A method and device for early warning of thermal runaway in an energy storage system, and a storage medium.

Abstract

This invention discloses a method, device, and storage medium for early warning of thermal runaway in energy storage systems, relating to the field of energy storage system monitoring technology. The method first employs a scheme where the control input signal changes over time within a preset signal range according to a preset signal variation rule, obtaining output signals corresponding to different input signals, thus acquiring richer parameter information. Subsequently, by constructing an output signal sequence, the aforementioned rich parameter information is simulated as image information, facilitating data analysis and processing. Finally, based on Hough transform processing of the output signal sequence, the characteristic structure in the parameter information can be efficiently and accurately identified. Based on this characteristic structure, the probability of thermal runaway is analyzed, thereby improving the accuracy of early warning of thermal runaway in energy storage systems.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of energy storage system monitoring, and particularly relates to an energy storage system thermal runaway early warning method and device and a storage medium. BACKGROUND

[0002] The energy storage system is used for storing electric energy. For an electrochemical energy storage system, it mainly includes the following types: lithium ion batteries, sodium-sulfur batteries, flow batteries (such as all-vanadium flow batteries), and lead-acid batteries. The above energy storage systems store a large amount of energy and also carry a huge risk. Thermal runaway is a major risk faced by the above energy storage systems.

[0003] Thermal runaway refers to a chain reaction phenomenon caused by various incentives. A large amount of heat and harmful gas emitted by thermal runaway can cause the energy storage system to catch fire and explode. For an energy storage system, the energy storage system parameters include gas parameters, temperature parameters, pressure parameters, sound parameters, and optical signal parameters. By analyzing the above energy storage system parameters, a certain degree of energy storage thermal runaway early warning can be achieved.

[0004] In the prior art, the energy storage system parameters of the energy storage system are mainly monitored from the following aspects to achieve energy storage system thermal runaway early warning.

[0005] Gas monitoring: The characteristic gas released in the early stage of battery thermal runaway is detected by hydrogen, carbon monoxide, and other gas sensors. For example, the hydrogen concentration mutation rate can be used as a core early warning factor.

[0006] Internal temperature monitoring: The internal temperature of the battery cell is obtained in real time by using an optical fiber sensor or a dynamic impedance phase shift method. For example, the optical fiber sensor is pre-buried in the internal battery cell, and early warning is achieved by the mutation of the temperature rise rate.

[0007] Acoustic and optical signal identification: The acoustic signal characteristics (such as frequency and amplitude) and optical characteristics (such as the light refraction change of electrolyte leakage) of the battery exhaust are analyzed to achieve energy storage system thermal runaway early warning.

[0008] However, practice shows that the above methods do not fully use the energy storage system parameters of the energy storage system, and the internal information of the energy storage system parameters is not fully tapped, resulting in insufficient accuracy of the energy storage system thermal runaway early warning. Therefore, it is necessary to propose a new energy storage system thermal runaway early warning method to improve the accuracy of the energy storage system thermal runaway early warning. SUMMARY

[0009] The present application provides an energy storage system thermal runaway early warning method and device and a storage medium for improving the accuracy of the energy storage system thermal runaway early warning.

[0010] To solve the above technical problems, the present application discloses a thermal runaway early warning method of an energy storage system, the method is used for monitoring the energy storage system by an energy storage system parameter sensor, the parameter sensor is used for changing the output signal of the input signal after the parameter sensor in response to the change of the energy storage system parameter, the method comprises:

[0011] Step one, for each of the plurality of preset continuous target time, taking the target time as the starting point, controlling the input signal to change in the preset signal range with the preset signal change rule along with time, and obtaining the output signal corresponding to the different input signals to obtain the output signal set; wherein the signal change rule corresponding to each target time is the same;

[0012] Step two, based on the output signal set obtained by all the target time, an output signal sequence is constructed, each row of the output signal sequence corresponds to a different output signal in a target time, and each column of the output signal sequence corresponds to the output signal of different target time arranged in time sequence;

[0013] Step three, mapping the output signal sequence to the Hough space, obtaining the feature structure corresponding to the signal sequence based on the Hough transform, and determining the probability of the thermal runaway phenomenon of the energy storage system according to the feature structure.

[0014] As an optional implementation, in the first aspect of the present application, for each of the plurality of preset continuous target time, taking the target time as the starting point, controlling the input signal to change in the preset signal range with the preset signal change rule along with time, and obtaining the output signal corresponding to the different input signals to obtain the output signal set, comprising:

[0015] For each of the plurality of preset continuous target time, taking the target time as the starting point, determining the time after each preset time difference as a node time, controlling the input signal to increase in the preset signal range along with the change of the node time, so that each node time corresponds to a different input signal; and obtaining the output signal corresponding to the input signal of each node time to obtain the output signal set corresponding to the target time; wherein the output signal set comprises the output signal of each node time corresponding to the target time.

[0016] As an optional implementation, in the first aspect of the present application, the output signal sequence is mapped to the Hough space, and the feature structure corresponding to the signal sequence is obtained based on the Hough transform, comprising:

[0017] The signal sequence is regarded as analog image data mapping to a Hough space, and a feature pattern is extracted from the analog image data based on a Hough transform;

[0018] Furthermore, the determining the probability of the thermal runaway of the energy storage system according to the feature structure comprises:

[0019] The feature pattern is matched with a plurality of target patterns stored in advance, a target pattern matched with the feature pattern is determined, and the probability of the thermal runaway of the energy storage system is determined according to the target pattern.

[0020] As an optional implementation, in the first aspect of the present application, the control of the input signal increasing in the preset signal range with the change of the node time for each of the plurality of continuous target times comprises:

[0021] The control of the input signal increasing in the preset signal range according to the trend of an exponential function with the change of the node time for each of the plurality of continuous target times.

[0022] As an optional implementation, in the first aspect of the present application, the method further comprises:

[0023] If it is determined that the probability of the thermal runaway of the energy storage system exceeds a preset first threshold, the steps one to three are retriggered, and the time interval between the two adjacent target times in the step one is shortened.

