Intelligent safety distributed energy storage device and method of use thereof

Abstract

The application discloses an intelligent safe distributed energy storage device and a use method thereof, relates to the technical field of distributed energy storage, and comprises a plurality of distributed energy storage units, a central management platform and a multi-stage cooperative safety system. A cloud end LSTM engine is used to predict thermal runaway risks. Unit-level sensor groups are used to monitor temperature gradients and gas concentration mutations in real time. An explosion fuse type rapid shutdown device and a high-temperature-resistant isolation cover are used to ensure that extreme faults are strictly limited within a single module, prevent disasters from spreading, and an intelligent gradient temperature control system is used to adjust the temperature and inhibit thermal runaway. Meanwhile, a double-target DDPG algorithm energy scheduling optimizer of the central management platform is used to synchronously optimize electricity cost and a battery life attenuation factor. A power conversion module supports multi-port access, improves flexibility, a fault self-healing control module automatically recombines a power supply network topology, and combines a local controller interface to guarantee that the system can still maintain key power supply when local faults occur, thereby improving the toughness and reliability.

Description

Technical Field

[0001] This application relates to the field of distributed energy storage technology, specifically to an intelligent and safe distributed energy storage device and its usage method. Background Technology

[0002] With the accelerated transformation of the global energy structure, the proportion of renewable energy generation, represented by wind and solar power, continues to increase. These energy sources have significant intermittent and fluctuating characteristics, posing a huge challenge to the stability, reliability, and power quality of the power grid. At the same time, the peak-valley difference in power load is increasing, and traditional power grid peak-shaving methods are costly, slow to respond, and lack flexibility. Against this backdrop, energy storage technology, especially distributed energy storage, is needed to achieve time-shifting of energy and provide peak-shaving and frequency regulation services.

[0003] Existing distributed energy storage devices suffer from various shortcomings. For example, patent document CN106941276B discloses a distributed, module-based DC data center, but the device in this document lacks a multi-level system protection mechanism, resulting in thermal runaway being unable to be confined to a single module and posing a high risk of disaster propagation. Patent document CN103577892A discloses a progressive scheduling method for an intelligent power distribution system, but the device in this document suffers from low accuracy in locating system faults and high response latency. Patent document US20150311713A1 discloses a service-based grid management method, but the device in this document struggles to accurately quantify battery degradation trajectories, leading to a disconnect between charge / discharge measurement and lifespan optimization. Patent document CN118920551A discloses a distributed energy storage device, but the device in this document cannot simultaneously meet the technical requirement of low-power access for a large number of nodes. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this application provides an intelligent, safe, distributed energy storage device and its usage method to solve the problems mentioned above in the prior art.

[0005] To achieve the above objectives, this application provides the following technical solution:

[0006] A smart and safe distributed energy storage device includes a central management platform, a multi-level collaborative safety system, and several distributed energy storage units;

[0007] Each of the distributed energy storage units (1) includes a modular battery module (3), a power conversion module (6), a local controller (10), and a dynamic isolation module (11). The modular battery module (3) is equipped with a gradient temperature control system, including a phase change material cooling layer (4) and an independent liquid cooling channel (5). The gradient temperature control system is configured to activate liquid cooling to enhance phase change material regeneration when the temperature gradient exceeds a first set threshold. The power conversion module (6) is equipped with a DC port (7), an AC port (8), and an emergency output port (9). The local controller (10) has a built-in consensus algorithm communication interface. The ports of the power conversion module (6), the dynamic isolation module (11) and the gradient temperature control system are respectively connected. The dynamic isolation module (11) includes an explosive fuse type fast shut-off device (12), a high temperature resistant isolation cover (13) and an inert gas release unit. The action time of the explosive fuse type fast shut-off device (12) is ≤1ms. The high temperature resistant isolation cover (13) is made of ceramic fiber material and maintains structural integrity for ≥120 seconds at a high temperature of 1600℃ and unfolds to form a closed compartment within 0.5 seconds. The inert gas release unit is built into the high temperature resistant isolation cover (13) and is used to release inert gas into the compartment when triggered.

[0008] The central management platform connects each distributed energy storage unit through a hybrid communication network, including an energy scheduling optimizer, a safety strategy generator, and a cloud-based safety analysis engine. The energy scheduling optimizer is used to simultaneously optimize electricity costs and battery life degradation factors using a dual-objective deep deterministic strategy gradient algorithm. The safety strategy generator is used to dynamically adjust the upper limit of charge and discharge rates. The cloud-based safety analysis engine is used to predict thermal runaway risks based on a long short-term memory neural network.

[0009] The multi-level collaborative safety system includes a unit-level sensor group (19), a regional-level safety node (20), and a fault self-healing control module (27). The unit-level sensor group (19) is connected to the signal output terminal of the local controller (10). The unit-level sensor group (19) is used to monitor the battery temperature gradient and gas concentration mutation rate in real time. The regional-level safety node (20) is connected to the dynamic isolation module (11). The regional-level safety node (20) integrates a flame detector and a directional fire extinguishing device. The fault self-healing control module (27) automatically reassembles the power supply network topology by controlling the on / off state of each power conversion module (6).

