Direct current arc discharge detection system and method for energy storage system

Abstract

The invention discloses a direct current arc discharge detection system and method for an energy storage system, and relates to the field of direct current arc discharge detection.In the system, a main control module is in communication connection with a battery management system and is used for obtaining operation parameters uploaded by the battery management system in real time; leakage current value signals, output by the data processing module, of all the leakage current sensors are received; the main control module is used for generating a plurality of arc discharge judgment results by using a preset multi-dimensional signal fusion analysis algorithm for each leakage current sensor based on the corresponding leakage current value signal and the operation parameter, and sending the arc discharge judgment results to the main control module; comprehensively judging whether a battery pack or a direct current bus corresponding to the leakage current sensor has an arc discharge fault according to a preset arc discharge judgment rule; a multi-dimensional signal fusion analysis algorithm is utilized to generate a plurality of arc discharge judgment results to realize fusion judgment of multi-source signals, and compared with a single-signal detection method, multi-dimensional fusion remarkably improves the recognition capability of low-amplitude and short-time high-frequency arc discharge signals, and reduces the risk of misjudgment.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of direct current arc detection, in particular to a direct current arc detection system and method for energy storage system. BACKGROUND

[0002] Compared with photovoltaic systems, arc detection in energy storage systems faces greater challenges. On the one hand, the frequency domain characteristics of arc signals are weak and easily covered by system noise, and traditional detection methods such as spectrum analysis and waveform analysis have low sensitivity and slow response; on the other hand, the structure of the energy storage system is complex, the components are diverse, and the interface standards are not unified, which further increases the technical difficulty of multi-source signal fusion and intelligent judgment. At present, the arc recognition technology of energy storage is still in the early exploration stage, and only a few companies (such as Yangguang Power) have launched active detection schemes based on converter integration, but the overall technology level is not mature, and there is a lack of industry unified standards and efficient multi-source signal fusion algorithm support. Therefore, how to build an arc detection system with high sensitivity, low false alarm rate and fast response capability has become a key technical problem to be solved in the field of energy storage safety.

[0003] In the prior art, most schemes still rely on detection mechanisms centered on filtering and amplification, such as extracting alternating disturbance components through differential sampling and band-pass filtering, and making judgments based on the amplification characteristics of arc voltage. Such schemes have certain advantages in structural integration, but when dealing with low-amplitude, short-time high-frequency arc signals in energy storage systems, they have problems such as insufficient detection sensitivity, low response accuracy, and weak noise immunity, especially in strong noise background, the misjudgment rate is high, which is difficult to meet the demand of early warning and accurate positioning of arc events in actual application. Therefore, the development of a high-precision, real-time arc detection system not only helps to improve the intrinsic safety of energy storage systems, but also provides strong support for enterprises to establish unique technical competitive advantage. SUMMARY

[0004] In order to realize accurate judgment of arc fault in energy storage system, the present application proposes a direct current arc detection system for energy storage system, the energy storage system comprising an electric energy conversion module PCS, a positive direct current bus, a negative direct current bus and a plurality of battery packs; the positive terminal of the PCS is connected with the positive terminals of the battery packs through the positive direct current bus, and the negative terminal of the PCS is connected with the negative terminals of the battery packs through the negative direct current bus, forming a main circuit; the direct current arc detection system comprises: a leakage current detection module, comprising a leakage current sensor sleeved on the positive direct current bus and the negative direct current bus, and a leakage current sensor sleeved on the positive and negative connection lines of each battery pack, for collecting the leakage current signals of the main circuit and each battery pack; The data processing module is configured to preprocess the leakage current signals collected by the leakage current detection module, obtain leakage current value signals corresponding to each leakage current sensor, and output the leakage current value signals. The master control module is in communication connection with the battery management system, configured to obtain running parameters uploaded by the battery management system in real time, and receive the leakage current value signals of each leakage current sensor output by the data processing module; the master control module is configured to, for each leakage current sensor, generate a plurality of arc drawing judgment results based on the leakage current value signal and the running parameters corresponding to the leakage current sensor, and comprehensively judge whether the battery pack or the DC bus corresponding to the leakage current sensor has an arc drawing fault according to a preset arc drawing judgment rule.

[0005] Further, the data processing module comprises: The high-pass filter unit is configured to filter the leakage current signals collected by the leakage current detection module to obtain high-frequency current signals. The data processing unit is configured to sequentially filter, amplify and analog-to-digital convert the high-frequency current signals to obtain the leakage current value signals.

[0006] Further, the running parameters comprise: The voltage and temperature data of each battery pack; The voltage and temperature data of the DC bus.

[0007] Further, the preset multi-dimensional signal fusion analysis algorithm comprises: The absolute value of the leakage current value signal is accumulated by using a sliding window integral method to obtain an integral value; if the integral value exceeds a first preset threshold, a first judgment result is marked; The leakage current value signal is subjected to frequency spectrum analysis by using a short-time Fourier transform to extract an energy mutation feature of a preset target frequency band; if the energy change rate of the preset target frequency band exceeds a second preset threshold, a second judgment result is marked; If the temperature rising rate of the battery pack or the DC bus exceeds a third preset threshold within a time period corresponding to the first judgment result, a third judgment result is marked; If the voltage of the battery pack or the DC bus deviates from the rated value by a magnitude exceeding a fourth preset threshold within the time period corresponding to the first judgment result, a fourth judgment result is marked.

[0008] Further, the preset arc drawing judgment rule is: The weighted score sum of each judgment result is calculated by using preset weight coefficients corresponding to the judgment results; if the sum exceeds a fifth preset threshold, it is determined that an arc drawing fault exists, and a fault confirmation signal is triggered.

