Energy storage battery cluster thermal runaway intelligent early warning and isolation method and system

Abstract

The present application relates to the field of electrical technology, and discloses a method and system for intelligent early warning and isolation of thermal runaway of energy storage battery clusters, comprising: calculating the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition benchmark, and constructing the thermal runaway feature vector of the energy storage battery cluster; calculating the spatial dispersion of the temperature field gradient of the energy storage battery cluster, and synchronously calculating the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster, to calculate the thermal runaway risk probability value of the energy storage battery cluster; when the thermal runaway risk probability value exceeds the preset threshold value, and the generation rate of characteristic gas and the temperature rise rate in the energy storage battery cluster present positive correlation matching, it is determined that the energy storage battery cluster is a precursor of thermal runaway, and an early warning signal and fault positioning coordinates of the energy storage battery cluster are output; in response to the early warning signal and fault positioning coordinates, a linkage isolation instruction of the energy storage battery cluster is generated to execute intelligent isolation of the energy storage battery cluster. The present application can improve the safety of densely arranged energy storage battery clusters.

Description

TECHNICAL FIELD

[0001] The present application relates to a method and system for intelligent early warning and isolation of thermal runaway of energy storage battery clusters, belonging to the technical field of electricity. BACKGROUND

[0002] The intelligent early warning and isolation of thermal runaway of energy storage battery clusters refers to a technical system that, by fusing multi-dimensional sensing data and advanced algorithms, accurately identifies early weak signs of battery thermal runaway and locates the fault source in real time throughout the life cycle, and then triggers electrical cutoff, chemical suppression, and physical blocking measures to contain disasters; its core significance lies in breaking the lagging limitations of traditional passive firefighting, moving the safety line from post-fire extinguishing to pre-warning and in-process control, effectively preventing the spread of thermal runaway leading to a chain disaster, and building a solid safety barrier for large-scale and highly reliable energy storage applications in new power systems.

[0003] The traditional method of intelligent early warning and isolation of thermal runaway of energy storage battery clusters usually relies on single BMS voltage / temperature threshold alarms or simple smoke detection, and cooperates with general total flooding gas fire extinguishing such as heptafluoropropane for passive response. This method cannot capture the early internal pressure and gas characteristics of thermal runaway, resulting in missing the best intervention opportunity. At the same time, traditional gas fire extinguishing can only extinguish open flames and lacks deep cooling and chemical suppression capabilities, making it difficult to block the continuous heat release reaction inside the battery, thus leading to the spread of fire between densely arranged battery clusters, resulting in high systematic safety risks. SUMMARY

[0004] The present application provides a method and system for intelligent early warning and isolation of thermal runaway of energy storage battery clusters, which aims to improve the safety of densely arranged energy storage battery clusters.

[0005] To achieve the above purpose, the method for intelligent early warning and isolation of thermal runaway of energy storage battery clusters provided by the present application comprises:

[0006] Real-time collection of battery pack external feature data and internal feature data of the energy storage battery cluster by a distributed sensor network deployed in the energy storage battery cluster;

[0007] According to the battery pack external feature data and internal feature data, the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference is calculated, and the internal feature data of the feature gas component concentration is combined to construct the thermal runaway feature vector of the energy storage battery cluster;

[0008] According to the thermal runaway feature vector, the spatial dispersion of the temperature field gradient of the energy storage battery cluster is calculated, and the time cross-correlation coefficient of the internal pressure oscillation and gas spectrum change of the energy storage battery cluster is calculated synchronously to calculate the thermal runaway risk probability value of the energy storage battery cluster;

[0009] determining that the energy storage battery cluster is in a precursor of thermal runaway when the thermal runaway risk probability value exceeds a preset threshold value and a generation rate of a characteristic gas in the energy storage battery cluster is positively correlated with a temperature rise rate, and outputting a warning signal and a fault positioning coordinate of the energy storage battery cluster;

[0010] In response to the warning signal and the fault positioning coordinate, a linkage isolation instruction of the energy storage battery cluster is generated to perform intelligent isolation of the energy storage battery cluster.

[0011] Optionally, the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster is calculated, including:

[0012] A pressure oscillation signal time sequence of the internal pressure oscillation of the energy storage battery cluster and a gas spectrum evolution time sequence of the gas spectrum change are constructed;

[0013] The time axis of the pressure oscillation signal sequence is kept unchanged, the gas spectrum evolution sequence is stepwise translated relative to the pressure oscillation signal sequence within a preset time lag range, a normalized cross-correlation function value between the gas spectrum evolution sequence and the pressure oscillation signal sequence at each time lag step is calculated to generate a cross-correlation function curve of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster;

[0014] A first time lag point of a maximum cross-correlation function value in the cross-correlation function curve and a second time lag point when the cross-correlation function value first exceeds a preset correlation threshold value are defined, and a characteristic time lag interval is determined based on the first time lag point and the second time lag point;

[0015] The time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster in the characteristic time lag interval is calculated.

[0016] Optionally, the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster in the characteristic time lag interval is calculated, including:

[0017] An envelope curve of the cross-correlation function curve of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster in the characteristic time lag interval is extracted;

[0018] A local energy entropy of the corresponding pressure oscillation signal sequence of the energy storage battery cluster is calculated to construct a causal decay kernel function of the characteristic time lag interval;

[0019] A spectral enhancement coefficient of the characteristic time lag interval is calculated;

[0020] According to the envelope curve, the causal attenuation kernel function, and the spectral enhancement coefficient, a time cross-correlation coefficient of internal pressure oscillation and gas spectrum change of the energy storage battery cluster in a characteristic time lag interval is calculated.

[0021] Optionally, according to the thermal runaway feature vector, a spatial dispersion degree of the temperature field gradient of the energy storage battery cluster is calculated, including:

[0022] The thermal runaway feature vector is mapped into a three-dimensional spatial topology structure of the energy storage battery cluster in combination with preset spatial coordinate information of each battery pack of the energy storage battery cluster, so as to construct a spatial mapping distribution matrix representing a thermal runaway risk degree of each spatial position of the energy storage battery cluster.

[0023] Based on the spatial mapping distribution matrix, a feature index difference value between any adjacent nodes in the three-dimensional spatial topology structure is calculated, and the feature index difference value is subjected to spatial weighted processing to obtain a local temperature gradient strength between the adjacent nodes.

[0024] A standard deviation and a coefficient of variation of the local temperature gradient strength are calculated, and the standard deviation and the coefficient of variation are subjected to weighted summation to obtain the spatial dispersion degree of the temperature field gradient of the energy storage battery cluster.

[0025] Optionally, the distributed sensor network adopts a hierarchical heterogeneous architecture, and the distributed sensor network structure is configured to correspond to a three-dimensional spatial layout of the energy storage battery cluster, wherein the distributed sensor network includes:

[0026] An external perception array includes a plurality of NTC temperature sensors attached to geometric centers of battery module housings of the energy storage battery cluster and a voltage sampling circuit arranged at a busbar;

[0027] An internal perception array includes an in-situ gas spectrum sensor implanted into an internal cavity of a battery pack of the energy storage battery cluster through a fiber optic airtight joint, and a high-temperature-resistant micro pressure sensor integrated at an explosion-proof valve interface of the battery pack and having a probe extending into the internal cavity of the battery pack;

[0028] The external perception array and the internal perception array are cascaded through a field bus, and the distributed sensor network is configured with a global synchronous clock module.

