Fault detection method and energy storage system

Abstract

The application discloses a fault detection method and an energy storage system, and relates to the technical field of energy storage systems.The fault detection method comprises the following steps: determining fault information of the energy storage system; constructing a fault causal relationship graph of the energy storage system corresponding to the fault information; and determining the final fault cause of the energy storage system according to the fault causal relationship graph. The method can determine the fault cause leading to the fault of the energy storage system.

Description

Technical Field

[0001] This invention relates to the field of energy storage system technology, and in particular to a fault detection method and an energy storage system. Background Technology

[0002] In related technologies, existing energy storage systems identify faults in the energy storage system through deep learning. Deep learning mainly learns the correlation between training data and fault labels, but does not take into account the causal relationship between training data, thus failing to determine the root cause of the fault in the energy storage system. Summary of the Invention

[0003] The present invention aims to at least solve one of the technical problems existing in the prior art. Therefore, one object of the present invention is to provide a fault detection method that can determine the cause of a fault in an energy storage system.

[0004] The second objective of this invention is to propose an energy storage system.

[0005] To address the aforementioned problems, a first aspect of the present invention provides a fault detection method for an energy storage system. The method includes: determining fault information of the energy storage system; constructing a fault causal relationship map of the fault information corresponding to the energy storage system; and determining the final cause of the fault in the energy storage system based on the fault causal relationship map.

[0006] According to the fault detection method of the present invention, by constructing a fault causal relationship map of the energy storage system corresponding to the fault information, the final fault cause of the energy storage system can be identified through the fault causal relationship map. Thus, compared with the prior art, which can only identify the fault type of the energy storage system, the present application can identify the fundamental fault cause of the energy storage system through the fault causal relationship map, thereby realizing the location of the cause of the energy storage system fault, while avoiding the problem of mistaking correlation for causation.

[0007] In some embodiments, constructing the fault causal relationship map of the energy storage system includes: constructing the fault causal relationship map based on the theoretical operating data and historical fault data of the energy storage system.

[0008] In some embodiments, constructing the fault causal relationship map based on the theoretical operating data and historical fault data of the energy storage system includes: determining an initial causal relationship based on the theoretical operating data; correcting the initial causal relationship based on the historical fault data to obtain a corrected causal relationship; determining a supplementary causal relationship based on the historical fault data; and constructing the fault causal relationship map based on the corrected causal relationship and the supplementary causal relationship.

[0009] In some embodiments, constructing the fault causality map based on the modified causality and the supplementary causality includes: determining the causal objects in the modified causality and the supplementary causality and the causal guidance relationship between the causal objects; and constructing the fault causality map based on the causal objects and the causal guidance relationship.

[0010] In some embodiments, determining the final cause of failure of the energy storage system based on the fault causality graph includes: obtaining multiple hypothetical causes of failure based on the fault causality graph; and determining the final cause of failure of the energy storage system based on the multiple hypothetical causes of failure.

[0011] In some embodiments, determining the final cause of failure of the energy storage system based on multiple hypothetical causes includes: constructing a digital twin model of the energy storage system, wherein the digital twin model is based on the actual operating data of the energy storage system; simulating each hypothetical cause of failure based on the digital twin model to obtain simulation operating data corresponding to each hypothetical cause of failure; and determining the cause of failure of the energy storage system based on the simulation operating data corresponding to each hypothetical cause of failure.

[0012] In some embodiments, determining the cause of failure of the energy storage system based on the simulation operation data corresponding to each hypothetical cause of failure includes: determining the causal strength of each hypothetical cause of failure; determining the data matching degree between the simulation operation data and the actual operation data corresponding to each hypothetical cause of failure; and determining the cause of failure of the energy storage system based on the causal strength of each hypothetical cause of failure and the data matching degree of each hypothetical cause of failure.

[0013] In some embodiments, simulating each hypothetical cause of failure based on the digital twin model includes: determining simulation control parameters for each hypothetical cause of failure; and controlling the digital twin model based on the simulation control parameters for each hypothetical cause of failure to simulate each hypothetical cause of failure.

[0014] In some embodiments, the method further includes: determining the fault propagation path based on the fault cause of the energy storage system and the fault causal relationship map.

