System for testing heat insulation performance of heat insulation pad of power battery

Abstract

The invention provides a power battery heat insulation pad heat insulation performance test system, and relates to the field of heat insulation pad performance testing, the system comprises a control module and a test module, the test module is arranged on the upper side of the control module, and the test module comprises a test box body (11), a driving rod (12), a lifting plate (13), a pressure assembly, an environment box body (15) and a heating assembly. According to the system, through cooperation of environment-temperature-pressure, multi-factor coupling evaluation of the heat insulation performance of the heat insulation pad is realized, so that the actual working condition faced by the heat insulation pad during thermal runaway of the power battery is simulated, dynamic, real-time and full-cycle evaluation of the heat insulation performance of the heat insulation pad is completed, and the accuracy and stability of performance evaluation are ensured.

Description

Technical Field

[0001] This invention relates to the field of thermal insulation pad performance testing technology, and specifically to a thermal insulation performance testing system for power battery thermal insulation pads. Background Technology

[0002] Thermal safety of power batteries is a core safety guarantee for new energy vehicles and energy storage systems. Among them, thermal insulation pads, as key components between modules, can delay or block the thermal runaway of cells and are an important barrier to maintain system safety.

[0003] Currently, the testing of thermal insulation performance of heat insulation pads mainly consists of a heating source, fixtures, and a temperature monitoring system. This system simulates thermal runaway by heating one side, monitoring the temperature rise of the cold side to evaluate the pad's insulation performance. Some devices add a pressure mechanism to simulate the pad under pressure, but due to temperature fluctuations caused by the edge effect of the heating source and the lag in pressure loading, the actual data measured by the sensors has dynamic errors (i.e., large deviations between test data and actual data). This not only affects the accuracy of the test data but also leads to poor test repeatability (i.e., poor consistency of results across multiple tests). Furthermore, existing technologies only control temperature or pressure changes and obtain the insulation performance of the pad using temperature test values ​​from the cold and hot sides, without considering the influence of the heat-load-humidity coupling effect on insulation performance. For example, theoretically, the basic thermal conductivity of the heat insulation pad is related to the material itself, but in actual working conditions, increased temperature accelerates molecular thermal motion (thermal field effect), increased pressure compresses porosity (force field effect), and increased humidity leads to moisture penetration (wet field effect). The combined effect of these three factors causes a change in the material's thermal conductivity. Furthermore, relying solely on static values ​​of cold and hot surface temperatures to evaluate the thermal insulation performance of thermal insulation pads not only fails to cover the entire lifecycle performance of the pads (it only reflects the temperature difference at the current instant; conducting full-cycle testing would be time-consuming and costly), but also fails to eliminate the ambiguity of subjective judgment (i.e., evaluation using only cold and hot surface temperatures is usually a qualitative description that lower cold surface temperatures and larger temperature differences are better, which cannot accurately distinguish the performance differences of thermal insulation pads of different thicknesses and materials); in addition, cold and hot surface temperatures are greatly affected by test conditions (such as heating power, pressure, humidity, etc.), and temperature data from different tests cannot be directly compared, resulting in poor test repeatability (i.e., the test performance of thermal insulation pads from different batches, models, and environments cannot be reliably compared). Summary of the Invention

[0004] To address the problems existing in the prior art, the present invention aims to provide a power battery heat insulation pad thermal insulation performance testing system. This system achieves multi-factor coupling evaluation of the heat insulation performance of the heat insulation pad through the coordinated operation of environment, temperature and pressure, thereby completing the dynamic, real-time and full-cycle evaluation of the heat insulation performance of the heat insulation pad, ensuring the accuracy and stability of the performance evaluation.

[0005] The objective of this invention is achieved through the following technical solution: A power battery heat insulation pad thermal insulation performance testing system includes a control module and a testing module. The testing module is located on the upper side of the control module and includes a test chamber, drive rods, a lifting plate, a pressure component, an environmental chamber, and a heating component. The test chamber is located at the end of the control module and contains the environmental chamber. The environmental chamber is fixedly mounted on the end face of the control module and is coaxially arranged with the test chamber. Two drive rods are symmetrically arranged on both sides of the environmental chamber. A lifting plate is located between the two drive rods and on the upper side of the environmental chamber. A pressure component is located on the bottom surface of the lifting plate, with the end of the pressure component away from the lifting plate located inside the environmental chamber. A heating component is located at the bottom of the environmental chamber, corresponding to the pressure component.

