Power battery track control method and system

Abstract

The invention relates to a power battery track control method and system in the technical field of vehicle safety, and the method comprises the following steps: constructing a track prediction initial model, training the track prediction initial model through a training set, and obtaining a track prediction training model; collecting an inertial parameter II, an electrochemical parameter and an environmental parameter of the power battery of the trajectory to be predicted, and performing denoising processing to obtain a preprocessed inertial parameter; inputting the preprocessed inertial parameters into a trajectory prediction training model to obtain predicted trajectory parameters, and outputting a damping braking scheme based on the predicted trajectory parameters; the moving track of the power battery is adjusted based on the damping braking scheme, and the problem that after an existing power battery is thrown away from the vehicle after thermal runaway, the sliding direction and sliding displacement of the power battery cannot be controlled, the final stop position is too close to the vehicle, and the vehicle safety cannot be guaranteed is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle safety, and particularly relates to a method and a system for controlling the trajectory of a power battery. Background Art

[0002] When a thermal runaway occurs in an electric vehicle in the prior art, a monitoring device detects a risk of thermal runaway in the power battery, and then the vehicle immediately controls the whole vehicle to power off and cuts off the electrical connection between the power battery and the vehicle. At this time, a fusing mechanism is activated, and a fusing bolt is melted by heating, so that the power battery is detached from the vehicle and falls to the ground.

[0003] Meanwhile, the system enables a super capacitor to be used as a temporary driving power source for the vehicle. Due to the fast charge and discharge characteristics of the super capacitor, it can provide power for the vehicle in a short time, enabling the vehicle to automatically or under the operation of the driver to drive to a safe distance away from the power battery with thermal runaway. If the power of the super capacitor is insufficient, the system will control the power battery to be quickly charged before detachment according to factors such as the slope of the vehicle where it is located, ensuring that the super capacitor can drive the vehicle to travel to a preset safe distance.

[0004] In addition, after detecting the risk of thermal runaway, the warning mechanism of the vehicle will remind the driver to stay away from the vehicle through a display or a speaker, further ensuring personal safety. The entire system monitors the states of the power battery and the super capacitor in real time through the monitoring device, ensuring that emergency avoidance operations can be quickly and effectively executed in an emergency, reducing the risk of the vehicle being ignited by the power battery with thermal runaway, and ensuring the life and property safety of the driver.

[0005] However, after the power battery is thrown away, due to inertia, the battery often continues to slide forward. If the stopping position of the battery is too close to the electric vehicle, there will still be a safety hazard. Summary of the Invention

[0006] Aiming at the deficiencies in the prior art, the present invention provides a method and a system for controlling the trajectory of a power battery, which solves the problem that after the power battery is thrown away from the vehicle due to thermal runaway in the prior art, the sliding direction and sliding displacement of the power battery cannot be controlled, resulting in the final stopping position being too close to the vehicle and still unable to ensure the safety of the vehicle.

[0007] To solve the above technical problems, the present invention is solved by the following technical solutions:

[0008] A method for controlling the trajectory of a power battery includes the following steps:

[0009] Construct an initial trajectory prediction model and train the initial trajectory prediction model with a training set to obtain a trained trajectory prediction model, where the training set is a set formed by the inertial parameters of a power battery and the actual trajectory parameters having a mapping relationship with the inertial parameters;

[0010] Collect the inertial parameters II, electrochemical parameters, and environmental parameters of the power battery of the trajectory to be predicted, and perform denoising processing to obtain preprocessed inertial parameters;

[0011] Input the preprocessed inertial parameters into the trained trajectory prediction model to obtain predicted trajectory parameters, and output a damping braking scheme based on the predicted trajectory parameters;

[0012] Adjust the operating trajectory of the power battery after detaching from the vehicle based on the damping braking scheme.

[0013] Optionally, constructing an initial trajectory prediction model and training the initial trajectory prediction model with a training set includes the following steps:

[0014] Initialize the state vector of the inertial parameters I, where the inertial parameters I include the three-axis acceleration, displacement, and angular velocity of the power battery;

[0015] Input the initialized state vector of the inertial parameters I into the initial trajectory prediction model to obtain the trajectory position parameters at the next moment;

[0016] Calculate the residual value between the current inertial parameters I and the predicted trajectory position parameters at the next moment, and dynamically update the feedback gain coefficient and the state vector based on the residual value;

[0017] Apply a tangent constraint to the three-axis acceleration for non-linear compensation;

[0018] Set a prediction completion condition. When the prediction result meets the prediction condition, end the prediction and output the trajectory position parameters.