[0015] A second aspect of the present invention provides an energy storage system, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the at least one processor executes the computer program to implement the fault detection method described in the above embodiments.

[0016] According to the energy storage system of the present invention, by executing the fault detection method of the above embodiments, the cause of the fault in the energy storage system can be determined.

[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0018] The above and / or additional aspects and advantages of the present invention will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is a flowchart of a fault detection method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a fault causal relationship diagram according to an embodiment of the present invention; Figure 3 This is a schematic diagram of data transmission according to an embodiment of the present invention; Figure 4 This is a flowchart of a fault detection method according to another embodiment of the present invention; Figure 5 This is a structural block diagram of an energy storage system according to an embodiment of the present invention.

[0019] Figure label: Energy storage system 10; Processor 1; Memory 2; Acquisition Module 3; Digital Twin Model Module 4; Fault Detection Module 5; Cause-Effect Graph Model Construction Module 6; Fault Cause Location Module 7; Fault Propagation Path Prediction Module 8; Maintenance Decision Generation Module 9. Detailed Implementation

[0020] The embodiments of the present invention are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary. The embodiments of the present invention are described in detail below.

[0021] Magnetic levitation flywheel energy storage systems, as a novel physical energy storage technology, offer advantages such as fast response speed, long cycle life, and environmental friendliness, and have been widely applied in fields such as power system frequency regulation, uninterruptible power supplies, and rail transportation. However, magnetic levitation flywheel energy storage systems have a complex structure, comprising multiple core components such as a rotor, magnetic bearings, a high-speed motor, a vacuum system, and a power electronic converter. Failure of any one of these components can lead to the shutdown of the entire energy storage system and even cause safety accidents.

[0022] To address the aforementioned problems, a first aspect of the present invention provides a fault detection method that can determine the cause of a fault in an energy storage system.

[0023] The following is for reference. Figure 1 Describe a fault detection method according to embodiments of the present invention, such as Figure 1 As shown, the method includes steps S1-S3.

[0024] Step S1: Determine the fault information of the energy storage system.

[0025] The energy storage system is a flywheel energy storage system, which can be a magnetic levitation flywheel energy storage system. Fault information includes the fault type and corresponding fault characteristic data.

[0026] Specifically, deep learning models are used to identify fault information in the energy storage system, including abnormal bearing vibration or decreased vacuum. However, since deep learning models only learn data correlations, they are insufficient to pinpoint the final cause of the fault. Therefore, a fault causal relationship graph is needed for final fault cause analysis.

[0027] Step S2: Construct a fault causal relationship map of the energy storage system corresponding to the fault information.

[0028] Among them, the fault causal relationship map of the energy storage system is a map constructed for different faults of the energy storage system. The fault causal relationship map includes the development link from the cause of the fault to the fault.

[0029] Step S3: Determine the final cause of failure of the energy storage system based on the fault causal relationship diagram.

[0030] Specifically, since the fault is located at the result node in the fault causal relationship graph, and the final cause of the fault is located at the cause node in the fault causal relationship graph, the result node is located at the end of the fault causal relationship graph, and the cause node is located at the beginning of the fault causal relationship graph. The fault causal relationship graph is a directed graph, where the directed edges in the fault causal relationship graph represent the direction of fault propagation. Then, the cause node in the fault causal relationship graph is searched in reverse by the direction of the edges. Thus, the cause node can be traced from the fault causal relationship graph through the fault information, so as to obtain the final cause of the energy storage system's fault through the cause node. The final cause of the fault is the root cause of the fault.

[0031] According to the fault detection method of the present invention, by constructing a fault causal relationship map of the energy storage system corresponding to the fault information, the final fault cause of the energy storage system can be identified through the fault causal relationship map. Thus, compared with the prior art, which can only identify the fault type of the energy storage system, the present application can identify the fundamental fault cause of the energy storage system through the fault causal relationship map, thereby realizing the location of the cause of the energy storage system fault, while avoiding the problem of mistaking correlation for causation.

[0032] In some embodiments, constructing a fault causal relationship map of an energy storage system includes: constructing a fault causal relationship map based on theoretical operating data and historical fault data of the energy storage system.