[0006] Based on further optimization of the above scheme, the pressure component includes a pressure rod and a pressure block. The pressure rod is fixedly installed on the bottom surface of the lifting plate and its end away from the lifting plate penetrates through the top surface of the environmental chamber. The pressure block is installed at the bottom end of the pressure rod. The heating component includes a heating source and a heat spreader. The heating source is fixedly installed on the end face of the control module and a heat spreader is installed on its end face.

[0007] Based on further optimization of the above scheme, a sensor system is installed inside the test chamber. The sensor system includes a hot surface temperature sensor, a cold surface temperature sensor, a piezoelectric pressure sensor, a capacitive humidity sensor, and a laser displacement sensor. The hot surface temperature sensors are evenly distributed on the end face of the heat spreader plate, the cold surface temperature sensors are evenly distributed on the end face of the heat insulation pad to be tested, and the piezoelectric pressure sensors are evenly distributed on the end face of the heat insulation pad to be tested, with the piezoelectric pressure sensors and cold surface temperature sensors being positioned opposite each other. The capacitive humidity sensor is installed inside the environmental chamber cavity to detect the internal environmental humidity. The laser displacement sensor is installed on the top surface of the inner cavity of the test chamber, with its transmitting end vertically aligned with the end face of the lifting plate (a laser signal receiving end can be installed on the end face of the lifting plate).

[0008] Based on further optimization of the above scheme, the lifting plate can be driven to move up and down by any one of the following drive sources: servo motor, pneumatic device, or hydraulic device.

[0009] Based on further optimization of the above scheme, the specific testing method for the thermal insulation performance of the test system includes: Step S1, Pre-test preparation: Deploy the sensor system and the heat insulation pad under test in the test chamber as required, wherein the heat insulation pad under test is placed on the end face of the heat spreader. Step S2, Data Acquisition: Turn on the heating and pressure components, conduct tests, and simultaneously collect data on cold / hot surface temperature, contact pressure, ambient humidity, and real-time thickness of the insulation pad. Step S3, Dynamic Error Compensation: Kalman filtering is used to fuse multiple sets of data from the sensor system to obtain the corrected true state value of the sensor. Step S4, Coupling effect quantification: Substitute the corrected sensor parameters into the coupling coefficient model to obtain the effective thermal conductivity and effective thermal insulation resistance. Step S5, Real-time performance evaluation: Evaluate the thermal insulation performance of the thermal insulation pad based on the effective thermal conductivity and effective thermal insulation resistance. Step S6, Long-term performance prediction: Based on the pre-trained LSTM (Long Short-Term Memory) time series model, predict the effective thermal insulation resistance of the insulation pad at future times and evaluate the performance degradation trend.

[0010] Based on further optimization of the above scheme, in step S2, an FPGA clock synchronization module (synchronizer) is used to control the synchronization of various sensor data to ensure that the sampling time difference of all sensors is no greater than 1ms; at the same time, the sampling frequency is 10Hz, that is, it is sampled once every 0.1s.

[0011] Based on further optimization of the above scheme, in step S3, the dynamic error compensation specifically includes: First, define the state vector. X k for: ,in, T hot-k express k The hot surface temperature is measured at all times. T cold-k express k The temperature of the cold surface is measured at all times. P k express k The pressure of constant feedback RH k express k The ambient humidity is measured at all times. express k The thickness of the insulation pad at any given time; Input vector U k for: ,in, Q k express k Heating power at any time F k express k Constant pressure driving force H in-k express k Humidification level at any given time; The system state equation is: ; In the formula: A This represents a 5×5 state transition matrix, used to describe the influence of the previous time step on the current time step; BThis represents a 5×3 input matrix used to describe the effect of the input on the state; W k-1 Indicates process noise; The observation equation is: ; In the formula: H This represents a 5×5 observation matrix; V k Indicates observation noise; The Kalman filter update equation is: Prediction step: ; In the formula: express k The predicted state vector at time t. express k The optimal state vector at time -1 (the final state estimate after combining the observations at time k-1). express k The prediction covariance matrix at time 1, express k The optimal covariance matrix at time -1 (description) (the degree of uncertainty) O Represents the process noise variance matrix (describing the uncertainty of "unknown disturbances during state transition"); Update steps: ; In the formula: K k express k Kalman gain at time step R This represents the observation noise variance matrix (describing the uncertainty of "sensor's own error"); express k The optimal state vector at time (the final state estimate combining the predicted state and the observed values). express k The optimal covariance matrix at time (description) (the degree of uncertainty) I Represents the identity matrix; The final output is the corrected true state value. By fusing multi-dimensional sensor data in real time through Kalman filtering, dynamic errors such as heating hysteresis, pressure fluctuations, and humidity penetration delays are compensated, thereby effectively reducing measurement errors of core parameters such as temperature and pressure and improving test accuracy.

