A method, apparatus, device, medium and product for handling energy storage system failure

Abstract

The embodiment of the application provides a kind of energy storage system fault handling method, device, equipment, medium and product, involve photovoltaic power generation field, this method is by constructing cloud-edge-end collaborative fault handling architecture, solve the problem of low efficiency in prior art fault handling, change traditional static threshold determination logic, adopt the decision mode of combination of multi-source information fusion and dynamic prediction: first, utilize edge computing to realize the preliminary diagnosis of real-time data, ensure the immediacy of response, then through cloud platform fusion historical trend analysis, equipment health prediction and multidimensional real-time state evaluation, generated the dynamic fault evaluation that can accurately reflect the actual severity and development trend of fault, so that the system can adapt to equipment aging and environmental change, and trigger the hierarchical response strategy that is accurately matched with fault level, realize the leap from passive alarm to active, accurate fault control, improve fault handling efficiency.

Description

Technical Field

[0001] This application relates to the field of photovoltaic power generation, and in particular to a method, apparatus, equipment, medium and product for handling faults in energy storage systems. Background Technology

[0002] Home energy storage systems have become key equipment for achieving energy self-sufficiency and promoting low-carbon transformation. The operating environment of home energy storage systems is affected by various factors such as temperature, humidity and grid conditions. Against this backdrop, troubleshooting energy storage systems has become a crucial link in ensuring their safe and stable operation, fully realizing their energy regulation value, and promoting the sustainable development of the new energy industry.

[0003] In existing technologies, sensors are deployed at key parts of the energy storage system to continuously collect operating data and compare it with a pre-set safety threshold. When any parameter continuously or momentarily exceeds the threshold range, the system determines it to be in an abnormal state and executes a preset protection action.

[0004] However, existing technologies suffer from low fault handling efficiency. This static threshold-based judgment method is difficult to adapt to the gradual degradation of equipment performance and dynamic changes in environmental conditions during the operation of energy storage systems. Furthermore, the fixed and singular response strategy leads to a mismatch between alarm frequency and the actual severity of the fault, thus limiting the overall efficiency of the fault handling process. Summary of the Invention

[0005] This application provides a method, apparatus, equipment, medium, and product for handling faults in an energy storage system, in order to solve the problem of low fault handling efficiency in the prior art.

[0006] In a first aspect, embodiments of this application provide a fault handling method for an energy storage system, applied to the cloud platform layer of the fault handling system, wherein the fault handling system further includes a sensor layer and an edge computing layer, and the method includes:

[0007] The system acquires preliminary diagnostic results, multi-dimensional data, and multiple historical time-series data. The preliminary diagnostic results are obtained by the edge computing layer based on the multi-dimensional data, which is collected by the sensor layer. The multiple historical time-series data are used to represent the operating status of the energy storage system within a preset first time period, where the end time of the first time period is earlier than the current time. The multi-dimensional data are used to represent the current operating status of the energy storage system and its environment.

[0008] The battery health status and remaining lifespan are predicted based on a preset prediction model and the multiple historical time-series data; wherein, the energy storage system includes the battery;

[0009] The preliminary diagnostic results and the multi-dimensional data are fused based on a preset fusion model to obtain a comprehensive fault evaluation result; wherein, the comprehensive fault evaluation result is used to represent the fault type and urgency of the fault in the energy storage system;

[0010] Based on the battery health status, remaining lifespan, and comprehensive fault evaluation results, a fault level is determined, and a target processing strategy corresponding to the fault level is triggered from among a plurality of preset processing strategies; wherein, the target processing strategy is used to instruct the sensor layer and the edge computing layer to perform a response action corresponding to the fault level and send an alarm signal corresponding to the fault level to a preset user terminal.

[0011] In one possible design, after triggering the target processing strategy corresponding to the fault level among a plurality of preset processing strategies, the following is also included:

[0012] Based on the battery health status, remaining lifespan, and comprehensive fault evaluation results, a collaborative optimization strategy is generated.

[0013] The parameters of the prediction model and the fusion model are optimized according to the collaborative optimization strategy.

[0014] The collaborative optimization strategy is sent to the sensor layer and the edge computing layer; wherein, the sensor layer is used to optimize multiple preset data acquisition parameters according to the collaborative optimization strategy, and the edge computing layer is used to optimize multiple preset fault detection parameters in the fault diagnosis according to the collaborative optimization strategy.

[0015] In one possible design, after sending the collaborative optimization strategy to the sensor layer and the edge computing layer, the method further includes:

[0016] Acquire strategy execution feedback data and system spatiotemporal information; wherein, the strategy execution feedback data includes parameter optimization data of the sensor layer after executing the collaborative optimization strategy, fault diagnosis results of the edge computing layer after executing the collaborative optimization strategy, and historical fault handling records of the energy storage system; the system spatiotemporal information includes deployment location information of multiple devices preset in the energy storage system;

[0017] Based on the spatiotemporal information of the system, the strategy execution feedback data is mapped to a preset spatiotemporal coordinate system to obtain a spatiotemporal correlation map;

[0018] The spatiotemporal correlation map is input into a preset causal analysis model for analysis to obtain a fault analysis report; wherein, the fault analysis report is used to represent the causes and propagation paths of the faults in the energy storage system.

[0019] In one possible design, the fault level includes a first level, a second level, and a third level, and the target processing strategy corresponding to the fault level among the multiple preset processing strategies includes:

[0020] In response to the fault level being the first level, the preset system log is updated according to the fault type and urgency of the fault in the energy storage system, and a preset first alarm message is sent to the user terminal; wherein, the first alarm message is sent via application push on the user terminal.

[0021] In response to the fault level being the second level, the edge computing layer is controlled to switch the energy storage system to a preset fault mitigation mode and send a preset second alarm message to the user terminal; wherein, the second alarm message is sent via SMS, and the fault mitigation mode refers to the preset derated operation state of the energy storage system;

[0022] In response to the fault level being the third level, the edge computing layer is controlled to perform emergency protection actions for the energy storage system and send a preset third alarm message to the user terminal; wherein, the third alarm message is sent via an automatic telephone call, and the emergency protection action includes cutting off the main power circuit of the energy storage system.

[0023] In one possible design, after sending the preset third alarm information to the user terminal, the method further includes:

[0024] Obtain the user's feedback behavior data to the alarm signal, and calculate the average user response time and ignore rate for the fault type based on the feedback behavior data;

[0025] Based on the average user response time and the ignore rate, the urgency thresholds for triggering the first, second, and third levels are dynamically adjusted to obtain the adjusted urgency thresholds.

[0026] Based on the adjusted urgency threshold and the battery health status, the mapping relationship between the fault level and the target processing strategy is updated, and the updated mapping relationship is sent to the edge computing layer.

[0027] In one possible design, the sensor layer includes multiple types of sensors, and the preliminary diagnostic results and the multi-dimensional data are fused based on a preset fusion model to obtain a comprehensive fault evaluation result, including:

[0028] Obtain historical acquisition data and data accuracy labels for each type of sensor; wherein, the data accuracy labels are used to represent the measurement error of each type of sensor;

[0029] Based on the historical data collected by each type of sensor and the data accuracy label, calculate the sensor accuracy corresponding to each dimension of the multi-dimensional data, and assign weight coefficients to each dimension of the data according to the accuracy.

[0030] Obtain the confidence threshold of historical diagnostic cases corresponding to the fault type, and filter the multi-dimensional data according to the weight coefficient and the confidence threshold to obtain an effective data set;

[0031] The preliminary diagnostic results and the effective data set are fused based on the fusion model to obtain the comprehensive fault evaluation result.

[0032] Secondly, embodiments of this application provide an energy storage system fault handling device, applied to the cloud platform layer of a fault handling system. The fault handling system further includes a sensor layer and an edge computing layer. The device includes:

[0033] The first acquisition module is used to acquire preliminary diagnostic results, multi-dimensional data, and multiple historical time-series data. The preliminary diagnostic results are obtained by the edge computing layer through fault diagnosis based on the multi-dimensional data. The multi-dimensional data is collected by the sensor layer. The multiple historical time-series data are used to represent the operating status of the energy storage system within a preset first time period. The end time of the first time period is earlier than the current time. The multi-dimensional data are used to represent the current operating status of the energy storage system and its environment.

[0034] A prediction module is used to predict the battery health status and remaining lifespan based on a preset prediction model and the multiple historical time-series data; wherein, the energy storage system includes the battery;

[0035] The fusion module is used to fuse the preliminary diagnostic results and the multi-dimensional data based on a preset fusion model to obtain a comprehensive fault evaluation result; wherein, the comprehensive fault evaluation result is used to represent the fault type and urgency of the fault in the energy storage system;

[0036] The determination module is used to determine the fault level based on the battery health status, the remaining lifespan, and the comprehensive fault evaluation result, and to trigger a target processing strategy corresponding to the fault level from a set of preset processing strategies; wherein, the target processing strategy is used to instruct the sensor layer and the edge computing layer to perform a response action corresponding to the fault level and send an alarm signal corresponding to the fault level to a preset user terminal.

[0037] In one possible design, the energy storage system fault handling device further includes:

[0038] The generation module is used to generate a collaborative optimization strategy based on the battery health status, the remaining lifespan, and the comprehensive fault evaluation results.

[0039] The optimization module is used to optimize the parameters of the prediction model and the fusion model according to the collaborative optimization strategy;

[0040] A sending module is used to send the collaborative optimization strategy to the sensor layer and the edge computing layer; wherein, the sensor layer is used to optimize multiple preset data acquisition parameters according to the collaborative optimization strategy, and the edge computing layer is used to optimize multiple preset fault detection parameters in the fault diagnosis according to the collaborative optimization strategy.