[0029] Optionally, according to the external feature data and the internal feature data of the battery pack, an acceleration deviation of the energy storage battery cluster relative to a dynamic working condition reference is calculated, including:

[0030] acquire battery characteristic data of the energy storage battery cluster in a preset historical time window before the current time, combine a current real-time operation parameter sequence of the energy storage battery cluster, and use an adaptive filtering algorithm to construct a dynamic reference curve of the energy storage battery cluster matched with a current working condition;

[0031] subtract the external real-time sampling value of the battery pack external characteristic data and the internal real-time sampling value in the internal characteristic data from reference values at corresponding moments on the dynamic reference curve point by point to acquire an original deviation sequence of the corresponding characteristic indicators of the energy storage battery cluster relative to the dynamic working condition, and perform low-pass filtering processing on the original deviation sequence to obtain a residual sequence of the characteristic indicators;

[0032] perform discrete second-order difference operation on the residual sequence in the time axis to calculate an acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference.

[0033] Optionally, the low-pass filtering processing on the original deviation sequence to obtain the residual sequence of the characteristic indicators comprises: performing low-pass filtering processing on the original deviation sequence at a preset cutoff frequency to filter out white noise caused by electromagnetic interference, while retaining characteristic frequency band signals representing a thermal runaway precursor, to obtain the residual sequence of the characteristic indicators.

[0034] Optionally, the thermal runaway characteristic vector of the energy storage battery cluster is constructed in combination with the characteristic gas component concentration in the internal characteristic data, comprising:

[0035] extracting a time-domain dynamic characteristic component of the acceleration deviation of the energy storage battery cluster;

[0036] extracting a frequency-domain evolution characteristic component of the characteristic gas component concentration;

[0037] splicing the time-domain dynamic characteristic component and the frequency-domain evolution characteristic component to generate the thermal runaway characteristic vector of the energy storage battery cluster.

[0038] Optionally, in response to the early warning signal and the fault positioning coordinates, a linkage isolation instruction of the energy storage battery cluster is generated, comprising:

[0039] in response to the early warning signal and the fault positioning coordinates, an electrical cut-off operation, a chemical inhibition operation and a physical blocking operation of the energy storage battery cluster are generated;

[0040] According to a preset linkage control strategy matrix, trigger priorities and time intervals of the electrical cut-off operation, the chemical inhibition operation and the physical blocking operation are determined to generate the linkage isolation instruction of the energy storage battery cluster.

[0041] In order to solve the above problems, the present application also provides an energy storage battery cluster thermal runaway intelligent early warning and isolation system, which comprises:

[0042] a battery data acquisition module, configured to collect, in real time, external feature data and internal feature data of the battery pack of the energy storage battery cluster through a distributed sensor network deployed in the energy storage battery cluster;

[0043] a thermal runaway feature determination module, configured to calculate an acceleration deviation of the energy storage battery cluster relative to a dynamic working condition benchmark according to the external feature data and the internal feature data of the battery pack, and construct a thermal runaway feature vector of the energy storage battery cluster in combination with a feature gas component concentration in the internal feature data;

[0044] a thermal runaway risk analysis module, configured to calculate a spatial dispersion degree of a temperature field gradient of the energy storage battery cluster according to the thermal runaway feature vector, and synchronously calculate a time cross-correlation coefficient of internal pressure oscillation and gas spectrum change of the energy storage battery cluster, so as to calculate a thermal runaway risk probability value of the energy storage battery cluster;

[0045] a battery cluster early warning module, configured to determine that the energy storage battery cluster is a precursor of thermal runaway when the thermal runaway risk probability value exceeds a preset threshold value, and a generation rate of a feature gas and a temperature rise rate of the energy storage battery cluster present a positive correlation match, and output an early warning signal and a fault positioning coordinate of the energy storage battery cluster;

[0046] a battery cluster isolation module, configured to generate a linkage isolation instruction of the energy storage battery cluster to perform intelligent isolation of the energy storage battery cluster in response to the early warning signal and the fault positioning coordinate.

[0047] The present application realizes real-time high-fidelity collection of external and internal feature data of the battery pack by deploying a hierarchical heterogeneous distributed sensor network, effectively captures the tiny physical and chemical changes in the early stage of thermal runaway, constructs a dynamic working condition benchmark based on an adaptive filtering algorithm and calculates an acceleration deviation, can accurately exclude the interference caused by high-rate charging and discharging and environmental fluctuations, significantly reduces the false positive rate of the existing fixed threshold method, establishes a multi-dimensional space-time correlation analysis model by introducing the spatial dispersion degree of the temperature field gradient and the time cross-correlation coefficient of internal pressure oscillation and gas spectrum change, greatly improves the accuracy and reliability of the calculation of the thermal runaway risk probability value, finally combines the positive correlation matching of the generation rate of the feature gas and the temperature rise rate, accurately identifies the precursor of thermal runaway and outputs the fault positioning coordinate, cooperates with the linkage isolation mechanism of electrical cut-off, chemical inhibition and physical blocking, realizes the closed-loop control from millisecond-level early warning to intelligent blocking, effectively contains the spread of faults, and greatly improves the safe operation level of the energy storage battery cluster. Therefore, the present application can improve the safety of the densely arranged energy storage battery cluster. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1A flowchart of a thermal runaway intelligent early warning and isolation method of a cluster of energy storage batteries according to an embodiment of the present application is shown in FIG.

[0049] Figure 2 A sensor network topology diagram for implementing the thermal runaway intelligent early warning and isolation method of the cluster of energy storage batteries according to an embodiment of the present application is shown in FIG.

[0050] Figure 3 A module diagram for implementing the thermal runaway intelligent early warning and isolation method of the cluster of energy storage batteries according to an embodiment of the present application is shown in FIG.

[0051] Figure 4 A schematic diagram of a computer device for implementing the thermal runaway intelligent early warning and isolation method of the cluster of energy storage batteries according to an embodiment of the present application is shown in FIG.

[0052] The object, functional features and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0053] It should be understood that the specific embodiments described herein are merely intended to explain the present application and are not intended to limit the present application.

[0054] The embodiments of the present application provide a thermal runaway intelligent early warning and isolation method of a cluster of energy storage batteries. The execution subject of the thermal runaway intelligent early warning and isolation method of the cluster of energy storage batteries includes, but is not limited to, at least one of electronic devices such as a server, a terminal, etc., which can be configured to execute the method provided by the embodiments of the present application. In other words, the thermal runaway intelligent early warning and isolation method of the cluster of energy storage batteries can be executed by software or hardware installed in a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster, etc.

[0055] Referring to FIG. Figure 1 A flowchart of a thermal runaway intelligent early warning and isolation method of a cluster of energy storage batteries according to an embodiment of the present application is shown in FIG.

[0056] S1, real-time collection of battery pack external feature data and internal feature data of the cluster of energy storage batteries by a distributed sensor network deployed in the cluster of energy storage batteries.

[0057] The present application can capture the tiny physical and chemical changes occurring in the battery at an early stage of thermal runaway of the cluster of energy storage batteries by real-time collection of battery pack external feature data and internal feature data of the cluster of energy storage batteries by a distributed sensor network deployed in the cluster of energy storage batteries, and provides a high-fidelity data basis for subsequent millisecond-level precursor early warning.

