Electric arc discharge detection method of parallel energy storage battery unit and related product

Abstract

The invention provides an electric arc discharge detection method of a parallel energy storage battery unit and a related product. A plurality of energy storage battery units are connected in parallel, microphones are arranged in the energy storage battery units, battery management systems of the energy storage battery units are configured to detect end currents and end voltages of the energy storage battery units, and the electric arc discharge detection method comprises the steps that the battery management systems judge whether the end currents and the end voltages have electric arc discharge characteristics or not; if yes, the battery management system controls the energy storage battery unit to be disconnected; collecting a sound signal detected by the microphone in the same time period when the electric arc discharge characteristic appears; identifying whether the sound signal has a sound feature in an electric arc generation state or not; and if yes, determining that the energy storage battery unit has an electric arc discharge event. According to the scheme provided by the invention, the electrical arcing identification of the sound signal dimension is carried out according to the sound signal detected by the microphone in the time period when the electrical arcing characteristic appears, so that the judgment is carried out by utilizing the multi-dimensional data, the identification accuracy is improved, and the false detection and the missing detection are reduced.

Description

Technical Field

[0001] The present invention relates to the field of energy storage, and particularly to an electric arc detection method and related products for a parallel energy storage battery unit. Background Art

[0002] Parallel connection of energy storage batteries is a technology that connects multiple battery units in parallel to increase the total system capacity and output power. This connection method is very common in energy storage systems, which can provide a higher current output capacity while maintaining a lower voltage level.

[0003] Through parallel connection, the capacity of the energy storage battery system is greatly improved. When in use, if one of the batteries fails, theoretically the other batteries can still continue to work, improving the reliability of the system. The parallel-connected battery units need to maintain voltage consistency, otherwise it may lead to current imbalance and affect the system performance.

[0004] An electric arc (or electric arcing) is a form of air ionization discharge, which is an instantaneous and rapidly developing process, generally lasting less than 1 s, but its arc column core temperature is extremely high and it has extremely high radiant energy. After long-term operation of the energy storage battery system, due to wire aging and vibration of connection terminals, it is easy to have unreliable electrical connections, thus there is a risk of arcing. Arcing may trigger a series of chain reactions in a short time, such as internal short circuit of the battery, electrolyte leakage, etc., and even cause a fire. Especially for high-energy-density energy storage batteries, the danger of arcing is extremely high.

[0005] However, due to the extremely short occurrence time of the electric arc, it is difficult to detect arcing between energy storage battery units, and the existing technology lacks means for detecting arcing in a parallel energy storage battery system. Summary of the Invention

[0006] An object of the present invention is to provide a method for accurately detecting abnormal electric arcing of a parallel energy storage battery unit.

[0007] A further object of the present invention is to reduce the influence of electric arcing of a parallel energy storage battery unit.

[0008] In particular, the present invention provides an electric arc detection method for a parallel energy storage battery unit. A plurality of energy storage battery units are connected in parallel, and a microphone is provided inside. The battery management system of the energy storage battery unit is configured to detect the terminal current and terminal voltage of the energy storage battery unit, and the electric arc detection method includes:

[0009] The battery management system determines whether the terminal current and terminal voltage have electric arc characteristics;

[0010] If so, the battery management system controls the energy storage battery unit to disconnect;

[0011] The sound signal detected by the acquisition microphone during the same period when the electric arc striking feature appears;

[0012] Identify whether the sound signal has the sound characteristics in the state of electric arc striking;

[0013] If so, determine that an electric arc striking event occurs in the energy storage battery unit.

[0014] Optionally, in the case of determining that no electric arc striking event occurs in the energy storage battery unit, it further includes:

[0015] Judge whether the terminal current and terminal voltage exceed their respective preset limits;

[0016] If either the terminal current or the terminal voltage exceeds the preset limit, maintain the disconnected state and report the over-limit abnormal state by the battery management system.

[0017] Optionally, in the case that both the terminal current and the terminal voltage do not exceed their respective preset limits, the above method for detecting electric arc striking of the parallel energy storage battery unit further includes:

[0018] The battery management system performs a reconnect recovery operation and restores the connection for a set duration;

[0019] Within the set duration, re-collect the terminal current and terminal voltage and perform abnormal detection.

[0020] Optionally, the energy storage battery unit is further provided with a DC-DC conversion module and a connection component. The connection component includes a main connection terminal and a spare connection terminal. The energy storage battery unit defaults to use the main connection terminal to connect to the DC-DC conversion module, and in the case of determining that an electric arc striking event occurs in the energy storage battery unit, it further includes:

[0021] The battery management system disconnects the connection between the main connection terminal and the DC-DC conversion module;

[0022] After an interval of the set duration, use the spare output terminal to connect to the DC-DC conversion module, so that the energy storage battery units are re-connected in parallel;

[0023] Re-collect the terminal current and terminal voltage and perform abnormal detection.

