Early warning method, system and equipment of lithium battery energy storage power station, medium and product

Abstract

The invention discloses an early warning method, system and device for a lithium battery energy storage power station, a medium and a product, and the method comprises the steps: obtaining the initial operation data of each layer of equipment in the lithium battery energy storage power station in a preset period, and determining a plurality of early warning factors based on the initial operation data; determining an association factor having an association relationship with an early warning factor in a preset early warning adjacent interval from a pre-constructed knowledge graph; if the correlation factor has an early warning record in a preset period, historical operation data and target operation data are extracted from record information of the early warning factor to calculate the operation state deviation degree of the early warning factor, and when the operation state deviation degree is higher than a preset deviation threshold value, the operation state deviation degree and the abnormal frequency of the correlation factor are calculated; the target early warning threshold value of the early warning factor is determined based on the abnormal frequency, and the early warning information of the lithium battery energy storage power station is generated based on the target operation data and the early warning threshold value, so that the early warning accuracy can be improved.

Description

Technical Field

[0001] This invention relates to the field of power systems, and more particularly to early warning methods, systems, equipment, media, and products for lithium battery energy storage power stations. Background Technology

[0002] Lithium-ion battery energy storage power stations are key facilities for building new power systems, and their safe and stable operation is crucial. However, lithium-ion batteries themselves carry the risk of thermal runaway, and energy storage power stations operate in complex environments with dynamically changing conditions. Multiple factors, such as battery aging, electrical equipment failure, and thermal management malfunctions, can easily trigger chain reactions, leading to severe equipment damage or even fires and explosions. Therefore, real-time and accurate online early warning systems for energy storage power stations are of great practical significance for ensuring the safety of power station assets, the stable operation of the power grid, and the safety of life and property of people in the surrounding area.

[0003] Currently, the commonly used early warning methods in the industry mainly rely on monitoring static thresholds for key operating parameters such as voltage, current, and temperature. When a parameter exceeds a fixed limit, the system triggers an alarm. However, this method has significant accuracy flaws. First, static thresholds cannot adapt to dynamic factors such as equipment aging, seasonal changes, and fluctuations in operating conditions. Setting them too high can lead to missed detections of early, minor anomalies, while setting them too low can easily generate numerous false alarms due to normal fluctuations. Second, traditional methods rely on single-point indicator judgments and lack system-level correlation analysis, failing to capture potential interconnected anomaly patterns between devices, thus resulting in untimely or inaccurate early warnings. Summary of the Invention

[0004] This invention provides a method, system, equipment, medium, and product for early warning of lithium battery energy storage power stations, which can dynamically adjust the early warning threshold of early warning factors in lithium battery energy storage power stations to improve the accuracy of early warning.

[0005] An embodiment of the present invention provides an early warning method for a lithium battery energy storage power station, comprising: Acquire initial operating data of each level of equipment in a lithium battery energy storage power station within a preset period, and determine several early warning factors based on the initial operating data; From the pre-constructed knowledge graph, identify the associated factors that are related to the warning factors located in the vicinity of the preset warning range; If the associated factor has a warning record within a preset period, then the historical operating data and target operating data of the warning factor within the historical period are extracted from the record information of the warning factor to calculate the operating status deviation of the warning factor. When the operating status deviation is higher than a preset deviation threshold, the operating status deviation and the abnormal frequency of the associated factor are calculated to determine the target warning threshold of the warning factor based on the abnormal frequency, and the warning information of the lithium battery energy storage power station is generated based on the target operating data and the warning threshold.

[0006] This invention, by acquiring initial operational data from multiple levels and determining early warning factors, can capture more comprehensive and nuanced operational status information, avoiding inaccurate judgments due to incomplete data. Through a pre-constructed knowledge graph, it automatically retrieves other related factors with dependencies or subordinate relationships to the early warning factors, elevating early warning judgment from isolated single-point indicators to a system-level network of connections. This enables early detection of potential risks arising from inter-device linkage effects, solving the problem of difficulty in capturing potential abnormal linkage patterns between devices and achieving a preventative effect. After identifying related factors, it further checks whether these factors have early warning records and calculates the deviation of the current early warning factor's operational curve from the historical baseline, achieving precise quantification and verification of abnormal states and effectively identifying abnormal states masked by improper threshold settings. By dynamically adjusting the early warning threshold, the threshold can adaptively adjust with dynamic factors such as equipment aging and changes in operating conditions, completely solving the inherent defects of "static early warning thresholds." Ultimately, the system is able to generate warning information based on the adjusted target warning thresholds that better reflect the current actual operating status, thereby significantly reducing missed and false alarms at the source and significantly improving the accuracy of warnings.

[0007] Furthermore, the determination of several early warning factors based on the initial operating data includes: The initial running data is standardized to obtain the processing result; Read the first warning factor from the processing result; The processing results are calculated using a predefined mathematical model to generate a corresponding second warning factor, wherein the warning factor includes the first warning factor and the second warning factor.

[0008] By acquiring initial operational data from multiple levels and determining early warning factors, more comprehensive and detailed operational status information can be captured, avoiding inaccurate judgments due to incomplete data.

[0009] Furthermore, the second warning factor includes a health score, the calculation process of which includes: Based on the first early warning factor, construct the feature vector of the component; The deviation of the feature vector from the baseline of similar devices is calculated to obtain the lateral deviation, and the deviation of the feature vector from its own historical baseline is calculated to obtain the longitudinal deviation. The baseline of similar devices is determined by statistically analyzing the average health feature vector of similar devices of the component within a historical period, and the historical baseline of the component is determined by statistically analyzing the average health feature vector of the component within several historical periods. The overall health score of the component is calculated based on the weighted sum of the lateral deviation and the longitudinal deviation.

