Temperature early warning method, system, equipment and medium for alternating current side equipment of energy storage system

Abstract

The invention discloses an energy storage system alternating current side equipment temperature early warning method, system, equipment and medium, and belongs to the technical field of electrochemical energy storage system safety monitoring. Equipment temperature, environment temperature, load power and historical operation multi-dimensional data of energy storage system alternating current side equipment are obtained; a load measuring point with the highest correlation with a temperature measuring point to be analyzed is screened out, topological relation and time sequence original data are obtained through a relational database and a time sequence database, the original data are preprocessed to obtain a low-frequency characteristic data set, equipment temperature rise is calculated, step length is set according to rated power, power intervals are divided, and multi-dimensional data are distinguished; and in each power interval, calculating a temperature rise group mean value and a standard deviation of equipment of the same type, evaluating a deviation degree of single equipment, converting into a standardized health score according to a preset rule, dividing temperature risk grades, realizing graded early warning, and realizing early recognition and graded warning on the over-temperature risk of the energy storage alternating current side equipment.

Description

Technical Field

[0001] This invention belongs to the field of safety monitoring technology for electrochemical energy storage systems, specifically relating to a method, system, equipment, and medium for early warning of temperature of AC side equipment in energy storage systems. Background Technology

[0002] With electrochemical energy storage systems serving as a key technology supporting the safe and stable operation of new power systems, they are being deployed and widely applied in various scenarios, including the grid side (such as peak shaving and frequency regulation, emergency backup), the power source side (such as energy storage for new energy sources, smoothing output fluctuations), and the user side (such as industrial and commercial energy storage, demand response). A typical electrochemical energy storage system usually consists of several core subsystems, including battery clusters (composed of multiple battery modules connected in series and parallel), a battery management system (BMS), a power conversion system (PCS), an AC-side prefabricated substation (referred to as "prefabricated substation"), and a comprehensive monitoring and energy management system. Among them, the PCS, as the core power electronic device connecting the DC-side battery system and the AC grid, undertakes the key function of bidirectional power conversion (AC / DC and DC / AC); while the prefabricated substation is mainly used to step up the low-voltage AC power output from the PCS to a voltage level that meets grid connection requirements (such as 10kV or 35kV), thereby achieving efficient power transmission and grid connection.

[0003] In actual operation, PCS and transformer substations, as key equipment with high power density and high current carrying capacity, are prone to localized overheating under complex conditions such as high load operation, high ambient temperature, poor ventilation and heat dissipation, or equipment aging. This abnormal temperature rise not only accelerates the aging of insulation materials and reduces the insulation performance of equipment, but may also cause loose electrical connections, component failure, and in severe cases, even equipment burnout, short circuits, and fires, posing a significant threat to the safe and stable operation of energy storage systems. Currently, the industry mainly relies on point temperature sensors (such as PT100 platinum resistance thermometers, K-type thermocouples, etc.) installed on the surface of the equipment casing, inside the windings, or near key heat-generating components for temperature monitoring of PCS and transformer substations. Their alarm logic usually adopts a static threshold mechanism, that is, when the measured temperature exceeds the preset upper limit (such as 85℃ or 105℃), an alarm or protection action is triggered.

[0004] However, this traditional monitoring method has certain limitations: 1) Response lag: Relying solely on measured temperature, it cannot provide early warnings when the temperature has not yet exceeded the limit but the trend is abnormal; 2) Fixed threshold: The alarm threshold is statically set and does not take into account the impact of factors such as ambient temperature, load conditions, and data from similar equipment on temperature rise; 3) High false alarm / missed alarm rate: False alarms are likely to occur when the ambient temperature changes suddenly or the load changes abruptly, while missed alarms may occur during the slow temperature rise process.