[0009] Further, the preset arc drawing judgment rule is: If only the first determination result is satisfied and the second, third and fourth determination results are not satisfied, a pre-warning state is marked and continuous monitoring is performed; if the first determination result and any one of the second or third or fourth determination results are satisfied at the same time, it indicates that there is an arc fault, and a fault confirmation signal is triggered immediately.

[0010] Further, the master control module comprises: The linkage protection unit in communication connection with the power conversion module PCS is used for sending an alarm code to the battery management system and sending a PWM shutdown instruction to the power conversion module PCS to cut off the connection between the DC bus and the power grid through CAN bus or RS485 communication when the fault confirmation signal is triggered.

[0011] The embodiment of the application further provides a DC arc detection method for an energy storage system, which is applied to the DC arc detection system. The total leakage current sensor sleeved on the positive DC bus and the negative DC bus and the module-level leakage current sensor sleeved on the positive and negative electrode connecting lines of each battery pack are used to collect leakage current signals of the main circuit and each battery pack; The collected leakage current signals are preprocessed to obtain leakage current numerical signals corresponding to each leakage current sensor; Real-time operation parameters of the DC bus and each battery pack monitored by the battery management system are acquired; for each leakage current sensor, based on the leakage current numerical signals and the operation parameters corresponding to the leakage current sensor, a plurality of arc determination results are generated by using a preset multi-dimensional signal fusion analysis algorithm, and whether the battery pack or the DC bus corresponding to the leakage current sensor has an arc fault is comprehensively judged according to a preset arc determination rule; Based on the judgment result of the arc fault, a pre-warning is performed.

[0012] Further, the operation parameters comprise: Voltage and temperature data of each battery pack and voltage and temperature data of the DC bus; The plurality of arc determination results are generated by using the preset multi-dimensional signal fusion analysis algorithm, and specifically: The absolute value of the leakage current numerical signal is accumulated by using a sliding window integral method to obtain an integral value; if the integral value exceeds a first preset threshold value, a first determination result is marked; The leakage current numerical signal is subjected to frequency spectrum analysis by using a short-time Fourier transform to extract an energy mutation feature of a preset target frequency band; if an energy change rate of the preset target frequency band exceeds a second preset threshold value, a second determination result is marked; If the temperature rising rate of the battery pack or the DC bus exceeds a third preset threshold value in a time period corresponding to the first determination result, a third determination result is marked; If the voltage of the battery pack or the DC bus deviates from the rated value by more than the fourth preset threshold in the time period corresponding to the first determination result, a fourth determination result is marked.

[0013] Further, the comprehensive judgment of whether the battery pack or the DC bus corresponding to the leakage current sensor has an arc fault according to the preset arc judgment rule specifically includes: Step 1, judge whether only the first determination result is satisfied, and the second determination result, the third determination result and the fourth determination result are not satisfied; if the condition is satisfied, a pre-warning state is marked, and a continuous monitoring process is entered; if not, step 2 is entered; Step 2, judge whether the first determination result and any one of the second or third or fourth determination result are satisfied at the same time; if the condition is satisfied, an arc fault is determined, and a fault confirmation signal is triggered immediately; The pre-warning based on the judgment result of the arc fault specifically includes: When the fault confirmation signal is triggered, an alarm code is sent to the battery management system through the CAN bus or RS485 communication, and a PWM shutdown instruction is sent to the power conversion module PCS to cut off the connection between the DC bus and the power grid.

[0014] Compared with the prior art, the present application has at least the following beneficial effects: (1) In the present application, the main control module is in communication connection with the battery management system, used for acquiring the running parameters uploaded by the battery management system in real time, and receiving the leakage current value signals of each leakage current sensor output by the data processing module; the main control module is used for generating a plurality of arc determination results based on the leakage current value signals and the running parameters corresponding to each leakage current sensor, and comprehensively judging whether the battery pack or the DC bus corresponding to the leakage current sensor has an arc fault according to the preset arc judgment rule; the present application uses a multi-dimensional signal fusion analysis algorithm to generate a plurality of arc determination results to realize the fusion judgment of multi-source signals, compared with the detection method of single signal (such as only relying on leakage current or frequency spectrum), the multi-dimensional fusion significantly improves the recognition ability of low amplitude, short time high frequency arc signal, and reduces the risk of misjudgment; (2) The present application realizes multi-dimensional feature extraction of arc events through sliding window integration of leakage current signals (capturing high frequency disturbance energy), short time Fourier transform (extracting frequency band energy mutation characteristics), temperature and voltage abnormal correlation analysis (combining thermodynamic characteristics), compared with the detection method of single signal, the multi-dimensional fusion significantly improves the recognition ability of low amplitude, short time high frequency arc signal, and reduces the risk of misjudgment; (3) The present application assigns corresponding weight coefficients to different determination results (such as leakage current integration, spectral energy, temperature rise rate), calculates the weighted score sum of each determination result, and if the sum exceeds the fifth preset threshold, it is determined that there is an arc fault, realizing the fusion determination of multi-dimensional signals and improving the determination accuracy; (4) The present application adopts a real-time linkage protection mechanism, and the main control module communicates with the battery management system (BMS) and the power conversion module PCS through CAN bus or RS485 communication. When the arc fault is determined, the linkage protection unit immediately sends a PWM shutdown instruction to the PCS to disconnect the DC bus from the power grid, with a response time of milliseconds. In addition, the leakage current sensor covers the DC bus total circuit and each battery pack, and combined with the multi-source signal determination result (such as locating the bus fault when only TMR0 is triggered abnormally, and locating the corresponding battery pack fault when TMR1 is triggered), the precise tracing of the fault location is realized. This layered positioning capability can support targeted maintenance, reduce system downtime and improve operation and maintenance efficiency; (5) The present application adopts a layered protection mechanism to balance safety and system availability through preset arc determination rules (such as entering a warning state when only the first determination result is triggered, and executing power-off when multiple conditions are triggered at the same time). The warning stage continuously monitors and records data to provide a basis for subsequent analysis; the fault confirmation stage immediately triggers a protection action to form a closed-loop control of "warning -> confirmation -> protection"; (6) The present application breaks through the limitations of traditional arc detection technology (such as relying on a single signal), and solves the industry pain points of low-amplitude arc signal being easily masked by noise and high misjudgment rate through multi-source signal fusion algorithm (leakage current + spectrum + temperature + voltage) and dynamic determination logic. Compared with the prior art, the system has significant improvements in sensitivity, response speed and anti-interference ability. This technology not only provides a high-precision, low-false alarm active protection scheme for energy storage systems, but also builds a unique multi-dimensional signal fusion algorithm and layered determination rules to establish a technical barrier for enterprises in the field of energy storage safety, promotes the industry from "passive protection" to "active early warning", and has significant industrial application value and market competitiveness. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 A DC arc detection system module diagram for an energy storage system in an embodiment of the present application; Figure 2 An installation structure diagram of an energy storage system and a leakage current sensor in an embodiment of the present application; Figure 3 A current signal superimposition diagram under normal and fault conditions in an embodiment of the present application; Figure 4 is Figure 3 A corresponding integral value curve diagram. DETAILED DESCRIPTION

