Cross-platform barrier-free communication method and system

Abstract

The invention discloses a cross-platform barrier-free communication method and system, and relates to the technical field of cross-platform communication, and the scheme comprises the steps: obtaining the electric quantity, load power, network bandwidth and delay data of energy storage equipment in real time, and generating a path load value through normalization and weighted fusion; predicting a future load through the LSTM model, and triggering early warning when the predicted load is insufficient and the delay exceeds a threshold; when the electric quantity is insufficient and in an early warning state, a protocol switching instruction and a dynamic scheduling table are automatically generated; a path quality evaluation matrix is established, a load pressure factor is calculated, an over-limit path is eliminated, and a scheduling table is updated based on an optimized path and issued to equipment. Through dynamic coupling of load prediction and path quality evaluation, real-time optimization and self-healing of a communication link are realized; and protocol switching and path reforming are automatically triggered during low-power and high-load early warning of the equipment, so that efficient and reliable barrier-free communication between the distributed energy storage equipment is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cross-platform communication, and more particularly, to a cross-platform barrier-free communication method and system. BACKGROUND

[0002] With the continuous development of the energy internet and smart grid, cross-platform barrier-free communication in the distributed energy storage device collaborative management of the energy grid is increasingly important. The distributed energy storage device communication system needs to realize efficient data exchange in a multi-protocol, multi-network environment to ensure device state monitoring, load balancing, and grid stable operation. If the communication process cannot be dynamically optimized and adaptively adjusted, it may lead to communication delay, data loss, or device collaboration failure, affecting the overall efficiency and reliability of the energy grid. Therefore, the optimization of the cross-platform barrier-free communication method has become a problem to be solved in the field of energy management.

[0003] Existing cross-platform barrier-free communication solutions mostly rely on fixed communication protocols or predefined routing strategies. The approach is to transmit data according to device basic state and network performance parameters through pre-set communication paths and protocol rules. The advantage of such technology is that it is simple to implement and has lower deployment costs, and can meet the needs of regular communication, such as providing basic connection services in a stable network environment.

[0004] However, the existing communication method is easily limited by static configuration in the process of path selection and protocol switching, and lacks real-time data-driven optimization, resulting in uneven load of communication paths, high transmission delay, and low communication efficiency. In terms of dynamically adapting to network fluctuations and device state changes, the existing technology often makes little use of historical load data and network performance data, resulting in its inability to intelligently adjust according to real-time predictions. This manifests itself in poor path selection and untimely protocol switching during communication, which can easily lead to communication interruption and data loss, affecting the collaborative performance of distributed energy storage devices. Therefore, the cross-platform barrier-free communication method and system are proposed to solve this problem. SUMMARY

[0005] To solve the above technical problems, the cross-platform barrier-free communication method and system are provided, which solve the problems raised in the background technology.

[0006] In order to achieve the above purpose, the technical scheme of the present application is as follows:

[0007] In a first aspect, the present application provides a cross-platform barrier-free communication method applied to a distributed energy storage device communication system in an energy grid, comprising:

[0008] The device state data, network performance data and historical load data of the distributed energy storage device are acquired in real time through an Internet of Things chip, the device state data includes device identification, power value and load power, and the network performance data includes bandwidth usage value and transmission delay value;

[0009] The power value and the load power are normalized, and the standardized data is generated by associating the device identification, the standardized data is compressed, and the processed load power and the bandwidth usage value in the network performance data are fused by weighting to generate a path load value;

[0010] The historical load data is input into a pre-trained LSTM model to output a predicted load value in a future period, and when the predicted load value is less than the path load value and the transmission delay value is greater than a delay threshold, a load warning state is triggered;

[0011] When the power value is less than a preset power threshold when in the load warning state, a protocol switching instruction is generated, and a dynamic scheduling table including device communication timing and communication paths is generated based on the protocol switching instruction;

[0012] A communication path quality evaluation matrix is established based on the coupling relationship between the predicted load value and the path load value, the difference between the predicted load value and the path load value in the communication path quality evaluation matrix is calculated as a load stress factor, and the communication path in the communication path quality evaluation matrix whose load stress factor exceeds a stress threshold is removed;

[0013] An optimized communication path quality evaluation matrix is constructed based on the removed communication path, an optimized dynamic scheduling table is constructed based on the optimized communication path quality evaluation matrix, and the optimized dynamic scheduling table is output to the corresponding distributed energy storage device.

[0014] In a second aspect, the application provides a cross-platform barrier-free communication system applied to a distributed energy storage device communication system in an energy grid, for implementing the cross-platform barrier-free communication method described in any of the above aspects, comprising:

[0015] A data acquisition module is configured to acquire device state data, network performance data and historical load data of the distributed energy storage device in real time, the device state data includes device identification, power value and load power, and the network performance data includes bandwidth usage value and transmission delay value;

[0016] A path load calculation module is configured to normalize the power value and the load power, generate standardized data by associating the device identification, compress the standardized data, and fuse the processed load power and the bandwidth usage value in the network performance data by weighting to generate a path load value;

[0017] a load early warning module configured to input historical load data into a pre-trained LSTM model, output a predicted load value of a future time period, and trigger a load early warning state when the predicted load value is less than a path load value and a transmission delay value is greater than a delay threshold value;

[0018] a schedule generation module configured to generate a protocol switching instruction when in the load early warning state and the power value is less than a preset power threshold value, and generate a dynamic schedule table including device communication timing and a communication path based on the protocol switching instruction;

[0019] a path optimization module configured to establish a communication path quality evaluation matrix based on a coupling relationship between the predicted load value and the path load value, calculate a difference between the predicted load value and the path load value in the communication path quality evaluation matrix as a load stress factor, and eliminate a communication path in the communication path quality evaluation matrix whose load stress factor exceeds a stress threshold value;

[0020] a schedule optimization module configured to construct an optimized communication path quality evaluation matrix based on the eliminated communication path, construct an optimized dynamic schedule table based on the optimized communication path quality evaluation matrix, and output the optimized dynamic schedule table to a corresponding distributed energy storage device.

[0021] In a third aspect, the present application provides a computer device, which comprises a memory and a processor, the memory stores a code, and the processor is configured to acquire the code and execute the cross-platform barrier-free communication method described above.

[0022] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the cross-platform barrier-free communication method described above.

