Search

CN-122017635-A - Information energy integrated battery unit fault early warning method and system

CN122017635ACN 122017635 ACN122017635 ACN 122017635ACN-122017635-A

Abstract

The invention provides an information energy integrated battery unit fault early warning method and system, wherein the method comprises the steps of firstly determining a mode of an energy storage system for inputting working current at a PCS side, 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 battery unit terminal voltage and loop current through the PCS, calculating impedance values corresponding to the key frequency points, constructing a group cluster set, constructing an individual cluster set, continuously monitoring real-time impedance values, and judging that a battery unit breaks down and giving early warning when the accumulated times of a certain battery unit deviating from the group cluster set or the individual cluster set reach preset times. According to the invention, the on-line injection and impedance detection of the excitation signal are realized through the PCS, the battery state information acquisition is completed while the energy is transmitted, the early warning precision, the instantaneity and the economy are both considered, and the operation safety of the energy storage system is effectively improved.

Inventors

  • LI RUI
  • PENG CHENG

Assignees

  • 上海交通大学

Dates

Publication Date
20260512
Application Date
20260129

Claims (10)

  1. 1. The information and energy integrated battery unit fault early warning method is characterized by comprising the following steps of: step S1, determining a working current form of an input energy storage PCS side when an energy storage system operates according to a topological structure of an energy storage converter, wherein the working current form comprises a direct current form or an alternating current form containing a fundamental wave component; s2, carrying out frequency domain analysis on the working current at the energy storage PCS side to obtain the frequency corresponding to the component with the largest amplitude in the frequency spectrum as a main frequency component; step S3, selecting a plurality of frequency points sensitive to impedance change according to the electrochemical type of the battery unit to form a key frequency point sequence for fault early warning; S4, constructing a multi-frequency point modulation signal containing the plurality of frequency points according to the main frequency component and the key frequency point sequence; step S5, synchronously sampling terminal voltage and loop current of the battery unit, and calculating impedance values of the battery unit in the key frequency point sequence; Step S6, based on the impedance values of each battery unit in the energy storage power station on the key frequency point sequence, a group clustering set and an individual clustering set of each battery unit are constructed, wherein the group clustering set summarizes the impedance values of all the battery units in the energy storage power station at different moments, and the individual clustering set summarizes the impedance values of each battery unit at a plurality of historical moments; And S7, continuously monitoring the real-time impedance value of each battery unit, and if the accumulated times of the deviation of the impedance value of each battery unit from the statistical range of the group cluster set or the individual cluster set of each battery unit reaches the preset times, judging that the battery unit fails and giving out early warning.
  2. 2. The information and energy integrated battery unit fault pre-warning method according to claim 1, wherein the step S1 includes: According to the topological structure of the energy storage PCS, the energy storage PCS is divided into a serial PCS, a distributed PCS and a cascade PCS, wherein in the serial PCS, each battery unit is directly connected to an alternating current bus through a single-stage DC-AC converter, in the distributed PCS, each battery unit is connected to a direct current bus through a DC-DC converter after being connected in parallel and then is connected to the alternating current bus through a high-capacity DC-AC converter, in the cascade PCS, each battery unit is connected in cascade through the single-stage converter, the working current form of the PCS side is determined according to the bus form of bridge arm connection obtained after cascading, if the working current form is a direct current form, and if the working current form is an alternating current form containing alternating current fundamental wave components.
  3. 3. The method for early warning of faults of an information and energy integrated battery unit according to claim 1, wherein the method for acquiring the main frequency component in the step S2 is to convert a time domain waveform of an operating current input to an energy storage PCS side into a frequency domain component by using a fast fourier transform, and according to a component with a maximum amplitude in the frequency domain component, taking a frequency corresponding to the maximum amplitude component as the main frequency component f 0 .
  4. 4. The information-energy integrated battery cell malfunction alerting method according to claim 1, wherein the step S3 comprises: And selecting a group of discrete frequency points in the frequency interval to form a key frequency point sequence for fault early warning, wherein the key frequency point sequence is denoted as { f1 x, f2 x, fn x }, and n is the number of frequency points in the key frequency point sequence.
  5. 5. The information and energy integrated battery unit fault pre-warning method according to claim 4, wherein the step S4 comprises the following sub-steps: Step S4.1, adding the main frequency component f 0 to each frequency point in the key frequency point sequence { f 1 * , f 2 * , …, f n * } to form a frequency { f 1 , f 2 , …,f i ,…, f n } corresponding to the multi-frequency point sinusoidal modulation sequence, wherein f i = f i * + f 0, i=1, 2, n; Step S4.2, generating a multi-frequency point sinusoidal signal f MFS (t) by using the corresponding frequency of the multi-frequency point sinusoidal modulation sequence: and S4.3, converting the multi-frequency point sinusoidal signals into multi-frequency point modulation signals S (t), wherein the form of the multi-frequency point modulation signals S (t) is determined according to the type of the converter for the full-bridge converter and the half-bridge converter.
