Manufacturing method of robot battery pack, thermal runaway suppression method, and power management method

Abstract

This solution relates to the manufacturing method, thermal runaway suppression method, and power management method of a robotic battery pack, specifically within the field of battery technology. The manufacturing method for this robotic battery pack pioneers a three-in-one manufacturing process integrating a "stiffness-adjustable step frame + micro / nano structure phase change layer + integrated pressure relief and purification module." In this solution, the composite frame with step-decreasing stiffness can actively switch from a "rigid protection" state to a "buffered energy absorption" state during thermal runaway, more effectively dissipating the impact energy generated by the violent expansion of the battery cells and preventing shell rupture caused by hard impacts. The phase change material layer with a micro / nano secondary structure achieves lower interfacial thermal resistance and more precise directional volume expansion during phase change, thereby significantly improving the cooling efficiency and uniformity of constraint pressure for faulty battery cells. The integrated module realizes a "one-stop" harmless treatment from pressure relief and instantaneous cooling to catalytic purification, completely solving the secondary safety problem of direct emission of toxic and flammable gases during thermal runaway.

Description

[0001] This solution relates to the field of battery technology, specifically to a method for manufacturing a robot battery pack, a method for suppressing thermal runaway, and a method for power management. Background Technology

[0002] With the widespread application of mobile robots in warehousing and logistics, outdoor inspection, and special operations, the safety, reliability, and environmental adaptability of their power battery systems have become core bottlenecks restricting their large-scale deployment and expansion into more complex scenarios. Robot operating environments are highly variable, often accompanied by stresses such as vibration, impact, and temperature and humidity changes, posing continuous challenges to the mechanical structure, thermal management, and electrical connection reliability of batteries. More importantly, in densely populated or high-value-added environments such as shopping malls, data centers, and industrial production lines, the consequences of battery thermal runaway are extremely severe. Therefore, developing highly reliable power systems that can fundamentally prevent and suppress battery thermal runaway and deeply integrate with the robot as a whole has become an urgent industry need.

[0003] Existing technologies primarily improve the safety of robot batteries through two paths: one is the hardware protection path, such as adding fire-resistant and heat-insulating materials inside the battery pack, using a more robust outer shell, installing simple pressure relief valves, or adding fire-fighting devices to the outside of the battery pack. The other is the management optimization path, such as monitoring voltage and temperature through the battery management system and cutting off the circuit in case of abnormalities, or using robot scheduling algorithms to prevent battery over-discharge. However, these solutions have significant limitations: hardware protection is mostly passive and has limited effectiveness in the face of severe internal thermal runaway, and external fire protection has response delays and secondary hazards (such as damage to equipment from sprinklers); management optimization is mostly based on post-event responses with fixed thresholds, and the battery management system is independent of the robot's motion control and task planning systems, resulting in information fragmentation and forming "islands".

[0004] The fundamental problems with existing technical solutions are as follows: First, the safety mechanisms are lagging and passive. Whether it's alarms based on voltage / temperature thresholds or physical pressure relief, these are passive responses after severe internal side reactions have occurred and thermal runaway has entered an irreversible stage, failing to provide preventative and early, precise intervention. Second, system coordination is lacking. The battery, as an independent "black box" integrated with the robot, has a disconnect between its output capacity, thermal management requirements, and the robot's real-time motion state and task load. This prevents the power system from performing optimally within safety boundaries and can easily lead to system-wide task interruption in the event of partial failures, resulting in insufficient system-level reliability and availability. Therefore, a novel solution capable of proactive prediction, active suppression, and deep coordination is needed. Summary of the Invention

[0005] This specification provides a method for manufacturing a robot battery pack, a method for suppressing thermal runaway, and a method for power management, which can solve the above-mentioned problems in the prior art.

[0006] Firstly, this specification provides a method for manufacturing a robot battery pack, comprising: A honeycomb-shaped compartmentalized structure is constructed inside the battery pack. The compartmentalized structure is composed of a three-dimensional frame made of shape memory alloy and high-damping elastomer. The frame is configured such that its macroscopic stiffness can decrease in a step manner within a specific temperature range. Each honeycomb unit is used to independently accommodate a battery cell module. On the inner wall of each of the cellular cells, a solid phase change material layer with a micro-nano secondary structure is laminated by a vapor deposition process. When the phase change material layer reaches the phase change temperature, its volume expansion rate is greater than 10% and the interfacial thermal resistance in surface contact with the cell surface decreases by at least 50%. Each of the cellular units is configured with a multi-stage intelligent pressure relief-purification integrated module, which includes: a pressure-temperature dual-parameter triggered pressure relief valve, a microchannel transient cooler directly connected to the outlet of the pressure relief valve, and a non-precious metal integral catalyst unit loaded on a porous metal foam substrate, for the staged treatment of the leaked gas.

[0007] In some embodiments, the macroscopic stiffness can be reduced in a stepwise manner by pre-embedding shape memory alloy wires with different austenitic phase transformation end temperatures (Af) at the key stress nodes of the composite frame. The alloy wires are woven in a three-dimensional interlacing manner in the frame. When the temperature exceeds the Af point of each alloy wire in sequence from low to high, the equivalent elastic modulus of the frame decreases in a stepwise manner, thereby achieving a progressive mechanical response of "rigid support - buffer energy absorption - flexible wrapping" during the thermal runaway of the battery cell.

[0008] In some embodiments, the micro / nano secondary structure is a micrometer-scale columnar array constructed by a template method in a phase change material layer, and a nanowire or nanosheet structure grown on the surface of the columnar array. This design is used to enhance the capillary adsorption force with the cell surface through the nanostructure during the phase change process, and to guide the volume expansion direction perpendicular to the cell surface through the microstructure, so as to apply uniform surface pressure.

[0009] In some embodiments, the microchannel transient cooler is pre-filled with inert working fluid microcapsules capable of undergoing endothermic phase change; the non-precious metal monolithic catalyst is a perovskite-type composite oxide, which has a catalytic oxidation efficiency of not less than 95% for carbon monoxide and hydrocarbons within a temperature window of 300-500°C.

[0010] Secondly, this specification provides a method for suppressing thermal runaway in a robot battery pack, characterized in that it is applied to a robot battery pack manufactured by the method described in any one of the first aspects, the method comprising: Based on the geometry, material properties, and cell chemistry model of the battery pack, a high-fidelity digital twin containing multi-field coupling of electrochemistry, heat, stress, fluid, and chemical reaction is constructed, and a lightweight proxy model is trained for real-time inference. During the operation of the battery pack, the state of the physical entity and the digital twin are synchronized in real time. Using the proxy model, the state vector of the previous moment is used as input to predict the time-series change curve of the "thermal runaway risk entropy" of each cell in the battery pack within a future time window. Define and dynamically update an adaptive risk probability threshold related to the current health status and historical operating conditions of the battery; When the predicted "thermal runaway risk entropy" of a certain cellular cell exceeds the adaptive risk probability threshold within the time window, a predictive suppression decision process is triggered. This process optimizes and generates the optimal suppression parameters for the cell based on the simulation results of the digital twin. The parameters include at least the type, dosage, release rate curve, and co-phase relationship with the charging current of the inhibitor.

[0011] In some embodiments, the thermal runaway risk entropy integrates the dimensionless indices of the predicted temperature gradient, internal pressure, characteristic gas yield, and inter-unit thermal interaction intensity within the unit, and its calculation incorporates the output of the physical model and the neural network correction terms trained based on historical fault data.

[0012] In some embodiments, the update logic of the adaptive risk probability threshold is as follows: when the battery is in a high-rate discharge, high ambient temperature, or has recently experienced a large mechanical shock, the threshold is automatically lowered to trigger a more conservative warning and suppression strategy.

[0013] In some embodiments, the coordinated phase relationship with the charging current specifically refers to: triggering the release of the inhibitor at a specific phase (such as a peak or trough) of the AC or pulse charging current, and utilizing the coupling effect of the current field and the stress field to enhance the penetration and uniform distribution of the inhibitor in the porous electrodes inside the cell.

[0014] Thirdly, this specification provides a power management method for a robot battery pack, characterized in that it is applied to a robot integrating a robot battery pack as described in the first aspect and a thermal runaway suppression method for the robot battery pack as described in the second aspect, the method being executed by a power domain controller, including: The system receives and integrates information from multiple sources to generate a global dynamic system state cognition. The information includes at least: a real-time thermal runaway risk entropy distribution map from the predictive suppression system, the real-time output power peak of the battery pack, the real-time posture and joint load of the robot body, and the future trajectory and action intention from the upper-level task planner. Based on the global state cognition, a rolling optimization problem with multiple objective functions such as task completion, system safety margin, and energy consumption is solved, and a set of dynamically executable instructions is output in real time. The instruction set includes at least an adaptive torque-speed limit curve issued to the motion controller, dynamic pump speed and valve opening instructions issued to the thermal management system, and task degradation or reconfiguration suggestions triggered when necessary.