[0041] In one possible design, the energy storage system fault handling device further includes:

[0042] The second acquisition module is used to acquire strategy execution feedback data and system spatiotemporal information; wherein, the strategy execution feedback data includes parameter optimization data of the sensor layer after executing the collaborative optimization strategy, fault diagnosis results of the edge computing layer after executing the collaborative optimization strategy, and historical fault handling records of the energy storage system; and the system spatiotemporal information includes deployment location information of multiple devices preset in the energy storage system.

[0043] The mapping module is used to map the strategy execution feedback data to a preset spatiotemporal coordinate system based on the spatiotemporal information of the system, so as to obtain a spatiotemporal correlation map;

[0044] The analysis module is used to input the spatiotemporal correlation map into a preset causal analysis model for analysis and obtain a fault analysis report; wherein, the fault analysis report is used to represent the causes and propagation paths of the faults in the energy storage system.

[0045] In one possible design, the fault levels include a first level, a second level, and a third level, and the determining module includes:

[0046] The first update unit is configured to, in response to the fault level being the first level, update the preset system log according to the fault type and urgency of the fault in the energy storage system, and send the preset first alarm information to the user terminal; wherein, the first alarm information is sent via application push on the user terminal.

[0047] A switching unit is configured to, in response to the fault level being the second level, control the edge computing layer to switch the energy storage system to a preset fault mitigation mode and send a preset second alarm message to the user terminal; wherein, the second alarm message is sent via SMS, and the fault mitigation mode refers to the preset derated operation state of the energy storage system;

[0048] An execution unit is configured to, in response to the fault level being the third level, control the edge computing layer to perform emergency protection actions for the energy storage system and send a preset third alarm message to the user terminal; wherein the third alarm message is sent via an automatic telephone call, and the emergency protection action includes cutting off the main power circuit of the energy storage system.

[0049] In one possible design, the determining module further includes:

[0050] The first acquisition unit is used to acquire the user terminal's feedback behavior data to the alarm signal, and calculate the user's average response time and ignore rate for the fault type based on the feedback behavior data;

[0051] An adjustment unit is used to dynamically adjust the urgency thresholds for triggering the first level, the second level, and the third level based on the average user response time and the ignore rate, so as to obtain the adjusted urgency thresholds.

[0052] The second update unit is used to update the mapping relationship between the fault level and the target processing strategy according to the adjusted urgency threshold and the battery health status, and send the updated mapping relationship to the edge computing layer.

[0053] In one possible design, the sensor layer includes multiple types of sensors, and the fusion module includes:

[0054] The second acquisition unit is used to acquire historical data and data accuracy labels of each type of sensor; wherein the data accuracy labels are used to represent the measurement error of each type of sensor.

[0055] The calculation unit is used to calculate the sensor accuracy corresponding to each dimension data in the multi-dimensional data based on the historical acquisition data of each type of sensor and the data accuracy label, and to assign weight coefficients to each dimension data according to the accuracy.

[0056] The filtering unit is used to obtain the confidence threshold of historical diagnostic cases corresponding to the fault type, and to filter the multi-dimensional data according to the weight coefficient and the confidence threshold to obtain an effective data set.

[0057] The fusion unit is used to fuse the preliminary diagnostic results and the effective data set based on the fusion model to obtain the comprehensive fault evaluation result.

[0058] Thirdly, embodiments of this application provide an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0059] The memory stores computer-executed instructions;

[0060] When the processor executes the computer execution instructions stored in the memory, it is used to implement the energy storage system fault handling method as described in any of the first aspects.

[0061] Fourthly, embodiments of this application provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the energy storage system fault handling method as described in any of the first aspects.

[0062] Fifthly, embodiments of this application provide a computer program product, including a computer program, which, when executed by a processor, is used to implement the energy storage system fault handling method as described in any of the first aspects.

[0063] This application provides a method, apparatus, device, medium, and product for handling faults in an energy storage system. By constructing a cloud-edge-device collaborative fault handling architecture, it solves the problem of low fault handling efficiency in existing technologies. It changes the traditional single, static threshold judgment logic and instead adopts a decision-making mode that combines multi-source information fusion and dynamic prediction. First, edge computing is used to achieve preliminary diagnosis of real-time data, ensuring the immediacy of the response. Then, by integrating historical trend analysis, equipment health prediction, and multi-dimensional real-time status evaluation through a cloud platform, a dynamic fault evaluation that accurately reflects the actual severity and development trend of the fault is generated. This enables the system to adapt to equipment aging and environmental changes and trigger a graded response strategy that precisely matches the fault level, reducing false alarms and missed alarms. It achieves a leap from passive alarm to proactive and precise fault control, improving overall fault efficiency and reliability. Attached Figure Description

[0064] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0065] Figure 1 This is a schematic diagram illustrating an application scenario of the energy storage system fault handling method provided in the embodiments of this application;

[0066] Figure 2 A flowchart illustrating the energy storage system fault handling method provided in this application embodiment. Figure 1 ;

[0067] Figure 3 A flowchart illustrating the energy storage system fault handling method provided in this application embodiment. Figure 2 ;

[0068] Figure 4 A flowchart illustrating the energy storage system fault handling method provided in this application embodiment. Figure 3 ;

[0069] Figure 5 A flowchart illustrating the energy storage system fault handling method provided in this application embodiment. Figure 4 ;

[0070] Figure 6 This application provides a cloud-edge-device collaborative operation architecture diagram for an energy storage system.

[0071] Figure 7 This application provides a flowchart of a closed-loop process for multi-level alarm and fault handling in an energy storage system.

[0072] Figure 8 A schematic diagram of the structure of the energy storage system fault handling device provided in the embodiments of this application;

[0073] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application.

[0074] The accompanying drawings have illustrated specific embodiments of this application, which will be described in more detail below. These drawings and descriptions are not intended to limit the scope of the concept in any way, but rather to illustrate the concept of this application to those skilled in the art through reference to specific embodiments. Detailed Implementation

[0075] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0076] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, use and processing of the relevant data must comply with relevant laws, regulations and standards, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0077] In the embodiments of this application, the terms "first" and "second" are used to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and that "first" and "second" do not necessarily imply difference. It should be noted that in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner. In the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more.

[0078] It should be noted that the phrase "at...time" in the embodiments of this application can refer to the instant at which a certain situation occurs, or to a period of time after the occurrence of a certain situation. The embodiments of this application do not specifically limit this. In addition, the energy storage system fault handling method, device, equipment, medium and product provided in the embodiments of this application are only examples. An energy storage system fault handling method, device, equipment, medium and product may also include more or less content.

[0079] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the invention as detailed in the appended claims.

[0080] The technical solution of the present invention will be described in detail below with reference to specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of the present invention will now be described with reference to the accompanying drawings.

[0081] To clearly understand the technical solution of this application, the existing technical solutions will first be described in detail. Home energy storage systems have become key equipment for achieving energy self-sufficiency and promoting low-carbon transformation. The operating environment of home energy storage systems is affected by various factors such as temperature, humidity, and grid conditions. Against this backdrop, troubleshooting of energy storage systems has become a crucial link in ensuring their safe and stable operation, fully realizing their energy regulation value, and promoting the sustainable development of the new energy industry.

[0082] In existing technologies, sensors are deployed at key components of the energy storage system to continuously collect operational data and compare it with pre-set safety thresholds. When any parameter continuously or momentarily exceeds the threshold range, the system determines an abnormal state and executes preset protection actions. This method struggles to adapt to the gradual degradation of equipment performance and dynamic changes in environmental conditions during energy storage system operation. Furthermore, the fixed and simplistic response strategy leads to a mismatch between alarm frequency and the actual severity of the fault, limiting the overall efficiency of fault handling. Therefore, existing technologies suffer from low fault handling efficiency.

[0083] Therefore, addressing the issue of low fault handling efficiency in existing technologies, this research found that to solve this problem, real-time equipment operating conditions and full-cycle historical operating data can be integrated. Accurate fault determination can be achieved through hierarchical diagnostic analysis and multi-source information fusion, establishing a dynamic matching mechanism between fault levels and differentiated handling strategies: ① Data collection and preliminary diagnosis tasks can be decentralized to edge nodes closer to the equipment, enabling rapid fault perception and initial screening. Simultaneously, non-real-time tasks such as massive historical data mining and complex model calculations can be moved to the cloud platform. This hierarchical division of labor resolves the contradiction between immediate response and in-depth analysis inherent in traditional centralized architectures, providing architectural support for improving fault handling efficiency. ② A fault level classification standard based on comprehensive fault evaluation results can be established, and differentiated hierarchical handling strategies can be configured for different fault levels. Alarm notifications, parameter adjustments, and equipment shutdowns can be precisely linked to fault severity, replacing the traditional one-size-fits-all fixed response model. This reduces meaningless frequent alarms and excessive handling, achieving optimal allocation of fault handling resources and improving the overall effectiveness of the fault handling process. ③ By combining historical operating sequence data of the entire life cycle of energy storage equipment, a state prediction model adapted to the gradual decline law of equipment performance can be constructed. By mining the correlation between equipment operating trends and decline characteristics, the health status and fault development trend of equipment can be predicted, so that the fault judgment logic can dynamically adapt to equipment aging and environmental changes, and solve the problem of insufficient alarm accuracy caused by fixed thresholds.