[0023] Compared with the prior art, the present application has the following beneficial effects:

[0024] The present application solves the problem of poor adaptability of static or fixed routing strategies in a dynamic network environment by establishing a communication path quality evaluation matrix and calculating a load stress factor, dynamically eliminating high load stress paths, and realizes real-time optimization and self-healing of communication paths, ensuring high availability and stability of cross-platform communication links.

[0025] The present application constructs and outputs a dynamic schedule table based on an optimized communication path quality evaluation matrix, automatically triggers protocol switching and path reorganization when the device power is insufficient and faces high load early warning, solves the problem of uneven resource allocation in a multi-protocol heterogeneous network, realizes intelligent matching of communication resources and device states, and ultimately achieves efficient and reliable barrier-free communication between distributed energy storage devices in an energy grid. BRIEF DESCRIPTION OF DRAWINGS

[0026] The disclosure of the present application is explained with reference to the accompanying drawings. It should be appreciated that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of the present application. Among them:

[0027] Figure 1 Flow chart of the cross-platform barrier-free communication method proposed in the present application;

[0028] Figure 2 Flow chart of the method for periodically updating the LSTM model in the present application;

[0029] Figure 3 Data flow chart of the cross-platform barrier-free communication method in the present application;

[0030] Figure 4 Structural block diagram of the cross-platform barrier-free communication system proposed in the present application. DETAILED DESCRIPTION

[0031] It is easy to understand that according to the technical solution of the present application, those skilled in the art can propose various structures and implementation manners which can be replaced with each other without changing the essential spirit of the present application. Therefore, the following specific embodiments and drawings are only exemplary descriptions of the technical solution of the present application, and should not be regarded as the whole or as the limitation or restriction of the technical solution of the present application.

[0032] In the prior art, the cross-platform barrier-free communication method has the problems of path selection static, protocol switching lag and lack of real-time data driven optimization. It usually depends on fixed communication protocol or pre-defined routing strategy, once the network performance fluctuates or the device state changes, the system cannot dynamically adjust the communication path and protocol, still highly depends on the pre-set configuration. Such rigidity and passivity not only reduces the communication efficiency, but also may cause data loss or device coordination failure due to uneven path load and high transmission delay, affecting the stable operation of the energy grid.

[0033] In order to solve the above problems, the application generates a path load value by real-time acquisition of equipment state data, network performance data and historical load data, normalization processing and weighted fusion; uses a pre-trained LSTM model to output a predicted load value, and triggers a load warning state when the predicted load value is less than the path load value and the transmission delay value exceeds the threshold; generates a protocol switching instruction and a dynamic scheduling table when in the load warning state and the power value is lower than the preset threshold; establishes a communication path quality evaluation matrix based on the coupling relationship between the predicted load value and the path load value, calculates the load pressure factor, eliminates the communication path with the load pressure factor exceeding the threshold, and constructs an optimized dynamic scheduling table and outputs it to the distributed energy storage device. Through dynamic evaluation and iterative optimization of the communication path, the adaptability and reliability of the communication process are ensured, the problems of uneven path load and untimely protocol switching under static configuration are solved, efficient operation of cross-platform barrier-free communication is realized, and the overall collaborative performance and stability of the energy grid are improved.

[0034] Embodiment 1

[0035] As shown in Figure 1 , a cross-platform barrier-free communication method is provided, applied to a distributed energy storage device communication system in an energy grid, comprising:

[0036] S1, real-time acquisition of equipment state data, network performance data and historical load data of the distributed energy storage device, the equipment state data including device identification, power value and load power, the network performance data including bandwidth usage value and transmission delay value;

[0037] S2, normalizing the power value and the load power, and generating standardized data by associating the device identification, data compression processing the standardized data, and weighted fusion of the processed load power and the bandwidth usage value in the network performance data to generate a path load value;

[0038] S3, inputting the historical load data into a pre-trained LSTM model to output a predicted load value in a future period, and triggering a load warning state when the predicted load value is less than the path load value and the transmission delay value is greater than the delay threshold;

[0039] S4, when in the load warning state and the power value is less than the preset power threshold, generating a protocol switching instruction, and generating a dynamic scheduling table containing device communication timing and communication path based on the protocol switching instruction;

[0040] S5, establishing a communication path quality evaluation matrix based on the coupling relationship between the predicted load value and the path load value, calculating the difference between the predicted load value and the path load value in the communication path quality evaluation matrix as a load pressure factor, and eliminating the communication path with the load pressure factor exceeding the pressure threshold in the communication path quality evaluation matrix;

[0041] S6, constructing an optimized communication path quality evaluation matrix based on the pruned communication path, and constructing an optimized dynamic scheduling table based on the optimized communication path quality evaluation matrix, and outputting the optimized dynamic scheduling table to the corresponding distributed energy storage device.

[0042] Regarding step S1:

[0043] Real-time collection of multiple types of data from distributed energy storage devices in the energy grid. Specifically, the device state data includes device identification, power value and load power: the device identification is the coding or address information for uniquely identifying each distributed energy storage device; the power value represents the real-time remaining power storage of the device, usually expressed in percentage or absolute power unit; the load power represents the power load borne by the device when it is currently running, measured in power unit (such as watt). Network performance data includes bandwidth usage value and transmission delay value: the bandwidth usage value represents the actual occupied network bandwidth resource in the current communication link, expressed in data transmission rate unit (such as Mbps); the transmission delay value represents the round-trip time of data packets in the communication path, measured in time unit (such as millisecond). Historical load data includes the load power record of the device in the past predetermined time period, for subsequent load prediction analysis. Data acquisition is realized through built-in sensors, network monitoring interfaces or communication protocols (such as Modbus, MQTT) in real time, and the timeliness and integrity of the data are ensured.

[0044] The collected raw data is preliminarily processed to ensure data quality and consistency. First, the device state data and network performance data are unified in format and checked to eliminate data format conflicts caused by device heterogeneity or communication protocol differences. For example, the device identification is mapped to a unified coding format, the power value and the load power are converted to standard units, and the bandwidth usage value and the transmission delay value are filtered for abnormal values (such as out-of-range or invalid data). The historical load data is aligned and archived according to time sequence to ensure that its timestamp matches that of the real-time data. This processing process does not involve data transformation or fusion, providing standardized input for subsequent steps.