[0005] The aforementioned problems severely restrict the ability of energy storage systems to proactively prevent and control thermal risks of key AC-side equipment. Therefore, it is urgent to develop a temperature anomaly early warning method that integrates multi-source operating data and has the ability to identify early anomalies. Summary of the Invention

[0006] The purpose of this invention is to overcome the problem of poor proactive prevention and control capabilities of energy storage systems for thermal risks of key AC-side equipment, and to propose a method, system, equipment, and medium for temperature early warning of AC-side equipment in energy storage systems.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for early warning of temperature of AC side equipment in an energy storage system, comprising the following steps: Acquire multi-dimensional operational data of AC-side equipment in the energy storage system. This multi-dimensional operational data includes equipment temperature measurement point data, ambient temperature data, load power data, and historical operational data. Based on statistical correlation analysis, the load measurement points with the highest correlation to the temperature measurement points to be analyzed are selected from multi-dimensional operational data. The topological relationships, load measurement point list and temperature measurement point list to be analyzed are obtained through relational database. The time series raw data of load measurement points and temperature measurement points to be analyzed are extracted by combining time series database. The raw time series data is preprocessed to obtain a low-frequency feature dataset; based on the low-frequency feature dataset, the temperature rise of the equipment is obtained by calculating the difference between the temperature measurement point to be analyzed and the ambient temperature data. The power step size is set according to the rated power of the AC side equipment of the energy storage system. Based on the load measurement point, the power range of similar loads or operating conditions is divided according to the active power of the AC side equipment of the energy storage system, and the multi-dimensional operating data is divided into each power range. For AC-side equipment of the same type of energy storage system within each power range, calculate the group mean and standard deviation of temperature rise based on equipment temperature rise, and assess the degree of deviation of the temperature rise of individual equipment relative to the group mean; The degree of deviation is converted into a standardized equipment health score through a preset nonlinear mapping rule, and the equipment temperature risk level is divided according to the standardized health score to complete the graded early warning.

[0008] Furthermore, statistical correlation analysis uses the Pearson correlation coefficient to calculate pairwise correlations between various types of data in historical operating data. Through visualization results, the load power data with the highest correlation to the temperature measurement point to be analyzed is selected as the load measurement point.

[0009] Furthermore, preprocessing includes cleaning, missing value completion, and temporal granular aggregation; The cleaning process distinguishes between charging and discharging states based on the positive and negative values ​​of active power, and identifies and removes dead value data that remain constant for a preset duration under non-zero power. Missing values ​​were filled using the average of a sliding window within a similar time period at the same measurement point using the same equipment. Time-granularity aggregation is performed by averaging according to a preset time window.

[0010] Furthermore, the power range includes a static range and several load power ranges, with the operating condition of equipment with zero active power being classified as the static range.

[0011] Furthermore, the calculation of the temperature rise group mean and standard deviation based on equipment temperature rise includes calculating the mean and standard deviation of equipment temperature rise for the same type of equipment within each power range, converting the equipment temperature rise data of a single equipment into a standard deviation multiple; when aggregating by equipment, if there are records with a standard deviation multiple > 3σ within the current power range, only the records with a standard deviation multiple > 3σ are averaged, otherwise all records are averaged to obtain the equipment-level temperature rise anomaly deviation.

[0012] Furthermore, the preset nonlinear mapping rules include: When the abnormal deviation of the equipment-level temperature rise is ≤3σ, linear mapping is performed according to the rule of preset upper limit of fraction - (deviation value / 3)×10; When the abnormal deviation of equipment temperature rise is greater than 3σ, the calculation is performed linearly decreasing according to the gradient of 3σ corresponding to the first score, 6σ corresponding to the second score, and 9σ corresponding to the third score, and finally the standardized equipment health score between the preset lower limit and the preset upper limit is obtained.

[0013] Furthermore, the criteria for classifying temperature risk levels include: When the standardized equipment health score is greater than or equal to the first preset score, the temperature risk level is the fault-free risk level. When the second preset score ≤ standardized equipment health score < the first preset score, the temperature risk level is a slight abnormality or a low probability risk level. When the third preset score is less than or equal to the standardized equipment health score and less than the second preset score, the temperature risk level is classified as a medium risk level. When the standardized equipment health score is less than the third preset score, the temperature risk level is classified as a severe overheating risk level. The system automatically triggers early warning information and maintenance suggestions based on temperature risk levels.