[0024] As an optional implementation, in the first aspect of the present application, the method is further used for monitoring different positions of the energy storage system by a plurality of parameter sensors, and the method further comprises:

[0025] The corresponding operations in the steps one and two are controlled to be performed by all the parameter sensors, so as to obtain the output signal sequence corresponding to each parameter sensor;

[0026] The three-dimensional output signal sequence is constructed according to the output signal sequences corresponding to all the parameter sensors;

[0027] The three-dimensional signal sequence is regarded as analog three-dimensional image data mapping to a Hough space, and a feature shape is extracted from the analog three-dimensional image data based on a Hough transform;

[0028] The feature shape is input into a pre-trained thermal runaway early warning model, and the probability of the thermal runaway of the energy storage system is determined according to the output result of the thermal runaway early warning model.

[0029] As an optional implementation, in the first aspect of the present application, the inputting of the feature shape into the pre-trained thermal runaway early warning model comprises:

[0030] applying a high-frequency filter and a low-frequency filter to the three-dimensional matrix corresponding to the feature shape along a preset direction to obtain a high-frequency sub-band and a low-frequency sub-band, wherein the filtering frequency corresponding to the high-frequency filter is higher than the filtering frequency corresponding to the low-frequency filter;

[0031] inputting the low-frequency sub-band into the pre-trained thermal runaway early warning model, if the output result of the thermal runaway early warning model indicates that the probability of the thermal runaway phenomenon of the energy storage system is greater than a preset second threshold, inputting the high-frequency sub-band into the pre-trained thermal runaway early warning model, and determining the probability of the thermal runaway phenomenon of the energy storage system according to the output result of the thermal runaway early warning model.

[0032] As an optional implementation, in the first aspect of the present application, the parameter sensor comprises one of a gas sensor, a pressure sensor, a temperature sensor, a sound sensor and an optical sensor.

[0033] The second aspect of the present application discloses a thermal runaway early warning device for an energy storage system, which is used for monitoring the energy storage system through an energy storage system parameter sensor, the parameter sensor is used to change the output signal of the input signal after the parameter sensor in response to the change of the energy storage system parameter, and the device comprises:

[0034] a signal acquisition module, configured to, for each of a plurality of preset continuous target time points, control the input signal to change in a preset signal change rule within a preset signal range over time from the target time point as a starting point, and obtain an output signal corresponding to different input signals to obtain an output signal set; wherein the signal change rule corresponding to each target time point is the same;

[0035] a sequence construction module, configured to construct an output signal sequence based on the output signal set obtained at all the target time points, each row of the output signal sequence corresponding to a different output signal in a target time point, and each column of the output signal sequence corresponding to an output signal of different target time points arranged in time sequence;

[0036] a thermal runaway analysis module, configured to map the output signal sequence to a Hough space, obtain a feature structure corresponding to the signal sequence based on Hough transform, and determine the probability of the thermal runaway phenomenon of the energy storage system according to the feature structure.

[0037] As an optional implementation form, in the second aspect of the present application, the signal acquisition module controls the input signal to change in the preset signal range according to a preset signal change rule starting from each of the preset plurality of continuous target time points, and obtains the output signal corresponding to the input signal at each of the target time points, to obtain an output signal set corresponding to the target time point. The specific operation mode of obtaining the output signal set corresponding to the target time point includes:

[0038] For each of the preset plurality of continuous target time points, a time point after the target time point by a preset time difference is determined as a node time point, the input signal is controlled to increase in the preset signal range according to the change of the node time point, so that each node time point corresponds to a different input signal; and the output signal corresponding to the input signal at each node time point is obtained, to obtain an output signal set corresponding to the target time point; wherein the output signal set includes the output signal of each node time point corresponding to the target time point.

[0039] As an optional implementation form, in the second aspect of the present application, the specific operation mode of the thermal runaway analysis module for mapping the output signal sequence to a Hough space and obtaining the feature structure corresponding to the signal sequence based on Hough transformation includes:

[0040] The signal sequence is regarded as analog image data, and the analog image data is mapped to a Hough space, and a feature pattern is extracted from the analog image data based on Hough transformation.

[0041] And the specific operation mode of determining the probability of the energy storage system occurring thermal runaway phenomenon according to the feature structure includes:

[0042] The feature pattern is matched with a plurality of target patterns stored in advance, a target pattern matched with the feature pattern is determined, and the probability of the energy storage system occurring thermal runaway phenomenon is determined according to the target pattern.

[0043] As an optional implementation form, in the second aspect of the present application, the specific operation mode of the signal acquisition module for controlling the input signal to increase in the preset signal range according to the change of the node time point for each of the preset plurality of continuous target time points includes:

[0044] For each of the preset plurality of continuous target time points, the input signal is controlled to increase in the preset signal range according to the trend of an exponential function according to the change of the node time point.

[0045] As an optional implementation form, in the second aspect of the present application, the device further includes:

[0046] a feedback adjustment module configured to, when determining that the probability of the thermal runaway of the energy storage system exceeds the first preset threshold, re-trigger the signal acquisition module and the sequence construction module to perform the corresponding operation steps, and shorten the time interval between two adjacent target time points in the corresponding operation of the signal acquisition module.

[0047] As an optional implementation form, in the second aspect, the device is further configured to monitor different positions of the energy storage system through a plurality of parameter sensors, and the device further comprises:

[0048] a joint signal acquisition module configured to control all the parameter sensors to perform the corresponding operations of the signal acquisition module and the sequence construction module, so as to obtain the output signal sequence corresponding to each parameter sensor;

[0049] a three-dimensional sequence construction module configured to construct a three-dimensional output signal sequence according to the output signal sequences corresponding to all the parameter sensors;

[0050] a Hough transform module configured to map the three-dimensional signal sequence to a Hough space as simulated three-dimensional image data, extract a feature shape from the simulated three-dimensional image data based on Hough transform;

[0051] a model analysis module configured to input the feature shape into a pre-trained thermal runaway early warning model, and determine the probability of the thermal runaway of the energy storage system according to an output result of the thermal runaway early warning model.

