An energy storage management system based on AI recognition model

Abstract

The present invention relates to the technical field of energy storage management, and discloses an energy storage management system based on an AI recognition model; the method comprises acquiring communication characteristics of each subunit, identifying an adapter, constructing an energy storage management scenario, acquiring comprehensive management parameters of the energy storage management scenario, calculating a management reliability index of the energy storage management scenario, determining whether to generate a re-identification prompt, formulating a re-identification instruction, and re-identifying the adapter; compared with the prior art, the present invention can specifically predict an adapter corresponding to a communication protocol and a point table through an AI recognition model, and can accurately calculate the adaptation operation reliability between the identified adapter and the subunit, and can re-identify the adapter when it is found that the degree of adaptation is not good, thereby ensuring that a fast and accurate data link and control chain conduction adaptation effect can be achieved between the subunits of the energy storage management system, thereby improving the operation efficiency of the energy storage management system.

Description

Technical Field

[0001] The present invention relates to the technical field of energy storage management, and more specifically, to an energy storage management system based on an AI recognition model. Background Art

[0002] At present, for many integrated products, several sub-units need to be linked together to achieve mutual communication and control. In the BMS, PCS, and EMS sub-units in the power energy storage system, when the sub-units are upgraded or updated, the data communication links between the sub-units need to be manually opened to achieve the overall management and control effect of the energy storage system. In order to improve the rapid adaptation of data links and control chains between sub-units, it is necessary to use AI recognition models to manage the energy storage system.

[0003] The patent application with reference publication number CN103227494A discloses an energy storage battery management system, which adopts a distributed three-layer management system, including multiple bottom-level battery management units BMU, multiple middle-level battery cluster management systems BCMS and a top-level battery system unit management system BAMS; each bottom-level battery management unit BMU manages the corresponding battery unit; each middle-level battery cluster management system BCMS manages multiple bottom-level battery management units BMU through a CAN bus; the top-level battery system unit management system BAMS manages all middle-level battery cluster management systems BCMS through a CAN bus;

[0004] In order to achieve the effect of connecting the data chain and control chain of each sub-unit in the energy storage system, the existing energy storage management system needs to obtain the communication protocol and point table of each sub-unit, and use manual analysis and calculation to obtain the adapter program. Although the effect of connecting the data chain and the control chain can be achieved, every time a sub-unit is upgraded or updated in the energy storage system, the overall development of the adapter program needs to be manually performed, resulting in a long development time and low accuracy of the adapter program, which cannot meet the efficient operation requirements of the energy storage system.

[0005] In view of this, the present invention proposes an energy storage management system based on an AI recognition model to solve the above problems. Summary of the invention

[0006] In order to overcome the above-mentioned defects of the prior art and to achieve the above-mentioned purpose, the present invention provides the following technical solution: an energy storage management system based on an AI recognition model, applied to an AI server, comprising:

[0007] The model recognition module is used to obtain the communication characteristics of each sub-unit, input the communication characteristics into the pre-trained AI recognition model, and identify the adapter corresponding to the communication characteristics. The communication characteristics include the communication protocol and the point table;

[0008] A scenario construction module is used to construct an initial management scenario with a control link and send the adapter to the control point of the control link to generate an energy storage management scenario;

[0009] An index calculation module is used to obtain comprehensive management parameters of the energy storage management scenario and calculate the management reliability index of the energy storage management scenario based on the comprehensive management parameters. The comprehensive management parameters include the standard format ratio, data delay duration, and abnormal state ratio.

[0010] A prompt determination module, used for comparing the management reliability index with a preset management reliability threshold, and determining whether to generate a re-identification prompt according to the comparison result;

[0011] The re-identification module is used to identify the optimization parameters from the comprehensive management parameters, formulate the re-identification instructions, and re-identify the adaptation program.

[0012] Furthermore, the training methods of AI recognition models include:

[0013] Pre-collecting multiple groups of communication characteristics and adaptation programs corresponding to the communication characteristics;

[0014] Mark each group of communication features as training features, annotate the adapter program for each group of training features, and divide the annotated training features into a training set and a test set; use the training set to train the AI ​​recognition model, and use the test set to test the AI ​​recognition model; preset an error threshold, and when the mean of the prediction errors of all training features in the test set is less than the error threshold, output the AI ​​recognition model that identifies the adapter program.

[0015] Furthermore, the method for constructing the initial management scenario includes:

[0016] Three units with a triangular structure are constructed by virtual reality technology, and the unit located at the vertex of the triangle is marked as the first unit, and the remaining two units are marked as the second unit and the third unit from left to right;

[0017] Establish data frames on the first unit, the second unit and the third unit respectively, and import the energy management system, the battery management system and the energy storage converter into the data frames on the first unit, the second unit and the third unit one by one to generate a basic management scenario;

[0018] In the basic management scenario, a data transfer point is established at a middle position between the first unit position, the second unit position and the third unit position, which is recorded as a transfer point;

[0019] Two-way conductive control links are established one by one between the transfer position and the first unit position, the second unit position and the third unit position to generate an initial management scenario.

