Fault early warning method and system for information and energy integrated battery unit

Abstract

The invention provides an information and energy integrated battery unit fault early warning method and system. The method comprises the following steps: firstly, determining the form of an energy storage system for inputting PCS side working current; extracting main frequency components; selecting frequency points sensitive to impedance change to form a key frequency point sequence; constructing a multi-frequency-point information energy integrated modulation signal; synchronously sampling a battery unit end voltage and a loop current through the PCS, and calculating an impedance value corresponding to each key frequency point; constructing a group clustering set, and constructing an individual clustering set; and continuously monitoring the real-time impedance value, and when the cumulative number of times that a certain battery unit deviates from the group clustering set or the individual clustering set reaches a preset number of times, judging that the battery unit has a fault and giving an early warning. According to the invention, excitation signal online injection and impedance detection are realized through the PCS, battery state information acquisition is completed while energy is transmitted, early warning precision, real-time performance and economical efficiency are considered, and the operation safety of the energy storage system is effectively improved.

Description

Technical Field

[0001] This invention relates to the field of energy storage system control and protection, specifically to a method and system for early warning of faults in an integrated information and energy battery cell. Background Technology

[0002] As the global energy transition accelerates and the penetration rate of renewable energy in power systems continues to increase, lithium-ion battery energy storage systems, with their advantages of fast response, high energy density, and long cycle life, have become core supporting equipment for scenarios such as peak shaving and valley filling in smart grids, renewable energy consumption, and microgrid stability control. Their operational safety and stability directly determine the overall operational efficiency of the power system. Currently, the single-unit capacity of energy storage power stations has reached a scale of tens to hundreds of megawatts. A single power station typically contains hundreds of thousands of battery cells, which are connected in series and parallel to form battery units, serving as the smallest operating unit of the energy storage system.

[0003] Due to factors such as natural degradation of cell performance, errors in battery management system (BMS) data acquisition, and complex and variable operating conditions (charge / discharge rate, ambient temperature), battery cells are prone to abnormal increases in internal resistance, degradation of cell consistency, and thermal runaway during long-term service. Because energy storage systems are large-scale and have numerous cells, the probability of failure increases significantly. If a localized fault fails to provide timely warning, leading to an unplanned system outage, the resulting power deficit will severely impact grid frequency stability and voltage quality, and may even trigger a cascading failure. Therefore, accurate and timely fault warnings for battery cells have become a key technical bottleneck restricting the improvement of energy storage system reliability.

[0004] Currently, the technical solutions for battery cell fault early warning in the industry are mainly divided into three categories: The first category is based on battery models and state estimation algorithms. Its core is to construct an equivalent circuit model of the battery and identify parameters online, combining this with deviations in key state quantities such as state of charge (SOC) to achieve fault diagnosis. The second category is based on data-driven and artificial intelligence methods. These methods rely on historical operating data and use machine learning and deep learning models to mine battery operating patterns and abnormal characteristics to achieve early warning. The third category is based on additional sensing hardware, which directly captures changes in the battery's physical characteristics by adding dedicated sensors.

[0005] In summary, existing technical solutions are either limited by the computing power and data accuracy of the BMS, or cannot balance practicality and widespread adoption due to hardware costs and algorithm complexity. It is worth noting that in string energy storage systems, each battery cluster corresponds to an independent PCS unit. During operation, the PCS can collect and process data such as voltage, current, and power from both the grid and battery sides in real time, and possesses powerful data processing and real-time control capabilities. However, existing technologies fail to fully utilize the inherent functions of the PCS. They cannot achieve deep perception of the battery cell's operating status and accurate early fault warnings through the PCS without adding extra sensing hardware or significantly increasing the BMS's computing load. This results in the PCS's intelligent diagnostic potential being wasted, and it is used merely as a power command execution device.

[0006] A patent search revealed invention patent CN115963408A, which discloses a fault early warning system and method for individual batteries in an energy storage power station. The system includes: a communication management unit for collecting historical equipment data from each individual battery in the energy storage power station; a server for processing and analyzing the historical equipment data and user behavior data to determine at least one initial predicted fault point, calculate the actual fault occurrence probability of each initial predicted fault point, and use the initial predicted fault points whose actual fault occurrence probability exceeds a preset probability threshold as the final predicted fault points, generating fault early warning information; and user behavior data generated based on user queries and management of historical equipment data and / or fault early warning information. This patent does not consider the electrochemical characteristics of the battery, relies on user behavior data which is susceptible to subjective influence, lacks support from core battery parameters for fault determination, and suffers from insufficient accuracy and specificity in early warning.

[0007] In summary, given the problems of the existing technologies, researching an integrated information and energy battery cell fault early warning method and system has become a critical task that urgently needs to be addressed. Summary of the Invention

[0008] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method and system for early warning of faults in an integrated information and energy battery cell.

[0009] A method for early warning of faults in an integrated information and energy battery cell according to the present invention includes the following steps:

[0010] Step S1: Based on the topology of the energy storage converter, determine the form of the operating current input to the energy storage PCS side during the operation of the energy storage system. The form of the operating current includes DC form or AC form containing fundamental frequency component. Step S2: Perform frequency domain analysis on the operating current of the energy storage PCS side to obtain the frequency corresponding to the component with the largest amplitude in the spectrum, which is taken as the main frequency component; Step S3: Based on the electrochemical type of the battery cell, select multiple frequency points that are sensitive to impedance changes to form a key frequency point sequence for fault early warning. Step S4: Construct a multi-frequency modulation signal containing multiple frequency points based on the main frequency components and key frequency point sequences; Step S5: Synchronously sample the terminal voltage and loop current of the battery cell, and calculate the impedance value of the battery cell at key frequency points. Step S6: Based on the impedance values ​​of each battery cell in the energy storage power station at key frequency points, construct a group cluster set and an individual cluster set for each battery cell. The group cluster set summarizes the impedance values ​​of all battery cells in the energy storage power station at different times, and the individual cluster set summarizes the impedance values ​​of each battery cell at multiple historical times. Step S7: Continuously monitor the real-time impedance value of each battery cell. If the cumulative number of times the impedance value of a battery cell deviates from the statistical range of the group cluster set or the individual cluster set of the battery cell itself reaches a preset number, it is determined that the battery cell has failed and an early warning is issued.