[0010] By constructing component feature vectors and calculating their deviation from the baseline of similar devices (horizontal) and their own historical baseline (vertical), a comprehensive health score can be generated, which can more comprehensively and deeply reflect the evolution of component health status and provide key input for subsequent knowledge graph association analysis.

[0011] Furthermore, the step of determining the correlation factors from the pre-constructed knowledge graph that are associated with the warning factors located in the preset warning proximity interval includes: By traversing the knowledge graph, factor nodes or health score nodes that are in the preset warning proximity range and are connected to the warning factor through edges are retrieved. The knowledge graph uses the warning factor and health score of the lithium battery energy storage power station as nodes and the relationship between nodes as edges. The early warning factors corresponding to the factor nodes and health score nodes are determined as the correlation factors.

[0012] By using a pre-built knowledge graph, other related factors that have a dependency or subordinate relationship with the warning factor can be automatically retrieved. This can elevate the warning judgment from an isolated single-point indicator to a system-level network of relationships, enabling the early detection of potential risks caused by the linkage effect between devices. This solves the problem of difficulty in capturing potential abnormal linkage patterns between devices and achieves a preventive effect of preventing problems before they occur.

[0013] Further, the step of extracting historical operating data and target operating data of the warning factor within a historical period from the recorded information of the warning factor to calculate the deviation of the operating state of the warning factor includes: Extract the historical operation data and target operation data of the warning factor within the historical period from the recorded information of the warning factor; A first operating curve and a second operating curve are generated based on the historical operating data and the target operating data, respectively. The deviation of the operating status of the early warning factor is determined based on the area difference or statistical deviation value between the first operating curve and the second operating curve.

[0014] After identifying the correlated factors, we further check whether these correlated factors have warning records and calculate the deviation of the current warning factor's operating curve from the historical baseline. This enables precise quantification and verification of abnormal states and effectively identifies those abnormal states that are masked by improper threshold settings.

[0015] Further, the step of calculating the deviation of the operating state and the abnormal frequency of the correlation factor, so as to determine the target early warning threshold of the early warning factor based on the abnormal frequency, includes: Calculate the abnormal frequency of each of the aforementioned correlation factors; Based on the deviation of the operating state and the frequency of each abnormality, the comprehensive impact value is calculated through the factor influence function; Based on the deviation of the operating status and the comprehensive impact value, the initial warning threshold is adjusted using a sensitivity factor to obtain the target warning threshold of the warning factor.

[0016] By dynamically adjusting the warning threshold, the threshold can be adaptively adjusted according to dynamic factors such as equipment aging and changes in operating conditions, thus completely solving the inherent defects of the "static warning threshold".

[0017] Another embodiment of the present invention provides an early warning system for a lithium battery energy storage power station, comprising: The acquisition module is used to acquire the initial operating data of each level of equipment in the lithium battery energy storage power station within a preset period, and to determine several early warning factors based on the initial operating data. The determination module is used to identify, from a pre-built knowledge graph, related factors that are associated with warning factors located in the vicinity of a preset warning range; The early warning module is used to extract historical operating data and target operating data of the early warning factor within a historical period from the record information of the early warning factor if the associated factor has an early warning record within a preset period, so as to calculate the operating status deviation of the early warning factor. When the operating status deviation is higher than a preset deviation threshold, the module calculates the operating status deviation and the abnormal frequency of the associated factor, so as to determine the target early warning threshold of the early warning factor based on the abnormal frequency, and generates early warning information for the lithium battery energy storage power station based on the target operating data and the early warning threshold.

[0018] Another embodiment of the present invention provides a terminal device, including: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements the steps of the early warning method for a lithium battery energy storage power station as described in the present invention.

[0019] Another embodiment of the present invention provides a computer-readable storage medium item, including: a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located executes the early warning method of the lithium battery energy storage power station as described in the present invention.

[0020] Another embodiment of the present invention provides a computer program product, including a computer program or instructions, characterized in that, when the computer program or instructions are executed by a communication device, they implement the steps of the early warning method for a lithium battery energy storage power station as described in the present invention. Attached Figure Description

[0021] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating one embodiment of the early warning method for a lithium battery energy storage power station provided in this application; Figure 2 This is a schematic diagram of the structure of one embodiment of the early warning method for a lithium battery energy storage power station provided in this application; Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application; the terms “comprising” and “having”, and any variations thereof, in the specification, claims, and foregoing description of the drawings are intended to cover non-exclusive inclusion.

[0025] In the description of the embodiments of this application, technical terms such as "first" and "second" are used only to distinguish different objects and should not be construed as indicating or implying relative importance or implicitly specifying the number, specific order, or primary and secondary relationship of the indicated technical features. In the description of the embodiments of this application, "multiple" means two or more, unless otherwise explicitly defined.

[0026] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0027] In the description of the embodiments in this application, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this document generally indicates that the preceding and following related objects have an "or" relationship.

[0028] In the description of the embodiments of this application, the term "multiple" refers to two or more (including two), similarly, "multiple sets" refers to two or more (including two sets), and "multiple pieces" refers to two or more (including two pieces).