[0020] Furthermore, the method for generating the energy storage management scenario includes:

[0021] The connection end of the control link and the transit position is recorded as the starting end, and the connection end of the control link and the first unit position, the second unit position and the third unit position is recorded as the terminating end, and the starting ends and terminating ends of the three control links are marked one by one;

[0022] Mark all points on the control link, and record the middle point between the starting end and the ending end on the control link as the import point, and obtain three import points;

[0023] A front contour line is established on the side close to the starting end at the three lead-in points, and a rear contour line is established on the side close to the ending end, and the front contour line, lead-in point and rear contour line located on the same control link are assigned the same number to obtain three control points;

[0024] At the same time, the adapter program is sent to three control points to generate an energy storage management scenario.

[0025] Furthermore, the method for obtaining the standard format ratio includes:

[0026] Query the attribute status of the data through the attribute management system, and record the data with the attribute status being processed as valid data;

[0027] The time when valid data first appears in the energy storage management scenario is queried through the timestamp and recorded as the starting time;

[0028] When no valid data appears within the preset time, the last time valid data appears is recorded as the end time, and the period from the start time to the end time is recorded as the packet capture period;

[0029] Mark out the medium duration during the packet capture period capture time, and use the capture tool to Each packet capture moment captures an equal amount of valid data;

[0030] The data format of valid data is queried through the format management system, and valid data with the same data format as the standard format is recorded as standard data, and the number of standard data is counted;

[0031] Will The number of standard data at each capture time is compared with the number of valid data in turn to obtain The percentage of children;

[0032] The expression of the sub-ratio is:

[0033] ;

[0034] In the formula, For the The percentage of sub-packets at the time of packet capture, For the The number of standard data at each capture moment, is the number of valid data;

[0035] Will The percentages of each child are accumulated and averaged to obtain the percentage in standard format;

[0036] The expression of the standard format ratio is:

[0037] ;

[0038] In the formula, is the standard format ratio, For the The percentage of children.

[0039] Furthermore, the method for obtaining the data delay duration includes:

[0040] At time T1, the p valid data are numbered in ascending order, and the numbered valid data are synchronously imported into the data processor;

[0041] Record the p valid data exported by the data processor one by one, and query the exporting time of the p valid data in order from small to large numbers to obtain p exporting times;

[0042] The duration between time T1 and p export times is recorded as the real-time processing duration, and the real-time processing duration greater than the standard processing duration is recorded as the delayed processing duration. Delay processing time;

[0043] Will The delayed processing time is subtracted from the standard processing time in turn, and then The differences are accumulated one by one to obtain the data delay time;

[0044] The expression of data delay time is:

[0045] ;

[0046] In the formula, is the data delay time, For the Delayed processing time, This is the standard processing time.

[0047] Furthermore, the method for obtaining the abnormal state ratio value includes:

[0048] A1: At each packet capture moment, the real-time current and temperature of the battery are detected by the current sensor and temperature sensor respectively, and the real-time current and temperature of the battery are obtained. The current value and temperature value, and The current value and the corresponding temperature value combinations, generating Power supply status;

[0049] A2: The power supply state in which the current value is greater than the standard current value and the temperature value is greater than the standard temperature value is recorded as an abnormal state, and the number of abnormal states is counted;

[0050] A3: After comparing the number of abnormal states with the number of power supply states, a sub-proportion value is obtained;

[0051] A4: Repeat Repeat steps A1-A3 to obtain The percentage of children;

[0052] The expression of the sub-ratio value is:

[0053] ;

[0054] In the formula, For the The percentage of children, For the The number of abnormal states, For the The number of power supply states;

[0055] Will The percentage values ​​of each child are accumulated and averaged to obtain the percentage value of abnormal state;

[0056] The expression of abnormal status ratio is:

[0057] ;

[0058] In the formula, is the abnormal status ratio, For the The percentage of children.

[0059] Furthermore, the expression of management reliability index is:

[0060] ;

[0061] In the formula, To manage the reliability index, , , is a proportional coefficient greater than 0;

[0062] The determination method of whether to generate a re-identification prompt includes:

[0063] Management Reliability Index With the preset management reliability threshold Compare;

[0064] when Greater than or equal to When , it is determined that no re-identification prompt is generated;

[0065] when Less than , it is determined that a re-identification prompt is generated.

[0066] Further, the identification method of the optimization parameters includes:

[0067] The standard format ratio is compared with a preset standard ratio, and when the standard format ratio is less than the preset standard ratio, the standard format ratio is identified as an optimization parameter;

[0068] Comparing the data delay time with a preset standard delay, and identifying the data delay time as an optimization parameter when the data delay time is greater than the preset standard delay;

[0069] The abnormal state proportion value is compared with a preset standard proportion value, and when the abnormal state proportion value is greater than the preset standard proportion value, the abnormal state proportion value is identified as an optimization parameter.

[0070] Further, the re-identification instruction includes a priority identification instruction and a normal identification instruction;

[0071] The method for formulating priority identification instructions and general identification instructions includes:

[0072] Count the number of optimization parameters and obtain the optimization value;

[0073] When the optimization value is 1, a normal recognition instruction is formulated;

[0074] When the optimization value is 2 or 3, a priority recognition instruction is formulated.