[0011] Preferably, step S1 includes: Based on the topology of energy storage PCS, energy storage PCS is divided into string PCS, distributed PCS, and cascaded PCS. In string PCS, each battery cell is directly connected to the AC bus through a single-stage DC-AC converter. In distributed PCS, each battery cell is connected to the DC bus in parallel through a DC-DC converter, and then connected to the AC bus through a high-capacity DC-AC converter. In cascaded PCS, each battery cell is cascaded through a single-stage converter. The operating current form on the PCS side is determined according to the bus configuration of the resulting bridge arm: if connected to a DC bus, the operating current form is DC; if connected to an AC bus, the operating current form is AC, including the fundamental AC component.

[0012] Preferably, the method for obtaining the main frequency component in step S2 is to use a fast Fourier transform to convert the time-domain waveform of the working current input to the energy storage PCS into frequency-domain components, and to take the frequency corresponding to the component with the largest amplitude as the main frequency component f0 based on the component with the largest amplitude in the frequency domain components.

[0013] Preferably, step S3 includes: Based on the electrochemical type of the battery cell, determine the frequency range in which the electrochemical impedance of the battery cell at different SOCs is sensitive to frequency changes; select a set of discrete frequency points within the frequency range to form a key frequency point sequence for fault early warning, denoted as {f1*, f2*, …, fn*}, where n is the number of frequency points in the key frequency point sequence.

[0014] Preferably, step S4 includes the following sub-steps: Step S4.1: Combine the main frequency component f0 with the key frequency point sequence {f1} * f2 * , …, f n * Each frequency point in the sequence is added together to form a multi-frequency sinusoidal modulation sequence corresponding to frequencies {f1, f2, …, f}. i ,…, f n}, where f i = f i * +f 0， i = 1, 2, …, n; Step S4.2: Generate a multi-frequency sinusoidal signal f from the corresponding frequencies of the multi-frequency sinusoidal modulation sequence. MFS (t):

[0015] Step S4.3: Convert the multi-frequency sinusoidal signal into a multi-frequency modulation signal S(t). For full-bridge and half-bridge converters, the form of the multi-frequency modulation signal S(t) is determined according to the converter type.

[0016] Preferably, in step S4.3, for the full-bridge converter, the multi-frequency modulation signal S(t) consists of -1 and 1, and the generation logic is as follows: Step S4.3.1: Normalize the multi-frequency sinusoidal signal to the range (-1, 1) to obtain the normalized signal f. MFS_nom (t):

[0017] Where n represents the number of key frequency points; Step S4.3.2: For each sampling time step k, calculate the error e[k] according to the following recursive relationship: e[k] = e[k - 1] + (f MFS_nom [k]–S[k – 1]); Step S4.3.3: Determine the multi-frequency modulation signal S[k] at the current moment based on the calculation error: If e[k] > 0, then S[k] = 1; otherwise, S[k] = -1. For a half-bridge converter, the multi-frequency modulation signal S(t) consists of 0 and 1. It is obtained by linear transformation based on the generation result of the full-bridge converter. Multiplying the multi-frequency modulation signal generated by the full-bridge converter by 0.5 and then adding 0.5 will convert it into a signal consisting of 0 and 1. Different values ​​of the multi-frequency modulation signal S(t) represent different control actions on the input PCS side current. S(t)1=1 represents conducting the current to the battery cell; S(t)1=-1 represents conducting the current in reverse; S(t)1=0 represents bypassing the current.

[0018] Preferably, step S5 includes the following sub-steps: Step S5.1: Synchronously acquire the terminal voltage v of the battery cell through the voltage and current sampling unit of the energy storage converter. DC (t) and loop current i DC (t), and store it in the data storage unit of PCS. The time length of the collected data meets the requirement of not less than the lowest frequency point in the key frequency point sequence; Step S5.2: Obtain the terminal voltage v from the data storage unit. DC (t) and loop current i DC (t), and calculate v using Fast Fourier Transform respectively. DC (t), i DC The frequency domain quantity V of (t) DC (f), I DC (f) Based on the key frequency sequence, the impedance value Z of the battery cell at the corresponding key frequency point f is calculated according to the following formula. EIS (f): .

[0019] Preferably, the method for constructing the group cluster set and the individual cluster set corresponding to step S6 is as follows: After the initial state equalization of the energy storage power station is completed, the battery system is subjected to one charge-discharge cycle, and the impedance value ZEIS(f) of each battery cell controlled by each energy storage converter PCS at each key frequency point f is obtained through step S5 under different states of charge (SOC). After uploading the impedance values ​​ZEIS(f) of the battery cells managed by each PCS at key frequency points to the cloud platform, box plot statistics are performed on the real and imaginary parts of the impedance at each key frequency point in the cloud platform. This allows us to obtain the group cluster set and individual cluster set of the real and imaginary parts of the impedance of the battery cells managed by each PCS under different SOCs in the initial state of the power plant.

[0020] Preferably, step S7 includes: Based on the group cluster set, if the real-time impedance value of a certain battery cell deviates from the statistical range of the group cluster set by more than a preset threshold, the group outlier warning count of the battery cell is increased by 1; otherwise, the group outlier warning count of the battery cell is decreased by 1. When the group outlier warning count reaches a preset number, the battery cell is determined to have failed. Based on the individual cluster set of each battery cell, if the real-time impedance value of the battery cell deviates from the statistical range of the corresponding individual cluster set by more than a preset threshold, the individual outlier warning count of the battery cell is increased by 1. When the individual outlier warning count reaches a preset number, the battery cell is determined to have malfunctioned.

[0021] This invention also provides an integrated information and energy battery cell fault early warning system, employing the aforementioned integrated information and energy battery cell fault early warning method, comprising: Module M1: Based on the topology of the energy storage converter, determine the form of the operating current input to the energy storage PCS side during the operation of the energy storage system. The form of the operating current includes DC form or AC form containing fundamental frequency component. Module M2: Performs frequency domain analysis on the operating current of the energy storage PCS side to obtain the frequency corresponding to the component with the largest amplitude in the spectrum, which is used as the main frequency component; Module M3: Based on the electrochemical type of the battery cell, select multiple frequency points that are sensitive to impedance changes to form a sequence of key frequency points for fault early warning. Module M4: Constructs a multi-frequency modulation signal containing multiple frequency points based on the main frequency components and key frequency point sequences; Module M5: Synchronously samples the terminal voltage and loop current of the battery cell and calculates the impedance value of the battery cell sequence at key frequency points; Module M6: Based on the impedance values ​​of each battery cell in the energy storage power station at key frequency points, construct a group cluster set and an individual cluster set of each battery cell. The group cluster set summarizes the impedance values ​​of all battery cells in the energy storage power station at different times, and the individual cluster set summarizes the impedance values ​​of each battery cell at multiple historical times. Module M7: Continuously monitors the real-time impedance value of each battery cell. If the cumulative number of times the impedance value of a battery cell deviates from the statistical range of the group cluster set or the individual cluster set of the battery cell itself reaches a preset number, the battery cell is determined to have malfunctioned and an early warning is issued.

