Battery stack with intelligent temperature control function

Abstract

This invention discloses a battery stack with intelligent temperature control, belonging to the field of battery stack technology. It includes: a battery module; a multimodal passive wireless sensor network embedded within the battery module, employing a SAW resonator to monitor temperature, gas concentration, and impedance parameters; an edge computing-driven LSTM-PID co-controller that predicts the probability of thermal runaway and dynamically adjusts PID parameters based on the temperature change rate, gas concentration abrupt change, and impedance fluctuation values ​​collected by the multimodal sensor network; a dynamic hybrid air-cooled / liquid-cooled temperature control device that controls the coordinated operation of an adjustable-speed fan and a coolant pump based on the thermal runaway probability value; and a TSN blockchain communication module that transmits control commands via IEEE 802.1Qbv time slot scheduling and verifies the legality of the commands through a blockchain smart contract. This invention solves the problems of complex wiring and difficult maintenance associated with traditional sensors, while simultaneously achieving multi-parameter fusion monitoring and improving the early warning capability of the temperature control system.

Description

Technical Field

[0001] This invention belongs to the field of battery stack technology, and in particular to a battery stack with intelligent temperature control function. Background Technology

[0002] As the core device for high-energy-density electrochemical energy storage and conversion, fuel cell stacks are widely used in new energy vehicles, large-scale energy storage power stations, aerospace, and portable power systems. Their typical structure consists of multiple individual cells integrated in series and parallel, relying on a thermal management system to maintain the operating temperature within a safe and efficient range (typically 20–45°C). However, with the continuous increase in power density (e.g., the power density of automotive fuel cell stacks has exceeded 4 kW / L), the rate of heat generation inside the battery has increased dramatically. Local hot spots are prone to triggering thermal runaway chain reactions, leading to serious safety accidents such as fires and explosions. Therefore, intelligent temperature control systems have become a critical infrastructure for ensuring the safe operation of fuel cell stacks.

[0003] Currently, K-type thermocouples or NTC thermistors are commonly used and attached to the surface of the battery stack. Such sensors have three inherent defects: (1) Spatial limitation: The sensor can only contact the outer surface of the battery cell, while the core heat-generating area inside the battery (the tab connection and the micropore area of ​​the separator) is 5–10 mm away from the surface. In the heat conduction path, the graphene thermal pad (thermal conductivity ≈ 5 W / m·K) and the air gap (thermal conductivity ≈ 0.026 W / m·K) form significant thermal resistance; (2) Response hysteresis: The thermocouple has a large heat capacity (typical value > 100 J / K). Under 5C discharge conditions, when the core temperature reaches 80℃, the surface only shows 65℃, with a time delay of 12–18 seconds; (3) It only outputs temperature signals and cannot capture the associated parameters of the precursors of thermal runaway (such as CO / HF gas evolution and sudden changes in interface impedance). Its wired temperature sensor has a single monitoring dimension and has physical blind spots. Summary of the Invention

[0004] The purpose of this invention is to provide a battery stack with intelligent temperature control function to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a battery stack with intelligent temperature control function, comprising: The preferred solution employs a multimodal passive wireless sensor network: a differential SAW resonator structure embedded within the battery stack to achieve real-time monitoring of multiple parameters, including temperature, gas concentration, and impedance. Compared to traditional wired sensors, passive wireless sensors require no battery power, avoiding issues related to battery leakage, aging, or replacement. Simultaneously, the differential design counteracts the effects of temperature and other interferences, ensuring stable operation over a wide temperature range of -40℃ to 85℃, with a total frequency variation of less than 100ppm. This sensor network can capture early fault signals such as temperature gradients, gas release, and impedance changes within the battery stack, providing more comprehensive data support for intelligent temperature control.

[0006] The preferred edge computing-driven LSTM-PID co-controller in this solution combines the advantages of deep learning prediction and traditional control algorithms. It uses an LSTM neural network to predict the probability of battery thermal runaway and dynamically adjusts the PID parameters. The controller employs an FPGA chip for hardware acceleration, reducing parameter adjustment latency from milliseconds to microseconds, while also supporting real-time processing of multi-channel sensor data. Compared to traditional PID control, this invention can predict thermal runaway risk 30 seconds in advance, improving temperature control accuracy to ±0.1℃ and reducing response time to less than 50ms, significantly enhancing the real-time performance and accuracy of the temperature control system.

[0007] The preferred embodiment of this solution is a dynamic hybrid temperature control device combining air cooling and liquid cooling: it intelligently switches between air cooling and liquid cooling modes based on the thermal runaway probability value output by the LSTM-PID co-controller. Air cooling is prioritized under low heat load scenarios to reduce energy consumption; liquid cooling is switched to under high heat load scenarios to ensure effective heat dissipation. Compared to existing hybrid temperature control technologies, this invention achieves seamless switching, balancing heat dissipation and energy consumption, and significantly reducing the operating cost of the temperature control system.