[0045] The processed data is used for subsequent calculation and analysis. Specifically, the device state data (including device identification, power value and load power) and network performance data (including bandwidth usage value and transmission delay value) are transmitted to the path load calculation module to generate path load value; the historical load data is transmitted to the load warning module as the input of the pre-trained LSTM model for predicting future load value. Data output is realized through internal message queue or data bus to ensure low delay and high reliability, thereby supporting real-time execution of cross-platform barrier-free communication methods.

[0046] Further, the acquired device state data and network performance data are time-synchronized and integrity-verified, including:

[0047] Timestamp information is added to the obtained device state data, network performance data and historical load data;

[0048] Timestamp information is added to the obtained device state data and network performance data;

[0049] Based on the timestamp information, the device state data and the network performance data are time-synchronized and verified, and when the time difference is greater than a preset synchronization threshold, the data is reacquired;

[0050] The device state data and the network performance data after time synchronization verification are subjected to integrity verification to check whether the data is missing or has abnormal values;

[0051] When missing data is found, linear interpolation is used to complete the missing data based on historical device state data and historical network performance data;

[0052] When an abnormal value is found, the abnormal value is removed and the data of the abnormal value is reacquired.

[0053] Through the above technical solutions, the present application realizes effective collaborative management and quality control of multi-source heterogeneous data. Through the timestamp addition and synchronization verification mechanism, the time deviation caused by device heterogeneity and network fluctuation is eliminated, providing time sequence consistency guarantee for subsequent fusion calculation. Through the integrity verification and abnormal data processing mechanism, a complete and reliable data foundation is constructed, improving the accuracy of load prediction and path evaluation. Under the synergistic action of linear interpolation completion and abnormal data reacquisition, the continuity of the data sequence is guaranteed, and the interference noise is effectively filtered, enhancing the robustness and data adaptive ability of the system in complex network environment. In addition, the dual verification mechanism of time synchronization and integrity verification provides a high-quality data input basis for cross-platform barrier-free communication, ensuring the reliability of system decision and the precision of communication scheduling.

[0054] Regarding step S2:

[0055] The real-time data of step S1 is obtained:

[0056] The device state data specifically includes power value (indicating the real-time residual power storage of the energy storage device) and load power (indicating the power load when the device is currently running), as well as the device identifier associated with the data.

[0057] The network performance data specifically includes bandwidth usage value (indicating the data transmission rate occupation of the current communication link).

[0058] The obtained original multi-source heterogeneous data is converted into a unified path load evaluation index:

[0059] (1) Normalization processing and standardized data generation: The electric quantity values and load powers of different devices with different dimensions are normalized to map their values uniformly into a preset same value range (for example, the interval [0, 1]) to eliminate the dimension influence and facilitate subsequent comprehensive comparison and calculation. Then, the normalized electric quantity values and load powers are associated and bound with the unique device identifier to generate the standardized data of each record.

[0060] (2) Data compression processing: The generated standardized data is subjected to data compression processing. This process aims to reduce data volume, reduce storage and transmission overhead, and retain key feature information of the data to adapt to the resource-limited distributed device communication environment.

[0061] (3) Weighted fusion to generate path load value: The load power data after compression processing and the bandwidth usage value in the network performance data are subjected to weighted fusion calculation. Specifically, the processed load power and bandwidth usage value are respectively assigned a preset weight coefficient, and a comprehensive path load value is calculated by weighted summation or similar fusion algorithm. The value comprehensively reflects the communication pressure level on a specific communication path determined by the device load and network bandwidth condition.

[0062] After the above processing, a key intermediate indicator, the path load value, is output. The value is output to the subsequent modules of the system, especially the load warning module, as the quantitative evaluation result of the current communication path load condition, and is used to compare with the predicted load value and transmission delay value output by the LSTM model to determine whether to trigger the load warning state.

[0063] Further, the process of calculating the path load value specifically includes:

[0064] obtaining the processed load power , bandwidth usage value , historical load power and , and historical bandwidth usage value and ;

[0065] calculating the load power change rate :

[0066] ;

[0067] wherein represents the load power value at the previous sampling time, represents the load power value at the previous two sampling times, represents a preset fixed sampling time interval;

[0068] calculating the bandwidth usage value change rate :

[0069] ;

[0070] wherein, denotes the bandwidth usage value at the previous sampling time, denotes the bandwidth usage value at the previous two sampling times;

[0071] calculating the weight coefficient of the load power :

[0072] ;

[0073] wherein, denotes the absolute value of the load power change rate, denotes the absolute value of the bandwidth usage value change rate;

[0074] calculating the weight coefficient of the bandwidth usage value :

[0075] ;

[0076] calculating the path load value :

[0077] ;

[0078] Through the above technical solutions, the present application realizes accurate quantification and dynamic evaluation of the communication path load. Through the change rate calculation based on historical data, a dynamic weight distribution mechanism of the load power and the bandwidth usage value is established, so that the path load value can reflect the change trend of the network state in real time. Through the cooperative processing of absolute value operation and normalized weight, not only the interference of change direction on the evaluation result is eliminated, but also the rationality and comparability of weight distribution are ensured. Through the adaptive weight adjustment driven by the change rate, the system can automatically optimize the evaluation strategy according to the dynamic change of the network load characteristics, improving the accuracy and real-time performance of the path load evaluation. In addition, the weight distribution mechanism based on historical changes enhances the adaptability to sudden load fluctuations, providing more accurate and reliable basis for subsequent communication scheduling decisions, thereby improving the intelligent level and operation efficiency of the entire communication system.

[0079] Regarding step S3:

[0080] The following data is obtained:

[0081] Historical load data: refers to the load power history records of the distributed energy storage equipment at a series of consecutive time points in the past collected and stored by the data acquisition module. This data constitutes a time series for model prediction.

[0082] Real-time computing data: including the current path load value generated by the path load computing module, and the transmission delay value collected by the data acquisition module in real time.

[0083] (1) Load prediction: input the obtained historical load data (i.e. historical load power time series) into a pre-trained LSTM (Long Short-Term Memory) model. The model outputs a predicted load value corresponding to a future specific period by learning and analyzing the time series dependence in the historical data. The predicted load value is an estimate of the future time device communication load level.

[0084] (2) Warning state judgment: the system compares the predicted load value output by the model with the real-time acquired path load value, and compares the real-time acquired transmission delay value with a preset delay threshold. When the following two conditions are met at the same time, the system triggers the load warning state:

[0085] Condition one: the predicted load value is less than the path load value. This indicates that the future load capacity predicted based on historical trends will not be enough to support the comprehensive load pressure borne by the current path.