[0022] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention relies on the high-precision sampling unit and voltage and current control module of the original PCS of the energy storage system to realize online detection and fault early warning of battery cell impedance. It eliminates the need to deploy dedicated detection hardware such as pressure sensors and resonant circuits, avoiding the procurement, deployment, wiring and maintenance costs of additional hardware, simplifying the system structure, and improving the compatibility of existing energy storage system transformation and upgrading.

[0023] 2. This invention achieves fault judgment through frequency domain analysis, multi-frequency modulation signal construction, and cluster set comparison. The core algorithm is designed based on the inherent data processing capabilities of PCS, and its computational complexity is lower than that of traditional data-driven algorithms and multi-level model estimation algorithms. It does not require additional BMS computing resources and can be directly deployed on the existing PCS hardware platform, meeting the real-time requirements of large-scale energy storage systems for fault early warning.

[0024] 3. This invention identifies faults based on changes in battery cell impedance. Impedance parameters are more sensitive to fault characteristics such as cell performance degradation and consistency deterioration than single voltage and temperature parameters. At the same time, by using a dual comparison mechanism of group clustering and individual clustering, combined with the cumulative number of deviations to determine faults, it can effectively avoid interference from operating conditions such as current acquisition errors and temperature fluctuations, reduce false alarms and missed alarms, and improve the accuracy of fault identification.

[0025] 4. The multi-frequency modulation signal constructed by this invention can complete the excitation signal injection and battery status information acquisition while the PCS performs power conversion and energy transmission, without occupying additional running time or energy resources. This realizes the coordinated operation of energy transmission and information acquisition, and improves the overall operating efficiency of the energy storage system.

[0026] 5. This invention can adaptively determine the operating current form and modulation signal generation logic according to different PCS topologies such as string, distributed, and cascaded, and is compatible with battery cells of different electrochemical types. It can be widely used in various large-scale energy storage power stations and has good engineering promotion value. Attached Figure Description

[0027] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a flowchart of a fault early warning method for an integrated information and energy battery cell according to an embodiment of the present invention; Figure 2 This invention presents a common battery energy storage system converter (PCS) architecture—a string energy storage system architecture. Figure 3 This invention presents a common battery energy storage system converter (PCS) architecture—a distributed energy storage system architecture. Figure 4 The present invention relates to a common battery energy storage system converter (PCS) architecture – a cascaded energy storage system architecture. Figure 5 This is a schematic diagram of the information-energy integrated modulation signal generation method in an embodiment of the present invention; Figure 6This invention includes a modulation signal segment, a multi-frequency sinusoidal signal segment, and a modulation signal spectrum (full-bridge modulation signal). Figure 7 An embodiment of the present invention includes a modulation signal segment, a multi-frequency sinusoidal signal segment, and a modulation signal spectrum (half-bridge modulation signal); Figure 8 The required operating states of the PCS under different PCS architectures in the embodiments of the present invention are the operating states of the switching transistors under the string energy storage system architecture when implementing the proposed modulation. Figure 9 The simulated DC-side voltage and current waveforms of the modulation method proposed under the string PCS architecture in the embodiments of the present invention are shown. Figure 10 The required operating states of the PCS under different PCS architectures in the embodiments of the present invention are the operating states of the switching transistors under the distributed energy storage system architecture when implementing the proposed modulation. Figure 11 The simulated waveforms of typical DC-side voltage and current of the modulation method proposed under the distributed PCS architecture in the embodiments of the present invention are shown. Figure 12 The required operating states of the PCS under different PCS architectures in the embodiments of the present invention are as follows: the operating states of the switching transistors under the cascaded energy storage system architecture. Figure 13 The above are typical DC-side voltage and current simulation waveforms (full-bridge modulation) of the modulation method proposed under the cascaded PCS architecture in the embodiments of the present invention. Figure 14 The simulated DC-side voltage and current waveforms (half-bridge modulation) of the modulation method proposed under the cascaded PCS architecture in the embodiments of the present invention are typical DC-side voltage and current waveforms. Figure 15 This is the online impedance identification result of information and energy integration, taking the cascaded PCS architecture as an example in this embodiment of the invention. Figure 16 This is a schematic diagram of a method for constructing an impedance group cluster set under a specific SOC in an embodiment of the present invention; Figure 17 This is a schematic diagram of a method for constructing an impedance individual cluster set under a specific SOC in an embodiment of the present invention. Detailed Implementation

[0028] The present invention will now be described in detail with reference to specific embodiments. These embodiments will help those skilled in the art to further understand the present invention, but do not limit the invention in any way. It should be noted that those skilled in the art can make several changes and improvements without departing from the concept of the present invention. These all fall within the protection scope of the present invention.

[0029] This invention provides an integrated information and energy battery cell fault early warning method and system, aiming to solve the problems of existing fault early warning schemes for lithium-ion battery energy storage systems, which rely on additional hardware, have high computational complexity, or lack sufficient early warning reliability. The method fully utilizes the high-precision, high-speed sampling and voltage / current control functions of the energy storage converter (PCS) itself, achieving accurate early warning of battery cell faults without adding additional hardware. The steps include: first, determining the form of the operating current input to the PCS side of the energy storage system based on the PCS topology; performing frequency domain analysis on the operating current to extract the main frequency components corresponding to the component with the largest amplitude; selecting frequency points sensitive to impedance changes to form a key frequency point sequence based on the electrochemical type of the battery cells; constructing a multi-frequency integrated information-energy modulation signal based on the main frequency components and the key frequency point sequence; calculating the impedance value corresponding to each key frequency point by synchronously sampling the battery cell terminal voltage and loop current through the PCS; constructing a group cluster set by summarizing the impedance data of all battery cells in the power station, and constructing an individual cluster set by summarizing the impedance data of each battery cell at different times; continuously monitoring the real-time impedance value, and determining that the battery cell has failed and issuing an early warning when the cumulative number of times a battery cell deviates from the group cluster set or the individual cluster set reaches a preset number. This invention realizes online injection of excitation signals and impedance detection through PCS, and completes the acquisition of battery status information while transmitting energy, taking into account the early warning accuracy, real-time performance and economy, effectively improving the operational safety of the energy storage system.