[0024] Optionally, after the step of re-collecting the terminal current and terminal voltage and performing abnormal detection in the above method, it further includes:

[0025] If the result of the abnormal detection is normal operation, the battery management system re-restores the connection between the main connection terminal and the DC-DC conversion module, so as to connect to the DC-DC conversion module using both the main connection terminal and the spare connection terminal at the same time;

[0026] Collect the terminal current and terminal voltage again and perform anomaly detection. If the result of the anomaly detection is still normal operation, disconnect the connection of the spare output terminal;

[0027] Detect whether the electric arc pulling event of the energy storage battery unit has been eliminated. If it has been eliminated, reconnect using the main connection terminal; if it has not been eliminated, disconnect the main connection terminal and use the spare output terminal for connection.

[0028] Optionally, the steps of identifying whether the sound signal has the sound characteristics in the state of electric arc pulling include:

[0029] Process the terminal current and terminal voltage into an electrical feature data set;

[0030] Process the sound signal into a noise feature data set;

[0031] Input the electrical feature data set and the noise feature data set into a pre-trained machine learning model;

[0032] The machine learning model makes an identification and judgment to determine whether an arc occurs.

[0033] Optionally, the machine learning model is obtained through supervised training based on a normal sample data set of the energy storage battery unit in the normal charge and discharge state and an arc pulling abnormal sample data set in the arc pulling state; and the normal sample data set is obtained by processing the terminal current and terminal voltage and the sound signal collected in the normal charge and discharge state; the arc pulling abnormal sample data set is obtained by processing the terminal current and terminal voltage and the sound signal collected in the arc pulling state.

[0034] Optionally, before the step of processing the sound signal into a noise feature data set, it further includes:

[0035] Filter the environmental noise and the background sound of the device operation in the sound signal to obtain an effective sound signal to be identified;

[0036] Process the effective sound signal to be identified to obtain a noise feature data set.

[0037] Optionally, the steps for the battery management system to judge whether the terminal current and terminal voltage have the characteristics of electric arc pulling include:

[0038] Perform time-domain feature analysis and frequency-domain feature analysis on the terminal current and terminal voltage respectively, so as to determine whether the terminal current and terminal voltage have the characteristics of electric arc pulling in the time domain or the frequency domain.

[0039] Optionally, the steps for the battery management system to judge whether the terminal current and terminal voltage have the characteristics of electric arc pulling include:

[0040] The battery management system obtains the charge and discharge state of the energy storage battery unit;

[0041] Input the charge and discharge state, terminal current, and terminal voltage into a pre-trained arcing electrical signal classification model, and perform classification and identification by the arcing electrical signal classification model;

[0042] When the arcing probability threshold determined by the arcing electrical signal classification model is greater than a preset threshold, it is determined that an arcing electrical feature appears.

[0043] According to another aspect of the present invention, there is also provided a computer program product, including a computer program, which when executed by a processor, implements the steps of any one of the above-mentioned arcing detection methods for parallel energy storage battery units.

[0044] According to still another aspect of the present invention, there is also provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the steps of any one of the above-mentioned arcing detection methods for parallel energy storage battery units.

[0045] According to still another aspect of the present invention, there is also provided a computer device, which includes a memory, a processor, and a computer program stored on the memory, and the processor executes the computer program to implement the steps of any one of the above-mentioned arcing detection methods for parallel energy storage battery units.

[0046] For the arcing detection method of the parallel energy storage battery unit of the present invention, the energy storage battery units are connected in parallel, and the battery management system of the energy storage battery unit is configured to detect the terminal current and terminal voltage of the energy storage battery unit. The microphone can detect the sound at its location. During the detection process, the battery management system determines whether the terminal current and terminal voltage have arcing electrical features; if so, it controls the energy storage battery unit to disconnect to avoid more serious impacts caused by arcing. This method further performs arcing identification in the dimension of sound signals based on the sound signals detected by the microphone during the same period when arcing electrical features appear, so as to make a judgment using multi-dimensional data, improve the identification accuracy, and reduce false detections and missed detections.

[0047] Furthermore, for the arcing detection method of the parallel energy storage battery unit of the present invention, a machine learning model is used to perform feature identification on sound signals, terminal current, and terminal voltage. The machine learning model is obtained through supervised training using a normal sample data set in the normal charge and discharge state and an arcing abnormal sample data set in the arcing state, with a fast convergence speed and high identification accuracy.

[0048] Furthermore, the electric arc pull detection method for the parallel energy storage battery unit of the present invention provides an accurate basis for arc pull judgment by virtue of the connection characteristics of the parallel energy storage battery unit and the charging and discharging electrical signal characteristics, can accurately capture and analyze the abnormal electric arc pull, greatly improves the accuracy and timeliness of fault detection, and provides a solid technical support for ensuring the efficient and stable operation of the parallel energy storage battery system.