[0012] Based on further optimization of the above scheme, in step S4, the coupling effect quantification specifically involves: Establish a coupling coefficient model of thermal-mechanical-humidity, including the effective thermal conductivity. With effective thermal insulation resistance : ; ; In the formula: k T , k P , k RH , k TP , k TRH , k PRH These represent the temperature field coupling coefficient, pressure field coupling coefficient, humidity field coupling coefficient, temperature-pressure coupling coefficient, temperature-humidity coupling coefficient, and pressure-humidity coupling coefficient, respectively. Indicates the pressure compressibility coefficient; This indicates the difference between the actual hot surface temperature and the standard temperature. T Deviation between 0 and 0: ; Indicates the difference between actual pressure and standard pressure P Deviation between 0 and 0: ; This indicates the difference between actual relative humidity and standard humidity. RH Deviation between 0 and 0: ; The true state value obtained in step S3 Real-time hot surface temperature in the vector T hot Real-time pressure P Real-time relative humidity RH and real-time insulation pad thickness Substituting into the coupling coefficient model, the real-time effective thermal conductivity is obtained. With effective thermal insulation resistance .

[0013] Based on further optimization of the above scheme, in step S5, the temperature rise rate is... r T Maximum thermal insulation temperature difference With insulation failure time t f A comprehensive evaluation of the thermal insulation performance of the insulation pad; Temperature rise rate r T : ; In the formula: C This indicates the heat capacity of the cold-side module (i.e., the pressure block that receives heat from the insulation pad). S This indicates the heat transfer area of ​​the insulation pad; Maximum thermal insulation temperature difference : ; In the formula: Q heat Indicates heating power; Insulation failure time t f : ; In the formula: T cold-0 Indicates the initial temperature of the cold surface. T cold-m This indicates the temperature reached by the cold side when the insulation fails.

[0014] Based on further optimization of the above scheme, the LSTM (Long Short-Term Memory) time-series model training process in step S6 is specifically as follows: With hot surface temperature T hot Real-time pressure P Real-time relative humidity RH The time series sequence is used as input, and the input feature matrix is ​​used as input. G =[ G 1, G 2,…, G N ] T Each sample G i for: ; in, L Indicates the length of the time window; With thermal resistance R eff The time series sequence is used as the output, and the output label is: ; in, M Indicates the prediction step size; The input features and output labels are normalized and then fed into the LSTM model structure for training. The LSTM model controls the flow of information through forget gates, input gates, and output gates to adapt to long-term time dependencies. The model training loss function is: ; In the formula: N represents the number of samples in the training set, and M represents the prediction step size for each sample; Indicates the first i In the sample, the first j The effective thermal insulation resistance value corresponding to each prediction step size; Indicates the first i In the sample, the first j The effective thermal resistance value predicted by the LSTM model for each prediction step size; Indicates the penalty coefficient; E () indicates an indicator function, which takes the value 1 when the condition is met and 0 when the condition is not met; Indicates the first i In the sample, the first j The actual hot surface temperature corresponding to each predicted step size; T ysm This indicates the preset extreme hot surface temperature of the battery runaway.

[0015] Based on further optimization of the above scheme, in step S6, evaluating the performance degradation trend specifically involves: inputting the latest sampling point data into the pre-trained LSTM time series model, and outputting the effective thermal insulation resistance value at future times through the LSTM time series model; obtaining the predicted thermal insulation resistance value by inversely normalizing the effective thermal insulation resistance value output by the LSTM time series model. R eff-yc Ultimately, the performance degradation rate is obtained. : ; In the formula: R eff-0 This indicates the initial measured effective thermal insulation resistance value of the insulation pad.