[0052] As an optional implementation form, in the second aspect, the specific operation mode in which the model analysis module inputs the feature shape into the pre-trained thermal runaway early warning model and determines the probability of the thermal runaway of the energy storage system according to the output result of the thermal runaway early warning model comprises:

[0053] applying a high-frequency filter and a low-frequency filter to the three-dimensional matrix corresponding to the feature shape along a preset direction to obtain a high-frequency subband and a low-frequency subband, wherein the filtering frequency corresponding to the high-frequency filter is higher than the filtering frequency corresponding to the low-frequency filter;

[0054] inputting the low-frequency subband into the pre-trained thermal runaway early warning model, if the output result of the thermal runaway early warning model indicates that the probability of the thermal runaway of the energy storage system is greater than a second preset threshold, inputting the high-frequency subband into the pre-trained thermal runaway early warning model, and determining the probability of the thermal runaway of the energy storage system according to the output result of the thermal runaway early warning model.

[0055] Compared with the prior art, the present application has the following beneficial effects:

[0056] The application firstly adopts a scheme of controlling input signals to change in a preset signal change rule within a preset signal range over time, and obtaining output signals corresponding to different input signals, so that more abundant parameter information can be obtained; subsequently, the abundant parameter information is simulated as image information by constructing an output signal sequence, thereby facilitating the analysis and processing of data; finally, the output signal sequence is processed based on Hough transformation, so that the characteristic structure in the parameter information can be efficiently and accurately recognized, and the analysis of the probability of thermal runaway is realized based on the characteristic structure, thereby improving the accuracy of the thermal runaway early warning of the energy storage system. BRIEF DESCRIPTION OF DRAWINGS

[0057] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiment description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative labor.

[0058] Figure 1 is a flow diagram of a thermal runaway early warning method of an energy storage system disclosed by the embodiments of the present application;

[0059] Figure 2 is a structural diagram of a thermal runaway early warning device disclosed by the embodiments of the present application. DETAILED DESCRIPTION

[0060] In order to make the personnel in the technical field better understand the present application scheme, the following will combine the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0061] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish different objects, not to describe a specific order. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product or end including a series of steps or units is not limited to the listed steps or units, but can optionally include steps or units not listed, or can optionally include other steps or units inherent to the process, method, product or end.

[0062] Reference to an "embodiment" in this document means that a particular feature, structure, or characteristic described in connection with the embodiment can be included in at least one embodiment of the application. The appearances of the phrase in various places in the specification are not necessarily all referring to the same embodiment, nor are they necessarily mutually exclusive of one another. As will be apparent to those of ordinary skill in the art, embodiments described herein can be combined with other embodiments.

[0063] Embodiment one

[0064] Reference is made to Figure 1 , Figure 1 is a flowchart of a method for thermal runaway early warning of an energy storage system according to an embodiment of the application. In the method, Figure 1 The method for thermal runaway early warning of an energy storage system can be integrated into a device for thermal runaway early warning of an energy storage system, which can be integrated into a cloud server or a local server. The method is used to monitor the energy storage system through an energy storage system parameter sensor, which is used to change the output signal of the input signal after the parameter sensor in response to the change of the energy storage system parameter. As Figure 1 The method for thermal runaway early warning of an energy storage system can include the following operations:

[0065] Step one, for each of a plurality of preset continuous target time points, the input signal is controlled to change in a preset signal change rule in a preset signal range with time as a starting point, and the output signal corresponding to different input signals is obtained, and an output signal set is obtained.

[0066] In the embodiment of the application, the signal change rule corresponding to each target time point is the same. The parameter sensor can include one of a gas sensor, a pressure sensor, a temperature sensor, a sound sensor, and an optical sensor. The temperature sensor is commonly used. The conventional energy storage system temperature detection scheme uses an optical fiber sensor or a dynamic impedance phase shift method to obtain the internal temperature of the battery cell in real time. For example, the optical fiber sensor is pre-buried in the battery cell, and the temperature is warned in advance by the sudden change of the temperature rise rate. However, the temperature information of the above scheme only collects the real-time temperature of the energy storage system, and the information amount is not large enough. In order to make the temperature information carry more effective data, in the embodiment of the application, the temperature sensor is taken as an example, which is different from the conventional temperature detection scheme, and the scheme of controlling the input signal to change in a preset signal change rule in a preset signal range with time is adopted, and the output signal corresponding to different input signals is obtained.

[0067] In an optional embodiment, the scheme in step one can include:

[0068] For each of the plurality of preset continuous target time moments, a node time moment is determined at each time moment after the target time moment by a preset time difference, the control input signal is increased in the preset signal range with the change of the node time moment, so that each node time moment corresponds to a different input signal, and an output signal corresponding to each node time moment input signal is obtained to obtain an output signal set corresponding to the target time moment; wherein the output signal set can include the output signal of each node time moment corresponding to the target time moment

[0069] For the scheme in step one, the following example is given, but this example is not a specific limitation on step one in the embodiments of the present application:

[0070] The plurality of preset continuous target time moments are the 10th, 20th, 30th, 40th, 50th and 60th seconds in each minute; for example, in the 10th second in each target time moment, the control input signal (voltage or current or frequency or other signals of the input signal) is increased or decreased from the 10th second to the 19th second, for example, exponentially increased, or increased at a constant rate, etc. The output signals corresponding to different input signals can be obtained randomly from the 10th second to the 19th second, or can be obtained once every 1 second; or the input signal can be changed every interval, and the output signal is obtained once after each change of the input signal. Finally, the output signal set is obtained.

[0071] Step two, based on the output signal set obtained from all target time moments, an output signal sequence is constructed.

[0072] In the embodiments of the present application, one of the inventive concepts of the embodiments of the present application is to construct an output signal sequence. Unlike traditional parameter monitoring (temperature, gas, pressure, sound and light, etc.), the embodiments of the present application construct a signal sequence, each row of the output signal sequence corresponds to a different output signal in a target time moment, and each column of the output signal sequence corresponds to an output signal of a different target time moment arranged in time sequence. The signal output sequence forms an array, which is plotted in a plane graph, each output signal value of each target time moment forms a pixel point, and a signal change image is simulated. Next, only the simulated image needs to be analyzed, and information that may indicate thermal runaway in the image can be mined.