[0008] The preferred TSN blockchain communication architecture of this solution employs Time-Sensitive Networking (TSN) to ensure low-latency communication, and uses the IEEE 802.1Qbv protocol to configure a GCL (Gated Control List) to transmit control commands in 1ms time slots. Simultaneously, the blockchain verification unit uses a Proof-of-Work (PoW) consensus algorithm to verify the signatures of the control commands, ensuring the trustworthiness of the command source. Compared to traditional CAN bus solutions, this invention reduces communication latency to below 20ms, significantly improving system real-time performance and effectively preventing system failures caused by network attacks.

[0009] The preferred approach in this scheme employs a fault diagnosis and safety protection linkage mechanism: It uses an SVM (Support Vector Machine) algorithm to analyze data collected by a multimodal sensor network to identify abnormal battery states (such as internal short circuits, lithium plating, and electrolyte decomposition). When a serious fault is detected, the safety protection unit immediately disconnects the battery stack's output circuit, enabling rapid fault response and handling. Simultaneously, the adaptive learning unit dynamically adjusts the parameter thresholds of the LSTM-PID co-controller based on the battery stack's historical operating data and current operating conditions, allowing the system to adapt to changes in different battery states and operating conditions.

[0010] This preferred solution utilizes a multi-system integration and digital twin interface: Through the distributed network architecture of the TSN blockchain communication module, seamless integration with BMS (Battery Management System), EMS (Energy Management System), and VMS (Vehicle Management System) is achieved, supporting access control functions of blockchain smart contracts to ensure the security and reliability of system integration. Simultaneously, the digital twin interface supports battery stack modeling and simulation, and through the auditing function of blockchain smart contracts, ensures the accuracy and reliability of digital twin data, providing support for remote monitoring and management of battery stacks.

[0011] Compared with the prior art, the technical effects and advantages of the present invention are as follows: This battery stack with intelligent temperature control solves the problems of complex wiring and difficult maintenance of traditional sensors by using a differential SAW resonator structure and phase change material encapsulation. It also achieves multi-parameter fusion monitoring, improving the early warning capability of the temperature control system. Compared with existing technologies, this sensor network exhibits a total frequency variation of less than 100ppm within a temperature range of -40℃ to 85℃, demonstrating significant technological advantages.

[0012] By predicting the probability of thermal runaway using an LSTM neural network and dynamically adjusting PID parameters, the problems of slow response speed and poor prediction accuracy in existing temperature control algorithms are solved. FPGA hardware acceleration technology reduces the parameter adjustment latency from milliseconds to microseconds, meeting real-time control requirements.

[0013] By configuring the GCL table and blockchain verification unit using the IEEE 802.1Qbv protocol, the problems of high communication latency and poor security in existing temperature control systems are solved. This communication architecture reduces communication latency to below 20ms while ensuring the integrity and trustworthiness of control commands.

[0014] By using the SVM algorithm to identify abnormal battery states and link them with the safety protection unit, the problem of existing temperature control systems lacking fault diagnosis and self-protection functions is solved. This mechanism can issue a thermal runaway warning more than 42 seconds in advance, significantly improving system safety. Attached Figure Description

[0015] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a schematic diagram of the differential SAW resonator structure of the present invention; Figure 2 This is a flowchart of the LSTM-PID collaborative controller of the present invention; Figure 3 This is a schematic diagram of the TSN GCL configuration of the present invention; Figure 4 This is a comparison chart of thermal runaway early warning times according to the present invention; Figure 5 This is a comparison chart of the battery cycle life of the present invention. Detailed Implementation

[0017] In the following description, numerous specific details are set forth in order to provide a more thorough understanding of the invention. However, it will be apparent to those skilled in the art that the invention can be practiced without one or more of these details. In other instances, certain technical features well-known in the art have not been described in order to avoid obscuring the invention.

[0018] This embodiment provides, for example Figures 1 to 5 The battery stack shown adopts a modular design with intelligent temperature control, mainly comprising five core components: a battery module, a multimodal passive wireless sensor network, an edge computing-driven LSTM-PID co-controller, an air-cooled / liquid-cooled dynamic hybrid temperature control device, and a TSN blockchain communication module. These components are interconnected through standardized interfaces to form a complete intelligent temperature control closed-loop system.

[0019] In this embodiment, the battery module is composed of multiple individual cells connected in series / parallel, and the individual cells are connected by a thermally conductive material to form an integral battery stack structure. The thermally conductive material is a graphene composite material with a thermal conductivity ≥5000W / m·K, ensuring uniform temperature distribution inside the battery stack. A multimodal passive wireless sensor network is embedded inside the battery module. The sensors are evenly distributed along the surface of the battery cells inside the stack, with one sensor embedded every 10cm², specifically in locations between the cell tabs and within the cooling plate channels.