[0033] Specifically, the causal relationships of energy storage system failures are obtained through theoretical operating data and historical fault data of the energy storage system. A fault causal relationship map is then drawn based on these causal relationships. Thus, this application constructs a fault causal relationship map using theoretical operating data and historical fault data, so that the fault causal relationship map has physical and logical support, fits the actual faults, and enables rapid location of the final cause of the energy storage system failure.

[0034] In some embodiments, a fault causality map is constructed based on the theoretical operating data and historical fault data of the energy storage system, obtained through the following steps: First, the theoretical operating data can be analyzed using analytical methods to extract the initial causal relationships within the data. For example, in the initial causal relationship, the cause could be rotor mass imbalance, and the effect could be increased radial vibration. In the initial causal relationship, the cause could be vacuum pump wear, and the effect could be decreased vacuum level.

[0035] The analysis method can be FMEA (Failure Mode and Effects Analysis), which can analyze all potential failure modes of energy storage systems, the causes of failures, and the consequences of failures.

[0036] Secondly, the initial causal relationship is corrected based on historical fault data to obtain a corrected causal relationship.

[0037] Specifically, the initial causal relationship is corrected based on historical fault data to obtain a corrected causal relationship. Since historical fault data is the fault data when the energy storage system has experienced faults during its operation over a period of time, the relationship between fault data can be determined through historical fault data. This relationship can then be used to correct the initial causal relationship. In other words, verifying or correcting the initial causal relationship through historical fault data can effectively improve the accuracy of the initial causal relationship.

[0038] For example, if the initial causal relationship is that rotor mass imbalance leads to an increase in first-harmonic vibration, which in turn leads to an increase in radial displacement, and the data on first-harmonic vibration and radial displacement in historical fault data show that the increase in first-harmonic vibration precedes the increase in radial displacement, then the relationship that the increase in first-harmonic vibration leads to the increase in radial displacement is retained in the initial causal relationship.

[0039] Then, supplementary causal relationships are determined based on historical failure data.

[0040] Specifically, since historical fault data represents data from when the energy storage system experienced faults during its operation over a period of time, causal relationships can be obtained by analyzing the relationships between fault data using the Peter-Clark Algorithm (PC) and Ridge Regression Causal Discovery Algorithm. These causal relationships can serve as supplementary causal relationships, where the supplementary causal relationships can be intermediate transmission relationships between causal objects in the initial causal relationship. For example, if the analysis of historical fault data reveals that the bearing current increases with the increase in radial displacement, then the increase in bearing current due to the increase in radial displacement can be used as a corrected causal relationship, which can effectively expand the number of nodes when constructing the fault causal relationship graph.

[0041] Finally, a fault causal relationship map is constructed based on the corrected causal relationships and the supplementary causal relationships.

[0042] Specifically, the corrected causal relationships and supplementary causal relationships are plotted into a fault causal relationship graph. That is, the causal objects in the corrected and supplementary causal relationships are used as nodes. For example, the causal objects and the resulting faults in the corrected and supplementary causal relationships can be used as nodes. The causal objects can be located in the middle or at both ends of the causal relationship. Then, the directed causal relationships in the corrected and supplementary causal relationships are connected. Therefore, this application constructs a fault causal relationship graph by using corrected and supplementary causal relationships, thereby greatly enriching the number of nodes and edge connection density in the fault causal relationship graph while ensuring the accuracy of the causal relationships, and can cover more scenarios for fault detection in energy storage systems.

[0043] In some embodiments, constructing a fault causality map based on corrected causal relationships and supplementary causal relationships includes: determining the causal objects and causal guidance relationships between the causal objects in the corrected causal relationships and supplementary causal relationships; and constructing a fault causality map based on the causal objects and causal guidance relationships.

[0044] Specifically, the causal objects and causal orientation relationships between them are determined in the modified causal relationship and the supplementary causal relationship. The causal objects can be cause nodes, intermediate state nodes, and effect nodes. The cause node is the inferred cause of the failure, and the effect node is the failure of the energy storage system. There are causal relationships between the cause node, intermediate state node, and effect node. Based on this, all causal objects in the modified causal relationship and the supplementary causal relationship are extracted first, and then the causal orientation relationships between the causal objects are established. Multiple causal objects are connected according to the causal orientation relationships between them to construct a failure causal relationship graph. The causal objects serve as nodes in the failure causal relationship graph, and the edges of the failure causal relationship graph are drawn through the causal orientation relationships. That is, the edges reflect the causal orientation between the causal objects. Thus, this application uses the failure causal relationship graph to intuitively reflect the causal relationship, so that the final failure cause can be quickly located through the failure causal relationship graph.