[0073] Step three, mapping the output signal sequence to the Hough space, obtaining the feature structure corresponding to the signal sequence based on the Hough transform, and determining the probability of thermal runaway phenomenon of the energy storage system according to the feature structure.

[0074] Hough Transform is a classic image processing algorithm, mainly used for detecting specific geometric shapes such as straight lines, circles, etc. from images. Its principle is based on mapping points in image space to parameter space, and accumulating the distribution of points in parameter space to detect geometric shapes in images. The core idea of Hough Transform is to map points in image space to parameter space. In image space, a point may belong to multiple geometric shapes; while in parameter space, a geometric shape corresponds to a specific parameter set. By mapping all points in the image to the parameter space and counting the distribution of points in the parameter space, the peak points in the parameter space can be found, and the parameters corresponding to these peak points are the parameters of the geometric shapes in the image.

[0075] During the operation of the energy storage system, various parameters of the energy storage system are constantly changing. For example, the higher the output power of the energy storage system, the higher the temperature. Once the energy storage system appears thermal runaway or is about to appear thermal runaway, the changes of various parameters of the energy storage system will be abnormal, such as abnormal change of temperature, appearance of abnormal sound, emission of abnormal gas, etc. In the present application, the above parameter changes can be well displayed in the signal sequence, but how to identify the abnormal parameter signal and analyze its relationship with thermal runaway becomes the key. Therefore, the embodiments of the present application simulate the signal sequence as image data based on the Hough Transform principle in image processing, map the output signal sequence to Hough space, obtain the feature structure corresponding to the signal sequence based on Hough Transform, and the feature structure can be different parameter sets in Hough space, which also represents a graph with special structure, such as sawtooth shape, abnormal sharp peak, thin and large slope straight line, etc. Different feature structures represent different states of the energy storage system, and according to the feature structure, the probability of thermal runaway phenomenon of the energy storage system can be determined.

[0076] In an alternative embodiment, the feature structure can be put into a pre-trained artificial intelligence model, and the feature structure is analyzed and identified based on the artificial intelligence model, so as to correspond the corresponding feature structure to a certain state of the energy storage system. Further, based on the feature structure, the probability of thermal runaway is evaluated.

[0077] It can be seen that, in the embodiment of the present application, firstly, the scheme of changing the control input signal in the preset signal range with the preset signal change rule over time and obtaining the output signal corresponding to different input signals is adopted, which can obtain more abundant parameter information; subsequently, the abundant parameter information is simulated as image information by constructing the output signal sequence, thereby facilitating the analysis and processing of data; finally, the output signal sequence is processed based on the Hough transform, which can efficiently and accurately identify the characteristic structure in the parameter information, and the probability of thermal runaway is analyzed based on the characteristic structure, thereby improving the accuracy of the thermal runaway early warning of the energy storage system.

[0078] In an optional embodiment, mapping the output signal sequence to the Hough space, obtaining the characteristic structure corresponding to the signal sequence based on the Hough transform can include:

[0079] Mapping the signal sequence to the Hough space as analog image data, and extracting the characteristic pattern from the analog image data based on the Hough transform;

[0080] And, determining the probability of the thermal runaway phenomenon of the energy storage system according to the characteristic structure can include:

[0081] Matching the characteristic pattern with a plurality of target patterns stored in advance, determining the target pattern matched with the characteristic pattern, and determining the probability of the thermal runaway phenomenon of the energy storage system according to the target pattern.

[0082] The optional embodiment finds that:

[0083] The Hough transform is commonly used in the field of image processing, and is mainly used for detecting specific shapes such as straight lines, circles, etc. After the time sequence of the parameter sequence signal changing over time is converted to the Hough space through the Hough transform, the following types of information can be obtained:

[0084] (1) Parameter information

[0085] For example, period and frequency, if the time sequence presents periodic change, after the Hough transform, specific parameter combinations in the Hough space will be obtained. For example, in the polar or Cartesian coordinate system of the Hough space, obvious peak values or aggregation areas may appear, and the parameter values corresponding to these areas can reflect the period length, frequency size, etc. of the signal. For example, the time sequence of a sinusoidal electrical signal, after processing, the Hough transform can help determine its fundamental frequency and corresponding period.

[0086] Amplitude information, the value of each unit in the accumulator of the Hough space is usually related to the number of times or energy of the corresponding parameter combination appearing in the time sequence. Therefore, the peak height in the accumulator can reflect the amplitude size of the signal at the corresponding frequency or period, thereby understanding the energy distribution of the signal at different frequency components.

[0087] Phase information: For electrical signals with clear phase characteristics, such as square waves, sine waves, etc., phase-related parameter information can also be extracted in the Hough space. By analyzing the results of Hough transform, the starting point, wave peak and trough position of the signal at different periods or frequencies, etc. phase characteristics are determined, and the phase information of the signal is obtained.

[0088] (2) Feature structure information

[0089] Main components: Hough transform can highlight the main change patterns in time series, that is, the main frequency components or periodic characteristics. In Hough space, these main components usually appear as significant peaks or continuous high-value regions. By identifying and extracting these regions, the most significant and key change rules in electrical signals can be clearly understood, so as to distinguish the main components from noise or other secondary components.

[0090] Harmonic relationship: If the electrical signal contains multiple harmonic components, the results of Hough transform can help reveal the relationship between these harmonics. In Hough space, the parameter positions corresponding to the fundamental frequency and its harmonics often have certain regularity and correlation. By analyzing these positions, the number of harmonics, the relative amplitude and phase of each harmonic, etc. information in the signal can be determined, and the harmonic structure of the signal can be further understood.

[0091] Inter-harmonic detection: For some electrical signals with inter-harmonics, Hough transform also has detection capability. Inter-harmonic refers to the frequency component between the fundamental frequency and the harmonic. Traditional Fourier transform may have difficulty in accurately detecting these components, while Hough transform can capture the existence of these inter-harmonics through detailed analysis of Hough space, and determine their frequency and amplitude parameters.