[0075] Technical effects and advantages of an energy storage management system based on an AI recognition model of the present invention:

[0076] The present invention obtains the communication characteristics of each subunit, inputs the communication characteristics into the AI ​​recognition model trained in advance, identifies the adapter corresponding to the communication characteristics, constructs an initial management scenario with a control link, and sends the adapter to the control point of the control link to generate an energy storage management scenario, obtains the comprehensive management parameters of the energy storage management scenario, and calculates the management reliability index of the energy storage management scenario based on the comprehensive management parameters, compares the management reliability index with a preset management reliability threshold, and determines whether to generate a re-identification prompt based on the comparison result, identifies the optimization parameters from the comprehensive management parameters, formulates a re-identification instruction, and re-identifies the adapter; Compared with the prior art, by obtaining the communication protocol and point table data of each subunit and combining the AI ​​recognition model, it can specifically predict the communication protocol and point table data. The adapter corresponding to the table is used to achieve the rapid adaptation effect of the data communication link of each sub-unit in the energy storage management system. By building an energy storage management scenario and collecting comprehensive management parameters to calculate the management reliability index, the adaptation operation reliability between the identified adapter and the sub-unit can be accurately analyzed and calculated, and the adaptation degree of the data transmission link between each sub-unit in the energy storage management system can be accurately obtained. At the same time, when it is found that the adaptation degree is not good, the adapter program can be re-identified as needed to ensure that the various sub-units of the energy storage management system can achieve fast and accurate data chain and control chain conduction adaptation effects, effectively avoiding the problems of slow data chain and control chain conduction adaptation speed and low accuracy caused by unit upgrades or version updates of each sub-unit in the integrated product, and effectively improving the operation efficiency of the energy storage management system. BRIEF DESCRIPTION OF THE DRAWINGS

[0077] Figure 1 A schematic diagram of an energy storage management system based on an AI recognition model provided in Embodiment 1 of the present invention;

[0078] Figure 2 A flowchart of an energy storage management method based on an AI recognition model is provided in Embodiment 2 of the present invention. DETAILED DESCRIPTION

[0079] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0080] Example 1: Please refer to Figure 1 As shown, the energy storage management system based on the AI ​​recognition model described in this embodiment is applied to an AI server, including:

[0081] The model recognition module obtains the communication characteristics of each sub-unit, inputs the communication characteristics into the pre-trained AI recognition model, and identifies the adapter corresponding to the communication characteristics;

[0082] Subunit refers to each specific component of the energy storage management system, which can represent the specific component modules or equipment in the energy storage management system separately;

[0083] The subunits include energy management system, battery management system and energy storage converter. As the decision-making center of the energy storage management system, the energy management system is not only the monitor of energy flow in the entire energy storage management system, but also the optimizer of operation strategy, which is used to formulate various control instructions of the energy storage management system. As the core component of the energy storage management system, the battery management system ensures the safe, stable and efficient operation of the battery through sophisticated architecture and advanced technology, and is used to sense and collect various specific data of the energy storage management system. The energy storage converter plays the role of a bridge connecting the AC power grid and the battery pack in the energy storage management system, and is used to execute various control instructions and generate various specific data.

[0084] Communication characteristics are related parameters for the transmission and control of communication data of each sub-unit, including communication protocol and point table; communication protocol refers to providing communication support for each sub-unit to connect different operating systems and different hardware architectures. Point table refers to the specific representation of hardware devices and software devices corresponding to each sub-unit. After obtaining the communication characteristics of the sub-unit, the communication characteristics can be input into the pre-trained AI recognition model to identify the adapter that can adapt to the communication characteristics, so that the adapter can effectively and accurately control the data transmission and control of each sub-unit in the energy storage management system; the communication protocol and point table are obtained after detecting the communication mode of each sub-unit one by one through the communication converter;

[0085] The AI ​​recognition model is a machine learning model that analyzes the acquired communication features and identifies the adapter corresponding to the communication features from the database, thereby playing an efficient role in data transmission and control between various sub-units in the energy storage management system;

[0086] The training methods of AI recognition models include:

[0087] Pre-collecting multiple groups of communication characteristics and adaptation programs corresponding to the communication characteristics;

[0088] Mark each group of communication features as training features, annotate the adapter of each group of training features, and divide the annotated training features into a training set and a test set; use 70% of the training features as the training set and 30% of the training features as the test set; use the training set to train the AI ​​recognition model and use the test set to test the AI ​​recognition model; preset an error threshold, and when the mean of the prediction errors of all training features in the test set is less than the error threshold, output the AI ​​recognition model that identifies the adapter;

[0089] Exemplarily, the AI ​​recognition model adopts any one of a support vector machine model or a random forest model; the preset error threshold is pre-set according to the accuracy actually required by the AI ​​recognition model;

[0090] By inputting the acquired communication features into the trained AI recognition model, the corresponding adapter can be identified based on the communication features, thereby achieving the effect of learning and recognition using the AI ​​recognition model.