[0084] Specifically, by integrating real-time operating data and full-cycle historical operating data of energy storage systems, preliminary fault screening and in-depth situation assessment can be carried out in a hierarchical manner. Combined with the dynamic prediction results of equipment health status and remaining lifespan, a multi-dimensional comprehensive fault evaluation system can be constructed, and a precise linkage mechanism between fault level and differentiated handling strategy can be established. This breaks through the limitations of the traditional static threshold judgment mode, adapts to the actual needs of equipment performance degradation and dynamic environmental changes, and improves the accuracy and efficiency of fault handling.

[0085] This application discloses a method, apparatus, device, medium, and product for handling faults in an energy storage system. By constructing a cloud-edge-device collaborative fault handling architecture, it solves the problem of low fault handling efficiency in the prior art. It changes the traditional single, static threshold judgment logic and instead adopts a decision-making mode that combines multi-source information fusion and dynamic prediction. First, edge computing is used to achieve preliminary diagnosis of real-time data, ensuring the immediacy of the response. Then, by integrating historical trend analysis, equipment health prediction, and multi-dimensional real-time status evaluation through the cloud platform, a dynamic fault evaluation that can accurately reflect the actual severity and development trend of the fault is generated. This enables the system to adapt to equipment aging and environmental changes and trigger a graded response strategy that is precisely matched with the fault level, reducing false alarms and missed alarms. It realizes a leap from passive alarm to proactive and precise fault control, improving overall fault efficiency and reliability.

[0086] Based on the above-mentioned inventive discovery, the technical solution of this application is proposed.

[0087] The following describes the application scenarios of the energy storage system fault handling method provided in the embodiments of the present invention. Figure 1 This is a schematic diagram illustrating an application scenario of the energy storage system fault handling method provided in the embodiments of this application. For example... Figure 1 As shown, this application scenario includes a cloud platform layer 101, a sensor layer 102, and an edge computing layer 103. The sensor layer 102 collects multi-dimensional data and sends it to the cloud platform layer 101 and the edge computing layer 103. The edge computing layer 103 performs fault diagnosis based on the multi-dimensional data to obtain a preliminary diagnosis result and sends the preliminary diagnosis result to the cloud platform layer 101. The cloud platform layer 101 acquires multiple historical time-series data. Based on a preset prediction model and the multiple historical time-series data, the cloud platform layer 101 predicts the battery health status and remaining lifespan. Based on a preset fusion model, the cloud platform layer 101 fuses the preliminary diagnosis result and the multi-dimensional data to obtain a comprehensive fault evaluation result. Based on the battery health status, remaining lifespan, and comprehensive fault evaluation result, the cloud platform layer 101 determines the fault level and triggers the target processing strategy corresponding to the fault level from among multiple preset processing strategies.

[0088] The embodiments of the present invention will now be described with reference to the accompanying drawings.

[0089] Figure 2 A flowchart illustrating the energy storage system fault handling method provided in this application embodiment. Figure 1 .like Figure 2 As shown, in this embodiment, the execution entity of this invention is the cloud platform layer. Therefore, the energy storage system fault handling method provided in this embodiment includes the following steps:

[0090] S201. Obtain preliminary diagnostic results, multi-dimensional data, and multiple historical time-series data. The preliminary diagnostic results are obtained by the edge computing layer based on the multi-dimensional data for fault diagnosis. The multi-dimensional data is collected by the sensor layer. The multiple historical time-series data are used to represent the operating status of the preset energy storage system within a preset first time period. The end time of the first time period is earlier than the current time. The multi-dimensional data are used to represent the current operating status of the energy storage system and its environment.

[0091] Specifically, the sensor layer can collect real-time operating status parameters and environmental parameters of the energy storage system, forming multi-dimensional data, which is then synchronously transmitted to the edge computing layer and the cloud platform layer. After receiving the multi-dimensional data, the edge computing layer performs real-time analysis and fault diagnosis, generating preliminary diagnostic results which are then uploaded to the cloud platform layer. The cloud platform layer retrieves pre-stored operating status records of the corresponding energy storage system within a specific time period, obtaining multiple historical time-series data. This step provides complete foundational data for subsequent prediction of battery health status and remaining lifespan based on historical time-series data, as well as for fusing the preliminary diagnostic results and multi-dimensional data.

[0092] Multi-dimensional data refers to a collection of heterogeneous data from multiple sources, collected by various types of sensing devices deployed at the sensor layer, which comprehensively reflects the operating status of the energy storage system and its surrounding environment. This data primarily includes three core categories: first, battery operating parameters, covering core operating indicators such as voltage, current, temperature, state of charge, and health status of individual battery cells and modules; second, environmental monitoring parameters, including temperature, humidity, smoke concentration, and air pressure within the energy storage compartment, as well as environmental characteristics such as light intensity and wind speed; and third, equipment collaboration and grid interaction parameters, involving related data such as photovoltaic inverter output power, grid-side voltage and frequency, energy storage converter operating conditions, and communication status between the edge computing layer and the cloud platform layer. In some scenarios, security monitoring parameters such as access control status and infrared detection signals can also be included. These multi-dimensional data complement and support each other, providing comprehensive and accurate basic sensing information for subsequent fault diagnosis, status prediction, and data fusion.

[0093] S202. Predict the battery health status and remaining lifespan based on a preset prediction model and multiple historical time series data; wherein, the energy storage system includes the battery.

[0094] Specifically, multiple historical time-series data can be input into a pre-trained prediction model. This model can be a Bi-LSTM with Attention model, consisting of a data preprocessing module, a feature extraction module, and a state prediction module. The historical time-series data first enters the data preprocessing module for data cleaning, normalization, and removal of redundant and outlier data. The preprocessed results flow into the feature extraction module, which extracts key features from the data, such as battery operating patterns and performance degradation trends. The extracted features are then fed into the state prediction module, which outputs the battery health status and remaining lifespan based on the input feature data. The entire prediction model takes multiple historical time-series data points from the energy storage system within a first time period as input and outputs the battery health status and remaining lifespan. This step provides a basis for determining the actual state of the battery by combining preliminary diagnostic results with multi-dimensional data to determine the corresponding processing strategy.

[0095] S203. Based on the preset fusion model, the preliminary diagnosis results and multi-dimensional data are fused to obtain a comprehensive fault evaluation result; wherein, the comprehensive fault evaluation result is used to represent the fault type and urgency of the fault in the energy storage system.

[0096] Specifically, based on the DS evidence algorithm, preliminary diagnostic results and multi-dimensional data can be synchronously input into a pre-trained fusion model. This fusion model consists of a data adaptation module, a feature association module, and a result output module. The preliminary diagnostic results and multi-dimensional data first enter the data adaptation module to unify data formats and align data, eliminating differences in source and structure between the two types of data. The adapted data then flows into the feature association module, which performs deep correlation matching between the fault features in the preliminary diagnostic results and the operating status and environmental features in the multi-dimensional data, uncovering the inherent connections between the two types of data. The correlated feature information is then sent to the result output module, which generates a comprehensive evaluation result based on the matching results, including specific types and urgency levels. The entire fusion model takes preliminary diagnostic results and multi-dimensional data as input and outputs a comprehensive evaluation result. This step provides a comprehensive and accurate system status reference for subsequent determination of corresponding processing strategies based on battery health status and remaining lifespan, ensuring that subsequent processes can be carried out in an orderly manner based on complete status information.

[0097] Among them, the Dempster evidence algorithm is a fusion method for handling uncertain, incomplete, and inconsistent multi-source information. Its core is to abandon the dependence of traditional probability theory on prior probability, define a basic probability allocation function to describe the degree of direct trust in the proposition, and then derive a trust function and a likelihood function. With the help of Dempster's combination rule, independent evidence from different information sources is fused, gradually narrowing the uncertainty interval, and finally obtaining a comprehensive and reliable trust evaluation result for the target proposition. This algorithm is widely used in fields that require the integration of multi-dimensional fuzzy information, such as multi-sensor data fusion, fault diagnosis, fault evaluation, and pattern recognition.

[0098] S204. Based on the battery health status, remaining lifespan, and comprehensive fault evaluation results, determine the fault level and trigger the target processing strategy corresponding to the fault level from among multiple preset processing strategies; wherein, the target processing strategy is used to instruct the sensor layer and edge computing layer to perform response actions corresponding to the fault level and send alarm signals corresponding to the fault level to the preset user terminal.

[0099] Specifically, the battery health status, remaining lifespan, and comprehensive evaluation results can be integrated and compared against preset grading standards to determine the appropriate grading level. Then, a target processing strategy matching the determined level is selected from multiple preset processing strategies. Instructions to execute corresponding response actions are then sent to the sensor layer and edge computing layer, while corresponding signals are sent to the preset user terminal. This step provides a clear direction for the energy storage system's operation, prompting the sensor layer and edge computing layer to execute corresponding actions, while allowing the user terminal to promptly grasp the relevant system status, ensuring the closed-loop implementation of subsequent processes.

[0100] The energy storage system fault handling method provided in this application can be applied to new energy and power system related fields such as smart grid, photovoltaic grid connection and microgrid operation. Through the full-process management and control capabilities of real-time diagnosis, status prediction, hierarchical disposal and closed-loop optimization, it can solve the safety, stability and economic problems of energy storage system operation in different scenarios.