[0049] Furthermore, through the detailed description of the specific embodiments of the present invention in conjunction with the following drawings, those skilled in the art will more clearly understand the above and other objects, advantages and features of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Some specific embodiments of the present invention will be described in detail hereinafter with reference to the drawings in an exemplary but not restrictive manner. The same reference numerals in the drawings denote the same or similar components or parts. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings:

[0051] Figure 1 is a schematic diagram of a parallel energy storage battery unit according to an embodiment of the present invention;

[0052] Figure 2 is a connection schematic diagram of a parallel energy storage battery unit according to an embodiment of the present invention;

[0053] Figure 3 is a schematic diagram of the electric arc pull detection method for a parallel energy storage battery unit according to an embodiment of the present invention;

[0054] Figure 4 is a schematic diagram of the processing procedure in the electric arc pull detection method for a parallel energy storage battery unit when the sound signal does not have the arc pull sound characteristics according to an embodiment of the present invention;

[0055] Figure 5 is a schematic diagram of the processing procedure after determining that an electric arc pull event has occurred in the electric arc pull detection method for a parallel energy storage battery unit according to an embodiment of the present invention;

[0056] Figure 6 is a schematic diagram of a computer program product according to an embodiment of the present invention;

[0057] Figure 7 is a schematic diagram of a computer-readable storage medium according to an embodiment of the present invention;

[0058] Figure 8 is a schematic block diagram of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0059] Those skilled in the art should understand that the embodiments described below are only a part of the embodiments of the present invention, rather than all the embodiments of the present invention. This part of the embodiments is intended to explain the technical principles of the present invention and is not intended to limit the protection scope of the present invention. Based on the embodiments provided by the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts should still fall within the protection scope of the present invention.

[0060] The energy storage battery cluster is formed by combining multiple energy storage battery units (energy storage battery monomers) in series and parallel to form a battery assembly with specific voltage, capacity, and power output capabilities. Figure 1 It is a schematic diagram of the parallel energy storage battery unit 110 according to an embodiment of the present invention. The energy storage battery unit 110 provides higher energy storage through parallel connection. Each parallel energy storage battery unit 110 has a battery management system 111 (Battery Management System, abbreviated as BMS), a direct current - direct current (DC / DC) conversion module 112, and a connection component 113.

[0061] The BMS 111 is used to collect the operating state of the energy storage battery unit 110 and perform state adjustment and battery state estimation of the parallel energy storage battery unit 110 according to the operating state. The operating state of the energy storage battery unit 110 may include voltage, current, temperature, etc.; the battery state estimation may include: state of charge (SOC) estimation, state of health (SOH) estimation; the state adjustment may include: equalization control, current control, on - off control, battery protection (such as over - charge over - discharge protection, over - current protection, short - circuit protection), etc.

[0062] The DC / DC conversion module 112 is controlled by the BMS 111, converts the electrical energy of the battery into a corresponding output voltage, and adjusts the output current. In some embodiments, the DC / DC conversion module 112 can convert the battery voltage from low voltage (typical value 51.2V) to high voltage (typical value 400V / 850V). The input end of the DC / DC conversion module 112 is connected to the battery body, and its output end serves as the external connection interface of the energy storage battery unit 110. The energy storage battery units are connected in parallel through electrical cables (an optional equivalent alternative is an electrical busbar).

[0063] For lithium iron phosphate battery cells, the voltage of a single energy storage battery unit is generally 51.2V. After conversion by the DC / DC conversion module 112, the voltage can reach 400V or 850V. As the number of parallel connections increases, the converging current increases accordingly. By adjusting the charge and discharge current through the BMS 111 and the DC / DC conversion module 112, balanced charge and discharge between the parallel energy storage battery units 110 can be achieved.

[0064] The output voltage of the parallel energy storage battery unit 110 is equal due to the parallel connection, and its output current is regulated by the BMS111 and the DC / DC conversion module 112. Compared with the series energy storage battery unit, the above-mentioned electrical connection characteristics of the parallel energy storage battery unit are more flexible in adjustment and control, and a single energy storage battery unit 110 can be flexibly replaced. Therefore, a single energy storage battery unit 110 generally adopts a detachable connection component 113 to connect the input end of the DC / DC conversion module 112, and the connection component 113 can use a connection component that is easy to disassemble and assemble, such as a quick plug connection terminal. During long-term operation, due to insulation aging, structural shaking and other reasons, the above-mentioned detachable plug-in terminal will not be tight. When charging and discharging, DC arcing will occur at the terminals that are not tightly connected. Compared with AC arcing, DC arcing does not pass through the zero point and will always maintain the arcing state. The energy of the series arcing is large, and the spacing between the energy storage battery units 110 is small. The arc burning is easy to affect other batteries around, which is very likely to cause serious failures.

[0065] The DC end of the inverter 30 is connected to the parallel end (electrical bus) of the parallel energy storage battery unit, and the AC end is connected to the power load or the AC power grid, and is used to convert the overall output power of the parallel energy storage battery unit 110 into AC power to supply power to the power load or transmit it to the power grid.