[0086] Condition two: the transmission delay value is greater than the delay threshold. This indicates that the current network transmission performance has deteriorated to an unacceptable level.

[0087] When the above two conditions are met, it is determined that the communication link is at risk, and the warning is triggered.

[0088] Output the load warning state result. The state is a flag signal, which indicates that the system is currently in the "load warning state" when it is triggered. This state signal will be output to the subsequent scheduling generation module as one of the conditions for determining whether to generate a protocol switching instruction and a dynamic scheduling table.

[0089] Further, as shown in Figure 2 , the LSTM model is periodically updated, specifically including:

[0090] Periodically obtain the latest historical load data in the calibration time window to retrain the LSTM model;

[0091] Obtain the data confidence of the retrained LSTM model in outputting the predicted load value. When the data confidence is less than the preset confidence threshold, retrain the LSTM model until the data confidence output by the retrained LSTM model is greater than or equal to the confidence threshold.

[0092] Further, during model training, the mean of the historical load data is calculated using the moving average method as the current predicted load value.

[0093] Further, in combination with the above content, the specific training process of the LSTM model includes:

[0094] Receive historical load data, perform time synchronization and integrity verification. Specifically, add timestamp information to historical load data, perform time synchronization check (reacquire data when time difference is greater than preset synchronization threshold), and perform integrity check (check for missing data or outliers). When missing data is found, linear interpolation is used to complete the missing data based on historical load data; when outliers are found, the data is discarded and reacquired. Clean and complete time series data is output for model training.

[0095] Receive preprocessed historical load data, and learn long-term dependencies in time series through an LSTM network structure. During training, model parameters are adjusted through backpropagation and optimization algorithms (such as gradient descent) to minimize prediction error. The output is a trained LSTM model that can output predicted load values for future time periods based on input historical load data. In addition, during initial training or retraining, the moving average method can be used to calculate the mean of historical load data as an initial reference for predicted load values, ensuring the stability of the model output.

[0096] Receive the trained LSTM model, periodically acquire the latest historical load data within the calibration time window for retraining to adapt to the dynamically changing network environment. After retraining, evaluate the data confidence of the model output predicted load values (for example, by calculating the error distribution of predicted values and actual values or using statistical methods). When the data confidence is less than the preset confidence threshold, retrain the LSTM model until the output confidence is greater than or equal to the confidence threshold. The output is a verified LSTM model with confidence that meets the requirements for real-time load prediction.

[0097] Through the above technical solutions, the present application realizes the continuous optimization and adaptive maintenance of the prediction model. By periodically retraining the LSTM model based on the latest historical data, it ensures that the load prediction capability can adapt to the dynamically changing network environment. Through the confidence evaluation and threshold comparison mechanism, an effective monitoring system for prediction quality is established, and the retraining process is started in time when the model confidence is insufficient. By introducing the moving average method as a backup prediction scheme, a dual protection mechanism is established, which not only maintains the continuity of the prediction output, but also ensures the stable operation of the system during model updating. In addition, the synergistic effect of model periodic updating and confidence verification improves the accuracy and reliability of load prediction, providing more accurate data support for subsequent communication scheduling decisions, thereby enhancing the adaptive ability and long-term stability of the entire communication system.

[0098] Further, when the load warning state is triggered, a cooperative verification of the warning credibility is also included:

[0099] When a single communication path triggers a load warning, obtain the transmission delay value and path load value of its N adjacent communication paths in the topology structure;

[0100] If more than a preset number of adjacent paths are also in a load warning state, a path reform instruction is triggered;

[0101] If more than a preset number of adjacent paths are normal, an optimization instruction for the single communication path is triggered.

[0102] For example, in a regional energy grid, four distributed energy storage devices (Device A, Device B, Device C, and Device D) are deployed and connected to form a mesh topology through multiple communication paths. Data exchange between devices is carried out through heterogeneous protocols such as MQTT and CoAP. Due to network fluctuations or device state changes, a single path may trigger a load warning, but the credibility of the warning needs to be confirmed through collaborative verification to avoid misjudgment and optimize resource allocation.

[0103] Assumptions: Preset number threshold: 2 (i.e., at least 2 adjacent paths are in a warning state to trigger path reform); Delay threshold: 40ms (transmission delay exceeding this value is considered network performance deterioration); Path load threshold: 0.2 (normalized value, used to determine if the path load is too high); Topology structure: Communication paths are connected with devices as nodes, and adjacent paths are defined as directly connected paths (e.g., path A-B, path A-C, and path A-D are adjacent paths of device A).

[0104] (1) Single path triggers load warning:

[0105] The communication path of device A (path A) triggers a load warning. The triggering condition is that the predicted load value output by the LSTM model (0.18) is less than the path load value (0.25), and the transmission delay value (45ms) is greater than the delay threshold (40ms).

[0106] The system records that path A is in a load warning state.

[0107] (2) Obtain adjacent path data:

[0108] The system identifies the adjacent paths of path A as path B (connecting device A and device B), path C (connecting device A and device C), and path D (connecting device A and device D) according to the topology structure.

[0109] Real-time acquisition of transmission delay values and path load values of these adjacent paths:

[0110] Path B: Transmission delay value = 42ms, path load value = 0.22;

[0111] Path C: Transmission Delay Value = 38ms, Path Load Value = 0.15;

[0112] Path D: Transmission Delay Value = 50ms, Path Load Value = 0.28.

[0113] (3) Assessing Adjacent Path Warning Status:

[0114] The system checks whether each adjacent path is in a load warning state (i.e., predicted load value is less than path load value and transmission delay value is greater than delay threshold):

[0115] Path B: Predicted Load Value = 0.20, Path Load Value = 0.22 (predicted value < path value), Transmission Delay Value = 42ms (> 40ms), triggering a warning.

[0116] Path C: Predicted Load Value = 0.18, Path Load Value = 0.15 (predicted value > path value, not meeting the condition), Transmission Delay Value = 38ms (≤ 40ms), no warning triggered.

[0117] Path D: Predicted Load Value = 0.25, Path Load Value = 0.28 (predicted value < path value), Transmission Delay Value = 50ms (> 40ms), triggering a warning.

[0118] Count the number of adjacent paths in the warning state: Path B and Path D trigger warnings, totaling 2.