[0101] In the field of smart grids, the method of this application can solve the following problems in the grid-connected operation of energy storage systems: First, it solves the problem of lagging fault diagnosis of energy storage systems during grid peak-valley regulation. In smart grids, energy storage systems play a core role in smoothing load peak-valley differences, participating in grid frequency and voltage regulation, and ensuring grid stability. Traditional fault handling methods mostly rely on single cloud-based diagnosis, which has the disadvantages of large data transmission latency and slow local emergency fault response. This application uses a cloud-edge-device collaborative architecture, where the edge computing layer realizes real-time preprocessing and preliminary diagnosis of multi-dimensional data, and the cloud platform layer combines historical time-series data to complete accurate prediction of battery health status and remaining life. It can respond to energy storage system anomalies under grid load fluctuations in milliseconds, avoiding grid frequency deviation, voltage fluctuations, and other related problems caused by untimely fault handling. The application addresses several key issues: First, it addresses the problem of energy storage system operation status and grid demand mismatch. Second, it resolves the issue of low matching between the operating status of energy storage systems and grid demand during grid dispatch. This application uses a fusion model to output comprehensive fault evaluation results, combined with battery health status and remaining lifespan, to determine the fault level. This allows for targeted triggering of differentiated handling strategies such as derating operation and emergency shutdowns. This ensures that the grid can support energy storage output during high-load periods while preventing overcharging and discharging of batteries in unhealthy states, achieving bidirectional adaptation between grid dispatch commands and the safe operation of energy storage systems. Third, it addresses the challenge of managing large-scale energy storage clusters within the grid. The application's tiered alarm and closed-loop handling mechanism enables unified status monitoring and differentiated management of multiple energy storage devices within the grid, providing the grid dispatch center with accurate energy storage cluster operation data and facilitating integrated optimization of the power grid, grid, load, and storage systems.

[0102] In the field of photovoltaic power generation, the method of this application can solve the following technical problems: First, it solves the problem of frequent charging and discharging of energy storage systems and rapid battery life degradation caused by the fluctuation and intermittency of photovoltaic output. Photovoltaic power generation is significantly affected by environmental factors such as light intensity and temperature, resulting in large output fluctuations. The supporting energy storage system needs to frequently respond to changes in photovoltaic output to adjust charging and discharging. Traditional methods lack real-time prediction and fault control of battery health status, which can easily lead to overcharging and over-discharging of batteries and shortened cycle life. This application analyzes the historical time series data of the energy storage system through a preset prediction model to accurately predict the battery health status and remaining life. Combined with real-time monitoring data of photovoltaic output, it dynamically adjusts the energy storage charging and discharging strategy. Under the premise of ensuring full consumption of photovoltaic power, it avoids the battery from operating in a high-fault state and significantly extends the battery life. Second, it solves the problem of weak fault early warning capability and poor grid connection safety of photovoltaic grid-connected energy storage systems. When the photovoltaic-energy storage system is connected to the grid, the photovoltaic inverter and energy storage transformer... The coordinated operation of photovoltaic and energy storage systems is prone to faults such as harmonic interference and islanding effects. Traditional methods often rely on threshold judgment for fault alarms, resulting in a high false alarm rate. This application, through multi-dimensional data fusion, covering photovoltaic module output parameters, energy storage system operating parameters, and environmental monitoring parameters, can accurately identify hidden faults in the coordinated operation of photovoltaic and energy storage systems, trigger early warnings, and take protective measures in advance, reducing the impact of photovoltaic grid connection on the power grid and improving the safety and stability of the photovoltaic-energy storage system grid connection. Thirdly, it solves the problem of low operation and maintenance efficiency of distributed photovoltaic energy storage systems. Distributed photovoltaic energy storage systems are widely distributed and have a large number of devices. Traditional manual operation and maintenance modes have the disadvantages of high cost and slow fault diagnosis. The cloud platform layer of this application can realize centralized monitoring and remote control of multiple distributed photovoltaic energy storage systems. Combined with a hierarchical alarm mechanism, fault information is accurately pushed to operation and maintenance personnel, enabling rapid fault location and handling, significantly reducing the operation and maintenance cost of distributed photovoltaic energy storage systems, and improving the overall operating efficiency of the system.

[0103] This embodiment provides a fault handling method for energy storage systems. By dynamically predicting query execution time and combining it with real-time optimization strategy selection, it solves the problem of low query efficiency caused by the reliance on static rules in existing technologies. When the prediction time exceeds a set threshold, a target optimization strategy suitable for the current query characteristics and system state is dynamically selected from multiple optimization strategies. The query information is then optimized according to the target optimization strategy to obtain a query plan, which is executed to obtain the target data. This method can flexibly adjust the query execution strategy based on real-time query characteristics and prediction results, overcoming the shortcomings of static optimization rules in adapting to data changes and load fluctuations. In scenarios with sudden changes in data distribution or high-concurrency queries, it improves data query efficiency by dynamically optimizing the query plan.

[0104] In one possible design, the fault levels include a first level, a second level, and a third level. In step S204, the target processing strategy corresponding to the fault level is triggered from among several preset processing strategies, including:

[0105] S2041. In response to a fault level of Level 1, update the preset system log according to the fault type and urgency of the fault in the energy storage system, and send a preset first alarm message to the user terminal; wherein, the first alarm message is sent via push notification from the user terminal application.

[0106] Specifically, after determining the level to be Level 1, relevant information regarding the corresponding type and urgency can be extracted. This information can then be supplemented and updated according to preset system log recording specifications. Simultaneously, the first alarm information can be retrieved from a preset alarm information database and sent to the preset user terminals via application push notifications. This step preserves a complete and clear operational record for subsequent system status viewing and record retrieval, while also ensuring that users are promptly informed of the system's current status.

[0107] S2042. In response to the fault level being the second level, the control edge computing layer switches the energy storage system to a preset fault mitigation mode and sends a preset second alarm message to the user terminal; wherein, the second alarm message is sent via SMS, and the fault mitigation mode refers to the preset derated operation state of the energy storage system.

[0108] Specifically, after determining the level to be Level 2, the cloud platform layer can issue a mode switching command to the edge computing layer, instructing the edge computing layer to adjust the energy storage system to a preset derated operation state. Simultaneously, it can retrieve a second alarm from a preset alarm information database and send it to a preset user terminal via SMS. This step ensures the energy storage system operates at a low load, guaranteeing stable system operation while allowing users to be promptly informed of the system's current status.

[0109] S2043. In response to the fault level being the third level, the control edge computing layer executes emergency protection actions for the energy storage system and sends a preset third alarm message to the user terminal; wherein, the third alarm message is sent via an automatic telephone call, and the emergency protection actions include cutting off the main power supply circuit of the energy storage system.

[0110] Specifically, after the level is determined to be Level 3, the cloud platform layer can issue an emergency protection command to the edge computing layer, instructing the edge computing layer to cut off the main power circuit of the energy storage system. Simultaneously, it can retrieve the third alarm information from a pre-set alarm database and send it to the pre-set user terminals via automatic telephone calls. This step is used to quickly terminate the current operation of the energy storage system, preventing further impact from the system's deteriorating state, while also ensuring that users are promptly informed of the system's real-time status.

[0111] The technical effect of this solution in this embodiment is as follows: By establishing a graded response and alarm mechanism that strictly matches the fault level, it solves the problem that the fault response strategy in the prior art cannot carry out differentiated and precise handling according to the actual severity of the fault. Based on the dynamically evaluated fault level, it triggers multi-level responses in sequence from low to high, such as log recording and application notification, system de-rated operation and SMS alarm, and emergency power outage and telephone call. This ensures that the handling measures and alarm intensity are highly adapted to the immediate threat level of the fault, avoids over-response and user alarm fatigue while ensuring safety, and achieves optimized allocation of fault management resources and precise improvement of response efficiency.

[0112] In one possible design, the sensor layer includes multiple types of sensors. In S203, the preliminary diagnostic results and multi-dimensional data are fused based on a preset fusion model to obtain a comprehensive fault evaluation result, including:

[0113] S2031. Obtain historical acquisition data and data accuracy labels for each type of sensor; wherein, the data accuracy labels are used to represent the measurement error of each type of sensor.

[0114] Specifically, the cloud platform can retrieve past acquisition records from pre-set storage modules to obtain historical acquisition data for each type of sensor, and retrieve pre-labeled accuracy tags representing the measurement errors of each type of sensor. This step provides complete foundational data for subsequent calculations of sensor accuracy corresponding to each dimension of the multi-dimensional data based on historical acquisition data and accuracy tags, as well as for assigning weighting coefficients to each dimension of the data.

[0115] S2032. Based on the historical data collected by various types of sensors and the data accuracy labels, calculate the sensor accuracy corresponding to each dimension of the multi-dimensional data, and assign weight coefficients to each dimension of the data according to the accuracy.

[0116] Specifically, by combining historical data and accuracy labels from various types of sensors, the deviation between the historical data and corresponding benchmark values ​​can be compared. Combined with the measurement errors indicated by the accuracy labels, the sensor accuracy for each dimension of the multi-dimensional data can be calculated. Then, weighting coefficients are assigned to each dimension based on its accuracy, with higher accuracy dimensions receiving larger weighting coefficients. This step provides a clear numerical basis for subsequent filtering of multi-dimensional data based on weighting coefficients and confidence thresholds, ensuring the reliability of the selected data.

[0117] S2033. Obtain the confidence threshold of historical diagnostic cases corresponding to the fault type, and filter the multi-dimensional data according to the weight coefficient and the confidence threshold to obtain an effective data set.

[0118] Specifically, the cloud platform can retrieve historical diagnostic cases matching the corresponding type from a pre-defined historical case storage module, extract the marked confidence thresholds, compare the weight coefficients of each dimension's data with these confidence thresholds, and retain the dimension data with weight coefficients higher than the thresholds. This data is then integrated to form a valid dataset. This step provides more reliable dimensional data support for subsequent fusion of preliminary diagnostic results and the valid dataset based on a fusion model.

[0119] S2034. Based on the fusion model, the preliminary diagnostic results and the effective data set are fused to obtain the comprehensive fault evaluation results.