[0016] The following are the technical effects of the present invention: This invention uses Kalman filtering to fuse multi-dimensional sensor data (including temperature, humidity, pressure, etc.) in real time during the evaluation of the thermal insulation performance of the heat insulation pad, compensating for dynamic errors such as heating lag, pressure fluctuations, and humidity penetration, thereby improving the accuracy of data measurement. At the same time, through a coupling coefficient model, temperature, pressure, and relative humidity are regarded as three coupled fields. The coupling coefficient is used to correct the effective thermal conductivity and effective thermal insulation resistance in real time, thereby avoiding the change in thermal conductivity caused by the interaction of temperature, humidity, and pressure. This makes it more consistent with the actual use environment of battery heat insulation pads affected by multi-dimensional and complex environmental factors, avoiding the problems of large errors and low accuracy caused by single-factor testing. Furthermore, this invention comprehensively evaluates the thermal insulation performance of thermal insulation pads by considering the temperature rise rate, maximum thermal insulation temperature difference, and thermal insulation failure time. The temperature rise rate quantifies the thermal insulation pad's ability to block heat accumulation and facilitates the evaluation of thermal insulation efficiency under dynamic conditions. The maximum thermal insulation temperature difference effectively eliminates interference from test conditions and facilitates the evaluation of thermal insulation limit capabilities under steady-state conditions. The thermal insulation failure time quantifies the safety and environmental protection window, enables wind direction prediction, and facilitates the evaluation of safety risks under extreme conditions. This avoids the problem that static values ​​at a single moment can only reflect instantaneous performance, eliminates the ambiguity of subjective judgment, and improves the accuracy, comprehensiveness, and stability of performance evaluation.

[0017] Ultimately, by using the time-series prediction model of LSTM, the entire life cycle of thermal insulation performance can be predicted, thereby capturing the time-series decay law of thermal insulation performance under coupled environment (such as the irreversible increase in thermal conductivity caused by long-term humidity penetration), avoiding potential safety hazards in advance, ensuring the foresight of performance evaluation, and avoiding problems such as increased cost, long time consumption, and poor stability and reliability caused by full-cycle testing. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the thermal insulation performance testing system in an embodiment of the present invention.

[0019] Figure 2 This is a flowchart of the thermal insulation performance testing method in an embodiment of the present invention.

[0020] Among them, 11, test chamber; 12, drive rod; 13, lifting plate; 141, pressure rod; 142, pressure block; 15, environmental chamber; 161, heating source; 162, temperature distribution plate; 21, servo control system; 22, industrial computer. Detailed Implementation

[0021] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0022] Example 1: A power battery heat insulation pad thermal insulation performance testing system includes a control module and a testing module, such as... Figure 1 As shown, the test module is located on top of the control module and includes a test chamber 11, a drive rod 12, a lifting plate 13, a pressure assembly, an environmental chamber 15, and a heating assembly. The test chamber 11 is located at the end of the control module, and the environmental chamber 15 is installed inside it. The environmental chamber 15 is fixedly installed on the end face of the control module and is coaxially arranged with the test chamber 11 (e.g., Figure 1 As shown, the test chamber 11 and the environmental chamber 15 are equipped with an inner wall insulation layer. The insulation layer can be made of conventional insulation materials in this field, avoiding inaccurate test results due to heat exchange between the external environment and the environmental chamber 15 during the test. Two drive rods 12 are symmetrically arranged on both sides of the environmental chamber 15 (e.g., Figure 1 As shown), a lifting plate 13 is installed between the two drive rods 12 and on the upper side of the environmental chamber 15. A pressure component is installed on the bottom surface of the lifting plate 13, and the end of the pressure component away from the lifting plate is located in the inner cavity of the environmental chamber 15. (Refer to...) Figure 1 As shown: The pressure assembly includes a pressure rod 141 and a pressure block 142. The pressure rod 141 is fixedly installed on the bottom surface of the lifting plate 13, and its end away from the lifting plate 13 passes through the top surface of the environmental chamber 15 (the pressure rod 141 is slidably connected to the top surface of the environmental chamber 15). The pressure block 142 is installed at the bottom end of the pressure rod 141 (i.e., at the end located in the inner cavity of the environmental chamber 15). A heating assembly is installed at the bottom of the environmental chamber 15 and corresponding to the pressure assembly (specifically, corresponding to the pressure block 142). The heating assembly includes a heating source 161 and a heat spreader 162. The heating source 161 is fixedly installed on the end face of the control module, and the heat spreader 162 is installed on its end face (the heating source 161 can be an existing conventional heating device).