[0045] For example, such as Figure 2 The diagram shows a fault causal relationship. It connects rotor mass imbalance with increased unbalanced centrifugal force; increases unbalanced centrifugal force with increased first-harmonic vibration; increases first-harmonic vibration with increased radial displacement; increases radial displacement with abnormal radial vibration and increased magnetic bearing current; rotor misalignment with increased first-harmonic vibration and increased radial displacement; bearing installation deviation with increased radial displacement; and improper control parameters with increased magnetic bearing current. When the fault in the fault causal relationship diagram is abnormal radial vibration, it will lead to a protective shutdown.

[0046] Furthermore, there are no specific restrictions on the number of causal objects, nor are there specific restrictions on the number of cause nodes, intermediate state nodes, and effect nodes.

[0047] In some embodiments, determining the final cause of failure of the energy storage system based on a fault causality graph includes: obtaining multiple hypothetical causes of failure based on the fault causality graph; and determining the final cause of failure of the energy storage system based on the multiple hypothetical causes of failure.

[0048] Among these, the hypothetical cause of failure is a candidate cause for the final cause of failure. The final cause of failure is the actual cause of failure determined from multiple hypothetical causes.

[0049] Specifically, since a certain fault is located at a result node in the fault causality graph, and the fault causality graph is a directed graph, the directions of the edges are traced back to multiple cause nodes in the fault causality graph. The contents of these multiple cause nodes are used as multiple inferred causes of the fault. Thus, multiple inferred causes of the fault can be obtained from the fault causality graph based on the specific fault. These inferred causes are then analyzed to determine the most probable cause as the final cause of the energy storage system's failure. Therefore, this application uses a fault causality graph to intuitively represent causal relationships, enabling rapid location of the final cause of the energy storage system's failure.

[0050] For example, after detecting radial vibration anomaly in the energy storage system, the fault propagation path is determined by the fault causal relationship graph corresponding to the radial vibration anomaly. In the fault causal relationship graph, the radial vibration anomaly is located at the result node. By tracing back along the fault causal relationship graph along the radial vibration anomaly, multiple inferred fault causes such as rotor mass imbalance, rotor misalignment, bearing installation deviation, and improper control parameters are generated.

[0051] In some embodiments, determining the final cause of failure of an energy storage system based on multiple hypothetical causes includes: constructing a digital twin model of the energy storage system, wherein the digital twin model is based on actual operating data of the energy storage system; simulating each hypothetical cause of failure based on the digital twin model to obtain simulation operating data corresponding to each hypothetical cause of failure; and determining the final cause of failure of the energy storage system based on the simulation operating data corresponding to each hypothetical cause of failure.

[0052] In this embodiment, the digital twin model includes a rotor dynamics model, a magnetic bearing model, a high-speed motor model, a vacuum system model, and a converter model. The rotor dynamics model is a 12-DOF rotor dynamics model established using the finite element method, which considers common fault modes such as rotor mass imbalance, misalignment, cracks, and bending. The magnetic bearing model is a closed-loop control model for magnetic bearings that includes electromagnetic force nonlinearity, PID (Proportional-Integral-Derivative) control algorithm, and displacement sensor noise. The high-speed motor model is a permanent magnet synchronous motor model that considers stator winding temperature distribution, rotor eddy current losses, and air gap eccentricity. The vacuum system model is a dynamic model of the vacuum system that includes vacuum pump pumping speed characteristics, pipe flow resistance, and leakage rate. The converter model is a three-phase full-bridge converter model that considers IGBT (Insulated Gate Bipolar Transistor) switching characteristics, bus voltage fluctuations, and dead-zone effects.

[0053] Specifically, the digital twin model of an energy storage system is a virtual model built in simulation software that is identical to the actual energy storage system. This virtual model can reproduce the actual operation of the energy storage system. Different parameters are set for the digital twin model to predict different causes of failure. Then, the actual operating data of the energy storage system is input into the digital twin model. The digital twin model simulates each predicted cause of failure, that is, it simulates the operating process corresponding to each predicted cause of failure to obtain simulated operating data for each predicted cause. The simulated operating data for each predicted cause of failure is compared with the actual operating data to determine the final cause of failure of the energy storage system. For example, the predicted cause of failure corresponding to simulated operating data similar to the actual operating data is taken as the final cause of failure of the energy storage system. Therefore, in this application, the final cause of failure of the energy storage system can be located through the simulated operating data corresponding to each predicted cause of failure.