[0010] Preferably, the dynamic isolation module further includes a negative pressure guiding channel for directionally exporting thermal runaway gas to a safe area.

[0011] Preferably, the local controller is configured to implement the following method: when a temperature gradient > 5℃ / s or a gas concentration mutation rate > 10% / min is detected, an explosion-type fast shutdown device is triggered; safety data is exchanged with adjacent units through a consensus algorithm to accurately locate the faulty unit; and when the cloud-based safety analysis engine outputs a thermal runaway risk warning, the phase change material cooling layer is actively activated for preventive cooling.

[0012] Preferably, the central management platform further includes a blockchain evidence storage module and a digital twin operation and maintenance interface. The blockchain evidence storage module is used to encrypt and store security event operation logs and response instructions, and the digital twin operation and maintenance interface is used to establish a full life cycle battery health status mirror model.

[0013] Preferably, the hybrid communication network includes: a time-sensitive network protocol layer, a 5G NB-IoT communication module, and an end-to-end encryption unit. The time-sensitive network protocol layer is used to transmit emergency control commands, the 5G NB-IoT communication module is used to transmit real-time monitoring data, and the end-to-end encryption unit is used to achieve two-way authentication between the device and the cloud.

[0014] A method of using the aforementioned intelligent and safe distributed energy storage device, based on the same inventive concept, includes the following steps:

[0015] S1 intelligent collaborative scheduling, also known as energy scheduling optimizer, dynamically allocates charging and discharging tasks for each unit based on the DDPG algorithm, so that the battery life degradation factor is ≤0.001% / cycle;

[0016] S2 multi-level collaborative security response

[0017] The first-level response means that when a unit-level sensor group detects an anomaly, the charge / discharge rate of that unit is limited to ≤0.5C.

[0018] A Level 2 response means that when a regional safety node detects a heat radiation signal using a flame detector, it activates the high-temperature resistant isolation cover of the adjacent unit and sprays aerosol extinguishing agent in a directional manner.

[0019] Level 3 response means that when the cloud security analysis engine predicts a thermal runaway probability of >90%, the dynamic isolation module is fully locked and the inert gas release unit is activated to release inert gas.

[0020] S3 fault self-healing reconfiguration means transferring the energy of the faulty unit to the healthy unit through the fault self-healing control module;

[0021] S4 predictive protection is based on an LSTM model to identify high-risk units 30 minutes in advance, proactively transfer the load and activate a gradient temperature control system to cool them down, and simultaneously update the blockchain evidence report and digital twin model parameters.

[0022] Preferably, the secondary response further includes energy transfer across units and the activation of directional spraying of aerosol extinguishing agents at regional safety nodes via a power conversion module.

[0023] Preferably, step S4 further includes: generating a blockchain evidence report to record the security incident response process; and updating the thermal runaway prediction model parameters based on the digital twin operation and maintenance interface.

[0024] Preferably, the method further includes the following steps:

[0025] Automatically switches to emergency output port for power supply when the power grid fails;

[0026] It also receives reverse power transmission from electric vehicles through the AC port to smooth out grid frequency fluctuations.

[0027] Preferably, the load forecasting model construction and optimization module is also used to manually correct the forecast data and upload it to the upper-level control system.

[0028] Preferably, the modular battery module is quickly replaceable through a standardized interface, and the emergency output port is connected to the virtual power plant and emergency power vehicle system.

[0029] Compared with existing technologies, the advantages of this solution are as follows: 1. This solution predicts the risk of thermal runaway through a cloud-based LSTM engine, monitors temperature gradients and gas concentration changes in real time with unit-level sensor groups, and provides rapid response from regional-level safety nodes. Through explosive fuse-type fast shutdown devices and high-temperature resistant isolation enclosures, extreme faults are strictly limited to a single module to prevent the spread of disasters. The gradient temperature control system intelligently adjusts the temperature to suppress thermal runaway. At the same time, the dual-objective DDPG algorithm energy scheduling optimizer of the central management platform simultaneously optimizes electricity costs and battery life degradation factors, significantly reducing the total life cycle cost. The power conversion module supports multi-port access, improving flexibility. In addition, the fault self-healing control module can automatically reorganize the power supply network topology. Combined with the consensus algorithm interface of the local controller, it ensures that the system can still maintain critical power supply during local faults, thereby improving resilience and reliability.

[0030] 2. This solution achieves a revolutionary breakthrough in safety protection capabilities through the coordinated upgrade of the dynamic isolation module and the local controller. The newly added negative pressure diversion channel of the dynamic isolation module actively directs the thermal runaway high-temperature gas to the safe zone, forming a "directional risk elimination-physical isolation" double insurance with the original explosion fuse-type fast shut-off device with a response of ≤1ms and the ceramic fiber isolation cover. The local controller triggers the fuse isolation instantaneously based on highly sensitive criteria, and at the same time achieves sub-meter level fault accurate location by exchanging data with adjacent units through consensus algorithms. When the cloud LSTM engine predicts risks, it actively activates the phase change material cooling layer for proactive prevention, improving the safety of use.