[0029] In the description of the embodiments of this application, unless otherwise expressly specified and limited, technical terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. For those skilled in the art, the specific meaning of the above terms in the embodiments of this application can be understood according to the specific circumstances.

[0030] Lithium-ion battery energy storage power stations are key facilities for building new power systems, and their safe and stable operation is of paramount importance. Real-time and accurate online early warning for energy storage power stations is of great practical significance for ensuring the safety of power station assets, the stable operation of the power grid, and the safety of people and property in the surrounding area. Currently, the early warning methods commonly used in the industry mainly rely on monitoring key operating parameters such as voltage, current, and temperature by setting static thresholds. When a parameter exceeds a fixed limit, the system triggers an alarm. This approach directly leads to untimely or inaccurate early warnings.

[0031] See Figure 1 To improve the accuracy of early warning, an embodiment of the present invention provides an early warning method for a lithium battery energy storage power station, including steps S101 to S103: Step S101: Obtain the initial operating data of each level of equipment in the lithium battery energy storage power station within a preset period, and determine several early warning factors based on the initial operating data; In some embodiments, the initial operating data of each level of equipment in the lithium battery energy storage power station within a preset period is obtained. Specifically, the initial operating data generated during the operation of the power station is collected in real time at a frequency of seconds or minutes by data acquisition modules integrated in each level of equipment such as the battery management system (BMS), energy management system (EMS), converter controller and temperature control device, and uploaded to a unified data processing platform.

[0032] It should be noted that the various levels of equipment include battery layer data, equipment layer data, and power station layer data. Specifically, the initial operating data includes, but is not limited to: BMS (battery layer data): individual cell voltage / current / temperature / internal resistance, module / cluster voltage range, temperature range, current consistency, system total voltage / current, grounding resistance, charge / discharge cycles, etc.; equipment layer data includes converter data, transformer data, and air conditioning / fan data. Converter data includes DC side voltage, DC side current, AC side voltage, AC side current, power module temperature, ambient temperature, operating mode, commissioning / shutdown status, and fault codes; transformer data includes primary side voltage, primary side current, secondary side voltage, secondary side current, grounding current, winding temperature, efficiency, power factor, operating status, and fault / alarm status; air conditioning / fan data includes operating mode (cooling / heating / ventilation / shutdown), temperature, air conditioning power / energy consumption, fan power / energy consumption, air conditioning fault codes, and fan fault status. Power station layer data includes location information, affiliated company, dispatch response, shutdown status, and damage status, etc.

[0033] In some embodiments, determining several early warning factors based on the initial operating data includes: standardizing the initial operating data to obtain a processing result; reading a first early warning factor from the processing result; and calculating the processing result using a predefined mathematical model to generate a corresponding second early warning factor, wherein the early warning factor includes the first early warning factor and the second early warning factor. Specifically, firstly, after obtaining the initial operating data, the collected initial operating data can be standardized, such as outlier removal, missing value imputation, unit conversion, and time alignment, to obtain a processing result with uniform quality that can be used for calculation. Then, after completing data standardization, the system directly reads real-time values ​​from the processing result as the first early warning factor. For example, the voltage and current of a single battery cell, the temperature of the positive and negative terminals of a battery module, the DC side voltage and current of the converter, and the temperature of the transformer windings are all examples of directly collected early warning factors. Finally, by automatically executing mathematical models and formula calculations on the processing result using methods such as sliding window methods and integral algorithms, the periodic calculation and real-time updating of the early warning factor are completed to generate the corresponding second early warning factor.

[0034] It should be noted that the warning factors correspond to different warning dimensions, including battery cell dimension, battery module dimension, battery cluster dimension, battery system dimension, battery box dimension, inverter dimension, transformer dimension, air conditioner / fan dimension, and power station dimension.

[0035] In some embodiments, at the individual battery cell level, the internal resistance of the individual battery cell is typically acquired directly using an AC impedance test or a DC internal resistance measuring instrument; the capacity decay of the individual battery cell is an important parameter used to measure the health status of the battery. The calculation formula is: In the formula, This is the actual battery capacity as currently measured. This is the battery's rated capacity at the time of manufacture. The average rate of change of the battery's internal resistance... The formula reflecting the worsening trend of internal resistance is as follows: ,in, This represents the internal resistance value of the i-th sample. For the corresponding time point, n is the total number of samples.

[0036] In some embodiments, at the battery module level, the temperature difference between individual battery cells within the module is extremely large. The formula reflecting the temperature consistency within the module is as follows: The voltage difference between individual battery cells within the module is extremely large. This reflects the voltage consistency within the module, and the calculation formula is: ,in, This represents the temperature sensor measurement value of the i-th battery cell in the module. This represents the voltage sensor measurement value of the i-th battery cell in the module. The overall internal resistance of the battery module can be calculated by averaging or equivalently summing the internal resistances of its individual cells. If the battery cells are predominantly connected in series, the module's internal resistance... The calculation formula is: ,in, Let be the internal resistance value of the i-th battery cell in the module, and k be the total number of battery cells in the battery module.

[0037] In some embodiments, at the battery cluster level, the location of the highest-temperature cell can be found by scanning all cell temperatures to locate the cell number corresponding to the maximum value. The calculation of voltage range is similar to that of modules. Current consistency is typically assessed by evaluating the parallel consistency of each module or cluster using standard deviation. The calculation formula is: ,in, It is the average current of all battery clusters. This is the current measurement value of the i-th battery cluster. The total number of battery clusters is represented by the number of clusters. The smaller the current difference between two clusters, the better the consistency.