[0120] Specifically, preliminary diagnostic results and valid datasets can be input into a pre-defined fusion model. This model consists of a data adaptation module, a feature fusion module, and a result output module. The preliminary diagnostic results and valid datasets first enter the data adaptation module, where format unification and dimensional alignment are achieved, eliminating structural differences between the two types of data. The adapted data then flows into the feature fusion module, which mines the inherent correlation between fault features in the preliminary diagnostic results and various state features in the valid dataset, completing deep fusion at the feature level. The fused feature information is then sent to the result output module, which generates a comprehensive evaluation result based on the fused features, including type and urgency. The entire fusion model takes preliminary diagnostic results and valid datasets as input and outputs a comprehensive evaluation result. This step provides accurate system status information for subsequent level determination and triggering of corresponding processing strategies, ensuring that subsequent processes align with the actual system state.

[0121] The technical effect of this solution in this embodiment is as follows: By introducing a dynamic screening mechanism based on the data credibility of sensor historical performance and diagnostic cases, the problem of inaccurate overall evaluation results caused by directly using raw multi-source data in fault fusion in the prior art, or by misjudging individual sensor errors or specific fault types, is solved. By combining the historical accuracy of sensors with dynamic data weight allocation and joint screening based on the diagnostic confidence threshold of similar faults, the data input to the fusion model is ensured to have high reliability. This improves the accuracy and robustness of multi-source information fusion at the source, making the final fault evaluation result more realistically reflect the actual safety status of the system.

[0122] Figure 3 A flowchart illustrating the energy storage system fault handling method provided in this application embodiment. Figure 2 In this embodiment, in Figure 2 Based on the provided embodiments, the energy storage system fault handling method is further explained. The energy storage system fault handling method includes:

[0123] S301. Obtain preliminary diagnostic results, multi-dimensional data, and multiple historical time series data. The preliminary diagnostic results are obtained by the edge computing layer based on the multi-dimensional data for fault diagnosis. The multi-dimensional data is collected by the sensor layer. The multiple historical time series data are used to represent the operating status of the preset energy storage system within a preset first time period. The end time of the first time period is earlier than the current time. The multi-dimensional data are used to represent the current operating status of the energy storage system and its environment.

[0124] S302. Predict the battery health status and remaining lifespan based on a preset prediction model and multiple historical time series data; wherein, the energy storage system includes the battery.

[0125] S303. Based on the preset fusion model, the preliminary diagnostic results and multi-dimensional data are fused to obtain a comprehensive fault evaluation result; wherein, the comprehensive fault evaluation result is used to represent the fault type and urgency of the fault in the energy storage system.

[0126] S304. Based on the battery health status, remaining lifespan, and comprehensive fault evaluation results, determine the fault level and trigger the target processing strategy corresponding to the fault level from among multiple preset processing strategies; wherein, the target processing strategy is used to instruct the sensor layer and edge computing layer to perform response actions corresponding to the fault level and send alarm signals corresponding to the fault level to the preset user terminal.

[0127] S301-S304 are similar to S201-S204, and will not be described again in this embodiment.

[0128] S305. Generate a collaborative optimization strategy based on the battery health status, remaining lifespan, and comprehensive fault evaluation results.

[0129] Specifically, this involves integrating battery health status, remaining lifespan, and comprehensive fault assessment results to analyze the inherent correlation among these three factors. Combined with the current operating status and environmental characteristics of the energy storage system, it clarifies the adjustment directions for prediction model parameters, fusion model parameters, sensor data acquisition parameters, and edge computing layer fault detection parameters, forming a collaborative optimization strategy covering both the cloud platform layer model and front-end node parameters. This step provides a clear and feasible action plan for subsequent parameter optimization of the prediction and fusion models, as well as guiding the sensor and edge computing layers to adjust their own operating parameters, ensuring that the operating parameters of the entire system always adapt to the current equipment status and environmental conditions.

[0130] S306. Optimize the parameters of the prediction model and the fusion model according to the collaborative optimization strategy.

[0131] Specifically, the parameter adjustment requirements for the prediction model and the fusion model in the collaborative optimization strategy can be analyzed. This involves extracting parameter adjustment instructions for the prediction model's data preprocessing module (cleaning threshold, normalization coefficient), feature weights for the feature extraction module, fitting coefficients for the state prediction module, and format conversion rules for the fusion model's data adaptation module, association thresholds for the feature association module, and decision coefficients for the result output module. Then, parameters are updated for each component of both the prediction model and the fusion model according to these instructions, completing the parameter optimization for both models. This step ensures that the operating parameters of the prediction model and the fusion model are adapted to the current equipment status and environmental characteristics of the energy storage system, guaranteeing the accuracy of the subsequent output results and supporting the stable operation of the entire system's subsequent processing flow.

[0132] S307. Send the collaborative optimization strategy to the sensor layer and the edge computing layer; wherein, the sensor layer is used to optimize multiple preset data acquisition parameters according to the collaborative optimization strategy, and the edge computing layer is used to optimize multiple preset fault detection parameters in fault diagnosis according to the collaborative optimization strategy.

[0133] Specifically, the cloud platform layer can send collaborative optimization strategies to the sensor layer and the edge computing layer via pre-defined communication links. The strategy for the sensor layer includes adjustment requirements for data acquisition parameters, while the strategy for the edge computing layer includes adjustment requirements for fault detection parameters. This step allows the sensor layer to optimize its data acquisition parameters according to the adjustment requirements, and the edge computing layer to optimize its fault detection parameters accordingly. This ensures that the front-end data acquisition and diagnostic work adapts to the current device status and environmental conditions, guaranteeing the coordinated operation of all components of the entire system.

[0134] The technical effect of this solution in this embodiment is as follows: by introducing a closed-loop optimization mechanism based on real-time fault evaluation and equipment health prediction, the technical problems of fixed fault handling models and parameters and inability to self-adjust with the dynamic evolution of system status in the prior art are solved. By generating and distributing collaborative optimization strategies, the prediction model, fusion model and even the underlying sensing and diagnostic parameters can be continuously iterated according to the latest battery status and fault situation, thereby ensuring that the accuracy of fault identification and evaluation can continuously evolve with the entire life cycle of the system.

[0135] Figure 4 A flowchart illustrating the energy storage system fault handling method provided in this application embodiment. Figure 3 In this embodiment, in Figure 2 Based on the provided embodiments, the energy storage system fault handling method is further explained. The energy storage system fault handling method includes:

[0136] S401. Generate a collaborative optimization strategy based on the battery health status, remaining lifespan, and comprehensive fault evaluation results.

[0137] S402. Optimize the parameters of the prediction model and the fusion model according to the collaborative optimization strategy.

[0138] S403. Send the collaborative optimization strategy to the sensor layer and the edge computing layer; wherein, the sensor layer is used to optimize multiple preset data acquisition parameters according to the collaborative optimization strategy, and the edge computing layer is used to optimize multiple preset fault detection parameters in fault diagnosis according to the collaborative optimization strategy.

[0139] S401-S403 are similar to S305-S307, and will not be described again in this embodiment.

[0140] S404. Obtain strategy execution feedback data and system spatiotemporal information; wherein, strategy execution feedback data includes parameter optimization data of the sensor layer after executing the collaborative optimization strategy, fault diagnosis results of the edge computing layer after executing the collaborative optimization strategy, and historical fault handling records of the energy storage system; and system spatiotemporal information includes deployment location information of multiple devices preset in the energy storage system.

[0141] Specifically, the cloud platform layer receives parameter optimization data generated by the sensor layer after executing the collaborative optimization strategy, and fault diagnosis results output by the edge computing layer after executing the collaborative optimization strategy. Simultaneously, it retrieves pre-stored historical handling records of the energy storage system, integrating them to form strategy execution feedback data. Then, it retrieves the deployment location information of multiple devices in the energy storage system from the cloud platform layer's preset storage module to obtain the system's spatiotemporal information. This step provides complete foundational data for subsequent mapping of the strategy execution feedback data to a preset spatiotemporal coordinate system, constructing a spatiotemporal correlation map, and inputting it into a causal analysis model for analysis.

[0142] S405. Based on the system's spatiotemporal information, the strategy execution feedback data is mapped to a preset spatiotemporal coordinate system to obtain a spatiotemporal correlation map.

[0143] Specifically, the process begins by parsing the deployment location information of multiple devices within the system's spatiotemporal information to determine the spatial dimension coordinates of a preset spatiotemporal coordinate system. Next, a corresponding time label is added to each item in the strategy execution feedback data, matching the time dimension coordinates of the coordinate system. Subsequently, the time-labeled feedback data is bound to the corresponding device spatial coordinates. Following preset graph construction rules, the data is located and associated within the spatiotemporal coordinate system, forming a spatiotemporal correlation graph that includes the correspondence between device location, time nodes, and feedback data. This step provides a structured and visualized data foundation for subsequently inputting the spatiotemporal correlation graph into a causal analysis model for analysis.

[0144] Among them, the spatiotemporal coordinate system is a three-dimensional or higher reference framework that integrates time and space dimensions for locating and associating events. Its core is to bind scattered physical location information with time nodes through unified coordinate rules, so that the data has clear attributes of "when and where it happened", in order to clearly present the spatiotemporal distribution characteristics and evolution logic of events.