[0058] It needs to be explained that the energy storage battery cluster refers to a group lithium ion battery system installed in a standard container or cabinet, which is electrically connected by a plurality of battery modules through series and parallel connection, and integrates a battery management system, a thermal management system and a fire extinguishing system. The distributed sensor network refers to a multi-physical field perception array arranged in multiple points according to the spatial topology structure inside the battery cluster by using a hierarchical heterogeneous architecture. The external characteristic data of the battery pack refers to physical and chemical quantities reflecting the running state of the battery outside the battery pack shell, at least including the voltage and current of the battery pole or busbar, the surface temperature of the key point of the battery pack shell, the environmental temperature and humidity of the battery pack installation area and the environmental smoke concentration. The internal characteristic data refers to the precursor data reflecting the microscopic physicochemical property changes of the battery cell, at least including the internal pressure and high-frequency oscillation component in the sealed cavity of the battery pack, and the concentration of characteristic gas components generated by electrolyte decomposition.

[0059] It needs to be explained that the distributed sensor network adopts a hierarchical heterogeneous architecture, and the distributed sensor network structure is configured to correspond to the three-dimensional spatial layout of the energy storage battery cluster, including:

[0060] The external perception array includes a plurality of NTC temperature sensors attached to the geometric center of the battery module shell of the energy storage battery cluster and a voltage sampling circuit arranged at the busbar;

[0061] The internal perception array includes an in-situ gas spectrum sensor implanted in the internal cavity of the battery pack of the energy storage battery cluster through a fiber optic airtight joint, and a high-temperature resistant micro pressure sensor integrated at the interface of the explosion-proof valve of the battery pack and the probe extending into the internal cavity of the battery pack;

[0062] The external perception array and the internal perception array are cascaded through a field bus, and the distributed sensor network is configured with a global synchronous clock module.

[0063] The layered heterogeneous architecture refers to the topological organization form of the sensor network, the geometric center of the battery module shell refers to the vertical projection position of the geometric center calculated based on the three-dimensional dimensions of the length, width and height of the battery module on the surface of the battery module shell, the multi-point NTC temperature sensor refers to a temperature collection unit adopting a negative temperature coefficient thermistor as a sensitive element and arranging at least two or more temperature measurement points on different geometric sides or different height layers of the same battery module, the busbar refers to a physical connection area for realizing electrical connection between different battery modules in the energy storage battery cluster, the voltage sampling circuit refers to an electronic circuit module directly collecting potential difference at both ends of the busbar and converting an analog voltage signal into a digital signal for transmission to a battery management unit, the optical fiber airtight joint refers to a transition connecting piece for passing through the sealing shell of the battery pack, the internal cavity of the battery pack refers to a sealed space enclosed by the upper cover, lower box and side plate of the energy storage battery cluster battery pack, the in-situ gas spectrum sensor refers to a spectrum technology sensor based on non-dispersive infrared (NDIR) technology, the battery pack explosion-proof valve interface refers to a mounting flange of an explosion-proof valve for automatically opening to release pressure when the internal pressure of the battery pack exceeds a threshold, and the high-temperature-resistant miniature pressure sensor refers to a capacitive pressure sensitive element manufactured by using a micro-electro-mechanical system (MEMS) technology.

[0064] Referring to Figure 2 As shown in the drawings, the sensor network topology schematic diagram for implementing the energy storage battery cluster thermal runaway intelligent early warning and isolation method provided by an embodiment of the present application, wherein the energy storage battery cluster standard container / cabinet refers to the physical carrier of the whole system, the battery pack A, the battery pack B, the battery pack C and the battery pack D refer to the core energy storage units in the container / cabinet, which are composed of a plurality of batteries in series and parallel connection, T1 and T2 refer to external temperature sensors for obtaining the external thermal characteristics of the energy storage battery cluster battery pack, V represents a voltage sampling point for monitoring the voltage of the energy storage battery cluster battery pack in real time, G represents an internal gas sensor for detecting the characteristic gas generated by the chemical reaction inside the battery of the energy storage battery cluster, P represents an internal pressure sensor for monitoring the pressure change inside the energy storage battery cluster battery pack caused by gas generation, the field bus cascade network is used to connect T1, T2, V, G and P together through a field bus, the global synchronous clock module provides a high-precision time reference for the whole system of the energy storage battery cluster, and the battery management system refers to the control core of the whole system.

[0065] It should be noted that the high-temperature-resistant micro pressure sensor is not only used for collecting static pressure values of the energy storage battery cluster, but also used for capturing internal pressure high-frequency oscillation components representing boiling or severe chemical reaction of electrolyte inside the battery of the energy storage battery cluster. In order to adapt to the bandwidth limitation of the field bus and ensure real-time performance, the internal sensing array is further integrated with a signal preprocessing unit configured with a microprocessor, which is used for performing edge side calculation processing such as time domain to frequency domain conversion, filtering or feature value extraction on the pressure high-frequency oscillation components locally at the sensor, and only uploading the extracted pressure feature data, so as to reduce the transmission bandwidth requirement while retaining the key feature information of the thermal runaway precursor.

[0066] S2, according to the battery pack external feature data and internal feature data, calculating the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference, and combining the feature gas component concentration in the internal feature data, constructing the thermal runaway feature vector of the energy storage battery cluster.

[0067] According to the battery pack external feature data and internal feature data, the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference is calculated, which effectively eliminates the interference caused by normal charging and discharging fluctuations or environmental temperature changes of the energy storage battery cluster, and solves the defect that the fixed threshold method in the prior art is prone to false alarm under large rate charging and discharging working conditions.

[0068] In detail, the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference is calculated according to the battery pack external feature data and internal feature data, including:

[0069] Obtaining the battery feature data of the energy storage battery cluster in a preset historical time window before the current time, combining the current real-time running parameter sequence of the energy storage battery cluster, and using an adaptive filtering algorithm to construct a dynamic reference curve matched with the current working condition of the energy storage battery cluster;

[0070] The external real-time sampling value of the battery pack external feature data and the internal real-time sampling value in the internal feature data are subtracted from the reference value at the corresponding time point on the dynamic reference curve point by point, the original deviation sequence of the corresponding feature index of the energy storage battery cluster relative to the dynamic working condition is obtained, and the original deviation sequence is subjected to low-pass filtering processing to obtain the residual sequence of the feature index;

[0071] The residual sequence is subjected to discrete second-order difference operation on the time axis to calculate the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference.