[0054] In this embodiment, the actual operating data is collected from multiple sources of sensors, and the time synchronization of this multi-source actual operating data is processed. For example, the IEEE 1588 PTP (Precision Time Protocol) is used to achieve microsecond-level time synchronization of all sensor data. The multiple sources of sensors include: mechanical parameter sensors, electrical parameter sensors, and special parameter sensors. The mechanical parameter sensors include vibration sensors, displacement sensors, and speed sensors. The sampling frequency of the vibration sensor is 20kHz, the displacement sensor is 10kHz, and the speed sensor is 1kHz. The electrical parameter sensors include current sensors, voltage sensors, and power sensors. The sampling frequency of the current sensor is 5kHz, the voltage sensor is 5kHz, and the power sensor is 100Hz. The special parameter sensors are a vacuum sensor and a temperature sensor. The vacuum sensor has a sampling frequency of 1Hz, and the temperature sensor has a sampling frequency of 1Hz.

[0055] In some embodiments, determining the final cause of failure of the energy storage system based on the simulation operation data corresponding to each hypothetical cause of failure includes: determining the causal strength of each hypothetical cause of failure; determining the data matching degree between the simulation operation data and the actual operation data corresponding to each hypothetical cause of failure; and determining the final cause of failure of the energy storage system based on the causal strength of each hypothetical cause of failure and the data matching degree of each hypothetical cause of failure.

[0056] Specifically, a higher causal strength of the inferred cause indicates a stronger causal relationship within the inferred cause. A higher data matching degree between the simulated and actual operating data corresponding to the inferred cause indicates a higher probability that the inferred cause is the final cause of failure in the energy storage system. Based on this, conditional probability is used to calculate the causal strength of each inferred cause, and dynamic time warping algorithm is used to calculate the data matching degree between the simulated and actual operating data corresponding to each inferred cause. Then, the probability of each inferred cause being the final cause is calculated by comprehensively considering both the causal strength and the data matching degree. This allows for the determination of the most probable inferred cause from among multiple inferred causes, which is then used as the final cause of failure in the energy storage system. Therefore, this application quantifies the probability of each inferred cause being the final cause by measuring its causal strength and data matching degree, effectively improving the accuracy of determining the final cause of failure.

[0057] For example, when an abnormal radial vibration fault is detected, multiple hypothetical causes include rotor mass imbalance, rotor misalignment, bearing installation deviation, and improper control parameters. The causal strength of rotor mass imbalance is 0.85, and the data matching degree between the simulated and actual operating data corresponding to rotor mass imbalance is determined to be 0.92. Therefore, the score for rotor mass imbalance is the product of causal strength and data matching degree, with a product value of 0.782. Similarly, the causal strength of rotor misalignment is 0.72, and the data matching degree between the simulated and actual operating data corresponding to rotor misalignment is determined to be 0.65. Therefore, the score for rotor misalignment is the product of causal strength and data matching degree, with a product value of 0.468. The causal strength of bearing installation deviation is 0.68, and the data matching degree between the simulation running data and the actual running data corresponding to bearing installation deviation is determined to be 0.58. Therefore, the score of bearing installation deviation is the product of causal strength and data matching degree, with a product value of 0.394. The causal strength of improper control parameters is 0.85, and the data matching degree between the simulation running data and the actual running data corresponding to improper control parameters is determined to be 0.92. Therefore, the score of improper control parameters is the product of causal strength and data matching degree, with a product value of 0.782. The inferred fault cause with the highest score among the above inferred fault causes is rotor mass imbalance. Therefore, rotor mass imbalance is taken as the final fault cause of the energy storage system.

[0058] In some embodiments, simulating each hypothetical cause of failure based on a digital twin model includes: determining simulation control parameters for each hypothetical cause of failure; and controlling the digital twin model based on the simulation control parameters for each hypothetical cause of failure to simulate each hypothetical cause of failure.