[0030] This invention utilizes a power conversion converter (PCS) within the energy storage system to achieve online injection of the excitation signal. Through high-speed, high-precision sampling within the PCS, the current and voltage of the battery cells are acquired, allowing for the calculation of the battery cell impedance. This enables the energy storage system to provide early warnings of battery cell faults based on abnormal changes in battery cell impedance. This method can acquire information while transmitting energy, improving the safety of the energy storage system without adding extra hardware.

[0031] Example 1: Figure 1 This is a flowchart of a fault early warning method for an integrated information and energy battery cell according to an embodiment of the present invention.

[0032] like Figure 1 As shown, this embodiment provides a fault early warning method for an integrated information and energy battery cell, including the following steps: Step S1: Based on the topology of the energy storage converter, determine the form of the operating current input to the energy storage PCS (PowerConversion System) during the operation of the energy storage system. The form of the operating current includes DC form or AC form containing fundamental frequency component. Specifically, based on the existing common energy storage PCS topologies, energy storage PCS can be divided into string PCS, distributed PCS, and cascaded PCS, with their specific characteristics as follows: like Figure 2 As shown, in a string PCS, each battery cell is directly connected to the AC bus via a single-stage DC-AC converter. During normal grid-connected operation, the operating current input to the energy storage PCS is three-phase AC; during off-grid operation, the specific current operating mode can be determined based on the output current of the parallel PCS.

[0033] like Figure 3 As shown, in a distributed PCS, each battery cell is connected to the DC bus via a parallel DC-DC converter, and then connected to the AC bus via a high-capacity DC-AC converter. When the system is connected to the grid, the current input to the energy storage PCS side (DC-DC) is DC; when operating off-grid, the current input to the energy storage PCS side (DC-DC) can be determined based on the specific output current of the parallel DC-DC converter.

[0034] like Figure 4 As shown, in a cascaded PCS, each battery cell is cascaded through a single-stage converter. The specific operating current form on the PCS side is determined according to the bus configuration of the bridge arm after cascading: when connected to a DC bus, the operating current form on the energy storage PCS side is DC; when connected to an AC bus, the operating current form on the energy storage PCS side is AC at power frequency.

[0035] Step S2: Perform frequency domain analysis on the operating current of the energy storage PCS side to obtain the frequency corresponding to the component with the largest amplitude in the spectrum, which is used as the main frequency component of the operating current input to the PCS side of the energy storage system.

[0036] Specifically, the method for obtaining the main frequency component is to use Fast Fourier Transform (FFT) to transform the time-domain waveform of the operating current input to the energy storage PCS into frequency-domain components, and then obtain the frequency corresponding to the component with the largest amplitude in the frequency domain as the main frequency component f0. Optionally, when the energy storage PCS architecture is string or distributed, since the switching signal can be freely controlled when the off-grid is running, the maximum frequency component is the DC component, so f0 = 0 Hz; for cascaded energy storage systems connected to the DC bus, the maximum frequency component of the operating current flowing into the PCS side is also 0 Hz. Optionally, for a system with a power frequency of 50 Hz, the cascaded connection of the AC bus is an energy storage system, and the maximum frequency component of the operating current flowing into the PCS side is obtained at 50 Hz, then f0 = 50 Hz. Optionally, for a system with a power frequency of 60 Hz, the cascaded connection of the AC bus is an energy storage system, and the maximum frequency component of the operating current flowing into the PCS side is obtained at 60 Hz, then f0 = 60 Hz. Step S3: Based on the electrochemical type of the battery cell, select multiple frequency points that are sensitive to impedance changes to form a key frequency point sequence for fault early warning.

[0037] Specifically, step S3 includes: determining the frequency range in which the electrochemical impedance of the battery cell is sensitive to frequency changes under different SOCs based on the electrochemical type of the battery cell; selecting a set of discrete frequency points within the frequency range to form a key frequency point sequence for fault early warning, denoted as {f1*, f2*, …, fn*}, where n is the number of frequency points in the key frequency point sequence.

[0038] For lithium iron phosphate batteries, the frequency range with sensitive characteristics is 0.1 Hz to 110 Hz. Therefore, seven points such as 1 Hz, 2 Hz, 4 Hz, 8 Hz, 16 Hz, 32 Hz, and 64 Hz can be selected as the key frequency point sequence {f1}. * f2 * , …,f n *}

[0039] Step S4: Based on the main frequency components and key frequency point sequence, according to... Figure 5 The method shown constructs a multi-frequency modulation signal containing multiple frequency points.

[0040] Specifically, step S4 includes the following sub-steps: Step S4.1: Combine the main frequency component f0 with the key frequency point sequence {f1} * f2 * , …, f n * Each frequency point in the sequence is added together to form a multi-frequency sinusoidal modulation sequence corresponding to frequencies {f1, f2, …, f}. n}, where f i = f i * + f0(i= 1, 2, …, n). Taking f0= 50 Hz and key frequency point sequences of 1 Hz, 2 Hz, 4 Hz, 8 Hz, 16 Hz, 32 Hz, 64 Hz as an example, the constructed multi-frequency point sinusoidal modulation sequence corresponds to frequencies {f1, f2, …, f n} = {51 Hz, 52 Hz, 54 Hz, 58 Hz, 66 Hz, 82 Hz, 114 Hz}.

[0041] Step S4.2: Convert the corresponding frequencies of the multi-frequency sinusoidal modulation sequence into a multi-frequency sinusoidal signal f. MFS (t):

[0042] Step S4.3: Convert the multi-frequency sinusoidal signal into a multi-frequency modulation signal S(t) (switching modulation signal). The form of the multi-frequency modulation signal S(t) is slightly different for full-bridge converters and half-bridge converters.