[0023] A sensor system is installed inside the test chamber 11. This system includes a hot surface temperature sensor, a cold surface temperature sensor, a piezoelectric pressure sensor, a capacitive humidity sensor, and a laser displacement sensor. The hot surface temperature sensors are evenly distributed on the end face of the heat spreader 162 (used to test the hot surface temperature of the insulation pad). The cold surface temperature sensors are evenly distributed on the end face of the insulation pad under test (i.e., the side away from the heat spreader 162, used to test the cold surface temperature of the insulation pad). The piezoelectric pressure sensors are evenly distributed on the end face of the insulation pad under test, with the piezoelectric pressure sensors and cold surface temperature sensors positioned opposite each other. The capacitive humidity sensor is located inside the environmental chamber 15 to detect the internal humidity. The laser displacement sensor is located on the top surface of the inner cavity of the test chamber 11, with its transmitting end vertically aligned with the end face of the lifting plate 13 (the end face of the lifting plate 13 can be equipped with a laser signal receiver, such as...). Figure 1 As shown, the protruding position in the middle of the top surface of the lifting device 13 is the signal receiving end.

[0024] The lifting platform 13 is driven to move up and down by any one of the following drive sources: servo motor, pneumatic device, or hydraulic device; for example... Figure 1As shown, this embodiment uses servo motor drive, that is, the two drive rods 12 are set as screws, and the top ends of the two screws are rotatably connected to the top surface of the inner cavity of the test chamber 11, and the bottom ends of the two screws are rotatably connected to the top surface of the control module. The two screws pass through the lifting plate 13 and are threaded together. The control module includes a servo control system 21 and an industrial computer 22. The servo control system 21 controls the two screws to rotate simultaneously and in the same direction through the servo motor. The industrial computer 22 is electrically connected to the servo control system 21, the sensor system, and the heating source 161.

[0025] Example 2: As another embodiment of the present invention, a method for testing the thermal insulation performance of a thermal insulation pad, using the testing system shown in Example 1, includes: Step S1, Pre-test preparation: Deploy the sensor system and the heat insulation pad under test in the test chamber as required. The heat insulation pad under test is placed on the end face of the heat spreader (with the side of the heat insulation pad with the cold surface temperature sensor and the piezoelectric pressure sensor facing upwards).

[0026] Step S2, Data Acquisition: Turn on the heating and pressure components, conduct tests, and simultaneously collect the cold / hot surface temperature, contact pressure, ambient humidity, and real-time thickness of the insulation pad; use an FPGA clock synchronization module (synchronizer) to synchronize the data from each sensor, ensuring that the sampling time difference of all sensors is no greater than 1ms; at the same time, the sampling frequency is 10Hz, that is, it is collected once every 0.1s.

[0027] Step S3, Dynamic Error Compensation: Kalman filtering is used to fuse multiple sets of data from the sensor system to obtain the corrected true sensor state value; specifically: First, define the state vector. X k for: ,in, T hot-k express k The hot surface temperature is measured at all times. T cold-k express k The temperature of the cold surface is measured at all times. P k express k The pressure of constant feedback RH k express k The ambient humidity is measured at all times. express k The thickness of the insulation pad at any given time; Input vector U k for: ,in, Q k expressk Heating power at any time F k express k Constant pressure driving force H in-k express k Humidification level at any given time; The system state equation is: ; In the formula: A This represents a 5×5 state transition matrix, used to describe the influence of the previous time step on the current time step; B This represents a 5×3 input matrix (obtained through experimental calibration) used to describe the effect of the input on the state; W k-1 The process noise is represented (Gaussian distribution, variance obtained experimentally). The state transition matrix can be: ; In the formula: Indicates the sampling period (generally no more than 0.1s); The time constant representing the decay of the hot surface temperature. Represents the time constant of heat-cold transfer. The time constant representing the temperature decay of the cold side. The pressure decay time constant, The time constant representing the decay of humidity. The time constant representing the effect of pressure and thickness. The springback time constants, which represent the thickness, were all obtained through experimental calibration. The observation equation is: ; In the formula: H H represents a 5×5 observation matrix (assuming the sensor is in a direct observation state, then H is the identity matrix). V k This represents the observation noise (Gaussian distribution, variance determined by sensor accuracy). The Kalman filter update equation is: Prediction step: ; In the formula: express k The predicted state vector at time t. express k The optimal state vector at time -1 (the final state estimate after combining the observations at time k-1). express k The prediction covariance matrix at time 1, expressk The optimal covariance matrix at time -1 (description) (the degree of uncertainty) O Represents the process noise variance matrix (describing the uncertainty of "unknown disturbances during state transition"); Update steps: ; In the formula: K k express k Kalman gain at time step R This represents the observation noise variance matrix (describing the uncertainty of "sensor's own error"); express k The optimal state vector at time (the final state estimate combining the predicted state and the observed values). express k The optimal covariance matrix at time (description) (the degree of uncertainty) I Represents the identity matrix; The final output is the corrected true state value. By fusing multi-dimensional sensor data in real time through Kalman filtering, dynamic errors such as heating hysteresis, pressure fluctuations, and humidity penetration delays are compensated, thereby effectively reducing measurement errors of core parameters such as temperature and pressure and improving test accuracy.

