Wide-temperature-range lithium ion battery system for robots and thermal management method thereof

Abstract

The application provides a wide-temperature-range lithium ion battery system for robots and a thermal management method thereof, and relates to the technical fields of energy storage and robot engineering. The wide-temperature-range lithium ion battery system for robots and the thermal management method thereof comprise the following steps: based on real-time data of an environmental temperature sensor, an enhanced learning algorithm is adopted, the mixing ratio of a low-freezing-point solvent and a high-temperature film-forming additive in electrolyte is dynamically adjusted through a microfluidic chip, the variation of internal resistance measured by electrochemical impedance spectroscopy is minimized, and adaptive electrolyte mixing ratio parameters are generated. By dynamically adjusting the mixing ratio of the low-freezing-point solvent and the high-temperature film-forming additive in the electrolyte, the electrochemical impedance characteristics are optimized by the enhanced learning algorithm, the temperature sensitivity of the electrolyte viscosity is significantly reduced, and the lower limit of the effective working temperature range of the battery is widened to a super-low temperature environment. The internal temperature field distribution of the battery is predicted in real time, and hotspot gradient coordinates are output through the cooperation of an electrochemical-thermal coupling model and a long short-term memory neural network.

Description

Technical Field

[0001] This invention relates to the fields of energy storage technology and robotics engineering technology, specifically to a wide-temperature-range lithium-ion battery system for robots and its thermal management method. Background Technology

[0002] The field of energy storage technology and robotics engineering focuses on the cross-application of highly reliable energy systems and intelligent mobile platforms, encompassing the development of electrochemical energy storage materials, thermal management engineering, power system integration, and environmentally adaptable design. Core challenges in this field include: ensuring the continuous output of robot power sources under complex operating conditions (such as polar expeditions, high-temperature workshops, and field inspections); and how to synergistically improve system energy efficiency through material modification (such as wide-temperature electrolytes and high-conductivity electrodes), structural innovation (lightweight thermal management components), and intelligent algorithms (dynamic temperature control strategies). One example is a wide-temperature-range lithium-ion battery system for robots and its thermal management method, which is a power supply solution specifically designed for robot platforms operating in extreme temperature environments. This system achieves stable power supply within a wide temperature range from tens of degrees below zero Celsius to seventy or eighty degrees Celsius through a three-level synergy of wide-temperature-resistant cell materials (such as alloyed negative electrodes and wide-range electrolytes), an integrated thermal management structure (a combination of graphene heating film and heat spreader), and intelligent control algorithms (fuzzy PID temperature control strategy).

[0003] Existing technologies rely on electrolyte systems with fixed formulations, whose solvent and additive ratios cannot be dynamically adjusted with ambient temperature. This leads to a significant decrease in ionic conductivity at low temperatures and an exacerbation of interfacial side reactions at high temperatures. For example, the viscosity of conventional electrolytes increases dramatically at low temperatures, causing frequent start-up failures of polar research robots. Traditional thermal management employs a passive response mechanism, initiating heating or heat dissipation only when the temperature exceeds a threshold. This results in control lag, often causing irreversible capacity loss in equipment operating in high-temperature environments. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a wide-temperature-range lithium-ion battery system for robots and its thermal management method. It solves the problem that existing technologies rely on electrolyte systems with fixed formulations, where the ratio of solvent to additives cannot be dynamically adjusted with ambient temperature. This leads to a significant decrease in ionic conductivity at low temperatures and an exacerbation of interfacial side reactions at high temperatures. For example, the viscosity of conventional electrolytes increases significantly at low temperatures, causing frequent start-up failures of polar research robots.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a wide-temperature-range lithium-ion battery system for robots and its thermal management method, comprising the following steps:

[0006] S1: Based on real-time data from an ambient temperature sensor, a reinforcement learning algorithm is used to dynamically adjust the mixing ratio of low freezing point solvent and high temperature film-forming additive in the electrolyte through a microfluidic chip, so as to minimize the change in internal resistance measured by electrochemical impedance spectroscopy and generate an adaptive electrolyte mixing ratio parameter.

[0007] S2: Based on adaptive electrolyte mixing ratio parameters and battery current load data, an electrochemical-thermal coupling simulation model is used, combined with a long short-term memory neural network to predict the future internal temperature field distribution of the battery, outputting the coordinates of hot spot areas and temperature gradient, and generating a thermal field spatiotemporal distribution prediction map.

[0008] S3: Based on the predicted thermal field spatiotemporal distribution map, a fuzzy proportional-integral-derivative control algorithm is adopted: if the low temperature zone is predicted, the graphene heating film is triggered and the power is modulated by PWM, with a target temperature of 10℃; if the high temperature zone is predicted, the staged phase change microcapsule heat absorption mechanism is activated, and the working fluid flow rate of the heat spreader is controlled synchronously; the phase change state of the PCM is monitored in real time by an ultrasonic sensor, and a real-time temperature control execution status code is generated.

[0009] S4: Based on the heat dissipation intensity data in the real-time temperature control execution status code, the Seebeck coefficient optimization algorithm is used to calculate the temperature difference power generation efficiency between the heat spreader and the thermoelectric module. Dynamically adjust the working fluid flow rate of the heat exchanger to maximize ΔT; recover electrical energy and store it in the auxiliary capacitor, giving priority to the heating film and generating a thermoelectric recovery energy quota;

[0010] S5: Based on the thermoelectric recovery energy quota and the spatiotemporal distribution prediction map of the thermal field, an improved auction algorithm is adopted; the battery temperature distribution of each robot in the cluster is obtained through the LoRa network; high-power tasks are assigned to robots in low-temperature zones and low-power tasks are assigned to robots in high-temperature zones; the wireless charging module is called to realize cross-robot thermal energy transfer according to the thermoelectric recovery energy quota, and a cluster thermal balance scheduling instruction set is generated.

[0011] Preferably, the generation of adaptive electrolyte mixing ratio parameters based on S1 includes the following steps:

[0012] S101: Based on real-time data from an ambient temperature sensor, a sliding window mean filtering algorithm is used to preprocess the temperature signal, eliminate measurement noise, and generate a filtered ambient temperature value.

[0013] S102: Based on the filtered ambient temperature value, the Q-learning reinforcement learning algorithm is used to calculate the optimal solvent / additive ratio and generate electrolyte mixing ratio decision parameters by using the change in internal resistance of electrochemical impedance spectroscopy as the reward function.

[0014] S103: Based on the electrolyte mixing ratio decision parameters, the injection ratio of ethyl acetate and fluoroethylene carbonate is precisely controlled by the microfluidic chip piezoelectric drive algorithm to generate adaptive electrolyte mixing ratio parameters.