[0145] For example, taking a small-scale residential photovoltaic energy storage system as an example, the spatial coordinates of the energy storage battery pack in this system are (X1, Y1, Z1). When the sensor layer collects data at 15:10:00 and the battery pack temperature is slightly higher than the normal threshold, the edge computing layer determines through preliminary diagnosis that the preliminary diagnosis result is a slight heat dissipation abnormality. If the historical fault handling record shows that the battery pack had a similar situation one month ago, the spatiotemporal coordinate system will map this information to the coordinate points of ((X1, Y1, Z1), 15:10:00, temperature abnormality) and ((X1, Y1, Z1), one month ago, historical similar abnormality). This presents the fault correlation of the same device at different times, so as to quickly determine whether the current abnormality is a reproduction of the historical problem and provide intuitive data support for subsequent optimization of heat dissipation-related parameters.

[0146] S406. Input the spatiotemporal correlation map into the preset causal analysis model for analysis to obtain a fault analysis report; wherein, the fault analysis report is used to represent the causes and propagation paths of the energy storage system faults.

[0147] Specifically, the spatiotemporal correlation graph can be input into a pre-defined causal analysis model. This model consists of a graph analysis module, a correlation mining module, and a report generation module. The spatiotemporal correlation graph first enters the graph analysis module, which extracts core elements such as device location, time nodes, and feedback data from the graph and performs structured processing. The processed elements then flow into the correlation mining module, which mines the inherent relationships between feedback data from different locations and time nodes, tracing the evolution of corresponding events. The mined information is then sent to the report generation module, which generates an analysis report containing the event causes and evolution paths based on this information. The input to the entire causal analysis model is the spatiotemporal correlation graph, and the output is the analysis report. This step provides specific evidence of event evolution for subsequent parameter adjustments and strategy improvements in the energy storage system, supporting the implementation of related fault handling work.

[0148] The technical effect of this solution in this embodiment is as follows: By constructing a spatiotemporal tracing and causal analysis mechanism for the effect of strategy execution, it solves the problem in the prior art of lacking determination of the effectiveness of the strategy, the root cause of the fault, and the propagation law after fault handling. By integrating strategy feedback data and equipment spatiotemporal information to generate a correlation map and performing causal analysis, it can clearly reveal the inducing source, evolution path, and actual suppression effect of different optimization strategies of the fault event. This provides a data-driven decision-making basis for the accurate correction of the subsequent fault model and the targeted strengthening of the strategy, realizing a closed loop from execution-response to execution-analysis-optimization.

[0149] Figure 5 A flowchart illustrating the energy storage system fault handling method provided in this application embodiment. Figure 4 In this embodiment, in Figure 2 Based on the provided embodiments, the energy storage system fault handling method is further explained. The energy storage system fault handling method includes:

[0150] S501. In response to a fault level of Level 1, the preset system log is updated according to the fault type and urgency of the fault in the energy storage system, and a preset first alarm message is sent to the user terminal; wherein, the first alarm message is sent via push notification from the user terminal application.

[0151] S502. In response to the fault level being the second level, the control edge computing layer switches the energy storage system to a preset fault mitigation mode and sends a preset second alarm message to the user terminal; wherein, the second alarm message is sent via SMS, and the fault mitigation mode refers to the preset derated operation state of the energy storage system.

[0152] S503. In response to a fault level of Level 3, the control edge computing layer executes emergency protection actions for the energy storage system and sends a preset third alarm message to the user terminal. The third alarm message is sent via an automatic telephone call, and the emergency protection actions include cutting off the main power supply circuit of the energy storage system.

[0153] S501-S503 are similar to S2041-S2043, and will not be described again in this embodiment.

[0154] S504. Obtain user feedback behavior data on alarm signals, and calculate the average user response time and ignore rate for each fault type based on the feedback behavior data.

[0155] Specifically, the cloud platform layer can receive and aggregate various operation records from the user terminal after receiving an alarm signal, forming feedback behavior data. This data is then categorized according to different signal types. For each type, the time from receiving the signal to completing the feedback action is calculated and averaged. Simultaneously, the proportion of times no feedback was given is calculated as the ignore rate. This step provides real user behavior data support for subsequent dynamic adjustments to the urgency thresholds corresponding to different levels, ensuring that adjustments accurately reflect actual user responses.

[0156] S505. Based on the average user response time and ignore rate, dynamically adjust the urgency thresholds for triggering the first, second, and third levels to obtain the adjusted urgency thresholds.

[0157] Specifically, the cloud platform layer can analyze the matching degree between the current urgency thresholds corresponding to the first, second, and third levels and the actual user feedback by combining the calculated average user response time and ignore rate. For level categories with longer response times or higher ignore rates, the corresponding urgency threshold values ​​can be appropriately reduced, ultimately determining the adjusted urgency thresholds. This step provides a numerical basis for subsequently updating the mapping relationship between levels and target processing strategies based on the adjusted urgency thresholds and battery health status, adapting to actual user feedback.

[0158] S506. Based on the adjusted urgency threshold and battery health status, update the mapping relationship between fault level and target processing strategy, and send the updated mapping relationship to the edge computing layer.

[0159] Specifically, the adjusted urgency threshold can be combined with the current battery health status to redefine the state ranges corresponding to different levels. Based on this, the mapping between levels and target processing strategies is updated, and the updated mapping is then sent to the edge computing layer via a pre-defined communication link. This step allows the edge computing layer to quickly match subsequent processing strategies based on the latest mapping, ensuring that the system's subsequent processing flow aligns with the current state settings.

[0160] The technical effect of this solution in this embodiment is as follows: By introducing an adaptive optimization mechanism for alarm strategies based on user feedback behavior data, the problem of fixed alarm strategies and inability to be personalized according to users' actual response habits in the prior art is solved, which leads to reduced alarm effectiveness or user interference. By analyzing users' historical response data to different faults, the personalized fault level trigger threshold and response strategy mapping relationship are dynamically adjusted, so that the system alarms match the user's cognition and behavior patterns. This ensures that critical faults are responded to in a timely manner while reducing the frequency of unnecessary alarms, realizing the evolution of the fault alarm system from system preset to human-machine collaboration.

[0161] This application also provides a fault handling system, which includes a cloud platform layer, a sensor layer, and an edge computing layer.

[0162] The cloud platform layer is used to aggregate, store, and deeply analyze all data collected by the sensor layer and uploaded by the edge computing layer. Through intelligent diagnostic algorithms, such as time series analysis, transfer learning, and attention mechanism networks, it can evaluate battery health status, predict faults, and analyze trends. At the same time, it can generate intelligent reminders and push them to the user end. It can also build a virtual mapping of the system based on digital twin technology and dynamically optimize the diagnostic model and alarm rules in combination with user feedback. Finally, the optimized model and rules are distributed to the edge computing layer to support the global management and iterative upgrade of the entire fault handling system.

[0163] The sensor layer is used to deploy various types of sensing devices to collect multi-source operating data of the energy storage system in real time, including core status parameters such as battery voltage, current, and temperature, as well as security monitoring data such as ambient temperature and humidity and smoke concentration. The collected raw data is synchronously transmitted to the edge computing layer to provide accurate and continuous basic sensing support for subsequent fault diagnosis, fault evaluation, and linkage protection.

[0164] The edge computing layer is used to preprocess the raw data transmitted from the sensor layer. It enables real-time fault diagnosis and local multi-level alarm triggering through lightweight intelligent models and rule engines. When a fault is detected, it can directly execute safety control actions, such as derated operation and cutting off the main power supply. At the same time, it uploads the processed data to the cloud platform layer and receives optimized models and rules from the cloud, realizing efficient linkage between local handling and cloud collaboration, and ensuring the real-time response and safe operation of the energy storage system.

[0165] The core of the energy storage system fault handling method provided in this application lies in the distributed diagnostic architecture, multimodal perception fusion, intelligent early warning and reminder mechanism, and integrated security linkage. Figure 6 The cloud-edge-device collaborative operation architecture diagram of the energy storage system provided in the embodiments of this application is as follows: Figure 6 As shown, the process of this architecture is as follows:

[0166] Sensor layer (end): Responsible for collecting data from multiple sources, including:

[0167] Cell data: voltage, current, temperature, internal resistance.

[0168] Power Conversion System (PCS) data: input / output voltage and current, frequency, and insulation resistance.

[0169] Environmental data: cabin temperature, humidity, smoke, water immersion, and concentration of combustible gases.

[0170] Security data: equipment displacement and vibration, door opening and closing status, perimeter intrusion detection (such as infrared) signals.

[0171] Edge computing layer (edge):

[0172] Local real-time diagnostics: Deploy lightweight AI algorithms (such as optimized FFRLS for online internal resistance identification) to perform millisecond-level rapid judgment and response to critical faults (such as internal short circuits and thermal runaway tendencies).

[0173] Execute local control policies: Based on diagnostic results or cloud instructions, immediately execute protective actions, such as disconnecting the load branch box, activating fire alarm linkage, and triggering audible and visual alarms.

[0174] Cloud platform layer (cloud):

[0175] Deep analytics and machine learning: Receive data uploaded from the edge layer and run more complex predictive maintenance algorithms (such as Bi-LSTM-based fault prediction and battery health status evaluation). The trained model can be distributed to edge nodes to update their local algorithms, achieving "cloud-edge collaboration".

[0176] Intelligent Alert Engine: Managing the alarm alert logic throughout the entire process is the core of the "Security Guardian" concept.

[0177] Existing technologies rely solely on cloud-based analysis, resulting in high latency, or rely on local threshold judgments, leading to high false alarm rates. This application addresses this issue at the edge: employing an improved Forget Factor Recursive Least Squares (FFRLS) method for online real-time estimation of cell internal resistance. Internal resistance is a sensitive indicator of battery state of health (SOH) and early faults (such as micro-short circuits). The algorithm formula is as follows:

[0178]

[0179] in, These are the estimated battery parameters at time k. The estimated values ​​of the battery parameters for the (k-1)th battery are as follows: Let be the gain matrix at time k. The battery system measurement output at time k. Let be the regression vector at time k.