[0091] The scenario construction module constructs an initial management scenario with a control link and sends the adapter to the control point of the control link to generate an energy storage management scenario;

[0092] The initial management scenario refers to a blank scenario that can be generated based on the association of each sub-unit, so that the initial management scenario can serve as the basic scenario for data transmission and control of each sub-unit in the energy storage management system. The control link refers to the channel used to realize data transmission and control between each sub-unit in the initial management scenario, which can ensure that each sub-unit with data transmission and control requirements can maintain association and facilitate subsequent data transmission and control requirements.

[0093] The construction method of the initial management scenario includes:

[0094] Three units with a triangular structure are constructed through virtual reality technology, and the unit located at the vertex of the triangle is marked as the first unit, and the remaining two units are marked as the second unit and the third unit from left to right. The unit is used to provide a large range of import positions for the energy management system, battery management system and energy storage converter, and the unit of the triangular structure can ensure that any two units maintain a certain correlation, thereby improving the stability of subsequent scene construction.

[0095] Establish data frames on the first unit, the second unit and the third unit respectively, and import the energy management system, the battery management system and the energy storage converter into the data frames on the first unit, the second unit and the third unit one by one to generate a basic management scenario; the data frame is used to provide a small range limit for the import position of the energy management system, the battery management system and the energy storage converter to ensure that the positions of the energy management system, the battery management system and the energy storage converter in the scenario are fixed to each other;

[0096] In the basic management scenario, a data transfer point is established at the middle position of the first unit, the second unit and the third unit, which is recorded as a transfer point; the data transfer point refers to a point that can transfer and interact the data in the first unit, the second unit and the third unit, so as to achieve a convenient interaction effect of the data in the first unit, the second unit and the third unit;

[0097] A bidirectional control link is established between the transfer position and the first unit position, the second unit position and the third unit position one by one to generate an initial management scenario. The control link refers to a channel that can bidirectionally interact data between the transfer position, the first unit position, the second unit position and the third unit position, thereby ensuring the stable flow of data within the scenario.

[0098] It should be noted that the initial management scenario is only the basic outline of the subsequent energy storage management system. By configuring different adapters in the initial management scenario, the initial management scenario can be converted into a corresponding scenario that meets various data transmission and control requirements of the energy storage management system.

[0099] The energy storage management scenario is a scenario generated after configuring the corresponding adapter program on the control link based on the initial management scenario. It can provide a simulation basis for the data transmission and control requirements of each sub-unit in the energy storage management system, so that the energy storage management scenario can serve as the basis for the subsequent communication characteristics of the energy storage management system and the consistency detection of the adapter program;

[0100] The generation method of energy storage management scenario includes:

[0101] The connection end of the control link and the transit position is recorded as the starting end, and the connection end of the control link and the first unit position, the second unit position and the third unit position is recorded as the terminating end, and the starting ends and terminating ends of the three control links are marked one by one;

[0102] Mark all points on the control link, and record the middle point between the starting end and the ending end on the control link as the import point, and obtain three import points; the import point is used to define the basic position of the control point to ensure the stability of the subsequent control point position;

[0103] A front contour line is established on the side close to the starting end at the three import points, and a rear contour line is established on the side close to the ending end, and the front contour line, the import point and the rear contour line located on the same control link are assigned the same number to obtain three control points; the front contour line and the rear contour line are used to define the boundaries of the influence range before and after the control point, so as to close and define the influence range of the control point;

[0104] At the same time, the adapter program is sent to three control points to generate an energy storage management scenario.

[0105] It should be noted that the energy storage management scenario is the basis for detecting the compatibility between the communication characteristics and the adapter program on the energy storage management system, thereby providing a basis for detecting the normal operation reliability of the energy storage management system.

[0106] An index calculation module obtains comprehensive management parameters of the energy storage management scenario and calculates the management reliability index of the energy storage management scenario based on the comprehensive management parameters;

[0107] Comprehensive management parameters refer to data that can be used to calculate the reliability of communication characteristics and adaptation programs in energy storage management scenarios, and serve as a basis for subsequent analysis of whether the communication data links between various sub-units in the energy storage management system are fully connected and can be controlled as a whole, and as a basis for subsequent calculation of management reliability coefficients;

[0108] Comprehensive management parameters include the percentage of standard formats, data delay duration, and percentage of abnormal status;

[0109] The standard format ratio refers to the ratio of the number of data in the standard format to the total number of data in the energy storage management scenario, which can be used to indicate the effectiveness of the data conversion format in the energy storage management scenario. The larger the standard format ratio, the larger the ratio of the number of data in the standard format to the total number of data in the energy storage management scenario. At this time, the energy storage management system has a higher safety and reliability, and the management reliability index is higher.