[0031] 3. This solution uses a blockchain-based evidence storage module to encrypt and store security event operation logs and response instructions in a distributed ledger, constructing an immutable, traceable, and auditable chain of security event evidence. A digital twin operation and maintenance interface simultaneously builds a full-lifecycle battery health (SOH) mirror model, accurately quantifying battery degradation trajectories and enabling predictive health management. Combined with cloud-based LSTM risk prediction and DDPG optimization algorithms, it dynamically adjusts charging and discharging strategies to mitigate lifespan degradation. Based on the SOH model, it proactively avoids high-load risk conditions and provides a scientific basis for spare parts replacement decisions, thereby improving the reliability, economy, and compliance of the distributed energy storage system.

[0032] 4. This solution provides ≤1ms deterministic transmission and zero packet loss guarantee for critical operations such as circuit breaker commands and charge / discharge regulation through the Time-Sensitive Networking (TSN) protocol layer, ensuring absolute reliability of millisecond-level security response. The 5G NB-IoT communication module supports the synchronous transmission of real-time monitoring data such as temperature gradient and gas concentration by 100,000 nodes with wide coverage and low power consumption, enabling full-dimensional analysis of cloud-based LSTM thermal runaway prediction. The end-to-end encryption unit uses national cryptographic algorithms to achieve two-way authentication between the device and the cloud, and combined with the blockchain evidence storage module, it eliminates the risk of data tampering and unauthorized access. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the overall process of one embodiment of the present solution;

[0034] Figure 2 This is a schematic diagram of an embodiment of the energy dispatch optimizer for the energy storage device in this scheme;

[0035] Figure 3 This is a schematic diagram of a process flow for an embodiment of the energy storage device unit-level sensor group in this solution.

[0036] Figure 4 This is a schematic diagram of the gas leakage handling process in one embodiment of the energy storage device of this scheme.

[0037] Figure 5 This is a schematic diagram of the risk prediction process for an embodiment of the energy storage device in this scheme.

[0038] The components include: 1. Distributed energy storage unit; 2. Central management platform; 3. Modular battery module; 4. Phase change material cooling layer; 5. Independent liquid cooling channel; 6. Power conversion module; 7. DC port; 8. AC port; 9. Emergency output port; 10. Local controller; 11. Dynamic isolation module; 12. Explosion-proof fuse type fast shutdown device; 13. High-temperature resistant isolation cover; 14. Hybrid communication network; 16. Energy scheduling optimizer; 17. Security policy generator; 18. Cloud security analysis engine; 19. Unit-level sensor group; 20. Regional-level security node; 23. Negative pressure diversion channel; 24. Blockchain evidence storage module; 25. Digital twin operation and maintenance interface; 27. Fault self-healing control module. Detailed Implementation

[0039] 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.

[0040] In the description of this application, it should be noted that the terms "upper," "lower," "inner," "outer," "front end," "rear end," "both ends," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0041] In the description of this application, it should be noted that, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0042] Example 1: See Figure 1 , Figure 2 and Figure 3 The present invention provides an embodiment of an intelligent and safe distributed energy storage device, comprising a central management platform 2, a multi-level collaborative safety system, and several distributed energy storage units 1 (only one is shown in the figure).

[0043] Each of the plurality of distributed energy storage units 1 comprises:

[0044] The modular battery module 3 is equipped with a gradient temperature control system, including a phase change material cooling layer 4 and an independent liquid cooling channel 5. The gradient temperature control system is configured to activate liquid cooling to enhance the regeneration of the phase change material when the temperature gradient exceeds a first set threshold.

[0045] The power conversion module 6 supports three-terminal connection: DC port 7, AC port 8, and emergency output port 9.

[0046] Local controller 10, with a built-in consensus algorithm communication interface;

[0047] The dynamic isolation module 11 integrates an explosive fuse-type fast shut-off device 12, a high-temperature resistant isolation cover 13, and an inert gas release unit. The action time of the explosive fuse-type fast shut-off device 12 is ≤1ms. The high-temperature resistant isolation cover 13 is made of ceramic fiber material. The isolation cover maintains structural integrity for ≥120 seconds at a high temperature of 1600℃ and unfolds to form a closed compartment within 0.5 seconds. The inert gas release unit is built into the high-temperature resistant isolation cover 13 and is used to release inert gas into the compartment when triggered.

[0048] The central management platform 2 connects to each unit via a hybrid communication network 14, including:

[0049] The energy scheduling optimizer 16 employs a dual-objective deep deterministic policy gradient (DDPG) algorithm to simultaneously optimize electricity costs and battery life degradation factors.

[0050] Safety policy generator 17 dynamically adjusts the upper limit of charge and discharge rates;

[0051] Cloud-based security analytics engine 18 predicts thermal runaway risk based on Long Short-Term Memory (LSTM) neural networks;

[0052] A multi-level collaborative security system includes:

[0053] Unit-level sensor group 19 monitors battery temperature gradient and gas concentration abrupt change rate in real time;

[0054] Regional-level safety node 20 integrates flame detectors and directional fire extinguishing devices;

[0055] Fault self-healing control module 27 automatically reassembles the power supply network topology;

[0056] In this embodiment, the gradient temperature control system combines the passive heat absorption of the phase change material cooling layer 4 with the active cooling of the independent liquid cooling channel 5 to achieve intelligent and efficient thermal management. The mechanism of initiating liquid cooling regeneration based on the temperature gradient threshold can effectively suppress heat diffusion in the battery pack, nip the risk of thermal runaway in the bud, and thus extend the service life and efficiency of the phase change material.

