An energy storage system fault diagnosis method, device, equipment and storage medium

Abstract

The application discloses a kind of energy storage system fault diagnosis method, device, equipment and storage medium.The technical scheme of the present application, by constructing the causal diagram model of the multi-source operating variable of energy storage system, establishes abnormal node identification, causal edge weight estimation, intervention effect quantification, root cause contribution score and fault chain automatic backtracking mechanism, realize the transformation of energy storage system fault from abnormal alarm to root cause positioning, can distinguish cause node and response node, improve diagnostic accuracy, to solve the problem that root cause and symptom are difficult to distinguish in the prior art, traceability result is one-sided, lack of interpretability and low efficiency of on-site disposal.Meanwhile, the present application has on-line closed-loop updating capability, can adapt to the fault diagnosis needs of different energy storage stations, different working conditions and different battery aging stages.

Description

Technical Field

[0001] This invention relates to the field of energy storage system operation and maintenance technology, and in particular to a method, device, equipment and storage medium for diagnosing faults in energy storage systems. Background Technology

[0002] With the widespread application of energy storage systems in peak shaving and frequency regulation, renewable energy consumption, and electricity trading, their operational scale continues to expand, and their system structure is becoming increasingly complex. A typical energy storage system consists of battery clusters, a battery management system (BMS), a power supply system (PCS), a thermal management subsystem, a fire protection system, a combiner unit, and an energy management system (EMS). In actual operation, faults such as abnormal individual cell voltage, cluster current deviation, insulation abnormalities, abnormal temperature rise, cooling circuit failure, contactor malfunction, sampling drift, equalization failure, and communication interruption often do not occur in isolation, but rather propagate gradually through multiple mechanisms involving heat, electricity, control, and structure.

[0003] Existing fault diagnosis methods often employ threshold alarms, rule matching, correlation analysis, or classification models. While effective in detecting single-point anomalies, they often struggle to determine the cause of the anomaly, how the fault propagates, and which root cause should be addressed first in complex, coupled fault scenarios. Furthermore, operational data is characterized by multiple variables, high sampling frequency, frequent operating condition changes, high noise levels, and significant lagged correlations between variables. If the source is traced solely based on the strength of correlations, the results tend to remain at the level of accompanying symptoms, making it difficult to distinguish between root cause variables and passive response variables. Summary of the Invention

[0004] This invention provides a method, apparatus, equipment, and storage medium for fault diagnosis of energy storage systems, so as to improve the accuracy, interpretability, and efficiency of fault diagnosis and engineering handling.

[0005] According to one aspect of the present invention, a fault diagnosis method for an energy storage system is provided. The method includes:

[0006] Acquire multi-source operational data of the energy storage system, and perform robust standardization processing on the multi-source operational data to obtain standard operational data;

[0007] Based on the standard operating data, a directed causal graph containing battery status nodes, thermal management nodes, electrical execution nodes, and control command nodes is constructed, and a structural causal model between the nodes is established according to the nodes in the directed causal graph.

[0008] Based on the structural causal model, the edge set is filtered by constrained time-delay regression and conditional independence test to generate the operational causal graph of the energy storage system under the operational scenario.

[0009] When an anomaly is detected in the energy storage system, based on the operating cause-effect graph and the normal operating condition model, the abnormal deviation of each node is calculated to obtain the residual abnormality of each node, and an anomaly candidate set is constructed according to the residual abnormality of each node.

[0010] For each anomaly candidate node in the anomaly candidate set, a virtual intervention analysis is performed on the anomaly candidate node to calculate the average causal intervention effect of each anomaly candidate node on the key fault characterization node.

[0011] Based on the residual abnormality, average causal intervention effect, and predetermined propagation contribution and prior engineering information corresponding to each node, the root cause comprehensive score corresponding to each node is determined.

[0012] The nodes are sorted from largest to smallest according to the root cause comprehensive score to determine the primary root cause node. Based on the primary root cause node, the strongest propagation path from the primary root cause node to the fault representation node is searched along the directed causal graph, and the fault diagnosis result is output. The fault diagnosis result includes at least the primary root cause node, secondary related nodes, main fault propagation path, node score, location reliability, and suggested action.