[0015] Preferably, the generation of the spatiotemporal distribution prediction map of the thermal field based on S2 includes the following steps:

[0016] S201: Based on adaptive electrolyte mixing ratio parameters, a three-dimensional geometric model of the battery is constructed using the finite element mesh generation method. Electrolyte viscosity and electrode thermal conductivity parameters are imported to generate a multiphysics simulation basic model.

[0017] S202: Based on the multiphysics simulation model and real-time current load data, the heat generation rate distribution is calculated using the electrochemical-thermal coupling equation. Generate a transient heat generation distribution map;

[0018] S203: Based on the transient heat generation distribution map, a long short-term memory neural network is used to predict the future temperature field evolution, output the coordinates of the maximum temperature difference gradient, and generate a spatiotemporal distribution prediction map of the thermal field.

[0019] Preferably, the generation of real-time temperature control execution status codes based on S3 includes the following steps:

[0020] S301: Based on the thermal field spatiotemporal distribution prediction map, a temperature region segmentation algorithm is used to identify the boundary coordinates of the low temperature region and the high temperature region, and generate a temperature control zone identification matrix.

[0021] S302: Based on the temperature control zone identifier matrix, a fuzzy proportional-integral-derivative control algorithm is adopted to output the PWM heating duty cycle in the low temperature zone; the phase change microcapsule activation threshold is calculated in the high temperature zone to generate a set of temperature control execution parameters.

[0022] S303: Based on the temperature control execution parameter set, it drives the graphene heating film and the phase change microcapsule working fluid pump, and uses an ultrasonic phase change monitoring algorithm to detect the capsule liquefaction rate in real time and generate a real-time temperature control execution status code.

[0023] Preferably, the generation of thermoelectric recovery energy quotas based on S4 includes the following steps:

[0024] S401: Based on the heat dissipation intensity data in the real-time temperature control execution status code, the equivalent thermal resistance RK of the heat spreader and thermoelectric module is calculated using a thermal resistance network model to generate the thermal resistance parameters of the thermoelectric system.

[0025] S402: Based on the thermal resistance parameters of the thermoelectric system, the Seebeck coefficient optimization algorithm is used to solve for the operating point of maximum power generation efficiency. Generate the optimal temperature difference setpoint;

[0026] S403: Based on the optimal temperature difference setpoint, the flow rate of the working fluid on the heat exchanger is adjusted by a proportional flow control valve to ensure the actual temperature difference is within the setpoint. Approaching Generate thermoelectric recovery energy quotas.

[0027] Preferably, the generation cluster hot balance scheduling instruction set based on S5 includes the following steps:

[0028] S501: Based on the thermoelectric recovery energy quota, the energy quota allocation algorithm is used to calculate the upper limit of energy that each robot can share and generate a cluster energy sharing permission table.

[0029] S502: Based on the thermal field spatiotemporal distribution prediction map, the thermal management urgency index of each robot is calculated using the thermal load weighted evaluation algorithm to generate a cluster thermal urgency ranking.

[0030] S503: Based on the cluster energy sharing permission table and the cluster thermal urgency ranking, an improved auction algorithm is used to allocate tasks. High-urgency robots are given priority to obtain energy sharing permissions; low-urgency robots execute high-power tasks, generating a cluster thermal balance scheduling instruction set.

[0031] A wide-temperature-range lithium-ion battery system for robots includes the following modules: an adaptive electrolyte module, an intelligent temperature control execution module, and a collaborative optimization decision-making module;

[0032] The adaptive electrolyte module dynamically optimizes the electrolyte formulation based on real-time data from an ambient temperature sensor and employs a reinforcement learning algorithm. It precisely controls the mixing ratio of low freezing point solvents and high-temperature film-forming additives through a microfluidic chip to minimize the change in internal resistance of the electrochemical impedance spectroscopy and generate electrolyte mixing ratio parameters.

[0033] The adaptive electrolyte module includes a temperature filtering submodule, a formulation decision submodule, and a microfluidic execution submodule;

[0034] The intelligent temperature control execution module, based on the electrolyte mixing ratio parameter, uses an electrochemical-thermal coupling simulation model to predict the battery thermal field distribution, drives the graphene heating film and phase change microcapsules through a fuzzy proportional-integral-derivative control algorithm, and adjusts the temperature control intensity in real time by combining an ultrasonic phase change monitoring algorithm to generate a thermal management status code.

[0035] The intelligent temperature control execution module includes a thermal field modeling submodule, a temperature control zoning submodule, and an execution feedback submodule;

[0036] The collaborative optimization decision module, based on thermal management status codes, uses the Seebeck coefficient optimization algorithm to recover thermal energy and generate energy quotas. By improving the auction algorithm and combining it with cluster thermal field distribution data, it dynamically allocates tasks and cross-machine energy transfer strategies, and generates a cluster scheduling instruction set.

[0037] The collaborative optimization decision-making module includes an energy recovery submodule, a thermal urgency assessment submodule, and a task allocation submodule.

[0038] Preferably, the temperature filtering submodule uses a sliding window mean filtering algorithm to eliminate noise interference based on the raw data from the ambient temperature sensor, and generates a filtered ambient temperature value.

[0039] The formulation decision submodule, based on the filtered ambient temperature value, uses the Q-learning reinforcement learning algorithm with the change in internal resistance of the electrochemical impedance spectroscopy ΔR as the reward function to calculate the optimal ratio of solvent / additive and generate electrolyte mixing ratio decision parameters.

[0040] The microfluidic execution submodule, based on the electrolyte mixing ratio decision parameters, adjusts the flow rates of ethyl acetate and fluoroethylene carbonate through a piezoelectric valve precision control algorithm to generate electrolyte mixing ratio parameters.

[0041] Preferably, the thermal field modeling submodule calculates the internal heat distribution of the battery using the electrochemical-thermal coupled finite element method based on the electrolyte mixing ratio parameters. Generate a spatiotemporal distribution map of the thermal field;

[0042] The temperature control zoning submodule, based on the spatiotemporal distribution map of the thermal field, uses a regional gradient segmentation algorithm to identify low-temperature and high-temperature zones and generate a temperature control zoning identification matrix.

[0043] The execution feedback submodule, based on the temperature control zone identifier matrix, uses fuzzy proportional-integral-derivative control to output the PWM duty cycle of the graphene film and the working fluid flow rate of the phase change microcapsule. The liquefaction rate is detected in real time through an ultrasonic phase change monitoring algorithm to generate a thermal management status code.

[0044] Preferably, the energy recovery submodule uses a Seebeck coefficient optimization algorithm to solve for the maximum power generation efficiency operating condition based on the heat dissipation intensity in the thermal management status code. Generate thermoelectric recovery energy quota;

[0045] The thermal urgency assessment submodule calculates the thermal management urgency index for each robot based on the spatiotemporal distribution map of the thermal field and using a thermal load weighted algorithm. Generate a ranking of cluster hotness urgency;

[0046] The task allocation submodule, based on the energy quota of thermoelectric recovery and the ranking of cluster thermal urgency, uses an improved auction algorithm to allocate tasks, achieving high efficiency. Robots have priority access to energy sharing; low The robot performs high-power tasks and generates a cluster scheduling instruction set.