[0057] The ≤1ms explosion-fusible fast circuit breaker 12 can completely cut off the electrical connection at the very early stage of the fault, fundamentally preventing the arc and energy from continuing to flow into the fault point;

[0058] The high-temperature resistant isolation cover 13 is made of ceramic fiber material, which forms a closed compartment within 0.5 seconds and releases inert gas to achieve physical explosion suppression and fireproof sealing under extreme conditions, ensuring that the thermal runaway of a single cell or module will not trigger a catastrophic chain reaction.

[0059] Unit-level sensor group 19: Real-time monitoring of temperature gradients and gas concentration abrupt change rates, the earliest signs of thermal runaway, realizes the transformation from monitoring voltage or current to early warning of intrinsic safety parameters;

[0060] Regional-level safety node 20: Integrates flame detectors and directional fire extinguishing devices, enabling accurate detection and rapid extinguishing of local fires and preventing the situation from escalating;

[0061] The fault self-healing control module 27 can automatically reorganize the power supply network after a unit fault is isolated, ensuring continuous power supply to critical loads and greatly improving the resilience and reliability of the system.

[0062] The Energy Dispatch Optimizer 16 adopts the dual-objective DDPG algorithm, whose core value lies in simultaneously optimizing electricity cost and battery life degradation factor. This means that it not only seeks the most cost-effective charging and discharging strategy, but also actively selects the working mode that is most friendly to battery health, thereby significantly delaying battery capacity degradation and extending the battery pack's lifespan.

[0063] The security strategy generator 17 dynamically adjusts the upper limit of the charge and discharge rate based on the prediction results of the cloud analysis engine. When the risk is low, it can fully release the system's capacity to participate in peak shaving and earn revenue. When the risk is high, it will actively reduce the rate to ensure safety, thus achieving a dynamic balance between economic benefits and safe operation.

[0064] The cloud-based security analysis engine 18 analyzes historical and real-time data based on LSTM neural networks, enabling it to predict thermal runaway risks in advance. This achieves a leap from "post-event handling" to "pre-event warning" and from "regular maintenance" to "predictive maintenance," greatly reducing unplanned downtime and unnecessary maintenance costs.

[0065] Each energy storage unit has a built-in local controller 10 with a consensus algorithm communication interface, which lays the foundation for decentralized or partially decentralized collaborative control. This enables the units to continue to operate collaboratively through the consensus algorithm when the central management platform 2 fails, making the system architecture more robust and flexible.

[0066] The combination of the hybrid communication network 14 ensures the reliability and real-time performance of data interaction, providing a communication foundation for the implementation of all advanced algorithms;

[0067] The power conversion module 6 supports three-terminal access: DC port 7, AC port 8, and emergency output port 9. This allows the device to flexibly adapt to various application scenarios and seamlessly switch between grid-connected and off-grid modes. Combined with the functions of the fault self-healing control module 27, its emergency output port 9 can become a reliable solution to ensure uninterrupted power supply to critical loads, and is particularly suitable for occasions with extremely high requirements for power supply reliability.

[0068] Example 2: See Figure 1 and Figure 3 This application also provides an embodiment in which the dynamic isolation module 11 further includes:

[0069] The negative pressure guiding channel 23 directs the thermal runaway gas to a safe area.

[0070] The local controller 10 is configured to perform:

[0071] When a temperature gradient > 5℃ / s or a gas concentration mutation rate > 10% / min is detected, the explosion fuse type fast shut-off device 12 is triggered.

[0072] By exchanging security data with neighboring units through a consensus algorithm, the faulty unit can be accurately located.

[0073] When the cloud-based security analysis engine 18 outputs a warning of thermal runaway risk, the phase change material cooling layer 4 is proactively activated for preventative cooling.

[0074] Furthermore, by directing and rapidly discharging flammable and explosive runaway gases to a safe area, the accumulation of dangerous gases in the battery compartment can be prevented, thereby preventing a violent explosion.

[0075] The local controller 10 is given autonomous execution rights based on explicit thresholds, such as temperature gradient > 5℃ / s or gas concentration mutation rate > 10% / min. It can bypass the delay of uploading to the cloud for decision-making and can directly trigger the explosive fuse-type fast shutdown device 12 within the initial millisecond window of thermal runaway, achieving the fastest fault point isolation. Its response speed is far superior to systems that rely on a central platform. Even if the central management platform 2 or the hybrid communication network 14 fails, each unit still has independent and powerful local safety decision-making and execution capabilities, which greatly enhances the robustness and survivability of the entire distributed system.

[0076] The local controller 10 exchanges security data with neighboring units through a consensus algorithm. Through cross-verification of data from multiple units, it can quickly and accurately locate the single faulty unit that has experienced thermal runaway, achieving sub-meter level fault location accuracy. The accurate location result can be directly provided to the fault self-healing control module 27, enabling it to purposefully reorganize the power supply network, accurately isolate the faulty unit, and ensure the continued operation of healthy units.