[0014] In a second aspect, the present invention provides a temperature warning system for AC side equipment of an energy storage system, comprising: The multi-dimensional data acquisition module is used to acquire multi-dimensional operating data of the AC side equipment of the energy storage system. The multi-dimensional operating data includes equipment temperature measurement point data, ambient temperature data, load power data, and historical operating data. The raw data extraction module is used to filter out the load measurement points with the highest correlation to the temperature measurement points to be analyzed from multi-dimensional operational data based on statistical correlation analysis, and obtain the topological relationship, load measurement point list and temperature measurement point list to be analyzed through a relational database, and extract the time series raw data of the load measurement points and temperature measurement points to be analyzed by combining the time series database. The equipment temperature rise calculation module is used to preprocess the raw time series data to obtain a low-frequency feature dataset; based on the low-frequency feature dataset, the equipment temperature rise is obtained by calculating the difference between the temperature measurement point to be analyzed and the ambient temperature data. The power range division module is used to set the power step size according to the rated power of the AC side equipment of the energy storage system. Based on the load measurement points, it divides the power range of similar loads or operating conditions according to the active power of the AC side equipment of the energy storage system, and distinguishes the multi-dimensional operating data into each power range. The temperature rise deviation assessment module is used to calculate the group mean and standard deviation of the temperature rise of AC side equipment of the same type of energy storage system in each power range, and to assess the degree of deviation of the temperature rise of a single device from the group mean. The temperature classification and early warning module is used to convert the degree of deviation into a standardized equipment health score through a preset nonlinear mapping rule, and classify the equipment temperature risk level according to the standardized health score to complete the classification and early warning.

[0015] Thirdly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a method for early warning of temperature of AC side equipment in an energy storage system.

[0016] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements a method for early warning of temperature of AC side equipment in an energy storage system.

[0017] Compared with the prior art, the present invention has the following beneficial technical effects: This invention proposes a temperature early warning method for AC-side equipment in energy storage systems. Through a comprehensive design that integrates multi-dimensional data on temperature, environment, and load; optimizes core indicators by replacing the original temperature with temperature rise data; constructs dynamic benchmarks based on power range group statistics; and combines deviation levels with health scores to predict abnormal trends, this method fundamentally solves the three major pain points of traditional monitoring methods: delayed response, fixed thresholds, and false alarms / missed alarms. It achieves more accurate, earlier, and adaptable temperature early warnings for complex operating conditions, addressing the problem of poor proactive prevention and control capabilities for thermal risks in critical AC-side equipment of energy storage systems. The method integrates ambient temperature, load power, and historical operating data for AC-side equipment temperature early warning. Utilizing correlation analysis, it can accurately screen load measurement points strongly correlated with the temperature measurement points to be analyzed, reducing interference from invalid data. It also meticulously divides power ranges for more targeted analysis. By evaluating deviations and then using nonlinear mapping to derive standardized health scores, it classifies risk levels and achieves tiered early warning. This allows for early detection of potential hazards, effectively ensuring the safe and stable operation of energy storage systems. It enables early identification and tiered warning of over-temperature risks in critical AC-side equipment of energy storage systems, improving the operational safety and reliability of energy storage systems. Attached Figure Description

[0018] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of the invention in any way. Furthermore, the shapes and proportions of the components in the drawings are merely illustrative to aid in understanding the invention and do not specifically limit the shapes and proportions of the components of the invention.

[0019] In the attached diagram: Figure 1 This is a flowchart of the temperature early warning method for AC side equipment in the energy storage system of the present invention.

[0020] Figure 2 This is a simplified structural diagram of the AC side equipment temperature early warning system of the energy storage system of the present invention.

[0021] Figure 3 This is a schematic diagram of the electronic device used in the temperature early warning method for the AC side equipment of the energy storage system according to the present invention.

[0022] Figure 4 This is a schematic diagram of the overall process of the temperature early warning method for AC side equipment of the energy storage system in an embodiment of the present invention.

[0023] Figure 5 This is a flowchart of a method for early warning of temperature of AC side equipment in an energy storage system according to an embodiment of the present invention.

[0024] Figure 6 This is a block diagram of the AC side equipment temperature early warning system of the energy storage system in an embodiment of the present invention. Detailed Implementation

[0025] 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.