[0038] In some embodiments, for the battery system, the grounding resistance can be directly obtained through an insulation monitoring device, and the equivalent charge-discharge count needs to be accumulated by adding all complete equivalent cycles. The calculation formula is: In the formula, This represents the net change (absolute value) in the state of charge (SOC) during the i-th complete charge or discharge cycle. The total system current and total voltage are directly acquired quantities provided by system-level current / voltage sensors.

[0039] In some embodiments, for the battery box dimension, SOC is an estimated energy value, which can be obtained by coulomb measurement, and the calculation formula is: In the formula, This refers to the total nominal capacity of the battery. The battery current (in A) measured at time point τ, with charging being positive and discharging being negative. The state of charge at time t-1; the state of charge (SOH) can also be calculated using the capacity decay ratio or internal resistance gain ratio; the energy retention rate reflects the system's energy storage capacity, and the calculation formula is... voltage control deviation It can be expressed by normalizing the difference between the actual output voltage and the set target voltage. The calculation formula is as follows: ,in, This is the actual measured value of the AC side output voltage of the converter (PCS). This is the target voltage setpoint required by the power grid or system for the converter output. The dispatch response success rate is the ratio of successful command responses to the total number of responses. The unplanned outage coefficient is automatically derived by analyzing the system's outage event logs, and the calculation formula is: In the formula, This refers to the total unplanned downtime caused by malfunctions within the statistical period. This represents the total running time of the statistical cycle. It should be noted that the PCS charge / discharge transition time is the current direction switching delay during the charge / discharge switching period, which can be calculated from the interruption point of the real-time current change curve.

[0040] In some embodiments, for the converter dimension, DC current, voltage, and equipment temperature are all conventionally acquired parameters. For the transformer dimension, the grounding current is a value acquired by the power protection device, and the winding temperature is obtained through embedded thermocouples or PT measuring points. Efficiency... The calculated value needs to be obtained by sampling the input and output electrical parameters. The calculation formula is as follows: ,in, This refers to the active power at the transformer output. This refers to the active power at the transformer input terminal; and These are the voltage and current at the transformer output terminals, respectively. and These represent the voltage and current at the transformer input terminals, respectively.

[0041] In some embodiments, for the air conditioning / fan dimension, the operating status is determined by switch signals and energy consumption curves; the fan fault status is determined by speed feedback or abnormal alarm logs. For the power plant dimension, the power plant location information is a static configuration parameter, the damage status can be obtained from accident reports or feedback information from on-site detection equipment, and the affiliated company is an operation and management attribute field.

[0042] It should be noted that a single early warning dimension record table includes the operational data of that early warning dimension uploaded at each time point within multiple cycles.

[0043] By acquiring initial operational data from multiple levels and determining early warning factors, more comprehensive and detailed operational status information can be captured, avoiding inaccurate judgments due to incomplete data.

[0044] In some embodiments, the second warning factor includes a health score, the calculation process of which includes: constructing a feature vector of the component based on the first warning factor; calculating the deviation of the feature vector from the baseline of similar devices to obtain a lateral deviation, and calculating the deviation of the feature vector from its own historical baseline to obtain a longitudinal deviation, wherein the baseline of similar devices is determined by statistically analyzing the average health feature vector of similar devices of the component over a historical period, and the own historical baseline is determined by statistically analyzing the average health feature vector of the component over several historical periods; and calculating the comprehensive health score of the component based on the weighted sum of the lateral deviation and the longitudinal deviation. Specifically, firstly, for the first warning factor in the energy storage power station... Each component (such as a battery module, inverter, or transformer) is extracted by the system in the current period t. Each of the first early warning factors (i.e., standardized basic operating parameters) constitutes a feature vector. This feature vector is used to comprehensively characterize the operating state of the component at the current moment. The mathematical expression of this feature vector is as follows: ,in, This represents the standardized warning factor value of the k-th component at time t. This indicates the number of warning factors contained in the component. Secondly, by comparing this feature vector with two preset baselines (baselines of similar devices and its own historical baseline), the degree of anomaly in its current state is quantified, yielding the lateral deviation. and longitudinal deviation Finally, the deviations from the two dimensions are combined, and a weighted sum is used to calculate the final comprehensive health score. The calculation formula directly reflects the overall health level of the component: ,in, Represents the overall health score of the k-th component; and These are weighting coefficients, representing the degree of influence of horizontal and vertical deviations, respectively.

[0045] It should be noted that the lateral deviation is used to measure the difference between the current component and the average health level of other components of the same type. Its calculation relies on a baseline of similar equipment, which is determined by statistically analyzing the mean of the health feature vectors of similar equipment over a historical period. Covariance Matrix The calculation formula is: In the formula, For the number of devices of the same type, For the historical health feature vector (baseline vector) of the kth similar device in a healthy and normal state: Subsequently, calculate the feature vector of the current component. The deviation from this baseline is used as the lateral deviation. The calculation formula is: .

[0046] It should be noted that longitudinal deviation is used to measure the difference between the current health status of a component and its historical health status. Its calculation relies on its own historical baseline, which is determined by statistically analyzing the mean of the component's health feature vectors over M historical sampling periods. Covariance Matrix The calculation formula is: In the formula, M represents the number of historical sampling periods used to establish its own historical baseline. Let k be the feature vector of the k-th component in the i-th historical sampling period. Let be the mean vector of the historical baseline of the k-th component. Let the covariance matrix of the historical baseline of the k-th component be used; then, the eigenvector of the current component is calculated. The deviation from its own historical baseline, as the longitudinal deviation. The calculation formula is: .