[0028] Step S4, Coupling Effect Quantification: Substitute the corrected sensor parameters into the coupling coefficient model to obtain the effective thermal conductivity and effective thermal resistance; specifically: Establish a coupling coefficient model of thermal-mechanical-humidity, including the effective thermal conductivity. With effective thermal insulation resistance : ; ; In the formula: k T , k P , k RH , k TP , k TRH , k PRH These represent the coupling coefficients of the temperature field, pressure field, humidity field, temperature-pressure, temperature-humidity, and pressure-humidity fields, respectively (obtained through experimental calibration). Indicates the pressure compressibility coefficient (obtained through experimental calibration); This indicates the difference between the actual hot surface temperature and the standard temperature. T Deviation between 0 and 0: ; Indicates the difference between actual pressure and standard pressure P Deviation between 0 and 0: ; This indicates the difference between actual relative humidity and standard humidity. RH Deviation between 0 and 0: ; The true state value obtained in step S3 Real-time hot surface temperature in the vector T hot Real-time pressure P Real-time relative humidity RH and real-time insulation pad thickness Substituting into the coupling coefficient model, the real-time effective thermal conductivity is obtained. With effective thermal insulation resistance .

[0029] Step S5, Real-time Performance Evaluation: Based on the effective thermal conductivity and effective thermal resistance, evaluate the thermal insulation performance of the insulation pad; by measuring the temperature rise rate. r T Maximum thermal insulation temperature difference With insulation failure time t f A comprehensive evaluation of the thermal insulation performance of the insulation pad; Temperature rise rate r T : ; In the formula: C This indicates the heat capacity of the cold-side module (i.e., the pressure block that receives heat from the insulation pad). S This indicates the heat transfer area of ​​the insulation pad; Maximum thermal insulation temperature difference : ; In the formula: Q heat Indicates heating power; Insulation failure time t f : ; In the formula: T cold-0 Indicates the initial temperature of the cold surface. T cold-m This indicates the temperature reached by the cold side when the insulation fails.