[0015] In some embodiments, the task degradation or reconfiguration suggestion is as follows: when the predictive suppression system is triggered or a battery pack unit is isolated, the power domain controller automatically generates a degradation operation plan that maintains the highest possible task completion rate based on the updated system capability model, while ensuring core safety constraints, and submits it to the upper-level task planner for confirmation and execution through a standard interface.

[0016] This solution has the following beneficial effects: The manufacturing method of this robot's battery pack pioneered a three-in-one manufacturing process: a "stiffness-adjustable step frame + micro / nano structure phase change layer + integrated pressure relief and purification module." The effects are explained as follows: The composite frame in S1, with its step-adjustable stiffness, can actively switch from a "rigid protection" state to a "buffered energy absorption" state during thermal runaway, more effectively dissipating the impact energy generated by the cell's violent expansion and preventing shell rupture due to hard impacts. The phase change material layer in S2, with its micro / nano secondary structure, achieves lower interfacial thermal resistance and more precise directional volume expansion during phase change compared to traditional filling methods, thus significantly improving the cooling efficiency and uniformity of constraint pressure for faulty cells. The integrated module in S3 achieves a "one-stop" harmless treatment from pressure relief and instantaneous cooling to catalytic purification, completely solving the secondary safety problem of direct emission of toxic and flammable gases during thermal runaway.

[0017] The battery pack manufactured using this method transcends the traditional concept of passive protection, creating an intelligent adaptive structure that can dynamically respond to internal disasters, actively intervene in runaway processes, and handle hazardous byproducts in a closed loop, thus achieving a revolutionary improvement in intrinsic safety.

[0018] The robotic battery pack thermal runaway suppression method constructs a predictive decision chain consisting of a high-fidelity digital twin, a lightweight surrogate model, risk entropy prediction, and adaptive thresholds. Effect reasoning: Through a multi-physics coupled digital twin, the complex electrochemical-thermal-mechanical interactions within the battery can be virtually reproduced and extrapolated, something simple data-driven models cannot achieve. Introducing a comprehensive "thermal runaway risk entropy" as a predictive indicator provides a more accurate assessment of system instability tendencies than a single temperature or voltage parameter. Combined with an adaptive threshold for battery health status, the system possesses "contextual awareness," enabling it to adopt different alert levels under varying operating conditions.

[0019] This method upgrades battery safety monitoring from "alarms based on simple thresholds" to "predictive diagnosis based on virtual simulation and comprehensive entropy criteria", enabling advanced, quantitative, and dynamic assessment of thermal runaway risks, thereby allowing interventions to be initiated much earlier than traditional methods.

[0020] This method is the core of the entire intelligent decision-making process. It receives rich sensor data from the battery pack manufactured according to claim 1, and through simulation prediction, provides the most critical inputs for subsequent power domain collaborative management—a "thermal runaway risk entropy distribution map" and "predictive suppression commands." It acts as the system's "early warning system and decision-making brain," transforming the hardware's state perception into advanced risk knowledge and control strategies, driving subsequent execution and global coordination.

[0021] The power management method for robot battery packs achieves "multi-source information fusion + multi-objective rolling optimization + autonomous task degradation and reconfiguration" through a power domain controller. Effect reasoning: By fusing information from the battery safety system, the robot's state, and task intent across all dimensions, it breaks down the information barriers between traditional subsystems. By solving the multi-objective rolling optimization problem, it can dynamically rebalance safety, performance, and energy consumption in each control cycle, providing the current optimal comprehensive instruction. In particular, the "task degradation and reconfiguration" capability ensures that the system does not passively stop after partial failure, but actively seeks the optimal operating solution under new constraints.

[0022] This method integrates the robot's power system from a series of passively executed components into an intelligent whole that can think proactively, optimize collaboratively, and gracefully respond to faults, achieving a qualitative leap in system-level reliability and task completion capabilities.

[0023] This method represents the ultimate embodiment and "enhancement" of the entire technical solution's value. It fully utilizes the advanced risk prediction information provided in the second aspect, transforming it into real-time constraints (adaptive power boundaries) and forward-looking planning (task degradation suggestions) for robot motion. This enables the material safety and predictive safety capabilities created by claims 1 and 5 to be seamlessly translated into the robot's overall high reliability, high availability, and high performance, completing a value loop from "component safety" to "system survivability."

[0024] Other functionalities of the solutions provided in this specification will be partially listed in the following description. The inventive aspects of the solutions provided in this specification can be fully understood through practice or use of the methods, apparatus, and combinations described in the detailed examples below. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of this specification, 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 this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0026] Figure 1 A schematic flowchart of a method for manufacturing a robot battery pack according to an embodiment of this specification is shown. Figure 2 A schematic flowchart of a method for suppressing thermal runaway of a robot battery pack according to an embodiment of this specification is shown. Figure 3 A power management method for a robot battery pack is shown according to an embodiment of this specification. Detailed Implementation

[0027] The following description provides specific application scenarios and requirements for this specification, intended to enable those skilled in the art to make and use the contents of this specification. Various partial modifications to the disclosed embodiments will be apparent to those skilled in the art, and the general principles defined herein can be applied to other embodiments and applications without departing from the spirit and scope of this specification. Therefore, this specification is not limited to the embodiments shown, but rather to the widest scope consistent with the claims.

[0028] The terminology used herein is for the purpose of describing particular exemplary embodiments only and is not restrictive. For example, unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “the” used herein may also include the plural forms. When used in this specification, the terms “comprising,” “including,” and / or “containing” mean that the associated integers, steps, operations, elements, and / or components are present, but do not exclude the presence of one or more other features, integers, steps, operations, elements, components, and / or groups, or that other features, integers, steps, operations, elements, components, and / or groups may be added to the system / method.

[0029] Considering the following description, these and other features of this specification, as well as the operation and function of the related components of the structure, and the economy of assembly and manufacture of the parts, can be significantly improved. All of these form part of this specification with reference to the accompanying drawings. However, it should be clearly understood that the drawings are for illustrative and descriptive purposes only and are not intended to limit the scope of this specification. It should also be understood that the drawings are not drawn to scale.

[0030] The flowcharts used in this specification illustrate operations implemented according to some embodiments of this specification. It should be clearly understood that the operations in the flowcharts may not be implemented in a sequential order. Instead, the operations may be implemented in reverse order or simultaneously. Furthermore, one or more additional operations may be added to the flowcharts. One or more operations may be removed from the flowcharts.

[0031] Figure 1 A schematic flowchart illustrating a method for manufacturing a robot battery pack according to an embodiment of this specification is shown; as follows: Figure 1 As shown. This specification provides a method for manufacturing a robot battery pack, including: S101. Construct a honeycomb-shaped compartment structure inside the battery pack, wherein the compartment structure is composed of a three-dimensional frame made of shape memory alloy and high-damping elastomer.

[0032] The three-dimensional frame is configured such that its macroscopic stiffness can decrease in a step within a specific temperature range, and each cellular unit is used to independently house a battery cell module.

[0033] This step aims to manufacture the core support structure of the battery pack. A "honeycomb" geometry divides the interior of the battery pack into numerous independent hexagonal units, each unit (honeycomb cell) acting like an independent "room" housing a group of battery cells. The material used to construct the walls of these units is not ordinary metal or plastic, but a special three-dimensional framework composed of a "shape memory alloy" and a "high-damping elastomer."

[0034] Explanation of proper nouns: Shape memory alloys: a type of smart metallic material that can undergo a reversible phase transition at a specific temperature, thereby restoring its preset shape or generating a huge restoring force.

[0035] High-damping elastomer: a polymer material with excellent ability to absorb mechanical vibration energy (damping) and convert it into heat energy for dissipation.

[0036] A step-like decrease in macroscopic stiffness refers to a significant and sudden decrease in the overall stiffness of a material or structure as the temperature changes through a specific range, rather than a gradual change, like descending a step.