[0013] According to another aspect of the present invention, a fault diagnosis device for an energy storage system is provided. The device includes:

[0014] The data standard processing module is used to acquire multi-source operating data of the energy storage system and perform robust standardization processing on the multi-source operating data to obtain standard operating data.

[0015] The causal model construction module is used to construct a directed causal graph containing battery status nodes, thermal management nodes, electrical execution nodes and control command nodes based on the standard operating data, and to establish a structural causal model between the nodes according to the nodes in the directed causal graph.

[0016] The causal graph generation module is used to generate an operational causal graph of the energy storage system under the operating scenario by filtering the edge set through constrained time-delay regression and conditional independence test based on the structural causal model.

[0017] An anomaly set construction module is used to, when an anomaly is detected in the energy storage system, calculate the anomaly deviation for each node based on the operating causal graph and the prediction of the normal operating condition model, obtain the residual anomaly degree corresponding to each node, and construct an anomaly candidate set based on the residual anomaly degree corresponding to each node.

[0018] The intervention effect determination module is used to perform virtual intervention analysis on each abnormal candidate node in the abnormal candidate set and calculate the average causal intervention effect of each abnormal candidate node on the key fault characterization node.

[0019] The comprehensive scoring module is used to determine the root cause comprehensive score for each node based on the residual abnormality, average causal intervention effect, pre-determined propagation contribution, and prior engineering information.

[0020] The diagnostic result output module is used to sort each node from largest to smallest according to the root cause comprehensive score, determine the primary root cause node, and search along the directed causal graph for the strongest propagation path from the primary root cause node to the fault characterization node based on the primary root cause node, and output the fault diagnosis result. The fault diagnosis result includes at least the primary root cause node, secondary related nodes, main fault propagation path, node score, location reliability, and suggested action.

[0021] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0022] At least one processor; and

[0023] A memory communicatively connected to the at least one processor; wherein,

[0024] The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the energy storage system fault diagnosis method according to any embodiment of the present invention.

[0025] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the energy storage system fault diagnosis method according to any embodiment of the present invention.

[0026] The technical solution of this invention constructs a causal graph model of multi-source operating variables of an energy storage system, and establishes mechanisms for anomaly node identification, causal edge weight estimation, intervention effect quantification, root cause contribution scoring, and automatic fault chain backtracking. This transforms energy storage system faults from anomaly alarms to root cause localization, distinguishing between causal nodes and response nodes, improving diagnostic accuracy, and thus solving the problems of difficulty in distinguishing root causes from symptoms, one-sided tracing results, insufficient interpretability, and low on-site handling efficiency in existing technologies. Furthermore, this invention possesses online closed-loop update capabilities, adapting to the fault diagnosis needs of different energy storage stations, different operating conditions, and different battery aging stages.

[0027] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0028] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a flowchart of a fault diagnosis method for an energy storage system provided according to an embodiment of the present invention;

[0030] Figure 2 This is a structural diagram of an energy storage system fault diagnosis device provided according to an embodiment of the present invention;

[0031] Figure 3 This is a schematic diagram of the structure of an electronic device that implements the energy storage system fault diagnosis method of the present invention. Detailed Implementation

[0032] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0033] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0034] Figure 1This is a flowchart illustrating a fault diagnosis method for an energy storage system provided in an embodiment of the present invention. This embodiment is applicable to the comprehensive diagnosis of faults in various types of energy storage systems. The method can be executed by an energy storage system fault diagnosis device, which can be implemented in hardware and / or software and can be configured in an electronic device. Figure 1 As shown, the method includes:

[0035] S101. Obtain multi-source operating data of the energy storage system, and perform robust standardization processing on the multi-source operating data to obtain standard operating data.

[0036] This invention collects energy storage system operation data at discrete time t and sampling period Δt. The operation data includes at least cluster voltage V(t), cluster current I(t), maximum temperature Tmax(t), minimum temperature Tmin(t), state of charge SOC(t), state of health SOH(t), insulation resistance Riso(t), voltage dispersion ΔV(t), temperature dispersion ΔT(t), cooling flow rate Fcool(t), PCS DC side power Pdc(t), dispatch power command Pcmd(t), contactor status K(t), and alarm tag A(t).