[0020] In this embodiment, a multimodal passive wireless sensor network is implemented: multiple differential SAW resonator sensors are arranged inside the battery stack. The sensors are encapsulated in a nickel-based alloy shell and a fluororubber sealing layer. Heat dissipation fins are provided on the shell surface, and the sealing layer is filled with a phase change material (phase change temperature 60℃±5℃) to adapt to the harsh environment inside the battery stack. The sensors transmit data wirelessly, eliminating the need for complex wiring and reducing system complexity and maintenance costs.

[0021] In this embodiment, the edge computing-driven LSTM-PID co-controller employs an FPGA chip for hardware acceleration and internally includes an LSTM neural network module and a PID controller module. The LSTM neural network module receives input parameters including the rate of temperature change, sudden changes in gas concentration, and impedance fluctuations, and outputs a thermal runaway probability value. The PID controller module dynamically adjusts the proportional gain Kp, integral time Ti, and derivative time Td based on the thermal runaway probability value. For example, when the thermal runaway probability value exceeds a preset threshold, Kp is adjusted by increasing the proportional gain by 5% for every 10% increase in probability, Ti is adjusted by decreasing the integral time by 20% for every 10% increase in probability, and Td is adjusted by increasing the derivative time by 15% for every 10% increase in probability.

[0022] In this embodiment, the air-cooled-liquid-cooled dynamic hybrid temperature control device includes an adjustable-speed fan, a coolant pump, and coolant piping. The control system controls the coordinated operation of the adjustable-speed fan and coolant pump based on the thermal runaway probability value. For example, under 5C discharge conditions, when the thermal runaway probability value is lower than a preset threshold, the control system prioritizes the operation of the air-cooled subsystem, and the speed of the adjustable-speed fan is adjusted according to the control signal output by the LSTM-PID co-controller; when the thermal runaway probability value exceeds the preset threshold, the control system activates the liquid-cooled subsystem, and the flow rate of the coolant pump is adjusted according to the control signal output by the LSTM-PID co-controller.

[0023] In this embodiment, the TSN blockchain communication module uses the IEEE 802.1Qbv protocol to configure the GCL (Gated Control List) and divides time slots into 1ms periods for transmitting control commands. Simultaneously, the blockchain verification unit uses the Proof-of-Work (PoW) consensus algorithm to verify the signatures of the control commands, ensuring the trustworthiness of the command source. The network management unit manages the network configuration of the TSN blockchain communication module, including the IEEE 802.1AS time synchronization protocol, IEEE 802.1Qca path control, and reservation protocols, ensuring network stability and reliability.

[0024] In this embodiment, the multimodal passive wireless sensor network is implemented using a differential SAW resonator structure, such as... Figure 1As shown, two interdigitated electrodes (IDTs) are arranged in opposite directions on a quartz substrate, with a spacing of 0.5 mm. When the temperature changes, the frequencies of the two resonators change in opposite directions, and temperature interference can be canceled out through differential calculation. The formula for calculating the differential frequency change is: Δf = (f2 - f1) / (f1 + f2) × 100%; where f1 and f2 are the frequency values ​​of the two resonators, respectively. Through this design, this technology can control the total frequency variation within 100 ppm in a temperature range of -40℃ to 85℃.

[0025] In this embodiment, the SAW resonator is encapsulated with a nickel-based alloy (such as Niloy52) housing and a fluororubber sealing layer. The nickel-based alloy housing is 1 mm thick and has a diameter of 5.0-5.5 × 10⁻⁶ mm. -6 The coefficient of thermal expansion (20-300℃) is highly compatible with the quartz substrate, avoiding structural deformation caused by thermal stress. The fluororubber sealing layer is 0.5mm thick, exhibiting excellent resistance to electrolyte corrosion and a wide operating temperature range of -20℃ to 250℃. The sealing layer is filled with a phase change material (ortho-capric acid-paraffin / expanded graphite composite material, mass ratio 8:1:1), with a phase change temperature of 60℃±5℃. This material can absorb the heat generated inside the battery and release it when needed, maintaining a stable sensor operating temperature.

[0026] In this embodiment, the sensors are evenly distributed along the surface of the cells inside the battery stack, with one sensor arranged every 10 cm². Specific locations include between cell tabs and within cooling plate channels. The sensors are fixed inside the stack using M4 screws and 3M 8810 thermally conductive adhesive to ensure good thermal conductivity. The sensor signals are wirelessly transmitted to the edge computing controller via 433MHz RFID over a distance of 50 cm, meeting the communication requirements within the battery stack.

