A safe battery replacing system applied to a new energy vehicle battery replacing station

Abstract

The application belongs to the technical field of battery replacement monitoring, and provides a safe battery replacement system applied to a new energy vehicle battery replacement station, comprising: a safety perception module, which collects stress concentration coefficients of contact points of a battery replacement work surface, constructs a stress concentration sequence, and identifies stress distribution concentration clusters; a grade division module, which collects electrical and thermal characteristic parameters of the stress distribution concentration clusters, adopts a radial basis function to fit to form a warning surface, and divides a battery replacement risk grade by calculating a boundary crossing ratio; an ability matching module, which constructs a protection vector and compares it with a target characteristic vector to determine whether the protection characteristic and the risk grade are matched; if matched, a contraction analysis module constructs a boundary crossing contraction model based on historical data, inputs a real-time protection vector to obtain a boundary crossing ratio contraction degree, and draws a contraction trajectory on the warning surface. The application realizes risk grading warning and dynamic matching of protection ability in the battery replacement process, and improves the battery replacement safety.

Description

Technical Field

[0001] This invention belongs to the field of battery swapping monitoring technology, specifically a safe battery swapping system for new energy vehicle battery swapping stations. Background Technology

[0002] With the rapid development of the new energy vehicle industry, battery swapping has become an important technological approach to address range anxiety in electric vehicles due to its efficient and convenient energy replenishment advantages. However, the battery swapping process involves high-risk operations such as battery insertion and removal, high-voltage electrical connections, and mechanical transmission, posing multiple safety challenges to battery swapping operations.

[0003] The existing battery swapping station protection system does not differentiate between active and passive protection mechanisms, resulting in excessive resource consumption in low-risk areas and insufficient protection capabilities in high-risk areas due to the uniform strategy. Active and passive protection operate independently, lacking dynamic matching and assessment. Under extreme conditions, passive protection cannot provide a sufficient safety window, easily leading to the spread of disasters. Existing technology lacks a three-dimensional risk trajectory display, making it difficult to pinpoint key causes during accident debriefing and resulting in a lack of targeted corrective measures. Summary of the Invention

[0004] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0005] The technical solution adopted by this invention to solve its technical problem is: a safe battery swapping system for new energy vehicle battery swapping stations, comprising the following modules:

[0006] Safety sensing module: Collects stress concentration coefficients at contact points and constructs stress concentration sequences, establishes dynamic baselines for stress concentration sequences and performs synchronous identification and analysis to obtain stress distribution clusters;

[0007] The risk level classification module collects the electrothermal characteristic parameters of the stress distribution clusters, performs function fitting to form an early warning surface, constructs a three-dimensional input vector for fitting analysis, determines whether the three-dimensional input vector breaks through the boundary of the early warning surface, and if it breaks through the boundary of the early warning surface, calculates the boundary ratio of the early warning surface and classifies the battery swapping risk level.

[0008] Capability matching module: Based on the determined battery swapping risk level, extract the protection features of the battery swapping interface and perform capability matching analysis to determine whether the protection features of the battery swapping interface can match the risk level.

[0009] Contraction Analysis Module: If a match is found, construct an out-of-bounds contraction model, input the current active and passive protection vectors into the out-of-bounds contraction model, obtain the degree of contraction of the out-of-bounds ratio, and draw the out-of-bounds contraction trajectory.

[0010] Furthermore, the method for establishing a dynamic baseline and performing synchronous identification analysis on the stress concentration sequence is as follows:

[0011] Obtain the stress concentration coefficients of each contact point on the battery swapping contact surface at the current moment and at multiple historical moments, and construct a stress concentration sequence;

[0012] An unsupervised learning process is employed to learn the stress concentration sequence at each contact point using a graph convolutional autoencoder, constructing a dynamic baseline model that incorporates temporal and spatial dependencies.

[0013] The dynamic baseline model is trained to reconstruct the stress concentration sequence, and the reconstruction error of the real-time sequence is used as the outlier score.

[0014] Stress distribution clusters are identified based on the anomaly scores of spatially adjacent contact points using a sliding time window.

[0015] Furthermore, the stress concentration factor is obtained as follows:

[0016] Numerical analysis was performed on the contact pressure factors at different contact points on the battery swapping contact surface to obtain the stress concentration factor at the contact point.

[0017] Furthermore, the method for forming the early warning surface through function fitting is as follows:

[0018] Obtain the stress concentration coefficient within the stress distribution cluster of the battery swapping contact surface, as well as the contact resistance of multiple contact points within the stress distribution cluster, and perform evidence consistency verification.

[0019] After obtaining the consistency verification of evidence, the stress concentration factor, contact resistance and temperature gradient between adjacent contact points of the stress concentration clusters on the historical battery swapping contact surface are used to construct a three-dimensional input vector.

[0020] The three-dimensional input vector is fitted with a function to obtain an early warning surface that includes stress concentration factor, resistance and temperature gradient.

[0021] Furthermore, the method for calculating the out-of-bounds ratio of the warning surface is as follows:

[0022] The three-dimensional input vector obtained in real time is input into the fitting function. If the result of the fitting function breaks through the surface boundary, the Euclidean distance between the fitting function output and the surface boundary is calculated to obtain the boundary distance.

[0023] Calculate the Euclidean distance from the center point of the surface to the output of the fitted function to obtain the out-of-bounds distance;

[0024] The outbound distance is calculated by comparing the outbound margin with the outbound length to obtain the outbound ratio.