[0092] (3) Pattern recognition information

[0093] Pattern classification: According to the parameter characteristics obtained after Hough transform, different types of electrical signal patterns can be classified. For example, by comparing and analyzing the manifestations of periodic signals, trend signals, random signals, etc. in Hough space, corresponding classification models or discrimination criteria can be established, so as to realize the automatic recognition and classification of unknown electrical signal patterns.

[0094] Pattern matching: When a specific electrical signal pattern is known to correspond to a specific parameter range or structure in Hough space, the time series to be tested can be converted to Hough space by Hough transform, and matched with the parameter characteristics of the known pattern. If the matching degree is high, it means that the signal to be tested may have the specific pattern. This method can be used in fault diagnosis, anomaly detection, etc. to quickly judge whether the signal conforms to a certain normal or abnormal pattern characteristics.

[0095] Based on the above findings in the optional embodiment, in the optional embodiment, the signal sequence is regarded as mapping the analog image data to the Hough space, and a feature pattern is extracted from the analog image data based on the Hough transform; the feature pattern contains more detailed parameter change information, such as various subtle changes in the temperature of the energy storage system.

[0096] In addition, according to the feature structure, the probability that the energy storage system has a thermal runaway phenomenon can be determined, which can include:

[0097] The feature pattern is matched with a plurality of target patterns stored in advance, and a target pattern matched with the feature pattern is determined, and the probability that the energy storage system has a thermal runaway phenomenon is determined according to the target pattern.

[0098] In the optional embodiment, for different energy storage systems, various experiments can be performed before leaving the factory, so as to determine various target patterns in advance, such as a plurality of target patterns corresponding to stable operation of the energy storage system, a plurality of target patterns corresponding to possible thermal runaway of the energy storage system, and a plurality of target patterns corresponding to thermal runaway of the energy storage system. Finally, matching the feature pattern with the target pattern can reduce the calculation amount of the thermal runaway probability analysis, so as to efficiently determine the probability that the energy storage system has a thermal runaway phenomenon.

[0099] It is also found in the embodiment of the application that the exponential function change in the signal sequence is more sensitive based on the principle of the Hough transform, and therefore, in order to further improve the accuracy of the thermal runaway early warning in the embodiment of the application, for each of a plurality of preset continuous target time points, the control input signal can be increased in a preset signal range according to the trend of the exponential function as the node time changes.

[0100] For each of a plurality of preset continuous target time points, the control input signal can be increased in a preset signal range according to the trend of the exponential function as the node time changes.

[0101] Although the above operation in the embodiment of the application can be more accurate than conventional detection, the construction of the signal sequence increases the calculation amount, especially in a real-time monitoring scenario, the method in the embodiment of the application is not advantageous in calculation amount. Therefore, in another optional embodiment, the method can further include:

[0102] If it is determined that the probability that the energy storage system has a thermal runaway phenomenon exceeds a preset first threshold, steps one to three are retriggered, and the time interval between the two adjacent target time points in step one is shortened.

[0103] In the optional embodiment, the control of the input signal within the preset signal range over time according to the preset signal change rule, and the obtaining of the output signal corresponding to the different input signals can be set in the conventional operation. The time interval between each target time is set to be larger, and the density of the output signal is larger. In this way, the calculation amount can be reduced in the conventional detection, which is similar to the idling operation state. Once it is found that the probability of thermal runaway of the energy storage system exceeds the preset first threshold value, the fine data acquisition stage is entered, that is, steps 1 to 3 are retriggered, and the time interval between the two adjacent target times in step 1 is shortened. In the optional embodiment, coarse detection is first performed. Once the coarse detection shows that an abnormality may occur, fine detection is immediately entered, so as to obtain more data and provide more accurate data for subsequent thermal runaway probability analysis. At the same time, the coarse detection is performed most of the time, which can reduce a large amount of calculation.

[0104] In the embodiment of the present application, for the energy storage system, common ones are lithium ion batteries, sodium-sulfur batteries, flow batteries (such as all-vanadium flow batteries), and lead-acid batteries. In different application scenarios, different chemical batteries are also designed to be different shapes or whole bodies composed of dispersed parts. Therefore, it is still not accurate to use only one parameter sensor to measure the parameters of one part of the energy storage system. For example, in some cases, thermal runaway may gradually spread from some local areas, and in other cases, thermal runaway often occurs in some special parts.

[0105] Therefore, in order to systematically perform thermal runaway early warning on the energy storage system, in an optional embodiment, the method is also used for monitoring different positions of the energy storage system by using multiple parameter sensors. The method can further include:

[0106] Controlling all parameter sensors to perform the corresponding operations in steps 1 and 2, so as to obtain the output signal sequence corresponding to each parameter sensor;

[0107] According to the output signal sequences corresponding to all parameter sensors, a three-dimensional output signal sequence is constructed. Similar to three-dimensional image data, each parameter sensor corresponds to a two-dimensional output signal sequence. The two-dimensional output signal sequences of multiple parameter sensors are combined together to form a three-dimensional output signal sequence similar to three-dimensional image data. For example, the x-axis of the three-dimensional output signal sequence is the parameter sensor number, the y-axis is the target time, and the z-axis is the node time corresponding to the target time. The value of each point of the three-dimensional output signal is the output signal value of a certain parameter sensor at a certain node time corresponding to a certain target time.

[0108] The three-dimensional signal sequence is regarded as simulated three-dimensional image data mapped to a Hough space. Based on the Hough transform, a feature shape is extracted from the simulated three-dimensional image data.

[0109] The characteristic shape is input into a pre-trained thermal runaway early warning model, and the probability of the energy storage system having a thermal runaway phenomenon is determined according to an output result of the thermal runaway early warning model.