[0110] Methods for obtaining the standard format ratio include:

[0111] The attribute status of the data is queried through the attribute management system, and the data with the attribute status of processed is recorded as valid data; the attribute status is used to indicate whether the data has been processed, and the attribute status includes processed status and unprocessed status;

[0112] The time when valid data first appears in the energy storage management scenario is queried through the timestamp and recorded as the starting time;

[0113] When no valid data appears within the preset time, the last time valid data appears is recorded as the end time, and the period from the start time to the end time is recorded as the packet capture period; the preset time is the maximum interval time for the appearance of valid data, so that the acquisition time period of valid data can be limited, thereby ensuring the accuracy of the acquisition of valid data;

[0114] Mark out the medium duration during the packet capture period capture time, and use the capture tool to Each packet capture moment captures an equal amount of valid data;

[0115] The data format of valid data is queried through the format management system, and valid data with the same data format as the standard format is recorded as standard data, and the number of standard data is counted;

[0116] Will The number of standard data at each capture time is compared with the number of valid data in turn to obtain The percentage of children;

[0117] The expression of the sub-ratio is:

[0118] ;

[0119] In the formula, For the The percentage of sub-packets at the time of packet capture, For the The number of standard data at each capture moment, is the number of valid data;

[0120] Will The percentages of each child are accumulated and averaged to obtain the percentage in standard format;

[0121] The expression of the standard format ratio is:

[0122] ;

[0123] In the formula, is the standard format ratio, For the The percentage of children.

[0124] Data delay time refers to the difference between the time of data transmission and processing in the energy storage management scenario and the standard time. It can be used to represent the performance of data transmission and processing in the energy storage management scenario. The longer the data delay time is, the greater the difference between the time of data transmission and processing in the energy storage management scenario and the standard time. At this time, the safety and reliability of the energy storage management system is lower, and the management reliability index is smaller.

[0125] Methods for obtaining data delay duration include:

[0126] At time T1, the p valid data are numbered in ascending order, and the numbered valid data are synchronously imported into the data processor;

[0127] Record the p valid data exported by the data processor one by one, and query the exporting time of the p valid data in order from small to large numbers to obtain p exporting times;

[0128] The duration between time T1 and p export times is recorded as the real-time processing duration, and the real-time processing duration greater than the standard processing duration is recorded as the delayed processing duration. The standard processing time is the time taken to process valid data once completely, and is used as the data basis for judging whether the real-time processing time is delayed. The standard processing time is obtained by collecting a large amount of valid data and calculating the average time taken to process the data once completely.

[0129] Will The delayed processing time is subtracted from the standard processing time in turn, and then The differences are accumulated one by one to obtain the data delay time;

[0130] The expression of data delay time is:

[0131] ;

[0132] In the formula, is the data delay time, For the Delayed processing time, This is the standard processing time.

[0133] The abnormal state ratio value refers to the ratio of the number of battery power supply states identified as abnormal to the total number in the energy storage management scenario, which can be used to indicate the stability of the battery operation state in the energy storage management scenario. The larger the abnormal state ratio value, the larger the ratio of the number of battery power supply states identified as abnormal to the total number in the energy storage management scenario. At this time, the safety reliability of the energy storage management system is lower, and the management reliability index is smaller.

[0134] Methods for obtaining the abnormal status ratio include:

[0135] A1: At each packet capture moment, the real-time current and temperature of the battery are detected by the current sensor and temperature sensor respectively, and the real-time current and temperature of the battery are obtained. The current value and temperature value, and The current value and the corresponding temperature value combinations, generating Power supply status;

[0136] A2: The power supply state in which the current value is greater than the standard current value and the temperature value is greater than the standard temperature value is recorded as an abnormal state, and the number of abnormal states is counted; the standard current value refers to the maximum value of the current when there is no battery and no abnormal phenomenon occurs, and the standard temperature value refers to the maximum value of the temperature when there is no battery and no abnormal phenomenon occurs, so as to serve as a numerical comparison basis for judging whether the current value and the temperature value are abnormal;

[0137] A3: After comparing the number of abnormal states with the number of power supply states, a sub-proportion value is obtained;

[0138] A4: Repeat Repeat steps A1-A3 to obtain The percentage of children;

[0139] The expression of the sub-ratio value is:

[0140] ;

[0141] In the formula, For the The percentage of children, For the The number of abnormal states, For the The number of power supply states;

[0142] Will The percentage values ​​of each child are accumulated and averaged to obtain the percentage value of abnormal state;

[0143] The expression of abnormal status ratio is:

[0144] ;

[0145] In the formula, is the abnormal status ratio, For the The percentage of children.

[0146] After the comprehensive management parameters are obtained, the management reliability index can be calculated based on the comprehensive management parameters, so that the management reliability index can numerically represent the reliability in the energy storage management scenario, and further represent the degree of adaptation between the communication characteristics and the adaptation program in the energy storage management system;

[0147] The expression of management reliability index is:

[0148] ;

[0149] In the formula, To manage the reliability index, , , is a proportional coefficient greater than 0;

[0150] in, , exemplary, is 0.36, is 0.35, is 0.29; , , The setting is to adjust the proportion of standard format, data delay duration and abnormal state proportion in the management reliability index, so that changes in the standard format proportion, data delay duration and abnormal state proportion can cause corresponding changes in the management reliability index.