[0119] (4) Collaborative Verification Decision:

[0120] The system compares the number of adjacent paths in warning with the preset number threshold (2):

[0121] Since the number of adjacent paths in warning (2) is equal to the preset number threshold (2), the system determines that "adjacent paths greater than the preset number threshold are also in a load warning state," thus triggering a path reorganization instruction.

[0122] The path reorganization instruction initiates a global path optimization process, re-evaluates all available paths, generates a new communication topology and scheduling table, and alleviates regional network congestion.

[0123] If the number of adjacent paths in warning is less than the preset threshold (e.g., only 1 warning), the system triggers an optimization instruction for a single communication path, making local adjustments (such as adjusting communication timing or switching protocols) only to Path A.

[0124] After the system triggers the path reorganization instruction, it recalculates the load pressure factor based on the communication path quality evaluation matrix, eliminates high-load paths (such as Path D), and generates an optimized dynamic scheduling table. The scheduling table is distributed to all devices to ensure the redistribution of communication resources and improve overall network stability.

[0125] Regarding step S4:

[0126] The following data is obtained:

[0127] System warning state: the load warning state output by the load warning module.

[0128] Device status data: the real-time power value collected by the data acquisition module, i.e. the real-time residual power storage of the distributed energy storage device.

[0129] Pre-set parameters: pre-stored pre-set power threshold in the system, which is used to determine whether the device is in a low power running state.

[0130] (1) Condition judgment and protocol switching instruction generation: the system makes logical judgments on the obtained state information. When the following two conditions are met at the same time, the system automatically generates a protocol switching instruction:

[0131] Condition one: the system is currently in a load warning state.

[0132] Condition two: the real-time power value of the device is less than the pre-set power threshold set by the system.

[0133] This protocol switching instruction aims to instruct the communication system to switch from the currently used communication protocol to another alternative protocol that is more suitable for the current low power, high load complex network environment, in order to optimize communication efficiency and reliability.

[0134] (2) Dynamic scheduling table generation: based on the generated protocol switching instruction, the system further constructs a dynamic scheduling table. This scheduling table is a set of instructions containing specific scheduling strategies, the content of which mainly defines:

[0135] Device communication timing: assign appropriate communication time slices or timing rules to related distributed energy storage devices to avoid channel conflicts and optimize resource use.

[0136] Communication path: specify the preferred communication path that the device should use after protocol switching.

[0137] The generation of this dynamic scheduling table aims to convert the macroscopic instructions of protocol switching into fine communication scheduling schemes that can be directly executed by the device.

[0138] The following data is output: protocol switching instruction: control command indicating system communication protocol switching; dynamic scheduling table: scheduling scheme containing specific device communication timing and communication path.

[0139] Specifically, the process of generating a dynamic scheduling table based on a protocol switching instruction includes:

[0140] Real-time acquisition of protocol switching instruction , device identifier set in device status data and a corresponding set of electrical quantity values , and a set of path load values of each communication path ;

[0141] based on the set of electrical quantity values determining a communication timing sequence , specifically:

[0142] (1) sorting the set of device identifiers in ascending order of the corresponding electrical quantity values to obtain an ordered device sequence ;

[0143] (2) assigning a communication time slot to each device according to the order of the ordered device sequence to generate a communication timing sequence , wherein represents the communication start time or timing priority of the device ;

[0144] based on the set of path load values determining a communication path sequence , specifically:

[0145] sorting the communication paths in ascending order of the path load values to generate an ordered communication path sequence ;

[0146] constructing an initial dynamic scheduling table , specifically:

[0147] (1) associating and mapping the communication timing sequence and the ordered communication path sequence to generate an initial scheduling scheme, denoted as a set , wherein each tuple represents the scheduling arrangement of using path at timing ;

[0148] (2) optimizing the initial dynamic scheduling table using a greedy algorithm to generate an optimized dynamic scheduling table , whose objective function is to minimize the total path load, denoted as:

[0149] ;

[0150] wherein represents the path load value of the path assigned to timing ;

[0151] (3) Based on the protocol switching command Switch to the corresponding protocol type And in the optimized dynamic scheduling table The protocol type is marked in the text. The output is labeled with the protocol type. Optimized dynamic scheduling table To the corresponding distributed energy storage device.

[0152] For example, the process of associating and mapping a communication time sequence with an ordered communication path sequence includes:

[0153] Obtain communication time sequence ,in Indicates the first The communication time slot or having the first Device identifier for communication priority. The length of the time series sequence;

[0154] Obtain the ordered communication path sequence The sequence is arranged in ascending order of path load value. Indicates the sorted order of the first... Bit-based communication path identifier, The path sequence length;

[0155] Establish communication timing sequence With ordered communication path sequence Mapping relationship between This mapping is for each communication timing sequence in the time series. Allocate a communication path Generate an initial set of scheduling schemes. ;

[0156] The mapping relationship The assigned path index is defined by the following formula. :

[0157] ;

[0158] in, It is time sequence In sequence The position index in the middle, ; The modulo operator is used to calculate... Divide by The remainder after calculation; The value is an ordered communication path sequence The path location index.

[0159] based on the mapping relationship , generate an initial dynamic scheduling table :

[0160] ;

[0161] wherein, indicates the communication path identifier allocated according to the mapping relationship for the time sequence , that is, the path at the th position in the ordered communication path sequence . .

[0162] Regarding step S5:

[0163] The following data is obtained:

[0164] Predicted load value: the load prediction value for the future period output by the pre-trained LSTM model in the load warning module.

[0165] Path load value: a quantitative value reflecting the current path comprehensive load pressure generated by the path load calculation module by weighted fusion of device state data and network performance data.

[0166] Pre-set parameter: a pressure threshold pre-stored in the system, used to determine whether the load pressure of the communication path is within an acceptable range.

[0167] (1) Establish a communication path quality evaluation matrix: based on the coupling relationship between the predicted load value and the path load value, a communication path quality evaluation matrix is established. This matrix takes the communication path in the system as the evaluation object, associates each path with its corresponding predicted load value and path load value, forming a two-dimensional data structure for systematic evaluation of path communication quality.

[0168] (2) Calculate the load pressure factor: for each path in the communication path quality evaluation matrix, calculate the difference between the predicted load value and the path load value, and define the difference as the load pressure factor of the path. This factor quantitatively reflects the load pressure level that the path may face in the future period (the degree of deficiency of the predicted load relative to the current path capacity).