[0037] The original state vector at time t is defined as follows:

[0038] To eliminate the impact of differences in units and ranges on subsequent causal estimation, robust standardization is performed on each variable. Let the median of the j-th variable be... Interquartile range is The multi-source operational data will then undergo robust standardization processing, which will be achieved in the following way: ;

[0039] in, This refers to standard operating data, This refers to preventing tiny positive numbers with a denominator of zero. It means Time of the first Multiple source operational data, It refers to the median of the j-th original state vector. This refers to the interquartile range. To enhance the time delay modeling capability, a delay feature set {Zj(t-1),Zj(t-2),…,Zj(tL)} is also constructed, where L is the maximum lag order.

[0040] S102. Based on the standard operating data, construct a directed causal graph that includes battery status nodes, thermal management nodes, electrical execution nodes, and control command nodes, and establish a structural causal model between the nodes according to the nodes in the directed causal graph.

[0041] Based on the electro-thermal-control coupling mechanism of energy storage systems, a directed causal graph G=(N,E) is constructed, where N is the set of nodes and E is the set of causal edges. Nodes include at least battery state nodes, thermal management nodes, electrical execution nodes, and control command nodes. For each node i, a structural causal model is established as follows: ;

[0042] in, This refers to the state of node i at time t. It refers to the set of parent nodes of node i. This refers to the exogenous disturbance term.

[0043] The initial direction of the causal edge is preset based on physical mechanism and control logic, such as "cooling flow decreases → maximum temperature increases", "maximum temperature increases → internal resistance increases → voltage dispersion increases", and "PCS power execution deviation → current surge → temperature rise accelerates".

[0044] S103. Based on the structural causal model, the edge set is filtered by constrained time-delay regression and conditional independence test to generate the operation causal graph of the energy storage system operation scenario.

[0045] Furthermore, constrained time-delay regression and conditional independence tests are used to screen the edge set of the structural causal model, and the weighted edge coefficient wij characterizes the causal influence strength of node i on node j. After screening, the causal graph under the energy storage system operation scenario is obtained. Preferably, the edge coefficients are estimated by the following objective function: ;

[0046] in, Let be the optimal time delay from node i to node j. The coefficient is the sparse regularization coefficient, which can be used to determine the edge coefficients.

[0047] S104. When an anomaly is detected in the energy storage system, based on the operating causal graph and the prediction of the normal operating condition model, the abnormal deviation of each node is calculated to obtain the residual abnormality corresponding to each node, and an anomaly candidate set is constructed according to the residual abnormality corresponding to each node.

[0048] When an anomaly is detected in the energy storage system, the anomaly deviation degree is calculated for each node based on the constructed operational causal graph. Let the health prediction value of node i be... If the actual value is Yi(t), then the residual constant definition formula is as follows: ;

[0049] in, It refers to the residual constant of node i at time t. This refers to the standard deviation of the residuals of node i under normal operating conditions. This refers to the health prediction value predicted by the normal operating condition model. This refers to the actual value.

[0050] Further, based on the residual abnormality corresponding to each node, an anomaly index Bi(t) is determined for each node. For nodes that satisfy the anomaly index Bi(t) = 1 within W consecutive sampling points, they are determined to enter the anomaly candidate set. For example, the anomaly index is determined through a node anomaly indication function: ;

[0051] in, This is the anomaly threshold for node i. Preferably, the threshold is set by the 95th percentile of normally operating samples or by historical operation and maintenance experience. It refers to the residual constant of node i at time t.

[0052] S105. For each abnormal candidate node in the abnormal candidate set, perform virtual intervention analysis on the abnormal candidate node and calculate the average causal intervention effect of each abnormal candidate node on the key fault characterization node.