[0016] The following is a specific embodiment of the present application and further describes the technical solutions of the present application in conjunction with the drawings, but the present application is not limited to these embodiments.

[0017] Example one With the widespread deployment of new energy storage systems, the voltage and current levels of direct current systems continue to rise, and the electrical structure is becoming increasingly complex, which puts higher requirements on the safety and reliability of system operation. As a key device in the storage system, the DC cabinet works in a high-voltage and high-current environment for a long time, and is easily affected by factors such as loose connection and insulation deterioration during operation, thereby causing arc discharge. Arcing is a sustained discharge behavior caused by high voltage breakdown of gas between conductors, which has characteristics such as high temperature, large energy, and long duration. In particular, in a direct current system, since the current does not have the natural zero-crossing characteristic, the arc is difficult to extinguish itself and is extremely easy to cause electrical fire and thermal runaway accidents. A considerable part of the major safety accidents of storage systems occurring worldwide is directly caused by arc, which has become one of the core risks restricting the large-scale safe application of storage systems.

[0018] Therefore, in order to accurately determine the arc fault in the storage system, as shown in Figure 1 , the present application proposes a DC arc detection system for a storage system, the storage system comprising an electric energy conversion module PCS, a positive DC bus, a negative DC bus, and a plurality of battery packs; the positive terminal of the PCS is connected to the positive terminals of the battery packs through the positive DC bus, and the negative terminal of the PCS is connected to the negative terminals of the battery packs through the negative DC bus, forming a main loop; the DC arc detection system comprises: a leakage current detection module, comprising a leakage current sensor sleeved on the positive DC bus and the negative DC bus, and a leakage current sensor sleeved on the positive and negative connection lines of each battery pack, for collecting the leakage current signals of the main loop and each battery pack; As shown in Figure 2 , the system sets a leakage current sensor, i.e. a total leakage current detection point TMR0, on the DC bus main loop output by the electric energy conversion module PCS, and respectively arranges TMR1, TMR2, and TMR3 sensors on the positive and negative connection lines of each battery pack (such as PACK1, PACK2, PACK3, etc.). During normal operation, the DC current output by the PCS to each battery pack is symmetrical between the positive and negative buses, and the leakage current signal is close to zero. However, when arc discharge occurs somewhere in the system, due to the asymmetry of the discharge path and the return path, a small difference will be generated in the current of the positive and negative buses. At this time, TMR0 can capture the total leakage current change of the main loop, and TMR1, TMR2, and TMR3 can monitor the weak leakage current (usually in the order of mA or even μA) on the positive and negative connection lines of the corresponding battery pack, thereby realizing the positioning and determination of the arc fault.

[0019] Figure 2 Middle represents the arc current, i.e. the leakage current signal, when an arc fault occurs, charges can flow between the battery packs through abnormal paths (such as high-frequency discharge channels formed by air breakdown), causing asymmetric disturbance to the originally balanced current. At this time, It is manifested as a high-frequency, short-time current surge, reflecting the abnormal flow of energy between the battery packs caused by the arc. By monitoring The sliding window integral value can identify the accumulation of high-frequency disturbance energy as one of the criteria for determining arc faults.

[0020] I pack3 represents the arc current, i.e. the leakage current signal, arc fault may be accompanied by insulation deterioration or the formation of discharge path, resulting in abnormal increase of battery pack leakage current (I pack3 ) to ground. Normally, I pack3 should be close to zero, but after arc, due to the return flow of charges through the grounding path, I pack3 will be significantly increased.

[0021] In view of the characteristics of low amplitude and high frequency of the arc current, the system uses a TMR sensor to replace the traditional Hall element or sampling resistor. The TMR sensor has high sensitivity (can detect nanampere-level current) and low noise characteristics, and can effectively capture high-frequency weak signals.

[0022] The data processing module is used for preprocessing the leakage current signal collected by the leakage current detection module, obtaining the leakage current numerical signal corresponding to each leakage current sensor and outputting; The data processing module comprises: A high-pass filter unit is used to filter the leakage current signal collected by the leakage current detection module to obtain a high-frequency current signal; A data processing unit is used to sequentially filter, amplify and analog-to-digital convert the high-frequency current signal to obtain a leakage current numerical signal.