[0047] In summary, this application includes at least one of the following beneficial technical effects:

[0048] This invention significantly reduces the temperature sensitivity of electrolyte viscosity by dynamically adjusting the mixing ratio of low-freezing-point solvents and high-temperature film-forming additives in the electrolyte, combined with reinforcement learning algorithms to optimize electrochemical impedance characteristics, thus broadening the effective operating temperature range of the battery to ultra-low temperature environments. It predicts the internal temperature field distribution of the battery in real time and outputs hotspot gradient coordinates. Through the collaboration of an electrochemical-thermal coupling model and a long short-term memory neural network, it identifies potential thermal runaway regions in advance, reducing the risk of capacity decay caused by sudden temperature changes. A fuzzy control algorithm drives the graphene heating film and the hierarchical phase change material, dynamically switching between heating and cooling modes based on temperature zone indicators. Monitoring the phase change state ensures temperature control uniformity, avoiding electrode lithium plating or electrolyte decomposition caused by localized overcooling or overheating. Optimizing the working fluid flow rate of the heat spreader and thermoelectric module improves Seebeck effect power generation efficiency, recovers heat dissipation energy and supplies heating demand, reducing the system's net power consumption. Based on thermal field distribution data and energy recovery quotas, an improved task allocation algorithm dynamically adjusts the robot cluster load, enabling cross-robot heat transfer between high and low temperature ranges and increasing the duration of multi-robot collaborative operations in extreme environments. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the main steps of this application;

[0050] Figure 2 This is a detailed schematic diagram of S1 in this application;

[0051] Figure 3 This is a detailed schematic diagram of S2 in this application;

[0052] Figure 4 This is a detailed schematic diagram of S3 in this application;

[0053] Figure 5 This is a detailed schematic diagram of S4 in this application;

[0054] Figure 6 This is a detailed schematic diagram of S5 in this application;

[0055] Figure 7 This is a system block diagram of the present invention. Detailed Implementation

[0056] The following is in conjunction with the appendix Figure 1-7 This application will be described in further detail.

[0057] See Figure 1 A wide-temperature-range lithium-ion battery system for robots and its thermal management method include the following steps:

[0058] S1: Based on real-time data from an ambient temperature sensor, a reinforcement learning algorithm is used to dynamically adjust the mixing ratio of low freezing point solvent and high temperature film-forming additive in the electrolyte through a microfluidic chip, so as to minimize the change in internal resistance measured by electrochemical impedance spectroscopy and generate an adaptive electrolyte mixing ratio parameter.

[0059] S2: Based on adaptive electrolyte mixing ratio parameters and battery current load data, an electrochemical-thermal coupling simulation model is used, combined with a long short-term memory neural network to predict the future internal temperature field distribution of the battery, outputting the coordinates of hot spot areas and temperature gradient, and generating a thermal field spatiotemporal distribution prediction map.

[0060] S3: Based on the predicted thermal field spatiotemporal distribution map, a fuzzy proportional-integral-derivative control algorithm is adopted: if the low temperature zone is predicted, the graphene heating film is triggered and the power is modulated by PWM, with a target temperature of 10℃; if the high temperature zone is predicted, the staged phase change microcapsule heat absorption mechanism is activated, and the working fluid flow rate of the heat spreader is controlled synchronously; the phase change state of the PCM is monitored in real time by an ultrasonic sensor, and a real-time temperature control execution status code is generated.

[0061] S4: Based on the heat dissipation intensity data in the real-time temperature control execution status code, the Seebeck coefficient optimization algorithm is used to calculate the temperature difference power generation efficiency between the heat spreader and the thermoelectric module. Dynamically adjust the working fluid flow rate of the heat exchanger to maximize ΔT; recover electrical energy and store it in the auxiliary capacitor, giving priority to the heating film and generating a thermoelectric recovery energy quota;

[0062] S5: Based on the thermoelectric recovery energy quota and the spatiotemporal distribution prediction map of the thermal field, an improved auction algorithm is adopted; the battery temperature distribution of each robot in the cluster is obtained through the LoRa network; high-power tasks are assigned to robots in low-temperature zones and low-power tasks are assigned to robots in high-temperature zones; the wireless charging module is called to realize cross-robot thermal energy transfer according to the thermoelectric recovery energy quota, and a cluster thermal balance scheduling instruction set is generated.

[0063] See Figure 2 The generation of adaptive electrolyte mixing ratio parameters based on S1 includes the following steps:

[0064] S101: Based on real-time data from an ambient temperature sensor, a sliding window mean filtering algorithm is used to preprocess the temperature signal, eliminate measurement noise, and generate a filtered ambient temperature value.

[0065] Based on the real-time acquisition of the raw data sequence [-25.3, -24.7, -26.1, -23.9, -25.5]℃ by an ambient temperature sensor, a sliding window mean filtering algorithm is adopted, with the window width set to 5 sampling points. The filtered value of the 3rd data point is calculated as (-24.7-26.1-23.9-25.5) / 4 = -25.05℃. Random fluctuation noise is eliminated, and the filtered ambient temperature value of -25.0℃ is output.

[0066] S102: Based on the filtered ambient temperature value, the Q-learning reinforcement learning algorithm is used to calculate the optimal solvent / additive ratio and generate electrolyte mixing ratio decision parameters by using the change in internal resistance of electrochemical impedance spectroscopy as the reward function.

[0067] Based on the filtered ambient temperature of -25.0℃, a Q-learning reinforcement learning algorithm is used, defining the state as the temperature range: low temperature zone. medium temperature zone High temperature zone The action range is the solvent ratio adjustment amount. reward function Initialize the Q-table as a zero matrix and set the learning rate. Discount factor Current state Belongs to the low temperature zone, select action (Increased solvent content), measurement after execution Calculate rewards Update Q value: The electrolyte mixing ratio decision parameter is ethyl acetate, which accounts for 72%.

[0068] S103: Based on the electrolyte mixing ratio decision parameters, the injection ratio of ethyl acetate and fluoroethylene carbonate is precisely controlled by the microfluidic chip piezoelectric drive algorithm to generate adaptive electrolyte mixing ratio parameters.