[0059] Specifically, simulation control parameters are determined for each hypothetical fault cause. These simulation control parameters include fault control parameters, which are determined by analyzing fault data, fault mechanisms, and normal operating data. For example, two value ranges are determined based on fault data and fault mechanisms. After correcting these value ranges using normal operating data, the overlapping portion of the two corrected value ranges is selected as the parameter range. For instance, simulation control parameters corresponding to rotor mass imbalance include unbalanced mass and eccentricity; simulation control parameters corresponding to bearing installation deviation include air gap deviation and installation position deviation; and simulation parameters corresponding to improper control parameters include PID parameter deviation. The operation of the digital twin model is controlled according to the simulation control parameters for each hypothetical fault cause, enabling the model to simulate hypothetical fault causes that closely resemble reality. Thus, this application uses simulation control parameters for hypothetical fault causes to control the digital twin model to simulate the operating process corresponding to the hypothetical fault cause, enabling the model to simulate faults in a real energy storage system.

[0060] In this embodiment, when the difference between the simulated operating data and the actual operating data of the digital twin model exceeds a preset threshold, an anomaly is determined in the energy storage system. A pre-trained CNN (Convolutional Neural Network) is then used to identify the fault type, such as abnormal radial vibration, abnormal axial displacement, or a decrease in vacuum. Therefore, even when fault samples are scarce, the digital twin model can be used to identify fault types, improving the reliability of fault diagnosis.

[0061] In this embodiment, the digital twin model can be calibrated online based on actual operating data and using a Bayesian optimization algorithm.

[0062] In some embodiments, the method further includes: determining the fault propagation path based on the final cause of failure and the fault causal relationship map of the energy storage system.

[0063] Specifically, after determining the ultimate cause of failure in the energy storage system, the failure propagation path is determined from the failure causal relationship graph based on the ultimate cause of failure. In the failure causal relationship graph, the ultimate cause of failure of the energy storage system is located at the cause node. By propagating the failure path forward along the failure causal relationship graph based on the ultimate cause of failure of the energy storage system, at least one intermediate state node is obtained, until the result node of the failure causal relationship graph is traced back. Thus, the failure propagation path is obtained through the ultimate cause of failure of the energy storage system, so as to generate targeted maintenance plans based on the failure propagation path. This allows for proactive measures to be taken through maintenance plans to prevent the failure from escalating and improve the reliability and safety of the energy storage system.

[0064] For example, for a fault caused by rotor mass imbalance, the fault propagation path is: rotor mass imbalance, increased radial vibration, magnetic bearing current fluctuations, and decreased control accuracy, ultimately leading to increased rotor axial displacement. This increased axial displacement triggers a protective shutdown of the energy storage system. The generated maintenance plan for this fault includes: shutting down the system for rotor dynamic balancing; or checking the wear of the magnetic bearings; or recalibrating the displacement sensor. After performing maintenance according to the plan, the energy storage system is subjected to an no-load test run to verify whether the fault has been eliminated.

[0065] In an embodiment, such as Figure 3 As shown, after the acquisition module 3 acquires the actual operating data, it inputs the actual operating data into the digital twin model module 4. The digital twin model module 4 simulates the operating process corresponding to each hypothetical fault cause through the digital twin model to obtain the simulation operating data corresponding to each hypothetical fault cause. The fault detection module 5 calculates the difference between the simulation operating data and the actual operating data corresponding to each hypothetical fault cause. After determining that the energy storage system has failed based on the difference, the cause-effect graph model construction module 6 obtains multiple hypothetical fault causes through the fault cause-effect graph. The fault cause localization module 7 locates the final fault cause of the energy storage system from the multiple hypothetical fault causes. The fault propagation path prediction module 8 obtains the fault propagation path through the final fault cause of the energy storage system, so that the maintenance decision generation module 9 can generate a targeted maintenance plan through the fault propagation path.

[0066] The following is for reference. Figure 4 The fault detection method of the present invention will be described in detail below.

[0067] Step S4: An anomaly in the energy storage system is detected and the fault type is identified.

[0068] Step S5: Generate multiple hypothetical causes of failure based on the fault causal relationship graph.

[0069] Step S6: Simulate the operational process corresponding to each hypothesized cause of failure using a digital twin model.