[0037] This step creates an intelligent framework with "temperature sensing" and "adaptive mechanical behavior" capabilities. When the battery is operating normally, the framework maintains high rigidity, providing stable support and protection for the cells and resisting daily vibrations and shocks. When a cell within a cellular unit experiences thermal runaway, an abnormal temperature rise, and enters a specific dangerous range, the shape memory alloy in that cell's framework is "activated," driving a step change in the mechanical properties of the entire framework, transforming it from a "rigid" state to a "flexible and highly damped" state.

[0038] Traditional battery packs have a rigid structure. During thermal runaway, the rapidly expanding cells collide head-on with the rigid outer casing, easily leading to casing rupture and the ejection of high-temperature substances. This solution proactively changes the rules of this "confrontation." When a localized overheating risk is detected, the intelligent frame automatically "flexes" in that area. This transformation brings two core advantages: 1) Buffering and energy absorption: The flexible, high-damping properties act like an "airbag," absorbing and dissipating the enormous impact energy generated by the cell expansion, preventing structural rupture. 2) Stress diversion: The flexible frame better adapts to and encloses the expanding cells, transforming localized stress concentration into a more uniform distribution, preventing stress concentration points from becoming the source of rupture. This achieves a fundamental shift in safety strategy from passively resisting internal destructive forces to actively adapting to and dissipating destructive forces.

[0039] S102. On the inner wall of each of the cellular units, a solid phase change material layer with a micro-nano secondary structure is laminated by a vapor deposition process. When the phase change material layer reaches the phase change temperature, its volume expansion rate is greater than 10% and the interfacial thermal resistance in surface contact with the cell surface decreases by at least 50%.

[0040] This step involves applying a special solid-state phase change material coating to the "walls" (inner surfaces) of each cellular cell using a precise vapor deposition process. This material has a unique "micro-nano secondary structure" and undergoes significant volume expansion upon reaching its designed phase change temperature, while simultaneously greatly reducing the contact thermal resistance with the cell surface.

[0041] Explanation of proper nouns: Vapor deposition process: a precision coating technology that vaporizes materials in a vacuum or a specific atmosphere and then condenses them on the surface of a substrate to form a film, which can achieve uniform, dense and strong-adhesion coatings.

[0042] Micro / nano secondary structures refer to material surfaces that simultaneously possess regular geometric morphologies at the micrometer and nanometer scales (such as nanowires grown on micrometer pillar arrays), forming a multi-level rough structure.

[0043] Interfacial thermal resistance: The additional resistance encountered by heat as it passes through the interface between two materials when they are in contact. The lower the thermal resistance, the better the thermal conductivity.

[0044] Solid-state phase change materials: materials that undergo a solid-solid phase transition (change in crystal structure) at a specific temperature, accompanied by the absorption / release of a large amount of latent heat.

[0045] This step constructs a "thermal-mechanical bidirectional intelligent response interface" between the battery cell and the protection structure. Its intelligence is reflected in two aspects: First, intelligent thermal management: when the battery cell overheats and triggers a phase change, the interface thermal resistance drops sharply, allowing the heat accumulated inside the battery cell to be efficiently conducted to the phase change material layer for absorption (utilizing the latent heat of phase change). Second, intelligent mechanical constraint: the large volume expansion accompanying the phase change applies a uniform and gentle surface pressure to the battery cell, mildly and effectively suppressing further bulging and deformation.

[0046] Traditional thermal management interfaces (such as silicone pads) have fixed performance and are prone to gaps after the battery cell swells, leading to increased thermal resistance and deteriorated heat dissipation. The logic of this solution lies in designing a dynamically optimized interface that becomes more robust and effective as the temperature rises. Its innovation lies in: 1) Achieving "adaptive bonding" through micro / nano structures: The micro / nano secondary structures enhance capillary adsorption with the battery cell surface during phase transitions, ensuring close contact during expansion and thus actively reducing thermal resistance, breaking the vicious cycle of "expansion-separation-increased thermal resistance-intensified temperature." 2) Transforming "point / line contact" into "uniform surface constraint": Large-area uniform expansion provides a constraint force that cannot be achieved by traditional point or line fixing methods, bonding to the entire surface of the battery cell, more effectively preventing localized tearing. This achieves synergistic enhancement of thermal management and mechanical constraint during thermal runaway, rather than the failure of traditional solutions.

[0047] S103. Configure a multi-level intelligent pressure relief-purification integrated module for each of the cellular units to perform cascade treatment on the leaked gas.

[0048] The module includes: a pressure-temperature dual-parameter triggered pressure relief valve, a microchannel transient cooler directly connected to the outlet of the pressure relief valve, and a non-precious metal monolithic catalyst unit loaded on a porous metal foam substrate, for the staged treatment of the leaked gas.

[0049] This step equips each cellular unit with a highly integrated secure processing terminal. This terminal module comprises three core components that operate in series: a pressure relief valve that opens under pressure and temperature control; a miniature, high-efficiency cooler immediately following the valve; and a purification unit housing a special non-precious metal catalyst.

[0050] Explanation of proper nouns: Pressure-temperature dual-parameter triggering: This means that the activation conditions of the pressure relief valve take into account both the internal pressure and temperature values. It is more accurate than single pressure triggering and can distinguish between normal pressure fluctuations and sudden pressure rises caused by thermal runaway.

[0051] Microchannel transient cooler: A high-efficiency heat exchange device with microchannels inside, which can rapidly cool the high-temperature fluid flowing through it in a very short time.

[0052] Porous metal foam substrate: a type of metal material with three-dimensional interconnected pores, high porosity, and large specific surface area, often used as a catalyst support.

[0053] Non-precious metal monolithic catalysts: These are catalytic devices that use non-precious metals (such as oxides of copper, manganese, cobalt, etc.) as active components and are integrated with the reactor structure. Their cost is lower than that of precious metal catalysts.

[0054] This step integrates the entire process of "release-cooling-purification" and renders harmless the high-temperature, high-pressure, flammable and toxic gases generated by thermal runaway. It transforms a potentially dangerous emission that could cause secondary fires or poisoning into a relatively safe, low-temperature, and harmless gas emission.

[0055] Traditional pressure relief valves simply release heat, directly venting high-temperature flammable gases into the environment surrounding the battery pack, posing a significant risk. This solution addresses this by implementing a "closed-loop management" approach to thermal runaway products. Its innovation lies in constructing a miniaturized, tiered treatment plant: 1) Dual-parameter triggering ensures precise release: avoiding false triggers and activating only in the event of actual thermal runaway. 2) Transient cooling immediately reduces risk: the gas is rapidly cooled after ejection, significantly reducing the likelihood of igniting surrounding combustibles and protecting subsequent catalysts. 3) Catalytic purification achieves intrinsic safety: the cooled gas passes through a non-precious metal catalyst bed, where combustible carbon monoxide and hydrogen are catalytically oxidized into harmless carbon dioxide and water. This not only eliminates the risk of combustion but also removes toxicity. The entire process is highly integrated into each cellular unit, achieving localized, immediate, and thorough treatment of safety risks, representing the final crucial link in the system's intrinsic safety design.

[0056] Optionally, the macroscopic stiffness can be reduced in a stepwise manner by pre-embedding shape memory alloy wires with different austenitic phase transformation end temperatures (Af) at the key stress nodes of the composite frame. The alloy wires are woven in a three-dimensional interlacing manner in the frame. When the temperature exceeds the Af point of each alloy wire in sequence from low to high, the equivalent elastic modulus of the frame decreases in a stepwise manner, thereby realizing a progressive mechanical response of "rigid support - buffer energy absorption - flexible wrapping" during the thermal runaway of the battery cell.

[0057] This detail illustrates how the intelligent characteristic of "step-down reduction in macroscopic stiffness" of the frame is specifically achieved. The core of this is to pre-embed multiple sets of shape memory alloy wires with different "austenitic phase transformation end temperatures" at key locations bearing the main forces within the composite frame, and then weave these alloy wires in a three-dimensional manner throughout the entire frame network.

[0058] Explanation of proper nouns: Austenitic transformation end temperature (Af): For shape memory alloys, this is the temperature at which they completely transform from the martensitic phase (soft state at low temperatures) to the austenitic phase (hard state at high temperatures and recovery of the shape memory). The Af point differs for alloys with different compositions.

[0059] Equivalent elastic modulus: A physical quantity that measures the ability of a material or structure to resist elastic deformation. The higher the value, the more "rigid" it is, and the lower the value, the more "soft" it is.

[0060] This design endows the battery pack frame with a multi-level mechanical response program that is linked to temperature, much like an "intelligent shock absorber." It is no longer simply a matter of being either "hard" or "soft," but can automatically and in stages adjust its mechanical properties according to the thermal runaway temperature rise process.