[0043] Specifically, for a full-bridge converter, such as Figure 6 As shown in the first sub-figure, the multi-frequency modulation signal S(t) consists of -1 and 1, and the specific generation logic is as follows: Step S4.3.1: Normalize the multi-frequency sinusoidal signal to the range (-1, 1) to obtain the normalized signal f. MFS_nom (t):

[0044] Where n represents the number of key frequency points; Step S4.3.2: For each sampling time step k, calculate the error e[k] according to the following recursive relationship: e[k] = e[k - 1] + (f MFS_nom [k] – S[k – 1]), S(0) can be selected according to the specific PCS architecture, and in this embodiment it is taken as -1.

[0045] Step S4.3.3: Determine the multi-frequency modulation signal S[k] at the current moment based on the calculation error: If e[k] > 0, then S[k] = 1; otherwise, S[k] = -1. After obtaining S[k], the modulated signal output remains unchanged until the next sampling time step triggers error calculation and re-determines the multi-frequency modulated signal output. The multi-frequency modulated signal obtained using this method, after spectral analysis, can be used to obtain... Figure 6 The frequency distribution shown in the second subplot indicates that the signal has a higher amplitude at the desired frequency point.

[0046] For half-bridge converters, such as Figure 7 As shown in the first sub-figure, the multi-frequency modulation signal consists of 0s and 1s. Its generation method can directly follow the full-bridge converter generation method. Simply multiply the multi-frequency modulation signal generated by the full-bridge converter by 0.5, then add 0.5 to convert it into a signal composed of 0s and 1s. Similarly, after spectral analysis, the multi-frequency modulation signal obtained using this method can be... Figure 7The frequency distribution shown in the second subplot indicates that the signal has a higher amplitude at the desired frequency point.

[0047] Different values ​​of the multi-frequency modulation signal S(t) represent different control actions on the input PCS side current. S(t)1=1 represents conducting the current to the battery cell; S(t)1=-1 represents conducting the current in reverse; S(t)1=0 represents bypassing the current.

[0048] Optionally, the multi-frequency modulation signal S(t) can be applied to different energy storage PCS architectures to perform current modulation in the following manner: like Figure 8 As shown, for a string PCS architecture, there are two methods for generating excitation signals. Both methods can be achieved by leveraging the mutual drag between parallel PCS under off-grid operation conditions.

[0049] When the voltage of the battery cell controlled by the two PCS is known, excitation signal generation can be performed using method 1. At this time, there is no need for synchronous control of PCS. It is only necessary to perform multi-frequency modulation on the PCS that controls the battery cell with higher open circuit voltage, and the other PCS can achieve mechanism injection by using the upper tube normally open. When the voltage of the battery cell controlled by the two PCS is not clearly defined, the excitation signal needs to be injected by simultaneously switching the upper and lower transistors. Both methods can obtain the following: Figure 9 The simulated waveforms of DC-side voltage and current in a typical string PCS architecture are shown. like Figure 10 As shown, for a distributed PCS architecture, under off-grid operation, the parallel DC / DC converters can be used for mutual drag, or a strategy of one controlling voltage and the other performing modulation, to inject excitation signals into the battery cells. However, since the DC / DC architecture typically includes inductors, this can lead to significant DC-side current and voltage oscillations after switching, such as... Figure 11 As shown.

[0050] like Figure 12 As shown, for the cascaded PCS architecture, under grid-connected operation, the complementary output between modules in the same bridge arm is used to achieve a constant output voltage without affecting the normal operation of the external circuit.

[0051] like Figure 13 As shown, when full-bridge modulation is used, the bridge arm current will continuously switch between positive and negative phases under the modulation of the PCS and be transmitted to the DC side, thus generating a corresponding response signal in the battery cell. In this case, the PCS is required to be an H-bridge circuit. Since the two modules have complementary outputs, the two modules have an equivalent output voltage of 0, which does not affect the normal external output of the system at all, and can realize the system's uninterrupted online electrochemical impedance excitation; like Figure 14As shown, when half-bridge modulation is used, the bridge arm current is bypassed by the PCS, so there may be zero current on the DC side. In this case, the PCS can be a half-bridge circuit. After the two modules complement each other, the DC side retains a stable positive level of one unit (average voltage of the battery cell). This can be controlled by the system to reduce the DC bias required by other modules in the same bridge arm.

[0052] Step S5: Synchronously sample the terminal voltage and loop current of the battery cell, and calculate the impedance value of the battery cell at the key frequency point sequence based on the key frequency point sequence.

[0053] Preferably, step S5 includes the following sub-steps: Step S5.1: Synchronously acquire the terminal voltage v of the battery cell through the high-precision, high-speed voltage and current sampling unit of the energy storage converter (PCS). DC (t) and loop current i DC (t), and store it in the PCS data storage unit. The time length of the collected data meets the requirement of not less than the lowest frequency point in the key frequency point sequence; Step S5.2: Obtain the terminal voltage v from the data storage unit. DC (t) and loop current i DC (t), and calculate v using Fast Fourier Transform (FFT) respectively. DC (t), i DC The frequency domain quantity V of (t) DC (f), I DC (f) Based on the key frequency sequence, the impedance value Z of the battery cell at the corresponding key frequency point f is calculated according to the following formula. EIS (f):

[0054] Based on the above method, taking a cascaded energy storage system architecture as an example, the following can be obtained: Figure 15 The figure shows the online impedance identification results of the integrated information and energy system. As can be seen from the figure, the identified impedance value almost perfectly matches the actual impedance value, indicating that the method has extremely high accuracy.

[0055] Step S6: Based on the impedance values ​​of each battery cell in the energy storage power station at key frequency points, construct a group cluster set and an individual cluster set for each battery cell: the group cluster set summarizes the impedance values ​​of all battery cells in the energy storage power station at different times, and the individual cluster set summarizes the impedance values ​​of each battery cell at multiple historical times.

[0056] Preferably, the method for constructing the group cluster set and individual cluster set corresponding to step S6 is as follows: like Figure 16As shown, after the energy storage power station completes the initial state equalization, the battery system is subjected to one charge-discharge cycle, and under different states of charge (SOC), the impedance value ZEIS(f) of the battery cell controlled by each energy storage converter PCS at each key frequency point f is obtained through step S5.