[0066] In some schemes, the DC end of the inverter 30 can be configured with a fault detection function, such as abnormal detection of electrical cables (electrical busbars) and connection terminals of the inverter 30, such as arc abnormality detection. If arc occurs, the inverter 30 cuts off the input. Therefore, the arc detection of the inverter 30 is for the circuit of the entire energy storage system. Once it is determined that arc occurs, the inverter 30 is protected and the parallel energy storage battery unit 110 is cut off as a whole. This will affect the entire battery energy storage system, and if a misjudgment occurs, it will also cause serious economic losses.

[0067] In comparison, the parallel energy storage battery unit 110 of this embodiment further detects arcing inside the energy storage battery unit 110 based on the abnormality detection function of the inverter 30, detects arcing inside a single energy storage battery unit 110, and uses the advantages of the parallel connection method to perform fault processing and promptly handle arcing abnormalities.

[0068] In some optional embodiments, the energy storage battery unit 110 also improves the connection component 113 so that the BMS 111 can process according to the abnormal detection situation. Figure 2It is a schematic connection diagram of the parallel energy storage battery unit 110 according to an embodiment of the invention. The connection component 113 between the energy storage battery unit 110 and the DC / DC conversion module 112 includes a plurality of connection terminals, and each connection terminal is connected with a switching switch K1, K2, K3. The switching switches K1, K2, K3 are controlled by the BMS111 to be connected or disconnected. Figure 2 Three mutually parallel connection terminals are shown in Figure 2 , and the switching switches K1, K2, K3 are respectively connected to the three groups of connection terminals. When normally connected, any connection terminal is connected, and power conduction can be achieved. When the BMS111 determines that an arcing situation occurs, the connection of the corresponding connection terminal to the DC / DC conversion module 112 can be disconnected, and the connection is switched to another parallel connection terminal, so as to solve the arcing problem through the switching of the connection terminals.

[0069] The connection terminals may include a main connection terminal 140 and a spare connection terminal 141. The energy storage battery unit 110 defaults to using the main connection terminal 140 to connect to the DC-DC conversion module 112. In the case of abnormal arcing, the spare connection terminal 141 is used for connection and switching. For example, to make the structure of the connection terminals more compact. The connection component 113 may include one main connection terminal 140 and two spare connection terminals 141. The power transmission capacity of the main connection terminal 140 is greater than that of a single spare connection terminal 141. During normal use, the main connection terminal 140 is default to be connected, and the spare connection terminal 141 is only connected when the connection of the main connection terminal 140 is unreliable and arcing occurs. This connection method can facilitate quick switching, the action is more timely, and it is also convenient for the connection component 113 to be more compact.

[0070] There are more or less tolerances in the manufacturing process of the connection component 113, which may cause virtual connection at the connection point. During long-term operation, the connection reliability decreases and arcing is likely to occur. The position of the microphone 133 can be set according to the structure of the energy storage battery unit and the internal sound transmission path.

[0071] The microphone 133 can be arranged near the connection component 113 and is used to detect the sound signal at the connection terminal to obtain a sound detection signal. The sound detection signal reflects the sound inside the energy storage battery unit 110.

[0072] The BMS111 can also be signal-connected to the inverter 30 to obtain the overall charge and discharge state of the parallel energy storage battery unit 110 from the inverter 30 (including but not limited to whether it is charging, whether it is discharging, charge and discharge current, charge and discharge voltage, charge and discharge power, etc.).

[0073] Figure 3Schematic diagram of the electrical arc detection method for a parallel energy storage battery unit according to an embodiment of the present invention. The electrical arc detection method for the parallel energy storage battery unit generally may include:

[0074] Step S301: The battery management system determines whether the terminal current and terminal voltage have electrical arc characteristics. The determination process of this step can be carried out in the time domain and / or frequency domain. The time-domain characteristics analysis and frequency-domain characteristics analysis are respectively performed on the terminal current and terminal voltage to determine whether the terminal current and terminal voltage have electrical arc characteristics in the time domain or frequency domain. The terminal current and terminal voltage can be detected by the detection units inside the energy storage battery unit, and these detection units are signal-connected to the battery management system.

[0075] The time-domain analysis detection method is based on the detection of signal waveform changes such as current and voltage, which is simple and fast. The frequency-domain analysis detection method performs Fourier transform analysis on the eigenvalue spectrum to obtain frequency-domain characteristics, thereby identifying whether there is an arc.

[0076] Both the time-domain analysis detection method and the frequency-domain analysis detection method have certain defects and the discrimination is not accurate enough. Therefore, in this embodiment, the time-domain analysis detection method and the frequency-domain analysis detection method are used with relatively broad judgment thresholds for the preliminary identification of electrical arc characteristics.