[0061] Traditional structures may experience stress concentration due to sudden performance changes under thermal shock. The ingenious logic of this design lies in breaking down a drastic change in mechanical properties into multiple mild and controllable step-like changes. The specific implementation path is as follows: 1) Temperature-encoded mechanical program: By setting Af points from low to high for different alloy wires (e.g., Af1=80°C, Af2=110°C, Af3=140°C), the frame stiffness change is effectively "programmed" with temperature. 2) Progressive response: When thermal runaway occurs, the temperature rises, first reaching the lowest Af1. The corresponding group of alloy wires is activated and contracts, causing the overall frame stiffness to decrease for the first time, transitioning from "rigid support" to "primary buffering" mode, and beginning to absorb energy. As the temperature continues to rise to Af2 and Af3, more alloy wire combinations are activated in sequence, causing the frame stiffness to decrease for the second and third time, gradually transitioning to a "deep energy absorption" and even a "flexible wrapping" mode. This step-by-step unloading and gradual softening strategy can dissipate the enormous energy of the cell expansion extremely smoothly, maximizing the avoidance of brittle fracture of the structure due to instantaneous impact, and achieving precise matching and self-adaptation between mechanical protection and the thermal event process.

[0062] Optionally, the micro / nano secondary structure is a micrometer-scale columnar array constructed in the phase change material layer by a template method, and a nanowire or nanosheet structure grown on the surface of the columnar array. This design is used to enhance the capillary adsorption force with the cell surface through the nanostructure during the phase change process, and to guide the volume expansion direction perpendicular to the cell surface through the microstructure, so as to apply uniform surface pressure.

[0063] This detail clarifies the specific composition and manufacturing method of the "micro-nano secondary structure". The structure consists of two levels: the first level is a regularly arranged array of micrometer-scale columnar structures prepared by the "template method"; the second level is the further growth of nanowire or nanosheet structures on the surface of these micrometer columnar structures.

[0064] Explanation of proper nouns: Template method: A common method for preparing ordered micro and nano structures, which uses a template with a specific pattern to guide the growth or deposition of materials, thereby replicating the morphology of the template.

[0065] Capillary adsorption: The strong adsorption force generated by liquid in narrow gaps at the micro- and nano-scale due to surface tension.

[0066] This design achieves active and directional enhancement of thermal contact and mechanical constraint performance during phase change by precisely controlling the micro-nano scale structure of the phase change material layer.

[0067] Traditional uniform coatings may exhibit uneven expansion upon thermal expansion and are prone to separation from the substrate. The logic behind this design lies in actively guiding and controlling the expansion behavior and interfacial interaction of the phase change material (PCM) through structural design. Its dual-structure has a clear division of labor: 1) Nanostructures enhance interfacial adhesion: The nanowires / sheets grown on the surface significantly increase the actual contact area and roughness with the battery cell surface. When the PCM softens upon heating, the molten precursor generates strong capillary adhesion forces in the nano-interstic gaps, acting like countless "nanoscale anchors," firmly "locking" the coating to the battery cell surface, thereby actively reducing interfacial thermal resistance and ensuring efficient heat dissipation. 2) Microstructures guide the expansion direction: A regularly arranged array of micron-sized pillars constitutes the main volume expansion unit and, like "guide pillars," constrains and guides the expansion primarily along a direction perpendicular to the battery cell surface (i.e., the axis of the pillars). This prevents the material from flowing randomly in all directions, thus highly concentrating and uniformly converting the expansion force into positive compressive stress applied to the entire surface of the battery cell. This collaborative design of "nanoscale adhesion and micron-level directional force application" solves the problems of interface failure and stress dispersion in the application of phase change materials, and maximizes the thermal-mechanical management efficiency.

[0068] Optionally, the microchannel transient cooler is pre-filled with inert working fluid microcapsules capable of undergoing endothermic phase change; the non-precious metal monolithic catalyst is a perovskite-type composite oxide, which has a catalytic oxidation efficiency of not less than 95% for carbon monoxide and hydrocarbons within a temperature window of 300-500°C.

[0069] This detail clarifies the implementation of the two core functional units in the pressure relief-purification integrated module. The microchannel transient cooler is not a traditional cavity; its interior is pre-filled with a special substance—an inert working fluid encapsulated in microcapsules that can undergo an endothermic phase change. The non-precious metal monolithic catalyst specifically refers to a class of composite oxide materials with a perovskite-type crystal structure, and specifies the extremely high catalytic efficiency it must achieve within the high operating temperature range.

[0070] Explanation of proper nouns: Endothermic phase change inert working fluid microcapsules: These are particles formed by encapsulating a working fluid that undergoes a phase change at a specific temperature (such as melting from a solid to a liquid, or vaporizing from a liquid to a gas) and absorbs a large amount of heat in the process (latent heat of phase change), and which is chemically stable and does not participate in the reaction, within a micron-sized polymer or inorganic shell.

[0071] Perovskite-type composite oxides: a class of composite metal oxide materials whose atoms are arranged according to a specific ABO3 structure. Their crystal structure is stable, and their electronic structure and catalytic performance can be flexibly controlled by adjusting the types of A and B site ions, making them an important system for high-performance non-noble metal catalysts.

[0072] Catalytic oxidation efficiency: The percentage of harmful gas molecules (such as carbon monoxide CO, hydrocarbons CxHy) that react with oxygen under the action of a catalyst and are completely converted into harmless carbon dioxide (CO2) and water (H2O).

[0073] This material and design choice together ensures that the pressure relief module can achieve "millisecond-level step cooling" and "near-complete transformation of harmful substances at high temperatures," thereby simultaneously reducing the thermal hazards and chemical toxicity of the thermal runaway ejecta to an acceptable safety level.

[0074] Traditional gas cooling relies on heat exchange between the gas and the wall (sensible heat cooling), which is slow and inefficient. This solution employs the principle of latent heat of phase change cooling, representing a qualitative leap in cooling technology. When high-temperature gas flows through a cooler filled with microcapsules, the working fluid inside the capsules rapidly absorbs heat from the gas and undergoes a phase change (e.g., from solid to liquid). The latent heat absorbed during the phase change process far exceeds the sensible heat absorbed by simply increasing the temperature, thus removing a massive amount of heat energy in a very short time, achieving a rapid, "step-like" drop in gas temperature.

[0075] This can cool the injected gas, which may reach temperatures of 800°C or higher, to 300°C or even lower within milliseconds. The direct effects are: a) greatly reducing the risk of the high-temperature gas itself or igniting surrounding combustibles; b) creating a suitable temperature window (300-500°C) for subsequent catalytic reactions, protecting the catalyst from overheating and sintering deactivation.

[0076] While noble metal catalysts are highly efficient, their high cost makes large-scale application difficult. This solution utilizes an optimized perovskite-type composite oxide, which is innovative in that: a) Intrinsically high activity: its unique crystal structure facilitates the formation of oxygen vacancies, resulting in excellent catalytic activity for the oxidation of CO and hydrocarbons. b) High thermal stability: the perovskite structure is not easily destroyed at high temperatures, ensuring structural and performance stability even under the high-temperature impact of thermal runaway gases. c) Integral design: the catalyst is directly grown on a porous metal foam framework, resulting in low gas flow resistance and full exposure of active sites.

[0077] Within the optimal operating temperature range of 300-500°C, optimized by the pre-cooler, the catalyst ensures that over 95% of harmful gases (CO, CxHy) are converted into non-toxic CO2 and H2O. This means that the gases discharged from the system are mainly nitrogen, carbon dioxide, and water vapor, fundamentally eliminating the risk of combustion of combustible gases and the risk of CO poisoning, thus achieving inherent safety of emissions.

[0078] Cooling and purification are not isolated processes, but rather deeply synergistic. "Transient phase change cooling" is the prerequisite and guarantee for "high-efficiency catalytic purification," providing the catalyst with a stable and efficient operating temperature environment. Conversely, "high-efficiency catalytic purification" is the value extension and ultimate safety loop of "transient cooling," completely rendering the cooled gas harmless. The combination of the two constitutes a complete, efficient, and reliable exhaust gas safety treatment chain from physical cooling to chemical transformation, which is the key technological support for achieving the top-level goal of "purification upon release, safety upon emission" in this solution.