[0019] Optionally, the formula for obtaining the trajectory position parameters at the next moment is:

[0020] x pred =x k +V k Δt+0.5a k Δt 2 , where X pred represents the predicted displacement value, X k represents the displacement of the power battery at time k, V k represents the speed of the power battery at time k; a k represents the acceleration of the power battery at time k; Δt represents the time difference.

[0021] Optionally, the update formula for dynamically updating the feedback gain coefficient based on the residual value is:

[0022] α = 0.6e -0.1t ; ; γ = 0.1ln(1 + t), where α represents the exponential decay coefficient in the feedback gain coefficient, β represents the coefficient matching the vibration frequency in the feedback gain coefficient, γ represents the logarithmic growth coefficient in the feedback gain coefficient, and t represents time.

[0023] Optionally, the update formula for dynamically updating the state vector based on the residual value is:

[0024] , where represents the predicted displacement at time k + 1; represents the predicted displacement value; residual represents the residual value; represents the updated velocity at time k + 1; represents the velocity of the power battery at time k; represents the updated acceleration at time k + 1; represents the acceleration of the power battery at time k.

[0025] Optionally, the prediction completion conditions include a time condition and a convergence condition, and the time condition and the convergence condition are satisfied simultaneously;

[0026] wherein, the time condition is: when the cumulative prediction duration is greater than or equal to the preset time value, the time condition is satisfied;

[0027] The convergence condition is: when the absolute value of the residual value for n consecutive times is less than the set value, the convergence condition is satisfied.

[0028] Optionally, outputting a damping braking scheme based on the predicted trajectory parameters includes the following steps:

[0029] Determine the three-dimensional acceleration, trajectory curvature, and absolute position coordinates of the power battery according to the predicted trajectory parameters;

[0030] Based on the three-dimensional acceleration, trajectory curvature, and absolute position coordinates, determine the damping pushing direction and distribute the damping force, and combine the sliding speed of the power battery to determine the form of damping braking as the air explosion braking damping method or the composite braking damping method;

[0031] Perform displacement compensation through preprocessing chemical parameters and preprocessing environmental parameters.

[0032] Optionally, the formula for distributing the damping force according to the three-dimensional acceleration direction is:

[0033] , where Fx Denotes the damping force in the first direction x, F y Denotes the damping force in the second direction y, F z Denotes the damping force in the third direction z, where the first direction, the second direction, and the third direction are perpendicularly arranged to each other Denotes the acceleration in the x direction Denotes the acceleration in the y direction Denotes the acceleration in the z direction

[0034] Optionally, before constructing the initial trajectory prediction model, the following steps are further included:

[0035] Perform real-time monitoring on the power battery in thermal runaway and collect thermal runaway data

[0036] Based on the thermal runaway data, determine whether to perform the detachment operation of the power battery. If so, unload and detach the power battery from the vehicle through the intelligent detachment mechanism

[0037] A power battery trajectory control system, the power battery trajectory control system executes the power battery trajectory control method described in any one of the above, and includes a multimodal thermal runaway detection unit, an intelligent detachment mechanism, a motion trajectory monitoring unit, a composite damping actuator, and a control center unit

[0038] The multimodal thermal runaway detection unit is used to perform real-time monitoring on the power battery after thermal runaway and collect thermal runaway data

[0039] The intelligent detachment mechanism is used to determine whether to perform the detachment operation of the power battery based on the thermal runaway data. If so, unload and detach the power battery from the vehicle through the intelligent detachment mechanism

[0040] The motion trajectory detection unit is used to collect the inertial parameters II, electrochemical parameters, and environmental parameters of the power battery after it detaches from the vehicle

[0041] The composite damping actuator is used to output a damping braking scheme according to the predicted trajectory parameters

[0042] Adopting the technical solution provided by the present invention, compared with the prior art, it has the following beneficial effects:

[0043] Through the trajectory prediction of the power battery after it detaches from the vehicle, the running trajectory of the power battery is positioned, which serves as the basis for adjusting the subsequent damping braking. At the same time, by means of damping braking of the power battery, the final stopping position of the power battery after it detaches from the vehicle is adjusted, thus avoiding the problem that the sliding direction and sliding displacement of the power battery cannot be controlled due to inertia, which affects the safety of the vehicle Description of the Drawings

[0044] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0045] Figure 1 Flowchart of a power battery trajectory control method proposed in Embodiment 1;

[0046] Figure 2 Training flowchart of the trajectory prediction training model proposed in Embodiment 1. Detailed implementation manners

[0047] The following further elaborates on the present invention in conjunction with embodiments. The following embodiments are explanations of the present invention, and the present invention is not limited to the following embodiments.

[0048] Embodiment 1

[0049] As Figure 1 shown, a power battery trajectory control method includes the following steps: constructing an initial trajectory prediction model, and training the initial trajectory prediction model through a training set to obtain a trajectory prediction training model, where the training set is a set formed by the inertial parameter 1 of the power battery and the actual trajectory parameters having a mapping relationship with the inertial parameter 1.

[0050] As Figure 2 shown, constructing an initial trajectory prediction model and training the initial trajectory prediction model through a training set includes the following steps: initializing the state vector of the inertial parameter 1, where the inertial parameter 1 includes the three-axis acceleration, displacement, and angular velocity of the power battery; inputting the initialized state vector of the inertial parameter 1 into the initial trajectory prediction model to obtain the trajectory position parameters at the next moment; calculating the residual value between the current inertial parameter 1 and the predicted trajectory position parameters at the next moment for model feedback correction, and dynamically updating the feedback gain coefficient and the state vector based on the residual value; applying a tangent constraint to the three-axis acceleration for non-linear compensation; setting a prediction completion condition, and when the prediction result meets the prediction condition, ending the prediction and outputting the trajectory position parameters.

[0051] Specifically, first define x0 = [x init , v init , a init T , representing the initial position x init , speed v init , acceleration a init ​, when the system is powered on, the position is initialized through the first frame of sensor data (such as displacement gauge readings), the speed and acceleration are set to zero, and then the state vector of the initialized inertial parameter one is input into the initial model of trajectory prediction. The formula for the trajectory position parameter at the next moment is: x pred =x k +v k Δt + 0.5a k Δt 2 , where, x pred represents the predicted displacement value, x k represents the displacement of the power battery at time k, V k represents the speed of the power battery at time k; a k represents the acceleration of the power battery at time k; Δt represents the time difference, thereby calculating the trajectory position parameter at the next moment.

[0052] Next, update the dynamic update feedback gain coefficient and the state vector. Among them, the update formula for dynamically updating the feedback gain coefficient based on the residual value is: α = 0.6e -0.1t ; β = 0.3(1 - cos t); γ = 0.1ln(1 + t), where, α represents the exponential decay coefficient in the feedback gain coefficient, β represents the coefficient matching the vibration frequency in the feedback gain coefficient, γ represents the logarithmic growth coefficient in the feedback gain coefficient, and t represents time.

[0053] Among them, it should be noted that the updated α is used to suppress the accumulation of historical errors, and the predicted displacement x pred is corrected by weighting the residual with α. The coefficient decays exponentially to reduce the impact of timely errors; the updated β is used to enhance the tracking of periodic motion. It contains a cosine term and matches the periodic speed fluctuations caused by mechanical vibrations (frequency ω) when the power battery detaches; the updated γ is used to improve the acceleration stability. It grows with the logarithmic function and enhances the response to sudden acceleration changes (such as ground collisions). Compared with the traditional fixed feedback gain coefficient, the present invention uses a dynamic update method to establish the model, thereby improving the accuracy of model prediction.

[0054] On the other hand, the update formula for dynamically updating the state vector based on the residual value is:

[0055] , where, represents the predicted displacement at time k + 1; represents the predicted displacement value; residual represents the residual value, residual = z measured −x pred , z measured represents inertial parameter one; Represents the updated speed at time k+1; Represents the speed of the power battery at time k; Represents the updated acceleration at time k+1; Represents the acceleration of the power battery at time k.