[0047] By constructing component feature vectors and calculating their deviation from the baseline of similar devices (horizontal) and their own historical baseline (vertical), a comprehensive health score can be generated, which can more comprehensively and deeply reflect the evolution of component health status and provide key input for subsequent knowledge graph association analysis.

[0048] Step S102: From the pre-constructed knowledge graph, identify the associated factors that are related to the warning factors located in the preset warning neighboring interval; In some embodiments, step S102 includes: traversing the knowledge graph to retrieve factor nodes or health score nodes that are in a preset warning proximity interval and connected to the warning factor via edges, wherein the knowledge graph uses the warning factor and health score of the lithium battery energy storage power station as nodes and the relationships between nodes as edges; and determining the warning factors corresponding to the factor nodes and health score nodes as the associated factors. Specifically, firstly, a traversal query of the pre-built knowledge graph is automatically triggered. All other nodes in the knowledge graph that are in the preset warning proximity interval and directly connected to the current "warning factor" node via edges are retrieved, wherein these nodes may represent basic operating parameters or higher-level health score nodes. Finally, the system determines the warning factors corresponding to these retrieved nodes as associated factors that have a relationship with the specified warning factor. For example, when the warning dimension is "battery module", by traversing the knowledge graph, its "associated warning dimension" can be automatically determined as "battery cluster" (its superior level) and "battery cell" (its subordinate level), and the relevant factors under these dimensions (such as cluster current, cell internal resistance, etc.) are used as associated factors.

[0049] It should be noted that the warning proximity range refers to a range of adjacent values ​​where the real-time value of a warning factor is close to, but has not yet been formally triggered, the warning threshold. For example, if the warning threshold of a certain warning factor is a lower limit of 20 and an upper limit of 30, and its warning range is below 20 or above 30, then its warning proximity range can be set to [20,22] or [28,30]. It should be noted that the warning proximity range for each warning factor can be preset by those skilled in the art based on actual operating conditions and safety margins.

[0050] It should be noted that each warning factor has a corresponding warning threshold and warning range. When the real-time value of a warning factor is close to its warning threshold but has not yet triggered a warning, it is considered to be in the warning proximity range. As the warning factor changes further, it may trigger a formal warning.

[0051] It should be noted that the knowledge graph is deeply integrated with the system structure and operational logic of energy storage power stations. Its foundation is the Unified Modeling Language (UML) class diagram for energy storage power stations. This UML class diagram clearly defines the hierarchical structure and subordinate relationships between various early warning dimensions (such as individual battery cells, battery modules, and battery clusters). Specifically: individual battery cells belong to battery modules, battery modules belong to battery clusters, battery clusters belong to battery systems, battery systems belong to energy storage battery boxes, and battery boxes belong to energy storage power stations. Additionally, converters, transformers, and air conditioners / fans also belong to energy storage battery boxes. These relationships are defined as one-to-many or one-to-one associations in the UML class diagram (for example, the relationship between an energy storage power station and a battery box is one-to-many, while the relationship between a battery box and a converter is typically one-to-one). Based on this UML class diagram, a knowledge graph can be constructed. The construction process is as follows: each warning factor is used as a node, and the relationship between each warning factor is used as an edge. The knowledge graph corresponding to each warning factor is established. That is, all the warning factors of the lithium battery energy storage power station are used as nodes in the graph, and the semantic relationship between the nodes is used as the connecting edge. These relationships are directly derived from the hierarchical structure in the UML class diagram and the physical and functional dependencies between devices. They mainly include subordinate relationships, dependency relationships (such as the internal resistance of the battery module depends on the internal resistance of the battery cell, and the temperature of the converter depends on the operating status of the air conditioner), and opposition relationships.

[0052] It should be noted that the early warning factor nodes include not only basic operating parameter factors (such as voltage, current, temperature, etc.), but also component-level health scores calculated based on a multi-factor fusion model. .

[0053] It should be noted that examples of dependencies are as follows: 1. Changes in the internal resistance of a battery module may depend on changes in the internal resistance of individual battery cells, because the module's internal resistance is a comprehensive reflection of the individual cell resistances. 2. The voltage range of individual cells within a battery cluster may depend on the voltage range of individual cells within a battery module, because the voltage range of individual cells within a cluster is the sum of the voltage ranges of individual cells within a module. 3. Current consistency between battery clusters may depend on the location of the highest temperature individual cells within a battery cluster, because current consistency may be affected by temperature unevenness. 4. The success rate of dispatch response of an energy storage power station may depend on the total current and total voltage of the battery system, because the system's charging and discharging capacity directly affects the power station's dispatch response. 5. The transformer's grounding current may depend on the DC current of the converter, because the operation of the converter may affect the transformer's current distribution. 6. The converter's temperature may depend on the operating status of the air conditioner / fan, because the air conditioner / fan is responsible for regulating the converter's heat dissipation. 7. Power station location information may affect the assessment of damage, because geographical location may affect the risk of disaster and the severity of damage. Furthermore, the dependency relationship of the health scoring node can be expressed as: module-level health scoring. Dependent on individual-level health scores Battery cluster health score Dependent on module-level health scores System health score Depends on battery cluster health score Power plant health score Dependent on system health score In addition, it includes health scores for equipment such as PCS and transformers. Through these expansions, knowledge graphs can establish a unified causal dependency system among multi-level and multi-type factors. By using a pre-built knowledge graph, other related factors that have a dependency or subordinate relationship with the warning factor can be automatically retrieved. This can elevate the warning judgment from an isolated single-point indicator to a system-level network of relationships, enabling the early detection of potential risks caused by the linkage effect between devices. This solves the problem of difficulty in capturing potential abnormal linkage patterns between devices and achieves a preventive effect of preventing problems before they occur.