[0079] Figure 2 A schematic flowchart of a method for suppressing thermal runaway of a robot battery pack according to an embodiment of this specification is shown; as follows: Figure 2 As shown. This specification provides a method for suppressing thermal runaway in a robot battery pack, characterized in that it is applied to a robot battery pack manufactured by the method described in any one of the first aspects, the method comprising: S201. Based on the geometric structure, material properties, and cell chemical model of the battery pack, construct a high-fidelity digital twin that includes multi-field coupling of electrochemistry, heat, stress, fluid, and chemical reaction, and train a lightweight proxy model for real-time inference.

[0080] This step is crucial for creating the "virtual brain" and "rapid-response core" of the entire prediction system. First, based on precise 3D blueprints of the battery pack, the physicochemical properties of all materials, and the electrochemical reaction principles of the battery cells, a highly realistic virtual battery pack model is built in a computer. This model is not a static blueprint, but a complex dynamic system capable of synchronously simulating changes in current and voltage, heat generation and transfer, structural stress and deformation, gas and fluid flow, and internal chemical reactions—a system called a "digital twin." Simultaneously, to meet the stringent computational speed requirements of real-time control, a simplified model—a "lightweight proxy model"—with minimal computational cost but maintaining key predictive capabilities, needs to be trained based on this high-fidelity twin.

[0081] Explanation of proper nouns: Digital twin: refers to a digital model constructed in a virtual information space that is fully mapped to and interacts with a physical entity in real time, and can reflect the entire life cycle of the entity.

[0082] Multi-field coupling refers to the situation in simulation where the equations and variables of multiple physical fields (such as electric field, temperature field, and stress field) are interrelated and mutually influential, and must be solved simultaneously to simulate the complex interactions in the real world.

[0083] Proxy model: A highly computationally efficient approximate model built based on original high-precision model data or physical laws, used to replace the original complex model for rapid analysis and optimization under acceptable accuracy conditions.

[0084] This step established a combination of a "high-fidelity digital lab" and an "embedded predictive chip." The high-fidelity twin is used for offline deep analysis and strategy verification, while the lightweight agent model is embedded in the actual battery management system, providing millisecond-level online extrapolation capabilities.

[0085] Traditional battery safety monitoring relies on the direct interpretation of data from a limited number of sensors (such as voltage and temperature), failing to "see" or "understand" the complex microscopic physicochemical processes inside the battery. This solution aims to create a "digital avatar" for the physical battery. This avatar can "run" and "preview" everything happening inside in real time. Its innovation lies in: 1) Panoramic virtual perspective: The digital twin enables the "visualization" and "computability" of states inside the battery that cannot be directly observed (such as internal temperature gradients, lithium-ion concentration, and local stress), breaking through the limitations of physical sensors. 2) Balancing computational efficiency: Through an architecture of "high-fidelity model + lightweight agent," the physical accuracy of model predictions is guaranteed while meeting the stringent requirements of real-time control systems for computational latency, making online prediction possible. This lays a crucial technological foundation for leaping from "perceiving appearances" to "understanding the essence and predicting the future."

[0086] S202. During the operation of the battery pack, the state of the physical entity and the digital twin are synchronized in real time, and the proxy model is used to predict the time-series change curve of the "thermal runaway risk entropy" of each cell in the battery pack within a future time window, using the state vector of the previous moment as input.

[0087] This step is the core operational process of the prediction system. While the battery pack is operating, readings from actual sensors (such as voltage, temperature, and pressure) are continuously input into the digital twin and its surrogate model to correct the current state of the virtual model, ensuring it aligns with the state of the real battery pack. This process is called "data assimilation." Then, starting from the current synchronized state, a lightweight surrogate model is used to continuously and iteratively predict how a comprehensive risk indicator called "thermal runaway risk entropy" will change over time within each cellular unit over a short period in the future.

[0088] Explanation of proper nouns: Data assimilation: A technique that dynamically combines observational data with mathematical models to correct model states, reduce prediction errors, and make models more realistic.

[0089] State vector: A mathematical set containing all variables (such as temperature, voltage, stress, etc. at each point) needed to describe the current complete state of the system.

[0090] Rolling forecasting: a continuous forecasting method that, with each new observation data obtained, re-predicts the situation for a future period starting from the latest state, and continuously updates the forecast results.

[0091] Thermal runaway risk entropy: an original, comprehensive indicator used to quantify the probability and urgency of thermal runaway. The higher the value, the greater the probability and speed at which the unit will experience severe thermal runaway within a short period of time.

[0092] This step enables continuous and quantitative prediction and visualization of future safety risks in various regions inside the battery. Its output is no longer a simple "normal / abnormal" alarm, but rather a series of predictive curves showing the risk evolution of each cell.

[0093] Traditional early warning systems rely on fixed thresholds (e.g., alarms for temperatures > 60℃), but the risk of thermal runaway in batteries is the result of dynamic coupling of multiple factors. This solution replaces isolated, static physical quantity thresholds with a dynamically evolving comprehensive indicator called "risk entropy." Its advantages are: 1) Prediction rather than detection: It doesn't only alarm when danger has already occurred (temperature is already high), but rather infers the risk level tens of seconds to minutes in advance based on the dynamic trends of the current state. 2) Comprehensive rather than singular: "Risk entropy" integrates multiple precursor signals such as temperature change rate, gas production trend, and inter-cell thermal interaction, enabling it to capture signs of system instability earlier and more reliably than a single temperature parameter. 3) Spatial resolution: It can perform independent risk assessment and prediction for each cellular cell, achieving risk localization. This marks a new stage in battery safety management, moving from "diagnosing based on current symptoms" to "predicting disease progression based on current signs."

[0094] S203. Define and dynamically update an adaptive risk probability threshold related to the current health status and historical operating conditions of the battery.

[0095] This step sets a dynamic, intelligent "early warning threshold" for the prediction system. This "risk probability threshold," used to determine whether to trigger a suppression action, is not a fixed value, but is dynamically calculated and adjusted based on the current health of the battery pack (such as capacity decay and increased internal resistance) and the recent workload (such as whether it has just undergone high-power discharge or been in a high-temperature environment).

[0096] Explanation of proper nouns: Adaptive threshold: A critical value that can be automatically adjusted according to changes in the system's own state or the external environment, making it more environmentally adaptable and robust than a fixed threshold.

[0097] Battery health status: used to characterize the degree of degradation of battery performance relative to its brand new state, usually expressed as a percentage of capacity or internal resistance.

[0098] This step endows the early warning system with the intelligence of "contextual awareness" and "personalized judgment." It enables the system to understand that the same predicted risk entropy value has completely different safety implications for a new battery and an aging battery, or for a battery that has just experienced harsh operating conditions and one that is in mild operating conditions.

[0099] Using fixed risk thresholds can lead to either overly conservative approaches resulting in frequent false alarms or insufficient sensitivity causing underreporting of real risks. This solution links warning standards to the battery's "physiological state" and "fatigue level." Its innovation lies in: 1) Health status correlation: When the battery's health declines (aging), its ability to withstand abuse and side reactions weakens. The system automatically lowers the risk threshold, adopting a more conservative warning strategy and becoming more sensitive to early risks. 2) Historical operating condition correlation: If the battery has recently experienced harsh operating conditions (such as high-rate discharge), its internal materials may be in a metastable or damaged state, making it more susceptible to thermal runaway. The system will also dynamically lower the threshold and raise the alert level. This achieves personalized and refined warning strategies, significantly improving the accuracy and reliability of warnings and reducing false alarms or underreporting caused by rigid standards.

[0100] S204. When the predicted "thermal runaway risk entropy" of a certain cellular unit exceeds the adaptive risk probability threshold within the time window, the predictive suppression decision process is triggered.

[0101] Based on the simulation results of the digital twin, the process optimizes and generates the optimal inhibition parameters for the unit. The parameters include at least the type, dosage, release rate curve, and co-phase relationship with the charging current of the inhibitor.

[0102] This step is the decision-making and action output stage of the entire prediction system. When step S202 predicts that the risk entropy curve of a certain cellular cell will exceed the dynamic threshold set in step S203 within a future time window, the system determines that the cell has a high probability of thermal runaway. At this time, the system will not immediately and simply release the extinguishing agent, but will initiate a complex "predictive suppression decision-making process". This process calls a high-fidelity digital twin, takes the current state and the predicted risk evolution path as input, performs rapid optimization calculations, and generates a set of tailored optimal suppression scheme parameters for that specific cell, instructing the execution mechanism to execute precisely before the predicted "risk outbreak point".

[0103] Explanation of proper nouns: Suppression parameters: These refer to the specific adjustable variables that control the suppression action, such as which extinguishing agent is used, the amount used, the release rate, and the timing of release in relation to the external charging current.