[0072] The battery characteristic data refers to historical time series data reflecting the state of the energy storage battery cluster collected by a distributed sensor network within a preset historical time window. It includes not only conventional electrical and thermal data such as voltage, current, and temperature, but also historical environmental parameter data. The real-time operating parameter sequence refers to a sequence of influencing factors that changes continuously over time and serves as an input variable for an adaptive filtering algorithm. It includes at least the real-time charge-discharge rate, the ambient temperature of the battery pack, and the real-time state of charge estimated by the battery management system. The adaptive filtering algorithm refers to a signal processing algorithm that can automatically adjust its internal parameter weights to minimize output error. Examples include the recursive least squares (RLS) algorithm or the least mean squares (LMS) algorithm. The dynamic reference curve refers to a real-time fitting curve of the theoretical evolution trajectory of the battery characteristic data under the current operating conditions of the energy storage battery cluster. The external real-time sampling value refers to the instantaneous value of the external characteristic data of the battery pack. The internal real-time sampling value refers to the instantaneous value of the internal characteristic data. The original deviation sequence refers to the initial deviation sequence obtained by numerically subtracting the reference value on the dynamic reference curve from the external real-time sampling value and the internal real-time sampling value. The residual sequence refers to the pure deviation sequence obtained by performing low-pass filtering on the original deviation sequence in a specific frequency band. The discrete second-order difference operation refers to a mathematical operation process of performing two consecutive first-order difference operations on the residual sequence in the discrete time domain. The acceleration deviation refers to a numerical indicator obtained by performing a discrete second-order difference operation on the residual sequence.

[0073] In this scheme, the battery characteristic data of the energy storage battery cluster within a preset historical time window before the current time is obtained, and the real-time operating parameter sequence of the energy storage battery cluster is combined to construct a dynamic reference curve matching the current operating conditions of the energy storage battery cluster using an adaptive filtering algorithm. Specifically, the real-time operating parameter sequence is used as an input signal, and the historical normal operating data in the battery characteristic data is used as a reference signal to construct a transfer function model between the input signal and the battery response. The transfer function model is used to calculate and output the theoretical normal value of the battery characteristic data under the current operating conditions, thereby forming the dynamic reference curve.

[0074] Further, the low-pass filtering of the original deviation sequence to obtain the residual sequence of the characteristic indicator includes low-pass filtering of the original deviation sequence at a preset cutoff frequency to filter out white noise caused by electromagnetic interference while retaining characteristic frequency band signals representing thermal runaway precursors, thereby obtaining the residual sequence of the characteristic indicator.

[0075] The preset cutoff frequency refers to a frequency threshold greater than the upper limit of the characteristic frequency band signal and used to filter out high-frequency electromagnetic noise. The characteristic frequency band signal refers to a signal component that contains physical and chemical changes within the battery and has a frequency lower than the preset cutoff frequency.

[0076] The application combines the concentration of characteristic gas components in the internal feature data to construct the thermal runaway feature vector of the energy storage battery cluster, which converts the non-stationary original data of the energy storage battery cluster into a characteristic index with physical meaning, significantly improves the characterization ability and anti-interference ability of the feature vector to the thermal runaway state, and provides accurate input basis for subsequent risk calculation.

[0077] In detail, the combination of the concentration of characteristic gas components in the internal feature data to construct the thermal runaway feature vector of the energy storage battery cluster comprises:

[0078] extracting the time-domain dynamic feature component of the acceleration deviation of the energy storage battery cluster;

[0079] extracting the frequency-domain evolution feature component of the concentration of characteristic gas components;

[0080] vector splicing the time-domain dynamic feature component and the frequency-domain evolution feature component to generate the thermal runaway feature vector of the energy storage battery cluster.

[0081] The time-domain dynamic feature component refers to a statistical index for characterizing the degree of change and mutation trend of physical quantities over time in the thermal runaway evolution process, the frequency-domain evolution feature component refers to an index for characterizing the distribution rule and evolution characteristics of the concentration of gas components at different frequency scales, and the thermal runaway feature vector refers to a multi-dimensional data structure obtained by arranging and combining the time-domain dynamic feature component of the acceleration deviation and the frequency-domain evolution feature component of the concentration of characteristic gas components after data normalization according to a preset dimension order.

[0082] Optionally, the extraction of the time-domain dynamic feature component of the acceleration deviation of the energy storage battery cluster comprises: statistical analysis of the acceleration deviation sequence in a preset sliding time window, calculation of the root mean square value and the peak factor of the acceleration deviation as the time-domain dynamic feature component for characterizing the time-domain mutation intensity of thermal runaway; wherein the peak factor is the ratio of the maximum absolute value of the acceleration deviation in the preset sliding time window to the root mean square value.

[0083] Optionally, the extraction of the frequency-domain evolution feature component of the concentration of characteristic gas components comprises: fast Fourier transform of the original time sequence of the concentration of characteristic gas components in the internal feature data to obtain frequency spectrum distribution data; calculation of the total energy spectrum density value of the frequency spectrum distribution data in a preset low-frequency frequency band and the offset of the main frequency spectrum peak as the frequency-domain evolution feature component for characterizing the frequency-domain evolution rule of thermal runaway.

[0084] S3, according to the thermal runaway feature vector, calculating the spatial dispersion of the energy storage battery cluster temperature field gradient, and synchronously calculating the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster, to calculate the thermal runaway risk probability value of the energy storage battery cluster.

[0085] According to the thermal runaway feature vector, the application can accurately locate the abnormal hot spot by calculating the spatial dispersion of the energy storage battery cluster temperature field gradient, which eliminates false interference caused by local sensor failure or global temperature rise of the system.

[0086] In detail, the calculation of the spatial dispersion of the energy storage battery cluster temperature field gradient according to the thermal runaway feature vector includes:

[0087] The thermal runaway feature vector is mapped to the three-dimensional spatial topology structure of the energy storage battery cluster in combination with the preset spatial coordinate information of each battery pack of the energy storage battery cluster, to construct a spatial mapping distribution matrix representing the thermal runaway risk degree of each spatial position of the energy storage battery cluster;

[0088] Based on the spatial mapping distribution matrix, the feature index difference value between any adjacent nodes in the three-dimensional spatial topology structure is calculated, and the feature index difference value is subjected to spatial weighting processing to obtain the local temperature gradient strength between the adjacent nodes.

[0089] The standard deviation and the coefficient of variation of the local temperature gradient strength are calculated, and the standard deviation and the coefficient of variation are weighted and summed to obtain the spatial dispersion of the energy storage battery cluster temperature field gradient.

[0090] The preset spatial coordinate information refers to geometric coordinate data for uniquely identifying the relative positions of each battery pack and module in the physical space in the energy storage battery cluster, the three-dimensional space topology refers to a mathematical graph structure capable of reflecting the physical connection relationship, relative position proximity, and heat transfer path between each component unit inside the energy storage battery cluster, each spatial position refers to a discrete sampling point in the three-dimensional space topology corresponding to the preset spatial coordinate information, the spatial mapping distribution matrix refers to a data matrix generated after filling the multi-dimensional feature values in the thermal runaway feature vector into the corresponding nodes of the three-dimensional space topology according to the preset spatial coordinate information, the arbitrary adjacent node refers to a pair of nodes in direct contact in the battery cluster arrangement structure, the feature index difference value refers to the numerical difference between the thermal runaway feature vectors corresponding to any adjacent nodes in the spatial mapping distribution matrix, the local temperature gradient intensity refers to a scalar value obtained by performing spatial weighting processing on the feature index difference value based on the physical distance, the standard deviation refers to a statistical quantity describing the degree of deviation of all local temperature gradient intensities in the energy storage battery cluster from the average value, the coefficient of variation refers to a dimensionless ratio value obtained by dividing the standard deviation by the average value of the local temperature gradient intensity, and the spatial dispersion degree refers to a degree of disorder for quantifying the temperature field distribution of the energy storage battery cluster as a whole.