[0069] Based on the electrolyte mixing ratio decision parameter of 72% ethyl acetate, and using a microfluidic chip piezoelectric drive algorithm, a base voltage of 3.0V for the piezoelectric valve corresponds to a flow rate of 0.5ml / min. The voltage-flow rate linear relationship is: Flow rate = 0.5 + 0.1 × (V - 3.0) Ethyl acetate required flow rate = total flow rate 0.8 ml / min × 72% = 0.576 ml / min, calculated required voltage V = 3.0 + (0.576 - 0.5) / 0.1 = 3.76 V, actual output flow rate 0.576 ml / min, fluoroethylene carbonate flow rate 0.224 ml / min, mixing error < 0.5%, generating adaptive electrolyte mixing ratio parameters.

[0070] See Figure 3 The generation of the spatiotemporal distribution prediction map of the thermal field based on S2 includes the following steps:

[0071] S201: Based on adaptive electrolyte mixing ratio parameters, a three-dimensional geometric model of the battery is constructed using the finite element mesh generation method. Electrolyte viscosity and electrode thermal conductivity parameters are imported to generate a multiphysics simulation basic model.

[0072] Based on the adaptive electrolyte mixing ratio parameters (ethyl acetate content 72%), the electrolyte viscosity was mapped to 0.45 Pa·s from the reference database, and the negative electrode thermal conductivity was taken as the measured value of 1.2 W / (m·K). The battery 3D model was generated by using tetrahedral meshes with a unit size of 2mm. Material properties were set in COMSOL to generate a multiphysics simulation base model containing 120,000 mesh nodes.

[0073] S202: Based on the multiphysics simulation model and real-time current load data, the heat generation rate distribution is calculated using the electrochemical-thermal coupling equation. Generate a transient heat generation distribution map;

[0074] Based on the multiphysics simulation model, with a real-time current load of 15A and an internal resistance of Rint=0.08Ω, the heat generation rate is calculated: The thermal conductivity λ is taken as the negative pole, 1.2 W / (m·K). Electrolyte 0.3 W / (m·K) The temperature gradient ∇T is calculated using the temperature difference between adjacent grids. For example, if grid A has a temperature of 35℃ and grid B has a temperature of 38℃, with a distance of 0.002m, then ∇T = (38-35) / 0.002 = 1500℃ / m, representing the conductive heat component. The Joule heat component I²Rint = 15² × 0.08 = 18 W, and the total heat production q = 1800 + 18 = 1818 W / m³. The transient heat production distribution map is generated by traversing all grids.

[0075] S203: Based on the transient heat generation distribution map, a long short-term memory neural network is used to predict the future temperature field evolution, output the coordinates of the maximum temperature difference gradient, and generate a spatiotemporal distribution prediction map of the thermal field.

[0076] Based on the transient heat generation distribution map, the temperature sequence [35, 36, 38, 40, 42]℃ from the past 5 minutes is extracted as the input to a Long Short-Term Memory (LSTM) neural network. The network structure contains 3 layers of LSTM units (64 nodes per unit). The output is a predicted temperature of 45℃ for the next 3 minutes. The temperature gradient between grids is calculated: coordinates (1, 2, 3) Temperature 45℃, adjacent (1,2,4) With a temperature of 42℃, the gradient = (45-42) / 0.002 = 1500℃ / m. Traverse all coordinate pairs to identify the point with the maximum gradient (1,2,3). Generate a thermal field spatiotemporal distribution prediction icon and mark the coordinates.

[0077] See Figure 4 The generation of real-time temperature control execution status codes based on S3 includes the following steps:

[0078] S301: Based on the thermal field spatiotemporal distribution prediction map, a temperature region segmentation algorithm is used to identify the boundary coordinates of the low temperature region and the high temperature region, and generate a temperature control zone identification matrix.

[0079] Based on the predicted coordinates of the spatiotemporal distribution of the thermal field, the coordinates (1, 2, 3) are marked on the icon. Temperature -5℃, (2,3,4) At a temperature of 48℃, a temperature region segmentation algorithm is used, defining a low-temperature region threshold of ≤0℃ and a high-temperature region threshold of ≥45℃. All grid points are traversed: temperatures ≤0℃ are marked as 1 (low-temperature region), temperatures ≥45℃ are marked as 2 (high-temperature region), and all others are marked as 0. For example, coordinates (1,2,3) Mark 1, (2,3,4) Mark 2, and generate a temperature control zone identifier matrix containing the marked values.

[0080] S302: Based on the temperature control zone identifier matrix, a fuzzy proportional-integral-derivative control algorithm is adopted to output the PWM heating duty cycle in the low temperature zone; the phase change microcapsule activation threshold is calculated in the high temperature zone to generate a set of temperature control execution parameters.

[0081] Based on the coordinates (1,2,3) in the temperature control zone identifier matrix Mark 1 (low temperature zone), using fuzzy proportional-integral-derivative control algorithm, set target temperature 10℃, current temperature -5℃, temperature difference 15℃, according to the temperature difference membership function: when temperature difference > 10℃, output duty cycle increment 30%, calculate PWM duty cycle = base value 50% + 30% = 80%, coordinates (2, 3, 4) Mark 2 (high temperature zone), set the phase change microcapsule activation threshold temperature difference gradient > 1000℃ / m, the current gradient is 1500℃ / m, activation level = min(3, int(1500 / 500)) = 3, generate temperature control execution parameter set {coordinate (1,2,3): PWM80%, coordinate (2,3,4): activation level 3}.

[0082] S303: Based on the temperature control execution parameter set, it drives the graphene heating film and the phase change microcapsule working fluid pump, and uses an ultrasonic phase change monitoring algorithm to detect the capsule liquefaction rate in real time and generate a real-time temperature control execution status code.

[0083] Based on the temperature control execution parameter set {coordinates (1,2,3): PWM80%, coordinates (2,3,4): activation level 3}, the graphene heating film is driven at coordinates (1,2,3) Applying a 12V voltage (80% duty cycle), the phase change microcapsule working fluid pump operates at coordinates (2,3,4). Injected at a flow rate of 0.8 ml / s corresponding to level 3, using an ultrasonic phase change monitoring algorithm, 40 kHz ultrasonic waves were emitted through the microcapsule layer, and the rate of change of sound velocity was measured. The sound velocity in solid PCM was 1500 m / s, and in liquid it was 1200 m / s. The measured sound velocity was 1300 m / s. The liquefaction rate was calculated as (1500-1300) / (1500-1200)×100%=66.7%. A real-time temperature control execution status code was generated: {Heating coordinates: (1,2,3), PWM:80%, Heat dissipation coordinates: (2,3,4), Liquefaction rate: 66.7%}.

[0084] See Figure 5 The generation of thermoelectric recovery energy quotas based on S4 includes the following steps:

[0085] S401: Based on the heat dissipation intensity data in the real-time temperature control execution status code, the equivalent thermal resistance RK of the heat spreader and thermoelectric module is calculated using a thermal resistance network model to generate the thermal resistance parameters of the thermoelectric system.