[0030] Step S6, Long-term performance prediction: Based on the pre-trained LSTM (Long Short-Term Memory) time series model, predict the effective thermal insulation resistance of the thermal insulation pad at future times and evaluate the performance degradation trend. The training process for the LSTM (Long Short-Term Memory) time-series model is as follows: With hot surface temperature T hot Real-time pressure P Real-time relative humidity RH The time series sequence is used as input, and the input feature matrix is ​​used as input. G =[ G 1, G 2,…, G N ] T Each sample G i for: ; in, L Indicates the length of the time window (generally) L (50-100) With thermal resistance R eff The time series sequence is used as the output, and the output label is: ; in, M Indicates the prediction step size (e.g.: M =10~50, indicating the predicted thermal insulation resistance value for the next 10~50 sampling points); The input features and output labels are normalized and then fed into the LSTM model structure for training. The LSTM model controls the flow of information through forget gates, input gates, and output gates to adapt to long-term time dependencies. Among them, the Gate of Oblivion: ; In the formula: This represents the Sigmoid activation function; W f This represents the weight matrix input to the forget gate; b f Indicates the forget gate bias; x t This indicates the input at the current time; h t-1 This indicates the hidden state in the previous moment; Input Gate: ; In the formula: W i This represents the input weight matrix, connected to the input at the current time step.x t The weight matrix to the input gate; b i Indicates input gate bias; Candidate state: ; In the formula: tanh() represents the Tanh activation function; W c This represents the weight matrix input to the candidate state; b c Indicates the candidate state bias; Cell status update: ; In the formula: Represents element-wise product; Output gate: ; In the formula: W o This represents the weight matrix from the input to the output gate; b o Indicates the forget gate bias; Output status: ; Predicted output: ; In the formula: W y This represents the output weight matrix; b y Indicates output bias; The model training loss function is: ; In the formula: N represents the number of samples in the training set, and M represents the prediction step size for each sample; Indicates the first i In the sample, the first j The effective thermal insulation resistance value corresponding to each prediction step size; Indicates the first i In the sample, the first j The effective thermal resistance value predicted by the LSTM model for each prediction step size; Indicates the penalty coefficient; E () indicates an indicator function, which takes the value 1 when the condition is met and 0 when the condition is not met; Indicates the first i In the sample, the first j The actual hot surface temperature corresponding to each predicted step size; T ysm This indicates the preset extreme hot surface temperature of the battery runaway.

[0031] The evaluation of performance degradation trend involves inputting the latest sampling point data into the pre-trained LSTM time series model, outputting the effective thermal insulation resistance value at future times through the LSTM time series model, and obtaining the predicted thermal insulation resistance value by inversely normalizing the effective thermal insulation resistance value output by the LSTM time series model. R eff-yc : ; In the formula: R max Indicates the effective thermal insulation resistance value in the training set. The maximum value, R min Indicates the effective thermal insulation resistance value in the training set. The minimum value (e.g., if the effective thermal insulation resistance value in the training set) The range is 0.2–0.8m. 2 ·K / W, then the minimum value is 0.2 m 2 K / W, with a maximum value of 0.8 m 2 (·K / W) This represents the predicted effective thermal insulation resistance value output by the LSTM time series model; The final performance degradation rate is obtained. : ; In the formula: R eff-0 This indicates the initial measured effective thermal insulation resistance value of the insulation pad.

Claims

1. A system for testing the thermal insulation performance of a thermal insulation pad for a power battery, characterized in that: The test system comprises a control module and a test module, the test module is arranged on the upper side of the control module, and comprises a test box body, a driving rod, a lifting plate, a pressure assembly, an environmental box body and a heating assembly.

2. The system of claim 1, wherein: The pressure assembly comprises a pressure rod and a pressure block, the pressure rod is fixedly arranged on the bottom surface of the lifting plate and penetrates through the top surface of the environmental box body at the end away from the lifting plate, and the pressure block is arranged at the bottom end of the pressure rod.

3. The thermal insulation performance testing system for a thermal insulation pad of a power battery according to claim 1 or 2, characterized in that: The test box body is provided with a sensor system, the sensor system comprises a hot surface temperature sensor, a cold surface temperature sensor, a piezoelectric pressure sensor, a capacitive humidity sensor and a laser displacement sensor, the hot surface temperature sensors are uniformly arranged on the end surface of the uniform temperature plate, the cold surface temperature sensors are uniformly arranged on the end surface of the to-be-tested thermal insulation pad, the piezoelectric pressure sensors are uniformly arranged on the end surface of the to-be-tested thermal insulation pad and are arranged at different positions from the cold surface temperature sensors, the capacitive humidity sensor is arranged in the inner cavity of the environmental box body and is used for detecting the internal environmental humidity, and the laser displacement sensor is arranged on the top surface of the inner cavity of the test box body and the emission end thereof is vertically aligned with the end surface of the lifting plate.