[0053] To distinguish between "root cause anomalies" and "passive response anomalies" among anomalous candidate nodes, a virtual intervention analysis is performed on them. Let node i be a candidate node, and node k be an observed key fault characterization node, such as maximum temperature, differential pressure, voltage dispersion, or protection action indicators. Then, the average causal intervention effect of node i on node k is defined as: ;

[0054] in, This refers to abnormal candidate nodes. Key Fault Characterization Nodes Average causal intervention effect This refers to intervening in node i. This refers to the value after node i is restored to the healthy baseline level, where yi is the current observed value. If the absolute value of ACEi→k is large, it indicates that when node i is corrected, the key fault characterization node can be significantly improved, the probability of node i being the root cause node of the fault increases, and it is more likely to be the root cause node.

[0055] S106. Based on the residual abnormality, average causal intervention effect, and predetermined propagation contribution and prior engineering information corresponding to each node, determine the root cause comprehensive score corresponding to each node.

[0056] Considering the path length and edge weight decay during anomaly propagation, it is necessary to determine the contribution of node i to the propagation of the system fault. For example, the formula for calculating the propagation contribution is as follows: ;

[0057] in, This refers to the propagation contribution of node i. It refers to the set of all directed paths from root cause node i to critical fault representation node k. This refers to the node anomaly indicator function of node i at time t. It refers to the path attenuation factor, which represents the cumulative attenuation effect when a fault propagates along path p.

[0058] Specifically, the root cause comprehensive score is determined in the following manner: ;

[0059] Among them, α, β, γ, and δ are weight coefficients, and α+β+γ+δ=1, The engineering prior weights for node i are used to characterize the impact of critical equipment failures on system security. This refers to the root cause comprehensive score of node i. This refers to the residual constant of node i. This refers to the absolute value of the average causal intervention effect. This refers to the propagation contribution of node i.

[0060] S107. Sort each node from largest to smallest according to the root cause comprehensive score, determine the primary root cause node, and search along the directed causal graph for the strongest propagation path from the primary root cause node to the fault characterization node based on the primary root cause node, and output the fault diagnosis result.

[0061] The fault diagnosis results include at least the primary root cause node, secondary related nodes, main fault propagation path, node score, location reliability, and suggested actions.

[0062] Specifically, the abnormal candidate nodes are sorted from largest to smallest according to the root cause comprehensive score Ri(t). The node with the highest score is the first-level root cause node, and the node with the second highest score that satisfies the causal link with the first-level root cause is the second-level associated root cause node.

[0063] After obtaining the first-level root cause node, the strongest propagation path from the first-level root cause node to the fault representation node is automatically searched along the directed causal graph, and the fault propagation chain is output. Preferably, the propagation path scoring function is defined as: ;

[0064] in, For path length, This represents the path length penalty coefficient. The path with the highest score is selected as the main path for fault propagation, and the corresponding node timing evidence, edge direction description, and recommended handling order are output simultaneously.

[0065] For example, after the output fault diagnosis result, the method further includes: based on the output fault diagnosis result, performing fault handling on the energy storage system, determining the system state improvement rate after fault handling, dynamically adjusting the prior weights and causal edge weights of relevant nodes in the causal graph, and realizing online closed-loop update.

[0066] To adapt to the differences in energy storage operation under different seasons, aging stages, and control strategies, changes in key indicators before and after each fault handling are collected, and the causal edge weights and threshold parameters are updated. Let the improvement rate of key characterization nodes after handling be Q(t), defined as: ;

[0067] If the processing node matches the positioning result of this invention and Q(t) is higher than the reference threshold Qref, then the prior weight Pi of the corresponding node is increased, and the relevant causal edges are weighted; if Q(t) is low, then the weight of the corresponding edge is reduced or the threshold is adjusted. This forms an online closed-loop mechanism of "location—processing—verification—update".

[0068] The technical solution of this invention constructs a causal graph model of multi-source operating variables of an energy storage system, and establishes mechanisms for anomaly node identification, causal edge weight estimation, intervention effect quantification, root cause contribution scoring, and automatic fault chain backtracking. This transforms energy storage system faults from anomaly alarms to root cause localization, distinguishing between causal nodes and response nodes, improving diagnostic accuracy, and thus solving the problems of difficulty in distinguishing root causes from symptoms, one-sided tracing results, insufficient interpretability, and low on-site handling efficiency in existing technologies. Furthermore, this invention possesses online closed-loop update capabilities, adapting to the fault diagnosis needs of different energy storage stations, different operating conditions, and different battery aging stages.

