Energy storage device processing method and intelligent diagnosis system

Abstract

The application discloses a kind of processing method and intelligent diagnosis system of energy storage equipment.The method comprises: at least one exception detection is carried out to first data by first engine, generates first content, the first engine includes the distributed real-time stream processing engine associated with first energy storage equipment, and the first data includes the real-time running data related to the first energy storage equipment reported using distributed message queue;In response to the deterioration trend that first content indicates the abnormal operation of the first energy storage equipment occurs, generate first description information based on the first content, call intelligent system based on first description information output second content, the intelligent system is used to execute fault diagnosis processing task to energy storage equipment, and the second content includes fault diagnosis result and fault processing suggestion.The application scheme solves the problem that fault diagnosis exists obvious lag and exists poor fault identification effect.

Description

Technical Field

[0001] This invention relates to the field of new energy technology, and in particular to a processing method and intelligent diagnostic system for energy storage devices. Background Technology

[0002] The data processing architecture of energy storage devices relies on the distributed IoT message server EMQX and the distributed message queue Kafka to complete data flow. It can also be paired with the time-series metric monitoring system Prometheus and the alarm management component Alertmanager to achieve metric monitoring and alarms. However, the above solution relies on fixed thresholds to trigger alarms, which can only identify anomalies that have occurred, resulting in significant lag in fault diagnosis and poor fault identification effectiveness. Summary of the Invention

[0003] This invention provides a processing method and intelligent diagnostic system for energy storage devices to solve the problems of significant lag in fault diagnosis and poor fault identification.

[0004] According to one aspect of the present invention, a method for processing energy storage devices is provided, the method comprising: The first engine performs at least one anomaly detection on the first data to generate the first content. The first engine includes a distributed real-time stream processing engine associated with the first energy storage device. The first data includes real-time operating data related to the first energy storage device reported using a distributed message queue. The system responds to a first content indicating that the first energy storage device is showing a deterioration trend of abnormal operation. Based on the first content, it generates first description information and calls an intelligent system to output second content based on the first description information. The intelligent system is used to perform fault diagnosis and processing tasks on the energy storage device. The second content includes fault diagnosis results and fault processing suggestions.

[0005] According to another aspect of the present invention, an intelligent diagnostic system is provided to implement the processing method of the energy storage device described in any embodiment of the present invention. The intelligent diagnostic system includes at least one energy storage device and is configured to perform the following operations: The first engine performs at least one anomaly detection on the first data to generate the first content. The first engine includes a distributed real-time stream processing engine associated with the first energy storage device. The first data includes real-time operating data related to the first energy storage device reported using a distributed message queue. The system responds to a first content indicating that the first energy storage device is showing a deterioration trend of abnormal operation. Based on the first content, it generates first description information and calls an intelligent system to output second content based on the first description information. The intelligent system is used to perform fault diagnosis and processing tasks on the energy storage device. The second content includes fault diagnosis results and fault processing suggestions.

[0006] In the technical solution of this invention embodiment, a distributed real-time stream processing engine associated with the first energy storage device is adopted to perform at least one anomaly detection on the energy storage device operation data transmitted and reported via a distributed message queue. The distributed real-time stream processing engine is configured with parallel streaming computing power and can continuously parse the full runtime sequence data of the first energy storage device pushed by the message queue. Compared with offline data analysis, it can effectively solve the defects of offline analysis such as delayed fault alarms and easy omission of progressive deterioration faults. When the deterioration trend of abnormal operation of the first energy storage device is identified, first descriptive information can be constructed based on the first content generated by at least one anomaly detection. This can provide input for the intelligent diagnostic system and automatically drive the intelligent system to perform fault diagnosis and early warning, outputting fault judgment conclusions and fault handling solutions. In addition, relying on the intelligent system for reasoning analysis to replace experience judgment, while constraining the reasoning logic to ensure the accuracy and reliability of the diagnostic results, directly outputting practical maintenance suggestions, and significantly improving the fault operation and maintenance efficiency of energy storage devices.

[0007] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a schematic flowchart of a processing method for an energy storage device provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the architecture of an intelligent diagnostic system provided in an embodiment of the present invention; Figure 3 This is a full-link timing diagram for intelligent fault diagnosis provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of a retrieval-enhanced reasoning process provided in an embodiment of the present invention. Detailed Implementation

[0010] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0011] It should be understood that the steps described in the method may be performed in different orders and / or in parallel. Furthermore, the method may include additional steps and / or omit the steps shown. The scope of this document is not limited in this respect.

[0012] The term "comprising" and its variations as used herein are open-ended inclusions, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one situation" means "at least one situation"; the term "another situation" means "at least one additional situation"; the term "some situations" means "at least some situations". Definitions of other terms will be given in the following description.

[0013] It should be noted that the concepts of "first" and "second" mentioned are only used to distinguish different devices, modules or units, and are not used to limit the order of the functions performed by these devices, modules or units or their interdependencies.

[0014] It should be noted that the terms "one" and "more" used in this document are illustrative rather than restrictive, and those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0015] The names of messages or information exchanged between multiple devices in this document are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0016] It is understandable that before using the technical solutions disclosed in each scenario in this document, users should be informed of the type, scope of use, and usage scenarios of the personal information involved in this document in an appropriate manner in accordance with relevant laws and regulations, and user authorization should be obtained.

[0017] For example, upon receiving a user's active request, a prompt message is sent to the user to explicitly inform them that the requested operation will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the software or hardware, such as electronic devices, applications, servers, or storage media, that perform the operations described herein, based on the prompt message.

[0018] As an optional but non-limiting implementation, in response to a user's active request, sending a prompt message to the user can be done via a pop-up window, where the prompt message can be presented in text format. Furthermore, the pop-up window can also include a selection control allowing the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0019] It is understood that the above notification and user authorization process is merely illustrative and does not limit the implementation method described in this article. Other methods that comply with relevant laws and regulations may also be applied to the implementation method described in this article.

[0020] It is understood that the data involved in the technical solutions in this article (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and related provisions.