[0023] To cope with the strong noise background in the energy storage system (such as load fluctuation, electromagnetic interference), the data processing module integrates a high-pass filter unit, which is specifically used to extract the high-frequency disturbance component in the 500Hz-20kHz frequency band of the leakage current signal, i.e. the high-frequency current signal, and through analog-to-digital conversion and amplification processing to enhance the signal-to-noise ratio of the signal.

[0024] Specifically, the data processing module adopts a high-speed analog-to-digital converter (ADC) with a sampling rate of not less than 1 MHz, supports multi-channel synchronous acquisition, and ensures high-precision real-time processing of the sensor output signal. The module extracts a high-frequency current signal (500 Hz-20 kHz) through a high-pass filter unit, which is used to capture the unique high-frequency disturbance energy characteristics (the filtered low-frequency component is the system discharge phenomenon caused by insulation failure) in the arc event. After filtering, amplification and analog-to-digital conversion, the captured signal forms a digitized leakage current numerical signal, which is finally transmitted to the main control module for multi-dimensional signal fusion analysis, thereby realizing differentiated identification of arc failure and insulation deterioration.

[0025] The main control module (STM32H7) is in communication connection with the battery management system, used for real-time acquisition of the running parameters uploaded by the battery management system, and receiving the leakage current numerical signals of each leakage current sensor output by the data processing module; the main control module is used for, for each leakage current sensor, based on the leakage current numerical signal and the running parameters corresponding thereto, generating a plurality of arc judgment results by using a preset multi-dimensional signal fusion analysis algorithm, and comprehensively judging whether the battery pack or the DC bus corresponding to the leakage current sensor has an arc failure according to a preset arc judgment rule.

[0026] The running parameters include: voltage and temperature data of each battery pack; voltage and temperature data of the DC bus.

[0027] The preset multi-dimensional signal fusion analysis algorithm includes: using a sliding window integration method to accumulate the absolute value of the leakage current numerical signal to obtain an integral value; if the integral value exceeds a first preset threshold, a first judgment result is marked; using short-time Fourier transform to perform frequency spectrum analysis on the leakage current numerical signal to extract the energy mutation characteristics of a preset target frequency band (500 Hz-20 kHz frequency band); if the energy change rate of the preset target frequency band exceeds a second preset threshold, a second judgment result is marked; The system combines time domain and frequency domain characteristics for comprehensive analysis: in the time domain, the absolute value of the leakage current signal is accumulated by using a sliding window integration method to evaluate the accumulation change of current energy in a specific time window; if the integral value exceeds a preset threshold, a potential arc failure is marked. In the frequency domain, the energy mutation characteristics of the 500 Hz-20 kHz frequency band in the leakage current signal are extracted by using short-time Fourier transform (STFT) to further identify high-frequency interference components. Through multi-dimensional signal fusion analysis, the system significantly improves the recognition accuracy and anti-interference ability of low-amplitude, short-time high-frequency arc events.

[0028] If the temperature rise rate of the battery pack or the DC bus exceeds the third preset threshold (for example, one of 1.5-2.5°C / s) within the time period corresponding to the first determination result, it is marked as the third determination result. If the temperature rise rate exceeds the third preset threshold, it can be determined that there is a potential local thermal abnormality phenomenon, and the specific threshold parameter can be dynamically adjusted according to the thermal capacity characteristics of the system.

[0029] If the voltage of the battery pack or the DC bus deviates from the rated value by more than the fourth preset threshold within the time period corresponding to the first determination result, it is marked as the fourth determination result.

[0030] Specifically, the system monitors the voltage variation trend of the battery pack or the DC bus in real time. If the voltage appears a short-term drop or rebound phenomenon within the time period corresponding to the first determination result (excessive leakage current integral value), and the deviation from the rated value is more than the fourth preset threshold (for example, ±5%), it is marked as the fourth determination result.

[0031] In this embodiment: First preset threshold (leakage current integral value): determined based on the amplitude energy mutation characteristics of high-frequency disturbance during arcing, combined with noise level.

[0032] Second preset threshold (frequency band energy change rate): needs to be set for the typical frequency band (such as 500Hz-20kHz) of the arcing signal, determined by spectrum comparison experiment (such as 0.02).

[0033] Third preset threshold (temperature rate): reference the critical temperature rise rate of battery thermal runaway (such as 1.5-2.5°C / s), combined with system thermal capacity parameters.

[0034] Fourth preset threshold (voltage deviation amplitude): set according to the tolerance range of the system rated voltage (such as ±5%), to avoid misjudgment of normal load fluctuation.

[0035] Data can be collected in typical arcing scenarios, and the statistical distribution of signal characteristics (such as integral value, frequency band energy mutation, temperature rise rate, and voltage deviation from rated value) is analyzed to determine the threshold that distinguishes between normal and abnormal: 1. Establish an experimental environment: create a test platform simulating a energy storage system, on which controlled arcing events can be safely triggered. By adjusting load conditions, temperature changes, and other factors, arcing situations under different working conditions can be simulated.

[0036] 2. Data collection: use high-precision sensors (such as TMR sensors) to monitor the leakage current signals of the DC bus and each battery pack in real time. At the same time, other related parameters are recorded, including but not limited to battery pack voltage, temperature, and DC bus voltage and temperature. These data will serve as the basis for subsequent analysis.

[0037] 3. Signal processing and feature extraction: Integral value calculation: The sliding window integration method is used to process the leakage current numerical signal, capturing the high-frequency disturbance energy accumulation effect caused by the arc.

[0038] Frequency band energy mutation analysis: Using short-time Fourier transform (STFT) or fast Fourier transform (FFT), analyze the energy distribution characteristics of a specific frequency band (such as 500Hz~20kHz), identify potential energy mutations caused by arc.