[0077] After receiving a long-term, predictive risk warning from the cloud-based security analysis engine 18 based on LSTM, the local controller 10 proactively activates the phase change material cooling layer 4 for preventative cooling. By intervening before thermal runaway occurs, the risk is proactively eliminated, greatly reducing the probability of catastrophic accidents. Moreover, this preventative cooling is only activated when a risk warning is issued, avoiding continuous high-power operation of the cooling system and improving safety.

[0078] Example 3: See Figure 1 and Figure 5 This application also provides an embodiment: the central management platform 2 further includes:

[0079] Blockchain evidence storage module 24 encrypts and stores security event operation logs and response instructions;

[0080] Digital twin operation and maintenance interface 25, establishes a full life cycle battery state of health (SOH) mirror model;

[0081] Furthermore, by encrypting and storing all security event operation logs and response instructions, and utilizing the distributed ledger characteristics of blockchain technology, the immutability, non-forgeability, and traceability of all critical security data are ensured. This provides a highly credible chain of evidence with strong legal force for post-incident root cause analysis and liability determination. All instructions executed through the local controller 10 or policies issued by the cloud security analysis engine 18 are permanently recorded, forming a transparent operation history that can be reviewed by all authorized parties. This helps improve the operational transparency of the system and simplifies the compliance audit and internal review process.

[0082] The digital twin operation and maintenance interface 25 enables accurate prediction and visualization of battery state of health (SOH). By establishing a digital twin model that is completely mirrored with the physical entity, this interface can accurately simulate and present the battery state of health (SOH) of each modular battery module 3 throughout its entire life cycle. This allows operation and maintenance personnel to intuitively and in real time understand the degree of battery degradation and remaining lifespan.

[0083] Based on historical and real-time data from digital twin models, it is possible to more accurately predict battery performance inflection points and maintenance needs, thereby enabling predictive maintenance and avoiding unexpected downtime. At the same time, through battery state of health (SOH) assessment, it is possible to optimize battery charging and discharging strategies, maximize the remaining value of batteries while ensuring safety, and provide authoritative scientific decision-making basis for the secondary use of batteries.

[0084] The LSTM risk prediction algorithm of the cloud-based security analytics engine 18 and the dual-objective DDPG algorithm of the energy scheduling optimizer 16 provide extremely rich and high-quality training and analysis data sources, which can further improve the accuracy and reliability of these artificial intelligence algorithms.

[0085] Example 4: See still Figure 1 One embodiment provided in this application: the hybrid communication network 14 includes:

[0086] The Time-Sensitive Networking (TSN) protocol layer transmits emergency control commands;

[0087] 5G NB-IoT communication module for transmitting real-time monitoring data;

[0088] End-to-end encryption unit enables two-way authentication between the device and the cloud;

[0089] Furthermore, the Time-Sensitive Networking (TSN) protocol layer is specifically used for transmitting emergency control commands, including commands to trigger the explosive fuse-type fast shut-off device 12. This provides deterministic, extremely low transmission latency and extremely high reliability, ensuring that the most critical safety commands can be delivered to the execution unit without interruption or collision at millisecond or even microsecond speeds, thus meeting the stringent timeliness requirements of the safety system.

[0090] By leveraging the wide coverage, massive connectivity, and low power consumption characteristics of 5G NB-IoT technology, it is possible to continuously and stably transmit the massive monitoring data generated by a large number of distributed energy storage units to the cloud security analysis engine 18 with extremely low energy consumption, providing an uninterrupted data stream for big data analysis and artificial intelligence prediction.

[0091] Two-way authentication ensures that only legitimate devices can access the network, and only legitimate cloud platforms can instruct devices, thus eliminating unauthorized access and fake base station attacks at the source. Encrypted transmission ensures that all transmitted instructions and data cannot be eavesdropped on or tampered with, effectively protecting the security of system operation data and user privacy. This enables the entire system to resist network attacks and meet the network security requirements of critical infrastructure.

[0092] Example 5: See Figure 1 , Figure 2 , Figure 3 , Figure 4 and Figure 5 This application provides an embodiment of a method for using an intelligent, safe, distributed energy storage device: The working steps of using the intelligent, safe, distributed energy storage device are as follows:

[0093] S1 Intelligent Cooperative Scheduling: The energy scheduling optimizer 16 dynamically allocates charging and discharging tasks for each unit based on the DDPG algorithm, so that the battery life degradation factor is ≤0.001% / cycle;

[0094] S2 Multi-Level Collaborative Security Response:

[0095] Level 1 response: When the unit-level sensor group 19 detects an anomaly, the charge / discharge rate of that unit is limited to ≤0.5C;

[0096] Level 2 response: When the area-level safety node 20 detects a heat radiation signal with a flame detector, it activates the high-temperature resistant isolation cover 13 of the adjacent unit and uses a directional fire extinguishing device to directionally spray aerosol fire extinguishing agent;

[0097] Level 3 Response: When the cloud security analysis engine 18 predicts a thermal runaway probability of >90%, it triggers the dynamic isolation module 11 to be fully locked and starts the inert gas release unit to release inert gas.