[0027] In this embodiment, the SAW resonator monitors temperature by frequency changes and monitors gas concentration and impedance changes by the cross-change between the resonant frequency and impedance. The specific monitoring principle is as follows: Temperature monitoring: Temperature changes cause changes in the sound velocity of the quartz substrate, which in turn changes the resonant frequency.

[0028] Gas concentration monitoring: Gas release causes changes in the mass load on the substrate surface, resulting in a shift in the resonant frequency.

[0029] Impedance monitoring: Changes in the internal impedance of the battery are reflected in the resonant frequency through changes in the electric field distribution on the substrate surface.

[0030] Sensor data is fused using a Kalman filter algorithm to eliminate noise interference and improve data reliability. The mathematical expression of the Kalman filter algorithm is: ; in, This is the optimal estimate at time k. Gain for Karlshah H represents the sensor measurement value, and H is the observation matrix.

[0031] LSTM Neural Network Module: The LSTM neural network uses a three-layer structure: 128 neurons in the input layer, 256 neurons in the hidden layer, and 1 neuron in the output layer. The network input parameters include the rate of temperature change (…). ), gas concentration mutation ( ) and impedance fluctuation value ( The LSTM network's predictions are normalized and then input into the network. The predictions are then converted into thermal runaway probability values ​​using the Sigmoid function. : ; Where h(t) is the output of the LSTM hidden layer, and w is the network weight.

[0032] PID controller module: The PID controller adopts a discrete form, and the control formula is: ; Where Kp is the proportional coefficient, Ki is the integral coefficient, Kd is the differential coefficient, and e(k) is the temperature error at the current moment.

[0033] FPGA Hardware Acceleration: The LSTM-PID co-controller employs a Xilinx Zynq UltraScale+ MPSoC chip for hardware acceleration. The LSTM network utilizes a deeply pipelined parallel computing architecture on the FPGA, decomposing network computation into multiple parallel processing units, each handling different time steps or neurons. The PID controller employs a fixed-point arithmetic module, converting floating-point operations into 8-bit fixed-point operations to improve computational efficiency.

[0034] The FPGA acceleration performance parameters are as follows: LSTM network throughput: 13.7 GOP / s PID parameter tuning delay: 9.6μs Hardware resource usage: 1.28%. Dynamic parameter adjustment mechanism: The PID parameters are dynamically adjusted based on the thermal runaway probability value. The adjustment formula is as follows: ; in, Initial PID parameters P is the probability value of thermal runaway. When P exceeds a preset threshold (e.g., 30%), the control system activates the liquid cooling subsystem; when P is below the threshold, the control system prioritizes controlling the air cooling subsystem.

[0035] Control algorithm flow: The workflow of the LSTM-PID co-controller is as follows Figure 2 As shown: Sensor networks collect temperature, gas concentration, and impedance data; The data fusion unit uses the Kalman filter algorithm to process multidimensional data and eliminate noise interference; The LSTM neural network module analyzes and processes the data to predict the probability value P of thermal runaway. The PID controller module dynamically adjusts the parameters Kp, Ti, and Td based on the P value; The adjusted PID parameters are transmitted to the air-cooled-liquid-cooled hybrid temperature control device via the TSN blockchain communication module; The temperature control device adjusts the speed of the adjustable fan or the flow rate of the coolant pump according to the control parameters. The control results are fed back to the LSTM-PID co-controller to form a closed-loop control.

[0036] TSN Blockchain Communication Module Configuration Method: TSN Network Interface Configuration: The TSN network interface uses the IEEE 802.1Qbv protocol to configure the GCL (Gated Control List), dividing time slots into 1ms periods for transmitting control commands. The GCL table parameters are as follows:

[0037] In this code, "O" indicates that the queue is open, allowing data transmission; "C" indicates that the queue is closed, prohibiting data transmission. Control commands are transmitted within a time slot of 0-200μs to ensure low latency; a guard band of 50μs is set to prevent other traffic from interfering with the transmission of control commands.

[0038] In this embodiment, the blockchain verification unit implements the following: The blockchain verification unit uses the Proof-of-Work (PoW) consensus algorithm to verify the control instruction signature, ensuring the trustworthiness of the instruction source. The ECDSA signature verification formula is: ; Where s is the signature value, k is the random number, e is the hash value, r is the signature parameter, d is the private key, and n is the elliptic curve order.

[0039] The verification process is as follows: The sender (such as an LSTM-PID controller) uses the private key d to sign the control command, generating a signature value s and parameter r; The receiver (such as an air-cooled / liquid-cooled hybrid temperature control device) uses the sender's public key Q=(x,y)=k·G, where G is the base point and k is the private key, to verify the signature: calculate ; calculate ; calculate ; verify .

[0040] If the verification passes, the receiver executes the control command; if the verification fails, the receiver ignores the command and sends an error signal.