[0056] After completing the dynamic update, apply a hyperbolic tangent constraint to the acceleration term for non-linear compensation to limit mutations: a compensated =tanh(a pred / 10)×10, thereby limiting the acceleration within ±10m / s² to prevent the predicted value from diverging due to sensor noise.

[0057] Finally, perform a judgment on the completion of the prediction. The prediction completion conditions include a time condition and a convergence condition, and the time condition and the convergence condition are satisfied simultaneously; among them, the time condition is: when the cumulative prediction duration is greater than or equal to the preset time value, the time condition is satisfied; the convergence condition is: when the absolute value of the residual difference for n consecutive times is less than the set value, the convergence condition is satisfied. It should be noted that the preset time value and n can both be set and modified according to the actual situation and are not specifically limited in this embodiment.

[0058] After obtaining the prediction training model, trajectory control can be carried out. Specifically, collect the inertial parameters, electrochemical parameters, and environmental parameters of the power battery of the trajectory to be predicted, and perform denoising processing through wavelet transform to obtain the preprocessed inertial parameters.

[0059] Among them, the electrochemical parameters refer to the real-time internal resistance (accuracy 0.1mΩ) and polarization voltage (sampling rate 1kHz), and the environmental parameters refer to the vibration spectrum.

[0060] Next, input the preprocessed inertial parameters into the trajectory prediction training model to obtain the predicted trajectory parameters, and output a damping braking scheme based on the predicted trajectory parameters. Specifically, output a damping braking scheme based on the predicted trajectory parameters, including the following steps: determine the three-dimensional acceleration, trajectory curvature, and absolute position coordinates of the power battery according to the predicted trajectory parameters; determine the damping pushing direction and distribute the damping force based on the three-dimensional acceleration, trajectory curvature, and absolute position coordinates, and determine the form of the damping braking as the air explosion braking damping method or the composite braking damping method in combination with the sliding speed of the power battery.

[0061] Among them, the formula for distributing the damping force according to the three-dimensional acceleration direction is:

[0062] , where F x Represents the damping force in the first direction x, F y Represents the damping force in the second direction y, F zrepresents the damping force in the third direction z, where the first direction, the second direction, and the third direction are perpendicular to each other. represents the acceleration at the prediction moment in the x direction. represents the acceleration at the prediction moment in the y direction. represents the acceleration at the prediction moment in the z direction.

[0063] For example, when the speed of the power battery obtained is greater than 5 m / s, air explosion damping braking is adopted. The air explosion damping will establish a 50 kN damping force within 0.5 ms (nitrogen pressure 35 MPa), and the energy consumption efficiency is to convert 85% of the impact kinetic energy into heat and gas work. When the operating speed of the power battery is in the medium-speed mode (0 - 5 m / s), a hybrid braking mode is adopted. At this time, the total braking damping force is:

[0064] F total = 0.7F eddy + 0.3F pneu where F eddy is the electromagnetic damping force, and F pneu is the air explosion damping force. The meaning of the formula is that the total braking force is composed of a combination of electromagnetic damping force (70%) and air explosion damping force (30%).

[0065] Among them, the electromagnetic composite braking is dynamically adjusted by the PWM duty cycle (20 - 100 kHz), and the response linearity can be improved to 0.98 (0.82 for traditional electromagnetic braking).

[0066] Thus, based on the damping braking scheme, the running trajectory of the power battery after detaching from the vehicle is adjusted.

[0067] Finally, this embodiment also provides fault diagnosis, and its architecture is as follows: input layer, 21-dimensional feature vector (including 8th-order IMF components after EMD decomposition); hidden layer structure, 5-layer restricted Boltzmann machine (RBM) stacked (1024 - 512 - 256 - 128 - 64). Compared with the prior art, DropConnect regularization (inactivation rate 0.3) and AdaBound optimizer are introduced.

[0068] The diagnosis steps are as follows: First, data input and preprocessing, input 21-dimensional feature vector, including: original sensor signals, such as real-time data of acceleration, displacement, current, pressure, etc.; then perform EMD decomposition components, 8th-order intrinsic mode functions (IMFs) extracted by empirical mode decomposition (EMD) to characterize the multi-scale vibration characteristics of the signal; then perform statistical features, including statistics of time-domain indicators such as mean, variance, peak-to-peak value, kurtosis, etc.