  6. 6. The method for early warning of failure of an information and energy integrated battery unit according to claim 5, wherein in the step S4.3, for a full-bridge converter, the multi-frequency point modulation signal S (t) is composed of-1 and 1, and the generation logic is as follows: Step S4.3.1, normalizing the multi-frequency point sinusoidal signal to be in a range (-1, 1) to obtain a normalized signal f MFS_nom (t): Wherein n represents the number of key frequency points; Step S4.3.2, for each sampling time step k, calculating an error e [ k ] according to the recurrence relation: e[k] = e[k - 1] + (f MFS_nom [k]–S[k – 1]); step S4.3.3, according to the calculated error, determining the multi-frequency point modulation signal S [ k ] at the current moment: if e [ k ] >0, S [ k ] = 1, otherwise S [ k ] = -1; For the half-bridge converter, the multi-frequency point modulation signal S (t) is composed of 0 and 1, is obtained by linear transformation based on the generation result of the full-bridge converter, and can be converted into a signal composed of 0 and 1 by multiplying the multi-frequency point modulation signal generated by the full-bridge converter by 0.5 and adding 0.5; The different values of the multi-frequency point modulation signal S (t) represent different control actions on the current at the input PCS side, s1=1 represents that the current is conducted to the battery unit, s1= -1 represents that the current is conducted reversely, and s1=0 represents that the current is bypassed.
  7. 7. The information and energy integrated battery unit fault pre-warning method according to claim 6, wherein the step S5 comprises the following sub-steps: Step S5.1, synchronously acquiring terminal voltage v DC (t) and loop current i DC (t) of a battery unit through a voltage and current sampling unit of an energy storage converter, and storing the terminal voltage v DC (t) and the loop current i DC (t) into a data storage unit of a PCS (personal digital System), wherein the time length for acquiring data meets the requirement of not less than the lowest frequency point in the key frequency point sequence; Step S5.2, obtaining the terminal voltage V DC (t) and the loop current i DC (t) from the data storage unit, calculating the frequency domain quantity V DC (f)、I DC (f) of V DC (t)、i DC (t) by adopting fast Fourier transform respectively, calculating the impedance value Z EIS (f) of the battery unit at the corresponding key frequency point f according to the following formula based on the key frequency point sequence: 。
  8. 8. the information and energy integrated battery unit fault early warning method according to claim 7, wherein the group cluster set and individual cluster set construction method corresponding to step S6 is as follows: After the initial state equalization of the energy storage power station is completed, carrying out one-time charge-discharge cycle on the battery system, and obtaining impedance values ZEIS (f) of battery units managed by the PCS of each energy storage converter at each key frequency point f through step S5 under different charge states SOC; After uploading impedance value ZEIS (f) of each PCS-controlled battery unit at the key frequency point to a cloud platform, carrying out box line diagram statistics on the real part and the imaginary part of impedance of each key frequency point in the cloud platform, so that a group clustering set and an individual clustering set of the real part and the imaginary part of impedance of each PCS-controlled battery unit under different SOCs (system of charge) can be obtained in the initial state of a power station.
  9. 9. The information-energy integrated battery cell malfunction alerting method according to claim 1, wherein the step S7 includes: Based on the group clustering set, if the real-time impedance value of a certain battery unit deviates from the statistical range of the group clustering set by more than a preset threshold value, the group outlier early warning count of the battery unit is increased by 1, otherwise, the group outlier early warning count of the battery unit is reduced by 1, and when the group outlier early warning count reaches the preset times, the battery unit is judged to have faults; Based on the individual clustering set of each battery unit, if the real-time impedance value of each battery unit deviates from the statistical range of the corresponding individual clustering set and exceeds a preset threshold value, the individual outlier early warning count of each battery unit is increased by 1, and when the individual outlier early warning count reaches the preset times, the battery unit is judged to be faulty.
  10. 10. An information-energy integrated battery unit fault early warning system, which adopts the information-energy integrated battery unit fault early warning method as set forth in any one of claims 1 to 9, and is characterized by comprising: The module M1 is used for determining a working current form of an input energy storage PCS side when the energy storage system operates according to the topological structure of the energy storage converter, wherein the working current form comprises a direct current form or an alternating current form containing a fundamental wave component; The module M2 is used for carrying out frequency domain analysis on the working current at the energy storage PCS side to obtain the frequency corresponding to the component with the largest amplitude in the frequency spectrum as a main frequency component; The module M3 is used for selecting a plurality of frequency points sensitive to impedance change according to the electrochemical type of the battery unit to form a key frequency point sequence for fault early warning; A module M4, which constructs a multi-frequency point modulation signal containing the frequency points according to the main frequency component and the key frequency point sequence; a module M5, synchronously sampling the terminal voltage and the loop current of the battery unit, and calculating the impedance value of the battery unit in the key frequency point sequence; The module M6 is used for constructing a group clustering set and an individual clustering set of each battery unit based on the impedance value of each battery unit in the energy storage power station on the key frequency point sequence, wherein the group clustering set is used for summarizing the impedance values of all the battery units in the energy storage power station at different moments, and the individual clustering set is used for summarizing the impedance values of each battery unit at a plurality of historical moments; And the module M7 is used for continuously monitoring the real-time impedance value of each battery unit, and judging that the battery unit fails and giving out early warning if the accumulated times of the impedance value of the battery unit deviating from the statistical range of the group cluster set or the individual cluster set of the battery unit reaches the preset times.