[0041] In this embodiment, the network security protection mechanism includes a TSN blockchain communication module comprising message signature verification, an access control list, and an intrusion detection system. Message signature verification ensures the trustworthiness of the instruction source; the access control list restricts the access permissions of specific nodes; and the intrusion detection system monitors network traffic, identifies abnormal behavior, and triggers protective measures.

[0042] In this embodiment, communication latency is optimized by ensuring that the time deviation of all nodes is less than 1 μs through the IEEE 802.1AS time synchronization protocol; and by optimizing the data transmission path through IEEE 802.1Qca path control and reservation protocols. In actual testing, the end-to-end latency of control commands is controlled to be less than 20 ms, which is about 25 times more reliable than the traditional CAN bus solution (500 ms).

[0043] In this embodiment, the fault diagnosis and safety protection linkage mechanism utilizes the SVM (Support Vector Machine) algorithm: The fault diagnosis unit employs the SVM algorithm to analyze data collected by the multimodal sensor network and identify abnormal battery states (such as internal short circuits, lithium plating, and electrolyte decomposition). The mathematical expression of the SVM algorithm is: ; Where, α i For the support vector coefficients, y i Training labels is the kernel function, and b is the bias term.

[0044] Abnormal State Identification: By training an SVM model, the following abnormal states are identified: (1) Internal short circuit: abnormal temperature rise ( And the impedance drops sharply.

[0045] (2) Lithium plating: abnormal temperature gradient (temperature difference between adjacent sensors > 3 K) and impedance fluctuation.

[0046] (3) Electrolyte decomposition: sudden change in gas concentration (CO>5ppm or HF>1ppm).

[0047] In this embodiment, when a serious fault is detected, the safety protection unit immediately cuts off the output circuit of the battery stack, enabling rapid fault response and handling. The working process of the safety protection unit is as follows: (1) The fault diagnosis unit identifies abnormal states and calculates the severity of faults; (2) When the severity of the fault exceeds the threshold, the safety protection unit generates an emergency control command; (3) Emergency control commands are transmitted through the high-priority channel of the TSN blockchain communication module; (4) The temperature control device that receives the instruction immediately performs an emergency operation (such as forcibly starting the liquid cooling or cutting off the output).

[0048] In this embodiment, the adaptive learning unit dynamically adjusts the parameter thresholds of the LSTM-PID collaborative controller based on the historical operating data and current operating conditions of the battery stack. The learning algorithm is based on the gradient descent method, and the parameter adjustment formula is as follows: Where θ is the controller parameter and η is the learning rate. The gradient of the cost function is given. The adaptive learning unit can dynamically adjust the control strategy based on the state of health (SOH) of the battery stack and the ambient temperature, thereby improving the system's adaptability.

[0049] In this embodiment, the multi-system integration interface is implemented as follows: The multi-system integration interface adopts the distributed network architecture of the TSN blockchain communication module, supports the IEEE 802.1Qcc protocol, and achieves seamless integration with BMS (Battery Management System), EMS (Energy Management System) and VMS (Vehicle Management System).

[0050] Data format: The integration interface uses a unified data format, which includes the following fields: (1) Timestamp (IEEE 802.1AS synchronization).

[0051] (2) Sensor data (temperature, gas concentration, impedance).

[0052] (3) Control commands (air-cooled / liquid-cooled mode, fan speed, pump flow rate).

[0053] (4) Blockchain verification information (signature, hash value, timestamp).

[0054] In this embodiment, the access control blockchain smart contract implements access control functionality, allowing only authorized nodes to access specific data or execute specific control commands. Access control rules are preset through the smart contract and include: (1) The BMS can read temperature data but cannot modify control commands; (2) The EMS can read temperature and gas concentration data but cannot modify control commands; (3) VMS can read all data but cannot modify control instructions.

[0055] In this embodiment, the digital twin interface is implemented as follows: The digital twin interface supports the modeling and simulation of the battery stack, simulating the operating state of the battery stack through an electrochemical-thermal coupling model. Model parameters include: (1) Battery electrochemical parameters (open circuit voltage, internal resistance, capacity); (2) Thermal management parameters (heat dissipation coefficient, heat capacity, phase change material characteristics); (3) Environmental parameters (temperature, humidity, air pressure).

[0056] In this embodiment, the digital twin interface utilizes the distributed network architecture of the TSN blockchain communication module to achieve data synchronization between the physical battery stack and the digital twin model. The synchronization frequency is 100Hz to ensure that the digital twin model can reflect the status of the physical battery stack in real time.

[0057] In this embodiment, a blockchain smart contract implements the auditing function of digital twin data, ensuring that the data is immutable. All data changes are recorded on the blockchain, forming an irreversible audit trail. The auditing function supports remote monitoring and management, providing a basis for the maintenance and upgrading of the battery stack.