[0091] The present application calculates the time cross-correlation coefficient of the pressure oscillation and the gas spectrum change in the energy storage battery cluster to calculate the thermal runaway risk probability value of the energy storage battery cluster as another indicator of the risk analysis of the energy storage battery cluster, thereby improving the reliability of the final risk analysis.

[0092] In detail, the calculation of the time cross-correlation coefficient of the pressure oscillation and the gas spectrum change in the energy storage battery cluster includes:

[0093] constructing a pressure oscillation signal time sequence of the pressure oscillation in the energy storage battery cluster and a gas spectrum evolution time sequence of the gas spectrum change;

[0094] keeping the time axis of the pressure oscillation signal sequence unchanged, stepwisely shifting the gas spectrum evolution sequence relative to the pressure oscillation signal sequence within a preset time lag range, calculating the normalized cross-correlation function value between the gas spectrum evolution sequence and the pressure oscillation signal sequence at each time lag step to generate a cross-correlation function curve of the pressure oscillation and the gas spectrum change in the energy storage battery cluster;

[0095] define a first time lag point as a maximum cross-correlation function value in the cross-correlation function curve, and a second time lag point as a time lag point when the cross-correlation function value first exceeds a preset correlation threshold, determine a characteristic time lag interval based on the first time lag point and the second time lag point;

[0096] calculate a time cross-correlation coefficient of internal pressure oscillation and gas spectrum change of the energy storage battery cluster in the characteristic time lag interval.

[0097] wherein, the internal pressure oscillation refers to a pressure fluctuation phenomenon caused by thermal runaway chemical reaction inside the energy storage battery cluster, the pressure oscillation signal time sequence refers to a discrete time data set representing the internal pressure oscillation phenomenon of the energy storage battery cluster, the gas spectrum change refers to a change in the component concentration of a characteristic gas inside the energy storage battery cluster over time, the gas spectrum evolution time sequence refers to time sequence data obtained after quantitative processing of the gas spectrum change, the time lag range refers to a maximum lag time for limiting the translation of the gas spectrum evolution sequence relative to the pressure oscillation signal sequence, the normalized cross-correlation function value refers to a Pearson correlation coefficient after normalization processing of the pressure oscillation signal time sequence and the gas spectrum evolution sequence, the cross-correlation function curve refers to a curve drawn with the lag time as the horizontal coordinate and the normalized cross-correlation function value as the vertical coordinate, the first time lag point refers to a time lag point at which the function value obtains a global maximum value on the cross-correlation function curve, the preset correlation threshold refers to a value for judging whether the correlation between the pressure oscillation signal time sequence and the gas spectrum evolution sequence has statistical significance, the second time lag point refers to a time lag point at which the normalized cross-correlation function value first rises from 0 and exceeds the preset correlation threshold on the cross-correlation function curve, and the characteristic time lag interval refers to a time interval jointly defined by the first time lag point and the second time lag point.

[0098] Further, the calculation of the time cross-correlation coefficient of internal pressure oscillation and gas spectrum change of the energy storage battery cluster in the characteristic time lag interval comprises:

[0099] extracting an envelope curve of the cross-correlation function curve of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster in the characteristic time lag interval;

[0100] calculating a local energy entropy of the pressure oscillation signal sequence corresponding to the energy storage battery cluster to construct a causal decay kernel function of the characteristic time lag interval;

[0101] calculating a spectral enhancement coefficient of the characteristic time lag interval;

[0102] According to the envelope curve, the causal decay kernel function and the spectral enhancement coefficient, a time cross-correlation coefficient of internal pressure oscillation and gas spectrum change of the energy storage battery cluster in a characteristic time lag interval is calculated.

[0103] Further, as another embodiment of the present application, the time cross-correlation coefficient is calculated by the following formula:

[0104]

[0105] wherein, represents the time cross-correlation coefficient of internal pressure oscillation and gas spectrum change of the energy storage battery cluster, represents a hyperbolic tangent activation function, represents a time domain weighting coefficient, represents a length of the characteristic time lag interval, represents a first time lag point of the characteristic time lag interval, represents a second time lag point of the characteristic time lag interval, represents an envelope curve, represents a causal decay kernel function, represents a frequency domain weighting coefficient, represents a spectral enhancement coefficient.

[0106] wherein, the envelope curve refers to a smooth curve extracted from the cross-correlation function curve after Hilbert transform or by signal processing methods such as moving average, low-pass filtering, etc., representing the change trend of the oscillation amplitude with time lag, the local energy entropy refers to the uniformity of the energy distribution of the pressure oscillation signal sequence in time-frequency domain after wavelet transform of the pressure oscillation signal sequence in the characteristic time lag interval, the causal decay kernel function refers to a function about time lag based on the local energy entropy, the spectral enhancement coefficient refers to a measure of the concentration of the energy of the cross-correlation function curve in the frequency domain, the time domain weighting coefficient refers to a contribution weight of the time domain weighted integral term in the final cross-correlation coefficient, which is calibrated according to the actual working condition, the frequency domain weighting coefficient refers to a contribution weight of the spectral enhancement coefficient in the final cross-correlation coefficient, which is also calibrated according to the actual working condition, but it needs to be noted that the time domain weighting coefficient and the frequency domain weighting coefficient add up to 1.

[0107] Optionally, the calculation of the spectral enhancement coefficient of the characteristic time lag interval performs a fast Fourier transform on the cross-correlation function curve to obtain the frequency spectrum energy distribution of the characteristic time lag interval, and calculates the ratio of the main frequency energy and the average energy in the frequency spectrum energy distribution as the spectral enhancement coefficient of the characteristic time lag interval.

[0108] The application calculates the thermal runaway risk probability value of the energy storage battery cluster based on a double-checking mechanism of spatial dispersion and time correlation coefficient, which greatly reduces the false positive rate compared with the existing single parameter superposition decision logic, so that the thermal runaway risk probability value of the energy storage battery cluster is calculated with high reliability. The thermal runaway risk probability value refers to the possibility of the current thermal runaway of the battery cluster. In detail, the thermal runaway risk probability value is obtained by linearly weighting and fusing the spatial dispersion and time correlation coefficient.

[0109] S4, when the thermal runaway risk probability value exceeds the preset threshold value, and the generation rate of characteristic gas in the energy storage battery cluster and the temperature rise rate present positive correlation matching, it is determined that the energy storage battery cluster is a precursor of thermal runaway, and an early warning signal and a fault positioning coordinate of the energy storage battery cluster are output.

[0110] When the thermal runaway risk probability value exceeds the preset threshold value, and the generation rate of characteristic gas in the energy storage battery cluster and the temperature rise rate present positive correlation matching, the application determines that the energy storage battery cluster is a precursor of thermal runaway, and outputs an early warning signal and a fault positioning coordinate of the energy storage battery cluster. The double trigger logic as the determination condition ensures that only when the self-heating chemical reaction inside the energy storage battery cluster is confirmed, the precursor is determined, which greatly improves the accuracy of the warning.

[0111] It should be explained that the preset threshold value refers to the numerical limit for determining whether the energy storage battery cluster has a thermal runaway risk, and the generation rate of characteristic gas in the energy storage battery cluster and the temperature rise rate present positive correlation matching refers to the concentration growth rate of the characteristic gas generated by the chemical reaction inside the energy storage battery cluster and the rising speed of the battery temperature present the same change trend, which represents that the battery inside the energy storage battery cluster is undergoing a violent exothermic chemical reaction, and a large amount of gas is generated and a large amount of heat is released to cause temperature rise. The early warning signal refers to the alarm instruction output after the energy storage battery cluster is determined as a precursor of thermal runaway, wherein the early warning signal includes thermal runaway risk level, fault type, etc. The fault positioning coordinate refers to the spatial position information for identifying the precursor of thermal runaway of the energy storage battery cluster.