[0021] In one embodiment, Figure 1 This is a flowchart illustrating a processing method for an energy storage device according to an embodiment of the present invention. This embodiment is applicable to situations involving fault diagnosis and analysis of energy storage devices, particularly scenarios involving intelligent analysis, fault diagnosis, and predictive maintenance of time-series data of portable energy storage power supplies and home photovoltaic energy storage inverters. This method can be executed by an intelligent diagnostic system, which can be implemented in hardware and / or software and can be configured in electronic devices. Figure 1 As shown, the processing method of the energy storage device in this embodiment may include the following steps: S110. Perform at least one anomaly detection on the first data through the first engine to generate first content. The first engine includes a distributed real-time stream processing engine associated with the first energy storage device. The first data includes real-time operating data related to the first energy storage device reported using a distributed message queue.

[0022] See Figure 2 and Figure 3 The IoT backend is a service cluster within the IoT platform that operates in the background on servers or in the cloud, without user interaction. It receives, processes, and stores operational data uploaded by all energy storage devices. The IoT backend includes a device access layer for connecting to the primary energy storage devices, which include battery cells and energy storage inverters. These primary energy storage devices can report their real-time operational data via a distributed message queue. A message gateway based on the MQTT protocol receives this data and sends it to a distributed message queue, such as Kafka. By asynchronously reporting, caching, and sorting the real-time operational data through the distributed message queue, the primary energy storage devices can resolve peak data congestion issues and ensure stable access to massive amounts of data.

[0023] Optionally, the real-time operating data related to the first energy storage device may include measurement data from at least one of the following: the battery management system (BMS), the energy storage inverter, the temperature control system, and the battery cells. The real-time operating data related to the first energy storage device may include time-series shadow data related to at least one of the following: cell temperature difference, DC internal resistance, charging / discharging current, battery voltage, power, device operating condition, and firmware operating status. The first data may be real-time dynamic time-series data of the first energy storage device. The first data carries a timestamp, possesses continuity and time sequence, and is distinct from static configuration data and offline statistical data. This allows for timely detection of at least one anomaly in the first data, addressing issues such as insufficient computing power, data latency, packet loss, and congestion.

[0024] See Figure 2 and Figure 3 For the IoT backend, this solution can introduce a first engine, which may include a distributed real-time stream processing engine. This engine possesses high concurrency, low latency, and uninterrupted streaming computing capabilities; for example, it could include Flink, adaptable to scenarios involving the real-time processing of massive amounts of time-series data. Upon receiving real-time operational data related to the first energy storage device, the distributed real-time stream processing engine consumes this data from a distributed message queue.

[0025] The first data stream utilizes a distributed message queue to smooth out peak flows and decouple asynchronous operations from the real-time operational data related to the first energy storage device, thus avoiding data congestion and loss caused by a large volume of such data. Combined with a distributed real-time stream processing engine integrated with the first energy storage device, this significantly reduces computational pressure and ensures real-time computation of the device's operational data.

[0026] At least one anomaly detection may include threshold-based anomaly detection and degradation trend anomaly detection. At least one anomaly detection may include detecting whether the operating data of the first energy storage device exhibits a gradual trend of deviating from the baseline value of the operating data of the first energy storage device over time. That is, even if the operating data of the first energy storage device does not exceed the baseline value of the operating data of the first energy storage device, the time-series trend of the corresponding operating data of the first energy storage device shows a trend of continuous increase and is about to exceed the baseline value of the operating data of the first energy storage device.

[0027] Anomaly detection based on degradation trends differs from threshold-based anomaly detection, which only checks whether the operating data indicators of the primary energy storage device exceed a fixed threshold (e.g., alarming only when cell temperature > 65°C). While identifying sudden and severe faults, degradation trend detection does not rely on instantaneous value exceeding limits. Instead, it determines whether there is a gradual trend of deviating from the baseline operating data values ​​of the primary energy storage device over time, based on the rate of change, evolution pattern, and continuous deviation of indicators within a continuous time window. Examples include latent degradation in its early stages, such as a slow and continuous increase in temperature difference or a gradual rise in internal resistance. These parameters may not reach the alarm threshold throughout the process, but a clear degradation pattern exists over time. For instance, for gradual degradation trends in the primary energy storage device, such as a slow increase in cell internal resistance and a continuous expansion of battery temperature difference, without instantaneous alarms, it can compensate for the lag in traditional threshold-based alarms.

[0028] In the above-described solution, the architecture of a distributed message queue combined with a distributed real-time stream processing engine can handle high-concurrency access and real-time parallel computing of energy storage device-related operational data. This avoids the drawbacks of data analysis lag. Furthermore, the distributed real-time stream processing engine continuously consumes operational data from the distributed message queue and performs at least one anomaly detection on the real-time operational data related to the first energy storage device. This not only detects whether the instantaneous value of the operational data exceeds the alarm threshold, enabling rapid identification of sudden and obvious faults related to short circuits and over-temperature, but also continuously tracks the changing trend of the operational data over time, determining whether the operational data deviates unidirectionally and continuously from the baseline value. This allows for the capture of gradual degradation trends in energy storage devices, such as a slow increase in cell internal resistance and a gradual widening of cell temperature difference, before the alarm threshold is reached. This breaks through the limitations of traditional detection methods that rely solely on a single threshold judgment, adding the ability to identify time-series trends. It can provide early warning of hidden risks before the fault fully erupts, effectively improving the shortcomings of traditional monitoring methods that can only identify over-limit obvious faults and are prone to missing potential equipment degradation hazards in their early stages.

[0029] S120. In response to the first content indicating that the first energy storage device has a deterioration trend of abnormal operation, generate first description information based on the first content, and call the intelligent system to output second content based on the first description information. The intelligent system is used to perform fault diagnosis and processing tasks on the energy storage device. The second content includes fault diagnosis results and fault processing suggestions.

[0030] The deterioration trend of abnormal operation can include a continuous, unidirectional degradation of real-time operating data related to the first energy storage device for a preset duration, exhibiting irreversible and gradual degradation characteristics. See also Figure 2 and Figure 3When the first energy storage device exhibits a deterioration trend indicating abnormal operation, if the relevant real-time operating data does not exceed the alarm threshold, the alarm mechanism will not be triggered, indicating a latent fault risk. The deterioration trend can be triggered by performance degradation or potential faults in at least one component of the first energy storage device, such as the battery cells or inverter. When a trend of continuously widening temperature difference and slowly increasing internal resistance in the battery cells of the first energy storage device is detected, the diagnostic process will be automatically triggered even if the alarm threshold is not reached, providing early fault warning.