[0039] Temperature rise rate monitoring: Real-time monitoring of the temperature change rate of the battery pack and DC bus, combined with thermodynamic model, to evaluate whether there is a local overheating phenomenon caused by arc.

[0040] Voltage deviation amplitude measurement: Detect the degree of deviation of the battery pack or DC bus voltage from the rated value, and determine whether there is voltage fluctuation caused by arc.

[0041] Statistical analysis of the above feature indicators (integral value, frequency band energy mutation, temperature rise rate, voltage deviation amplitude, etc.), determine the upper limit value of the 95% confidence interval as the standard to distinguish between normal and abnormal. For example, for the frequency band energy mutation, if the upper limit of the 95% confidence interval is 0.02, any change exceeding this value will be considered abnormal. In this way, the first, second, third, and fourth preset thresholds are obtained.

[0042] This embodiment proposes two arc determination rules: The preset arc determination rule is: Calculate the weighted score sum of each determination result through the preset weight coefficient corresponding to each determination result, if the sum exceeds the fifth preset threshold, it is determined that there is an arc fault, and a fault confirmation signal is triggered.

[0043] The following is an example of weighted calculation with specific data (1 represents triggering, 0 represents not triggering):

[0044] First determination result (leakage current integral): Trigger state: 1 indicates that the leakage current integral exceeds the preset threshold, which may be caused by high-frequency disturbance caused by arc.

[0045] Dynamic weight coefficient: 1.0; Because the system load is high, the leakage current is more sensitive to arc phenomenon, and the weight remains the highest value.

[0046] Weighted score: 1.0; Reflects the importance of leakage current integral in arc identification.

[0047] Second determination result (spectrum energy mutation): Trigger state: 1; indicates that the spectrum energy has a significant change, which may be caused by the increase of high-frequency components due to arcing.

[0048] Dynamic weight coefficient: 0.8; the recent spectrum noise is high, so the weight is moderately reduced to avoid misjudgment.

[0049] Weighted score: 0.8; the possibility of arcing is evaluated by combining the spectrum energy change and the noise level.

[0050] Third determination result (temperature rate): Trigger state: 0; indicates that the current temperature rate does not exceed the preset range, and there is no obvious thermal anomaly.

[0051] Dynamic weight coefficient: 0.5; the current temperature is normal, and the weight is maintained at a low level.

[0052] Weighted score: 0.0; the temperature rate is not triggered, and is not included in the weighted score total.

[0053] Fourth determination result (voltage deviation amplitude): Trigger state: 1; indicates that the amplitude of voltage deviation from the rated value exceeds the preset threshold, which may be caused by short-term drop or rebound caused by arcing.

[0054] Dynamic weight coefficient: 0.8; the voltage fluctuation is large, and the weight is appropriately increased to reflect the change in system stability.

[0055] Weighted score: 0.8; the arcing risk is evaluated by combining the voltage deviation amplitude and the system state.

[0056] Fifth preset threshold: 2.5; Weighted score total: 1.0 + 0.8 + 0.0 + 0.8 = 2.6 Conclusion: the total 2.6>2.5, trigger fault confirmation signal.

[0057] Through the weighted score total (2.5) of the above four determination results, the system meets the fourth preset threshold condition (≥2.5), triggers the fault confirmation signal, and links the PCS to perform the pre-warning information reporting or power-off protection action, to ensure the safe and stable operation of the energy storage system.

[0058] The present application assigns corresponding weight coefficients to different determination results (such as leakage current integral, spectrum energy, temperature rise rate), calculates the weighted score total of each determination result, and if the total exceeds the fifth preset threshold, it is determined that there is an arcing fault, realizing the fusion determination of multi-dimensional signals and improving the determination accuracy.

[0059] The preset arcing determination rule is: If only the first judgment result is met and the second, third and fourth judgment results are not met, it is marked as a warning state and continuously monitored; if the first judgment result and any one of the second, third or fourth judgment results are met at the same time, it indicates that an arcing fault exists and a fault confirmation signal is immediately triggered.

[0060] The main control module includes: The interlocking protection unit is connected to the power conversion module PCS and is used to send an alarm code to the battery management system through the CAN bus or RS485 communication when the fault confirmation signal is triggered, and send a PWM shutdown command to the power conversion module PCS to cut off the connection between the DC bus and the power grid.

[0061] The arcing phenomenon is usually accompanied by the formation of a high-frequency discharge path. When the charge breaks down through the air in the internal or external path of the battery pack (PACK), a high-frequency leakage backflow current will be generated. This leakage current eventually returns to the battery pack ( Figure 2 In the example, "Shell" represents grounding, and the connection channel from the Shell to the negative terminal of the battery pack represents the ground wire channel. This causes a slight asymmetry in the current between the positive and negative busbars of the DC system, and a significant high-frequency component (500Hz-20kHz) is superimposed on the current signal. The system uses leakage current sensors (TMR sensors) to collect leakage current signals from the main circuit and each battery pack in real time. High-pass filtering is used to extract high-frequency disturbance characteristics, and multi-dimensional identification of arcing events is achieved using sliding window integration, short-time Fourier transform, and temperature / voltage anomaly correlation analysis. This detection mechanism, based on multi-source signal fusion, not only captures low-amplitude, short-time, high-frequency arcing signals, but also significantly improves detection sensitivity and anti-interference capabilities through dynamic weight adjustment and hierarchical judgment rules, thereby providing accurate arcing fault warning and location capabilities for energy storage systems.