[0110] In this optional embodiment, the three-dimensional signal sequence is regarded as mapping the simulated three-dimensional image data to the Hough space, and the characteristic shape can be extracted based on the Hough transform. In the Hough transform process, an accumulator array will be involved. For the three-dimensional Hough transform, the dimension of the accumulator depends on the number of parameters of the target shape to be detected. After the three-dimensional data is mapped to the Hough space, each data point will vote at the corresponding parameter position of the accumulator, so that the value of the parameter position corresponding to the data point in the accumulator gradually increases. Finally, the voting value distribution of each position in the accumulator reflects the aggregation of the three-dimensional data in the Hough space, and the area with a higher voting value is the parameter area where the target shape may exist. By analyzing the voting values in the accumulator, the region with specific geometric characteristics in the three-dimensional data can be determined, and thus the detection and positioning of the target shape can be realized. In addition, the results of the Hough transform can directly obtain the characteristic parameters of the target shape in the three-dimensional data. For example, for straight line detection, the starting point coordinates and direction vector of the straight line can be obtained; for plane detection, the normal vector and distance to the origin of the plane can be obtained; for sphere detection, the sphere center coordinates and radius can be obtained, etc. These characteristic parameters can accurately describe the position, direction and size of the target shape in the three-dimensional space, and provide a strong basis for subsequent analysis and processing.

[0111] In this optional embodiment, the characteristic shape is input into a pre-trained thermal runaway early warning model, and the probability of the energy storage system having a thermal runaway phenomenon is determined according to an output result of the thermal runaway early warning model. Different characteristic shapes can reflect the change of specific parameters on the energy storage system, and abnormal parameter change corresponds to abnormal characteristic shape. Taking temperature as an example, the temperatures at different positions on the energy storage system are usually not much different, but when the temperature spreads from a certain position to other positions, it usually indicates the occurrence of an abnormality or even a thermal runaway site, at which time the characteristic shape will also change to a top hat shape. Based on the pre-trained thermal runaway model, the change of the characteristic shape can be analyzed to evaluate the probability of the occurrence of thermal runaway. It can be seen that this optional embodiment can more systematically and comprehensively detect the energy storage system, thereby further improving the accuracy of the thermal runaway early warning of the energy storage system.

[0112] In the optional embodiment, the three-dimensional signal sequence is regarded as mapping the analog three-dimensional image data to the Hough space, and the amount of calculation still has room for optimization. In the optional embodiment, it is found that for the three-dimensional signal sequence, the low-frequency component contains the basic features of the signal, and the high-frequency signal gives the detailed information of the signal, such as the edge contour of some image information in the analog three-dimensional image data. Therefore, in the optional embodiment, further optionally, the feature shape is input into the pre-trained thermal runaway early warning model, and according to the output result of the thermal runaway early warning model, the probability of the thermal runaway phenomenon of the energy storage system is determined, which can include:

[0113] The three-dimensional matrix corresponding to the feature shape is filtered along the preset direction by using a high-frequency filter and a low-frequency filter to obtain a high-frequency sub-band and a low-frequency sub-band, wherein the filtering frequency corresponding to the high-frequency filter is higher than the filtering frequency corresponding to the low-frequency filter; wherein the low-frequency component contains the basic features of the signal, and the high-frequency signal gives the detailed information of the signal.

[0114] The low-frequency sub-band is input into the pre-trained thermal runaway early warning model, and if the output result of the thermal runaway early warning model indicates that the probability of the thermal runaway phenomenon of the energy storage system is greater than a preset second threshold, then the high-frequency sub-band is input into the pre-trained thermal runaway early warning model, and according to the output result of the thermal runaway early warning model, the probability of the thermal runaway phenomenon of the energy storage system is determined, and the probability obtained at this time is a more accurate probability. If the output result of the thermal runaway early warning model indicates that the probability of the thermal runaway phenomenon of the energy storage system is less than or equal to the preset second threshold, then the current obtained probability value is directly output, and the probability value is a relatively rough probability value.

[0115] In the optional embodiment, when filtering the image, the frequency on the image represents the speed of change of the gray value in the image, which is as follows:

[0116] Low-frequency component: corresponding to the slowly changing gray component in the image, such as the large flat area in the image, the gray value of these areas is continuous and similar, the change amplitude is small, which forms the basic gray level of the image, determines the overall brightness and large light and dark trend of the image, and is a comprehensive measure of the intensity of the whole image, such as the large area and uniform color area in a landscape photo, such as the sky or the lake surface.

[0117] High-frequency component: corresponding to the gray component that changes sharply in the image, that is, the edge, contour, texture details and noise of the image. In these places, the difference between the gray values of adjacent pixels is large, and the change speed is fast, such as the junction of objects and backgrounds, the edges of buildings, etc. The high-frequency component reflects the detailed information of the image and the sudden change of the objects in the image.

[0118] It can be seen that in the optional embodiment, the three-dimensional matrix corresponding to the feature shape is filtered along the preset direction by using the high-frequency filter and the low-frequency filter to obtain a high-frequency sub-band and a low-frequency sub-band. When calculating, the low-frequency sub-band containing the basic features of the signal is analyzed first to roughly determine the thermal runaway probability. Only when the rough thermal runaway probability exceeds a certain threshold, the high-frequency sub-band is input into the model for analysis to obtain a more accurate thermal runaway probability result, so that the calculation amount of the method implementation is reduced under the premise of ensuring the thermal runaway early warning effect.

[0119] Embodiment two

[0120] The embodiment of the application discloses a thermal runaway early warning device of an energy storage system. The device is used for monitoring the energy storage system through an energy storage system parameter sensor. The parameter sensor is used to change the output signal of the input signal after the parameter sensor in response to the change of the energy storage system parameter. As shown in the figure, the device can include: Figure 2

[0121] The signal acquisition module is used to control the input signal to change in the preset signal range with a preset signal change rule from each target time as a starting point, and obtain the output signal corresponding to different input signals to obtain an output signal set. The signal change rule corresponding to each target time is the same.

[0122] The sequence construction module is used to construct an output signal sequence based on the output signal set obtained by all target times. Each row of the output signal sequence corresponds to different output signals in a target time, and each column of the output signal sequence corresponds to the output signals of different target times arranged in time sequence.

[0123] The thermal runaway analysis module is used to map the output signal sequence to the Hough space, obtain the feature structure corresponding to the signal sequence based on the Hough transform, and determine the probability of the thermal runaway phenomenon of the energy storage system according to the feature structure.