[0077] Alternatively, the determination process of step S301 can also be implemented by machine learning. For example, the battery management system obtains the charge and discharge state of the energy storage battery unit; inputs the charge and discharge state, terminal current, and terminal voltage into a pre-trained arc electrical signal classification model, and the arc electrical signal classification model performs classification and identification; when the arc probability threshold determined by the arc electrical signal classification model is greater than the preset threshold, it is determined that electrical arc characteristics occur. The charge and discharge state is used to indicate whether the energy storage battery unit is charging or discharging, the target voltage, current, etc. of the charge and discharge. The arc electrical signal classification model can use classification models such as convolutional neural network (CNN) and support vector machine (SVM). The terminal current and terminal voltage can be processed into input features such as spectral energy, mutation frequency, signal entropy, etc. as the input of the arc electrical signal classification model. The arc electrical signal classification model is also used for preliminary classification to perform the preliminary identification of electrical arc characteristics. The above threshold (probability threshold) can be set relatively small, such as 50%, that is, when judging the possibility of arc abnormality, it is determined that electrical arc characteristics occur.

[0078] Step S302: When the battery management system determines that the terminal current and terminal voltage have electrical arc characteristics, the battery management system controls the energy storage battery unit to disconnect. The battery management system can achieve the connection and disconnection of the energy storage battery unit by controlling the switching switch.

[0079] Step S303: Collect the sound signal detected by the microphone during the same period when the electric arc drawing feature appears.

[0080] Step S304: Identify whether the sound signal has the sound characteristics in the state of electric arc drawing.

[0081] Step S305: If the sound signal has the sound characteristics in the state of electric arc drawing, it is determined that an electric arc drawing event occurs in the energy storage battery unit. The step of identifying whether the sound signal has the sound characteristics in the state of electric arc drawing may specifically include: processing the terminal current and terminal voltage into an electrical feature data set; processing the sound signal into a noise feature data set; inputting the electrical feature data set and the noise feature data set into a pre-trained machine learning model; and performing identification and judgment by the machine learning model to determine whether an arc occurs. By adding the identification of the sound signal, the identification can be more accurate, making up for the defects of single-terminal voltage and terminal current judgment.

[0082] The above machine learning model is obtained through supervised training based on a normal sample data set of the energy storage battery unit in the normal charge and discharge state and an arc abnormal sample data set in the arc state; and the normal sample data set is obtained by processing the terminal current, terminal voltage, and sound signal collected in the normal charge and discharge state; the arc abnormal sample data set is obtained by processing the terminal current, terminal voltage, and sound signal collected in the arc state.

[0083] In some embodiments, the normal sample data set and the arc abnormal sample data set also respectively include charge and discharge state data. Considering the specific application scenario and layout difficulty, the machine learning model can adopt a neural network model, such as using a convolutional neural network (CNN). By processing the detection signal into data with a grid-like topological structure, key features are extracted, which not only effectively reduces the model parameters but also improves the network's grasping and recognition speed of key information. The convolutional neural network uses processes such as convolution, batch normalization, activation, and pooling to form a convolutional layer for key feature extraction, and then sends it to a fully connected network for identification and reasoning.

[0084] The method of the above embodiment uses a machine learning model to identify the detection signal. The machine learning model is trained based on a first detection signal sample data set in a reliable connection state between battery packs and a second detection signal sample data set in an arc state, and can accurately identify DC arc abnormalities with high identification accuracy, which is conducive to wide application. Compared with the arc electrical signal classification model used in the preliminary determination of electric arc drawing characteristics, the machine learning model has more input data dimensions and a higher judgment threshold.

[0085] Figure 4It is a schematic diagram of the processing procedure in the electric arc detection method for a parallel energy storage battery unit according to an embodiment of the present invention when the sound signal does not have the characteristics of an arc sound. The electric arc detection method for the parallel energy storage battery unit may further include, when it is determined that no electric arc event has occurred in the energy storage battery unit:

[0086] Step S401: Determine whether the terminal current and terminal voltage exceed their respective preset limits; the limits can be set according to the normal current value and voltage value of the energy storage battery unit in the current charge and discharge state. For example, when setting the current limit, if the terminal current signal exceeds the current limit, it is determined that the abnormality is serious.

[0087] Step S402: If either the terminal current or the terminal voltage exceeds the preset limit, maintain the disconnected state, and the battery management system reports the over-limit abnormal state. The judgment of the set limit can determine the severity of the fault abnormality. If the limit is not exceeded, through the self-recovery process, an attempt can be made to restore the energy storage battery unit to normal.

[0088] When both the terminal current and the terminal voltage do not exceed their respective preset limits, it may further include:

[0089] Step S403: The battery management system performs a reconnect and recovery operation for a set duration of recovery connection. The set duration only needs to meet the requirements for the acquisition and processing of the terminal current and terminal voltage. The duration is relatively short, generally about 10 ms, so that even if an arc occurs again after reconnecting, it will not cause serious consequences.

[0090] Step S404: Within the set duration, re-acquire the terminal current and terminal voltage and perform abnormal detection. This reconnect and recovery operation can automatically eliminate some abnormalities. If the abnormality fails to be eliminated, the battery is completely disconnected, and an abnormal prompt message is output.