[0062] The multi-condition joint discrimination strategy employed in this invention is based on multidimensional signal fusion analysis algorithms (such as sliding window integration, short-time Fourier transform, and operating parameter correlation analysis). It does not rely on complex model training and is easy to deploy on embedded platforms. Its parameters (such as the fourth preset threshold and weight coefficient) can be dynamically adjusted based on the energy storage system structure (such as the number of battery packs and bus voltage level) and operating conditions (such as load fluctuation range and ambient temperature), making it highly interpretable and configurable. Through the collaborative analysis of multi-source signals, the system significantly improves the recognition accuracy of early, low-energy arcing events (such as mA-level leakage current and energy mutations in the 500Hz-20kHz frequency band), while also shortening fault response time (to the millisecond level), providing reliable protection for the intrinsic safety of energy storage systems.

[0063] Example 2 The embodiment of the present invention further provides a DC arc detection method for an energy storage system, which is applied to the DC arc detection system described above. The DC arc detection method includes: The leakage current signals of the main circuit and each battery pack are collected through the total leakage current sensor installed on the positive and negative DC busbars, and the module-level leakage current sensors installed on the positive and negative connecting lines of each battery pack; Preprocessing the collected leakage current signals to obtain leakage current numerical signals corresponding to each leakage current sensor; The system acquires the operating parameters of the DC bus and each battery pack monitored by the battery management system in real time. For each leakage current sensor, based on its corresponding leakage current numerical signal and operating parameters, a preset multi-dimensional signal fusion analysis algorithm is used to generate multiple arcing determination results. The system then comprehensively determines whether the battery pack or DC bus corresponding to the leakage current sensor has an arcing fault based on the preset arcing determination rules. In the present invention, the main control module is communicatively connected with the battery management system, and is used to obtain the operating parameters uploaded by the battery management system in real time, and receive the leakage current numerical signal of each leakage current sensor output by the data processing module; the main control module is used to generate multiple arcing judgment results for each leakage current sensor based on its corresponding leakage current numerical signal and operating parameters, using a preset multi-dimensional signal fusion analysis algorithm, and comprehensively judge whether the battery pack or DC bus corresponding to the leakage current sensor has an arcing fault according to the preset arcing judgment rules; the present invention uses a multi-dimensional signal fusion analysis algorithm to generate multiple arcing judgment results to realize the fusion judgment of multiple source signals. Compared with the detection method of a single signal (such as relying only on leakage current or spectrum), multi-dimensional fusion significantly improves the recognition ability of low-amplitude, short-time high-frequency arcing signals, and reduces the risk of misjudgment.

[0064] The operating parameters include: The voltage and temperature data of each battery pack, as well as the voltage and temperature data of the DC bus; The preset multi-dimensional signal fusion analysis algorithm is used to generate multiple arc determination results, specifically: Accumulating the absolute value of the leakage current numerical signal using a sliding window integration method to obtain an integral value; if the integral value exceeds a first preset threshold, marking it as a first determination result; Performing spectrum analysis on the leakage current numerical signal using short-time Fourier transform to extract energy mutation characteristics of a preset target frequency band; if the energy change rate of the preset target frequency band exceeds a second preset threshold, marking it as a second determination result; If the temperature rise rate of the battery pack or the DC bus exceeds a third preset threshold value within the time period corresponding to the first determination result, it is marked as a third determination result; If the voltage of the battery pack or the DC bus deviates from the rated value by more than the fourth preset threshold in the time period corresponding to the first determination result, a fourth determination result is marked.

[0065] The determination result is determined according to the preset arc determination rule, and whether the battery pack or the DC bus corresponding to the leakage current sensor has an arc fault is determined, and specifically includes: Step 1, determine whether only the first determination result is satisfied, and the second determination result, the third determination result and the fourth determination result are not satisfied; if the condition is satisfied, a pre-warning state is marked, and a continuous monitoring process is entered; if not, step 2 is entered; Step 2, determine whether the first determination result and any one of the second or third or fourth determination result are satisfied at the same time; if the condition is satisfied, the arc fault is determined, and a fault confirmation signal is triggered immediately; The present application realizes multi-dimensional feature extraction of arc event through sliding window integration of leakage current signal (capture high frequency disturbance energy), short time Fourier transform (extract frequency band energy mutation feature), temperature and voltage abnormal correlation analysis (combine thermodynamic characteristics), compared with single signal detection method, multi-dimensional fusion significantly improves the recognition ability of low amplitude, short time high frequency arc signal, and reduces the risk of misjudgment.

[0066] Based on the determination result of the arc fault, a pre-warning is performed, specifically: When the fault confirmation signal is triggered, an alarm code is sent to the battery management system through CAN bus or RS485 communication, and a PWM shutdown instruction is sent to the power conversion module PCS, and the connection between the DC bus and the power grid is cut off.

[0067] The present application adopts real-time linkage protection mechanism, the main control module communicates with the battery management system (BMS) and the power conversion module PCS through CAN bus or RS485 communication. When the arc fault is determined, the linkage protection unit immediately sends a PWM shutdown instruction to the PCS, cuts off the connection between the DC bus and the power grid, and the response time can reach milliseconds. In addition, the leakage current sensor covers the DC bus total loop and each battery pack, and combined with the multi-source signal determination result (such as positioning the bus fault when only TMR0 exception is triggered, and positioning the corresponding battery pack fault when TMR1 is triggered), the accurate tracing of the fault position is realized. This layered positioning capability can support targeted maintenance, reduce system downtime and improve operation and maintenance efficiency.