[0069] Second, deep feature extraction (hidden layer processing), the restricted Boltzmann machine (RBM) is trained layer by layer. The first layer of RBM (1024 nodes): learns the underlying features (such as vibration frequency components) from the input data; the second layer of RBM (512 nodes): extracts higher-order features (such as the time-frequency distribution of impact events); the third layer of RBM (256 nodes): captures fault-related feature patterns (such as abnormal current harmonics); the fourth layer of RBM (128 nodes): further abstracts features (such as the dynamic pressure change of air path leakage); the fifth layer of RBM (64 nodes): generates the final feature encoding for classification decision-making.

[0070] Improved regularization: DropConnect: randomly disconnects 30% of the neuron connections during training (for example, about 307 connections fail in the 1024-node layer) to prevent the model from overfitting to specific noise; feature retention mechanism: the remaining 70% of the connections force the model to learn robust features.

[0071] Third, fine-tuning and classification output. The fine-tuning layer (a 64-node fully connected layer) uses the SELU activation function (SELU(x)=λ⋅max(0,x)+min(0,αex−α)) to solve the vanishing gradient problem; outputs a 64-dimensional compressed feature vector to represent the potential patterns of faults; for classification decision-making, it is mapped to 3 types of fault outputs (damper jamming, electromagnetic coil short circuit, air path leakage) through the Softmax layer.

[0072] It should be noted that the object of fault diagnosis in this embodiment is: fault recognition rate: 98.7% (89.2% for the traditional BP network); diagnostic response time: <80ms (required by ISO13849-1 PLd level); supported fault types include but are not limited to: damper jamming (alarm when the displacement anomaly is 0.1mm), electromagnetic coil short circuit (alarm when the current ripple > 30%), air path leakage (trigger when the pressure drop rate > 5MPa / s).

[0073] In addition, before constructing the initial model for trajectory prediction, the following steps are also included: real-time monitoring of the thermal runaway power battery and collecting thermal runaway data; determining whether to perform the detachment operation of the power battery based on the thermal runaway data, and if so, unloading and detaching the power battery from the vehicle through the intelligent detachment mechanism.

[0074] Embodiment 2

[0075] A power battery trajectory control system, the power battery trajectory control system executes the power battery trajectory control method shown in Embodiment 1, including a multi-modal thermal runaway detection unit, an intelligent detachment mechanism, a motion trajectory monitoring unit, a composite damping actuator, and a control center unit.

[0076] The multi-modal thermal runaway detection unit is used to monitor the power battery in real time after thermal runaway and collect thermal runaway data;

[0077] The intelligent disengagement mechanism is used to determine whether to perform the disengagement operation of the power battery based on the thermal runaway data. If so, the power battery is unloaded and disengaged from the vehicle through the intelligent disengagement mechanism;

[0078] The motion trajectory detection unit is used to collect the inertial parameters, electrochemical parameters and environmental parameters of the power battery after it is disengaged from the vehicle;

[0079] The composite damping actuator is used to output a damping braking scheme according to the predicted trajectory parameters.

[0080] Specifically, the multi-modal thermal runaway detection unit is responsible for real-time monitoring of the thermal runaway state of the battery pack. It adopts a variety of sensor fusion technologies to ensure high-precision and fast-response detection. Specifically, it includes: Fiber Bragg Grating sensors, which are used to accurately measure the internal temperature change of the battery pack, with a detection accuracy of ±0.5°C and a response time of less than 50 milliseconds; an array of pressure sensors, which monitors the internal pressure change of the battery pack, with a range of 0-10 MPa and a resolution of 0.1 kPa; MEMS gas sensors, which detect the internal gas concentration of the battery pack, including hydrogen (H2) and carbon monoxide (CO), with detection limits of 0.1% and 50 ppm respectively; a data fusion processor, which uses the Kalman filtering algorithm to process the data collected by the sensors in real time, eliminate signal interference, and improve the detection accuracy.