[0086] Based on the heat dissipation coordinates (2,3,4) in the real-time temperature control execution status code Corresponding to a liquefaction rate of 66.7%, the mapped heat dissipation power is 18W. A thermal resistance network model is used, with a copper substrate thickness of 0.5mm and a thermal conductivity of 400W / (m·K). Calculate the thermal resistance of the substrate Thermoelectric module ceramic substrate thermal resistance Thermal resistance of interface materials Total thermal resistance in series The resulting thermoelectric system has a thermal resistance parameter of 0.025 K / W.

[0087] S402: Based on the thermal resistance parameters of the thermoelectric system, the Seebeck coefficient optimization algorithm is used to solve for the operating point of maximum power generation efficiency. Generate the optimal temperature difference setpoint;

[0088] Based on the thermoelectric system's thermal resistance parameter of 0.025 K / W, the Seebeck coefficient α is taken as the measured value of the thermoelectric module. maximizing efficiency Set to 5%, calculate the optimal temperature difference setpoint. Based on the temperature limit of the thermoelectric module, which is corrected to 60K, the optimal temperature difference setting value of 60K is generated.

[0089] S403: Based on the optimal temperature difference setpoint, the flow rate of the working fluid on the heat exchanger is adjusted by a proportional flow control valve to ensure the actual temperature difference is within the setpoint. Approaching Generate thermoelectric recovery energy quotas.

[0090] Based on the optimal temperature difference setting of 60K, the current hot end temperature of the heat spreader is 48℃, and the cold end temperature of the radiator is 30℃. The actual... Temperature difference 42K, proportional flow control valve set adjustment coefficient Flow adjustment amount The original working fluid flow rate was increased from 0.8 ml / s to 5.0 ml / s, and ΔT was remeasured at 55 K. The generated heat and electricity recovery energy quota of 55 K temperature difference corresponds to a power generation of 3.3 W.

[0091] See Figure 6 The generation of cluster hot balance scheduling instruction set based on S5 includes the following steps:

[0092] S501: Based on the thermoelectric recovery energy quota, the energy quota allocation algorithm is used to calculate the upper limit of energy that each robot can share and generate a cluster energy sharing permission table.

[0093] Based on the thermoelectric recovery energy quota of 3.3W for robot A and 2.1W for robot B, an energy quota allocation algorithm is adopted, setting the upper limit of the shareable energy ratio to 70%. The upper limit of shareable energy for robot A is calculated as 3.3 × 0.7 = 2.31W, and the upper limit for robot B is 2.1 × 0.7 = 1.47W. The cluster energy sharing permission table {RobotA:2.31W, RobotB:1.47W} is generated.

[0094] S502: Based on the thermal field spatiotemporal distribution prediction map, the thermal management urgency index of each robot is calculated using the thermal load weighted evaluation algorithm to generate a cluster thermal urgency ranking.

[0095] Based on the predicted thermal field spatiotemporal distribution map of Robot A (high temperature 48℃, low temperature -5℃) and Robot B (high temperature 42℃, low temperature 2℃), a heat load weighted evaluation algorithm is adopted, defining the weights of temperature ranges as follows: 45℃ = 1.2, <0℃ = 1.5, and others = 1.0. The thermal urgency index of Robot A is calculated as 48 × 1.2 + (-5) × 1.5 = 57.6 - 7.5 = 50.1, and the index of Robot B is calculated as 42 × 1.0 + 2 × 1.0 = 44. After normalization, the urgency ranking is obtained as {RobotA: 0.532, RobotB: 0.468}.

[0096] S503: Based on the cluster energy sharing permission table and the cluster thermal urgency ranking, an improved auction algorithm is used to allocate tasks. High-urgency robots are given priority to obtain energy sharing permissions; low-urgency robots execute high-power tasks, generating a cluster thermal balance scheduling instruction set.

[0097] Based on the cluster energy sharing permission table {RobotA: 2.31W, RobotB: 1.47W} and the cluster thermal urgency ranking {RobotA: 0.532, RobotB: 0.468}, an improved auction algorithm is adopted, setting a high urgency threshold of 0.5 and a low urgency threshold of 0.4. RobotA's urgency (0.532 > 0.5) is assigned an energy receiving permission of 2.31W, while RobotB's urgency (0.468 < 0.4) is assigned a transport task (estimated power consumption of 15W). A cluster thermal balance scheduling instruction set is generated: {RobotA: Receive 2.31W of energy, RobotB: Execute transport task}.

[0098] See Figure 7 A wide-temperature-range lithium-ion battery system for robots includes the following modules: an adaptive electrolyte module, an intelligent temperature control execution module, and a collaborative optimization decision-making module;

[0099] The adaptive electrolyte module dynamically optimizes the electrolyte formulation based on real-time data from an ambient temperature sensor and employs a reinforcement learning algorithm. It precisely controls the mixing ratio of low freezing point solvents and high-temperature film-forming additives through a microfluidic chip to minimize the change in internal resistance of the electrochemical impedance spectroscopy and generate electrolyte mixing ratio parameters.

[0100] The adaptive electrolyte module includes a temperature filtering submodule, a formulation decision submodule, and a microfluidic execution submodule;

[0101] The intelligent temperature control execution module, based on the electrolyte mixing ratio parameter, uses an electrochemical-thermal coupling simulation model to predict the battery thermal field distribution, drives the graphene heating film and phase change microcapsules through a fuzzy proportional-integral-derivative control algorithm, and adjusts the temperature control intensity in real time by combining an ultrasonic phase change monitoring algorithm to generate a thermal management status code.

[0102] The intelligent temperature control execution module includes a thermal field modeling submodule, a temperature control zoning submodule, and an execution feedback submodule;

[0103] The collaborative optimization decision module, based on thermal management status codes, uses the Seebeck coefficient optimization algorithm to recover thermal energy and generate energy quotas. By improving the auction algorithm and combining it with cluster thermal field distribution data, it dynamically allocates tasks and cross-machine energy transfer strategies, and generates a cluster scheduling instruction set.

[0104] The collaborative optimization decision-making module includes an energy recovery submodule, a thermal urgency assessment submodule, and a task allocation submodule.

[0105] The temperature filtering submodule, based on the raw data from the ambient temperature sensor, uses a sliding window mean filtering algorithm to eliminate noise interference and generate filtered ambient temperature values.

[0106] Based on the original data sequence [-25.3, -24.7, -26.1, -23.9, -25.5]℃ collected by the ambient temperature sensor, and with a sliding window width of 5 sampling points, the filtered value of the 3rd data point is calculated: the mean of the data from index 1 to 5 is calculated as (-24.7-26.1-23.9-25.5) / 4=-25.05℃, and the filtered ambient temperature value is output as -25.0℃ after rounding.