[0098] S3 Fault Self-Healing Reconfiguration: Transfers energy from the faulty unit to the healthy unit via the fault self-healing control module 27;

[0099] S4 Predictive Protection: Based on the LSTM model, high-risk units are identified 30 minutes in advance, the load is actively transferred and the gradient temperature control system is activated to cool down, and the blockchain evidence report and digital twin model parameters are updated simultaneously.

[0100] The secondary response includes:

[0101] Energy transfer across units is achieved through power conversion module 6;

[0102] Regional safety node 20 initiates directional spraying of aerosol extinguishing agent;

[0103] S4 further includes:

[0104] Generate blockchain evidence reports and record the security incident response process;

[0105] Update thermal runaway prediction model parameters based on digital twin operation and maintenance interface 25;

[0106] The method further includes:

[0107] Automatically switches to emergency output port 9 for power supply when the power grid fails;

[0108] It also receives reverse power transmission from electric vehicles through AC port 8 to smooth out grid frequency fluctuations;

[0109] The modular battery module 3 can be quickly replaced through a standardized interface, and the emergency output port 9 supports access to virtual power plants and emergency power vehicle systems.

[0110] Furthermore, by controlling battery degradation to an extremely low level, the replacement cost and cost per kilowatt-hour over the entire life cycle are significantly reduced, directly improving the project's return on investment. This transforms the traditional "experience-based scheduling" into "data-driven scheduling" based on artificial intelligence, enabling the project to earn electricity revenue while proactively ensuring battery health, thus achieving the best balance between economy and longevity.

[0111] The role of a Level 1 response is to proactively reduce the heat load and buy time for the system to recover;

[0112] The Level 2 response is triggered by the regional security node 20, which activates the high-temperature resistant isolation cover 13 of the adjacent unit to build a physical firewall in advance to prevent the spread of the disaster. At the same time, it sprays aerosol fire extinguishing agent in a directional manner to carry out precise and efficient local fire suppression of the fire source, avoiding the impact of total flooding fire suppression on health equipment.

[0113] The power conversion module 6 enables energy transfer across units, quickly stripping energy from faulty units and fundamentally eliminating the source of danger.

[0114] The Level 3 response is triggered by the cloud-based security analysis engine 18 when the predicted probability is >90%, which causes the dynamic isolation module 11 to be fully locked and release inert gas. This enables the most comprehensive suppression measures to be activated at the highest warning level, ensuring that the system enters an absolutely safe state.

[0115] The fault self-healing control module 27 automatically transfers energy from the faulty unit to the healthy unit and reconstructs the network, which can greatly improve the power supply reliability of the system and ensure that the overall system can continue to supply power to the critical load when some units fail, realizing "uninterrupted power supply" self-repair. The control logic of the fault self-healing control module 27 is as follows: when a faulty unit is detected, the central processing unit starts the reconstruction process, communicates with adjacent units based on the consensus algorithm to obtain available backup capacity information, selects a reconstruction scheme according to threshold rules and real-time status, and cuts off the power output of the faulty unit through the power switch controller, controls the power conversion module 6 of the remaining healthy units to adjust the output power and operating mode, realizes the redistribution of energy and load transfer, continuously monitors the grid frequency and voltage fluctuations during the reconstruction process to ensure smooth power supply switching, and after the reconstruction is completed, updates the system topology information and synchronizes it to the digital twin operation and maintenance interface 25 of the central management platform 2.

[0116] Based on the LSTM model, risks can be identified 30 minutes in advance, and the load can be actively transferred and the gradient temperature control system can be activated to cool down. This helps to eliminate risks before thermal runaway occurs and minimizes the probability of safety accidents.

[0117] Generating blockchain-based evidence reports can provide an immutable, auditable, and authoritative record of operations for every security incident response, meeting compliance requirements and providing irrefutable evidence for incident tracing and liability determination.

[0118] By updating model parameters based on the digital twin operation and maintenance interface 25, the digital twin model can continuously evolve in actual combat, becoming more and more accurate with use, forming a virtuous cycle of self-learning and self-optimization, and continuously improving the predictive accuracy of the cloud security analysis engine 18.

[0119] Automatic switching to emergency output port 9 ensures uninterrupted power supply to critical loads in the event of a power grid failure.

[0120] Receiving reverse power transmission from electric vehicles via AC port 8 is beneficial for transforming distributed energy storage devices into regional energy hubs. This not only allows them to draw power from the grid but also aggregates distributed resources such as electric vehicles, providing services such as frequency regulation to the grid, smoothing grid frequency fluctuations, and enhancing grid stability.

[0121] The modular battery module 3 can be quickly replaced through standardized interfaces, which helps to simplify the maintenance process, shorten downtime, and reduce the difficulty and cost of operation and maintenance.

[0122] Emergency output port 9 supports access to virtual power plants and emergency power vehicles, which can greatly expand the external synergy and application flexibility of the system. It can participate in market transactions as a node of the virtual power plant, and can also be connected in parallel with emergency power vehicles in emergency situations to form a larger capacity power supply system.