[0104] Co-phase relationship: refers to the precise coordination between the release action of the inhibitor and the battery charging current waveform on the time axis, such as release at a specific moment of the charging current pulse.

[0105] This step creates a closed loop from "risk prediction" to "proactive, precise, and optimized intervention." Its output is no longer a general "firefighting" command, but a "customized prescription" for specific risk scenarios.

[0106] Traditional firefighting is a "post-disaster response," while the ultimate logic of this solution is to "pre-simulate disasters in virtual space and prevent disasters in physical space." Its revolutionary nature is reflected in: 1) Proactive intervention: Action triggering is based on "risk prediction" rather than "disaster confirmation," gaining a valuable window of time that may prevent disasters from occurring. 2) Precise customization: Utilizing the simulation capabilities of digital twins, the "optimal" combination of suppression parameters that maximizes suppression effectiveness and minimizes side effects can be calculated for specific predicted risk scenarios (such as heating rate and expected eruption time). For example, the most suitable extinguishing agent dosage can be calculated to avoid over- or under-dosing. 3) Multi-field coordination: Considering the coordinated phase with the charging current is a very high-order control strategy. By releasing the current at a specific phase, the coupling effect of the current field and the flow field can be used to optimize the penetration and distribution of inhibitors in the porous structure of the electrode, thereby greatly improving suppression efficiency. This marks a complete shift in battery safety from "disaster remediation" to "risk surgical elimination," elevating the initiative, precision, and intelligence of safety control to an unprecedented level.

[0107] Optionally, the thermal runaway risk entropy integrates the dimensionless indices of the predicted temperature gradient, internal pressure, characteristic gas yield, and inter-unit thermal interaction intensity within the unit, and its calculation incorporates the output of the physical model and the neural network correction term trained based on historical fault data.

[0108] This detail clarifies the specific composition and calculation method of the core early warning indicator, "thermal runaway risk entropy." This indicator is a comprehensive, unitless number whose value is determined by four key prediction parameters: the temperature difference distribution within the unit, pressure changes, the rate of specific gas generation, and the intensity of heat transfer between the unit and adjacent units. Its final calculated value is not a simple weighted sum, but rather a fusion of model predictions based on physical laws and the corrected output of an artificial intelligence neural network trained on a large amount of historical failure case data.

[0109] Explanation of proper nouns: Dimensionless quantities are pure numbers without physical units. They are obtained by combining multiple physical quantities with units through a certain mathematical combination (such as ratios or calculations using specific formulas), which facilitates comparisons across scenarios and scales and the setting of uniform thresholds.

[0110] Characteristic gas yield: refers to the amount of specific characteristic gases (such as CO, H2, and electrolyte solvent gases) produced per unit time in the early stages of battery thermal runaway, and is a key indicator of the severity of side reactions.

[0111] Thermal interaction intensity: an indicator that quantifies the rate and total amount of heat transfer between two adjacent units due to the temperature difference.

[0112] This definition creates a "super warning signal" that far surpasses a single parameter, enabling earlier and more reliable characterization of system instability precursors. It provides a comprehensive numerical representation of a battery cell's overall "instability tendency," encompassing heat accumulation, internal pressure buildup, chemical reactions, and interactions with the surrounding environment.

[0113] Traditional early warning systems rely on single or a few parameters (e.g., temperature > 60℃ at a certain point). However, thermal runaway is the result of the malignant coupling of multiple physicochemical processes, and a single parameter may be delayed or produce false alarms. This solution uses a "multi-parameter fusion decision" to replace a "single-parameter threshold decision." Its advantages are: 1) Comprehensiveness: Simultaneously monitoring anomalies in four dimensions—thermal, mechanical, chemical, and diffusion—avoids a "blind men and the elephant" scenario; any early anomaly in any aspect will be captured and contributed to the risk entropy. 2) Advancement: For example, characteristic gas yields and inter-unit thermal interaction intensity can show anomalies before a significant rise in overall temperature, thus achieving earlier warnings. 3) Intelligent correction: Introducing a neural network trained on historical fault data for correction, enabling this theoretically based physical model to absorb "hidden knowledge" and experience that are difficult to model in real complex systems, bringing its early warning accuracy and reliability closer to, or even surpassing, human expert experience. This achieves a leap from "simple, delayed, and prone to false alarms" to "comprehensive, advanced, and highly reliable" early warning signals.

[0114] Optionally, the update logic of the adaptive risk probability threshold is as follows: when the battery is in a high-rate discharge, high ambient temperature, or has recently experienced a large mechanical shock, the threshold is automatically lowered to trigger a more conservative early warning and suppression strategy.

[0115] This detail clarifies the specific rules for the system's automatic adjustment of the "alert level." The system continuously monitors the battery's operating conditions. When it detects that the battery is in or has just experienced one of these specific severe conditions—such as high current discharge, excessively high ambient temperature, or a severe mechanical impact—the "risk probability threshold" used to determine whether to trigger suppression will be automatically lowered.

[0116] Explanation of proper nouns: High-rate discharge: refers to discharging with a current that is very large relative to the battery capacity, which will exacerbate internal polarization, heat generation and material stress in the battery.

[0117] Adaptive risk probability threshold: refers to the risk entropy threshold that serves as the standard for triggering early warnings. This value can be automatically adjusted according to the system status, rather than remaining fixed.

[0118] This logic endows the safety system with "contextual awareness," enabling it to intelligently adjust its safety sensitivity based on the battery's "fatigue state" and "stress history," and to adopt a more conservative defense strategy when the battery is "vulnerable."

[0119] Fixed thresholds cannot adapt to the dynamic changes in a battery's "physical condition." This solution establishes a correlation model between "operating conditions, battery physical condition, and threshold," based on the following scientific principles: 1) High-rate discharge and high ambient temperatures directly lead to increased internal heat accumulation and accelerated side reaction kinetics, significantly lowering the "threshold" for inducing thermal runaway. 2) Large mechanical shocks can cause internal micro-short circuits and structural micro-damage, directly creating the conditions for thermal runaway. In these three situations, the battery is in a "highly vulnerable" state; the same internal state change (the same risk entropy) actually represents a higher probability of real danger. Therefore, by actively lowering the threshold, the system effectively raises the alert level, enabling intervention at an earlier stage with fewer risk indicators. This achieves an evolution in safety strategies from a "one-size-fits-all" approach to "time-specific, personalized protection," significantly improving the accuracy and robustness of safety protection under complex real-world conditions.

[0120] Optionally, the coordinated phase relationship with the charging current specifically refers to: triggering the release of the inhibitor at a specific phase (such as a peak or trough) of the AC or pulse charging current, and utilizing the coupling effect of the current field and the stress field to enhance the penetration and uniform distribution of the inhibitor in the porous electrodes inside the cell.

[0121] This detail specifies a very high-order control strategy for implementing predictive suppression: the release of the inhibitor needs to be precisely synchronized in time with the battery's charging current waveform. Specifically, the system is designed to trigger release at a specific phase point of the AC charging current or pulse charging current (e.g., the instant the current reaches its maximum or minimum value). The aim is to optimize the suppression effect by utilizing the interaction between the electric field generated by the current and the fluid stress field that may be induced by the release of the inhibitor.

[0122] Explanation of proper nouns: Phase: In a periodically changing waveform, a parameter used to describe the position of a specific point within one cycle.

[0123] The coupling effect between the current field and the stress field refers to the interaction between the flow of charge (current field) and the flow of electrolyte or gas, or pressure changes (stress field) inside a battery. For example, current affects ion migration, thereby indirectly affecting the permeation behavior of fluids.

[0124] This control strategy enables the active guidance and enhancement of the diffusion path and distribution mode of the inhibitor within the battery, thereby maximizing the "fire extinguishing" or "cooling" efficiency of the inhibitor while reducing the required dosage.

[0125] Traditional fire suppression involves external spraying, which has poor penetration into the porous and tortuous electrode structure inside the battery. This solution utilizes the inherent physical field within the battery as a "delivery guide" and "efficiency-enhancing tool." Its ingenuity lies in: 1) Utilizing the current field to drive penetration: At the peak or trough of the charging current, the effects of ion migration and electroosmotic flow within the battery are most significant. Releasing liquid inhibitors at this time allows the current field to act like an "invisible pump," drawing and driving inhibitor droplets or vapors deeper into the micropores of the electrodes, reaching the reaction hotspots, rather than merely remaining on the surface. 2) Optimizing distribution uniformity: Specific current phases may be associated with the expansion / contraction state of the electrodes; releasing inhibitors at this time helps to form a more uniform coverage. This achieves a fundamental change from "blind spraying" to "field-assisted precise delivery," resulting in an order-of-magnitude improvement in the efficiency of internal suppression. This is a key technological guarantee that predictive suppression can achieve maximum effect at minimal cost.