[0058] In this embodiment, the energy efficiency optimization unit implements the following process: 1. Energy consumption monitoring: The energy efficiency optimization unit monitors the energy consumption of the air-cooled subsystem and the liquid-cooled subsystem, including fan power consumption, pump power consumption and coolant circulation energy consumption.

[0059] 2. Dynamic Optimization Algorithm: The energy efficiency optimization unit adopts a multi-objective optimization algorithm based on heat load prediction to balance heat dissipation effect and energy consumption. Optimization objectives include: minimizing total energy consumption, maximizing heat dissipation effect, and maintaining uniform temperature distribution.

[0060] 3. Distributed Computing: The energy efficiency optimization unit utilizes distributed computing resources from the TSN blockchain communication module to achieve global energy efficiency optimization. Distributed computing nodes include: edge computing units (FPGA chips), cloud servers, and controllers for other battery stacks.

[0061] In this embodiment, the state estimation unit is implemented as follows: 1. SOC Estimation: The state estimation unit uses the Kalman filter algorithm to estimate the SOC (State of Charge) of the battery stack. The formula is as follows: ; in, Let be the charge / discharge amount at time k. This is the maximum capacity of the battery.

[0062] 2. SOH Estimation: The state estimation unit adopts an SOH (Health State) estimation algorithm based on impedance change, the formula of which is: ; Among them, Zk Z is the impedance value at time k, and Z0 is the initial impedance value.

[0063] 3. Predictive Capability: The state estimation unit utilizes the predictive capability of the LSTM-PID collaborative controller to improve estimation accuracy. The LSTM network predicts the SOC and SOH trends for the next 10 minutes, and the state estimation unit adjusts the estimation parameters based on the prediction results.

[0064] In this embodiment, the thermal equalization unit is implemented as follows: 1. Temperature distribution monitoring: The thermal equalization unit monitors the temperature distribution data collected by the multimodal passive wireless sensor network and identifies local hot spots or cold spots.

[0065] 2. Zoned Cooling Control: The thermal equalization unit controls the zoned cooling function of the cooling system based on temperature distribution data. Zoned cooling control includes: Local airflow control of the air-cooled subsystem: adjusting the fan speed in specific areas. Flow distribution control of the liquid cooling subsystem: adjusting the coolant flow rate in specific areas. 3. Equalization Algorithm: The thermal equalization unit adopts an equalization algorithm based on temperature gradient, the formula of which is: ; Among them, Q i Let Q0 be the cooling amount for the i-th region, and T be the base cooling amount. i Let T be the temperature of the i-th region. a This represents the average temperature.

[0066] In this embodiment, the security isolation unit is implemented as follows: 1. Distributed Architecture: The security isolation unit adopts a distributed architecture of TSN blockchain communication module to achieve secure isolation between the battery stack and other systems.

[0067] 2. Access Control: The secure isolation unit supports access control functions for blockchain smart contracts to prevent unauthorized access. Access control rules include: only authorized nodes are allowed to access specific data or execute specific control commands; nodes with different permission levels have different access permissions; access records are recorded on the blockchain, forming a traceable audit trail.

[0068] In this embodiment, the process of implementing the remote monitoring interface includes: 1. The remote monitoring interface adopts the distributed network architecture of the TSN blockchain communication module, supporting multiple access methods such as Web, APP and PC.

[0069] 2. The remote monitoring interface displays the real-time status and historical data of the battery stack, including: temperature distribution heat map, thermal runaway probability curve, charge / discharge status (SOC, SOH), and fault history records.

[0070] 3. The remote monitoring interface receives control commands input by the user and executes them after verification through the TSN blockchain communication module. Control commands include: temperature control target value, charge / discharge limits, and maintenance commands.

[0071] In this embodiment, the OTA upgrade interface adopts a distributed network architecture of the TSN blockchain communication module to realize remote upgrades of the battery stack's software and firmware. The upgrade process includes: confirming the differences between the current version and the target version; downloading the encrypted upgrade package from the cloud; verifying and decrypting the upgrade package signature using a blockchain smart contract; executing the upgrade in stages to ensure system stability; and verifying the upgrade result and recording it on the blockchain. The OTA upgrade interface supports the verification function of the blockchain smart contract to ensure the security and reliability of the upgrade. The upgrade package must contain a valid signature and hash value; otherwise, the upgrade will be rejected.

[0072] In this embodiment, the multi-protocol conversion interface includes: (1) the multi-protocol conversion interface supports multiple communication protocols such as CAN, Modbus, and MQTT, enabling the battery stack to interface with systems using different communication protocols. (2) the multi-protocol conversion interface converts a unified data format into corresponding protocol messages according to the communication protocol of the target system. The conversion logic includes: converting the unified data format into a message format of a specific protocol, adjusting parameters such as message length, priority, and transmission frequency, and optimizing communication performance according to the characteristics of the target system.