[0151] A prompt determination module compares the management reliability index with a preset management reliability threshold, and determines whether to generate a re-identification prompt based on the comparison result;

[0152] After the management reliability index is calculated, it can be compared with the corresponding threshold value, and the reliability of the energy storage management scenario can be analyzed and judged based on the comparison result, and then the communication characteristics in the energy storage management system and the reliability of the adaptation operation of the adaptation program can be judged;

[0153] When the reliability of the adaptation operation of the communication characteristics in the energy storage management system and the adapter program does not meet the predetermined reliability requirements, it means that the energy storage management system is not highly adaptable to the adapter program, and the operation reliability between the identified adapter program and the energy storage management system is low. At this time, a re-identification prompt needs to be issued, thereby providing an instruction basis for the subsequent re-identification of the adapter program;

[0154] The determination method of whether to generate a re-identification prompt includes:

[0155] Management Reliability Index With the preset management reliability threshold Comparison; The preset management reliability threshold is used for the maximum value of the management reliability index corresponding to the issuance of the re-identification prompt, which can provide a comparison basis for whether to generate the re-identification prompt; The preset management reliability threshold is obtained by collecting a large number of maximum values ​​of the management reliability index corresponding to the issuance of the re-identification prompt and calculating the average value;

[0156] when Greater than or equal to , it indicates that the management reliability index is greater than or equal to the preset management reliability threshold. At this time, the reliability of the adaptation operation of the communication characteristics and the adaptation program in the energy storage management system can meet the predetermined reliability requirements, and it is determined that no re-identification prompt is generated;

[0157] when Less than When , it indicates that the management reliability index is less than the preset management reliability threshold. At this time, the reliability of the adaptation operation of the communication characteristics and the adaptation program in the energy storage management system does not meet the preset reliability requirements, and it is determined to generate a re-identification prompt.

[0158] A re-identification module identifies the optimization parameters from the comprehensive management parameters, formulates the re-identification instructions, and re-identifies the adaptation procedures;

[0159] The optimization parameter refers to a specific parameter in the comprehensive management parameter whose value is too large or too small, so that the optimization parameter can have a greater impact on the value of the management reliability index and serve as the basis for the subsequent formulation of re-identification instructions;

[0160] Methods for identifying optimization parameters include:

[0161] Compare the standard format ratio with the preset standard ratio; the preset standard ratio is used to represent the minimum value of the standard format ratio corresponding to the issuance of the re-identification prompt, so as to distinguish the size of the standard format ratio and serve as a basis for determining whether the standard format ratio is an optimization parameter;

[0162] When the standard format ratio is less than the preset standard ratio, the standard format ratio has a greater impact on the management reliability index, and the standard format ratio is identified as an optimization parameter;

[0163] Compare the data delay time with a preset standard delay; the preset standard delay is used to represent the maximum value of the data delay time corresponding to the issuance of a re-identification prompt, thereby distinguishing the size of the data delay time and serving as a basis for determining whether the data delay time is an optimization parameter;

[0164] When the data delay time is greater than the preset standard delay, the data delay time has a greater impact on the management reliability index, and the data delay time is identified as an optimization parameter;

[0165] Compare the abnormal state ratio with the preset standard ratio; the preset standard ratio is used to represent the maximum value of the abnormal state ratio corresponding to the issuance of the re-identification prompt, so as to distinguish the size of the abnormal state ratio and serve as a basis for determining whether the abnormal state ratio is an optimization parameter;

[0166] When the abnormal state proportion value is greater than the preset standard proportion value, the abnormal state proportion value has a greater impact on the management reliability index, and the abnormal state proportion value is identified as an optimization parameter.

[0167] The re-identification instruction is specific guidance information for the AI ​​recognition model to re-identify the adapter program, so as to ensure that the AI ​​recognition model can re-identify the adapter program in an orderly and accurate manner, thereby meeting the use requirements of the energy storage management system;

[0168] Re-identification instructions include priority identification instructions and common identification instructions; priority identification instructions mean that the measures for re-identification through the AI ​​identification model need to be executed immediately and cannot be waited; common identification instructions mean that the measures for re-identification through the AI ​​identification model do not need to be executed immediately and can be waited;

[0169] The method for formulating priority identification instructions and general identification instructions includes:

[0170] Count the number of optimization parameters and obtain the optimization value;

[0171] When the optimization value is 1, the number of parameters that need to be optimized is small, and the abnormality between the communication characteristics in the energy storage management system and the adapter is low, so there is no need to re-identify immediately, and a normal identification instruction is formulated;

[0172] When the optimization value is 2 or 3, the number of parameters that need to be optimized is large, and the abnormality between the communication characteristics and the adapter in the energy storage management system is high, requiring immediate re-identification, so a priority identification instruction is formulated.

[0173] When a priority recognition instruction is generated, the adapter needs to be re-recognized immediately through the AI ​​recognition model. When a normal recognition instruction is generated, the adapter needs to be re-recognized through the AI ​​recognition model to identify the latest adapter, and perform subsequent operations based on the identified adapter. Steps S2-S5 are executed here to achieve the management effect of the energy storage management system and ultimately achieve stable and efficient communication links between the various sub-units in the energy storage management system.