[0068] Embodiment three In order to further illustrate in the time domain, the absolute value of the leakage current signal is accumulated by the sliding window integration method, the accumulation change of current energy in a specific time window is evaluated, and if the integral value exceeds the preset threshold, the feasibility of potential arc fault is marked, and the present embodiment is realized by Figure 3 And Figure 4A specific description is made: As shown in Figure 3 , the system is in normal operation state, the leakage current signal amplitude is stable and the fluctuation is small; when arc fault occurs, high frequency disturbance (such as 500Hz~20kHz frequency band) will be superimposed in the current waveform, which shows short-time amplitude sudden increase and periodic oscillation. In order to quantify this abnormal feature, the main control module uses sliding window integral method to process the leakage current numerical signal, and the specific calculation formula is as follows: (Formula 1); in the formula: : the value of the leakage current numerical signal at the corresponding position of k time; w: sliding window length (dynamically adjusted according to system noise level and arc characteristic frequency, typical value is 10ms~50ms); k : summation variable (moving in the window); n: position of the current time point; : integral result in the sliding window centered on n (representing the cumulative value of disturbance energy).

[0069] The leakage current detection module (such as TMR0, TMR1~TMR3) collects the leakage current signals of the main circuit and each battery pack in real time, and the data processing module extracts the high frequency disturbance component in the frequency band of 500Hz~20kHz through the high pass filter unit, and then filters, amplifies and digitizes it in turn to obtain the leakage current numerical signal.

[0070] The main control module performs sliding window integration (Formula 1) on the leakage current numerical signal, and adjusts the window length w (such as 20ms) to smooth the local noise and enhance the recognition robustness of arc events.

[0071] If the integral result exceeds the first preset threshold (for example, 1.2A·ms), it is marked as the first determination result; Combined with other determination results (such as spectrum energy mutation, temperature rate anomaly), multi-dimensional signal fusion analysis is carried out, if the preset arc determination rule is met, the fault confirmation signal is triggered, and the power conversion module PCS is linked to execute power-off protection.

[0072] As shown in Figure 4 , there is a significant difference in the curve after the integral before and after the arc fault: Normal state (see normal operation curve): The integral value is maintained in the low level interval (such as 0.1A·ms~0.5A·ms); Fault state (see the fault occurrence curve): the integral value rises rapidly in a short time (such as At = 0.172 seconds) and exceeds the first preset threshold value, and the system can trigger a warning within 172 ms after the arc occurs, reserving response time for subsequent protection measures (such as PWM shutdown instructions).

[0073] That is, in the normal state, the integral value always fluctuates in a low level interval. After the arc fault occurs, due to the significant increase in disturbance energy, the integral value rises rapidly in a short time and exceeds the set threshold level. The At = 0.172 shown in the figure represents the time interval from the start of the arc to the set alarm threshold, which is 0.172 seconds. The occurrence of the arc phenomenon is judged in a short time.

[0074] The present application adopts a layered protection mechanism, balances safety and system availability through preset arc judgment rules (such as entering a warning state when only the first judgment result is triggered, and executing power-off when multiple conditions are triggered). The warning stage continuously monitors and records data to provide a basis for subsequent analysis; the fault confirmation stage immediately triggers protection actions, forming a closed-loop control of "warning -> confirmation -> protection"; At the same time, the present application breaks through the limitations of traditional arc detection technology (such as relying on a single signal), solves the industry pain points of low-amplitude arc signals being easily masked by noise and high misjudgment rate through multi-source signal fusion algorithm (leakage current + frequency spectrum + temperature + voltage) and dynamic judgment logic. Compared with the prior art, the system has significantly improved sensitivity, response speed and anti-interference ability. This technology not only provides a high-precision, low-false alarm active protection scheme for energy storage systems, but also builds a unique multi-dimensional signal fusion algorithm and layered judgment rules for enterprises in the field of energy storage safety, promotes the industry from "passive protection" to "active warning" transformation, and has significant industrial application value and market competitiveness.

[0075] It should be noted that all the direction indications (such as up, down, left, right, front, back, etc.) in the embodiments of the present application are only used to explain the relative position relationship, movement condition, etc. between components in a certain specific posture (as shown in the drawings), and if the specific posture changes, the direction indications will also change accordingly. In addition, the descriptions such as "first", "second", "one", etc. in the present application are only for the purpose of description, and cannot be understood as indicating or implying the relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" can be explicitly or implicitly included at least one of the features. In the description of the present application, the meaning of "a plurality of" is at least two, for example, two, three, etc., unless otherwise specifically limited. In the present application, unless otherwise specifically defined and limited, the terms "connection", "fixation", etc. should be understood in a broad sense, for example, "fixation" can be fixed connection, or detachable connection, or integral; can be mechanical connection, or electrical connection; can be directly connected, or indirectly connected through an intermediate medium; can be the internal communication of two elements or the interaction relationship between two elements, unless otherwise specifically limited. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances. In addition, the technical solutions of each embodiment of the present application can be combined with each other, but it must be based on the realization of the ordinary skilled in the art, when the combination of technical solutions appears contradictory or unachievable, it should be considered that the combination of technical solutions does not exist, nor within the scope of protection required by the present application.

Claims

1. A DC arc detection system for an energy storage system, characterized in that: The energy storage system includes a power conversion module PCS, a positive DC bus, a negative DC bus, and multiple battery packs; the positive terminal of the PCS is connected to the positive electrode of each battery pack through the positive DC bus, and the negative terminal is connected to the negative electrode of each battery pack through the negative DC bus, forming a main circuit; the DC arc detection system includes: The leakage current detection module includes leakage current sensors mounted on the positive and negative DC busbars, as well as leakage current sensors mounted on the positive and negative connecting wires of each battery pack, for collecting leakage current signals of the main circuit and each battery pack; The data processing module is used to pre-process the leakage current signal collected by the leakage current detection module, obtain the leakage current numerical signal corresponding to each leakage current sensor and output it; The main control module is connected to the battery management system for real-time acquisition of operating parameters uploaded by the battery management system and for receiving leakage current numerical signals of each leakage current sensor output by the data processing module. The main control module is used to generate multiple arcing determination results for each leakage current sensor based on its corresponding leakage current numerical signal and operating parameters using a preset multi-dimensional signal fusion analysis algorithm, and to comprehensively determine whether there is an arcing fault in the battery pack or DC bus corresponding to the leakage current sensor according to the preset arcing determination rules.