[0073] (3) The multi-protocol conversion interface supports the authentication function of blockchain smart contracts to ensure the security of protocol conversion. All protocol conversion operations need to be verified by the blockchain to prevent security vulnerabilities in the protocol conversion process.

[0074] Example 1 The intelligent temperature control system of this invention, applied in the fuel cell stack of a hydrogen fuel cell vehicle, can stably control the stack temperature within the optimal operating range of 70-90°C. The specific implementation steps are as follows: Sensor arrangement: 100 differential SAW resonator sensors are arranged inside the fuel cell stack, evenly distributed along the interior of the stack, every 10 cm. 2 One sensor is installed. The sensor is fixed inside the fuel cell stack using M4 screws and 3M 8810 thermally conductive adhesive to ensure good thermal conductivity.

[0075] Controller configuration: An LSTM-PID collaborative controller is implemented using a Xilinx Zynq UltraScale+MPSoC chip, with the following configuration parameters: LSTM network: 128 neurons in the input layer, 256 neurons in the hidden layer, and 1 neuron in the output layer. PID controller: Initial parameters Kp0=0.4, Ti0=5s, Td0=0.1s FPGA resource usage: 1.28% Parameter adjustment delay: 9.6μs Temperature control device configuration: The air-cooled-liquid-cooled dynamic hybrid temperature control device includes: Three adjustable speed fans, maximum speed 12000rpm Liquid cooling pump, maximum flow rate 5L / min The coolant piping is made of copper alloy and has an inner diameter of 8mm. TSN communication configuration: Configure the GCL table of the IEEE 802.1Qbv protocol to divide the time slots into 1ms periods for transmitting control commands. Control commands are transmitted within time slots of 0-200μs to ensure low latency.

[0076] Test results: Under 5C discharge conditions, the system can predict the risk of thermal runaway 30 seconds in advance and automatically switch to liquid cooling mode, keeping the maximum temperature of the fuel cell stack below 40°C and the temperature difference below 1.5K. Compared with traditional liquid cooling systems, this invention can reduce communication latency to below 20ms and improve communication reliability by approximately 25 times compared with traditional CAN bus solutions.

[0077] Example 2 The intelligent temperature control system of this invention, when applied to lithium-ion battery stacks in energy storage power stations, can effectively prevent thermal runaway and improve cycle life. The specific implementation steps are as follows: Sensor arrangement: 200 differential SAW resonator sensors are arranged inside the lithium battery stack, evenly distributed along the interior of the stack, every 10 cm. 2 One sensor is installed. The sensor is fixed inside the fuel cell stack using M4 screws and 3M 8810 thermally conductive adhesive.

[0078] Controller configuration: An LSTM-PID collaborative controller is implemented using a Xilinx Zynq UltraScale+MPSoC chip, with the following configuration parameters: LSTM network: 128 neurons in the input layer, 256 neurons in the hidden layer, and 1 neuron in the output layer.

[0079] PID controller: initial parameters Kp0=0.3, Ti0=6 s, Td0=0.05 s.

[0080] FPGA resource usage: 1.28%; Parameter adjustment delay: 9.6μs; Temperature control device configuration: The air-cooled-liquid-cooled dynamic hybrid temperature control device includes: Five adjustable speed fans, maximum speed 10,000 rpm Liquid cooling pump, maximum flow rate 10L / min The coolant piping is made of stainless steel and has an inner diameter of 12mm. TSN communication configuration: Configure the GCL table for the IEEE 802.1Qbv protocol to divide time slots into 1ms periods for transmitting control commands. Control commands are transmitted within time slots ranging from 0 to 200μs.

[0081] Test results: Under 5C discharge conditions, the system can control the maximum temperature of the battery stack below 40°C and the temperature difference below 1.5K, reducing temperature fluctuation by approximately 5K compared to traditional liquid cooling systems. Experimental data shows that this invention can extend battery cycle life by 12%-15%, significantly outperforming traditional temperature control systems.

[0082] Example 3 The intelligent temperature control system of this invention, applied in drone battery packs, enables rapid heat dissipation during high-rate discharge, preventing performance degradation. Specific implementation steps are as follows: Sensor Layout: Fifty differential SAW resonator sensors are placed inside the drone's battery pack, evenly distributed along the inside of the battery pack, spaced 10cm apart. 2 One sensor is installed. The sensor is secured inside the battery pack using M4 screws and 3M 8810 thermally conductive adhesive.