[0054] Step S103: If the associated factor has a warning record within a preset period, then extract the historical operating data and target operating data of the warning factor within the historical period from the record information of the warning factor to calculate the operating state deviation of the warning factor. When the operating state deviation is higher than a preset deviation threshold, calculate the operating state deviation and the abnormal frequency of the associated factor to determine the target warning threshold of the warning factor based on the abnormal frequency, and generate the warning information of the lithium battery energy storage power station based on the target operating data and the warning threshold.

[0055] In some embodiments, if the associated factor has a warning record within a preset period, specifically, firstly, after determining that the target warning factor and the specified warning factor are associated, the system will further determine whether the number of failures of the associated factor exceeds a preset number threshold, or whether its number is higher than a preset number threshold; if either condition is met, the system will extract the current period and historical period operation data from the record table of the specified warning factor.

[0056] In some embodiments, extracting historical operating data and target operating data of the warning factor within a historical period from the record information of the warning factor to calculate the deviation of the warning factor's operating state includes: extracting historical operating data and target operating data of the warning factor within a historical period from the record information of the warning factor; generating a first operating curve and a second operating curve based on the historical operating data and the target operating data, respectively; and determining the deviation of the warning factor's operating state based on the area difference or statistical deviation value between the first operating curve and the second operating curve. Specifically, firstly, two types of time series data are extracted from the record table corresponding to the specified warning factor: one is the historical operating data of the warning factor within a historical period, wherein the historical period is usually selected as several normal operating periods (e.g., the previous M periods) before the current warning period, and the warning factor has not triggered any warning records within this historical period; the other is the target operating data (i.e., real-time operating data) of the warning factor within the current warning period. Secondly, a first operating curve is generated based on historical operating data, reflecting the historical operating baseline of the warning factor under normal conditions. A second operating curve is generated based on target operating data, reflecting the actual operating status of the warning factor in the current period. The operating curves are generated by continuously plotting time-series data in a time-numerical coordinate system, using linear interpolation or moving average methods to smooth the curves, ensuring continuity and comparability. Then, the system quantifies the deviation of the warning factor's operating status by calculating the difference between the first and second operating curves.

[0057] In some embodiments, determining the operational deviation of the early warning factor includes two methods: the first method is deviation calculation based on area difference: calculating the historical area value enclosed by the first operational curve and the horizontal axis. And the current area value enclosed by the second running curve and the horizontal axis. The normalized area difference is calculated as the deviation using the following formula. The calculation formula is: The first method involves calculating the area value by integrating the curve over the time window; the second method involves calculating the deviation based on statistical deviations: calculating the mean of historical operating data. and standard deviation and the mean of the current running data. The following formula in Z-score form is used. Calculate the statistical deviation as the degree of deviation. This method is suitable for scenarios where the data distribution is relatively stable and conforms to the normality assumption, and can reflect the statistical significance deviation of the current data from the historical baseline.

[0058] After identifying the correlated factors, we further check whether these correlated factors have warning records and calculate the deviation of the current warning factor's operating curve from the historical baseline. This enables precise quantification and verification of abnormal states and effectively identifies those abnormal states that are masked by improper threshold settings.

[0059] In some embodiments, calculating the deviation of the operating state and the abnormal frequency of the associated factors, and determining the target warning threshold of the warning factor based on the abnormal frequency, includes: calculating the abnormal frequency of each associated factor; calculating a comprehensive influence value using a factor influence function based on the deviation of the operating state and each abnormal frequency; and adjusting the initial warning threshold using a sensitivity factor according to the deviation of the operating state and the comprehensive influence value to obtain the target warning threshold of the warning factor. Specifically, the calculated deviation... Compared with a preset deviation threshold, if If the deviation exceeds the preset threshold, the operating status of the warning factor is determined to be significantly abnormal. In this case, the abnormal frequency of each associated factor within the last N preset periods is calculated. The set of correlation factors is: ,in, This represents the dependency strength weight between the target factor i and the current factor. This weight is pre-set based on the attributes of the edges in the knowledge graph or statistical analysis of historical data. This represents the frequency at which target factor i triggers an anomaly over the past N periods. Next, based on the aforementioned operational state deviation... and the frequency of anomalies in each correlation factor A comprehensive impact value is obtained by calculating the factor impact function. The formula for calculating the factor influence function is: Finally, based on the deviation from the operating state... and comprehensive impact value Using a preset sensitivity factor To dynamically adjust the initial warning threshold Thus, the target warning threshold is obtained. The formula is adjusted as follows: It should be noted that the preset deviation threshold can be set based on empirical data from multiple real-world scenarios obtained by those skilled in the art.

[0060] It should be noted that adjustments to the upper limit threshold are made using subtraction to ensure that the warning threshold is lowered when risk increases. The formula is: For adjusting the lower threshold, addition is used based on the direction of deviation, and the formula is as follows: .