[0180] By introducing a forgetting factor (Usually taken as 0.95-0.99), which makes the algorithm give higher weight to the latest data and more responsive to the time-varying characteristics of battery parameters.

[0181] On the cloud side: a Bi-LSTM with Attention network is used to deeply mine massive historical time-series data to predict battery state of health (SOH) and remaining lifespan (RUL) and achieve predictive maintenance.

[0182] Advantages: Bi-LSTM can simultaneously capture past and future information from time series, and the Attention mechanism can focus on key fault characteristic periods, improving prediction accuracy and interpretability. It overcomes the limitations of single-level diagnosis, achieving a unification of rapid edge response and deep cloud insights.

[0183] Existing technologies for fault diagnosis and security monitoring are isolated, with data not being shared. This application introduces multi-sensor data fusion technology (such as the DS evidence algorithm) to comprehensively determine faults. For example, when diagnosing a "fire fault," it simultaneously integrates three pieces of evidence: "sudden rise in battery temperature," "excessive smoke concentration," and "increased concentration of combustible gas in the cabin." This provides higher confidence than a single sensor judgment and significantly reduces the false alarm rate. Security linkage: When both displacement and infrared sensors detect an anomaly, it is determined to be a highly probable theft fault, immediately triggering a local audible and visual alarm and sending a remote alarm to the user. This deep integration of traditional fault diagnosis and physical security monitoring truly achieves integrated protection as a "security guardian."

[0184] Existing technology alarm methods are too simplistic and easily overlooked by users. Figure 7 The flowchart of the multi-level alarm and fault handling closed loop of the energy storage system provided in the embodiments of this application is as follows: Figure 7 As shown, the architecture's process is as follows: When a fault is triggered, it is first confirmed by edge or cloud diagnostics, and then an initial alarm is issued via APP push or SMS; if it is not handled within the set time limit, a second alarm is triggered by a phone call or strong APP reminder; if it is still not handled within a longer period of time, the alarm will be escalated and transferred to the background for manual intervention, ultimately forming a closed loop of handling; if the handling is completed at any stage, the alarm is cleared, thereby realizing hierarchical response and closed-loop management of fault events, ensuring that problems are followed up and resolved in a timely manner.

[0185] Existing technologies often employ single-mode diagnostics, either cloud-based or local, which struggles to balance real-time performance with in-depth analysis. This application addresses this by employing a cloud-edge-device collaborative architecture, achieving a balance between real-time edge response and in-depth cloud analysis. Existing technologies rely on threshold methods for diagnostic accuracy, resulting in high false alarm rates and a lack of predictive capabilities. This application significantly reduces false alarm rates and enables predictive maintenance through multimodal data fusion and intelligent algorithms. In existing technologies, fault monitoring and security systems are typically independent; this application deeply integrates them, achieving coordinated protection for both environmental and equipment operational safety. Existing technologies suffer from limited alarm methods, are easily overlooked, and lack follow-up mechanisms. This application forms a closed-loop alarm processing system through intelligent multi-level alerts and escalated alarms, significantly improving user experience and security. Existing technologies have fixed functions and are difficult to update; this application, with its open API and rule engine, supports third-party function extensions and custom alarm rules, better adapting to future needs.

[0186] Figure 8 This is a schematic diagram of the structure of the energy storage system fault handling device provided in the embodiments of this application. Figure 8 As shown, the energy storage system fault handling device includes:

[0187] The first acquisition module 801 is used to acquire preliminary diagnostic results, multi-dimensional data, and multiple historical time series data. The preliminary diagnostic results are obtained by the edge computing layer based on the multi-dimensional data for fault diagnosis. The multi-dimensional data is collected by the sensor layer. The multiple historical time series data are used to represent the operating status of the preset energy storage system within a preset first time period. The end time of the first time period is earlier than the current time. The multi-dimensional data are used to represent the current operating status of the energy storage system and its environment.

[0188] The prediction module 802 is used to predict the battery health status and remaining life based on a preset prediction model and multiple historical time series data; wherein, the energy storage system includes a battery.

[0189] The fusion module 803 is used to fuse the preliminary diagnostic results and multi-dimensional data based on a preset fusion model to obtain a comprehensive fault evaluation result; wherein, the comprehensive fault evaluation result is used to represent the fault type and urgency of the fault in the energy storage system.

[0190] The determination module 804 is used to determine the fault level based on the battery health status, remaining life and comprehensive fault evaluation results, and trigger the target processing strategy corresponding to the fault level from a set of preset processing strategies; wherein, the target processing strategy is used to instruct the sensor layer and the edge computing layer to perform the response action corresponding to the fault level and send the alarm signal corresponding to the fault level to the preset user terminal.

[0191] In one possible design, the energy storage system fault handling device also includes:

[0192] The generation module is used to generate collaborative optimization strategies based on battery health status, remaining lifespan, and comprehensive fault evaluation results.

[0193] The optimization module is used to optimize the parameters of the prediction model and the fusion model according to the collaborative optimization strategy.

[0194] The sending module is used to send the collaborative optimization strategy to the sensor layer and the edge computing layer. The sensor layer is used to optimize multiple preset data acquisition parameters according to the collaborative optimization strategy, and the edge computing layer is used to optimize multiple preset fault detection parameters in fault diagnosis according to the collaborative optimization strategy.

[0195] In one possible design, the energy storage system fault handling device also includes:

[0196] The second acquisition module is used to acquire strategy execution feedback data and system spatiotemporal information. The strategy execution feedback data includes parameter optimization data of the sensor layer after executing the collaborative optimization strategy, fault diagnosis results of the edge computing layer after executing the collaborative optimization strategy, and historical fault handling records of the energy storage system. The system spatiotemporal information includes the deployment location information of multiple devices preset in the energy storage system.

[0197] The mapping module is used to map the strategy execution feedback data to a preset spatiotemporal coordinate system based on the system's spatiotemporal information, thereby obtaining a spatiotemporal correlation map.

[0198] The analysis module is used to input the spatiotemporal correlation map into a preset causal analysis model for analysis and to obtain a fault analysis report. The fault analysis report is used to represent the causes and propagation paths of the energy storage system's faults.

[0199] In one possible design, the fault levels include a first level, a second level, and a third level. The determination module 804 includes:

[0200] The first update unit is used to update the preset system log in response to a fault level of Level 1, based on the fault type and urgency of the fault in the energy storage system, and to send a preset first alarm message to the user terminal; wherein, the first alarm message is sent via push notification from the user terminal application.

[0201] The switching unit is used to control the edge computing layer to switch the energy storage system to a preset fault mitigation mode in response to a fault level of level 2, and to send a preset second alarm message to the user terminal; wherein, the second alarm message is sent via SMS, and the fault mitigation mode refers to the preset derated operation state of the energy storage system.

[0202] The execution unit is used to respond to a fault level of level 3, control the edge computing layer to perform emergency protection actions for the energy storage system, and send a preset third alarm message to the user terminal; wherein, the third alarm message is sent via an automatic telephone call, and the emergency protection actions include cutting off the main power supply circuit of the energy storage system.

[0203] In one possible design, module 804 also includes:

[0204] The first acquisition unit is used to acquire user feedback behavior data on alarm signals, and calculate the average user response time and ignore rate for each fault type based on the feedback behavior data.

[0205] The adjustment unit is used to dynamically adjust the urgency thresholds for triggering the first, second, and third levels based on the user's average response time and ignore rate, thus obtaining the adjusted urgency thresholds.

[0206] The second update unit is used to update the mapping relationship between fault level and target processing strategy based on the adjusted urgency threshold and battery health status, and send the updated mapping relationship to the edge computing layer.

[0207] In one possible design, the sensor layer includes multiple types of sensors, and the fusion module 803 includes:

[0208] The second acquisition unit is used to acquire historical data and data accuracy labels of various types of sensors; wherein, the data accuracy labels are used to represent the measurement error of each type of sensor.

[0209] The calculation unit is used to calculate the sensor accuracy corresponding to each dimension of the multi-dimensional data based on the historical data collected by each type of sensor and the data accuracy label, and to assign weight coefficients to each dimension of the data according to the accuracy.

[0210] The filtering unit is used to obtain the confidence threshold of historical diagnostic cases corresponding to the fault type, and to filter the multi-dimensional data according to the weight coefficient and the confidence threshold to obtain an effective data set.

[0211] The fusion unit is used to fuse the preliminary diagnostic results and the effective data set based on the fusion model to obtain a comprehensive fault evaluation result.

[0212] The energy storage system fault handling device provided in this embodiment can perform... Figures 2 to 5 The technical solution of the embodiment of the energy storage system fault handling method shown herein, its implementation principle and technical effect are similar to Figures 2 to 5 The embodiment of the energy storage system fault handling method shown is similar and will not be described in detail here.

[0213] Figure 9This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of this application. Figure 9 As shown, the electronic device 90 includes at least one processor 901 and a memory 902. The electronic device 90 also includes a communication component 903. The processor 901, memory 902, and communication component 903 are connected via a bus 904.

[0214] In a specific implementation, at least one processor 901 executes computer execution instructions stored in memory 902, causing at least one processor 901 to implement a fault handling method for an energy storage system according to the above embodiment.

[0215] The specific implementation process of processor 901 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0216] In the above embodiments, it should be understood that the processor 901 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0217] The memory 902 may include high-speed RAM memory, and may also include non-volatile memory (NVM), such as at least one disk storage.

[0218] Bus 904 can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Bus 904 can be divided into address bus, data bus, control bus, etc. For ease of illustration, the bus 904 in the accompanying drawings of this application is not limited to only one bus or one type of bus.