[0081] The intelligent disengagement mechanism is designed to quickly disengage the battery pack from the vehicle when a thermal runaway risk is detected to ensure safety. Its main components include a two-stage unlocking device: primary unlocking, which uses a shape memory alloy trigger mechanism with a response time of less than 200 milliseconds to ensure quick start; secondary unlocking, which uses a servo motor to drive a wedge mechanism to provide a disengagement force of ≥5 kN to ensure reliable disengagement of the battery pack; anti-rebound design, which automatically unfolds a graphene buffer layer after disengagement to prevent the battery pack from rebounding during disengagement and ensure the smoothness and safety of the disengagement action.

[0082] The motion trajectory detection unit is used to monitor the motion state of the battery pack in real time after disengagement to ensure precise intervention of the damping control system, including: a three-axis MEMS accelerometer, which monitors the acceleration of the battery pack in three axes, with a range of ±50 g and a bandwidth of 1 kHz; a gyroscope, which measures the angular velocity of the battery pack, with a range of ±2000° / s; a UWB positioning module, which provides high-precision positioning with a positioning accuracy of ±10 cm; a data preprocessing unit, which uses wavelet transform technology to denoise the collected data and improve the reliability of the data.

[0083] The composite damping actuator effectively suppresses the inertial motion of the battery pack through pneumatic and electromagnetic damping mechanisms to ensure its rapid stability. It includes: a vector air explosion damping device; an annular array nozzle, with 8 nozzles that can be independently controlled to provide damping forces in multiple directions; a high-pressure gas tank with a gas storage pressure of 35 MPa to ensure that the nozzles can generate sufficient thrust; a piezoelectric ceramic valve with an opening time less than 5 milliseconds to achieve rapid response; an electromagnetic eddy current damping module, a Halbach permanent magnet array with a surface magnetic flux density reaching 1.2 T to provide a strong magnetic field; a copper brake disc with a thickness of 8 mm and a diameter of 300 mm for generating electromagnetic damping force; a power semiconductor controller that uses PWM control technology with a frequency of 20 kHz to achieve precise damping force adjustment.

[0084] The control center unit serves as the control core of the entire system and is responsible for data processing, risk assessment, and the execution of control strategies. It includes: a main control unit that uses a dual-core heterogeneous processor (ARM Cortex-M7 + FPGA) to ensure efficient data processing and real-time control; control algorithms: motion trajectory prediction: based on an improved α-β-γ filtering algorithm to predict the motion trajectory of the battery pack; damping force distribution: using a fuzzy PID controller to dynamically adjust the damping force to ensure the best control effect; fault diagnosis: using a deep belief network (DBN) model to monitor the system status in real time and detect and handle faults in a timely manner; a data recorder: storing time series data such as temperature, pressure, and motion parameters for subsequent analysis and optimization.

[0085] As described above, it is only a preferred embodiment of the present invention and does not impose any formal or substantial limitations on the present invention. It should be noted that for ordinary technical personnel in the technical field, without departing from the method of the present invention, several improvements and supplements can still be made, and these improvements and supplements should also be regarded as the protection scope of the present invention. For those skilled in the art, without departing from the spirit and scope of the present invention, any equivalent changes made by slightly modifying and evolving the above-disclosed technical content are equivalent embodiments of the present invention; at the same time, any equivalent changes made by modifying and evolving the above embodiments based on the substantial technology of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A power battery trajectory control method, characterized in that, It includes the following steps: Construct an initial trajectory prediction model, and train the initial trajectory prediction model through a training set to obtain a trajectory prediction training model. The training set is a set formed by the inertial parameters of a power battery and the actual trajectory parameters having a mapping relationship with the inertial parameters; Collect the inertial parameters II, electrochemical parameters, and environmental parameters of the power battery of the trajectory to be predicted, and perform denoising processing to obtain preprocessed inertial parameters; Input the preprocessed inertial parameters into the trajectory prediction training model to obtain predicted trajectory parameters, and output a damping braking scheme based on the predicted trajectory parameters; Adjust the running trajectory of the power battery after it detaches from the vehicle based on the damping braking scheme.

2. A power battery trajectory control method according to claim 1, characterized in that, Constructing an initial trajectory prediction model and training the initial trajectory prediction model through a training set includes the following steps: Initialize the state vector of the inertial parameters I, where the inertial parameters I include the three-axis acceleration, displacement, and angular velocity of the power battery; Input the initialized state vector of the inertial parameters I into the initial trajectory prediction model to obtain the trajectory position parameters at the next moment; Calculate the residual value between the current inertial parameters I and the predicted trajectory position parameters at the next moment, and dynamically update the feedback gain coefficient and the state vector based on the residual value; Apply a tangent constraint to the three-axis acceleration for non-linear compensation; Set the prediction completion condition. When the prediction result meets the prediction condition, end the prediction and output the trajectory position parameters.