[0091] Figure 5 It is a schematic diagram of the processing procedure in the electric arc detection method for a parallel energy storage battery unit according to an embodiment of the present invention after it is determined that an electric arc event has occurred in the energy storage battery unit. In an embodiment where the energy storage battery unit is further provided with a spare output terminal, the method of this embodiment may further include, when it is determined that an electric arc event has occurred in the energy storage battery unit:

[0092] Step S501: The battery management system disconnects the connection between the main connection terminal and the DC-DC conversion module.

[0093] Step S502: Use the spare output terminal to connect to the DC-DC conversion module, so that the energy storage battery units are re-paralleled;

[0094] Step S503: Re-acquire the terminal current and the terminal voltage and perform abnormal detection.

[0095] After the steps of re-acquiring the terminal current and terminal voltage and performing anomaly detection, the following steps may further be included:

[0096] Step S504, if the result of the anomaly detection is normal operation, the battery management system restores the connection between the main connection terminal and the DC-DC conversion module, so as to simultaneously use the main connection terminal and the spare connection terminal to connect to the DC-DC conversion module;

[0097] Step S505, re-acquire the terminal current and terminal voltage again and perform anomaly detection. If the result of the anomaly detection is still normal operation, disconnect the connection between the spare output terminal and the DC-DC conversion module. Thus, the connection state of only the main connection terminal is restored.

[0098] Step S506, detect whether the electric arc pulling event of the energy storage battery unit is eliminated;

[0099] Step S507, if it is eliminated, reconnect using the main connection terminal;

[0100] Step S508, if it is not eliminated, disconnect the main connection terminal and use the spare output terminal for connection.

[0101] Through the above switching process, some loose connections causing arcing can be eliminated, thereby eliminating arcing. If the arcing anomaly fails to be eliminated, the battery is completely disconnected, and an arcing prompt message is output.

[0102] After determining that arcing occurs between the energy storage battery units, the following steps may further be included: preprocessing the terminal voltage, terminal current, and sound signal corresponding to the period when arcing is likely to occur between the energy storage battery units, and using the preprocessed results to iteratively train a machine learning model. Through continuous iterative training of the machine learning model, the recognition can be made more accurate.

[0103] When using the machine learning model for recognition and data training, the processing process of the sound signal may include: performing spectral analysis on the sound signal; making a preliminary judgment on the sound signal according to the spectral analysis result to determine whether the sound signal is an ambient sound; if it is not an ambient sound, perform the steps of preprocessing the sound signal.

[0104] In some embodiments, before the step of processing the sound signal into a noise feature data set, the following steps may further be included: filtering out the ambient noise and the background sound of device operation in the sound signal to obtain an effective sound signal to be recognized; processing the effective sound signal to be recognized to obtain the noise feature data set. After the parallel energy storage battery units are installed, the detection of ambient sound and the background sound of device operation can be continuously performed, so as to filter out the noise of the device itself and ambient noise in the normal state for subsequent use, thereby forming an arcing sound signal recognition library. Through continuous correction, the recognition library can be made more accurate and comprehensive.

[0105] The preprocessing process of the sound signal is used to extract features such as the pitch (frequency level), volume (loudness), and timbre (spectral distribution) of the signal. In this embodiment, the noise feature dataset obtained from the preprocessing process can reflect features such as the frequency and energy density distribution of the sound signal. For example, the noise feature dataset can be in the form of a feature vector that is more suitable for use in a convolutional neural network.

[0106] The process of preprocessing the sound signal may include: pre-emphasis, framing and windowing, and discrete Fourier transform to generate a spectrogram. Among them, pre-emphasis is used to suppress noise in non-set frequency bands and highlight noise in the set frequency band. The above-mentioned set frequency band can be set according to the sound frequency emitted by the arc. For example, for the discharge sound of some DC arc strikes, the high-energy part is distributed around 10 kHz. A filter with a set transfer function can be used to achieve pre-emphasis, effectively highlighting the characteristics of the arc discharge sound. Framing is used to organize the sound data of multiple consecutive acquisition points into one frame. The frame length represents the time length of the signal, and the sound characteristics are considered stationary within the frame length to enable frame-by-frame analysis. Framing is usually achieved by weighting with a shiftable window function. The length of the window function is called the window length, and the signal values outside the window length are all 0. Windowing can maintain the integrity of the speech signal and prevent leakage of signal spectral energy, making the non-periodic sound signal exhibit the characteristics of a periodic signal.