[0107] The formulation decision submodule, based on the filtered ambient temperature value, uses the Q-learning reinforcement learning algorithm with the change in internal resistance of the electrochemical impedance spectroscopy ΔR as the reward function to calculate the optimal ratio of solvent / additive and generate electrolyte mixing ratio decision parameters.

[0108] Based on the filtered ambient temperature of -25.0℃, the Q-learning state space is defined as follows: Low temperature region (-40℃, -20℃] encoding 1, Medium temperature region (-20℃, 0℃] encoding 2, High temperature region (0℃, 60℃] encoding 3, Action space {solvent percentage -5%, 0%, +5%}, and Reward function. Initialize the Q-table as a 3×3 zero matrix, with a learning rate α=0.8 and a discount factor γ=0.9. Current state... (Low Temperature Zone), Select Action Actual measurement after execution Calculate R = 1 / (1+0.8) = 0.555, and update. The output electrolyte mixing ratio decision parameter is ethyl acetate, which accounts for 72%.

[0109] The microfluidic execution submodule, based on the electrolyte mixing ratio decision parameters, adjusts the flow rates of ethyl acetate and fluoroethylene carbonate through a piezoelectric valve precision control algorithm to generate electrolyte mixing ratio parameters.

[0110] Based on the electrolyte mixing ratio decision parameter of 72% ethyl acetate, and setting the total flow rate of the microfluidic chip to 0.8 ml / min, the required flow rate of ethyl acetate is... Piezoelectric valve voltage-flow relationship: Flow rate Calculate the required voltage The applied voltage was 3.76V, the measured flow rate of ethyl acetate was 0.578 ml / min, the flow rate of fluoroethylene carbonate was 0.222 ml / min, and the mixing ratio error was ( Output electrolyte mixing ratio parameters.

[0111] The thermal field modeling submodule, based on the electrolyte mixing ratio parameters, uses the electrochemical-thermal coupled finite element method to calculate the internal heat distribution of the battery. Generate a spatiotemporal distribution map of the thermal field;

[0112] Based on the electrolyte mixing ratio parameters of 72% ethyl acetate, the mapped electrolyte viscosity is 0.45 Pa·s, and the negative electrode thermal conductivity is 1.2 W / (m·K). A 3D mesh model of the battery was constructed with a cell size of 2mm. The input real-time current was 15A, and the internal resistance was 0.08Ω. The distance between mesh A (35℃) and mesh B (38℃) was 0.002m, and the temperature gradient was calculated. Conductive heat component The Joule heat component is 15² × 0.08 = 18 W, and the total heat production is... Iterate through all grids to generate a spatiotemporal distribution map of the thermal field.

[0113] The temperature control zoning submodule, based on the spatiotemporal distribution map of the thermal field, uses a regional gradient segmentation algorithm to identify low-temperature and high-temperature zones and generate a temperature control zoning identification matrix.

[0114] Based on the grid coordinates (1,2,3) in the thermal field spatiotemporal distribution map Temperature -5℃, (2,3,4) Temperature 48℃, low temperature threshold defined as ≤0℃, high temperature threshold defined as ≥45℃, traversing grid points: mark 1 when temperature ≤0℃, mark 2 when temperature ≥45℃, coordinates (1,2,3) Mark 1, (2,3,4) Mark 2, and mark the remaining temperature ranges as 0 to generate a temperature control zone identifier matrix.

[0115] The execution feedback submodule, based on the temperature control zone identifier matrix, uses fuzzy proportional-integral-derivative control to output the PWM duty cycle of the graphene film and the working fluid flow rate of the phase change microcapsule. The liquefaction rate is detected in real time through an ultrasonic phase change monitoring algorithm to generate a thermal management status code.

[0116] Based on the coordinates (1,2,3) in the temperature control zone identifier matrix Mark 1 (low temperature zone), target temperature set at 10℃, current temperature at -5℃, temperature difference at 15℃. Fuzzy proportional-integral-derivative control temperature difference membership function: when temperature difference > 10℃, output duty cycle increment is 30%. Calculate PWM duty cycle = 50% + 30% = 80%, coordinates (2, 3, 4). Mark 2 (high temperature zone), set the phase change microcapsule activation threshold temperature difference gradient > 1000℃ / m, current gradient 1500℃ / m, activation level = int(1500 / 500) = 3, working fluid flow rate = 0.2×3 = 0.6ml / s, ultrasonic monitoring solid PCM sound velocity 1500m / s, liquid 1200m / s, actual measured sound velocity 1300m / s, liquefaction rate = (1500-1300) / (1500-1200)×100% = 66.7%, generate thermal management status code {heating coordinates: (1,2,3), PWM: 80%, heat dissipation coordinates: (2,3,4), liquefaction rate: 66.7%}.

[0117] The energy recovery submodule, based on the heat dissipation intensity in the thermal management status code, uses a Seebeck coefficient optimization algorithm to solve for the maximum power generation efficiency operating condition. Generate thermoelectric recovery energy quota;

[0118] Based on the liquefaction rate of 66.7% corresponding to the heat dissipation coordinates in the thermal management status code, which maps to a heat dissipation power of 18W, a thermal resistance network model is adopted. The thickness of the copper substrate of the heat spreader is 0.5 mm, the thermal conductivity is 400 W / m Kelvin, and the area is 0.01 square meters. The thermal resistance of the substrate is calculated as thickness divided by thermal conductivity multiplied by area, i.e., 0.0005 divided by 400 multiplied by 0.01 equals 0.000125 Kelvin per watt. The thermal resistance of the ceramic substrate of the thermoelectric module is 0.02 Kelvin per watt, the thermal resistance of the interface material is 0.005 Kelvin per watt, and the total series thermal resistance RK is equal to 0.000125 plus 0.02 plus 0.005 equals 0.025125 Kelvin per watt, approximately 0.025 Kelvin per watt. The Seebeck coefficient α is taken as the measured value of 200 microvolts per Kelvin of the thermoelectric module, i.e., 0.0002 volts per Kelvin, for maximum efficiency. Set the value to 5%, or 0.05%, and calculate the optimal temperature difference setpoint. Equals four times RK times Dividing by α squared, i.e., four times 0.025 times 0.05 divided by 0.0002 times 0.0002, first calculate the numerator: four times 0.025 times 0.05 equals 0.005, the denominator is 0.00000004, the result 0.005 divided by 0.00000004 equals 125000 Kelvin. Based on the maximum temperature difference limit of 100 Kelvin for thermoelectric modules, adjust accordingly. The temperature difference is 60 Kelvin, corresponding to the generated thermoelectric recovery energy quota.