[0031] See Figure 2 and Figure 3 The first descriptive information can be prompt word information for the intelligent supply system generated by parsing and encapsulating the first content. The first content may include relevant features used to characterize the gradual degradation trend formed by the continuous shift of the real-time operating data related to the first energy storage device over time when the first energy storage device exhibits an abnormal operational degradation trend. The first content may also include the data identifier of the real-time operating data related to the first energy storage device, the offset of the real-time operating data related to the first energy storage device relative to the baseline value of the operating data related to the first energy storage device, and the duration of the gradual degradation of the real-time operating data related to the first energy storage device.

[0032] A progressive degradation trend can be described as the temporal variation pattern of real-time operating data related to the primary energy storage device deviating from the baseline value over time. This progressive degradation trend represents irreversible performance decline in the primary energy storage device, distinct from short-term fluctuations or sudden failures. It is a latent risk characteristic accumulated over a long period during the initial stages of a potential failure. This progressive degradation trend manifests as the instantaneous value of the real-time operating data related to the primary energy storage device not exceeding the alarm threshold, but the data exhibiting a continuous, unidirectional, and subtly deteriorating abnormal trend over time.

[0033] In some cases, at least one intelligent system can be pre-configured. An intelligent system can represent a system capable of performing table processing, invoking tools, or autonomous control based on machine learning models. For example, an intelligent system can make decisions and autonomously execute actions based on machine learning models to achieve preset goals or complete preset tasks—a virtual object or physical entity. An intelligent system can be an automated program that understands the user's intent and can utilize models or invoke tools to complete various types of tasks. In some cases, examples of intelligent systems may include, but are not limited to: agents, bots, chatbots, digital avatars, intelligent customer service, digital assistants, etc. Alternatively, an intelligent system can also be an intelligent role implemented based on a machine learning model. An "intelligent system" can process user requests based on generative models (e.g., language models, multimodal models) to perform specified types of tasks. In some cases, an intelligent system may also relate to virtual accounts, which may have corresponding avatars or nicknames.

[0034] Optionally, the fault diagnosis and processing task can be used to generate fault diagnosis results and corresponding fault handling suggestions. Generating fault diagnosis results may include at least one of the following: locating the faulty component (such as a battery cell or energy storage inverter in the first energy storage device) corresponding to the real-time operating data of the first energy storage device exhibiting a deterioration trend of abnormal operation; quantifying the degradation risk level corresponding to the real-time operating data of the first energy storage device exhibiting a deterioration trend of abnormal operation; or, deriving the cause of performance degradation of the faulty component corresponding to the real-time operating data of the first energy storage device exhibiting a deterioration trend of abnormal operation.

[0035] In the technical solution of this invention embodiment, a distributed real-time stream processing engine associated with the first energy storage device is adopted to perform at least one anomaly detection on the energy storage device operation data transmitted and reported via a distributed message queue. The distributed real-time stream processing engine is configured with parallel streaming computing power and can continuously parse the full runtime sequence data of the first energy storage device pushed by the message queue. Compared with offline data analysis, it can effectively solve the defects of offline analysis such as delayed fault alarms and easy omission of progressive deterioration faults. When the deterioration trend of abnormal operation of the first energy storage device is identified, first descriptive information can be constructed based on the first content generated by at least one anomaly detection. This can provide input for the intelligent diagnostic system and automatically drive the intelligent system to perform fault diagnosis and early warning, outputting fault judgment conclusions and fault handling solutions. In addition, relying on the intelligent system for reasoning analysis to replace experience judgment, while constraining the reasoning logic to ensure the accuracy and reliability of the diagnostic results, directly outputting practical maintenance suggestions, and significantly improving the fault operation and maintenance efficiency of energy storage devices.

[0036] Based on the above embodiments, variations of the above embodiments are proposed. It should be noted that, for the sake of brevity, the variations only describe the differences from the above embodiments. In one embodiment, generating first description information based on first content may include, but is not limited to, the following steps: At least one third content is determined from the first information based on the first content. The first information includes multiple fourth contents, which indicate various historical faults associated with the second energy storage device and corresponding maintenance plans. The first content matches the third content. The second energy storage device and the first energy storage device are of the same type, or the components in the first energy storage device and the components in the second energy storage device are of the same type. The multiple fourth contents are generated by processing the second data through a second engine, which includes a distributed batch offline data processing engine. The second data includes historical business data and historical log data associated with the second energy storage device. First descriptive information is generated based on at least one third content and the first content.

[0037] See Figure 2 and Figure 3 The fourth content can indicate historical faults associated with the second energy storage device and corresponding maintenance plans. Each fourth content corresponds to a primary equipment fault event of the second energy storage device. Among them, historical faults can include the deterioration characteristics, deteriorated components, and deterioration risk level when the second energy storage device exhibits an abnormal operation deterioration trend during a historical period. The maintenance plan corresponding to the historical fault can include the control parameters, inspection steps, and fault handling strategies used to repair the deteriorated components in the second energy storage device corresponding to the historical fault.

[0038] The second energy storage device can be identical to the first energy storage device in at least one of the following aspects: hardware model, cell specifications, inverter parameters, and control logic. The operational data related to the second energy storage device is consistent with the operational and fault attenuation patterns of the first energy storage device. Therefore, multiple fourth contents can be used to generate first information, thereby constructing a fault knowledge base used by the first energy storage device of the same model. This allows the faults recorded in the first information and the corresponding maintenance solutions to be reused for diagnostic and processing tasks on the first energy storage device. The first information can be stored in a vector database and a business database, providing matching historical fault samples for the execution of fault diagnosis and processing tasks. The first information can carry all historical faults of the same model of energy storage devices and corresponding maintenance solutions for those historical faults.