[0107] In this embodiment, by comparing the pushing of different window functions, the Hamming window can be used, making the sidelobes smaller, and the spectrum after windowing can be closer to the original signal spectrum. The discrete Fourier transform analyzes converts the signal from the time domain to the frequency domain. A spectrogram is an application of the short-time Fourier transform (STFT), which converts the one-dimensional speech waveform into three-dimensional image features, describing the sound information from three aspects: time, frequency, and energy density. The spectrogram retains more balanced and complete sound feature information on the basis of visualizing speech. Its generation process can be: the sound signal after framing and windowing is divided into multiple frames, converted into a frequency amplitude diagram through Fourier transform, and the frequency amplitude diagram is rotated and mapped to the gray values of pixel points to form a spectrogram. The preprocessed sound signal can contain features in three dimensions: time, frequency, and energy distribution, extracting more complete sound frequency distribution and energy density information, and enabling the accurate identification of the arc strike phenomenon with the help of a convolutional neural network.

[0108] Considering reducing the processing difficulty of the sound signal, and there are significant differences in the spectra between the ambient sound and the sound of the energy storage battery unit, through simple spectral analysis of the sound signal, obvious ambient sounds can be excluded through spectral analysis and judgment, reducing the data processing pressure on the machine learning model.

[0109] Considering that the sound characteristics of energy storage battery units in different packaging forms also vary due to differences in sound transmission paths and sound transmission media. The solution of this embodiment can train machine learning models for energy storage battery units of different specifications respectively, and the obtained machine learning models are smaller in scale and more targeted.

[0110] A specific recognition process of a machine learning model may include: obtaining a sound signal during the charging and discharging process of an energy storage battery unit; performing noise spectrum analysis on the sound signal to identify environmental noise and arcing noise; for arcing noise, recording the electrical signal simultaneous with the arcing noise, where the electrical signal includes terminal voltage, terminal current, and other signals such as power and harmonics obtained by calculating and processing the terminal voltage and terminal current; using the arcing noise and the electrical signal to form an arcing recognition library; collecting the electrical signal to be recognized and the sound to be recognized during the operation of the energy storage battery unit; if it is determined that there may be an arcing anomaly according to the time-domain and frequency-domain characteristics of the electrical signal to be recognized, using the machine learning model to perform feature recognition on the electrical signal to be recognized and the sound to be recognized; if the recognition is successful, it is determined that the energy storage battery unit has an arc, and then use BM to control the switching switch to realize the processing of the arcing anomaly.

[0111] This embodiment also provides a computer program product 80, a computer-readable storage medium 820, and a computer device 830. Figure 6 It is a schematic diagram of a computer program product 80 according to an embodiment of the present invention. Figure 7 It is a schematic diagram of a computer-readable storage medium 820 according to an embodiment of the present invention. Figure 8 It is a schematic block diagram of a computer device 830 according to an embodiment of the present invention.

[0112] The computer program product 80 includes a computer program 811, and when the computer program 811 is executed by a processor 831, it implements the steps of the arcing detection method between energy storage battery units in the stacked energy storage system described above. The computer-readable storage medium 820 stores the above computer program 811, and when the computer program 811 is executed by a processor 831, it implements the steps of the arcing detection method between energy storage battery units in the stacked energy storage system described in any of the above embodiments. The computer device 830 may include a memory 832, a processor 831, and a computer program 811 stored on the memory 832 and running on the processor 831.

[0113] The computer program 811 for performing the operations of the present invention may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, configuration data of an integrated circuit, or source code or object code written in any combination of one or more programming languages and procedural programming languages.

[0114] The computer program 811 may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the latter case, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider via the Internet). In some embodiments, to perform aspects of the present invention, an electronic circuit, including, for example, a programmable logic circuit, a field-programmable gate array (FPGA), or a programmable logic array (PLA), may execute computer-readable program instructions by utilizing the state information of the computer-readable program instructions to personalize the electronic circuit.

[0115] For the description of this embodiment, the computer program product 80 is a related product that includes the computer program 811.

[0116] For the description of this embodiment, the computer-readable storage medium 820 is a tangible device capable of retaining and storing the computer program 811, which may be any device that can contain, store, communicate, propagate, or transport the program 811 for use by or in connection with an instruction execution system, apparatus, or device. More specific examples (non-exhaustive list) of the computer-readable storage medium 820 include the following: portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), static random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital versatile disc (DVD), memory stick, floppy disk, mechanically encoded device, and any suitable combination of the above.

[0117] At this point, those skilled in the art should recognize that although numerous exemplary embodiments of the present invention have been shown and described in detail herein, many other variations or modifications that conform to the principles of the present invention can still be directly determined or derived from the disclosed content of the present invention without departing from the spirit and scope of the present invention. Therefore, the scope of the present invention should be understood and recognized as covering all such other variations or modifications.

Claims

1. An electric arc detection method for a parallel energy storage battery unit, wherein a plurality of energy storage battery units are connected in parallel, and a microphone is provided inside. The battery management system of the energy storage battery unit is configured to detect the terminal current and terminal voltage of the energy storage battery unit, and the electric arc detection method includes: Judging, by the battery management system, whether the terminal current and terminal voltage have electric arc characteristics; If so, controlling, by the battery management system, the energy storage battery unit to disconnect; Collecting the sound signal detected by the microphone during the same period when the electric arc characteristics appear; Identifying whether the sound signal has the sound characteristics in the state of electric arc occurrence; If so, determining that an electric arc event occurs in the energy storage battery unit.