[0119] The thermal urgency assessment submodule calculates the thermal management urgency index for each robot based on the spatiotemporal distribution map of the thermal field and using a thermal load weighted algorithm. Generate a ranking of cluster hotness urgency;

[0120] Based on the spatiotemporal distribution map of the thermal field, robot A has a high temperature of 48 degrees Celsius and a low temperature of -5 degrees Celsius, while robot B has a high temperature of 42 degrees Celsius and a low temperature of 2 degrees Celsius. A heat load weighted algorithm is used, defining temperature range weights: temperatures above 45 degrees Celsius have a weight of 1.2, temperatures below 0 degrees Celsius have a weight of 1.5, and other temperatures have a weight of 1.0. The thermal stress index of robot A is calculated as follows: the high temperature of 48 degrees Celsius (greater than 45 degrees Celsius) has a weight of 1.2, contributing 48 multiplied by 1.2 equals 57.6; the low temperature of -5 degrees Celsius (less than 0 degrees Celsius) has a weight of 1.5, contributing -5 multiplied by 1.5, and so on. The total is 57.6 plus -7.5, which equals 50.1. Robot B has a weight of 1.0 at its high temperature of 42 degrees Celsius, contributing 42 multiplied by 1.0, which equals 42. At its low temperature of 2 degrees Celsius, the weight is 1.0, contributing 2 multiplied by 1.0, which equals 2. The total is 42 plus 2, which equals 44. After normalizing the urgency, the total index is 50.1 plus 44, which equals 94.1. Robot A's urgency is 50.1 divided by 94.1, which is approximately 0.532. Robot B's urgency is 44 divided by 94.1, which is approximately 0.468. The resulting cluster heat urgency ranking is: Robot A first, Robot B second.

[0121] The task allocation submodule, based on the energy quota of thermoelectric recovery and the ranking of cluster thermal urgency, uses an improved auction algorithm to allocate tasks, achieving high efficiency. Robots have priority access to energy sharing; low The robot performs high-power tasks and generates a cluster scheduling instruction set.

[0122] Based on the energy quotas for thermoelectric recovery, Robot A has a quota of 3.3 watts and Robot B has a quota of 2.1 watts. The cluster's thermal urgency ranking shows Robot A's urgency at 0.532 and Robot B's at 0.468. An improved auction algorithm is used, setting a high urgency threshold of 0.5 and a low urgency threshold of 0.4. Urgency greater than 0.5 is considered high urgency, and urgency less than 0.4 is considered low urgency. Robot A's urgency of 0.532 is greater than 0.5, so it is assigned energy receiving rights, with a maximum shareable energy ratio of 70%. Robot A's shareable energy equals 3.3 x 0.7, which equals 2.31 watts. Robot B's urgency of 0.468 is between 0.4 and 0.5, not meeting the high / low threshold definition. Therefore, it is assigned a high-power task, such as a handling task with an estimated power consumption of 15 watts. A cluster scheduling instruction set is generated, with Robot A receiving 2.31 watts of energy and Robot B executing the handling task.

[0123] The above are all preferred embodiments of this application, and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made in accordance with the structure, shape and principle of this application should be covered within the scope of protection of this application.

Claims

1. A thermal management method for a wide-temperature-range lithium-ion battery for robots, characterized in that... Includes the following steps: S1: Based on real-time data from an ambient temperature sensor, a reinforcement learning algorithm is used to dynamically adjust the mixing ratio of low freezing point solvent and high temperature film-forming additive in the electrolyte through a microfluidic chip, so as to minimize the change in internal resistance measured by electrochemical impedance spectroscopy and generate an adaptive electrolyte mixing ratio parameter. S2: Based on adaptive electrolyte mixing ratio parameters and battery current load data, an electrochemical-thermal coupling simulation model is used, combined with a long short-term memory neural network to predict the future internal temperature field distribution of the battery, outputting the coordinates of hot spot areas and temperature gradient, and generating a thermal field spatiotemporal distribution prediction map. S3: Based on the predicted thermal field spatiotemporal distribution map, a fuzzy proportional-integral-derivative control algorithm is adopted: if the low temperature zone is predicted, the graphene heating film is triggered and the power is modulated by PWM, with a target temperature of 10℃; if the high temperature zone is predicted, the staged phase change microcapsule heat absorption mechanism is activated, and the working fluid flow rate of the heat spreader is controlled synchronously; the phase change state of the PCM is monitored in real time by an ultrasonic sensor, and a real-time temperature control execution status code is generated. S4: Based on the heat dissipation intensity data in the real-time temperature control execution status code, the Seebeck coefficient optimization algorithm is used to calculate the temperature difference power generation efficiency between the heat spreader and the thermoelectric module. Dynamically adjust the working fluid flow rate of the heat exchanger to maximize ΔT; recover electrical energy and store it in the auxiliary capacitor, giving priority to the heating film and generating a thermoelectric recovery energy quota; S5: Based on the thermoelectric recovery energy quota and the spatiotemporal distribution prediction map of the thermal field, an improved auction algorithm is adopted; the battery temperature distribution of each robot in the cluster is obtained through the LoRa network; high-power tasks are assigned to robots in low-temperature zones and low-power tasks are assigned to robots in high-temperature zones; the wireless charging module is called to realize cross-robot thermal energy transfer according to the thermoelectric recovery energy quota, and a cluster thermal balance scheduling instruction set is generated.

2. The thermal management method for a wide-temperature-range lithium-ion battery for robots according to claim 1, characterized in that: The generation of adaptive electrolyte mixing ratio parameters based on S1 includes the following steps: S101: Based on real-time data from an ambient temperature sensor, a sliding window mean filtering algorithm is used to preprocess the temperature signal, eliminate measurement noise, and generate a filtered ambient temperature value. S102: Based on the filtered ambient temperature value, the Q-learning reinforcement learning algorithm is used to calculate the optimal solvent / additive ratio and generate electrolyte mixing ratio decision parameters by using the change in internal resistance of electrochemical impedance spectroscopy as the reward function. S103: Based on the electrolyte mixing ratio decision parameters, the injection ratio of ethyl acetate and fluoroethylene carbonate is precisely controlled by the microfluidic chip piezoelectric drive algorithm to generate adaptive electrolyte mixing ratio parameters.

3. The thermal management method for a wide-temperature-range lithium-ion battery for robots according to claim 1, characterized in that: The generation of the spatiotemporal distribution prediction map of the thermal field based on S2 includes the following steps: S201: Based on adaptive electrolyte mixing ratio parameters, a three-dimensional geometric model of the battery is constructed using the finite element mesh generation method. Electrolyte viscosity and electrode thermal conductivity parameters are imported to generate a multiphysics simulation basic model. S202: Based on the multiphysics simulation model and real-time current load data, the heat generation rate distribution is calculated using the electrochemical-thermal coupling equation. Generate a transient heat generation distribution map; S203: Based on the transient heat generation distribution map, a long short-term memory neural network is used to predict the future temperature field evolution, output the coordinates of the maximum temperature difference gradient, and generate a spatiotemporal distribution prediction map of the thermal field.