[0169] (3) Path screening and elimination: compare the calculated load pressure factor of all paths with the system's pre-set pressure threshold. From the communication path quality evaluation matrix, eliminate all communication paths with a load pressure factor greater than the pressure threshold. This operation aims to actively exclude high-risk paths that may cause deterioration or interruption of communication quality due to excessive load pressure in the future period.

[0170] The optimized communication path set is output. Specifically, the communication paths that remain after the above filtering process and whose load pressure factors do not exceed the limit. These paths constitute a pool of available paths that are more reliable in quality and have load pressure within the expected range.

[0171] Further, the process of calculating the pressure threshold specifically includes:

[0172] Obtain a plurality of historical load pressure factor values stored in the historical communication path quality evaluation matrix to constitute a historical load pressure factor data set , wherein represents the historical load pressure factor, is a preset historical data quantity;

[0173] Calculate the average value of the historical load pressure factor :

[0174] ;

[0175] , wherein represents the sum operation on ;

[0176] Calculate the standard deviation of the historical load pressure factor :

[0177] ;

[0178] Calculate the pressure threshold :

[0179] .

[0180] Further, after removing the communication paths whose load pressure factors in the communication path quality evaluation matrix are greater than the pressure threshold, the removed communication paths are marked and reserved as backup paths. When the optimized communication path fails, the backup path is automatically switched to.

[0181] Through the technical solution, the application realizes intelligent determination and dynamic optimization of the communication path pressure threshold. Based on the statistical distribution characteristics of the historical load pressure factor, a pressure threshold calculation model with average value and standard deviation as the core is established, so that the threshold setting reflects the historical load characteristics and has the ability to adapt to abnormal fluctuations. Through the scientific calculation method of "average value + 2 times standard deviation", the effectiveness of path screening is ensured while the risk of false rejection is reduced. Through the innovative mechanism of marking the rejected path as a backup path, the optimization screening of the current communication path is completed, and an emergency communication guarantee resource is established for the system. In addition, the dynamic switching mechanism of the main and backup paths cooperates with the adaptive adjustment of the pressure threshold, forming a complete path fault tolerance protection system, improving the robustness and fault recovery capability of the communication system, and ensuring the continuous and stable operation of cross-platform communication.

[0182] Regarding step S6:

[0183] The following data is obtained:

[0184] Optimized communication path set: the available communication paths remaining after load pressure factor screening.

[0185] Dynamic scheduling table: an initial scheduling scheme containing device communication timing and communication paths.

[0186] (1) Construct an optimized communication path quality evaluation matrix: based on the optimized communication path set, an optimized communication path quality evaluation matrix is reconstructed. This matrix only contains high-quality paths with acceptable load pressure factors, providing accurate and reliable path quality information for subsequent scheduling.

[0187] (2) Construct an optimized dynamic scheduling table: based on the optimized communication path quality evaluation matrix, the initial dynamic scheduling table is optimized and reconstructed. Specifically, using the high-quality path information in the optimized matrix, the communication path allocation in the dynamic scheduling table is updated, and the device communication timing is adjusted accordingly to ensure that the scheduling scheme matches the current optimal communication path resources, forming the final optimized dynamic scheduling table.

[0188] Output the optimized dynamic scheduling table. This scheduling table is a complete scheduling scheme containing accurate device communication timing and optimized communication path instructions. The system will issue this optimized dynamic scheduling table to the corresponding distributed energy storage devices through the communication network.

[0189] After receiving the scheduling table, the device will perform data reception and transmission operations according to the specified communication timing and communication paths in the scheduling table, thereby realizing efficient, reliable and barrier-free communication across platforms.

[0190] Further, the optimized dynamic scheduling table based on the optimized communication path quality evaluation matrix is dynamically adjusted, specifically including:

[0191] Real-time monitoring of device state data and network performance data, when the power value or transmission delay value changes, recalculate the path load value and predicted load value;

[0192] According to the results of the recalculation, update the communication path quality evaluation matrix and the load pressure factor;

[0193] Regenerate the dynamic scheduling table every fixed time interval to ensure that the scheduling table matches the current network state;

[0194] The optimized dynamic scheduling table is transmitted to the distributed energy storage device in encrypted form, and the device adjusts the communication behavior according to the scheduling table.

[0195] As Figure 3 shown, the data flow diagram of the cross-platform barrier-free communication method is introduced; in order to facilitate understanding of the above embodiments, the following will take the data flow as an example Figure 3 to explain:

[0196] Scenario: In a certain regional energy grid, three distributed energy storage devices (Device A, Device B, Device C) are deployed for storing and dispatching electric energy. These devices communicate through heterogeneous network protocols (such as MQTT, CoAP) to collaboratively complete load balancing and energy scheduling tasks. Due to fluctuations in device power, changes in network bandwidth, and increases in transmission delay, communication links may become unstable.

[0197] Assumed parameters:

[0198] Pre-set power threshold: 40% (when the device power is lower than this value, it is considered to be low power);

[0199] Delay threshold: 40ms (when the transmission delay exceeds this value, it is considered to be network performance degradation);

[0200] Pressure threshold θ: 0.1 (calculated from historical load pressure factors, used for path elimination);

[0201] LSTM model predicts the next minute as the future time period;

[0202] Data normalization range: [0, 1];

[0203] Maximum load power reference value: 10kW;

[0204] Maximum bandwidth usage reference value: 100Mbps;

[0205] Step process:

[0206] (1) Real-time data acquisition

[0207] Real-time collection of device status data, network performance data, and historical load data through IoT chips:

[0208] Device A: Device ID = "ID_A", Power Value = 30%, Load Power = 5kW, Bandwidth Usage Value = 10Mbps, Transmission Delay Value = 50ms;

[0209] Device B: Device ID = "ID_B", Power Value = 50%, Load Power = 3kW, Bandwidth Usage Value = 8Mbps, Transmission Delay Value = 30ms;

[0210] Device C: Device ID = "ID_C", Power Value = 70%, Load Power = 2kW, Bandwidth Usage Value = 5Mbps, Transmission Delay Value = 20ms;

[0211] Historical Load Data: Device A's Load Power Time Series for the past 1 minute is [4.5kW, 4.8kW, 5.0kW] (for LSTM model input).

[0212] (2) Generate Path Load Value

[0213] Normalization: Map Power Value and Load Power to the range [0, 1].