2. The electric arc detection method for a parallel energy storage battery unit according to claim 1, further including, in the case of determining that no electric arc event occurs in the energy storage battery unit: Judging whether the terminal current and the terminal voltage exceed their respective preset limits; If any of the terminal current and the terminal voltage exceeds the preset limit, maintaining the disconnected state, and reporting the over-limit abnormal state by the battery management system.

3. The electric arc detection method for a parallel energy storage battery unit according to claim 2, further including, in the case that both the terminal current and the terminal voltage do not exceed their respective preset limits: Performing, by the battery management system, a reconnect recovery operation for a set duration of recovery connection; Within the set duration, re-collecting the terminal current and the terminal voltage and performing abnormal detection.

4. The electric arc detection method for the parallel energy storage battery unit according to claim 2, wherein, The energy storage battery unit is further provided with a DC-DC conversion module and a connection component. The connection component includes a main connection terminal and a standby connection terminal. The energy storage battery unit defaults to use the main connection terminal to connect the DC-DC conversion module, and in the case of determining that an electric arc event occurs in the energy storage battery unit, further including: Disconnecting, by the battery management system, the connection between the main connection terminal and the DC-DC conversion module; Using the standby output terminal to connect the DC-DC conversion module, so that the energy storage battery unit is re-connected in parallel; Re-collecting the terminal current and the terminal voltage and performing abnormal detection.

5. The electric arc detection method for a parallel energy storage battery unit according to claim 4, wherein, After the step of re-collecting the terminal current and the terminal voltage and performing abnormal detection, further including: If the result of the abnormal detection is normal operation, reconnecting, by the battery management system, the main connection terminal and the DC-DC conversion module, so as to connect the main connection terminal and the standby connection terminal to the DC-DC conversion module simultaneously; Re-recollecting the terminal current and the terminal voltage again and performing abnormal detection. If the result of the abnormal detection is still normal operation, disconnecting the connection of the standby output terminal; Detecting whether the electric arc event of the energy storage battery unit is eliminated. If it is eliminated, reconnecting using the main connection terminal; if not, disconnecting the main connection terminal and using the standby output terminal for connection.

6. The electric arc detection method for the parallel energy storage battery unit according to claim 1, wherein, The step of identifying whether the sound signal has the sound characteristics in the state of electric arc occurrence includes: Processing the terminal current and the terminal voltage into an electrical characteristic data set; Process the sound signal into a noise feature data set; Input the electrical feature data set and the noise feature data set into a pre-trained machine learning model; Have the machine learning model perform identification and judgment to determine whether arcing occurs.

7. The method for detecting electric arcing of a parallel energy storage battery unit according to claim 6, wherein, The machine learning model is obtained through supervised training based on a normal sample data set of the energy storage battery unit in a normal charge and discharge state and an arcing abnormal sample data set in an arcing state; And the normal sample data set is obtained by processing the terminal current, terminal voltage, and sound signal collected in the normal charge and discharge state; The arcing abnormal sample data set is obtained by processing the terminal current, terminal voltage, and sound signal collected in the arcing state.

8. The electric arc detection method for the parallel energy storage battery unit according to claim 1, wherein, Before the step of processing the sound signal into a noise feature data set, it further includes: Filter out ambient noise and device operation background noise in the sound signal to obtain an effective sound signal to be identified; Process the effective sound signal to be identified to obtain the noise feature data set.

9. The electric arc detection method for a parallel energy storage battery unit according to claim 1, wherein, The step of having the battery management system determine whether the terminal current and terminal voltage exhibit electric arcing characteristics includes: Perform time-domain feature analysis and frequency-domain feature analysis on the terminal current and the terminal voltage respectively, so as to determine whether the terminal current and the terminal voltage have the electric arcing characteristics in the time domain or the frequency domain.

10. The electric arc pull detection method for the parallel energy storage battery unit according to claim 1, wherein, The step of having the battery management system determine whether the terminal current and terminal voltage exhibit electric arcing characteristics includes: Have the battery management system obtain the charge and discharge state of the energy storage battery unit; Input the charge and discharge state, the terminal current, and the terminal voltage into a pre-trained arcing electrical signal classification model, and have the arcing electrical signal classification model perform classification and identification; In the case where the arcing probability threshold determined by the arcing electrical signal classification model is greater than a preset threshold, it is determined that the electric arcing characteristics occur.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method for detecting electric arcing of a parallel energy storage battery unit according to any one of claims 1 to 10.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that When the computer program is executed by a processor, it implements the steps of the method for detecting electric arcing of a parallel energy storage battery unit according to any one of claims 1 to 10.

13. A computer device, comprising a memory, a processor, and a computer program stored on the memory, characterized in that, The processor executes the computer program to implement the steps of the method for detecting electric arcing of a parallel energy storage battery unit according to any one of claims 1 to 10.

Images