[0126] Figure 3 This specification illustrates a power management method for a robot battery pack according to embodiments thereof; such as Figure 3 As shown. Thirdly, this specification provides a power management method for a robot battery pack, characterized in that it is applied to a robot integrating a robot battery pack as described in the first aspect and a thermal runaway suppression method for the robot battery pack as described in the second aspect, the method being executed by a power domain controller, comprising: S301. Receive and fuse information from multiple sources to generate a global understanding of the dynamic system state.

[0127] The multi-source information includes at least: a real-time thermal runaway risk entropy distribution map from the predictive suppression system, the real-time output power peak of the battery pack, the real-time posture and joint load of the robot body, and the future trajectory and action intention from the upper-level task planner.

[0128] This step is the core perception and information fusion stage of the dynamic domain controller. The controller continuously receives real-time or forward-looking data from four key sources: 1) the distribution map of "thermal runaway risk entropy" from the predictive suppression system; 2) the real-time output peak power of the battery pack provided by the battery management system; 3) the real-time posture and joint load feedback from the robot's sensor network; and 4) the predetermined motion trajectory and action sequence for a future period issued by the upper-level task planner. The controller fuses, aligns, and interprets this heterogeneous, multi-source, and time-varying information to form a unified and comprehensive understanding of the current and near-term state of the dynamic system, namely, "global state cognition."

[0129] Explanation of proper nouns: Global dynamic system state cognition: refers to a comprehensive and structured digital representation that integrates energy state, safety state, mechanical state and mission intent, and serves as the data foundation for high-level decision-making.

[0130] Thermal runaway risk entropy distribution map: a spatial distribution information that reflects the real-time thermal runaway risk probability and trend of each region (cell cell) within the battery pack, presented in the form of visualization or data matrix.

[0131] Action intent: refers to the specific sequence of operations that the task planner sets for the robot to execute, such as "accelerate forward", "raise the robotic arm", "rotate the gimbal", etc., which usually includes a description of the power requirements.

[0132] This step enables "holographic situational awareness" of the power system's operating environment. It breaks down information silos between subsystems such as batteries, mechanics, and missions, constructing a four-dimensional panoramic view that includes "energy supply capability," "real-time safety situation," "physical state of the system," and "future mission requirements."

[0133] In traditional robot control, batteries, motors, controllers, and task planners typically operate independently or only communicate in simple ways, leading to fragmented decision-making. This solution's logic lies in implementing cross-domain information fusion to establish a unified, high-dimensional "situation map" for decision-making. Its advantages are: 1) Introducing a safety situation dimension: Using predictive "risk entropy" as a key input, it endows power management with a proactive "safety awareness" to avoid risks—a dimension completely lacking in traditional systems. 2) Integrating future intentions: Knowing not only what the robot is "doing," but also what it "is about to do," it enables power management to have forward-looking planning capabilities, allowing it to reserve energy for high-power actions or adjust thermal management strategies in advance. 3) Unifying spatiotemporal references: Unifying all information to the same timestamp and spatial coordinate system ensures consistency in decision-making. This provides accurate, comprehensive, and timely "battlefield intelligence" for subsequent global optimization decisions.

[0134] S302. Based on the global state cognition, solve a rolling optimization problem with multiple objective functions such as task completion degree, system safety margin, and energy consumption, and output a dynamically executable instruction set in real time.

[0135] The instruction set includes at least an adaptive torque-speed limit curve issued to the motion controller, dynamic pump speed and valve opening commands issued to the thermal management system, and task degradation or reconfiguration suggestions triggered when necessary.

[0136] This step is the core decision-making and instruction generation stage of the dynamic domain controller. Based on the global state awareness generated by S301, the controller solves a complex optimization problem in real time within each control cycle. This optimization problem simultaneously considers three core objectives: maximizing task completion, maximizing system safety margin, and minimizing overall energy consumption. The result of the solution is not a single instruction, but a set of dynamically executable instructions, which includes three types of outputs: 1) adaptive torque and speed limit curves that vary over time and are sent to the robot motion controller; 2) control instructions sent to the thermal management system for dynamically adjusting coolant flow and valve opening; and 3) degradation or refactoring suggestions to the upper-level task planner, aimed at adjusting task objectives or paths, when system capabilities are limited due to safety or faults.

[0137] Explanation of proper nouns: Multi-objective rolling optimization: an optimization method that recalculates a series of control actions over a finite future time period based on the latest state in each control cycle, aiming to simultaneously optimize multiple (potentially conflicting) objective functions.

[0138] Adaptive torque-speed limit curve: refers to the rules that dynamically limit the output capacity (torque) and maximum speed of the robot's drive motor. This limit curve changes in real time according to the system state, rather than being a fixed value.

[0139] Task downgrade / refactoring suggestions: These refer to alternative task solutions automatically generated by the system when the original high-requirement task cannot be fully executed. Examples include reducing the accuracy of the task, slowing down the movement speed, or planning a smoother path.

[0140] This step represents a leap in power system control from "rule-driven, single-objective optimization" to "model prediction, multi-objective collaborative optimization." It outputs a comprehensive control scheme that achieves the optimal real-time balance between safety, performance, and energy efficiency.

[0141] In traditional control, safety, performance, and energy consumption are often conflicting objectives. Typically, fixed, conservative rules prioritize safety at the expense of performance. This solution formalizes the dynamic trade-off between multiple objectives into a real-time solvable mathematical optimization problem. Its innovations lie in: 1) Dynamic safety boundary: The "safety margin" objective in the optimization problem is directly coupled to the "risk entropy distribution map" and battery state. When the risk entropy increases, the optimization algorithm automatically tends to generate a more conservative torque-speed limit curve, smoothly shrinking the performance boundary in real-time to ensure safety, rather than abruptly cutting it off. 2) Proactive energy-thermal synergy: Based on future mission intent, the optimization algorithm can pre-schedule the thermal management system (e.g., increase cooling power in advance) or adjust the motor control strategy to ensure that battery and motor temperatures do not exceed limits during high-power operations. 3) Graceful degradation and system resilience: When optimization calculations find that the performance requirements of the original mission cannot be fully met under current safety constraints, it does not report an error and shut down, but proactively generates a "mission degradation suggestion." This gives the system the "resilience" to maintain core functions and continue operating when faced with internal failures (such as partial battery cell isolation) or external challenges. This marks the evolution of robot power management from "best effort, stop when encountering obstacles" to a higher-order form of "global optimization, adapt when encountering obstacles".

[0142] Optionally, the task degradation or reconfiguration suggestion is as follows: when the predictive suppression system is triggered or a battery pack unit is isolated, the power domain controller, based on the updated system capability model, automatically generates a degradation operation plan that maintains the highest possible task completion rate while ensuring core safety constraints, and submits it to the upper-level task planner for confirmation and execution through a standard interface.

[0143] In summary, the method provided in this specification constructs a new paradigm of robot dynamic system with inherent intelligence through deep interlocking and functional coupling of three levels of creation points, generating a powerful synergistic multiplication effect.

[0144] The vertical integration and enhancement of technological effects: the adaptive structural battery pack provides the physical carrier for intelligent response and precise execution; the predictive suppression method provides the intelligent core for risk insight and generation strategy; and the dynamic domain collaborative management provides the system-level brain for overall coordination and value realization. The output of each layer is the input of the next layer, and the function of each layer is multiplied by the existence of the next layer, forming a complete intelligent chain from microscopic material response to macroscopic system behavior.

[0145] Systemic Synergistic Effects in Solving Core Problems: This solution systematically addresses the two major pain points of "safety lag" and "lack of coordination." Its synergistic effects are reflected in: 1) Advancing the safety closed-loop in time and space: The combination of structural adaptation and predictive suppression enables a leap from "post-disaster emergency response" to "in-disaster intervention" and then to "pre-disaster prevention," significantly advancing the safety time window. 2) Dynamic unification of performance and safety: Dynamic domain collaborative management transforms the predicted safety boundary into a dynamic performance boundary in real time, enabling the robot to consistently perform at its maximum potential within the safety red line in complex dynamic environments, transforming contradictions into unity. 3) Fundamental construction of system survivability: When failures are unavoidable, a three-layered, progressive resilience defense is formed, from partial hardware isolation to predictive fault tolerance in software, and then to system task degradation, ensuring the robot survives "with damage" and continues to complete its core mission. This is crucial for the reliability of unmanned systems. The entire solution constitutes a complete and leading solution that combines "hard power" (adaptive structure), "soft intelligence" (predictive algorithm), and "strong coordination" (dynamic domain control).