[0214] Device A: Power Value Normalization = 0.3, Load Power Normalization = 0.5 (5 / 10);

[0215] Device B: Power Value Normalization = 0.5, Load Power Normalization = 0.3;

[0216] Device C: Power Value Normalization = 0.7, Load Power Normalization = 0.2;

[0217] Data Compression: Compress normalized data to reduce storage overhead (e.g., use lossless compression algorithm).

[0218] Weighted Fusion to Generate Path Load Value (Example for Device A):

[0219] Get Historical Data: Load Power Previous Two Time Values , ; Bandwidth Usage Value Previous Two Time Values , ; Sampling Time Interval seconds.

[0220] Calculate Load Power Change Rate ;

[0221] Calculate Bandwidth Usage Value Change Rate ;

[0222] Calculate weight coefficient:

[0223] ;

[0224] ;

[0225] Path load value (where bandwidth usage value normalization = 10 / 100 = 0.1);

[0226] Path load value of similar computing devices B and C: Device B: Z = 0.15, Device C: Z = 0.10.

[0227] (3) Load warning

[0228] Input the historical load data of device A into the pre-trained LSTM model, and output the predicted load value = 0.2 (normalized value).

[0229] Comparison condition:

[0230] Predicted load value (0.2) < Path load value (0.25)? Yes.

[0231] Transmission delay value (50ms) > Delay threshold (40ms)? Yes.

[0232] Satisfy both conditions, trigger load warning state (high load warning).

[0233] (4) Generate protocol switching instruction and dynamic scheduling table

[0234] Condition judgment: Device A is in high load warning state and power value (30%) < preset power threshold (40%), meet the condition.

[0235] Generate protocol switching instruction: switch from current MQTT protocol to CoAP protocol (more suitable for low power and high load environment).

[0236] Generate dynamic scheduling table:

[0237] Communication timing: based on power value, devices with low power have priority in communication. Device A has the lowest power, so it is assigned the first communication timing.

[0238] Communication path: based on path load value from small to large, preferentially select paths with low path load value. The order is device C (0.10), device B (0.15), device A (0.25).

[0239] Use greedy algorithm to optimize the scheduling table to avoid conflicts and maximize resource utilization.

[0240] Mark the protocol type as CoAP in the dynamic scheduling table.

[0241] (5) Path optimization

[0242] Establish a communication path quality evaluation matrix: contains all communication paths (path A, path B, path C), each path is associated with a predicted load value and a path load value.

[0243] Path A: predicted load value = 0.2, path load value = 0.25;

[0244] Path B: predicted load value = 0.25, path load value = 0.15;

[0245] Path C: predicted load value = 0.15, path load value = 0.10;

[0246] Calculate the load pressure factor (defined as the path load value minus the predicted load value to reflect the load pressure):

[0247] Path A: 0.25-0.2 = 0.05;

[0248] Path B: 0.15-0.25 = 0.10 (in this example, the pressure factor is taken as an absolute value to ensure a positive number);

[0249] In actual files, the load pressure factor can be the absolute value of the difference or a positive difference, but according to the context, the pressure factor should represent the size of the pressure, so in this example, the absolute value is taken: |0.15-0.25|=0.10;

[0250] Path C: |0.10-0.15|=0.05;

[0251] Eliminate paths with load pressure factors exceeding the threshold θ = 0.1: path B pressure factor 0.10 equals the threshold, not eliminated; but to demonstrate elimination, assume path B pressure factor is calculated as 0.12 (based on actual data), then eliminate path B.

[0252] Optimized to retain path A and path C.

[0253] (6) Output the optimized dynamic scheduling table

[0254] Based on the eliminated paths (path A and path C), construct an optimized communication path quality evaluation matrix.

[0255] Update the dynamic scheduling table: reassign communication timing and paths, preferentially use path C (low load pressure factor), device A uses path C according to timing.

[0256] Output the optimized dynamic scheduling table to devices A, B, C. Devices perform communication according to the scheduling table, for example, device A transmits data through path C using CoAP protocol at the specified timing.

[0257] Example 2

[0258] This invention also provides a cross-platform barrier-free communication system applied to a distributed energy storage device communication system in an energy grid, for implementing the above-mentioned cross-platform barrier-free communication method, including:

[0259] The data acquisition module 100 is used to acquire in real time the device status data, network performance data and historical load data of the distributed energy storage device. The device status data includes device identification, power value and load power, and the network performance data includes bandwidth usage value and transmission delay value.

[0260] The path load calculation module 200 is used to normalize the power consumption value and load power, associate the device identifier to generate standardized data, compress the standardized data, and weight and fuse the processed load power and the bandwidth usage value in the network performance data to generate the path load value.

[0261] The load warning module 300 is used to input historical load data into a pre-trained LSTM model and output the predicted load value for future periods. When the predicted load value is less than the path load value and the transmission delay value is greater than the delay threshold, the load warning state is triggered.

[0262] The scheduling generation module 400 is used to generate a protocol switching instruction when the load warning state is in effect and the power value is less than a preset power threshold, and to generate a dynamic scheduling table containing device communication timing and communication path based on the protocol switching instruction.

[0263] The path optimization module 500 is used to establish a communication path quality evaluation matrix based on the coupling relationship between the predicted load value and the path load value, calculate the difference between the predicted load value and the path load value in the communication path quality evaluation matrix as the load pressure factor, and eliminate communication paths in the communication path quality evaluation matrix whose load pressure factor exceeds the pressure threshold.

[0264] The scheduling optimization module 600 is used to construct an optimized communication path quality evaluation matrix based on the eliminated communication paths, construct an optimized dynamic scheduling table based on the optimized communication path quality evaluation matrix, and output the optimized dynamic scheduling table to the corresponding distributed energy storage device.

[0265] Example 3

[0266] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the cross-platform barrier-free communication method described above.

[0267] Example 4

[0268] The application further provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to construct a cross-platform barrier-free communication method.

[0269] Those skilled in the art can clearly understand that each embodiment can be constructed by means of software and necessary general hardware platform, and of course, can also be constructed by hardware through the above description of the embodiments. Based on such understanding, the above technical solutions or the part of the prior art which contributes essentially can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, etc.) execute the method described in each embodiment or some part of the embodiment.

[0270] The technical scope of the application is not limited to the above description, and those skilled in the art can make various modifications and changes to the above embodiments without departing from the technical idea of the application, and these modifications and changes should be within the protection scope of the application.