[0112] S5, in response to the early warning signal and the fault positioning coordinate, a linkage isolation instruction of the energy storage battery cluster is generated to execute intelligent isolation of the energy storage battery cluster.

[0113] In response to the early warning signal and the fault positioning coordinate, the application generates a linkage isolation instruction of the energy storage battery cluster to execute intelligent isolation of the energy storage battery cluster. According to the fault positioning coordinate, the electrical connection of the energy storage battery cluster fault cluster is cut off, and the linkage isolation through the linkage isolation instruction can effectively block the fault propagation and improve the safety index of the energy storage battery cluster.

[0114] In detail, the response to the early warning signal and the fault positioning coordinates, the linkage isolation instruction of the energy storage battery cluster is generated, including:

[0115] In response to the early warning signal and the fault positioning coordinates, the electrical cut-off operation, the chemical suppression operation and the physical blocking operation of the energy storage battery cluster are generated;

[0116] According to the preset linkage control strategy matrix, the trigger priority and time interval of the electrical cut-off operation, the chemical suppression operation and the physical blocking operation are determined to generate the linkage isolation instruction of the energy storage battery cluster.

[0117] The linkage control strategy matrix refers to a pre-set multi-dimensional logical mapping table, which establishes a corresponding relationship between the thermal runaway risk level, the fault type and the multiple emergency disposal operations in the linkage control strategy matrix, which is used to define the strategy set of the combination mode, the starting condition and the mutual cooperation rule of the electrical cut-off, the chemical suppression and the physical blocking operation means that should be taken under different thermal runaway evolution stages. The trigger priority refers to the execution level weight of the electrical cut-off operation, the chemical suppression operation and the physical blocking operation pre-defined in the linkage control strategy matrix. The time interval refers to the delay parameter between adjacent priority operations set in the linkage control strategy matrix. The linkage isolation instruction refers to the comprehensive control command set containing specific control parameters and execution time sequence.

[0118] It should be noted that the electrical cut-off operation includes analyzing the fault positioning coordinates to determine the connection position of the target battery pack in the energy storage battery cluster that occurs thermal runaway precursor in the electrical topology network; based on the connection position, control instructions are sent to the power electronic switching devices configured at the input and output ends of the target battery pack, and the power electronic switching devices are controlled by the control instructions to jump from the closed state to the open state, so as to realize the electrical isolation of the fault unit and the main loop of the energy storage battery cluster, cut off the short circuit current and continuous charging and discharging energy injection;

[0119] Among them, the electronic switching device refers to a power semiconductor device configured in the electrical main loop of the energy storage battery cluster, which is used to realize the circuit on-off control function.

[0120] The chemical suppression operation includes starting the chemical suppression device integrated in the energy storage battery cluster in response to the early warning signal; according to the representation information of the feature gas component concentration in the thermal runaway feature vector, the required chemical suppression dose of the energy storage battery cluster is calculated, and the chemical suppression device is controlled to direct the preset type and preset flow of chemical suppression agent to the internal cavity of the target battery pack of the energy storage battery cluster, which is used to capture free radicals in the battery internal reaction chain and interrupt the chain heat release reaction.

[0121] The chemical inhibition device refers to an electromechanical integrated system integrated in the energy storage battery cluster, used for storing and accurately releasing chemical inhibitors to the target area of the battery pack in thermal runaway, including a high-pressure resistant storage tank storing chemical inhibitors, an electromagnetic valve group controlled by a linkage isolation instruction, and a spray head extending to the internal cavity of the battery pack of the energy storage battery cluster, the chemical inhibitor amount refers to the mass of the chemical inhibitor required to neutralize the current thermal runaway reaction, and the chemical inhibitor refers to a chemical medium for inhibiting the thermal runaway chemical reaction of the lithium ion battery, such as perfluorohexanone, heptafluoropropane, aerosol generator, etc.

[0122] The physical blocking operation includes controlling the action of a physical blocking mechanism deployed in the energy storage battery cluster according to the fault positioning coordinates, and the physical blocking mechanism action at least includes opening or closing a heat insulation blocking valve between adjacent battery packs of the energy storage battery cluster, triggering the heat absorption expansion of the phase change material to fill the heat dissipation gap between the battery packs, or starting the liquid cooling spray system to deeply cool the affected area of the energy storage battery cluster, to build a physical barrier to prevent heat from transferring to the non-fault area of the energy storage battery cluster.

[0123] The present application realizes real-time high-fidelity acquisition of external and internal feature data of the battery pack by deploying a hierarchical heterogeneous distributed sensor network, effectively captures the tiny physical and chemical changes in the early stage of thermal runaway, builds a dynamic working condition benchmark based on an adaptive filtering algorithm and calculates the acceleration deviation, can accurately exclude the interference caused by high-rate charging and discharging and environmental fluctuations, significantly reduces the false alarm rate of the existing fixed threshold method, establishes a multi-dimensional space-time correlation analysis model by introducing the time cross-correlation coefficient of the spatial dispersion degree of the temperature field gradient, internal pressure oscillation and gas spectrum change, greatly improves the accuracy and reliability of the calculation of the thermal runaway risk probability value, and finally combines the positive correlation matching of the feature gas generation rate and the temperature rise rate Dual decision logic accurately identifies the thermal runaway precursor and outputs the fault positioning coordinates, cooperates with the linkage isolation mechanism of electrical cut-off, chemical inhibition and physical blocking, realizes the closed-loop control from millisecond-level early warning to intelligent blocking, effectively contains the spread of faults, and greatly improves the safe operation level of the energy storage battery cluster. Therefore, the present application can improve the safety of the densely arranged energy storage battery cluster.

[0124] As Figure 3 shown, it is a functional module diagram of the energy storage battery cluster thermal runaway intelligent early warning and isolation system of the present application.

[0125] The energy storage battery cluster thermal runaway intelligent early warning and isolation system 300 can be installed in an electronic device. According to the functions implemented, the energy storage battery cluster thermal runaway intelligent early warning and isolation system can include a battery data acquisition module 301, a thermal runaway feature determination module 302, a thermal runaway risk analysis module 303, a battery cluster early warning module 304, and a battery cluster isolation module 305. The modules in the present application can also be referred to as units, which refer to a series of computer program segments that can be executed by an electronic device processor and can complete a fixed function, which are stored in the memory of the electronic device.

[0126] In the embodiments of the present application, the functions of each module / unit are as follows:

[0127] The battery data acquisition module 301 is configured to collect battery pack external feature data and internal feature data of the energy storage battery cluster in real time through a distributed sensor network deployed in the energy storage battery cluster.

[0128] The thermal runaway feature determination module 302 is configured to calculate the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference according to the battery pack external feature data and internal feature data, and construct a thermal runaway feature vector of the energy storage battery cluster in combination with the feature gas component concentration in the internal feature data.