2. A DC arc detection system for an energy storage system according to claim 1, characterized in that: The data processing module includes: A high-pass filter unit is used to filter the leakage current signal collected by the leakage current detection module to obtain a high-frequency current signal; The data processing unit is used to filter, amplify and perform analog-to-digital conversion on the high-frequency current signal in sequence to obtain a leakage current numerical signal.

3. A DC arc detection system for an energy storage system according to claim 1, characterized in that: The operating parameters include: Voltage and temperature data of each battery pack; DC bus voltage and temperature data.

4. A DC arc detection system for an energy storage system according to claim 3, characterized in that: The preset multi-dimensional signal fusion analysis algorithm includes: Accumulating the absolute value of the leakage current numerical signal using a sliding window integration method to obtain an integral value; if the integral value exceeds a first preset threshold, marking it as a first determination result; Performing spectrum analysis on the leakage current numerical signal using short-time Fourier transform to extract energy mutation characteristics of a preset target frequency band; if the energy change rate of the preset target frequency band exceeds a second preset threshold, marking it as a second determination result; If the temperature rise rate of the battery pack or the DC bus exceeds a third preset threshold value within the time period corresponding to the first determination result, it is marked as a third determination result; If, within the time period corresponding to the first determination result, the voltage of the battery pack or the DC bus deviates from the rated value by more than a fourth preset threshold, it is marked as a fourth determination result.

5. A DC arc detection system for an energy storage system according to claim 4, characterized in that: The preset arcing judgment rule is: The weighted score sum of each determination result is calculated by the preset weight coefficient corresponding to each determination result. If the sum exceeds the fifth preset threshold, it is determined that an arcing fault exists and a fault confirmation signal is triggered.

6. A DC arc detection system for an energy storage system according to claim 4, characterized in that: The preset arcing judgment rule is: If only the first judgment result is met and the second, third and fourth judgment results are not met, it is marked as a warning state and continuously monitored; if the first judgment result and any one of the second, third or fourth judgment results are met at the same time, it indicates that an arcing fault exists and a fault confirmation signal is immediately triggered.

7. A DC arc detection system for an energy storage system according to claim 5 or 6, characterized in that: The main control module includes: The interlocking protection unit is connected to the power conversion module PCS and is used to send an alarm code to the battery management system through the CAN bus or RS485 communication when the fault confirmation signal is triggered, and send a PWM shutdown command to the power conversion module PCS to cut off the connection between the DC bus and the power grid.

8. A DC arc detection method for an energy storage system, characterized in that: Applicable to the DC arc detection system according to any one of claims 1 to 7; the DC arc detection method comprises: The leakage current signals of the main circuit and each battery pack are collected through the total leakage current sensor installed on the positive and negative DC busbars, and the module-level leakage current sensors installed on the positive and negative connecting lines of each battery pack; Preprocessing the collected leakage current signals to obtain leakage current numerical signals corresponding to each leakage current sensor; The system acquires the operating parameters of the DC bus and each battery pack monitored by the battery management system in real time. For each leakage current sensor, based on its corresponding leakage current numerical signal and operating parameters, a preset multi-dimensional signal fusion analysis algorithm is used to generate multiple arcing determination results. The system then comprehensively determines whether the battery pack or DC bus corresponding to the leakage current sensor has an arcing fault based on the preset arcing determination rules. An early warning is issued based on the judgment results of the arc fault.

9. A DC arc detection method for an energy storage system according to claim 8, characterized in that: The operating parameters include: The voltage and temperature data of each battery pack, as well as the voltage and temperature data of the DC bus; The preset multi-dimensional signal fusion analysis algorithm is used to generate multiple arc determination results, specifically: Accumulating the absolute value of the leakage current numerical signal using a sliding window integration method to obtain an integral value; if the integral value exceeds a first preset threshold, marking it as a first determination result; Performing spectrum analysis on the leakage current numerical signal using short-time Fourier transform to extract energy mutation characteristics of a preset target frequency band; if the energy change rate of the preset target frequency band exceeds a second preset threshold, marking it as a second determination result; If the temperature rise rate of the battery pack or the DC bus exceeds a third preset threshold value within the time period corresponding to the first determination result, it is marked as a third determination result; If, within the time period corresponding to the first determination result, the voltage of the battery pack or the DC bus deviates from the rated value by more than a fourth preset threshold, it is marked as a fourth determination result.

10. A DC arc detection method for an energy storage system according to claim 9, characterized in that: The step of comprehensively judging whether the battery pack or DC bus corresponding to the leakage current sensor has an arc fault according to a preset arc judgment rule specifically includes: Step 1: Determine whether only the first determination result is satisfied, and the second, third, and fourth determination results are not satisfied; if the condition is satisfied, mark it as an early warning state and enter the continuous monitoring process; if not, proceed to step 2; Step 2: Determine whether the first determination result and any one of the second, third, or fourth determination results are simultaneously met; if the condition is met, it is determined to be an arc fault and a fault confirmation signal is immediately triggered; The early warning based on the judgment result of the arc fault is specifically: When the fault confirmation signal is triggered, an alarm code is sent to the battery management system via the CAN bus or RS485 communication, and a PWM shutdown command is sent to the power conversion module PCS to cut off the connection between the DC bus and the grid.

Images