Claims

1. A cross-platform accessibility communication method, characterized by, A distributed energy storage device communication system applied to an energy grid comprises: Real-time acquisition of device state data, network performance data and historical load data of the distributed energy storage device through an Internet of Things chip, wherein the device state data comprises device identification, power value and load power, and the network performance data comprises bandwidth usage value and transmission delay value; Normalization processing of the power value and the load power, generation of standardized data by associating the device identification, data compression processing of the standardized data, and weighted fusion of the processed load power and the bandwidth usage value in the network performance data to generate a path load value; Input of the historical load data into a pre-trained LSTM model to output a predicted load value in a future period, triggering of a load warning state when the predicted load value is less than the path load value and the transmission delay value is greater than a delay threshold; Generation of a protocol switching instruction when in the load warning state and the power value is less than a preset power threshold, and generation of a dynamic scheduling table comprising device communication timing and communication paths based on the protocol switching instruction; Establishment of a communication path quality evaluation matrix based on the coupling relationship between the predicted load value and the path load value, calculation of the difference between the predicted load value and the path load value in the communication path quality evaluation matrix as a load stress factor, and elimination of communication paths in the communication path quality evaluation matrix whose load stress factor exceeds a stress threshold; Construction of an optimized communication path quality evaluation matrix based on the eliminated communication paths, construction of an optimized dynamic scheduling table based on the optimized communication path quality evaluation matrix, and output of the optimized dynamic scheduling table to the corresponding distributed energy storage device.

2. The cross-platform accessibility communication method of claim 1, wherein, Time synchronization and integrity verification of the acquired device state data and network performance data, comprising: Appending timestamp information to the acquired device state data and network performance data; Time synchronization verification of the device state data and network performance data based on the timestamp information, reacquisition of data when the time difference is greater than a preset synchronization threshold; Integrity verification of the device state data and network performance data after time synchronization verification, checking whether the data is missing or contains abnormal values; Linear interpolation completion of the missing data based on historical device state data and historical network performance data when the data is found to be missing; Elimination of the abnormal value and reacquisition of the data of the abnormal value when the abnormal value is found.

3. The cross-platform accessibility communication method of claim 1, wherein, Periodic updating of the LSTM model, specifically comprising: Periodic acquisition of the latest historical load data in a calibration time window to retrain the LSTM model; Acquisition of the data confidence of the retrained LSTM model in outputting the predicted load value, retraining of the LSTM model when the data confidence is less than a preset confidence threshold, until the data confidence of the retrained LSTM model is greater than or equal to the confidence threshold.

4. The cross-platform accessibility communication method of claim 1, wherein, Weighted fusion of the processed load power and the bandwidth usage value in the network performance data to generate a path load value, specifically comprising: acquiring the processed load power , a bandwidth usage value , a historical load power and as well as a historical bandwidth usage value and ; Computing load power rate of change : ； wherein, represents a load power value at a previous sampling time point, represents a load power value at a previous two sampling time points, represents a preset fixed sampling time interval; Computing a bandwidth usage value change rate : ； wherein, denotes the bandwidth usage value of the preceding sampling instant, denotes the bandwidth usage value of the two preceding sampling instants; Weighting factor for computing load power : ； wherein, represents an absolute value of a rate of change of the load power, represents an absolute value of a rate of change of the bandwidth usage value; Weighting coefficients for computing bandwidth usage values : ； Computing path load values : 。 5. The cross-platform accessibility communication method of claim 1, wherein, Collaborative verification of the warning credibility when the load warning state is triggered: When a single communication path triggers the load warning, transmission delay values and path load values of N adjacent communication paths in the topology structure are acquired. If the adjacent paths greater than the preset number threshold are also in the load warning state, a path reorganization instruction is triggered; If the adjacent path states greater than the preset number threshold are normal, an optimization instruction of the single communication path is triggered.

6. The cross-platform accessibility communication method of claim 1, wherein, The process of the pressure threshold calculation specifically comprises: Obtain a plurality of historical load stress factor values stored in the historical communication path quality evaluation matrix to form a historical load stress factor data set wherein represents the i th historical load stress factor, is a preset historical data quantity; calculating an average of the historical load pressure factors : ； wherein denotes a summing operation over denotes a summing operation over calculating a standard deviation of the historical load pressure factor : ； Computing pressure threshold : 。 7. The cross-platform accessibility communication method of claim 6, wherein, After removing the communication paths with load pressure factors greater than the pressure threshold in the communication path quality evaluation matrix, the removed communication paths are marked and reserved as backup paths. When the optimized communication path fails, the backup path is automatically switched to.

8. A cross-platform accessibility communication system, characterized by, The distributed energy storage device communication system applied to the energy grid is used to implement the cross-platform barrier-free communication method in any one of claims 1-7, comprising: A data acquisition module is configured to acquire device state data, network performance data, and historical load data of the distributed energy storage device in real time. The device state data includes device identification, power value, and load power. The network performance data includes bandwidth usage value and transmission delay value. A path load calculation module is configured to normalize the power value and load power, associate the device identification to generate standardized data, compress the standardized data, and weight and fuse the processed load power and bandwidth usage value in the network performance data to generate a path load value. A load warning module is configured to input the historical load data into a pre-trained LSTM model to output a predicted load value in a future period. When the predicted load value is less than the path load value and the transmission delay value is greater than a delay threshold, a load warning state is triggered. A scheduling generation module is configured to generate a protocol switching instruction when the load warning state and the power value are less than a preset power threshold, and generate a dynamic scheduling table containing device communication timing and communication paths based on the protocol switching instruction. A path optimization module is configured to establish a communication path quality evaluation matrix based on the coupling relationship between the predicted load value and the path load value, calculate the difference between the predicted load value and the path load value in the communication path quality evaluation matrix as a load pressure factor, and remove the communication paths with load pressure factors exceeding the threshold in the communication path quality evaluation matrix. A scheduling optimization module is configured to construct an optimized communication path quality evaluation matrix based on the removed communication paths, construct an optimized dynamic scheduling table based on the optimized communication path quality evaluation matrix, and output the optimized dynamic scheduling table to the corresponding distributed energy storage device. 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8. The processor executes the computer program to implement the cross-platform barrier-free communication method in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 9. The computer program is executed by the processor to implement the cross-platform barrier-free communication method in any one of claims 1-7.

Images