[0025] Furthermore, the method for performing capability matching analysis is as follows:

[0026] The protection of the battery swapping interface is divided into active protection and passive protection. The active redundancy coefficient and active response coefficient under active protection are obtained and combined to obtain the active protection vector.

[0027] Obtain the thermal resistance coefficient and thermal redundancy coefficient in the passive protection, and combine the thermal resistance coefficient and thermal redundancy coefficient to construct the passive protection vector;

[0028] The feature exceedance calculation is performed on the active protection vector and the passive protection vector to obtain the safety matching value;

[0029] The comparison is performed based on the safety matching value to determine whether the protection features of the battery swapping interface can match the risk level.

[0030] Furthermore, the active response coefficient and active redundancy coefficient are obtained as follows:

[0031] The content of extinguishing agent in the active protection of the battery swapping operation area is obtained, and the average amount of extinguishing agent used under the current risk level is obtained from historical data.

[0032] The active redundancy coefficient is obtained by comparing the average amount of extinguishing agent used under the current risk level with the content of extinguishing agent in active protection.

[0033] The active response coefficient is obtained by comparing the number of successful warnings and the total number of warnings at the current risk level with historical data and performing ratio processing.

[0034] Furthermore, the method for performing feature over-limit calculation is as follows:

[0035] Obtain the target feature vector and calculate the Chebyshev distance between the target feature vector and the active protection vector, which is used as the feature overlimit value between the target feature vector and the active protection vector;

[0036] Calculate the Chebyshev distance between the target feature vector and the passive protection vector, and use it as the feature overlimit value of the target feature vector and the passive protection vector.

[0037] Furthermore, the method for obtaining the degree of contraction of the out-of-bounds ratio is as follows:

[0038] Multiple active and passive protection vectors with different risk levels are obtained from historical data to establish a protection feature matrix;

[0039] Based on the protection feature matrix and the corresponding out-of-bounds ratio, an out-of-bounds contraction model is constructed and trained.

[0040] The real-time active protection vector and passive protection vector are input into the out-of-bounds contraction model, and the out-of-bounds contraction model outputs the predicted out-of-bounds ratio under the current active protection vector and passive protection vector.

[0041] The degree of shrinkage of the out-of-bounds ratio is obtained by calculating the deviation based on the predicted out-of-bounds ratio. Based on the degree of shrinkage of the out-of-bounds ratio, the shrinkage trajectory of the out-of-bounds ratio is drawn on the warning surface.

[0042] Furthermore, the deviation is calculated as follows:

[0043] Obtain the predicted out-of-bounds ratio output by the out-of-bounds contraction model, calculate the deviation ratio with the current out-of-bounds ratio of the battery swapping operation area, and obtain the degree of contraction of the out-of-bounds ratio.

[0044] The beneficial effects of this invention are as follows:

[0045] 1. Collect the stress concentration coefficient of the battery swapping operation surface to construct a dynamic baseline model, realize the spatiotemporal correlation analysis of stress anomalies at the contact interface, identify stress distribution clusters in advance, and warn of risk precursors such as sudden increase in contact resistance and material fatigue. In addition to safety monitoring, use radial basis function to fit the warning surface, quantify the risk level by the out-of-bounds ratio, realize the nonlinear coupling relationship modeling of high-risk scenarios such as thermal runaway, and improve the accuracy of risk prediction.

[0046] 2. The protection system is divided into active protection and passive protection. By constructing active protection vectors and passive protection vectors, differentiated protection capability assessments for different risk levels are achieved. Through characteristic over-limit value calculation and safety matching value determination, the adaptability of protection measures to risk levels is verified in real time, reducing resource waste (such as reducing fire extinguishing agent consumption in low-risk situations) or insufficient capabilities (such as triggering redundant protection activation in high-risk situations), and achieving optimal allocation of protection resources.

[0047] 3. Based on historical data, a boundary shrinkage model is constructed to quantify the suppression effect of the protection vector on the boundary ratio. It supports the prediction of the shrinkage degree of the boundary ratio and the actual boundary ratio, and can provide data support for dynamically adjusting the early warning threshold and optimizing local protection strategies. The risk point shrinkage trajectory is plotted on the early warning surface, which intuitively shows the evolution path of risk indicators (stress, resistance, temperature) to the safe area under protection intervention, assisting operation and maintenance personnel to quickly locate weak links and improve decision-making efficiency. Attached Figure Description

[0048] The invention will now be further described with reference to the accompanying drawings.

[0049] Figure 1 This is a block diagram of a safe battery swapping system for new energy vehicles according to an embodiment of the present invention;

[0050] Figure 2 This is a flowchart for determining whether the protection features of the battery swapping interface described in this embodiment of the invention can match the risk level;

[0051] Figure 3This is a flowchart of a safe battery swapping method for new energy vehicles using a battery swapping station, according to an embodiment of the present invention. Detailed Implementation

[0052] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0053] Example 1

[0054] Please see Figure 1 As shown in the embodiment of the present invention, a safe battery swapping system for new energy vehicles using a battery swapping station includes the following modules:

[0055] Safety sensing module: In the battery swapping operation area, the stress concentration coefficient of the contact point is collected and a stress concentration sequence is constructed. A dynamic baseline is established for the stress concentration sequence and synchronous identification and analysis are performed to obtain the stress distribution concentration clusters;

[0056] The method for collecting the stress concentration factor of the battery swapping work surface is as follows:

[0057] It should be explained that the battery swapping operation area refers to the core operation area for replacing batteries in new energy vehicles, including the vehicle compartment (a closed or semi-closed space where vehicles are parked and batteries are loaded and unloaded) and the battery transportation and installation facility area (such as the area where equipment such as battery swapping robotic arms, battery transportation tracks, and loading and unloading interfaces are located).