[0069] This embodiment uses a grid-side energy storage power station as an example. The power station has a rated power of 50MW and a rated capacity of 100MWh. Each battery compartment is equipped with a liquid cooling system, a cluster-level BMS, and compartment-level fire suppression and temperature control devices. The system sampling period is 10s. During operation, a composite anomaly was detected in one battery compartment: "continuously rising maximum temperature, widening temperature difference, increased cluster voltage dispersion, and PCS power limitation alarm." Automatic root cause identification is required.

[0070] (a) Parameter settings

[0071] In this embodiment, the maximum hysteresis order L is set to 6, corresponding to a maximum propagation delay of 60s; the anomaly detection window W is set to 6; the sparsity regularization coefficient λ is set to 0.08; the root cause scoring weights are set to α=0.30, β=0.35, γ=0.25, and δ=0.10, respectively; the path length penalty coefficient η is set to 0.12; and the key fault characterization nodes are selected as the maximum temperature Tmax, the temperature dispersion ΔT, and the power limitation flag Plimit.

[0072] Candidate root cause nodes selected in the cause-effect graph include: cooling flow rate (Fcool), circulating pump status (Pump), unit internal resistance characterization (Rcell), temperature sensor offset (Tsensor), PCS execution deviation (Pdev), and contactor status (K). The engineering prior weights are set as follows: Pcool = 0.90, Ppump = 0.95, Pcell = 0.85, Psensor = 0.55, Pdev = 0.70, and Pk = 0.65.

[0073] (II) Cause-effect graph and anomaly node identification process

[0074] Suppose the following key data were collected at fault time t0: Fcool(t0) = 18 L / min, significantly lower than the healthy baseline of 32 L / min; Tmax(t0) = 41.8℃, compared to the healthy baseline of 33.5℃; ΔT(t0) = 7.1℃, compared to the healthy baseline of 2.4℃; ΔV(t0) = 86mV, compared to the healthy baseline of 31mV; Pdc(t0) is 9.6% lower than Pcmd(t0). Based on the normal operating condition model prediction, the corresponding residual abnormalities are as follows: ;

[0075] Based on the anomaly threshold determination, Fcool, Pump, Rcell, and Pdev are identified as anomaly candidates. According to the physical mechanism and the constrained time-delay regression results, the main causal edge weights are: w(Fcool,Tmax)=0.82, w(Pump,Fcool)=0.79, w(Tmax,ΔT)=0.74, w(Tmax,ΔV)=0.61, w(ΔT,Plimit)=0.56, w(ΔV,Plimit)=0.48.

[0076] (III) Intervention effect calculation and root cause scoring process

[0077] The key failure characterization node was selected as Plimit, and a virtual intervention analysis was performed on each candidate node. Assuming that the corresponding node is restored to a healthy baseline, the average causal intervention effect is as follows: ;

[0078] Based on the definition of contribution to propagation and considering the main propagation paths, we can conclude that: ; ; ;

[0079] Substituting the above results into the formula (root cause score) yields: ; ; ; ;

[0080] Therefore, the root cause ranking result is: Fcool > Pump > Rcell > Pdev. Combining the causal chain direction "Pump→Fcool→Tmax→ΔT / ΔV→Plimit", it can be further determined that the first-level root cause is abnormal cooling flow, and the second-level related root cause is abnormal operation of the circulation pump.

[0081] (iv) Generation of propagation paths and explanation of results

[0082] Based on the path scoring function, the main propagation path was selected as "abnormal circulation pump status → decreased cooling flow rate → increased maximum temperature → increased temperature dispersion → power limitation alarm". On-site maintenance personnel prioritized checking the liquid-cooled circulation pump and cooling branches, discovering that a filter component in one branch was clogged, causing the flow rate decrease. After cleaning and restoration, the key indicators changed within 30 minutes as follows: Tmax decreased from 41.8℃ to 35.2℃, ΔT decreased from 7.1℃ to 2.8℃, ΔV decreased from 86mV to 39mV, and the power limitation alarm disappeared.