[0124] In an optional embodiment, the signal acquisition module controls the input signal to change in the preset signal range with a preset signal change rule from each target time as a starting point, and obtains the output signal corresponding to different input signals to obtain the specific operation mode of the output signal set, which can include:

[0125] ​For each of the plurality of preset continuous target time points, a node time point is determined as a time point after the target time point by a preset time difference, and the input signal is controlled to increase in the preset signal range with the change of the node time point, so that each node time point corresponds to a different input signal; and an output signal corresponding to each node time point input signal is obtained to obtain an output signal set corresponding to the target time point; wherein the output signal set can include the output signal of each node time point corresponding to the target time point.

[0126] In another optional embodiment, the thermal runaway analysis module maps the output signal sequence to a Hough space, and the specific operation mode of obtaining the feature structure corresponding to the signal sequence based on the Hough transform can include:

[0127] The signal sequence is regarded as analog image data mapped to the Hough space, and the feature pattern is extracted from the analog image data based on the Hough transform;

[0128] And the probability of the thermal runaway phenomenon of the energy storage system is determined according to the feature structure, which can include:

[0129] The feature pattern is matched with a plurality of preset target patterns, and the target pattern matched with the feature pattern is determined, and the probability of the thermal runaway phenomenon of the energy storage system is determined according to the target pattern.

[0130] In yet another optional embodiment, the signal acquisition module can include, for each of the plurality of preset continuous target time points, the specific operation mode of controlling the input signal to increase in the preset signal range with the change of the node time point can include:

[0131] For each of the plurality of preset continuous target time points, the input signal is controlled to increase in the preset signal range according to the trend of the exponential function with the change of the node time point.

[0132] In yet another optional embodiment, the device can further include:

[0133] The feedback adjustment module is configured to, when the probability of the thermal runaway phenomenon of the energy storage system is determined to exceed the preset first threshold, re-trigger the signal acquisition module and the sequence construction module to perform the respective corresponding operation steps, and shorten the time interval between the adjacent two target time points in the operation of the signal acquisition module.

[0134] In yet another optional embodiment, the device is further configured to monitor different positions of the energy storage system through a plurality of parameter sensors, and the device can further include:

[0135] The joint signal acquisition module controls all parameter sensors to perform operations corresponding to the signal acquisition module and the sequence construction module, so as to obtain an output signal sequence corresponding to each parameter sensor.

[0136] The three-dimensional sequence construction module is configured to construct a three-dimensional output signal sequence according to the output signal sequences corresponding to all parameter sensors.

[0137] The Hough transform module is configured to map the three-dimensional signal sequence to a Hough space as analog three-dimensional image data, extract a feature shape from the analog three-dimensional image data based on Hough transform.

[0138] The model analysis module is configured to input the feature shape into a pre-trained thermal runaway early warning model, and determine a probability of occurrence of a thermal runaway phenomenon of the energy storage system according to an output result of the thermal runaway early warning model.

[0139] In another optional embodiment, the model analysis module inputs the feature shape into a pre-trained thermal runaway early warning model, and determines the probability of occurrence of the thermal runaway phenomenon of the energy storage system according to an output result of the thermal runaway early warning model, and the specific operation mode can include:

[0140] The three-dimensional matrix corresponding to the feature shape is filtered along a preset direction by using a high-frequency filter and a low-frequency filter to obtain a high-frequency sub-band and a low-frequency sub-band, wherein the filtering frequency corresponding to the high-frequency filter is higher than the filtering frequency corresponding to the low-frequency filter.

[0141] The low-frequency sub-band is input into the pre-trained thermal runaway early warning model, if the output result of the thermal runaway early warning model indicates that the probability of occurrence of the thermal runaway phenomenon of the energy storage system is greater than a preset second threshold, then the high-frequency sub-band is input into the pre-trained thermal runaway early warning model, and the probability of occurrence of the thermal runaway phenomenon of the energy storage system is determined according to the output result of the thermal runaway early warning model.

[0142] In another optional embodiment, the parameter sensor can include one of a gas sensor, a pressure sensor, a temperature sensor, a sound sensor and an optical sensor.

[0143] Embodiment three

[0144] The embodiment of the application discloses a computer storage medium, which stores computer instructions, and when the computer instructions are called, the steps in the energy storage system thermal runaway early warning method described in the embodiment one are executed.

[0145] Embodiment four

[0146] The embodiment of the present application discloses a computer program product, which comprises a non-transitory computer-readable storage medium storing a computer program, and the computer program is operable to cause a computer to execute the steps in the energy storage system thermal runaway early warning method described in embodiment one.

[0147] The device embodiments described above are only schematic, wherein the modules illustrated as separate components can or can not be physically separated, and the components illustrated as modules can or can not be physical modules, i.e., can be located in one place or distributed on multiple network modules. Part or all of the modules can be selected to achieve the purpose of the embodiment scheme according to actual needs. Those skilled in the art can understand and implement it without creative labor.

[0148] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be realized by means of software and the necessary general hardware platform, and of course, it can also be realized by hardware. Based on such understanding, the above technical solutions can be embodied in the form of software products, and the computer software product can be stored in a computer readable storage medium, including a read-only memory (Read-Only Memory, ROM), a random access memory (Random Access Memory, RAM), a programmable read-only memory (Programmable Read-only Memory, PROM), an erasable programmable read-only memory (Erasable Programmable Read Only Memory, EPROM), a one-time programmable read-only memory (One-time Programmable Read-Only Memory, OTPROM), an electrically erasable programmable read-only memory (Electrically-Erasable Programmable Read-Only Memory, EEPROM), a compact disc read-only memory (Compact Disc Read-Only Memory, CD-ROM) or other optical disk storage, a magnetic disk storage, a magnetic tape storage, or any other computer readable medium that can be used to carry or store data.

[0149] Finally, it should be noted that: the energy storage system disclosed in the embodiment of the application thermal runaway early warning method and device, the storage medium disclosed is only the preferred embodiment of the application, only for the purpose of describing the technical solutions of the application, not for its limitation; although the application is described in detail with reference to the foregoing embodiments, those skilled in the art should understand; the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the application.