[0058] The battery swapping work area is a high-risk area for operations such as battery insertion and removal, electrical connection, and mechanical transmission. It needs to be monitored in real time by intelligent monitoring and early warning systems (such as six-dimensional force sensors and infrared thermal imaging) to monitor mechanical stress (insertion and removal force, torque) and temperature anomalies (precursors of thermal runaway). At the same time, it is equipped with all-scenario sprinkler facilities (such as follow-up sprinklers and targeted sprinkler pipelines) to suppress fires in the early stages. It is the core control area for the safety protection of the battery swapping station and is directly related to the safety of vehicles, batteries and operators during the battery swapping process.

[0059] Preferably, a six-dimensional force sensor is integrated into the end effector of the battery swapping robotic arm on the battery swapping work surface to simultaneously measure the z-axis force at the contact point of the battery swapping contact surface during new energy battery swapping, i.e., F. z (t);

[0060] Through the formula: Obtain the contact pressure factor P at the contact point, where A is the electrode contact area;

[0061] Through the formula: Obtain the stress concentration factor K at contact point i on the battery swapping contact surface. t (i);

[0062] in, Let be the contact pressure at the i-th contact point on the battery swapping contact surface at time t. This represents the average contact pressure factor at all contact points on the contact interface at time t.

[0063] It needs to be explained that the physical meaning of the stress concentration factor is: to measure stress non-uniformity: K t It reflects the degree of stress concentration at the contact interface or in a localized area of ​​a component. When K t When K = 1, it indicates a uniform stress distribution; t A value greater than 1 indicates localized stress concentration, and the larger the value, the more significant the stress concentration. For example, during the insertion and removal process of the battery swapping robotic arm, if K... t A value greater than 1.8 indicates that the local stress at the contact interface is far above the average level.

[0064] Assessing structural risk: The stress concentration factor is a key indicator for judging the safety and reliability of a structure. A high Kt value can accelerate material fatigue, induce microcrack propagation (such as metal fatigue at the interface of a battery swapping device), and even lead to plastic deformation or fracture. Monitoring Kt... t It can provide early warning of abnormalities at the contact interface, avoiding a chain reaction caused by stress concentration leading to a sudden increase in contact resistance and thermal runaway;

[0065] Obtain the stress concentration coefficients of each contact point on the battery swapping contact surface at the current moment and at multiple historical moments, and construct a stress concentration sequence;

[0066] A dynamic baseline was established for the stress concentration sequence and synchronous identification analysis was performed to identify the stress distribution concentration clusters on the battery swapping contact surface;

[0067] The method for establishing a dynamic baseline and performing synchronous identification analysis on the stress concentration sequence is as follows:

[0068] Preferably, a graph convolutional autoencoder (GCAE) is used to perform unsupervised learning on the stress concentration sequence at each contact point to construct a dynamic baseline model that includes temporal dependence and spatial correlation (graph convolution aggregating neighborhood features):

[0069] Under normal operating conditions, a dynamic baseline model is trained to reconstruct the stress concentration sequence, and the reconstruction error of the real-time stress concentration sequence is used as the outlier score.

[0070] Identify clusters of concentrated stress distribution based on the anomaly score of spatially adjacent contact points using a sliding time window;

[0071] It needs to be explained that each contact point on the battery swapping contact surface is modeled as a graph structure, with contact points as nodes and spatial adjacency relationships as edges, to construct spatiotemporal graph data containing the stress concentration sequence of nodes. Then, a graph convolutional autoencoder (GCAE) is designed to aggregate the spatial correlation features of adjacent nodes through a graph convolutional layer (GCN), and combine it with a long short-term memory network (LSTM) or a temporal convolutional layer to capture the temporal dependence of the sequence. The model is trained unsupervised under normal operating conditions. The stress concentration sequence is reconstructed through the autoencoder and the reconstruction error is minimized to establish a dynamic baseline. During real-time monitoring, the stress concentration sequence of the current contact point is input into the trained model, and the reconstruction error between the real-time sequence and the baseline is calculated as an anomaly score. Based on a sliding spatiotemporal window, when the anomaly scores of ≥3 adjacent nodes exceed 3 times the standard deviation of the baseline within 3 consecutive windows, it is identified as a stress distribution concentration cluster, thereby realizing the spatiotemporal correlation analysis of contact point stress anomalies and the construction of a dynamic baseline.

[0072] For example, when the anomaly scores of ≥3 spatially adjacent contact points all exceed 3 times the baseline standard deviation within 3 consecutive sliding time-space windows, the 3 spatially adjacent contact points are identified as stress distribution concentration clusters.

[0073] The risk classification module collects the electrothermal characteristic parameters of the stress distribution clusters on the battery swapping contact surface, performs function fitting to form an early warning surface, constructs a three-dimensional input vector for fitting analysis, determines whether the three-dimensional input vector breaks through the boundary of the early warning surface, and if it does, calculates the boundary crossing ratio of the early warning surface and classifies the battery swapping risk level.