[0057] After uploading the impedance values ​​ZEIS(f) of the battery cells managed by each PCS at key frequency points to the cloud platform, box plot statistics are performed on the real and imaginary parts of the impedance at each key frequency point in the cloud platform. This allows us to obtain the group cluster set distribution and individual cluster set distribution of the real and imaginary parts of the impedance of the battery cells managed by each PCS under different SOCs in the initial state of the power plant.

[0058] The impedance values ​​of all battery cells in the energy storage power station at the same SOC and the same critical frequency point are aggregated to form a group cluster set; for each battery cell, the impedance values ​​of the battery cell at different historical times and the same critical frequency point are aggregated to form an individual cluster set of the battery cell.

[0059] Step S7: Continuously monitor the real-time impedance value of each battery cell. If the cumulative number of times the impedance value of a battery cell deviates from the statistical range of the group cluster set or the individual cluster set of the battery cell itself reaches a preset number, it is determined that the battery cell has failed and an early warning is issued.

[0060] In this embodiment, in the initial stage of the energy storage power station's operation, based on the group cluster set established in step S6, it is determined whether there are abnormal battery cells in the initial state of the power station, and these battery cells are marked. After several charge-discharge cycles (e.g., 3 times), the impedance of the marked battery cells is checked again. If the impedance outlier is greater, that is, if the impedance value deviates from the interquartile range of the upper or lower edge of the box plot by a higher multiple, then its initial anomaly is further confirmed, providing a key target for subsequent continuous monitoring.

[0061] like Figure 17 As shown, for each battery cell managed by PCS, at the initial stage of the power station, the battery impedance at key frequency points under different states of charge (SOC) is monitored and obtained. Each time the electrochemical impedance is detected, the impedance value (including real and imaginary parts) at the key frequency point under the corresponding SOC is recorded again according to the SOC at the time of detection, and statistical analysis and clustering are performed. When the impedance data of a certain battery cell under a specific SOC exceeds 5, the box plot can be drawn and the individual cluster set can be determined. Each subsequent electrochemical impedance detection is performed, and the new data is added to the corresponding individual cluster set, thereby realizing the dynamic updating and improvement of the set and ensuring that it can reflect the latest normal state baseline of the battery.

[0062] Specifically, step S7 includes: Based on the group cluster set, if the real-time impedance value of a certain battery cell deviates from the statistical range of the group cluster set by more than a preset threshold, the group outlier warning count of the battery cell is increased by 1; otherwise, the group outlier warning count of the battery cell is decreased by 1. When the group outlier warning count reaches a preset number, the battery cell is determined to have failed. Based on the individual cluster set of each battery cell, if the real-time impedance value of the battery cell deviates from the statistical range of the corresponding individual cluster set by more than a preset threshold, the individual outlier warning count of the battery cell is increased by 1. When the individual outlier warning count reaches a preset number, the battery cell is determined to have malfunctioned.

[0063] Example 2: The present invention also provides an integrated information and energy battery cell fault early warning system. The integrated information and energy battery cell fault early warning system can be implemented by executing the process steps of the integrated information and energy battery cell fault early warning method. That is, those skilled in the art can understand the integrated information and energy battery cell fault early warning method as the preferred embodiment of the integrated information and energy battery cell fault early warning system.

[0064] Specifically, the integrated information and energy battery cell fault early warning system includes: Module M1: Based on the topology of the energy storage converter, determine the form of the operating current input to the energy storage PCS side during the operation of the energy storage system. The form of the operating current includes DC form or AC form containing fundamental frequency component. Module M2: Performs frequency domain analysis on the operating current of the energy storage PCS side to obtain the frequency corresponding to the component with the largest amplitude in the spectrum, which is used as the main frequency component; Module M3: Based on the electrochemical type of the battery cell, select multiple frequency points that are sensitive to impedance changes to form a sequence of key frequency points for fault early warning. Module M4: Constructs a multi-frequency modulation signal containing multiple frequency points based on the main frequency components and key frequency point sequences; Module M5: Synchronously samples the terminal voltage and loop current of the battery cell and calculates the impedance value of the battery cell sequence at key frequency points; Module M6: Based on the impedance values ​​of each battery cell in the energy storage power station at key frequency points, construct a group cluster set and an individual cluster set of each battery cell. The group cluster set summarizes the impedance values ​​of all battery cells in the energy storage power station at different times, and the individual cluster set summarizes the impedance values ​​of each battery cell at multiple historical times. Module M7: Continuously monitors the real-time impedance value of each battery cell. If the cumulative number of times the impedance value of a battery cell deviates from the statistical range of the group cluster set or the individual cluster set of the battery cell itself reaches a preset number, the battery cell is determined to have malfunctioned and an early warning is issued.

[0065] Specifically, module M4 includes the following sub-modules: Module M4.1: Combines the main frequency component f0 with the key frequency point sequence {f1} * f2 * , …, f n * Each frequency point in the sequence is added together to form a multi-frequency sinusoidal modulation sequence corresponding to frequencies {f1, f2, …, f}. n}, where f i = f i * + f0(i= 1, 2, …, n). Taking f0= 50 Hz and key frequency point sequences of 1 Hz, 2 Hz, 4 Hz, 8 Hz, 16 Hz, 32 Hz, 64 Hz as an example, the constructed multi-frequency point sinusoidal modulation sequence corresponds to frequencies {f1, f2, …, f n} = {51 Hz, 52 Hz, 54 Hz, 58 Hz, 66 Hz, 82 Hz, 114 Hz}.

[0066] Module M4.2: Converts the corresponding frequencies of a multi-frequency sinusoidal modulation sequence into a multi-frequency sinusoidal signal f. MFS (t):

[0067] Module M4.3: Converts multi-frequency sinusoidal signals into multi-frequency modulated signals S(t). The form of the multi-frequency modulated signal S(t) differs slightly between full-bridge and half-bridge converters.

[0068] Specifically, for a full-bridge converter, the multi-frequency modulation signal S(t) consists of -1 and 1, and the specific generation logic is as follows: Module M4.3.1: Normalizes multi-frequency sinusoidal signals to the range (-1, 1) to obtain the normalized signal f. MFS_nom (t):

[0069] Where n represents the number of key frequency points; Module M4.3.2: For each sampling time step k, calculate the error e[k] according to the following recursive relationship: e[k] = e[k - 1] + (f MFS_nom [k] – S[k – 1]), S(0) can be selected according to the specific PCS architecture, and in this embodiment it is taken as -1.