Claims

1. A method for thermal runaway early warning of an energy storage system, the method comprising: The method is used for monitoring the energy storage system by an energy storage system parameter sensor, the parameter sensor is used for changing an output signal of an input signal after the input signal passes through the parameter sensor in response to a change of an energy storage system parameter, and the method comprises the following steps: Step one, for each of a plurality of preset continuous target time points, taking the target time point as a starting point, controlling the input signal to change in a preset signal change rule within a preset signal range over time, and obtaining an output signal corresponding to different input signals to obtain an output signal set; wherein the signal change rule corresponding to each target time point is the same; Step two, based on the output signal set obtained at all target time points, an output signal sequence is constructed, each row of the output signal sequence corresponds to a different output signal at a target time point, and each column of the output signal sequence corresponds to an output signal at different target time points arranged in time sequence; Step three, mapping the output signal sequence to a Hough space, obtaining a feature structure corresponding to the signal sequence based on Hough transform, and determining a probability of thermal runaway of the energy storage system according to the feature structure.

2. The energy storage system thermal runaway early warning method of claim 1, wherein, The method for each of a plurality of preset continuous target time points, taking the target time point as a starting point, controlling the input signal to change in a preset signal change rule within a preset signal range over time, and obtaining an output signal corresponding to different input signals to obtain an output signal set, comprises the following steps: For each of a plurality of preset continuous target time points, taking the target time point as a starting point, determining that each time point after a preset time difference as a node time point, controlling the input signal to increase in a preset signal range over time with the change of the node time point, so that each node time point corresponds to a different input signal; and obtaining an output signal corresponding to the input signal at each node time point to obtain an output signal set corresponding to the target time point; wherein the output signal set comprises the output signal of each node time point corresponding to the target time point.

3. The energy storage system thermal runaway early warning method of claim 1 or 2, wherein, The method for each of a plurality of preset continuous target time points, taking the target time point as a starting point, controlling the input signal to change in a preset signal change rule within a preset signal range over time, and obtaining an output signal corresponding to different input signals to obtain an output signal set, comprises the following steps: The method for each of a plurality of preset continuous target time points, taking the target time point as a starting point, controlling the input signal to change in a preset signal change rule within a preset signal range over time, and obtaining an output signal corresponding to different input signals to obtain an output signal set, comprises the following steps: The method for each of a plurality of preset continuous target time points, taking the target time point as a starting point, controlling the input signal to change in a preset signal change rule within a preset signal range over time, and obtaining an output signal corresponding to different input signals to obtain an output signal set, comprises the following steps: The method for each of a plurality of preset continuous target time points, taking the target time point as a starting point, controlling the input signal to change in a preset signal change rule within a preset signal range over time, and obtaining an output signal corresponding to different input signals to obtain an output signal set, comprises the following steps:

4. The energy storage system thermal runaway early warning method of claim 2, wherein, The method for each of a plurality of preset continuous target time points, taking the target time point as a starting point, controlling the input signal to change in a preset signal change rule within a preset signal range over time, and obtaining an output signal corresponding to different input signals to obtain an output signal set, comprises the following steps: The method further comprises:

5. The energy storage system thermal runaway early warning method of claim 2 or 4, wherein, ​ If it is determined that the probability of the energy storage system having thermal runaway is greater than a first threshold, steps one to three are re-triggered and the time interval between two adjacent target time points in step one is shortened.

6. The energy storage system thermal runaway early warning method of claim 1, wherein, The method further comprises: controlling all parameter sensors to perform corresponding operations in steps one and two, so as to obtain a corresponding output signal sequence of each parameter sensor; constructing a three-dimensional output signal sequence according to the output signal sequences of all parameter sensors; mapping the three-dimensional output signal sequence to a Hough space as simulated three-dimensional image data, extracting a feature shape from the simulated three-dimensional image data based on Hough transform; inputting the feature shape into a pre-trained thermal runaway early warning model, and determining the probability of the energy storage system having thermal runaway according to an output result of the thermal runaway early warning model.

7. The energy storage system thermal runaway early warning method of claim 6, wherein, The method further comprises: applying a high-frequency filter and a low-frequency filter to a three-dimensional matrix corresponding to the feature shape along a preset direction to obtain a high-frequency sub-band and a low-frequency sub-band, wherein the filter frequency of the high-frequency filter is higher than that of the low-frequency filter; if the output result of the thermal runaway early warning model indicates that the probability of the energy storage system having thermal runaway is greater than a second threshold, inputting the high-frequency sub-band into the pre-trained thermal runaway early warning model, and determining the probability of the energy storage system having thermal runaway according to an output result of the thermal runaway early warning model.

8. The energy storage system thermal runaway early warning method of any of claims 1-2, 4, 6-7, wherein, The parameter sensor comprises one of a gas sensor, a pressure sensor, a temperature sensor, a sound sensor, and an optical sensor.

9. An energy storage system thermal runaway early warning device, characterized in that, The device is used for monitoring an energy storage system through an energy storage system parameter sensor, the parameter sensor is used for responding to changes in energy storage system parameters, so that the output signal of the input signal changes after passing through the parameter sensor, and the device comprises: a signal acquisition module, configured to, for each of a plurality of preset continuous target time points, control the input signal to change in a preset signal change rule within a preset signal range over time starting from the target time point, and obtain output signals corresponding to different input signals to obtain an output signal set; wherein the signal change rule corresponding to each target time point is the same; a sequence construction module, configured to construct an output signal sequence based on the output signal sets obtained for all target time points, each row of the output signal sequence corresponding to a different output signal in a target time point, and each column of the output signal sequence corresponding to output signals of different target time points arranged in time sequence; a feature shape extraction module, configured to map the output signal sequence to a Hough space as simulated three-dimensional image data, extract a feature shape from the simulated three-dimensional image data based on Hough transform, and input the feature shape into a pre-trained thermal runaway early warning model to determine the probability of the energy storage system having thermal runaway according to an output result of the thermal runaway early warning model. A thermal runaway analysis module is configured to map the output signal sequence to a Hough space, obtain a feature structure corresponding to the signal sequence based on a Hough transform, and determine a probability of thermal runaway of the energy storage system according to the feature structure.

10. A computer storage medium, characterized in that, The computer storage medium stores computer instructions, which are invoked to perform the energy storage system thermal runaway early warning method according to any one of claims 1-8.

Images