[0074] The method for collecting the electrothermal characteristic parameters of the stress distribution concentration clusters on the battery swapping contact surface is as follows:

[0075] Obtain the stress concentration coefficient within the stress distribution cluster of the battery swapping contact surface, as well as the contact resistance of multiple contact points within the stress distribution cluster, and perform evidence consistency verification to remove invalid data;

[0076] For example, when K t When the value is >2.0 but the contact resistance is below 50 microohms, it violates the physical law that stress concentration is accompanied by an increase in resistance. It is determined to be a sensor data conflict, and cross-calibration of the six-dimensional force sensor and the contact resistance measurement value is performed.

[0077] After obtaining evidence consistency verification, the stress concentration factor, contact resistance, and temperature gradient between adjacent contact points of the historical battery swapping contact surface stress concentration clusters are used as electrothermal characteristic parameters. A function fitting is then used to obtain the stress concentration factor K. t Contact resistance Rc, temperature gradient The warning surface;

[0078] Understandably, by collecting historical battery swapping data to construct a three-dimensional dataset containing stress concentration factor, contact resistance, and temperature gradient, and after denoising and normalizing abnormal samples (such as pre-thermal runaway conditions), the radial basis function (RBF) is used to fit the coupling relationship between parameters. With the runaway critical temperature and the upper limit threshold of resistance safety during historical faults as constraints, an early warning surface with stress, resistance, and temperature gradient as coordinate axes is constructed in three-dimensional space. The surface parameters are optimized through cross-validation, and finally, the comparison between the real-time input parameters and the surface boundary is realized to trigger graded early warning.

[0079] When fitting the radial basis function (RBF), the stress concentration factor, contact resistance, and temperature gradient from historical data are first used as three-dimensional input vectors. Using the fault labels corresponding to the temperature and resistance thresholds at the thermal runaway critical state as output constraints, a Gaussian kernel RBF basis function is selected. ,in The cluster centers are generated, and j is the cluster center number. The weight coefficients w of the basis functions are solved by minimizing the weighted mean square error (e.g., weighted mean square error) that includes historical fault threshold constraints as the loss function. j Construct the fitting function The fitting function is represented in three-dimensional space as a surface formed by a nonlinear combination of basis functions, and the surface boundary is dynamically constrained by the fault threshold.

[0080] The real-time acquired 3D input vector is input into the fitting function, and the output of the fitting function is compared with the surface boundary.

[0081] If the result of the fitting function breaks through the surface boundary, it is considered that there is a battery swapping safety risk at the new energy battery swapping contact surface.

[0082] The Euclidean distance to the nearest surface boundary is calculated based on the output of the fitting function to obtain the boundary distance.

[0083] Calculate the Euclidean distance from the center point of the surface to the output of the fitted function to obtain the out-of-bounds distance;

[0084] The ratio of the out-of-bounds margin to the out-of-bounds length is calculated to obtain the out-of-bounds ratio.

[0085] It is understandable that the physical meaning of constructing the early warning surface is as follows: the early warning surface fits the coupling relationship of risk parameters in historical data through radial basis functions (RBF), and its boundary is dynamically constrained by fault thresholds such as thermal runaway critical temperature and upper limit of resistance safety (e.g., the thermal runaway critical temperature of ternary lithium battery is about 200℃, and the upper limit of contact resistance safety is 50μΩ). The inside of the early warning surface is the safe area, and the outside of the boundary is the risk area. Breaking the boundary means that the parameter combination triggers the coordinated anomaly of "stress-resistance-temperature", which may lead to the occurrence of thermal runaway accident.

[0086] RBF fitting constructs a nonlinear mapping using a Gaussian kernel function, enabling the surface to accurately capture the complex relationships of multi-parameter coupling. Cross-validation optimization ensures the reliability of the boundaries.

[0087] Boundary distance (Euclidean distance from a point to the boundary of a curved surface): reflects the absolute degree of deviation between the current risk state and the safety boundary. The smaller the distance, the closer it is to the critical value of danger.

[0088] Outbound distance (Euclidean distance from a point to the center of the surface): measures the overall deviation of the current risk state in three-dimensional space, reducing misjudgments caused by single-dimensional anomalies (such as when only the temperature gradient is abnormal but the stress and resistance are normal, the distance is small).

[0089] Out-of-boundary ratio: By eliminating dimensional differences through dimensionless conversion, the risk location in three-dimensional space is transformed into a relative risk index in the 0-1 range, which facilitates cross-scenario comparison and level classification (e.g., an out-of-boundary ratio > 0.6 is judged as Level III high risk).

[0090] Based on the magnitude of the out-of-bounds ratio, the battery swapping risk of the battery swapping contact surface is divided into multiple levels, and different levels of battery swapping early warning signals are sent to the system.

[0091] The technical solution of this embodiment is as follows: In the battery swapping operation surface, the stress concentration coefficient of the contact point is collected and a stress concentration sequence is constructed. A dynamic baseline is established for the stress concentration sequence and synchronous identification and analysis are performed to obtain the stress distribution concentration cluster. The electrothermal characteristic parameters of the stress distribution concentration cluster of the battery swapping contact surface are collected and a function is fitted to form an early warning surface. A three-dimensional input vector is constructed for fitting analysis to determine whether the three-dimensional input vector breaks through the boundary of the early warning surface. If it breaks through the boundary of the early warning surface, the boundary crossing ratio of the early warning surface is calculated and the battery swapping risk level is classified. The early warning surface is fitted with a radial basis function, and the risk level is quantified by the boundary crossing ratio to realize the nonlinear coupling relationship modeling of high-risk scenarios such as thermal runaway, thereby improving the accuracy of risk prediction.