[0174] In this embodiment, by acquiring the communication characteristics of each subunit, the communication characteristics are input into the pre-trained AI recognition model, the adapter corresponding to the communication characteristics is identified, the initial management scenario with a control link is constructed, and the adapter is sent to the control point of the control link to generate an energy storage management scenario, obtain the comprehensive management parameters of the energy storage management scenario, and calculate the management reliability index of the energy storage management scenario based on the comprehensive management parameters, compare the management reliability index with the preset management reliability threshold, and determine whether to generate a re-identification prompt based on the comparison result, identify the optimization parameters from the comprehensive management parameters, formulate a re-identification instruction, and re-identify the adapter; Compared with the prior art, by acquiring the communication protocol and point table data of each subunit, and combining the AI ​​recognition model, it is possible to predict the communication protocol and point table data in a targeted manner. The adapter program corresponding to the point table is used to achieve the rapid adaptation effect of the data communication link of each sub-unit in the energy storage management system. By building an energy storage management scenario and collecting comprehensive management parameters to calculate the management reliability index, the adaptation operation reliability between the identified adapter program and the sub-unit can be accurately analyzed and calculated, and the adaptation degree of the data transmission link between each sub-unit in the energy storage management system can be accurately obtained. At the same time, when it is found that the adaptation degree is not good, the adapter program can be re-identified as needed to ensure that the data link and control chain conduction adaptation effect can be achieved quickly and accurately between the sub-units of the energy storage management system, effectively avoiding the problems of slow adaptation speed and low accuracy of data link and control chain conduction caused by unit upgrades or version updates of each sub-unit in the integrated product, and effectively improving the operation efficiency of the energy storage management system.

[0175] Example 2: Please refer to Figure 2 As shown, the part not described in detail in this embodiment is described in the first embodiment, and a method for energy storage management based on an AI recognition model is provided, which is applied to an AI server and implemented based on an energy storage management system based on an AI recognition model, including:

[0176] S1: Obtain the communication features of each subunit, input the communication features into the pre-trained AI recognition model, and identify the adapter corresponding to the communication features. The communication features include the communication protocol and the point table.

[0177] S2: construct an initial management scenario with a control link, and send the adapter to the control point of the control link to generate an energy storage management scenario;

[0178] S3: Obtain comprehensive management parameters of the energy storage management scenario, and calculate the management reliability index of the energy storage management scenario based on the comprehensive management parameters. The comprehensive management parameters include the standard format ratio, data delay duration, and abnormal state ratio.

[0179] S4: comparing the management reliability index with a preset management reliability threshold, and determining whether to generate a re-identification prompt according to the comparison result; if a re-identification prompt is generated, executing S5; if no re-identification prompt is generated, repeating S3-S4;

[0180] S5: Identify optimization parameters from the comprehensive management parameters, formulate re-identification instructions, and re-identify the adaptation program.

[0181] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An energy storage management system based on AI recognition model, applied to AI server, characterized in that: include: The model recognition module is used to obtain the communication characteristics of each sub-unit, input the communication characteristics into the pre-trained AI recognition model, and identify the adapter corresponding to the communication characteristics. The communication characteristics include the communication protocol and the point table; A scenario construction module is used to construct an initial management scenario with a control link and send the adapter to the control point of the control link to generate an energy storage management scenario; An index calculation module is used to obtain comprehensive management parameters of the energy storage management scenario and calculate the management reliability index of the energy storage management scenario based on the comprehensive management parameters. The comprehensive management parameters include the standard format ratio, data delay duration, and abnormal state ratio. A prompt determination module, used for comparing the management reliability index with a preset management reliability threshold, and determining whether to generate a re-identification prompt according to the comparison result; The re-identification module is used to identify the optimization parameters from the comprehensive management parameters, formulate the re-identification instructions, and re-identify the adaptation program.

2. The energy storage management system based on the AI ​​recognition model according to claim 1 is characterized in that: The training method of the AI ​​recognition model includes: Pre-collecting multiple groups of communication characteristics and adaptation programs corresponding to the communication characteristics; Mark each group of communication features as training features, annotate the adapter program for each group of training features, and divide the annotated training features into a training set and a test set; use the training set to train the AI ​​recognition model, and use the test set to test the AI ​​recognition model; preset an error threshold, and when the mean of the prediction errors of all training features in the test set is less than the error threshold, output the AI ​​recognition model that identifies the adapter program.

3. The energy storage management system based on the AI ​​recognition model according to claim 2 is characterized in that: The method for constructing the initial management scenario includes: Three units with a triangular structure are constructed by virtual reality technology, and the unit located at the vertex of the triangle is marked as the first unit, and the remaining two units are marked as the second unit and the third unit from left to right; Establish data frames on the first unit, the second unit and the third unit respectively, and import the energy management system, the battery management system and the energy storage converter into the data frames on the first unit, the second unit and the third unit one by one to generate a basic management scenario; In the basic management scenario, a data transfer point is established at a middle position between the first unit position, the second unit position and the third unit position, which is recorded as a transfer point; Two-way conductive control links are established one by one between the transfer position and the first unit position, the second unit position and the third unit position to generate an initial management scenario.