[0061] For example, if the warning factor is "battery module temperature," its current value is 59.7℃, which is within the upper limit range [58, 60], and the original upper limit threshold is 60℃. Its associated "cell internal resistance" (i.e., the associated factor) has shown anomalies 3 times in the last 5 cycles, with an anomaly frequency of... Dependence weight Deviation of the current cycle temperature operating curve The sensitivity parameter is set to The upper limit adjustment for the early warning is then: The new warning upper limit threshold is: At this point, the current value of 59.7℃ will trigger the corrected warning, avoiding the false negative problem that exists with traditional static thresholds.

[0062] By dynamically adjusting the warning threshold, the threshold can be adaptively adjusted according to dynamic factors such as equipment aging and changes in operating conditions, thus completely solving the inherent defects of the "static warning threshold".

[0063] In some embodiments, early warning information for lithium battery energy storage power stations is generated based on the target operating data and the early warning threshold. Specifically, the target operating data for the current period is compared with a new threshold. Real-time comparison is performed, and the warning threshold typically includes an upper limit threshold. and lower threshold If real-time data exceeds or below If the warning condition is triggered, the system will immediately generate a structured warning message, which includes the warning subject identifier, abnormal warning factor, real-time data, warning threshold, warning level, correlation factor analysis, and timestamp. At the same time, the generated warning message will be pushed to the mobile terminal or computer client of the operation and maintenance personnel in real time through the preset communication interface (such as API call, message queue), and displayed on the large screen or web interface of the centralized monitoring system (such as EMS) of the energy storage power station in the form of highlight, pop-up window or list item. At the same time, the abnormal component will be marked as abnormal (such as turning red flashing) on ​​the relevant system structure diagram to realize the rapid location of the fault.

[0064] It should be noted that in the above scheme, if the specified early warning factor of the specified early warning dimension is not in the preset early warning proximity interval of the specified early warning factor, or the associated early warning dimension does not have early warning record information within the preset period, or the target early warning factor is not associated with the specified early warning factor, or the number of failures of a single target early warning factor does not exceed a preset number threshold, or the number of target early warning factors is not higher than a preset number threshold, or the deviation is not higher than a preset deviation threshold, then the energy storage power station will be directly given an online early warning based on the early warning record information of the target early warning factor. However, the specified early warning factor will be given special attention so that the operating data of the specified early warning factor can be obtained in a short period of time for timely monitoring.

[0065] This invention, by acquiring initial operational data from multiple levels and determining early warning factors, can capture more comprehensive and nuanced operational status information, avoiding inaccurate judgments due to incomplete data. Through a pre-constructed knowledge graph, it automatically retrieves other related factors with dependencies or subordinate relationships to the early warning factors, elevating early warning judgment from isolated single-point indicators to a system-level network of connections. This enables early detection of potential risks arising from inter-device linkage effects, solving the problem of difficulty in capturing potential abnormal linkage patterns between devices and achieving a preventative effect. After identifying related factors, it further checks whether these factors have early warning records and calculates the deviation of the current early warning factor's operational curve from the historical baseline, achieving precise quantification and verification of abnormal states and effectively identifying abnormal states masked by improper threshold settings. By dynamically adjusting the early warning threshold, the threshold can adaptively adjust with dynamic factors such as equipment aging and changes in operating conditions, completely solving the inherent defects of "static early warning thresholds." Ultimately, the system is able to generate warning information based on the adjusted target warning thresholds that better reflect the current actual operating status, thereby significantly reducing missed and false alarms at the source and significantly improving the accuracy of warnings.

[0066] like Figure 2 As shown, based on the above method embodiments, corresponding apparatus embodiments are provided; One embodiment of the present invention provides an early warning system for a lithium battery energy storage power station, comprising: The acquisition module 100 is used to acquire the initial operating data of each level of equipment in the lithium battery energy storage power station within a preset period, and to determine several early warning factors based on the initial operating data. The determination module 200 is used to determine the associated factors that are related to the warning factors located in the preset warning neighbor interval from the pre-built knowledge graph; The early warning module 300 is used to extract historical operating data and target operating data of the early warning factor within a historical period from the record information of the early warning factor if the associated factor has an early warning record within a preset period, so as to calculate the operating state deviation of the early warning factor. When the operating state deviation is higher than a preset deviation threshold, the module calculates the operating state deviation and the abnormal frequency of the associated factor, so as to determine the target early warning threshold of the early warning factor based on the abnormal frequency, and generates early warning information for the lithium battery energy storage power station based on the target operating data and the early warning threshold.

[0067] It is understood that the above-described device embodiments correspond to the method embodiments of the present invention, and can implement the early warning method for the lithium battery energy storage power station provided by any of the above-described method embodiments of the present invention.

[0068] It should be noted that the device embodiments described above are merely illustrative, and some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Furthermore, in the accompanying drawings of the device embodiments provided by this invention, the connection relationships between modules indicate that they have communication connections, which can specifically be implemented as one or more communication buses or signal lines. Those skilled in the art can understand and implement this without any creative effort.

[0069] Based on the above-described embodiments of the early warning method for lithium battery energy storage power stations, another embodiment of the present invention provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the early warning method for lithium battery energy storage power stations according to any embodiment of the present invention.

[0070] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in the terminal device.

[0071] The terminal device may be a desktop computer, laptop, handheld computer, or cloud server, etc. The terminal device may include, but is not limited to, a processor and a memory.