[0092] Example 2

[0093] like Figure 1 As shown, a safe battery swapping method using a new energy vehicle battery swapping station includes the following steps:

[0094] Capability matching module: Based on the determined battery swapping risk level, extract the protection features of the battery swapping interface and perform capability matching analysis to determine whether the protection features of the battery swapping interface can match the risk level.

[0095] The method for extracting the protection features of the battery swapping interface is as follows:

[0096] The protection of the battery swapping interface is divided into active protection and passive protection;

[0097] It should be explained that active protection includes: the intelligent monitoring and early warning system described in Example 1, as well as the full-scene sprinkler system and active fire extinguishing agent intervention at the battery swapping interface;

[0098] Passive protection includes: outer thermal insulation materials and new thermal insulation and fireproof structures;

[0099] The protection is divided into active and passive protection, and the mechanism and response logic of the safety protection of the battery swapping interface are classified: active protection uses stress, resistance and temperature sensors to monitor in real time, obtain the risk level and make corresponding dynamic interventions (such as spray response, pressurization system) to actively block the risk evolution chain (such as cooling and extinguishing fire in the early stage of thermal runaway), focusing on pre-event prevention and in-event control.

[0100] Passive protection, through material properties and structural design, passively delays the spread of disasters. For example, low thermal conductivity insulation layers and full coverage of spray pipes between battery racks are used to isolate the thermal runaway of individual batteries. Passive protection focuses on post-disaster suppression and disaster scale control. Active and passive protection form a complementary system of "actively suppressing the occurrence of risks and passively limiting the expansion of risks," which meets the real-time response requirements for high dynamic risks (such as local overheating caused by a sudden increase in contact resistance) during battery swapping, ensures the basic safety baseline under extreme conditions (such as spray system failure), and constructs the key logical division of the "prevention-control-disaster reduction" three-level protection architecture.

[0101] The content of extinguishing agent in the active protection of the battery swapping operation area is obtained, and the average amount of extinguishing agent used under the current risk level is obtained from historical data.

[0102] The active redundancy coefficient is obtained by comparing the average amount of extinguishing agent used under the current risk level with the content of extinguishing agent in active protection.

[0103] The number of successful warnings and the total number of warnings at the current risk level are obtained from historical data, and the ratio is calculated to obtain the proactive response coefficient.

[0104] The active redundancy coefficient and the active response coefficient are combined as active protection features to construct the active protection vector XA;

[0105] Obtain the thermal resistance coefficient and thermal redundancy coefficient in the passive protection, use the thermal resistance coefficient and thermal redundancy coefficient as passive protection features and combine them to construct the passive protection vector XB.

[0106] Those skilled in the art will understand that the thermal resistance coefficient is calculated by measuring the physical parameters of the insulation layer at the battery swapping interface, i.e., the insulation layer thickness divided by the material's thermal conductivity, where the thermal conductivity is obtained through material thermal conductivity testing and the thickness is measured by measuring instruments.

[0107] The thermal redundancy coefficient is the ratio of the actual thermal resistance of the battery swapping interface to the critical thermal resistance under the target risk level. The critical thermal resistance is set according to the risk level, and the redundancy is quantified by comparing the actual thermal resistance with the preset critical value.

[0108] For example, XA=[x1,x2], XB=[x3,x4];

[0109] Where x1-x4 are the active response coefficient, active response coefficient, thermal resistance coefficient, and thermal redundancy coefficient, respectively;

[0110] Obtain the preset target feature vectors YA and YB, YA=[y1,y2], YB=[y3,y4], where the elements in the target feature vectors correspond one-to-one with the active protection vector and the passive protection vector;

[0111] Through the formula: Obtain the feature over-limit value D of the target feature vector and the active protection vector XA. A ;

[0112] Through the formula: Obtain the feature over-limit value D of the target feature vector and the passive protection vector XB. B ;

[0113] The feature exceeds the limit value D A and D B The safe matching value is obtained by summing the results.

[0114] Understandably, the purpose of calculating the safety matching value is to: quantify the deviation of the active protection vector and the passive protection vector from the preset feature vector of the target risk level, systematically evaluate the adaptability of the battery swapping interface protection capability to the risk level, and form a synergistic complement between active protection and passive protection to "suppress the occurrence of risks and limit the expansion of risks", providing a basis for dynamic adjustment of protection strategies (such as resource optimization and maintenance work order generation), while verifying the safety redundancy baseline under extreme operating conditions.

[0115] like Figure 2 If the safety matching value meets the preset safety matching range, it is determined whether the protection features of the battery swapping interface can match the risk level; otherwise, they do not match.

[0116] Contraction Analysis Module: If a match is found, obtain historical active and passive protection vectors and establish a protection feature matrix. Combine the out-of-bounds ratio to construct an out-of-bounds contraction model. Input the current active and passive protection vectors into the out-of-bounds contraction model to obtain the degree of contraction of the out-of-bounds ratio and draw the out-of-bounds contraction trajectory.