[0039] See Figure 2 and Figure 3The second engine may include a distributed batch offline data processing engine, such as Spark, which can be used to process historical business data and historical log data associated with the second energy storage device. Unlike the real-time streaming computing of the first engine, the second engine can periodically read historical business data and historical log data associated with the second energy storage device in batches, continuously iterating and updating the fourth content in the first information. The historical business data in the second data may include at least one of the following: fault work orders, maintenance records, equipment configuration, or charging / discharging condition files. The historical log data in the second data may include shadow time-series data of the second energy storage device, alarm log data, and operating logs of components in the second energy storage device, which may include energy storage inverters and battery cells.

[0040] See Figure 2 and Figure 3 The third content can be historical faults retrieved and filtered from multiple fourth contents included in the first information, matching the degradation trend indicated by the first content in terms of feature similarity. Each piece of third content can include historical faults and corresponding maintenance plans selected from multiple fourth contents. The third content can be used to provide historical faults matching the degradation trend of real-time operating data related to the first energy storage device, serving as a diagnostic description for the intelligent system to perform fault diagnosis and processing tasks. The first description information can integrate the first content and the third content, thereby ensuring that the first description information can include the detection results of at least one anomaly detection of the first data, as well as historical faults related to the first energy storage device and corresponding maintenance plans.

[0041] This solution preprocesses massive amounts of historical business and log data from the same type of energy storage equipment using an offline distributed batch engine to build a standardized historical fault knowledge base. After identifying the degradation trend of energy storage equipment in real time, it matches highly similar historical fault records of the same model based on real-time anomaly quantification features. It integrates real-time data with historical similar fault cases to generate diagnostic input descriptions. On the one hand, it uses offline computing to distribute the processing pressure of massive historical data, ensuring low latency and low computing power consumption for real-time diagnosis. On the other hand, it relies on the reuse of fault features of the same model equipment to supplement the fault reference basis for early hidden degradation trends, improve the accuracy of similar fault matching and diagnostic credibility, reduce reliance on the experience of senior operation and maintenance experts, and constrain the inference boundary of the large model from the input layer to reduce the output of false diagnostic conclusions.

[0042] Furthermore, this solution can deeply mine multi-source operation and maintenance data accumulated over a long period in energy storage operation and maintenance scenarios, such as historical fault cases, maintenance work orders, and abnormal equipment shadow trajectories. Through unified temporal feature extraction and standardized vectorization processing, the temporal feature vectors corresponding to various faults are persistently stored, constructing a fault knowledge base specifically for energy storage devices. When the first energy storage device exhibits an abnormal operation and deterioration trend, it automatically matches and recalls highly similar historical fault samples and corresponding standardized handling solutions to generate primary descriptive information. This allows historical operation and maintenance experience to be injected into the intelligent system for reasoning, changing the reliance on general knowledge for reasoning. By using analogy reasoning based on historical fault cases, fault diagnosis and handling are achieved, effectively addressing the technical shortcomings of scarce rare fault samples and high dependence on expert experience in energy storage scenarios. This significantly improves the identification accuracy of hidden and niche faults, effectively enhancing the intelligent diagnostic capabilities for unknown early faults in energy storage devices.

[0043] In one embodiment, optionally, determining at least one third content from the first information based on the first content may include determining at least one third content from the first information based on the first content through a retrieval-enhanced generation engine.

[0044] In one embodiment, optionally, see [link to relevant documentation]. Figure 2 , Figure 3 and Figure 4 Determining at least one third content from the first information based on the first content may include: vectorizing the first content to generate a first vector; reading the pre-constructed first information and traversing the feature vectors corresponding to multiple fourth contents in the first information; calculating the feature similarity between the first vector and the feature vectors corresponding to the fourth content, and selecting the fourth content with a feature similarity higher than a preset similarity threshold as the third content. After determining the third content from multiple fourth contents, the method further includes at least one of the following: filtering out third content that is inconsistent with the model of the first energy storage device; or, reordering the multiple third contents after filtering according to feature similarity from high to low, and selecting the first number of third contents that are ranked first.

[0045] In the above-mentioned solution, the failure attenuation patterns of energy storage devices of the same model are consistent, and the historical failures retrieved and matched have high reference value, solving the problem of scarce failure samples and lack of diagnostic basis for energy storage devices; furthermore, relying on vector quantization matching, historical failures that match the degradation trend indicated by the first content can be accurately identified, which can improve the accuracy of failure recall.

[0046] In one embodiment, optionally, see [link to relevant documentation]. Figure 2 , Figure 3 and Figure 4 Determining at least one third piece of content from the first information based on the first content may include the following steps: The first content is vectorized by the third engine to generate the first vector. The third engine is used to extract the temporal change features in the first data and encode them into vectors. Based on the first vector, similarity detection is performed on multiple fourth contents of the first information to obtain the third content. The multiple fourth contents are stored in vector form.

[0047] The third engine can be an embedded engine or a vectorized engine. This engine can extract key features from unstructured or semi-structured data and map them into a fixed-dimensional, unified numerical space, floating-point high-dimensional array, enabling digital representation of different types of data. The embedded engine or vectorized engine can be an engine adapted to the operational data related to the cells or inverters in the first energy storage device. When the real-time operational data of the first energy storage device is the real-time operational data of the cells or inverters, the first content can include the temporal variation patterns of cell temperature difference, DC internal resistance, and inverter power that continuously deviate from the baseline values ​​of the operational data related to the first energy storage device over time. For example, it can include the gradual variation patterns related to unidirectional increase in cell parameters, continuous expansion of temperature difference, and slow power decay. In this case, the embedded engine or vectorized engine can be used to extract continuous temporal variation features from the first content, complete normalization, and encode them to generate a first vector representing the abnormal degradation trend.

[0048] Optionally, multiple fourth contents can be persistently stored in a vector database using high-dimensional floating-point vector arrays, such as a Milvus-like vector database, instead of the original time-series text. The vector database used to store multiple fourth contents can be configured with an approximate nearest neighbor retrieval algorithm, enabling rapid batch calculation of distances between vectors and achieving fast similarity matching. Similarity detection can include calculating the spatial distance between the first vector and the corresponding feature vector of the fourth contents using at least one vector distance algorithm, such as cosine distance or Euclidean distance; a smaller distance indicates a more similar fault time-series trend. Similarity detection can quantify the degree of matching between the degradation trend indicated by the first content and historical faults in the fourth content, thereby filtering highly similar historical faults as auxiliary references for the intelligent system.