[0026] Example 1 See Figure 1 A method for early warning of temperature on the AC side of an energy storage system includes the following steps: Acquire multi-dimensional operational data from the AC-side equipment of the energy storage system. This equipment includes the power converter system (PCS), box-type transformer, high-voltage fuses, contactors, etc. The multi-dimensional operational data includes equipment temperature measurement point data, ambient temperature data, load power data, and historical operational data. Equipment temperature measurement point data covers the temperatures of core heat-generating components of each key AC-side device, including PCS power module temperature, transformer winding temperature, transformer core temperature, fuse contact temperature, and contactor contact temperature. Ambient temperature data includes the ambient temperature inside the energy storage compartment, the inlet temperature of the equipment's cooling duct, and the ambient temperature of the outdoor energy storage station. Load power data covers electrical parameters directly related to the operating load of the AC-side equipment, including PCS... Output active power, transformer input / output active power, bus voltage, etc.; historical operating data is the historical accumulation of the above three types of data, and the data storage format is consistent with the real-time acquired data, and includes the equipment's cumulative running time and related fields of historical maintenance records; based on statistical correlation analysis, the load measurement points with the highest correlation to the temperature measurement points to be analyzed are selected from the multi-dimensional operating data. The statistical correlation analysis uses the Pearson correlation coefficient to calculate the pairwise correlation of various types of data in the historical operating data. Through visualization results, the load power data with the highest correlation to the temperature measurement points to be analyzed are selected as load measurement points. In data selection, the Pearson correlation coefficient is used to accurately identify strongly correlated load power data, improving the targeting of the analysis. Visualization is used for each type The correlation results between load power data and temperature measurement point data are used to create scatter plots and correlation coefficient heatmaps. The scatter plots visually illustrate the correlation trend, while the heatmaps compare the correlation coefficients between all load power data and temperature measurement points. Furthermore, a relational database is used to retrieve topological relationships, a list of load measurement points, and a list of temperature measurement points to be analyzed. The relational database stores basic equipment information (equipment ID, model, rated parameters, installation location), topological relationships (physical connections between load and temperature measurement points, equipment belonging to a compartment / branch), a list of measurement points (IDs and names of temperature and load measurement points), maintenance records, and other structured data. The load measurement point list and the list of temperature measurement points to be analyzed are retrieved from the relational database. The list of temperature measurement points and their topological relationships are established to ensure that the extracted load measurement points and temperature measurement points belong to the same device or directly related branches. The raw time-series data of the load measurement points and the temperature measurement points to be analyzed are extracted using a time-series database. The time-series database stores real-time and historical operating data of the aforementioned temperature measurement point data, ambient temperature data, and load power data. A time range for data extraction is set, and the following data are extracted from the time-series database according to a unified timestamp: temperature data of the temperature measurement points to be analyzed, power data of the corresponding strongly correlated load measurement points, and ambient temperature data from the same period. During extraction, the timestamps of the three types of data are ensured to be aligned. The raw time-series data is cleaned, missing values ​​are filled in, and time-granularity aggregation is preprocessed to obtain a low-frequency feature dataset.Data cleaning distinguishes between charging and discharging states based on the positive and negative values ​​of active power. It identifies and removes dead value data where the value remains constant for a preset duration under non-zero power conditions. The operating state is determined based on the positive and negative values ​​of active power at the load measurement point: active power > 0 indicates discharging; active power < 0 indicates charging; and active power = 0 indicates a static state. Dead value data is marked and removed, and the resulting data gaps are subsequently filled using missing value imputation. Missing value imputation uses the average value of a sliding window within a similar time period at the same measurement point and equipment. Time-granular aggregation is performed by averaging within a preset time window. Data cleaning distinguishes between charging and discharging states and removes dead values, ensuring data validity. Missing value imputation and granular aggregation make the data more complete and standardized. The power range is reasonably divided, covering both static and different load conditions, making the analysis more comprehensive. Based on the low-frequency feature dataset, the temperature rise of the equipment is obtained by calculating the difference between the temperature measurement point to be analyzed and the ambient temperature data. The power step size is set according to the rated power of the AC side equipment of the energy storage system. Based on the load measurement points, the power ranges for similar loads or operating conditions are divided according to the active power of the AC side equipment of the energy storage system. The multi-dimensional operating data is distinguished into each power range, including the static range and several load power ranges. The operating condition of equipment with zero active power is divided into the static range. Each record in the low-frequency feature dataset is matched to the corresponding power range according to its average load value and operating status. For the same type of energy storage system in each power range, the power range is further divided into the static range. For flow-side equipment, the system calculates the group mean and standard deviation of temperature rise based on equipment temperature rise, assessing the degree of deviation of individual equipment temperature rise relative to the group mean. This includes calculating the mean and standard deviation of temperature rise for similar equipment within each power range, converting individual equipment temperature rise data into standard deviation multiples, accurately assessing abnormal temperature rise deviations and highlighting anomalous data. When aggregating by equipment, if a equipment has records with a standard deviation multiple > 3σ within the current power range, only these records are averaged; otherwise, all records are averaged to obtain equipment-level abnormal temperature rise deviations. The degree of deviation is converted into a standardized equipment health score through a preset nonlinear mapping rule. This preset nonlinear mapping rule reasonably divides the health score, clearly reflecting the equipment status. The rules for determining the temperature rise include: when the abnormal deviation of the equipment-level temperature rise is ≤3σ, a linear mapping is performed according to the rule of the upper limit of the preset score - (deviation value / 3)×10; when the abnormal deviation of the equipment-level temperature rise is >3σ, a linear decreasing calculation is performed according to the gradient of 3σ corresponding to the first score, 6σ corresponding to the second score, and 9σ corresponding to the third score, finally obtaining a standardized equipment health score between the lower limit and the upper limit of the preset score, and classifying the equipment temperature risk level according to the standardized health score. The temperature risk level classification criteria include: when the standardized equipment health score is ≥ the first preset score, the temperature risk level is the no-fault risk level; when the second preset score is ≤ the standardized equipment health score < the first preset score, the temperature risk level is the slight abnormality or low probability risk level.When the third preset score is less than or equal to the standardized equipment health score but less than the second preset score, the temperature risk level is classified as moderate; when the standardized equipment health score is less than the third preset score, the temperature risk level is classified as severe overheating risk. The system automatically triggers early warning information and maintenance suggestions for the temperature risk level, completing the graded early warning. The temperature risk level classification standard is clear, enabling rapid judgment of the risk degree. Furthermore, the system automatically pushes early warnings and maintenance suggestions, allowing maintenance personnel to promptly grasp the situation and take proactive measures, effectively preventing malfunctions in the AC side equipment of the energy storage system caused by abnormal temperatures, ensuring stable system operation, and reducing maintenance costs.