4. The thermal management method for a wide-temperature-range lithium-ion battery for robots according to claim 1, characterized in that: The steps for generating real-time temperature control execution status codes based on S3 are as follows: S301: Based on the thermal field spatiotemporal distribution prediction map, a temperature region segmentation algorithm is used to identify the boundary coordinates of the low temperature region and the high temperature region, and generate a temperature control zone identification matrix. S302: Based on the temperature control zone identifier matrix, a fuzzy proportional-integral-derivative control algorithm is adopted to output the PWM heating duty cycle in the low temperature zone; the phase change microcapsule activation threshold is calculated in the high temperature zone to generate a set of temperature control execution parameters. S303: Based on the temperature control execution parameter set, it drives the graphene heating film and the phase change microcapsule working fluid pump, and uses an ultrasonic phase change monitoring algorithm to detect the capsule liquefaction rate in real time and generate a real-time temperature control execution status code.

5. A thermal management method for a wide-temperature-range lithium-ion battery for robots according to claim 1, characterized in that: The generation of thermoelectric recovery energy quotas based on S4 includes the following steps: S401: Based on the heat dissipation intensity data in the real-time temperature control execution status code, the equivalent thermal resistance RK of the heat spreader and thermoelectric module is calculated using a thermal resistance network model to generate the thermal resistance parameters of the thermoelectric system. S402: Based on the thermal resistance parameters of the thermoelectric system, the Seebeck coefficient optimization algorithm is used to solve for the operating point of maximum power generation efficiency. Generate the optimal temperature difference setpoint; S403: Based on the optimal temperature difference setpoint, the flow rate of the working fluid on the heat exchanger is adjusted by a proportional flow control valve to ensure the actual temperature difference is within the setpoint. Approaching Generate thermoelectric recovery energy quotas.

6. A thermal management method for a wide-temperature-range lithium-ion battery for robots according to claim 1, characterized in that: The generation of a cluster hot balance scheduling instruction set based on S5 includes the following steps: S501: Based on the thermoelectric recovery energy quota, the energy quota allocation algorithm is used to calculate the upper limit of energy that each robot can share and generate a cluster energy sharing permission table. S502: Based on the thermal field spatiotemporal distribution prediction map, the thermal management urgency index of each robot is calculated using the thermal load weighted evaluation algorithm to generate a cluster thermal urgency ranking. S503: Based on the cluster energy sharing permission table and the cluster thermal urgency ranking, an improved auction algorithm is used to allocate tasks. High-urgency robots are given priority to obtain energy sharing permissions; low-urgency robots execute high-power tasks, generating a cluster thermal balance scheduling instruction set.

7. A wide-temperature-range lithium-ion battery system for robots, characterized in that... It includes the following modules: adaptive electrolyte module, intelligent temperature control execution module, and collaborative optimization decision-making module; The adaptive electrolyte module dynamically optimizes the electrolyte formulation based on real-time data from an ambient temperature sensor and employs a reinforcement learning algorithm. It precisely controls the mixing ratio of low freezing point solvents and high-temperature film-forming additives through a microfluidic chip to minimize the change in internal resistance of the electrochemical impedance spectroscopy and generate electrolyte mixing ratio parameters. The adaptive electrolyte module includes a temperature filtering submodule, a formulation decision submodule, and a microfluidic execution submodule; The intelligent temperature control execution module, based on the electrolyte mixing ratio parameter, uses an electrochemical-thermal coupling simulation model to predict the battery thermal field distribution, drives the graphene heating film and phase change microcapsules through a fuzzy proportional-integral-derivative control algorithm, and adjusts the temperature control intensity in real time by combining an ultrasonic phase change monitoring algorithm to generate a thermal management status code. The intelligent temperature control execution module includes a thermal field modeling submodule, a temperature control zoning submodule, and an execution feedback submodule; The collaborative optimization decision module, based on thermal management status codes, uses the Seebeck coefficient optimization algorithm to recover thermal energy and generate energy quotas. By improving the auction algorithm and combining it with cluster thermal field distribution data, it dynamically allocates tasks and cross-machine energy transfer strategies, and generates a cluster scheduling instruction set. The collaborative optimization decision-making module includes an energy recovery submodule, a thermal urgency assessment submodule, and a task allocation submodule.

8. A wide-temperature-range lithium-ion battery system for robots according to claim 7, characterized in that: The temperature filtering submodule, based on the raw data from the ambient temperature sensor, uses a sliding window mean filtering algorithm to eliminate noise interference and generate filtered ambient temperature values. The formulation decision submodule, based on the filtered ambient temperature value, uses the Q-learning reinforcement learning algorithm with the change in internal resistance of the electrochemical impedance spectroscopy ΔR as the reward function to calculate the optimal ratio of solvent / additive and generate electrolyte mixing ratio decision parameters. The microfluidic execution submodule, based on the electrolyte mixing ratio decision parameters, adjusts the flow rates of ethyl acetate and fluoroethylene carbonate through a piezoelectric valve precision control algorithm to generate electrolyte mixing ratio parameters.

9. A wide-temperature-range lithium-ion battery system for robots according to claim 7, characterized in that: The thermal field modeling submodule, based on the electrolyte mixing ratio parameters, uses the electrochemical-thermal coupled finite element method to calculate the internal heat distribution of the battery. Generate a spatiotemporal distribution map of the thermal field; The temperature control zoning submodule, based on the spatiotemporal distribution map of the thermal field, uses a regional gradient segmentation algorithm to identify low-temperature and high-temperature zones and generate a temperature control zoning identification matrix. The execution feedback submodule, based on the temperature control zone identifier matrix, uses fuzzy proportional-integral-derivative control to output the PWM duty cycle of the graphene film and the working fluid flow rate of the phase change microcapsule. The liquefaction rate is detected in real time through an ultrasonic phase change monitoring algorithm to generate a thermal management status code.

10. A wide-temperature-range lithium-ion battery system for robots according to claim 7, characterized in that: The energy recovery submodule, based on the heat dissipation intensity in the thermal management status code, uses a Seebeck coefficient optimization algorithm to solve for the maximum power generation efficiency operating condition. Generate thermoelectric recovery energy quota; The thermal urgency assessment submodule calculates the thermal management urgency index for each robot based on the spatiotemporal distribution map of the thermal field and using a thermal load weighted algorithm. Generate a ranking of cluster hotness urgency; The task allocation submodule, based on the energy quota of thermoelectric recovery and the ranking of cluster thermal urgency, uses an improved auction algorithm to allocate tasks, achieving high efficiency. Robots have priority access to energy sharing; low The robot performs high-power tasks and generates a cluster scheduling instruction set.

Images