4. The energy storage management system based on the AI ​​recognition model according to claim 3 is characterized in that: The method for generating the energy storage management scenario includes: The connection end of the control link and the transit position is recorded as the starting end, and the connection end of the control link and the first unit position, the second unit position and the third unit position is recorded as the terminating end, and the starting ends and terminating ends of the three control links are marked one by one; Mark all points on the control link, and record the middle point between the starting end and the ending end on the control link as the import point, and obtain three import points; A front contour line is established on the side close to the starting end at the three lead-in points, and a rear contour line is established on the side close to the ending end, and the front contour line, lead-in point and rear contour line located on the same control link are assigned the same number to obtain three control points; At the same time, the adapter program is sent to three control points to generate an energy storage management scenario.

5. The energy storage management system based on the AI ​​recognition model according to claim 4 is characterized in that: The method for obtaining the standard format ratio includes: Query the attribute status of the data through the attribute management system, and record the data with the attribute status being processed as valid data; The time when valid data first appears in the energy storage management scenario is queried through the timestamp and recorded as the starting time; When no valid data appears within the preset time, the last time valid data appears is recorded as the end time, and the period from the start time to the end time is recorded as the packet capture period; Mark out the medium duration during the packet capture period capture time, and use the capture tool to Each packet capture moment captures an equal amount of valid data; The data format of valid data is queried through the format management system, and valid data with the same data format as the standard format is recorded as standard data, and the number of standard data is counted; Will The number of standard data at each packet capture moment is compared with the number of valid data in turn to obtain The percentage of children; The expression of the sub-ratio is: ； In the formula, For the The percentage of sub-packets at the time of packet capture, For the The number of standard data at each capture moment, is the number of valid data; Will The percentages of each child are accumulated and averaged to obtain the percentage in standard format; The expression of the standard format ratio is: ； In the formula, is the standard format ratio, For the The percentage of children.

6. The energy storage management system based on the AI ​​recognition model according to claim 5, characterized in that: The method for obtaining the data delay duration includes: At time T1, the p valid data are numbered in ascending order, and the numbered valid data are synchronously imported into the data processor; Record the p valid data exported by the data processor one by one, and query the exporting time of the p valid data in order from small to large numbers to obtain p exporting times; The duration between time T1 and p export times is recorded as the real-time processing duration, and the real-time processing duration greater than the standard processing duration is recorded as the delayed processing duration. Delayed processing time; Will The delayed processing time is subtracted from the standard processing time in turn, and then The differences are accumulated one by one to obtain the data delay time; The expression of data delay time is: ； In the formula, is the data delay time, For the Delayed processing time, This is the standard processing time.

7. The energy storage management system based on the AI ​​recognition model according to claim 6 is characterized in that: The method for obtaining the abnormal state ratio value includes: A1: At each packet capture moment, the real-time current and real-time temperature of the battery are detected by the current sensor and the temperature sensor respectively, and the The current value and temperature value, and The current value and the corresponding temperature value combinations, generating Power supply status; A2: The power supply state in which the current value is greater than the standard current value and the temperature value is greater than the standard temperature value is recorded as an abnormal state, and the number of abnormal states is counted; A3: After comparing the number of abnormal states with the number of power supply states, a sub-proportion value is obtained; A4: Repeat Repeat steps A1-A3 to obtain The percentage of children; The expression of the sub-ratio value is: ； In the formula, For the The percentage of children, For the The number of abnormal states, For the The number of power supply states; Will The percentage values ​​of each child are accumulated and averaged to obtain the percentage value of abnormal state; The expression of abnormal status ratio is: ； In the formula, is the abnormal status ratio, For the The percentage of children.

8. The energy storage management system based on the AI ​​recognition model according to claim 7 is characterized in that: The expression of the management reliability index is: ； In the formula, To manage the reliability index, , , is a proportional coefficient greater than 0; The determination method of whether to generate a re-identification prompt includes: Management Reliability Index With the preset management reliability threshold Compare; when Greater than or equal to When , it is determined that no re-identification prompt is generated; when Less than , it is determined that a re-identification prompt is generated.

9. The energy storage management system based on the AI ​​recognition model according to claim 8, characterized in that: The identification method of the optimization parameters includes: The standard format ratio is compared with a preset standard ratio, and when the standard format ratio is less than the preset standard ratio, the standard format ratio is identified as an optimization parameter; Comparing the data delay time with a preset standard delay, and identifying the data delay time as an optimization parameter when the data delay time is greater than the preset standard delay; The abnormal state proportion value is compared with a preset standard proportion value, and when the abnormal state proportion value is greater than the preset standard proportion value, the abnormal state proportion value is identified as an optimization parameter.

10. The energy storage management system based on AI recognition model according to claim 9, characterized in that: The re-identification instruction includes a priority identification instruction and a common identification instruction; The method for formulating priority identification instructions and general identification instructions includes: Count the number of optimization parameters and obtain the optimization value; When the optimization value is 1, a normal recognition instruction is formulated; When the optimization value is 2 or 3, a priority recognition instruction is formulated.

Images