[0027] This embodiment acquires multi-dimensional operational data and filters strongly correlated load measurement points to accurately locate key factors affecting equipment temperature, providing a reliable basis for subsequent analysis. Preprocessing the raw time-series data and calculating equipment temperature rise effectively eliminates data noise, improves data quality, and makes the analysis results more accurate. Dividing power intervals by power step size and differentiating multi-dimensional operational data enables detailed analysis for different operating conditions, enhancing the pertinence of early warnings. Assessing the degree of temperature rise deviation of individual devices allows for timely detection of abnormal devices. Converting the degree of deviation into a standardized health score and classifying risk levels enables tiered early warning, helping maintenance personnel quickly assess equipment status, take proactive measures, ensure the safe and stable operation of the energy storage system, and reduce the risk of failure.

[0028] Example 2 See Figure 2 A temperature warning system for AC side equipment of an energy storage system, characterized in that it includes: The multi-dimensional data acquisition module is used to acquire multi-dimensional operating data of the AC side equipment of the energy storage system. The multi-dimensional operating data includes equipment temperature measurement point data, ambient temperature data, load power data, and historical operating data. The raw data extraction module is used to filter out the load measurement points with the highest correlation to the temperature measurement points to be analyzed from multi-dimensional operational data based on statistical correlation analysis, and obtain the topological relationship, load measurement point list and temperature measurement point list to be analyzed through a relational database, and extract the time series raw data of the load measurement points and temperature measurement points to be analyzed by combining the time series database. The equipment temperature rise calculation module is used to preprocess the raw time series data to obtain a low-frequency feature dataset; based on the low-frequency feature dataset, the equipment temperature rise is obtained by calculating the difference between the temperature measurement point to be analyzed and the ambient temperature data. The power range division module is used to set the power step size according to the rated power of the AC side equipment of the energy storage system. Based on the load measurement points, it divides the power range of similar loads or operating conditions according to the active power of the AC side equipment of the energy storage system, and distinguishes the multi-dimensional operating data into each power range. The temperature rise deviation assessment module is used to calculate the group mean and standard deviation of the temperature rise of AC side equipment of the same type of energy storage system in each power range, and to assess the degree of deviation of the temperature rise of a single device from the group mean. The temperature classification and early warning module is used to convert the degree of deviation into a standardized equipment health score through a preset nonlinear mapping rule, and classify the equipment temperature risk level according to the standardized health score to complete the classification and early warning.