[0049] In the above-mentioned solution, the degradation trend and related features are extracted by a vectorization engine and encoded into a unified vector. High-speed similarity matching of historical faults is completed by relying on pre-stored historical faults. This can solve the problems of difficulty in quickly comparing the operating data of energy storage equipment and difficulty in finding similar cases of early hidden trend faults. It can also provide objective and quantifiable historical faults for fault diagnosis, thereby improving the accuracy of fault diagnosis and providing reliable data support for intelligent diagnosis of early degradation of energy storage equipment.

[0050] In one embodiment, optionally, see [link to relevant documentation]. Figure 2 , Figure 3 and Figure 4 Determining at least one third content from the first information based on the first content may include the following steps: obtaining the third content by performing a keyword search on multiple fifth contents of the first information based on the first content, wherein the multiple fifth contents are determined based on at least one of the historical fault case documents, business rules, historical fault diagnosis reports, or question-and-answer knowledge bases of common problems associated with the second energy storage device; obtaining the third content by searching on multiple sixth contents of the first information based on the first content, wherein the multiple sixth contents are constructed and generated based on historical business data and historical log data associated with the second energy storage device.

[0051] The fifth content may include text fragments obtained by parsing and splitting at least one type of content from the historical fault case documents, business rules, historical fault diagnosis reports, or question-and-answer knowledge bases for common problems corresponding to the second energy storage device. The fault causes, fault handling solutions, hardware constraint rules, and historical maintenance experience stored in the fifth content can be used for precise keyword matching.

[0052] The second energy storage device is the same model as the first energy storage device, or the components included in the first energy storage device are of the same type as those included in the second energy storage device, and the first and second energy storage devices can share common fault knowledge. The second energy storage device and the first energy storage device belong to the same device type, or the second energy storage device and the first energy storage device have the same type of hardware architecture, the same type of battery management system, or the same type of battery cells or inverter components.

[0053] The first content is parsed to extract keywords. These keywords may include those related to the battery cells and energy storage inverters in the first energy storage device. Key detection related to the battery cells in the first energy storage device may include the temperature difference and internal resistance of the cells. Key detection related to the energy storage inverters in the first energy storage device may include the power attenuation of the inverters. A search query is generated by combining the keywords extracted from the first content. Based on the search query, keyword matching is performed in a text knowledge base composed of multiple fifth contents. Text fragments with keyword matching scores higher than a preset matching score threshold are selected to generate the third content.

[0054] In the above-described solution, keyword retrieval can quickly match relevant rules governing the hardware constraints of energy storage devices. This can constrain the output logic of the intelligent system, thereby avoiding the generation of disposal solutions that do not conform to the design of the energy storage devices. Moreover, the keyword retrieval process can reuse the knowledge base of a second type of energy storage device, compensating for the scarcity of fault samples in the first energy storage device, expanding the available operational experience, and improving the completeness of early degradation fault assessment. This deterministic rule-based rapid recall based on keyword matching has low retrieval computational overhead and fast response speed. Furthermore, the output third content serves as reasoning evidence, providing traceable basis for the subsequent intelligent system to execute diagnostic and processing tasks, reducing the probability of the model outputting non-compliant disposal solutions. This ensures that every step of reasoning in the intelligent system is based on evidence, improving the compliance of diagnostic and processing tasks.

[0055] Based on the above embodiments, variations of the embodiments are proposed. It should be noted that, for the sake of brevity, the variations only describe the differences from the above embodiments. In one embodiment, see [link to embodiment]. Figure 2 , Figure 3 and Figure 4 Generating first description information based on at least one third piece of content and the first content may include the following steps: The seventh content is obtained by performing a first feature extraction on the first data based on the feature calculation engine. The first feature extraction includes at least one of time domain feature extraction, frequency domain feature extraction, or time window feature extraction. First description information is generated based on the seventh content, at least one third content, and the first content.

[0056] The feature calculation engine can be used for at least one of time-domain feature extraction, frequency-domain feature extraction, or time-window feature extraction. It can uniformly complete feature extraction of the first data and provide general time-domain, frequency-domain, and window feature operators. Specifically, time-domain feature extraction can include calculating at least one of the mean, standard deviation, range, skewness, maximum difference, or minimum difference of the real-time operating data related to the first energy storage device. Frequency-domain feature extraction can include calculating at least one of the spectral amplitude, harmonic components, or oscillation frequency of the real-time operating data related to the first energy storage device. Time-window feature extraction can include setting a sliding time window and calculating at least one of the slope, increment, duration of continuous offset, or rate of change of the real-time operating data related to the first energy storage device within the sliding time window. The first data can include the real-time operating data of the first energy storage device, such as time-series data related to cell voltage, internal resistance, temperature difference, inverter power, and current in the first energy storage device.

[0057] In the above-described solution, the seventh content is generated from multiple dimensions, including the time domain, frequency domain, and sliding window, through a feature calculation engine. This allows for in-depth mining of the hidden periodic operating patterns and latent fault characteristics in the first data. The third content is then fused with the first content to generate the first descriptive information. This not only fills the information gaps of single data through multi-dimensional feature engineering and can cover as many types of fault characteristics of the first energy storage device as possible, but also integrates quantitative indicators, historical fault cases, and basic device context to construct the input of the intelligent system. This provides a basis for subsequent fault diagnosis and processing by the intelligent system, effectively suppresses invalid reasoning by the intelligent system, and significantly improves the accuracy and response efficiency of early-stage deterioration fault diagnosis in energy storage.

[0058] In one embodiment, optionally, see [link to relevant documentation]. Figure 2 , Figure 3 and Figure 4 After determining at least one third content from the first information based on the first content, the method further includes the following steps: determining the similarity between each third content and the first content; filtering each third content according to the similarity between each third content and the first content, wherein the similarity between the filtered and retained third content and the first content is greater than the similarity between the filtered and removed third content and the first content.