Description

Information energy integrated battery unit fault early warning method and system Technical Field The invention relates to the field of energy storage system control and protection, in particular to an information and energy integrated battery unit fault early warning method and system. Background Along with the acceleration of global energy transformation process, the permeability of renewable energy sources in an electric power system is continuously improved, and the lithium ion battery energy storage system has become core support equipment for scenes such as peak load regulation, renewable energy source consumption, micro-grid stable control and the like of a smart grid by virtue of the advantages of high response speed, high energy density, long cycle life and the like, and the operation safety and stability of the lithium ion battery energy storage system directly determine the overall operation efficiency of the electric power system. Currently, the single-machine capacity of an energy storage power station has been developed on a large scale of tens to hundreds of megawatts, and a single power station generally contains hundreds of thousands of battery cells inside, and the cells form battery cells in series-parallel connection to serve as the minimum operation unit of an energy storage system. The battery cell is influenced by factors such as natural attenuation of battery cell performance, data acquisition errors of a Battery Management System (BMS), complex and changeable operation conditions (charge-discharge multiplying power and ambient temperature), and the like, and fault hidden dangers such as abnormal increase of internal resistance, deterioration of consistency of monomers, thermal runaway and the like are easily caused in the long-term service process of the battery cell. Because the energy storage system is large in scale and numerous in units, the occurrence probability of faults is obviously increased, once local faults fail to timely early warn to cause unplanned outage of the system, the generated power shortage will cause huge impact on the stable frequency and voltage quality of a power grid and even cause cascading faults, and therefore accurate and timely fault early warning of battery units becomes a key technical bottleneck for restricting the improvement of the reliability of the energy storage system. The technical scheme of early warning for the faults of the battery units in the industry is mainly divided into three types, wherein the first type is a method based on a battery model and a state estimation algorithm, and the core of the method is to realize fault diagnosis by constructing a battery equivalent circuit model and identifying parameters on line and combining the deviation of key state quantities such as a state of charge (SOC). The second type is a method based on data driving and artificial intelligence, and early warning is realized by mining the battery operation rule and abnormal characteristics through a machine learning and deep learning model by relying on historical operation data. The third category is a method based on additional sensing hardware, which directly captures the physical characteristic change of the battery by adding a special sensor. In summary, the prior art schemes are limited by BMS computing power and data accuracy, or cannot be considered to be practical and popular due to hardware cost and algorithm complexity. It is worth noting that in the series-type parallel connection power storage system, each group of batteries corresponds to an independent PCS unit, and the PCS can acquire and process voltage, current, power and other data of a power grid side and a battery side in real time in the running process, and has strong data processing and real-time control capabilities. However, the prior art fails to fully utilize the inherent functions of the PCS, and can not realize the deep sensing and early fault accurate early warning of the running state of the battery unit through the PCS on the premise of not adding additional sensing hardware and not remarkably increasing the computing load of the BMS, so that the intelligent diagnosis potential of the PCS is wasted and the intelligent diagnosis potential is only used as power instruction execution equipment. The invention patent with publication number CN115963408A is found by searching patent literature, and discloses a single battery fault early warning system and method of an energy storage power station, wherein the system comprises a communication manager, a communication manager and a control unit, wherein the communication manager is used for respectively acquiring equipment history data of each single battery in the energy storage power station; the system comprises a server, a user behavior data processing module and a user behavior data processing module, wherein the server is used for processing and analyzing the equipment history data and the user behavior data, determining