[0123] Working principle: A cloud-based LSTM engine predicts thermal runaway risk; unit-level sensor group 19 monitors temperature gradients and gas concentration changes in real time; regional-level safety nodes 20 respond quickly; and explosive fuse-type fast shut-off devices 12 and high-temperature resistant isolation enclosures 13 ensure extreme faults are strictly confined to a single module to prevent disaster spread. A gradient temperature control system intelligently adjusts the temperature to suppress thermal runaway. Simultaneously, the dual-objective DDPG algorithm energy scheduling optimizer 16 of the central management platform 2 simultaneously optimizes electricity costs and battery life degradation factors, significantly reducing the total lifecycle cost. The power conversion module 6 supports multi-port access for increased flexibility. Furthermore, the fault self-healing control module 27 can automatically reorganize the power supply network topology, combined with the local controller 10... The consensus algorithm interface ensures that the system can maintain critical power supply even during partial failures, thereby improving resilience and reliability. Through the coordinated upgrade of the dynamic isolation module 11 and the local controller 10, a revolutionary breakthrough in safety protection capabilities has been achieved. The newly added negative pressure diversion channel 23 of the dynamic isolation module 11 actively directs the thermal runaway high-temperature gas to the safe zone, forming a "directional risk elimination-physical isolation" double insurance with the original ≤1ms response explosion fuse-type fast shut-off device 12 and ceramic fiber isolation cover. The local controller 10 triggers fuse isolation instantaneously based on high-sensitivity criteria, and at the same time achieves sub-meter level fault accuracy by exchanging data with adjacent units through the consensus algorithm. When the cloud LSTM engine predicts risks, it actively activates the phase change material cooling layer 4 for advanced pre-emptive protection. To enhance security, a blockchain-based evidence storage module 24 encrypts and stores security event operation logs and response instructions using a distributed ledger, constructing an immutable, traceable, and auditable security event evidence chain. A digital twin operation and maintenance interface 25 simultaneously builds a full-lifecycle battery health status (SOH) mirror model, accurately quantifying battery degradation trajectories and enabling predictive health management. Combined with cloud-based LSTM risk prediction and DDPG optimization algorithms, it dynamically adjusts charging and discharging strategies to mitigate lifespan degradation. Based on the SOH model, it proactively avoids high-load risk conditions and provides a scientific basis for spare parts replacement decisions, thereby improving the reliability, economy, and compliance of the distributed energy storage system. The Time-Sensitive Networking (TSN) protocol layer provides a basis for circuit breaker commands and charging... Key operations such as discharge regulation provide ≤1ms deterministic transmission and zero packet loss guarantee, ensuring absolute reliability of millisecond-level security response. The 5GNB-IoT communication module, with its wide coverage and low power consumption characteristics, supports the synchronous transmission of real-time monitoring data such as temperature gradient and gas concentration from 100,000 nodes, enabling full-dimensional analysis of cloud-based LSTM thermal runaway prediction. The end-to-end encryption unit uses national cryptographic algorithms to achieve two-way authentication between the device and the cloud, combined with a blockchain evidence storage module 24 to eliminate the risk of data tampering and unauthorized access. In terms of security, a five-level active defense system is constructed: the unit-level sensor group 19 limits the charge and discharge rate of abnormal units, the area-level synchronous activation of the high-temperature resistant isolation cover 13 and directional fire suppression, and the power conversion module 6 realizes energy transfer across units.The cloud-based LSTM predicts a full lockout when the probability of thermal runaway exceeds 90%, triggering predictive protection by initiating a gradient temperature control system 30 minutes in advance. Blockchain-based evidence storage and a digital twin model work in real-time to form a closed-loop evolutionary defense strategy. For lifespan optimization, the energy scheduling optimizer 16 uses the DDPG algorithm to compress the battery lifespan degradation factor to ≤0.001% / cycle, improving lifespan. Regarding system resilience, the fault self-healing control module 27 achieves power interruption recovery in <100ms, automatically switching to emergency output port 9 for power supply during grid outages and supporting reverse power transmission from electric vehicles to suppress grid fluctuations. The modular battery module 3 supports standardized and rapid replacement, and the emergency port seamlessly connects to virtual power plants and emergency power vehicle systems, constructing a multi-source energy mutual support ecosystem.