4. The thermal insulation performance testing system for a thermal insulation pad of a power battery according to claim 2 or 3, characterized in that: The specific test method of the test system for the thermal insulation performance comprises the following steps. Step S1, preparation before test: the sensor system and the to-be-tested thermal insulation pad are arranged in the test box body according to requirements, wherein the to-be-tested thermal insulation pad is placed on the end surface of the uniform temperature plate; Step S2, data acquisition: the heating assembly and the pressure assembly are turned on, test is performed, and the cold / hot surface temperature, the contact pressure, the environmental humidity and the real-time thickness of the thermal insulation pad are synchronously acquired; Step S3, dynamic error compensation: the multiple groups of data of the sensor system are fused by using Kalman filtering to obtain a corrected real state value of the sensor; Step S4, coupling effect quantification: the corrected sensor parameters are substituted into a coupling coefficient model to obtain an effective thermal conductivity coefficient and an effective thermal insulation resistance value; Step S5, real-time performance evaluation: the thermal insulation performance of the thermal insulation pad is evaluated based on the effective thermal conductivity coefficient and the effective thermal insulation resistance value; Step S6, long-term performance prediction: the effective thermal insulation resistance value of the thermal insulation pad at a future time is predicted based on a pre-trained LSTM time sequence model, and the performance attenuation trend is evaluated.

5. The thermal insulation performance test system for a thermal insulation pad of a power battery according to claim 4, characterized in that: In step S2, an FPGA clock synchronization module is used for synchronous control of various sensing data, so that the time difference of sampling of all sensors is not greater than 1 ms; meanwhile, the sampling frequency is 10 Hz.

6. The thermal insulation performance test system for a thermal insulation pad of a power battery according to claim 4, characterized in that: In step S3, the dynamic error compensation is specifically as follows: First, define the state vector X k is: where T hot-k represents k the hot surface temperature measured at the time instant T cold-k represents k the cold surface temperature measured at the time instant P k represents k the pressure fed back at the time instant RH k represents k the ambient humidity measured at the time instant represents k the insulation pad thickness at the time instant input vector U k is: wherein Q k represents k the heating power at the time point, F k represents k the pressure driving force at the time point, H in-k represents k the humidification amount at the time point; The system state equation is as follows: ； wherein: A represents a 5x5 state transition matrix describing the influence of the previous time instant on the current time instant; B represents a 5x3 input matrix describing the influence of the input on the state; W k-1 represents the process noise; The observation equation is as follows: ； wherein: H represents a 5 x 5 observation matrix; V k represents an observation noise; The Kalman filtering update equation is as follows: Prediction step: ； wherein: represents k the predicted state vector at time k, represents k the optimal state vector at time k-1; represents k the predicted covariance matrix at time k, represents k the optimal covariance matrix at time k-1, O represents a process noise variance matrix; Update step: ； wherein: K k represents k Kalman gain at time t, R represents an observation noise variance matrix; represents k optimal state vector at time t; represents k optimal covariance matrix at time t, I represents an identity matrix; The final output is the corrected real state value Through Kalman filtering, multi-dimensional sensor data is fused in real time to compensate for dynamic errors of heating lag, pressure fluctuation, and humidity penetration delay, and to improve test accuracy.

7. The system of claim 6, wherein: In step S4, the coupling effect quantification is specifically as follows: A thermal-force-humidity coupling coefficient model is established, including effective thermal conductivity coefficient and effective thermal resistance value : ； ； wherein: k T , k P , k RH , k TP , k TRH , k PRH denote the temperature field coupling coefficient, the pressure field coupling coefficient, the humidity field coupling coefficient, the temperature-pressure coupling coefficient, the temperature-humidity coupling coefficient, the pressure-humidity coupling coefficient, respectively; denotes the pressure compressibility coefficient; represents the deviation between the actual hot face temperature and the standard temperature T 0 between the actual hot face temperature and the standard temperature ; represents the deviation between the actual pressure and the standard pressure P 0: ; represents the deviation between the actual relative humidity and the standard humidity RH 0: ; The real state value obtained in step S3 Real-time hot face temperature in the vector T hot Real-time pressure P Real-time relative humidity RH And real-time insulation pad thickness Substitute into the coupling coefficient model to obtain the real-time effective thermal conductivity And the effective insulation resistance value .

8. The thermal insulation performance testing system of the thermal insulation pad of the power battery according to claim 7, characterized in that: In the step S5, the temperature rise rate r T the maximum temperature difference and the insulation failure time t f comprehensively evaluate the insulation performance of the insulation pad; Temperature rise rate r T : ； wherein: C represents the heat capacity of the cold side module, S represents the heat transfer area of the thermal insulation mat; maximum temperature difference : ； In the formulae: Q heat denotes the heating power; Thermal insulation failure time t f : ； where: T cold-0 Tc0represents the cold face initial temperature, T cold-m Tcmaxrepresents the temperature reached by the cold face at the time of heat shield failure.

Images