[0117] It needs to be explained that the method of drawing the out-of-bounds contraction trajectory on the warning surface based on the degree of contraction of the out-of-bounds ratio is as follows: extract historical and real-time three-dimensional risk data points (corresponding to stress concentration coefficient, contact resistance, and temperature gradient at different times), and calculate the out-of-bounds ratio and degree of contraction of each point after protective intervention.

[0118] The degree of shrinkage of the out-of-bounds ratio is mapped to the distance that the risk point moves to the safe area (inside) of the curved surface in three-dimensional space (e.g., for every 10% shrinkage of the out-of-bounds ratio, the corresponding coordinates of each dimension are adjusted by 5% towards the safe threshold direction). The historical points and the current points of the same risk area are connected by arrow lines to form a shrinkage trajectory.

[0119] By color-coding the degree of contraction (e.g., red indicates high risk with no contraction, and green indicates low risk with significant contraction), key nodes (such as thermal runaway critical threshold points and protective intervention initiation points) are marked on the surface. Finally, the path of risk points contracting from the boundary of the warning surface to the safe area under the action of active and passive protection is dynamically displayed, presenting the suppression effect and evolution trend of protective measures on three-dimensional risk indicators.

[0120] It is understandable that the purpose of calculating the degree of contraction in the out-of-bounds ratio is:

[0121] Function 1: Quantifying protection effectiveness: By comparing the predicted and real-time out-of-bounds ratios, the effectiveness of active and passive protection measures in suppressing risk indicators is assessed, and the degree to which risks converge toward safe areas is measured.

[0122] Function 2: To provide data support for dynamically adjusting the boundary of the early warning surface, optimizing sensor layout or protection resource configuration, and to achieve adaptive adjustment of protection strategies;

[0123] Thirdly, it can draw the contraction trajectory on the early warning surface, intuitively present the migration path of risk points, and help operation and maintenance personnel quickly locate key risk areas and weak links in protection.

[0124] The technical solution of this embodiment is as follows: Based on the determined battery swapping risk level, the protection features of the battery swapping interface are extracted and a capability matching analysis is performed to determine whether the protection features of the battery swapping interface can match the risk level. If they match, historical active and passive protection vectors are obtained and a protection feature matrix is ​​established. An out-of-bounds contraction model is constructed by combining the out-of-bounds ratio. The current active and passive protection vectors are input into the out-of-bounds contraction model to obtain the degree of contraction of the out-of-bounds ratio and to draw the out-of-bounds contraction trajectory. The risk point contraction trajectory is drawn on the early warning surface to intuitively show the evolution path of risk indicators (stress, resistance, temperature) to the safe area under protection intervention, assisting operation and maintenance personnel in quickly locating weak links and improving decision-making efficiency.

[0125] Example 3

[0126] like Figure 3 As shown, a safe battery swapping method using a new energy vehicle battery swapping station includes the following steps:

[0127] S1. In the battery swapping operation area, the stress concentration coefficient of the contact point is collected and a stress concentration sequence is constructed. A dynamic baseline is established for the stress concentration sequence and synchronous identification and analysis are performed to obtain the stress distribution concentration cluster.

[0128] The method for collecting the stress concentration factor of the battery swapping work surface is as follows:

[0129] Preferably, a six-dimensional force sensor is integrated into the end effector of the battery swapping robotic arm on the battery swapping work surface to simultaneously measure the z-axis force at the contact point of the battery swapping contact surface during new energy battery swapping, i.e., F. z (t);

[0130] Through the formula: Obtain the contact pressure factor P at the contact point, where A is the electrode contact area;

[0131] Through the formula: Obtain the stress concentration factor K at contact point i on the battery swapping contact surface. t (i);

[0132] in, Let be the contact pressure at the i-th contact point on the battery swapping contact surface at time t. This represents the average contact pressure factor at all contact points on the contact interface at time t.

[0133] Obtain the stress concentration coefficients of each contact point on the battery swapping contact surface at the current moment and at multiple historical moments, and construct a stress concentration sequence;

[0134] A dynamic baseline was established for the stress concentration sequence and synchronous identification analysis was performed to identify the stress distribution concentration clusters on the battery swapping contact surface;

[0135] The method for establishing a dynamic baseline and performing synchronous identification analysis on the stress concentration sequence is as follows:

[0136] Preferably, a graph convolutional autoencoder (GCAE) is used to perform unsupervised learning on the stress concentration sequence at each contact point to construct a dynamic baseline model that includes temporal dependence and spatial correlation (graph convolution aggregating neighborhood features):

[0137] Under normal operating conditions, a dynamic baseline model is trained to reconstruct the stress concentration sequence, and the reconstruction error of the real-time stress concentration sequence is used as the outlier score.

[0138] Stress distribution clusters are identified based on the anomaly scores of spatially adjacent contact points using a sliding time window.

[0139] S2. Collect the electrothermal characteristic parameters of the stress distribution clusters on the battery swapping contact surface, and perform function fitting to form an early warning surface. Construct a three-dimensional input vector for fitting analysis, and determine whether the three-dimensional input vector breaks through the boundary of the early warning surface. If it breaks through the boundary of the early warning surface, calculate the boundary ratio of the early warning surface and classify the battery swapping risk level.