[0070] Step S7: Calculate the data matching degree between the simulation running data and the actual running data corresponding to each speculated cause of failure.

[0071] Step S8: Determine if there are any unsimulated hypothetical causes of the fault. If yes, proceed to step S9; otherwise, proceed to step S10.

[0072] Step S9: Select the next possible cause of the fault and proceed to step S7.

[0073] Step S10: Determine the final cause of failure of the energy storage system by data matching degree and causal strength.

[0074] A second aspect of the present invention provides an energy storage system, such as... Figure 5 As shown, the energy storage system 10 includes at least one processor 1 and at least one memory 2.

[0075] In this embodiment, at least one processor is communicatively connected to at least one processor's memory; the memory stores a computer program that can be executed by at least one processor, and when the at least one processor executes the computer program, it implements the fault detection method of the above embodiment.

[0076] According to the energy storage system of the present invention, by executing the fault detection method of the above embodiments, the cause of the fault in the energy storage system can be determined.

[0077] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example.

[0078] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A fault detection method, characterized in that, For use in energy storage systems, the method includes: Determine the fault information of the energy storage system; Construct a fault causal relationship map of the energy storage system corresponding to the fault information; The final cause of failure of the energy storage system is determined based on the fault causal relationship diagram.

2. The fault detection method according to claim 1, characterized in that, The construction of the fault causal relationship map of the energy storage system includes: The fault causal relationship map is constructed based on the theoretical operating data and historical fault data of the energy storage system.

3. The fault detection method according to claim 2, characterized in that, The step of constructing the fault causal relationship map based on the theoretical operating data and historical fault data of the energy storage system includes: Determine the initial causal relationship based on the theoretical operating data; The initial causal relationship is corrected based on the historical fault data to obtain a corrected causal relationship; Supplementary causal relationships were determined based on the historical fault data. The fault causal relationship map is constructed based on the corrected causal relationship and the supplementary causal relationship.

4. The fault detection method according to claim 3, characterized in that, The step of constructing the fault causal relationship map based on the modified causal relationship and the supplementary causal relationship includes: Determine the causal objects and the causal orientation relationships between the causal objects in the modified causal relationship and the supplementary causal relationship; The fault causal relationship map is constructed based on the causal object and the causal guidance relationship.

5. The fault detection method according to claim 1, characterized in that, Determining the final cause of failure of the energy storage system based on the fault causal relationship diagram includes: Based on the fault causality diagram, several hypothetical causes of the fault were obtained; The final cause of failure of the energy storage system was determined based on multiple hypothetical causes of failure.

6. The fault detection method according to claim 5, characterized in that, The process of determining the final cause of failure of the energy storage system based on multiple hypothetical causes includes: A digital twin model of the energy storage system is constructed, wherein the digital twin model is based on the actual operating data of the energy storage system as input; The digital twin model is used to simulate each hypothetical cause of failure in order to obtain simulation running data corresponding to each hypothetical cause of failure. The final cause of failure of the energy storage system is determined based on the simulation operation data corresponding to each hypothesized cause of failure.

7. The fault detection method according to claim 6, characterized in that, The step of determining the final cause of failure of the energy storage system based on the simulation operation data corresponding to each hypothetical cause of failure includes: Determine the causal strength of each hypothetical cause of failure; Determine the data matching degree between the simulation running data and the actual running data corresponding to each speculated cause of failure; The final cause of failure of the energy storage system is determined based on the causal strength of each hypothetical cause and the data matching degree of each hypothetical cause.

8. The fault detection method according to claim 6, characterized in that, The simulation of each hypothetical cause of failure based on the digital twin model includes: Determine the simulation control parameters for each hypothesized cause of the fault; The digital twin model is controlled according to the simulation control parameters of each hypothetical cause of failure in order to simulate each hypothetical cause of failure.

9. The fault detection method according to claim 1, characterized in that, The method further includes: The fault propagation path is determined based on the final cause of failure of the energy storage system and the fault causal relationship diagram.

10. An energy storage system, characterized in that, include: At least one processor; A memory that is communicatively connected to at least one of the processors; The memory stores a computer program that can be executed by at least one of the processors, and when the at least one processor executes the computer program, it implements the fault detection method according to any one of claims 1-9.

Images