[0070] Module M4.3.3: Determine the multi-frequency modulation signal S[k] at the current moment based on the calculation error: If e[k]>0, then S[k] = 1; otherwise, S[k] = -1. After obtaining S[k], the modulation signal output remains unchanged until the next sampling time step triggers error calculation and redetermines the switching modulation signal output.

[0071] For a half-bridge converter, the multi-frequency modulation signal consists of 0 and 1. Its generation method can be directly adopted by the full-bridge converter. Simply multiply the multi-frequency modulation signal generated by the full-bridge converter by 0.5 and add 0.5 to convert it into a signal composed of 0 and 1.

[0072] Different values ​​of the multi-frequency modulation signal S(t) represent different control actions on the input PCS side current. S(t)1=1 represents conducting the current to the battery cell; S(t)1=-1 represents conducting the current in reverse; S(t)1=0 represents bypassing the current.

[0073] Preferably, module M5 includes the following sub-modules: Module M5.1: Synchronously acquires the terminal voltage V of the battery cell through the high-precision, high-speed voltage and current sampling unit of the energy storage converter (PCS). DC (t) and loop current i DC (t), and store it in the PCS data storage unit. The time length of the collected data meets the requirement of not less than the lowest frequency point in the key frequency point sequence; Module M5.2: Obtains the terminal voltage V from the data storage unit. DC (t) and loop current i DC (t), and calculate v using Fast Fourier Transform (FFT) respectively. DC (t), i DC The frequency domain quantity V of (t) DC (f), I DC (f) Based on the key frequency sequence, the impedance value Z of the battery cell at the corresponding key frequency point f is calculated according to the following formula. EIS (f):

[0074] The aforementioned functional modules are implemented collaboratively through the following hardware platforms: the energy storage converter (PCS) is responsible for signal modulation and control in modules M1 and M4, and data acquisition in module M5; the energy storage cloud platform is responsible for algorithm execution and data storage in modules M2, M3, M5 (partial calculation), M6, and M7. Specifically, the PCS has DC-side current and voltage control, high-speed and high-precision DC-side current and voltage sampling, and communication functions; the energy storage cloud platform has communication functions, data storage functions, and strong computing capabilities; the battery cell fault early warning algorithm is deployed in the energy storage cloud platform and executed by the processor in the energy storage cloud platform. The voltage and current sampling signals of each battery cell are transmitted to the data storage unit through communication between the energy storage converters. The cloud platform processor calls these data and performs calculation operations such as fast Fourier transform and impedance clustering. The data storage unit further stores the calculated electrochemical impedance, impedance cluster sets (including group cluster sets and individual cluster sets), and fault early warning counts (including group outlier fault early warning counts and individual outlier fault early warning counts).

[0075] Those skilled in the art will understand that, besides implementing the system and its various devices, modules, and units provided by this invention in the form of purely computer-readable program code, the same functions can be achieved entirely through logical programming of the method steps, making the system and its various devices, modules, and units of this invention function in the form of logic gates, switches, application-specific integrated circuits, programmable logic controllers, and embedded microcontrollers. Therefore, the system and its various devices, modules, and units provided by this invention can be considered as a hardware component, and the devices, modules, and units included therein for implementing various functions can also be considered as structures within the hardware component; alternatively, the devices, modules, and units for implementing various functions can be considered as both software modules implementing the method and structures within the hardware component.

[0076] Specific embodiments of the present invention have been described above. It should be understood that the present invention is not limited to the specific embodiments described above, and those skilled in the art can make various changes or modifications within the scope of the claims, which do not affect the essence of the present invention. Unless otherwise specified, the embodiments and features described in this application can be arbitrarily combined with each other.

Claims

1. A method for early warning of faults in an integrated information and energy battery cell, characterized in that, Includes the following steps: Step S1: Based on the topology of the energy storage converter, determine the form of the operating current input to the energy storage PCS side during the operation of the energy storage system. The form of the operating current includes DC form or AC form containing fundamental frequency component. Step S2: Perform frequency domain analysis on the operating current of the energy storage PCS side to obtain the frequency corresponding to the component with the largest amplitude in the spectrum, which is taken as the main frequency component; Step S3: Based on the electrochemical type of the battery cell, select multiple frequency points that are sensitive to impedance changes to form a key frequency point sequence for fault early warning. Step S4: Construct a multi-frequency modulation signal containing the multiple frequency points based on the main frequency components and the key frequency point sequence; Step S5: Synchronously sample the terminal voltage and loop current of the battery cell, and calculate the impedance value of the battery cell at the key frequency point sequence; Step S6: Based on the impedance values ​​of each battery cell in the energy storage power station at the key frequency point sequence, construct a group cluster set and an individual cluster set of each battery cell. The group cluster set summarizes the impedance values ​​of all battery cells in the energy storage power station at different times, and the individual cluster set summarizes the impedance values ​​of each battery cell at multiple historical times. Step S7: Continuously monitor the real-time impedance value of each battery cell. If the cumulative number of times the impedance value of a battery cell deviates from the statistical range of the group cluster set or the individual cluster set of the battery cell itself reaches a preset number, it is determined that the battery cell has failed and an early warning is issued.

2. The information-energy integrated battery cell fault early warning method according to claim 1, characterized in that, Step S1 includes: Based on the topology of energy storage PCS, energy storage PCS is divided into string PCS, distributed PCS, and cascaded PCS. In the string PCS, each battery cell is directly connected to the AC bus through a single-stage DC-AC converter. In the distributed PCS, each battery cell is connected to the DC bus in parallel through a DC-DC converter, and then connected to the AC bus through a high-capacity DC-AC converter. In the cascaded PCS, each battery cell is cascaded through a single-stage converter. The operating current form on the PCS side is determined according to the bus configuration of the resulting bridge arm: if connected to a DC bus, the operating current form is DC; if connected to an AC bus, the operating current form is AC, including the fundamental AC component.

3. The information-energy integrated battery cell fault early warning method according to claim 1, characterized in that, The method for obtaining the main frequency component in step S2 is to use a fast Fourier transform to convert the time-domain waveform of the working current input to the energy storage PCS into frequency-domain components, and to take the frequency corresponding to the component with the largest amplitude in the frequency-domain components as the main frequency component f0.

