instruction set in the obtained heterogeneous protocol instruction group, inputting each type of protocol field in the heterogeneous protocol instruction set into a pre-trained protocol feature extraction model to generate a protocol feature vector set; For each protocol feature vector set in the obtained protocol feature vector group, integrating the feature vectors in the protocol feature vector set using a dynamic weighted fusion mechanism to generate a fused protocol feature; For each fused protocol feature in the obtained fused protocol feature set, combining a protocol conversion model library and the fused protocol feature to generate a corresponding standardized OCPP instruction set and its operation type identifier; According to the basic protocol data set, the protocol conversion model library, the generated standardized OCPP instruction group and its operation type identifier, a bidirectional communication adaptation channel between the charging pile and the background system is constructed; The method for generating a corresponding standardized OCPP instruction set and its operation type identifier for each fused protocol feature in the obtained fused protocol feature set comprises: Obtaining the conversion demand coefficient of each model in the protocol conversion model library; Selecting a protocol conversion model with a conversion demand coefficient reaching an activation threshold as a core conversion engine; Inputting the fused protocol feature into the core conversion engine to output a primary OCPP instruction set and its operation type identifier; Inputting the fused protocol feature into a group of auxiliary conversion models in the protocol conversion model library to generate a secondary OCPP instruction set and its operation type identifier; For each instruction in the primary OCPP instruction set, the following verification steps are performed: detecting whether there is an instruction with a matching operation type identifier in the secondary OCPP instruction set; When a matching instruction is detected, the frequency of the operation type in the secondary OCPP instruction set is counted; If the frequency exceeds a preset consensus threshold, the corresponding instruction in the primary OCPP instruction set is confirmed as a standardized OCPP instruction.

2. The intelligent algorithm based charging pile OCPP protocol conversion adaptation method according to claim 1, characterized in that, The protocol feature analysis on the original communication data stream to obtain a basic protocol data set comprises: Identifying protocol syntax structure features from the original communication data stream; Cleaning abnormal data frames based on the protocol syntax structure features to generate purified protocol data; Through a preset protocol integrity verification rule, a subset meeting the communication reliability condition is selected from the purified protocol data as a basic protocol data set. 3.The smart algorithm based charging pile OCPP protocol conversion adaptation method according to claim 1, characterized in that, The integration of the feature vectors in the protocol feature vector set using a dynamic weighted fusion mechanism to generate a fused protocol feature comprises: Constructing a main fusion channel based on protocol syntax tree topology analysis to extract the logical dependency relationship between protocol fields; A secondary fusion channel based on communication timing characteristics is established to analyze the timestamp distribution pattern of instruction interaction; By calculating the semantic correlation measure value of the primary and secondary fusion channels, the fusion weight coefficient of the feature vector is dynamically adjusted.

4. The intelligent algorithm based charging pile OCPP protocol conversion adaptation method according to claim 1, characterized in that, The two-way communication adaptation channel between the charging pile and the background system comprises: According to the basic protocol data set and the protocol conversion model library, an intermediate protocol conversion instruction set and its operation type identifier are generated; The operation type identifier of the intermediate protocol conversion instruction set is associated and mapped with the operation type identifier of the standardized OCPP instruction group; Based on the mapping result, an instruction conversion routing table is established, and the configuration parameters of the two-way communication adaptation channel are generated.

5. The intelligent algorithm based charging pile OCPP protocol conversion adaptation method according to claim 4, characterized in that, The generation of the intermediate protocol conversion instruction set comprises: Calculate the historical call priority of each model in the protocol conversion model library; Screen the protocol conversion models whose historical call priority reaches the execution standard as target conversion models; Input the basic protocol data set into the target conversion model to generate an initial intermediate instruction set; Input the basic protocol data set into the candidate conversion model group in the protocol conversion model library to generate a reference intermediate instruction set; For each instruction in the initial intermediate instruction set, the following verification steps are performed: verify whether there is an instruction with the same operation type in the reference intermediate instruction set; When there is, aggregate all reference intermediate instructions with the same operation type to form a verification instruction group; If the size of the verification instruction group reaches the reliability threshold, confirm that the corresponding instruction in the initial intermediate instruction set is valid.

6. The intelligent algorithm based charging pile OCPP protocol conversion adaptation method according to claim 1, characterized in that, Further comprising: Monitoring the real-time data throughput of the two-way communication adaptation channel; When the data throughput exceeds the load threshold, starting the protocol conversion resource scheduler; According to the output instruction of the protocol conversion resource scheduler, optimizing the concurrent processing capacity of the two-way communication adaptation channel.

7. The intelligent algorithm based charging pile OCPP protocol conversion adaptation method according to claim 6, characterized in that, The starting of the protocol conversion resource scheduler comprises: Parsing the operation type identifier of each instruction in the currently running standardized OCPP instruction group; Based on the operation type identifier matching degree, the conversion tasks are clustered and grouped; According to the resource occupancy rate sorting of the clustered groups, the conversion tasks in the high resource occupancy rate groups are preferentially executed; The conversion tasks in the low resource occupancy rate groups are placed in the delay processing queue.

8. The intelligent algorithm based charging pile OCPP protocol conversion adaptation method according to claim 7, characterized in that, The optimization of the concurrent processing capacity of the two-way communication adaptation channel comprises: Establishing a state value evaluation function of the protocol conversion model; When the state value evaluation function is lower than the active threshold, releasing the computing resources occupied by the corresponding protocol conversion model; According to the historical conversion task load law, predicting the resource demand peak value; Reserving a computing resource buffer for high-priority protocol conversion models.

9. The intelligent algorithm based charging pile OCPP protocol conversion adaptation method according to claim 8, characterized in that, The release of the computing resources occupied by the corresponding protocol conversion model comprises: Identifying shared weight modules among protocol conversion models; Eliminating duplicate weight modules through a communication feature fingerprint comparison algorithm; Only unloading the memory space corresponding to the non-shared weight module; When the idle duration of the shared weight module exceeds the recycling threshold, triggering a distributed memory recycling mechanism.

Images