[0146] The foregoing has described specific embodiments of this specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than that shown in the embodiments and may still achieve the desired result. Furthermore, the processes depicted in the drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0147] In summary, after reading this detailed disclosure, those skilled in the art will understand that the foregoing detailed disclosure is presented by way of example only and is not restrictive. Although not explicitly stated herein, those skilled in the art will understand that this specification requires various reasonable changes, improvements, and modifications to the embodiments. These changes, improvements, and modifications are intended to be made by this specification and are within the spirit and scope of the exemplary embodiments described herein.

[0148] Furthermore, certain terms in this specification have been used to describe embodiments of this specification. For example, "an embodiment," "an embodiment," and / or "some embodiments" mean that a particular feature, structure, or characteristic described in connection with that embodiment may be included in at least one embodiment of this specification. Therefore, it is to be emphasized and understood that two or more references to "an embodiment" or "an embodiment" or "alternative embodiment" in various parts of this specification do not necessarily refer to the same embodiment. Moreover, specific features, structures, or characteristics may be suitably combined in one or more embodiments of this specification.

[0149] It should be understood that in the foregoing description of the embodiments in this specification, various features are combined in a single embodiment, drawing, or description for the purpose of simplifying the description and to aid in understanding a feature. However, this does not mean that the combination of these features is necessary, and those skilled in the art, upon reading this specification, may readily identify some of the devices as separate embodiments. That is, the embodiments in this specification can also be understood as an integration of multiple secondary embodiments. And the content of each secondary embodiment is valid even if it contains fewer than all the features of a single foregoing disclosed embodiment.

[0150] Every patent, patent application, publication of a patent application, and other material such as articles, books, specifications, publications, documents, articles, etc., cited herein, except for those inconsistent with or conflicting with this document, or those having a restrictive effect on the widest scope of the claims, may be incorporated herein by reference for all purposes now or hereafter associated with this document. Furthermore, in the event of any inconsistency or conflict between the description, definition, and / or use of relevant terms in any material and the description, definition, and / or use of relevant terms in this document, the terms in this document shall prevail.

[0151] Finally, it should be understood that the embodiments disclosed herein are illustrative of the principles of the embodiments described in this specification. Other modified embodiments are also within the scope of this specification. Therefore, the embodiments disclosed in this specification are merely examples and not limitations. Those skilled in the art can implement the applications described in this specification using alternative configurations based on the embodiments in this specification. Therefore, the embodiments in this specification are not limited to the embodiments precisely described in the applications.

Claims

1. A method for manufacturing a robot battery pack, characterized in that, include: A honeycomb-shaped compartmentalized structure is constructed inside the battery pack. The compartmentalized structure is composed of a three-dimensional frame made of shape memory alloy and high-damping elastomer. The frame is configured such that its macroscopic stiffness can decrease in a step manner within a specific temperature range. Each honeycomb unit is used to independently accommodate a battery cell module. On the inner wall of each of the cellular cells, a solid phase change material layer with a micro-nano secondary structure is laminated by a vapor deposition process. When the phase change material layer reaches the phase change temperature, its volume expansion rate is greater than 10% and the interfacial thermal resistance in surface contact with the cell surface decreases by at least 50%. Each of the cellular units is configured with a multi-stage intelligent pressure relief-purification integrated module, which includes: a pressure-temperature dual-parameter triggered pressure relief valve, a microchannel transient cooler directly connected to the outlet of the pressure relief valve, and a non-precious metal integral catalyst unit loaded on a porous metal foam substrate, for the staged treatment of the leaked gas.

2. The method according to claim 1, characterized in that, The stepwise decrease in macroscopic stiffness is achieved by pre-embedding shape memory alloy wires with different austenitic phase transformation end temperatures (Af) at the key stress nodes of the composite frame. The alloy wires are woven in a three-dimensional interlacing manner in the frame. When the temperature exceeds the Af point of each alloy wire in turn from low to high, the equivalent elastic modulus of the frame decreases in a stepwise manner, thereby achieving a progressive mechanical response of "rigid support - buffer energy absorption - flexible wrapping" during the thermal runaway of the battery cell.

3. The method according to claim 1, characterized in that, The micro / nano secondary structure is a micrometer-scale columnar array constructed in the phase change material layer by a template method, and a nanowire or nanosheet structure grown on the surface of the columnar array. This design is used to enhance the capillary adsorption force with the cell surface through the nanostructure during the phase change process, and to guide the volume expansion direction perpendicular to the cell surface through the microstructure, so as to apply uniform surface pressure.

4. The method according to claim 1, characterized in that, The microchannel transient cooler is pre-filled with inert working fluid microcapsules that can undergo endothermic phase change; the non-precious metal monolithic catalyst is a perovskite-type composite oxide, which has a catalytic oxidation efficiency of not less than 95% for carbon monoxide and hydrocarbons within a temperature window of 300-500°C.

5. A method for suppressing thermal runaway in a robot battery pack, characterized in that, Applied to a robot battery pack manufactured by the method according to any one of claims 1-4, the method comprising: Based on the geometry, material properties, and cell chemistry model of the battery pack, a high-fidelity digital twin containing multi-field coupling of electrochemistry, heat, stress, fluid, and chemical reaction is constructed, and a lightweight proxy model is trained for real-time inference. During the operation of the battery pack, the state of the physical entity and the digital twin are synchronized in real time. Using the proxy model, the state vector of the previous moment is used as input to predict the time-series change curve of the "thermal runaway risk entropy" of each cell in the battery pack within a future time window. Define and dynamically update an adaptive risk probability threshold related to the current health status and historical operating conditions of the battery; When the predicted "thermal runaway risk entropy" of a certain cellular cell exceeds the adaptive risk probability threshold within the time window, a predictive suppression decision process is triggered. This process optimizes and generates the optimal suppression parameters for the cell based on the simulation results of the digital twin. The parameters include at least the type, dosage, release rate curve, and co-phase relationship with the charging current of the inhibitor.

6. The method according to claim 5, characterized in that, The thermal runaway risk entropy is a dimensionless index that integrates the predicted temperature gradient, internal pressure, characteristic gas yield, and inter-unit thermal interaction intensity within the unit. Its calculation combines the output of the physical model with the correction term of the neural network trained based on historical fault data.

7. The method according to claim 5, characterized in that, The update logic for the adaptive risk probability threshold is as follows: when the battery is in a high-rate discharge, high ambient temperature, or has recently experienced a large mechanical impact, the threshold is automatically lowered to trigger a more conservative early warning and suppression strategy.

8. The method according to claim 5, characterized in that, The specific phase relationship with the charging current is as follows: the release of the inhibitor is triggered at a specific phase (such as peak or trough) of the AC or pulse charging current, and the coupling effect of the current field and stress field is used to enhance the penetration and uniform distribution of the inhibitor in the porous electrodes inside the cell.

9. A power management method for a robot battery pack, characterized in that, A system applied to a robot integrating the robot battery pack as described in claims 1-4 and the thermal runaway suppression method for the robot battery pack as described in claims 5-8, wherein the method is executed by a power domain controller, including: The system receives and integrates information from multiple sources to generate a global dynamic system state cognition. The information includes at least: a real-time thermal runaway risk entropy distribution map from the predictive suppression system, the real-time output power peak of the battery pack, the real-time posture and joint load of the robot body, and the future trajectory and action intention from the upper-level task planner. Based on the global state cognition, a rolling optimization problem with multiple objective functions such as task completion, system safety margin, and energy consumption is solved, and a set of dynamically executable instructions is output in real time. The instruction set includes at least an adaptive torque-speed limit curve issued to the motion controller, dynamic pump speed and valve opening instructions issued to the thermal management system, and task degradation or reconfiguration suggestions triggered when necessary.

10. The method according to claim 9, characterized in that, The task degradation or reconfiguration suggestion is as follows: when the predictive suppression system is triggered or a certain unit of the battery pack is isolated, the power domain controller, based on the updated system capability model, automatically generates a degradation operation plan that maintains the highest possible task completion rate under the premise of ensuring core safety constraints, and submits it to the upper-level task planner for confirmation and execution through a standard interface.

Images