[0059] Optionally, invoking the intelligent system to output the second content based on the first description information may include the following steps: splitting the first task into multiple second tasks and determining the first scheduling information of the multiple second tasks through the intelligent system orchestration engine, wherein the first task is a task for fault diagnosis processing of the first energy storage device determined based on the first content, the first scheduling information is used to invoke at least two intelligent systems and instruct at least two intelligent systems to collaboratively execute multiple second tasks, and the second task is a sub-task that can be executed independently, rolled back, and retried; and invoking at least two intelligent systems to execute multiple second tasks and output the second content based on the first scheduling information and the first description information.

[0060] The intelligent system orchestration engine can be used for at least one of the following: task lifecycle management or collaborative scheduling of multiple intelligent systems. It enables task decomposition, dependency topology, or fault tolerance control for the first task. Upon receiving the first task, the intelligent system orchestration engine performs task decomposition, dependency generation, allocation to multiple intelligent systems, and generation of scheduling strategies for those systems.

[0061] The first scheduling information may include scheduling instructions generated by the intelligent system orchestration engine. It may also include the execution order of the various second tasks, the dependencies between the second tasks, or the identifiers of the intelligent systems invoked by the second tasks. This first scheduling information can be used to instruct at least two different intelligent systems to cooperate in parallel and sequentially to complete the second tasks to be performed by each intelligent system.

[0062] In the above-described solution, a multi-intelligent system task orchestration and collaboration mechanism is adopted. The first task corresponding to the fault diagnosis of energy storage equipment is decomposed into tasks, and multiple intelligent systems work together to complete the entire fault diagnosis process. This not only breaks down the fault diagnosis process into sub-tasks with independent execution, retry, and rollback capabilities, improving the overall fault tolerance and high concurrency processing limit of the fault diagnosis link, but also distributes the computing load and alleviates the computational pressure on a single intelligent system. In addition, by using unified scheduling information to standardize the interaction logic between various intelligent systems, parallel collaborative computation of various second tasks is achieved, shortening the response time for energy storage equipment degradation fault diagnosis.

[0063] Based on the above embodiments, variations of the embodiments are proposed. It should be noted that, for the sake of brevity, the variations only describe the differences from the above embodiments. In one embodiment, see [link to embodiment]. Figure 2 , Figure 3 and Figure 4 After invoking the intelligent system to output the second content based on the first description information, the following steps may also be included: The second content is converted into a first control instruction. The first control instruction is a control instruction that can be recognized by the MQTT protocol instruction executor. The first control instruction is used to instruct the energy storage device to be adjusted according to the fault diagnosis results and fault handling suggestions in the second content. The first control instruction is then sent to the first energy storage device so that the first energy storage device executes the first control instruction.

[0064] The second content may include the fault diagnosis results and fault handling suggestions output by the intelligent system after performing the fault diagnosis and handling task; for example, it may include fault location results, degradation risk level, degradation cause analysis, or fault handling suggestions corresponding to the fault diagnosis results. The fault handling suggestions may include control processing logic for the first energy storage device, such as control parameters and handling logic related to at least one of the following: cell temperature control, power current limiting of the energy storage inverter, or charge / discharge rate control of the first energy storage device. The first control instruction may include hardware-executable operation instructions generated by converting the fault handling suggestions corresponding to the fault diagnosis results indicated by the second content. The first control instruction can be used to drive control actions on the cells and energy storage inverter in the first energy storage device according to the fault handling suggestions corresponding to the fault diagnosis results.

[0065] In the above scenario, the control commands can be directly parsed and converted into control commands that can be recognized by the MQTT protocol executor. The existing MQTT-based control links can be reused and automatically sent to the first energy storage device, thus completing the regulation of the battery cells and energy storage inverter in the first energy storage device without manual intervention.

[0066] Optionally, in one embodiment, the processing method of the energy storage device includes the following steps: adding second content to first information and updating the first information. By storing the second content, which includes fault diagnosis results and fault handling suggestions based on those results, into the first information to complete the update, historical fault cases can be continuously expanded, the retrieval knowledge base can be enriched, and new real cases can be added to match subsequent similar deterioration searches, thereby improving the accuracy of fault reasoning, forming a self-accumulation and iteration mechanism for diagnostic data, and continuously optimizing the system's fault diagnosis capabilities.

[0067] In one embodiment, this embodiment is applicable to situations involving fault diagnosis and analysis of energy storage devices, particularly scenarios involving intelligent analysis, fault diagnosis, and predictive maintenance of time-series data of portable energy storage power supplies and home photovoltaic energy storage inverters. This intelligent diagnostic system can implement the energy storage device processing method in hardware and / or software. The intelligent diagnostic system of this embodiment may include at least one energy storage device, and the intelligent diagnostic system is configured to perform the following operations: The first engine performs at least one anomaly detection on the first data to generate the first content. The first engine includes a distributed real-time stream processing engine associated with the first energy storage device. The first data includes real-time operating data related to the first energy storage device reported using a distributed message queue. The system responds to a first content indicating that the first energy storage device is showing a deterioration trend of abnormal operation. Based on the first content, it generates first description information and calls an intelligent system to output second content based on the first description information. The intelligent system is used to perform fault diagnosis and processing tasks on the energy storage device. The second content includes fault diagnosis results and fault processing suggestions.

[0068] In one embodiment, the solution of this embodiment can be combined with various optional solutions in one or more of the above embodiments to generate first description information based on the first content, including: Based on the first content, at least one third content is determined from the first information. The first information includes multiple fourth contents, which indicate various historical faults associated with the second energy storage device and the corresponding maintenance plans. The first content matches the third content. The second energy storage device and the first energy storage device are of the same type, or the components in the first energy storage device and the components in the second energy storage device are of the same type. The multiple fourth contents are generated by processing the second data through a second engine. The second engine includes a distributed batch offline data processing engine. The second data includes historical business data and historical log data associated with the second energy storage device. First description information is generated based on the at least one third content and the first content.