[0124] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A smart, safe, distributed energy storage device, characterized in that: It includes a central management platform (2), a multi-level collaborative security system, and several distributed energy storage units (1); Each of the distributed energy storage units (1) includes a modular battery module (3), a power conversion module (6), a local controller (10), and a dynamic isolation module (11). The modular battery module (3) is equipped with a gradient temperature control system, including a phase change material cooling layer (4) and an independent liquid cooling channel (5). The gradient temperature control system is configured to activate liquid cooling to enhance phase change material regeneration when the temperature gradient exceeds a first set threshold. The power conversion module (6) is equipped with a DC port (7), an AC port (8), and an emergency output port (9). The local controller (10) has a built-in consensus algorithm communication interface. The ports of the power conversion module (6), the dynamic isolation module (11) and the gradient temperature control system are respectively connected. The dynamic isolation module (11) includes an explosive fuse type fast shut-off device (12), a high temperature resistant isolation cover (13) and an inert gas release unit. The action time of the explosive fuse type fast shut-off device (12) is ≤1ms. The high temperature resistant isolation cover (13) is made of ceramic fiber material and maintains structural integrity for ≥120 seconds at a high temperature of 1600℃ and unfolds to form a closed compartment within 0.5 seconds. The inert gas release unit is built into the high temperature resistant isolation cover (13) and is used to release inert gas into the compartment when triggered. The central management platform (2) connects each distributed energy storage unit (1) through a hybrid communication network (14), including an energy scheduling optimizer (16), a safety strategy generator (17), and a cloud-based safety analysis engine (18). The energy scheduling optimizer (16) is used to simultaneously optimize electricity costs and battery life degradation factors using a dual-objective deep deterministic strategy gradient algorithm. The safety strategy generator (17) is used to dynamically adjust the upper limit of the charge and discharge rate. The cloud-based safety analysis engine (18) is used to predict thermal runaway risk based on a long short-term memory neural network. The multi-level collaborative safety system includes a unit-level sensor group (19), a regional-level safety node (20), and a fault self-healing control module (27). The unit-level sensor group (19) is connected to the signal output terminal of the local controller (10). The unit-level sensor group (19) is used to monitor the battery temperature gradient and gas concentration mutation rate in real time. The regional-level safety node (20) is connected to the dynamic isolation module (11). The regional-level safety node (20) integrates a flame detector and a directional fire extinguishing device. The fault self-healing control module (27) automatically reassembles the power supply network topology by controlling the on / off state of each power conversion module (6).

2. The intelligent and safe distributed energy storage device according to claim 1, characterized in that: The dynamic isolation module (11) also includes a negative pressure guiding channel (23) for directionally discharging thermal runaway gas to a safe area.

3. The intelligent and safe distributed energy storage device according to claim 1, characterized in that, The local controller (10) is configured to implement the following method: when a temperature gradient > 5℃ / s or a gas concentration mutation rate > 10% / min is detected, the explosion fuse type fast shut-off device (12) is triggered, the fault unit is accurately located by exchanging safety data with adjacent units through a consensus algorithm, and when the cloud safety analysis engine (18) outputs a thermal runaway risk warning, the phase change material cooling layer (4) is actively activated to perform preventive cooling.

4. The intelligent and safe distributed energy storage device according to claim 1, characterized in that: The central management platform (2) also includes a blockchain evidence storage module (24) and a digital twin operation and maintenance interface (25). The blockchain evidence storage module (24) is used to encrypt and store security event operation logs and response instructions. The digital twin operation and maintenance interface (25) is used to establish a full life cycle battery health status mirror model.

5. The intelligent and safe distributed energy storage device according to claim 1, characterized in that: The hybrid communication network (14) includes: a time-sensitive network protocol layer, a 5G NB-IoT communication module, and an end-to-end encryption unit. The time-sensitive network protocol layer is used to transmit emergency control commands, the 5G NB-IoT communication module is used to transmit real-time monitoring data, and the end-to-end encryption unit is used to achieve two-way authentication between the device and the cloud.

6. A method of using an intelligent, safe, distributed energy storage device according to claim 5, characterized in that, Includes the following steps: S1 intelligent collaborative scheduling, namely the energy scheduling optimizer (16), dynamically allocates the charging and discharging tasks of each unit based on the DDPG algorithm, so that the battery life decay factor is ≤0.001% / cycle; S2 multi-level collaborative security response The first-level response means that when the unit-level sensor group (19) detects an abnormality, the charge / discharge rate of the unit is limited to ≤0.5C; When the secondary response, i.e., the regional safety node (20), detects a heat radiation signal with a flame detector, it activates the high-temperature resistant isolation cover (13) of the adjacent unit and sprays aerosol extinguishing agent in a directional manner. Level 3 response means that when the cloud security analysis engine (18) predicts that the probability of thermal runaway is >90%, the dynamic isolation module (11) is fully locked and the inert gas release unit is started to release inert gas. S3 fault self-healing reconfiguration is to transfer the energy of the faulty unit to the healthy unit through the fault self-healing control module (27); S4 predictive protection is based on an LSTM model to identify high-risk units 30 minutes in advance, proactively transfer the load and activate a gradient temperature control system to cool them down, and simultaneously update the blockchain evidence report and digital twin model parameters.

7. The method of using an intelligent and safe distributed energy storage device according to claim 6, characterized in that: The secondary response also includes the transfer of energy across units through the power conversion module (6) and the activation of directional spraying of aerosol extinguishing agent by the regional safety node (20).

8. The method of using an intelligent and safe distributed energy storage device according to claim 6, characterized in that, Step S4 also includes: generating a blockchain evidence report to record the security incident response process; and updating the thermal runaway prediction model parameters based on the digital twin operation and maintenance interface (25).

9. The method of using an intelligent and safe distributed energy storage device according to claim 6, characterized in that... It also includes the following steps: When the power grid fails, it automatically switches to the emergency output port (9) for power supply; And receive reverse power transmission from electric vehicles through AC port (8) to smooth out grid frequency fluctuations.

10. The method of using an intelligent and safe distributed energy storage device according to claim 6, characterized in that: The modular battery module (3) can be quickly replaced through a standardized interface, and the emergency output port (9) is connected to the virtual power plant and emergency power vehicle system.

Images