[0129] The thermal runaway risk analysis module 303 is configured to calculate the spatial dispersion of the temperature field gradient of the energy storage battery cluster according to the thermal runaway feature vector, and simultaneously calculate the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster to calculate the thermal runaway risk probability value of the energy storage battery cluster.

[0130] The battery cluster early warning module 304 is configured to determine that the energy storage battery cluster is a precursor to thermal runaway when the thermal runaway risk probability value exceeds a preset threshold and the generation rate of the feature gas and the temperature rise rate in the energy storage battery cluster present a positive correlation match, and output an early warning signal and a fault positioning coordinate of the energy storage battery cluster.

[0131] The battery cluster isolation module 305 is configured to generate a linkage isolation instruction of the energy storage battery cluster in response to the early warning signal and the fault positioning coordinate to execute intelligent isolation of the energy storage battery cluster.

[0132] In detail, the modules in the energy storage battery cluster thermal runaway intelligent early warning and isolation system 300 in the embodiments of the present application use the same technical means as the energy storage battery cluster thermal runaway intelligent early warning and isolation method described in the above Figure 1 , and can produce the same technical effects, which will not be described here again.

[0133] In one embodiment, a computer device is provided, which may be a server or a client, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile and / or volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface is used for communication with external clients via a network connection. When the computer program is executed by the processor, it implements the functions or steps of a smart early warning and isolation method for thermal runaway of energy storage battery clusters on the server or client side.

[0134] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0135] The distributed sensor network deployed within the energy storage battery cluster collects external and internal characteristic data of the battery pack in real time.

[0136] Based on the external and internal characteristic data of the battery pack, the acceleration deviation of the energy storage battery cluster relative to the dynamic operating condition benchmark is calculated, and the thermal runaway characteristic vector of the energy storage battery cluster is constructed by combining the concentration of characteristic gas components in the internal characteristic data.

[0137] Based on the thermal runaway feature vector, the spatial dispersion of the temperature field gradient of the energy storage battery cluster is calculated, and the time cross-correlation coefficient between the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster is calculated simultaneously to calculate the thermal runaway risk probability value of the energy storage battery cluster.

[0138] When the probability value of thermal runaway exceeds a preset threshold, and the generation rate of the characteristic gas in the energy storage battery cluster is positively correlated with the temperature rise rate, the energy storage battery cluster is determined to be a precursor to thermal runaway, and an early warning signal and fault location coordinates of the energy storage battery cluster are output.

[0139] In response to the warning signal and fault location coordinates, a linkage isolation command for the energy storage battery cluster is generated to execute the intelligent isolation of the energy storage battery cluster.

[0140] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor:

[0141] The distributed sensor network deployed in the energy storage battery cluster collects the battery pack external feature data and internal feature data of the energy storage battery cluster in real time;

[0142] According to the battery pack external feature data and internal feature data, the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference is calculated, and the thermal runaway feature vector of the energy storage battery cluster is constructed in combination with the feature gas component concentration in the internal feature data;

[0143] According to the thermal runaway feature vector, the spatial dispersion of the temperature field gradient of the energy storage battery cluster is calculated, and the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster is calculated synchronously to calculate the thermal runaway risk probability value of the energy storage battery cluster;

[0144] When the thermal runaway risk probability value exceeds the preset threshold value, and the generation rate of the characteristic gas in the energy storage battery cluster and the temperature rise rate present a positive correlation matching, it is determined that the energy storage battery cluster has a thermal runaway precursor, and a warning signal and a fault positioning coordinate of the energy storage battery cluster are output;

[0145] In response to the warning signal and the fault positioning coordinate, a linkage isolation instruction of the energy storage battery cluster is generated to execute intelligent isolation of the energy storage battery cluster.

[0146] It should be noted that the functions or steps that the computer readable storage medium or computer device can achieve above can correspond to the related description of the server side and the client side in the foregoing method embodiments. To avoid repetition, they will not be described one by one here.

[0147] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, storage, database or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0148] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the division of the above-mentioned functional units and modules is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above.

[0149] It is obvious for those skilled in the art that the present application is not limited to the details of the above-mentioned exemplary embodiments, and the present application can be realized in other specific forms without departing from the spirit or essential characteristics of the present application.

[0150] Finally, it should be noted that in the above embodiments, each embodiment can be combined or independent, and deleting any one of them does not affect the technical implementation of the other embodiments. The above embodiments are only used to illustrate the technical solutions of the present application and are not limited. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A method for intelligent early warning and isolation of thermal runaway of energy storage battery clusters, characterized in that, The method comprises: Real-time collection of battery pack external feature data and internal feature data of the energy storage battery cluster through a distributed sensor network deployed in the energy storage battery cluster; According to the battery pack external feature data and internal feature data, the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference is calculated, and the internal feature data of the characteristic gas component concentration is combined to construct a thermal runaway feature vector of the energy storage battery cluster; According to the thermal runaway feature vector, the spatial dispersion of the temperature field gradient of the energy storage battery cluster is calculated, and the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster is calculated synchronously to calculate the thermal runaway risk probability value of the energy storage battery cluster, wherein the calculation of the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster comprises: constructing a pressure oscillation signal time sequence of the internal pressure oscillation and a gas spectrum evolution time sequence of the gas spectrum change, keeping the time axis of the pressure oscillation signal time sequence unchanged, shifting the gas spectrum evolution sequence relative to the pressure oscillation signal time sequence in a preset time lag range step by step, calculating the normalized cross-correlation function value between the gas spectrum evolution sequence and the pressure oscillation signal time sequence at each time lag step to generate a cross-correlation function curve of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster, defining the first time lag point as the maximum cross-correlation function value in the cross-correlation function curve, and the second time lag point as the time lag point when the cross-correlation function value first exceeds the preset correlation threshold, determining the characteristic time lag interval based on the first time lag point and the second time lag point, and calculating the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster in the characteristic time lag interval, wherein the calculation of the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster in the characteristic time lag interval comprises: extracting the envelope curve of the cross-correlation function curve of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster in the characteristic time lag interval, calculating the local energy entropy of the corresponding pressure oscillation signal time sequence of the energy storage battery cluster to construct the causal decay kernel function of the characteristic time lag interval, calculating the spectral enhancement coefficient of the characteristic time lag interval, and calculating the time cross-correlation coefficient of the internal pressure oscillation and the gas spectrum change of the energy storage battery cluster in the characteristic time lag interval according to the envelope curve, the causal decay kernel function and the spectral enhancement coefficient; When the thermal runaway risk probability value exceeds the preset threshold value, and the generation rate of the characteristic gas in the energy storage battery cluster and the temperature rise rate present a positive correlation matching, it is determined that the energy storage battery cluster has a thermal runaway precursor, and a warning signal and a fault positioning coordinate of the energy storage battery cluster are output; In response to the warning signal and the fault positioning coordinate, a linkage isolation instruction of the energy storage battery cluster is generated to execute intelligent isolation of the energy storage battery cluster.