3. The method for controlling the trajectory of a power battery according to claim 2, wherein, The formula for obtaining the trajectory position parameters at the next moment is: X pred = X k + V k Δt + 0.5a k Δt 2 , where X pred represents the predicted displacement value, X k represents the displacement of the power battery at time k, V k represents the velocity of the power battery at time k; a k represents the acceleration of the power battery at time k; Δt represents the time difference.

4. A power battery trajectory control method according to claim 2, characterized in that, The update formula for dynamically updating the feedback gain coefficient based on the residual value is: α = 0.6e -0.1t ; β = 0.3(1 - cos t); γ = 0.1ln(1 + t), where α represents the exponential decay coefficient in the feedback gain coefficient, β represents the coefficient for matching the vibration frequency in the feedback gain coefficient, γ represents the logarithmic growth coefficient in the feedback gain coefficient, and t represents time.

5. A power battery trajectory control method according to claim 4, characterized in that The update formula for dynamically updating the state vector based on the residual value is: , where represents the predicted displacement at time k + 1; represents the predicted displacement value; residual represents the residual value; represents the updated velocity at time k + 1; represents the velocity of the power battery at time k; represents the updated acceleration at time k + 1; represents the acceleration of the power battery at time k.

6. A power battery trajectory control method according to claim 2, characterized in that The prediction completion condition includes a time condition and a convergence condition, and the time condition and the convergence condition are satisfied simultaneously; Among them, the time condition is: when the cumulative prediction duration is greater than or equal to a preset time value, the time condition is satisfied; The convergence condition is: when the absolute value of the residual value for n consecutive times is less than a set value, the convergence condition is satisfied.

7. A power battery trajectory control method according to claim 1, characterized in that Outputting a damping braking scheme based on the predicted trajectory parameters includes the following steps: Determine the three-dimensional acceleration, trajectory curvature, and absolute position coordinates of the power battery according to the predicted trajectory parameters; Determine the damping push direction and distribute the damping force based on the three-dimensional acceleration, trajectory curvature, and absolute position coordinates, and combine the sliding speed of the power battery to determine the form of the damping braking as the air explosion braking damping method or the composite braking damping method; Perform displacement compensation through the preprocessed chemical parameters and preprocessed environmental parameters.

8. A power battery trajectory control method according to claim 7, characterized in that The formula for distributing the damping force according to the three-dimensional acceleration direction is: , where F x represents the damping force in the first direction x, F y represents the damping force in the second direction y, F z represents the damping force in the third direction z, where the first direction, the second direction, and the third direction are perpendicular to each other, represents the acceleration in the x direction, represents the acceleration in the y direction, represents the acceleration in the z direction.

9. A method for controlling the trajectory of a power battery according to claim 1, characterized in that, Before constructing the initial trajectory prediction model, it further includes the following steps: Monitor the power battery with thermal runaway in real time and collect thermal runaway data; Determine whether to perform the detachment operation of the power battery based on the thermal runaway data. If so, unload and detach the power battery from the vehicle through an intelligent detachment mechanism.

10. A power battery trajectory control system, characterized in that, The power battery trajectory control system executes the power battery trajectory control method according to any one of claims 1-9, and includes a multi-modal thermal runaway detection unit, an intelligent disconnection mechanism, a motion trajectory monitoring unit, a composite damping actuator, and a control center unit; The multi-modal thermal runaway detection unit is used to monitor the power battery after thermal runaway in real time and collect thermal runaway data; The intelligent disconnection mechanism is used to determine whether to perform a disconnection operation on the power battery based on the thermal runaway data. If so, the power battery is unloaded and disconnected from the vehicle through the intelligent disconnection mechanism; The motion trajectory detection unit is used to collect the inertial parameters II, electrochemical parameters, and environmental parameters of the power battery after it is disconnected from the vehicle; The composite damping actuator is used to output a damping braking scheme according to the predicted trajectory parameters.

Images