[0072] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor. The processor is the control center of the terminal device, connecting all parts of the terminal device via various interfaces and lines.

[0073] Based on the above-described method embodiments, another embodiment of the present invention provides a computer-readable storage medium including a stored computer program, wherein, when the computer program is executed, it controls the device where the computer-readable storage medium is located to execute the early warning method for a lithium battery energy storage power station as described in any of the above-described method embodiments of the present invention.

[0074] The modules / units integrated in the device / terminal equipment, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0075] The above description represents the preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principles of the present invention, and these improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A method for early warning of lithium battery energy storage power stations, characterized in that, include: Acquire initial operating data of each level of equipment in a lithium battery energy storage power station within a preset period, and determine several early warning factors based on the initial operating data; From the pre-constructed knowledge graph, identify the associated factors that are related to the warning factors located in the vicinity of the preset warning range; If the associated factor has a warning record within a preset period, then the historical operating data and target operating data of the warning factor within the historical period are extracted from the record information of the warning factor to calculate the operating status deviation of the warning factor. When the operating status deviation is higher than a preset deviation threshold, the operating status deviation and the abnormal frequency of the associated factor are calculated to determine the target warning threshold of the warning factor based on the abnormal frequency, and the warning information of the lithium battery energy storage power station is generated based on the target operating data and the warning threshold.

2. The early warning method for a lithium battery energy storage power station according to claim 1, characterized in that, The determination of several early warning factors based on the initial operating data includes: The initial running data is standardized to obtain the processing result; Read the first warning factor from the processing result; The processing results are calculated using a predefined mathematical model to generate a corresponding second warning factor, wherein the warning factor includes the first warning factor and the second warning factor.

3. The early warning method for a lithium battery energy storage power station according to claim 2, characterized in that, The second warning factor includes a health score, the calculation process of which includes: Based on the first early warning factor, construct the feature vector of the component; The deviation of the feature vector from the baseline of similar devices is calculated to obtain the lateral deviation, and the deviation of the feature vector from its own historical baseline is calculated to obtain the longitudinal deviation. The baseline of similar devices is determined by statistically analyzing the average health feature vector of similar devices of the component within a historical period, and the historical baseline of the component is determined by statistically analyzing the average health feature vector of the component within several historical periods. The overall health score of the component is calculated based on the weighted sum of the lateral deviation and the longitudinal deviation.

4. The early warning method for a lithium battery energy storage power station according to claim 2, characterized in that, The step of determining the correlation factors from the pre-constructed knowledge graph that are associated with the warning factors located in the preset warning proximity interval includes: By traversing the knowledge graph, factor nodes or health score nodes that are in the preset warning proximity range and are connected to the warning factor through edges are retrieved. The knowledge graph uses the warning factor and health score of the lithium battery energy storage power station as nodes and the relationship between nodes as edges. The early warning factors corresponding to the factor nodes and health score nodes are determined as the correlation factors.

5. The early warning method for a lithium battery energy storage power station according to claim 1, characterized in that, The step of extracting historical operating data and target operating data of the warning factor within a historical period from the recorded information of the warning factor, in order to calculate the deviation of the operating status of the warning factor, includes: Extract the historical operation data and target operation data of the warning factor within the historical period from the recorded information of the warning factor; A first operating curve and a second operating curve are generated based on the historical operating data and the target operating data, respectively. The deviation of the operating status of the early warning factor is determined based on the area difference or statistical deviation value between the first operating curve and the second operating curve.

6. The early warning method for a lithium battery energy storage power station according to claim 1, characterized in that, The step of calculating the deviation of the operating state and the abnormal frequency of the correlation factor, and determining the target early warning threshold of the early warning factor based on the abnormal frequency, includes: Calculate the abnormal frequency of each of the aforementioned correlation factors; Based on the deviation of the operating state and the frequency of each abnormality, the comprehensive impact value is calculated through the factor influence function; Based on the deviation of the operating status and the comprehensive impact value, the initial warning threshold is adjusted using a sensitivity factor to obtain the target warning threshold of the warning factor.

7. An early warning system for a lithium battery energy storage power station, characterized in that, include: The acquisition module is used to acquire the initial operating data of each level of equipment in the lithium battery energy storage power station within a preset period, and to determine several early warning factors based on the initial operating data. The determination module is used to identify, from a pre-built knowledge graph, related factors that are associated with warning factors located in the vicinity of a preset warning range; The early warning module is used to extract historical operating data and target operating data of the early warning factor within a historical period from the record information of the early warning factor if the associated factor has an early warning record within a preset period, so as to calculate the operating status deviation of the early warning factor. When the operating status deviation is higher than a preset deviation threshold, the module calculates the operating status deviation and the abnormal frequency of the associated factor, so as to determine the target early warning threshold of the early warning factor based on the abnormal frequency, and generates early warning information for the lithium battery energy storage power station based on the target operating data and the early warning threshold.

8. A terminal device, characterized in that, include: One or more processors; A memory, coupled to the processor, for storing one or more programs; When the one or more programs are executed by the one or more processors, the one or more processors implement the steps of the early warning method for a lithium battery energy storage power station as described in any one of claims 1-7.

9. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform the steps of the early warning method for a lithium battery energy storage power station as described in any one of claims 1-6.

10. A computer program product, comprising a computer program or instructions, characterized in that, When the computer program or instructions are executed by the communication device, they implement the steps of the early warning method for a lithium battery energy storage power station as described in any one of claims 1 to 6.

Images