[0140] The method for collecting the electrothermal characteristic parameters of the stress distribution concentration clusters on the battery swapping contact surface is as follows:

[0141] Obtain the stress concentration coefficient within the stress distribution cluster of the battery swapping contact surface, as well as the contact resistance of multiple contact points within the stress distribution cluster, and perform evidence consistency verification to remove invalid data;

[0142] After obtaining evidence consistency verification, the stress concentration factor, contact resistance, and temperature gradient between adjacent contact points of the historical battery swapping contact surface stress concentration clusters are used as electrothermal characteristic parameters. A function fitting is then used to obtain the stress concentration factor K. t Contact resistance Rc, temperature gradient The warning surface;

[0143] The real-time acquired 3D input vector is input into the fitting function, and the output of the fitting function is compared with the surface boundary.

[0144] If the result of the fitting function breaks through the surface boundary, it is considered that there is a battery swapping safety risk at the new energy battery swapping contact surface.

[0145] The Euclidean distance to the nearest surface boundary is calculated based on the output of the fitting function to obtain the boundary distance.

[0146] Calculate the Euclidean distance from the center point of the surface to the output of the fitted function to obtain the out-of-bounds distance;

[0147] The outbound distance is calculated by comparing the outbound margin with the outbound length to obtain the outbound ratio.

[0148] S3. Based on the determined battery swapping risk level, extract the protection features of the battery swapping interface and perform capability matching analysis to determine whether the protection features of the battery swapping interface can match the risk level.

[0149] The method for extracting the protection features of the battery swapping interface is as follows:

[0150] The protection of the battery swapping interface is divided into active protection and passive protection;

[0151] The content of extinguishing agent in the active protection of the battery swapping operation area is obtained, and the average amount of extinguishing agent used under the current risk level is obtained from historical data.

[0152] The active redundancy coefficient is obtained by comparing the average amount of extinguishing agent used under the current risk level with the content of extinguishing agent in active protection.

[0153] The number of successful warnings and the total number of warnings at the current risk level are obtained from historical data, and the ratio is calculated to obtain the proactive response coefficient.

[0154] The active redundancy coefficient and the active response coefficient are combined as active protection features to construct the active protection vector XA;

[0155] Obtain the thermal resistance coefficient and thermal redundancy coefficient in the passive protection, use the thermal resistance coefficient and thermal redundancy coefficient as passive protection features and combine them to construct the passive protection vector XB.

[0156] The thermal redundancy coefficient is the ratio of the actual thermal resistance of the battery swapping interface to the critical thermal resistance under the target risk level. The critical thermal resistance is set according to the risk level, and the redundancy is quantified by comparing the actual thermal resistance with the preset critical value.

[0157] Where x1-x4 are the active response coefficient, active response coefficient, thermal resistance coefficient, and thermal redundancy coefficient, respectively;

[0158] Obtain the preset target feature vectors YA and YB, YA=[y1,y2], YB=[y3,y4], where the elements in the target feature vectors correspond one-to-one with the active protection vector and the passive protection vector;

[0159] Through the formula: Obtain the feature over-limit value D of the target feature vector and the active protection vector XA. A ;

[0160] Through the formula: Obtain the feature over-limit value D of the target feature vector and the passive protection vector XB. B ;

[0161] The feature exceeds the limit value D A and D BThe safe matching value is obtained by summing the results.

[0162] If the safety matching value meets the preset safety matching range, it is determined whether the protection features of the battery swapping interface can match the risk level; otherwise, they do not match.

[0163] S4. If a match is found, obtain the historical active and passive protection vectors and establish a protection feature matrix. Combine the out-of-bounds ratio to construct an out-of-bounds contraction model. Input the current active and passive protection vectors into the out-of-bounds contraction model to obtain the degree of contraction of the out-of-bounds ratio and draw the out-of-bounds contraction trajectory.

[0164] The method for establishing the protection feature matrix is ​​as follows:

[0165] Multiple active and passive protection vectors with different risk levels are obtained from historical data to establish a protection feature matrix;

[0166] Based on the protection feature matrix and the corresponding out-of-bounds ratio, an out-of-bounds contraction model is constructed and trained.

[0167] The real-time active protection vector and passive protection vector are input into the out-of-bounds contraction model, and the out-of-bounds contraction model outputs the predicted out-of-bounds ratio under the current active protection vector and passive protection vector.

[0168] Obtain the predicted out-of-bounds ratio output by the out-of-bounds contraction model, calculate the deviation ratio with the current out-of-bounds ratio of the battery swapping operation area, and obtain the degree of contraction of the out-of-bounds ratio.

[0169] Based on the degree of contraction of the out-of-bounds ratio, the contraction trajectory of the out-of-bounds area is plotted on the warning surface.