[0083] Controller configuration: An LSTM-PID collaborative controller is implemented using a Xilinx Zynq UltraScale+MPSoC chip, with the following configuration parameters: LSTM network: 64 neurons in the input layer, 128 neurons in the hidden layer, and 1 neuron in the output layer. PID controller: Initial parameters Kp0=0.5, Ti0=4s, Td0=0.2s FPGA resource usage: 1.28%; Parameter adjustment delay: 9.6μs; Temperature control device configuration: The air-cooled-liquid-cooled dynamic hybrid temperature control device includes: Two adjustable speed fans, with a maximum speed of 15,000 rpm; Miniature liquid cooling pump, maximum flow rate 1L / min; The coolant piping is made of aluminum alloy and has an inner diameter of 4mm. TSN communication configuration: Configure the GCL table for the IEEE 802.1Qbv protocol to divide time slots into 1ms periods for transmitting control commands. Control commands are transmitted within time slots ranging from 0 to 200μs.

[0084] Test results: Under the condition of rapid drone rotation (equivalent to 3C discharge), the system can control the maximum battery temperature below 40°C, which is 5°C lower than that of traditional liquid cooling systems. At the same time, the system weight of this embodiment does not increase by more than 10% compared to traditional liquid cooling systems, meeting the lightweight requirements of drones.

[0085] Experimental verification shows that the intelligent temperature control system of the present invention has the following significant effects:

[0086] It should be noted that, in this document, relational terms such as "one" and "two" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, the phrase "comprising an element defined as..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0087] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A battery stack with intelligent temperature control function, characterized in that, include: A battery module is composed of multiple individual battery cells connected in series or parallel. A multimodal passive wireless sensor network, embedded inside a battery module, uses a SAW resonator to monitor temperature, gas concentration, and impedance parameters; An edge computing-driven LSTM-PID co-controller predicts the probability of thermal runaway and dynamically adjusts PID parameters based on the temperature change rate, gas concentration abrupt change, and impedance fluctuation values ​​collected by a multimodal sensor network. The air-cooled-liquid-cooled dynamic hybrid temperature control device controls the coordinated operation of the adjustable speed fan and coolant pump based on the thermal runaway probability value; The TSN blockchain communication module transmits control commands via IEEE 802.1Qbv time slot scheduling and verifies the legality of the commands through blockchain smart contracts.

2. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The multimodal passive wireless sensor network adopts a differential SAW resonator structure. By arranging two resonators at different positions on the substrate, the effects of temperature and interference factors are offset. The resonators use an AT-cut quartz substrate, and the total frequency change is less than 100ppm in the temperature range of -40℃ to 85℃.

3. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The edge computing-driven LSTM-PID co-controller includes: The LSTM neural network module takes temperature change rate, gas concentration abrupt change and impedance fluctuation value as input parameters, and outputs thermal runaway probability value. The PID controller module dynamically adjusts the proportional coefficient Kp, integral time Ti, and derivative time Td based on the thermal runaway probability value. FPGA chips are used for hardware acceleration of LSTM models and PID parameter calculations, reducing parameter tuning latency from milliseconds to microseconds. The data fusion unit uses the Kalman filter algorithm to process multi-dimensional sensor data and eliminate noise interference.

4. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The air-cooled-liquid-cooled dynamic hybrid temperature control device includes: Adjustable speed fan for air cooling; Coolant pump and coolant piping for liquid cooling heat dissipation; The control module prioritizes the operation of the air-cooled subsystem when the probability of thermal runaway is below a preset threshold, and activates the liquid-cooled subsystem when the probability of thermal runaway exceeds the preset threshold.

5. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The TSN blockchain communication module includes: The TSN network interface uses the IEEE 802.1Qbv protocol to configure the GCL and divides the time slots into 1ms periods to transmit control commands. The blockchain verification unit uses a proof-of-work consensus algorithm to verify the signature of control instructions, ensuring the credibility of the instruction source. Network security protection mechanisms, including message signature verification, access control lists, and intrusion detection systems, prevent network attacks.

6. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The control system also includes a fault diagnosis unit, which uses the SVM support vector machine algorithm to analyze the data collected by the multimodal sensor network and identify abnormal battery states, including internal short circuits, lithium plating, and electrolyte decomposition.

7. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The control system also includes an adaptive learning unit that dynamically adjusts the parameter thresholds of the LSTM-PID co-controller based on the historical operating data of the battery stack and the current operating conditions.

8. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The control system also includes a safety protection unit that controls the charging and discharging state of the battery stack based on the output signal of the LSTM-PID co-controller and the diagnostic results of the fault diagnosis unit. When a serious fault is detected, the output circuit of the battery stack is immediately cut off.

9. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The control system also includes a multi-system integration interface, adopts a distributed network architecture with TSN blockchain communication modules, achieves seamless integration with BMS, EMS and VMS, and supports access control functions for blockchain smart contracts.

10. A battery stack with intelligent temperature control function according to claim 1, characterized in that: The control system also includes a digital twin interface, which adopts a distributed network architecture of TSN blockchain communication module to realize digital twin modeling and simulation of battery stack, supports the auditing function of blockchain smart contracts, and ensures the accuracy and reliability of digital twin data.

Images