[0083] Based on the improvement rate formula, and taking the power limitation deviation corresponding to Plimit as the key characteristic, the deviation before treatment was 9.6%, and the deviation after treatment was 1.8%, then: ;

[0084] Since Q is greater than the reference threshold Qref=0.40, it indicates that the positioning result of this invention is consistent with the on-site handling result, and the positioning is effective. Statistics show that in 26 compound anomaly events at this energy storage station over the past three months, the accuracy rate of primary root cause localization using the method of this invention reached 88.46%, and the average positioning time was reduced from 27 minutes for manual analysis to less than 4 minutes.

[0085] (v) Example of online closed-loop update

[0086] In this embodiment, when multiple consecutive events demonstrate that the "Pump→Fcool" link has a significant causal contribution and the handling verification is effective, the edge weight w(Pump,Fcool) is increased, for example, from 0.79 to 0.83; simultaneously, the engineering prior weights of relevant nodes in the cooling system are appropriately increased. If a false temperature rise alarm caused by sensor drift subsequently occurs, the causal edge confidence between Tsensor and Tmax is reduced based on the feedback of ineffective handling. This allows for continuous correction of model parameters, improving the stability of root cause localization under different seasons and equipment conditions.

[0087] Figure 2 This is a schematic diagram of the structure of an energy storage system fault diagnosis device provided in an embodiment of the present invention. Figure 2 As shown, the device includes:

[0088] The data standard processing module 301 is used to acquire multi-source operating data of the energy storage system and perform robust standardization processing on the multi-source operating data to obtain standard operating data.

[0089] The causal model construction module 302 is used to construct a directed causal graph containing battery status nodes, thermal management nodes, electrical execution nodes and control command nodes based on the standard operating data, and to establish a structural causal model between the nodes according to the nodes in the directed causal graph.

[0090] The causal graph generation module 303 is used to generate an operational causal graph of the energy storage system operation scenario by filtering the edge set through constrained time-delay regression and conditional independence test based on the structural causal model.

[0091] The anomaly set construction module 304 is used to, when an anomaly is detected in the energy storage system, calculate the anomaly deviation for each node based on the operating causal graph and the prediction of the normal operating condition model, obtain the residual anomaly degree corresponding to each node, and construct an anomaly candidate set according to the residual anomaly degree corresponding to each node.

[0092] The intervention effect determination module 305 is used to perform virtual intervention analysis on each abnormal candidate node in the abnormal candidate set and calculate the average causal intervention effect of each abnormal candidate node on the key fault characterization node.

[0093] The comprehensive scoring module 306 is used to determine the root cause comprehensive score for each node based on the residual abnormality, average causal intervention effect, pre-determined propagation contribution, and prior engineering information.

[0094] The diagnostic result output module 307 is used to sort each node from largest to smallest according to the root cause comprehensive score, determine the primary root cause node, and search along the directed causal graph for the strongest propagation path from the primary root cause node to the fault characterization node based on the primary root cause node, and output the fault diagnosis result. The fault diagnosis result includes at least the primary root cause node, secondary related nodes, main fault propagation path, node score, location reliability, and suggested action.

[0095] The technical solution of this invention constructs a causal graph model of multi-source operating variables of an energy storage system, and establishes mechanisms for anomaly node identification, causal edge weight estimation, intervention effect quantification, root cause contribution scoring, and automatic fault chain backtracking. This transforms energy storage system faults from anomaly alarms to root cause localization, distinguishing between causal nodes and response nodes, improving diagnostic accuracy, and thus solving the problems of difficulty in distinguishing root causes from symptoms, one-sided tracing results, insufficient interpretability, and low on-site handling efficiency in existing technologies. Furthermore, this invention possesses online closed-loop update capabilities, adapting to the fault diagnosis needs of different energy storage stations, different operating conditions, and different battery aging stages.