4. The information-energy integrated battery cell fault early warning method according to claim 1, characterized in that, Step S3 includes: Based on the electrochemical type of the battery cell, determine the frequency range in which the electrochemical impedance of the battery cell at different SOCs is sensitive to frequency changes; select a set of discrete frequency points within the frequency range to form a key frequency point sequence for fault early warning, denoted as {f1*, f2*, …, fn*}, where n is the number of frequency points in the key frequency point sequence.

5. The information-energy integrated battery cell fault early warning method according to claim 4, characterized in that, Step S4 includes the following sub-steps: Step S4.1: Combine the main frequency component f0 with the key frequency point sequence {f1} * f2 * , …, f n * Each frequency point in the sequence is added together to form a multi-frequency sinusoidal modulation sequence corresponding to frequencies {f1, f2, …, f}. i ,…, f n }, where f i = f i * + f 0， i = 1, 2, …, n; Step S4.2: Generate a multi-frequency sinusoidal signal f from the frequencies corresponding to the multi-frequency sinusoidal modulation sequence. MFS (t): Step S4.3: Convert the multi-frequency sinusoidal signal into the multi-frequency modulation signal S(t). For full-bridge converters and half-bridge converters, the form of the multi-frequency modulation signal S(t) is determined according to the converter type.

6. The information-energy integrated battery cell fault early warning method according to claim 5, characterized in that, In step S4.3, for the full-bridge converter, the multi-frequency modulation signal S(t) consists of -1 and 1, and the generation logic is as follows: Step S4.3.1: Normalize the multi-frequency sinusoidal signal to the range (-1, 1) to obtain the normalized signal f. MFS_nom (t): Where n represents the number of key frequency points; Step S4.3.2: For each sampling time step k, calculate the error e[k] according to the following recursive relationship: e[k] = e[k - 1] + (f MFS_nom [k]–S[k – 1]); Step S4.3.3: Based on the calculated error, determine the multi-frequency modulation signal S[k] at the current moment: If e[k] > 0, then S[k] = 1; otherwise, S[k] = -1. For a half-bridge converter, the multi-frequency modulation signal S(t) consists of 0 and 1. It is obtained by linear transformation based on the generation result of the full-bridge converter. Multiplying the multi-frequency modulation signal generated by the full-bridge converter by 0.5 and then adding 0.5 will convert it into a signal consisting of 0 and 1. The different values ​​of the multi-frequency modulation signal S(t) represent different control actions on the input PCS side current. S(t)1=1 represents conducting the current to the battery cell; S(t)1=-1 represents conducting the current in reverse; S(t)1=0 represents bypassing the current.

7. The information-energy integrated battery cell fault early warning method according to claim 6, characterized in that, Step S5 includes the following sub-steps: Step S5.1: Synchronously acquire the terminal voltage v of the battery cell through the voltage and current sampling unit of the energy storage converter. DC (t) and loop current i DC (t), and store it in the data storage unit of PCS. The time length of the collected data meets the requirement that it is not less than the lowest frequency point in the key frequency point sequence; Step S5.2: Obtain the terminal voltage v from the data storage unit. DC (t) and loop current i DC (t), and calculate v using Fast Fourier Transform respectively. DC (t), i DC The frequency domain quantity V of (t) DC (f), I DC (f) Based on the key frequency point sequence, the impedance value Z of the battery cell at the corresponding key frequency point f is calculated according to the following formula. EIS (f): 。 8. The information-energy integrated battery cell fault early warning method according to claim 7, characterized in that, The methods for constructing the group cluster set and individual cluster set corresponding to step S6 are as follows: After the initial state equalization of the energy storage power station is completed, the battery system is subjected to one charge-discharge cycle, and the impedance value ZEIS(f) of each battery cell controlled by each energy storage converter PCS at each key frequency point f is obtained through step S5 under different states of charge (SOC). After uploading the impedance value ZEIS(f) of the battery cells controlled by each PCS at the key frequency point to the cloud platform, the real and imaginary parts of the impedance at each key frequency point are statistically analyzed by box plot in the cloud platform. This allows us to obtain the group cluster set and individual cluster set of the real and imaginary parts of the impedance of the battery cells controlled by each PCS under different SOCs in the initial state of the power plant.

9. The information-energy integrated battery cell fault early warning method according to claim 1, characterized in that, Step S7 includes: Based on the group cluster set, if the real-time impedance value of a certain battery cell deviates from the statistical range of the group cluster set by more than a preset threshold, the group outlier warning count of the battery cell is increased by 1; otherwise, the group outlier warning count of the battery cell is decreased by 1. When the group outlier warning count reaches a preset number, the battery cell is determined to have malfunctioned. Based on the individual cluster set of each battery cell, if the real-time impedance value of the battery cell deviates from the statistical range of the corresponding individual cluster set by more than a preset threshold, the individual outlier warning count of the battery cell is increased by 1. When the individual outlier warning count reaches a preset number, the battery cell is determined to have malfunctioned.

10. An integrated information and energy battery cell fault early warning system, employing the integrated information and energy battery cell fault early warning method according to any one of claims 1-9, characterized in that, include: Module M1: Based on the topology of the energy storage converter, determine the form of the operating current input to the energy storage PCS side during the operation of the energy storage system. The form of the operating current includes DC form or AC form containing fundamental frequency component. Module M2: Performs frequency domain analysis on the operating current of the energy storage PCS side to obtain the frequency corresponding to the component with the largest amplitude in the spectrum, which is used as the main frequency component; Module M3: Based on the electrochemical type of the battery cell, select multiple frequency points that are sensitive to impedance changes to form a sequence of key frequency points for fault early warning. Module M4: Constructs a multi-frequency modulation signal containing the multiple frequency points based on the main frequency components and the key frequency point sequence; Module M5: Synchronously samples the terminal voltage and loop current of the battery cell and calculates the impedance value of the battery cell at the key frequency point sequence; Module M6: Based on the impedance values ​​of each battery cell in the energy storage power station at the key frequency point sequence, construct a group cluster set and an individual cluster set of each battery cell. The group cluster set summarizes the impedance values ​​of all battery cells in the energy storage power station at different times, and the individual cluster set summarizes the impedance values ​​of each battery cell at multiple historical times. Module M7: Continuously monitors the real-time impedance value of each battery cell. If the cumulative number of times the impedance value of a battery cell deviates from the statistical range of the group cluster set or the individual cluster set of the battery cell itself reaches a preset number, the battery cell is determined to have malfunctioned and an early warning is issued.

Images