2. The energy storage battery cluster thermal runaway intelligent early warning and isolation method of claim 1, wherein, According to the thermal runaway feature vector, the spatial dispersion of the temperature field gradient of the energy storage battery cluster is calculated, which comprises: The thermal runaway feature vector is mapped to the three-dimensional space topology of the energy storage battery cluster according to preset spatial coordinate information of each battery pack of the energy storage battery cluster, so as to construct a spatial mapping distribution matrix representing the thermal runaway risk degree of each spatial position of the energy storage battery cluster. Based on the spatial mapping distribution matrix, the feature index difference value between any adjacent nodes in the three-dimensional space topology is calculated, and the feature index difference value is subjected to spatial weighting processing to obtain the local temperature gradient strength between the adjacent nodes. The standard deviation and the coefficient of variation of the local temperature gradient strength are calculated, and the standard deviation and the coefficient of variation are subjected to weighted summation to obtain the spatial dispersion degree of the temperature field gradient of the energy storage battery cluster.

3. The energy storage battery thermal runaway intelligent early warning and isolation method of claim 1, wherein, The distributed sensor network adopts a layered heterogeneous architecture, and the distributed sensor network structure is configured to correspond to the three-dimensional spatial layout of the energy storage battery cluster, wherein the distributed sensor network comprises: An external perception array comprising a plurality of NTC temperature sensors attached to the geometric center of the battery module housing of the energy storage battery cluster and a voltage sampling circuit arranged at the busbar; An internal perception array comprising an in-situ gas spectrum sensor implanted in the internal cavity of the battery pack of the energy storage battery cluster through a fiber optic airtight joint, and a high-temperature-resistant micro pressure sensor integrated at the interface of the explosion-proof valve of the battery pack and extending into the internal cavity of the battery pack; The external perception array and the internal perception array are cascaded through a field bus, and the distributed sensor network is configured with a global synchronous clock module.

4. The energy storage battery cluster thermal runaway intelligent early warning and isolation method of claim 1, wherein, According to the external feature data and the internal feature data of the battery pack, the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference is calculated, comprising: Obtaining the battery feature data of the energy storage battery cluster within a preset historical time window before the current time, combining the current real-time running parameter sequence of the energy storage battery cluster, and using an adaptive filtering algorithm to construct a dynamic reference curve matched with the current working condition of the energy storage battery cluster; The external real-time sampling value of the external feature data of the battery pack and the internal real-time sampling value in the internal feature data are subtracted from the reference value at the corresponding time point on the dynamic reference curve point by point to obtain the original deviation sequence of the corresponding feature index of the energy storage battery cluster relative to the dynamic working condition, and the original deviation sequence is subjected to low-pass filtering processing to obtain the residual sequence of the feature index; The residual sequence is subjected to discrete second-order difference operation on the time axis to calculate the acceleration deviation of the energy storage battery cluster relative to the dynamic working condition reference.

5. The energy storage battery thermal runaway intelligent early warning and isolation method of claim 4, wherein, The residual sequence of the feature index is obtained by low-pass filtering processing of the original deviation sequence, comprising: low-pass filtering processing of the original deviation sequence with a preset cutoff frequency to filter out white noise caused by electromagnetic interference, while retaining the feature band signal representing the precursor of thermal runaway, to obtain the residual sequence of the feature index.

6. The energy storage battery thermal runaway intelligent early warning and isolation method of claim 1, wherein, The thermal runaway feature vector of the energy storage battery cluster is constructed in combination with the feature gas component concentration in the internal feature data, comprising: Extracting the time-domain dynamic feature component of the acceleration deviation of the energy storage battery cluster; extracting a frequency domain evolution characteristic component of the concentration of the characteristic gas component; vector splicing the time domain dynamic characteristic component and the frequency domain evolution characteristic component to generate a thermal runaway characteristic vector of the energy storage battery cluster.

7. The energy storage battery thermal runaway intelligent early warning and isolation method of claim 1, wherein, In response to the early warning signal and the fault positioning coordinates, a linkage isolation instruction of the energy storage battery cluster is generated, including: In response to the early warning signal and the fault positioning coordinates, an electrical cut-off operation, a chemical inhibition operation and a physical blocking operation of the energy storage battery cluster are generated; According to a preset linkage control strategy matrix, a trigger priority and a time interval of the electrical cut-off operation, the chemical inhibition operation and the physical blocking operation are determined to generate the linkage isolation instruction of the energy storage battery cluster.

8. A thermal runaway intelligent early warning and isolation system for energy storage battery clusters, characterized in that, The system comprises: a battery data acquisition module configured to collect battery pack external characteristic data and internal characteristic data of the energy storage battery cluster in real time through a distributed sensor network deployed in the energy storage battery cluster; a thermal runaway characteristic determination module configured to calculate an acceleration deviation of the energy storage battery cluster relative to a dynamic working condition reference according to the battery pack external characteristic data and the internal characteristic data, and construct a thermal runaway characteristic vector of the energy storage battery cluster in combination with a concentration of a characteristic gas component in the internal characteristic data. a thermal runaway risk analysis module configured to calculate a spatial dispersion of the temperature field gradient of the energy storage battery cluster according to the thermal runaway feature vector, and simultaneously calculate a time cross-correlation coefficient of pressure oscillation and gas spectrum variation in the energy storage battery cluster to calculate a thermal runaway risk probability value of the energy storage battery cluster, wherein the calculation of the time cross-correlation coefficient of pressure oscillation and gas spectrum variation in the energy storage battery cluster includes: constructing a pressure oscillation signal time sequence of the pressure oscillation in the energy storage battery cluster and a gas spectrum evolution time sequence of the gas spectrum variation, keeping a time axis of the pressure oscillation signal time sequence unchanged, stepwise shifting the gas spectrum evolution sequence relative to the pressure oscillation signal time sequence within a preset time lag range, calculating a normalized cross-correlation function value between the gas spectrum evolution sequence and the pressure oscillation signal time sequence at each time lag step to generate a cross-correlation function curve of the pressure oscillation and the gas spectrum variation in the energy storage battery cluster, defining a first time lag point as a maximum cross-correlation function value in the cross-correlation function curve and a second time lag point as a time lag point at which a cross-correlation function value first exceeds a preset correlation threshold, determining a feature time lag interval based on the first time lag point and the second time lag point, and calculating the time cross-correlation coefficient of the pressure oscillation and the gas spectrum variation in the feature time lag interval of the energy storage battery cluster, wherein the calculation of the time cross-correlation coefficient of the pressure oscillation and the gas spectrum variation in the feature time lag interval of the energy storage battery cluster includes: extracting an envelope curve of the cross-correlation function curve of the pressure oscillation and the gas spectrum variation in the feature time lag interval of the energy storage battery cluster, calculating a local energy entropy of the pressure oscillation signal time sequence of the energy storage battery cluster to construct a causal decay kernel function of the feature time lag interval, calculating a spectral enhancement coefficient of the feature time lag interval, and calculating the time cross-correlation coefficient of the pressure oscillation and the gas spectrum variation in the feature time lag interval of the energy storage battery cluster according to the envelope curve, the causal decay kernel function, and the spectral enhancement coefficient; a battery cluster early warning module configured to determine that the energy storage battery cluster is a precursor of thermal runaway when the thermal runaway risk probability value exceeds a preset threshold and a generation rate of a characteristic gas and a temperature rise rate in the energy storage battery cluster present a positive correlation match, and output an early warning signal and a fault positioning coordinate of the energy storage battery cluster; a battery cluster isolation module configured to generate a linkage isolation instruction of the energy storage battery cluster to perform intelligent isolation of the energy storage battery cluster in response to the early warning signal and the fault positioning coordinate.

Images