[0219] The above description of the functions implemented by electronic devices and main control devices has introduced the solutions provided by the embodiments of the present invention. It is understood that, in order to implement the above functions, the electronic device or main control device includes hardware structures and / or software modules corresponding to the execution of each function. By combining the units and algorithm steps of the various examples described in the embodiments of the present invention, the embodiments of the present invention can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the technical solutions of the embodiments of the present invention.

[0220] This application also provides a computer-readable storage medium storing computer-executable instructions. When executed by a processor, these instructions are used to implement a fault handling method for an energy storage system as described in the above embodiments. In the specific implementation of the aforementioned fault handling method for an energy storage system, each module can be implemented as a processor.

[0221] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0222] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Of course, the readable storage medium can also be a component of the processor. The processor and the readable storage medium can reside in application-specific integrated circuits (ASICs). Alternatively, the processor and the readable storage medium can exist as discrete components in an electronic device or a host device.

[0223] This application also provides a computer program product, including a computer program, which, when executed by a processor, is used to implement a fault handling method for an energy storage system as described in the above embodiments.

[0224] The computer program is stored in a readable storage medium, and at least one processor can read the computer program from the readable storage medium and execute the computer program to perform the scheme provided in any of the above embodiments.

[0225] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0226] The technical solutions of this application have been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it is readily understood by those skilled in the art that the scope of protection of this application is obviously not limited to these specific embodiments. The above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A fault handling method for an energy storage system, characterized in that, A cloud platform layer applied to a fault handling system, the fault handling system further including a sensor layer and an edge computing layer, the method comprising: The system acquires preliminary diagnostic results, multi-dimensional data, and multiple historical time-series data. The preliminary diagnostic results are obtained by the edge computing layer based on the multi-dimensional data, which is collected by the sensor layer. The multiple historical time-series data are used to represent the operating status of the energy storage system within a preset first time period, where the end time of the first time period is earlier than the current time. The multi-dimensional data are used to represent the current operating status of the energy storage system and its environment. The battery health status and remaining lifespan are predicted based on a preset prediction model and the multiple historical time-series data; wherein, the energy storage system includes the battery; The preliminary diagnostic results and the multi-dimensional data are fused based on a preset fusion model to obtain a comprehensive fault evaluation result; wherein, the comprehensive fault evaluation result is used to represent the fault type and urgency of the fault in the energy storage system; Based on the battery health status, remaining lifespan, and comprehensive fault evaluation results, a fault level is determined, and a target processing strategy corresponding to the fault level is triggered from among a plurality of preset processing strategies; wherein, the target processing strategy is used to instruct the sensor layer and the edge computing layer to perform a response action corresponding to the fault level and send an alarm signal corresponding to the fault level to a preset user terminal; After triggering the target processing strategy corresponding to the fault level among the preset multiple processing strategies, the method further includes: Based on the battery health status, remaining lifespan, and comprehensive fault evaluation results, a collaborative optimization strategy is generated. The parameters of the prediction model and the fusion model are optimized according to the collaborative optimization strategy. The collaborative optimization strategy is sent to the sensor layer and the edge computing layer; wherein, the sensor layer is used to optimize multiple preset data acquisition parameters according to the collaborative optimization strategy, and the edge computing layer is used to optimize multiple preset fault detection parameters in the fault diagnosis according to the collaborative optimization strategy; Acquire strategy execution feedback data and system spatiotemporal information; wherein, the strategy execution feedback data includes parameter optimization data of the sensor layer after executing the collaborative optimization strategy, fault diagnosis results of the edge computing layer after executing the collaborative optimization strategy, and historical fault handling records of the energy storage system; the system spatiotemporal information includes deployment location information of multiple devices preset in the energy storage system; Based on the spatiotemporal information of the system, the strategy execution feedback data is mapped to a preset spatiotemporal coordinate system to obtain a spatiotemporal correlation map; The spatiotemporal correlation map is input into a preset causal analysis model for analysis to obtain a fault analysis report; wherein, the fault analysis report is used to represent the causes and propagation paths of the faults in the energy storage system.

2. The energy storage system fault handling method according to claim 1, characterized in that, The fault levels include a first level, a second level, and a third level. The target processing strategy corresponding to the fault level among the multiple preset processing strategies includes: In response to the fault level being the first level, the preset system log is updated according to the fault type and urgency of the fault in the energy storage system, and a preset first alarm message is sent to the user terminal; wherein, the first alarm message is sent via application push on the user terminal. In response to the fault level being the second level, the edge computing layer is controlled to switch the energy storage system to a preset fault mitigation mode and send a preset second alarm message to the user terminal; wherein, the second alarm message is sent via SMS, and the fault mitigation mode refers to the preset derated operation state of the energy storage system; In response to the fault level being the third level, the edge computing layer is controlled to perform emergency protection actions for the energy storage system and send a preset third alarm message to the user terminal; wherein, the third alarm message is sent via an automatic telephone call, and the emergency protection action includes cutting off the main power circuit of the energy storage system.

3. The energy storage system fault handling method according to claim 2, characterized in that, After sending the preset third alarm information to the user terminal, the method further includes: Obtain the user's feedback behavior data to the alarm signal, and calculate the average user response time and ignore rate for the fault type based on the feedback behavior data; Based on the average user response time and the ignore rate, the urgency thresholds for triggering the first, second, and third levels are dynamically adjusted to obtain the adjusted urgency thresholds. Based on the adjusted urgency threshold and the battery health status, the mapping relationship between the fault level and the target processing strategy is updated, and the updated mapping relationship is sent to the edge computing layer.

4. The energy storage system fault handling method according to claim 1, characterized in that, The sensor layer includes multiple types of sensors. The preliminary diagnostic results and the multi-dimensional data are fused based on a preset fusion model to obtain a comprehensive fault evaluation result, including: Obtain historical acquisition data and data accuracy labels for each type of sensor; wherein, the data accuracy labels are used to represent the measurement error of each type of sensor; Based on the historical data collected by each type of sensor and the data accuracy label, calculate the sensor accuracy corresponding to each dimension of the multi-dimensional data, and assign weight coefficients to each dimension of the data according to the accuracy. Obtain the confidence threshold of historical diagnostic cases corresponding to the fault type, and filter the multi-dimensional data according to the weight coefficient and the confidence threshold to obtain an effective data set; The preliminary diagnostic results and the effective data set are fused based on the fusion model to obtain the comprehensive fault evaluation result.

5. A fault handling device for an energy storage system, characterized in that, A cloud platform layer applied to a fault handling system, the fault handling system further including a sensor layer and an edge computing layer, the device comprising: The first acquisition module is used to acquire preliminary diagnostic results, multi-dimensional data, and multiple historical time-series data. The preliminary diagnostic results are obtained by the edge computing layer through fault diagnosis based on the multi-dimensional data. The multi-dimensional data is collected by the sensor layer. The multiple historical time-series data are used to represent the operating status of the energy storage system within a preset first time period. The end time of the first time period is earlier than the current time. The multi-dimensional data are used to represent the current operating status of the energy storage system and its environment. A prediction module is used to predict the battery health status and remaining lifespan based on a preset prediction model and the multiple historical time-series data; wherein, the energy storage system includes the battery; The fusion module is used to fuse the preliminary diagnostic results and the multi-dimensional data based on a preset fusion model to obtain a comprehensive fault evaluation result; wherein, the comprehensive fault evaluation result is used to represent the fault type and urgency of the fault in the energy storage system; The determination module is used to determine the fault level based on the battery health status, the remaining lifespan, and the comprehensive fault evaluation result, and to trigger a target processing strategy corresponding to the fault level from a set of preset processing strategies; wherein, the target processing strategy is used to instruct the sensor layer and the edge computing layer to perform a response action corresponding to the fault level and send an alarm signal corresponding to the fault level to a preset user terminal; The generation module is used to generate a collaborative optimization strategy based on the battery health status, the remaining lifespan, and the comprehensive fault evaluation results. The optimization module is used to optimize the parameters of the prediction model and the fusion model according to the collaborative optimization strategy; A sending module is used to send the collaborative optimization strategy to the sensor layer and the edge computing layer; wherein, the sensor layer is used to optimize multiple preset data acquisition parameters according to the collaborative optimization strategy, and the edge computing layer is used to optimize multiple preset fault detection parameters in the fault diagnosis according to the collaborative optimization strategy; The second acquisition module is used to acquire strategy execution feedback data and system spatiotemporal information; wherein, the strategy execution feedback data includes parameter optimization data of the sensor layer after executing the collaborative optimization strategy, fault diagnosis results of the edge computing layer after executing the collaborative optimization strategy, and historical fault handling records of the energy storage system; and the system spatiotemporal information includes deployment location information of multiple devices preset in the energy storage system. The mapping module is used to map the strategy execution feedback data to a preset spatiotemporal coordinate system based on the spatiotemporal information of the system, so as to obtain a spatiotemporal correlation map; The analysis module is used to input the spatiotemporal correlation map into a preset causal analysis model for analysis and obtain a fault analysis report; wherein, the fault analysis report is used to represent the causes and propagation paths of the faults in the energy storage system.

6. An electronic device, characterized in that, include: A processor, and a memory communicatively connected to the processor; The memory stores computer-executed instructions; When the processor executes the computer execution instructions stored in the memory, it is used to implement the energy storage system fault handling method as described in any one of claims 1 to 4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the energy storage system fault handling method as described in any one of claims 1 to 4.

8. A computer program product, characterized in that, Includes a computer program, which, when executed by a processor, is used to implement the energy storage system fault handling method as described in any one of claims 1 to 4.

Images