[0069] In one embodiment, the solution of this embodiment can be combined with various optional solutions in one or more of the above embodiments to determine at least one third content from the first information based on the first content, including: The first content is vectorized by a third engine to generate a first vector. The third engine is used to extract the temporal change features in the first data and encode them into a vector. Based on the first vector, similarity detection is performed on multiple fourth contents of the first information to obtain the third content, and the multiple fourth contents are stored in vector form.

[0070] In one embodiment, the solution of this embodiment can be combined with various optional solutions in one or more of the above embodiments to determine at least one third content from the first information based on the first content, including: Based on the first content, the third content is obtained by keyword retrieval of multiple fifth contents of the first information. The multiple fifth contents are determined based on at least one of the following: historical fault case documents, business rules, historical fault diagnosis reports, or question and answer knowledge bases of common problems associated with the second energy storage device. The third content is obtained by searching multiple sixth contents of the first information based on the first content. The multiple sixth contents are constructed and generated based on historical business data and historical log data associated with the second energy storage device.

[0071] In one embodiment, the solution of this embodiment can be combined with various optional solutions in one or more of the above embodiments to generate first description information based on the at least one third content and the first content, including: The seventh content is obtained by performing a first feature extraction on the first data based on the feature calculation engine. The first feature extraction includes at least one of time domain feature extraction, frequency domain feature extraction, or time window feature extraction. First description information is generated based on the seventh content, the at least one third content, and the first content.

[0072] In one embodiment, the solution of this embodiment can be combined with various optional solutions in one or more of the above embodiments, and after determining at least one third content from the first information based on the first content, it further includes: Determine the similarity between each of the at least one third content and the first content; Each third content is filtered based on its similarity to the first content. The third content retained after filtering has a greater similarity to the first content than the third content removed after filtering has a greater similarity to the first content.

[0073] In one embodiment, the solution of this embodiment can be combined with various optional solutions in one or more of the above embodiments, and the intelligent system is invoked to output the second content based on the first description information, including: The first task is divided into multiple second tasks by an intelligent system orchestration engine, and the first scheduling information of the multiple second tasks is determined. The first task is a task for fault diagnosis of the first energy storage device determined based on the first content. The first scheduling information is used to call at least two intelligent systems and instruct the at least two intelligent systems to cooperate in executing the multiple second tasks. The second tasks are sub-tasks that can be executed independently, rolled back, and retried. Based on the first scheduling information and the first description information, the at least two intelligent systems are invoked to execute the plurality of second tasks and output the second content.

[0074] In one embodiment, the solution of this embodiment can be combined with various optional solutions in one or more of the above embodiments, and after calling the intelligent system to output the second content based on the first description information, it further includes: The second content is converted into a first control instruction, which is a control instruction that can be recognized by the MQTT protocol instruction executor. The first control instruction is used to instruct the energy storage device to be adjusted according to the fault diagnosis results and fault handling suggestions in the second content. The first control command is sent to the first energy storage device so that the first energy storage device executes the first control command.

[0075] In one embodiment, the solution of this embodiment can be combined with various optional solutions in one or more of the above embodiments, and the system further includes: adding and storing the second content to the first information.

[0076] In the technical solution of this invention embodiment, a distributed real-time stream processing engine associated with the first energy storage device is adopted to perform at least one anomaly detection on the energy storage device operation data transmitted and reported via a distributed message queue. The distributed real-time stream processing engine is configured with parallel streaming computing power and can continuously parse the full runtime sequence data of the first energy storage device pushed by the message queue. Compared with offline data analysis, it can effectively solve the defects of offline analysis such as delayed fault alarms and easy omission of progressive deterioration faults. When the deterioration trend of abnormal operation of the first energy storage device is identified, first descriptive information can be constructed based on the first content generated by at least one anomaly detection. This can provide input for the intelligent diagnostic system and automatically drive the intelligent system to perform fault diagnosis and early warning, outputting fault judgment conclusions and fault handling solutions. In addition, relying on the intelligent system for reasoning analysis to replace experience judgment, while constraining the reasoning logic to ensure the accuracy and reliability of the diagnostic results, directly outputting practical maintenance suggestions, and significantly improving the fault operation and maintenance efficiency of energy storage devices.

[0077] The energy storage device processing system provided in the embodiments of the present invention can execute the energy storage device processing method provided in any of the embodiments of the present invention, and has the corresponding functions and beneficial effects of executing the energy storage device processing method. For details, please refer to the relevant operations of the energy storage device processing method in the foregoing embodiments.

[0078] It is worth noting that the various units and modules included in the above-mentioned device are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be realized; in addition, the specific names of each functional unit are only for easy differentiation and are not used to limit the protection scope of the embodiments of the present invention.

[0079] According to embodiments of the present invention, the process described in the flowchart above can be applied to an intelligent diagnostic system. For example, an embodiment of the present invention includes an intelligent diagnostic system comprising: at least one battery module; at least one battery management system, wherein the at least one battery management system corresponds one-to-one with the at least one battery module, and the battery management system is electrically connected to the corresponding battery module; the intelligent diagnostic system includes at least one energy storage device, wherein the energy storage device includes battery cells and an energy storage inverter, the battery cells perform energy storage functions, and the inverter realizes AC / DC power conversion and grid-connected charging and discharging control; the intelligent diagnostic system is used to execute program code of the processing method of the energy storage device shown in the flowchart.

[0080] In this embodiment, the intelligent diagnostic system provided by this application is applicable to various application scenarios, such as grid-connected power generation and energy storage, off-grid photovoltaic and energy storage (for powering electrical equipment in homes, RVs, and yachts), wind power generation and energy storage, and electric equipment. The specific application scenario can be determined according to the actual application scenario, and no limitation is made here. The following will use the off-grid photovoltaic and energy storage field as an example for explanation. Other application scenarios are basically similar and will not be described in detail.