[0096] Optionally, the device further includes an online closed-loop update module. This module is used for:

[0097] After the output fault diagnosis results are obtained, the energy storage system is subjected to fault handling based on the output fault diagnosis results, and the system state improvement rate after fault handling is determined. The prior weights and causal edge weights of relevant nodes in the causal graph are dynamically adjusted to achieve online closed-loop update.

[0098] Optionally, the structural causal model is as follows: ;

[0099] in, This refers to the state of node i at time t. It refers to the set of parent nodes of node i. This refers to the exogenous disturbance term.

[0100] Optionally, the calculation of abnormal deviations for each node to obtain the residual abnormality corresponding to each node is achieved in the following way: ;

[0101] in, It refers to the residual constant of node i at time t. This refers to the standard deviation of the residuals of node i under normal operating conditions. This refers to the health prediction value predicted by the normal operating condition model. This refers to the actual value.

[0102] Optionally, the formula for calculating the average causal intervention effect is as follows: ;

[0103] in, This refers to abnormal candidate nodes. Key Fault Characterization Nodes Average causal intervention effect This refers to intervening in node i. This refers to the value after node i is restored to the healthy baseline level, where yi is the current observation value. If the absolute value of ACEi→k is large, it indicates that when node i is corrected, the key fault characterization node can be significantly improved, and the probability that node i is the root cause node of the fault increases.

[0104] Optionally, the formula for calculating the propagation contribution is as follows: ;

[0105] in, This refers to the propagation contribution of node i. It refers to the set of all directed paths from root cause node i to critical fault representation node k. This refers to the node anomaly indicator function of node i at time t. It refers to the path attenuation factor, which represents the cumulative attenuation effect when a fault propagates along path p.

[0106] Optionally, the root cause comprehensive score is determined in the following manner: ;

[0107] Among them, α, β, γ, and δ are weight coefficients, and α+β+γ+δ=1, The engineering prior weights for node i are used to characterize the impact of critical equipment failures on system security. This refers to the root cause comprehensive score of node i. This refers to the residual constant of node i. This refers to the absolute value of the average causal intervention effect. This refers to the propagation contribution of node i.

[0108] The energy storage system fault diagnosis device provided in the embodiments of the present invention can execute the energy storage system fault diagnosis method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.

[0109] Figure 3 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0110] like Figure 3As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) 12 or a random access memory (RAM) 13, communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded from storage unit 18 into the RAM 13. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0111] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0112] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as energy storage system fault diagnosis methods.

[0113] In some embodiments, the energy storage system fault diagnosis method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded into and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the energy storage system fault diagnosis method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the energy storage system fault diagnosis method by any other suitable means (e.g., by means of firmware).

[0114] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), system-on-a-chip (SoCs), complex programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0115] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0116] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0117] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0118] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0119] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through a communication network. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0120] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0121] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. A fault diagnosis method for an energy storage system, characterized in that, include: Acquire multi-source operational data of the energy storage system, and perform robust standardization processing on the multi-source operational data to obtain standard operational data; Based on the standard operating data, a directed causal graph containing battery status nodes, thermal management nodes, electrical execution nodes, and control command nodes is constructed, and a structural causal model between the nodes is established according to the nodes in the directed causal graph. Based on the structural causal model, the edge set is filtered by constrained time-delay regression and conditional independence test to generate the operational causal graph of the energy storage system under the operational scenario. When an anomaly is detected in the energy storage system, based on the operating cause-effect graph and the normal operating condition model, the abnormal deviation of each node is calculated to obtain the residual abnormality of each node, and an anomaly candidate set is constructed according to the residual abnormality of each node. For each anomaly candidate node in the anomaly candidate set, a virtual intervention analysis is performed on the anomaly candidate node to calculate the average causal intervention effect of each anomaly candidate node on the key fault characterization node. Based on the residual abnormality, average causal intervention effect, and predetermined propagation contribution and prior engineering information corresponding to each node, the root cause comprehensive score corresponding to each node is determined. The nodes are sorted from largest to smallest according to the root cause comprehensive score to determine the primary root cause node. Based on the primary root cause node, the strongest propagation path from the primary root cause node to the fault representation node is searched along the directed causal graph, and the fault diagnosis result is output. The fault diagnosis result includes at least the primary root cause node, secondary related nodes, main fault propagation path, node score, location reliability, and suggested action.