[0170] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A safe battery swapping system for new energy vehicles using battery swapping stations, characterized in that: Includes the following modules: Safety sensing module: Collects stress concentration coefficients at contact points and constructs stress concentration sequences, establishes dynamic baselines for stress concentration sequences and performs synchronous identification and analysis to obtain stress distribution clusters; The method for establishing a dynamic baseline and performing synchronous identification and analysis on the stress concentration sequence is as follows: Obtain the stress concentration coefficients of each contact point on the battery swapping contact surface at the current moment and at multiple historical moments, and construct a stress concentration sequence; An unsupervised learning process is employed to learn the stress concentration sequence at each contact point using a graph convolutional autoencoder, constructing a dynamic baseline model that incorporates temporal and spatial dependencies. The dynamic baseline model is trained to reconstruct the stress concentration sequence, and the reconstruction error of the real-time sequence is used as the outlier score. Identify stress distribution clusters based on the anomaly score of spatially adjacent contact points using a sliding time window; The stress concentration factor is obtained as follows: Numerical analysis was performed on the contact pressure factors at different contact points on the battery swapping contact surface to obtain the stress concentration factor at the contact point; The risk level classification module collects the electrothermal characteristic parameters of the stress distribution clusters, performs function fitting to form an early warning surface, constructs a three-dimensional input vector for fitting analysis, determines whether the three-dimensional input vector breaks through the boundary of the early warning surface, and if it breaks through the boundary of the early warning surface, calculates the boundary ratio of the early warning surface and classifies the battery swapping risk level. Among them, the electrothermal characteristic parameters include: the stress concentration factor of the stress concentration cluster of the historical battery swapping contact surface, the contact resistance, and the temperature gradient between adjacent contact points as electrothermal characteristic parameters. The method for calculating the out-of-bounds ratio of the warning surface is as follows: The three-dimensional input vector obtained in real time is input into the fitting function. If the result of the fitting function breaks through the surface boundary, the Euclidean distance between the fitting function output and the surface boundary is calculated to obtain the boundary distance. Calculate the Euclidean distance from the center point of the surface to the output of the fitted function to obtain the out-of-bounds distance; The ratio of the out-of-bounds margin to the out-of-bounds length is calculated to obtain the out-of-bounds ratio. Capability matching module: Based on the determined battery swapping risk level, extract the protection features of the battery swapping interface and perform capability matching analysis to determine whether the protection features of the battery swapping interface can match the risk level. The method for performing capability matching analysis is as follows: The protection of the battery swapping interface is divided into active protection and passive protection. The active redundancy coefficient and active response coefficient under active protection are obtained and combined to obtain the active protection vector. Obtain the thermal resistance coefficient and thermal redundancy coefficient in the passive protection, and combine the thermal resistance coefficient and thermal redundancy coefficient to construct the passive protection vector; The feature exceedance calculation is performed on the active protection vector and the passive protection vector to obtain the safety matching value; The comparison is performed based on the safety matching value to determine whether the protection features of the battery swapping interface can match the risk level. Contraction Analysis Module: If a match is found, construct an out-of-bounds contraction model, input the current active and passive protection vectors into the out-of-bounds contraction model, obtain the degree of contraction of the out-of-bounds ratio, and draw the out-of-bounds contraction trajectory.

2. The safe battery swapping system for new energy vehicles using a battery swapping station according to claim 1, characterized in that: The method for forming the early warning surface through function fitting is as follows: Obtain the stress concentration coefficient within the stress distribution cluster of the battery swapping contact surface, as well as the contact resistance of multiple contact points within the stress distribution cluster, and perform evidence consistency verification. After obtaining the consistency verification of evidence, the stress concentration factor, contact resistance and temperature gradient between adjacent contact points of the stress concentration clusters on the historical battery swapping contact surface are used to construct a three-dimensional input vector. The three-dimensional input vector is fitted with a function to obtain an early warning surface that includes stress concentration factor, resistance and temperature gradient.

3. A safe battery swapping system for new energy vehicles using a battery swapping station as described in claim 1, characterized in that: The active response coefficient and active redundancy coefficient are obtained as follows: The content of extinguishing agent in the active protection of the battery swapping operation area is obtained, and the average amount of extinguishing agent used under the current risk level is obtained from historical data. The active redundancy coefficient is obtained by comparing the average amount of extinguishing agent used under the current risk level with the content of extinguishing agent in active protection. The active response coefficient is obtained by comparing the number of successful warnings and the total number of warnings at the current risk level with historical data and performing ratio processing.

4. A safe battery swapping system for new energy vehicles using a battery swapping station as described in claim 1, characterized in that: The method for performing feature over-limit calculation is as follows: Obtain the target feature vector and calculate the Chebyshev distance between the target feature vector and the active protection vector, which is used as the feature overlimit value between the target feature vector and the active protection vector; Calculate the Chebyshev distance between the target feature vector and the passive protection vector, and use it as the feature overlimit value of the target feature vector and the passive protection vector.

5. A safe battery swapping system for new energy vehicles using a battery swapping station as described in claim 1, characterized in that: The method for obtaining the degree of contraction of the out-of-bounds ratio is as follows: Multiple active and passive protection vectors with different risk levels are obtained from historical data to establish a protection feature matrix; Based on the protection feature matrix and the corresponding out-of-bounds ratio, an out-of-bounds contraction model is constructed and trained. The real-time active protection vector and passive protection vector are input into the out-of-bounds contraction model, and the out-of-bounds contraction model outputs the predicted out-of-bounds ratio under the current active protection vector and passive protection vector. The degree of shrinkage of the out-of-bounds ratio is obtained by calculating the deviation based on the predicted out-of-bounds ratio. Based on the degree of shrinkage of the out-of-bounds ratio, the shrinkage trajectory of the out-of-bounds ratio is drawn on the warning surface.

6. A safe battery swapping system for new energy vehicles using a battery swapping station as described in claim 5, characterized in that: The method for calculating the deviation is as follows: Obtain the predicted outbound ratio output by the outbound shrinkage model, calculate the deviation ratio with the current outbound ratio of the battery swapping operation area, and obtain the degree of shrinkage of the outbound ratio.

Images