[0081] In off-grid photovoltaic (PV) and energy storage (ESS) applications, a complete PV-ESS system includes at least a photovoltaic (PV) power generation system, a power conversion system, an intelligent diagnostic system, and a power consumption system. The PV power generation system consists of several solar panels connected in series and parallel to convert solar energy into electrical energy. The power conversion system injects the electrical energy generated by the PV power generation system into the intelligent diagnostic system for storage. The power consumption system then adapts the electrical energy stored in the intelligent diagnostic system to the power required by the electrical equipment. The aforementioned power conversion system can typically be implemented using a DC / DC converter with MPPT (Multi-Level Photovoltaic) functionality, while the power consumption system can typically be implemented using a DC / DC converter or a DC / AC converter. This section focuses on the intelligent diagnostic system. An intelligent diagnostic system is usually composed of multiple battery packs connected together. Connecting battery packs in series increases the output voltage of the battery pack, while connecting them in parallel achieves a larger battery capacity. Therefore, to obtain an intelligent diagnostic system with the target voltage level and capacity, users will connect multiple battery packs in series and parallel to obtain a high-voltage, high-capacity intelligent diagnostic system for energy storage and power supply. In addition, for convenient centralized control, each battery pack in the intelligent diagnostic system is usually equipped with a communication interface, allowing the battery packs to communicate with each other via methods such as CAN bus, RS485 bus, and RS232 bus. Currently, with the further development of the new energy industry and the expansion of application scenarios, higher requirements are being placed on the performance of intelligent diagnostic systems, including system stability and consistency, waterproofing, dustproofing, and safety levels.

[0082] In an optional but not limited implementation, this application also provides a photovoltaic power generation system, which includes: a photovoltaic power generation module, a controller, and an intelligent diagnostic system according to any one of the above embodiments; the photovoltaic power generation module is used to convert solar energy into direct current (DC) power; the controller is used to store the DC power in the intelligent diagnostic system.

[0083] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0084] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A method for processing energy storage devices, characterized in that, The method includes: The first engine performs at least one anomaly detection on the first data to generate the first content. The first engine includes a distributed real-time stream processing engine associated with the first energy storage device. The first data includes real-time operating data related to the first energy storage device reported using a distributed message queue. The system responds to a first content indicating that the first energy storage device is showing a deterioration trend of abnormal operation. Based on the first content, it generates first description information and calls an intelligent system to output second content based on the first description information. The intelligent system is used to perform fault diagnosis and processing tasks on the energy storage device. The second content includes fault diagnosis results and fault processing suggestions.

2. The method according to claim 1, characterized in that, First description information is generated based on the first content, including: Based on the first content, at least one third content is determined from the first information. The first information includes multiple fourth contents, which indicate various historical faults associated with the second energy storage device and the corresponding maintenance plans. The first content matches the third content. The second energy storage device and the first energy storage device are of the same type, or the components in the first energy storage device and the components in the second energy storage device are of the same type. The multiple fourth contents are generated by processing the second data through a second engine. The second engine includes a distributed batch offline data processing engine. The second data includes historical business data and historical log data associated with the second energy storage device. First description information is generated based on the at least one third content and the first content.

3. The method according to claim 2, characterized in that, Determining at least one third piece of content from the first information based on the first content includes: The first content is vectorized by a third engine to generate a first vector. The third engine is used to extract the temporal change features in the first data and encode them into a vector. Based on the first vector, similarity detection is performed on multiple fourth contents of the first information to obtain the third content, and the multiple fourth contents are stored in vector form.

4. The method according to claim 2, characterized in that, Determining at least one third piece of content from the first information based on the first content includes: Based on the first content, the third content is obtained by keyword retrieval of multiple fifth contents of the first information. The multiple fifth contents are determined based on at least one of the following: historical fault case documents, business rules, historical fault diagnosis reports, or question and answer knowledge bases of common problems associated with the second energy storage device. The third content is obtained by searching multiple sixth contents of the first information based on the first content. The multiple sixth contents are constructed and generated based on historical business data and historical log data associated with the second energy storage device.

5. The method according to claim 2, characterized in that, Generate first description information based on the at least one third content and the first content, including: The seventh content is obtained by performing a first feature extraction on the first data based on the feature calculation engine. The first feature extraction includes at least one of time domain feature extraction, frequency domain feature extraction, or time window feature extraction. First description information is generated based on the seventh content, the at least one third content, and the first content.

6. The method according to claim 2, characterized in that, After determining at least one third piece of content from the first information based on the first content, the method further includes: Determine the similarity between each of the at least one third content and the first content; Each third content is filtered based on its similarity to the first content. The third content retained after filtering has a greater similarity to the first content than the third content removed after filtering has a greater similarity to the first content.

7. The method according to claim 1, characterized in that, Based on the first description information, the intelligent system is invoked to output the second content, including: The first task is divided into multiple second tasks by an intelligent system orchestration engine, and the first scheduling information of the multiple second tasks is determined. The first task is a task for fault diagnosis of the first energy storage device determined based on the first content. The first scheduling information is used to call at least two intelligent systems and instruct the at least two intelligent systems to cooperate in executing the multiple second tasks. The second tasks are sub-tasks that can be executed independently, rolled back, and retried. Based on the first scheduling information and the first description information, the at least two intelligent systems are invoked to execute the plurality of second tasks and output the second content.

8. The method according to claim 1, characterized in that, After invoking the intelligent system to output the second content based on the first description information, it also includes: The second content is converted into a first control instruction, which is a control instruction that can be recognized by the MQTT protocol instruction executor. The first control instruction is used to instruct the energy storage device to be adjusted according to the fault diagnosis results and fault handling suggestions in the second content. The first control command is sent to the first energy storage device so that the first energy storage device executes the first control command.

9. The method according to claim 2, characterized in that, The method further includes: The second content is added and stored in the first information.

10. An intelligent diagnostic system, characterized in that, A processing method for an energy storage device according to any one of claims 1 to 9, wherein the intelligent diagnostic system includes at least one energy storage device, and the intelligent diagnostic system is configured to perform the following operations: The first engine performs at least one anomaly detection on the first data to generate the first content. The first engine includes a distributed real-time stream processing engine associated with the first energy storage device. The first data includes real-time operating data related to the first energy storage device reported using a distributed message queue. The system responds to a first content indicating that the first energy storage device is showing a deterioration trend of abnormal operation. Based on the first content, it generates first description information and calls an intelligent system to output second content based on the first description information. The intelligent system is used to perform fault diagnosis and processing tasks on the energy storage device. The second content includes fault diagnosis results and fault processing suggestions.

Images