2. The method according to claim 1, characterized in that, After the output fault diagnosis results, the method further includes: based on the output fault diagnosis results, performing fault handling on the energy storage system, determining the system state improvement rate after fault handling, dynamically adjusting the prior weights and causal edge weights of relevant nodes in the causal graph, and realizing online closed-loop update.

3. The method according to claim 1, characterized in that, The structural causal model is as follows: ； in, This refers to the state of node i at time t. It refers to the set of parent nodes of node i. This refers to the exogenous disturbance term.

4. The method according to claim 1, characterized in that, The calculation of abnormal deviations for each node, resulting in the residual abnormality for each node, is achieved in the following manner: ； in, It refers to the residual constant of node i at time t. This refers to the standard deviation of the residuals of node i under normal operating conditions. This refers to the health prediction value predicted by the normal operating condition model. This refers to the actual value.

5. The method according to claim 1, characterized in that, The formula for calculating the average causal intervention effect is as follows: ； in, This refers to abnormal candidate nodes. Key Fault Characterization Nodes Average causal intervention effect This refers to intervening in node i. This refers to the value after node i is restored to the healthy baseline level, where yi is the current observation value. If the absolute value of ACEi→k is large, it indicates that when node i is corrected, the key fault characterization node can be significantly improved, and the probability that node i is the root cause node of the fault increases.

6. The method according to claim 1, characterized in that, The formula for calculating the propagation contribution is as follows: ； in, This refers to the propagation contribution of node i. It refers to the set of all directed paths from root cause node i to critical fault representation node k. This refers to the node anomaly indicator function of node i at time t. It refers to the path attenuation factor, which represents the cumulative attenuation effect when a fault propagates along path p.

7. The method according to claim 1, characterized in that, The root cause comprehensive score is determined in the following manner: ； Among them, α, β, γ, and δ are weight coefficients, and α+β+γ+δ=1, The engineering prior weights for node i are used to characterize the impact of critical equipment failures on system security. This refers to the root cause comprehensive score of node i. This refers to the residual constant of node i. This refers to the absolute value of the average causal intervention effect. This refers to the propagation contribution of node i.

8. A fault diagnosis device for an energy storage system, characterized in that, include: The data standard processing module is used to acquire multi-source operating data of the energy storage system and perform robust standardization processing on the multi-source operating data to obtain standard operating data. The causal model construction module is used to construct a directed causal graph containing battery status nodes, thermal management nodes, electrical execution nodes and control command nodes based on the standard operating data, and to establish a structural causal model between the nodes according to the nodes in the directed causal graph. The causal graph generation module is used to generate an operational causal graph of the energy storage system under the operating scenario by filtering the edge set through constrained time-delay regression and conditional independence test based on the structural causal model. An anomaly set construction module is used to, when an anomaly is detected in the energy storage system, calculate the anomaly deviation for each node based on the operating causal graph and the prediction of the normal operating condition model, obtain the residual anomaly degree corresponding to each node, and construct an anomaly candidate set based on the residual anomaly degree corresponding to each node. The intervention effect determination module is used to perform virtual intervention analysis on each abnormal candidate node in the abnormal candidate set and calculate the average causal intervention effect of each abnormal candidate node on the key fault characterization node. The comprehensive scoring module is used to determine the root cause comprehensive score for each node based on the residual abnormality, average causal intervention effect, pre-determined propagation contribution, and prior engineering information. The diagnostic result output module is used to sort each node from largest to smallest according to the root cause comprehensive score, determine the primary root cause node, and search along the directed causal graph for the strongest propagation path from the primary root cause node to the fault characterization node based on the primary root cause node, and output the fault diagnosis result. The fault diagnosis result includes at least the primary root cause node, secondary related nodes, main fault propagation path, node score, location reliability, and suggested action.

9. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the energy storage system fault diagnosis method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the energy storage system fault diagnosis method according to any one of claims 1-7.

Images