[0029] Example 3 See Figure 3 An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements a method for temperature early warning of AC-side equipment in an energy storage system: acquiring multi-dimensional operational data of the AC-side equipment of the energy storage system, including equipment temperature measurement point data, ambient temperature data, load power data, and historical operational data; based on statistical correlation analysis, selecting the load measurement points with the highest correlation to the temperature measurement points to be analyzed from the multi-dimensional operational data, and obtaining the topological relationships, a list of load measurement points, and a list of temperature measurement points to be analyzed from a relational database; and extracting the time-series raw data of the load measurement points and the temperature measurement points to be analyzed from a time-series database; and pre-processing the time-series raw data. The data is processed to obtain a low-frequency feature dataset. Based on this dataset, the temperature rise of the equipment is calculated by comparing the temperature measurement points with ambient temperature data. A power step size is set according to the rated power of the AC-side equipment in the energy storage system. Based on the load measurement points and the active power of the AC-side equipment, power intervals for similar loads or operating conditions are divided, thus differentiating the multi-dimensional operating data into these intervals. For the same type of AC-side equipment in each power interval, the group mean and standard deviation of the temperature rise are calculated based on the equipment temperature rise, and the degree of deviation of the temperature rise of a single equipment relative to the group mean is assessed. The degree of deviation is converted into a standardized equipment health score using a preset nonlinear mapping rule, and the equipment temperature risk level is classified according to the standardized health score to complete the graded early warning.

[0030] Example 4 A computer-readable storage medium storing a computer program, which, when executed by a processor, implements a method for temperature early warning of AC-side equipment in an energy storage system: The method involves acquiring multi-dimensional operational data of the AC-side equipment in the energy storage system, including equipment temperature measurement point data, ambient temperature data, load power data, and historical operational data; based on statistical correlation analysis, selecting the load measurement points with the highest correlation to the temperature measurement points to be analyzed from the multi-dimensional operational data, and obtaining the topological relationships, a list of load measurement points, and a list of temperature measurement points to be analyzed from a relational database; and extracting the time-series raw data of the load measurement points and the temperature measurement points to be analyzed from a time-series database; and preprocessing the time-series raw data to obtain low... The system employs a low-frequency feature dataset. Based on this dataset, the temperature rise of the equipment is calculated by comparing the temperature measurement points with ambient temperature data. A power step size is set according to the rated power of the AC-side equipment in the energy storage system. Based on load measurement points and the active power of the AC-side equipment, power intervals for similar loads or operating conditions are divided, thus differentiating multi-dimensional operating data into these intervals. For AC-side equipment of the same type within each power interval, the group mean and standard deviation of the temperature rise are calculated based on the equipment temperature rise, assessing the degree of deviation of the temperature rise of a single device relative to the group mean. This deviation is then converted into a standardized equipment health score using a preset nonlinear mapping rule. Based on the standardized health score, the equipment temperature risk level is classified, and a graded early warning system is established.

[0031] Example 5 See Figure 4 This paper proposes a temperature early warning method for AC-side equipment in energy storage systems. Based on multi-source data fusion, an algorithm for early warning of AC-side equipment in energy storage systems is developed. By deeply integrating multi-dimensional information such as ambient temperature, load power, historical operating data, and temperature measurement point data, a high-precision and robust temperature early warning algorithm model is constructed. This enables early identification, dynamic assessment, and tiered alarm of abnormal temperature states in key AC-side equipment of energy storage systems (such as box-type transformers and PCS equipment). This not only significantly improves the safety and reliability of energy storage systems under complex operating conditions but also provides maintenance personnel with scientific and quantitative decision-making basis, effectively preventing failures and even safety accidents caused by equipment overheating. See also... Figure 5 Includes the following steps: S1. Based on the statistical correlation analysis between variables, identify load measurement points (such as the active power measurement points of equipment) that are strongly correlated with the temperature measurement points to be analyzed, and determine the model input; S2. Obtain the topological relationships between devices within the site and the list of measurement points corresponding to each device through a relational database (such as PostgreSQL), and further extract the second-level raw data of all relevant measurement points of the transformer substations under the site through a time series database (such as TaosDB) to ensure the integrity and timeliness of the data source; S3. Perform rigorous data cleaning on the original data. Then, use the mean imputation strategy to fill in the missing values ​​generated after cleaning, and aggregate the second-level data into a 5-minute granularity to improve the efficiency of model training and inference. S4. Based on this, key features are constructed. For equipment with temperature measuring points, the temperature rise column corresponding to the temperature measuring point to be analyzed needs to be obtained according to the temperature measuring point to be analyzed and the ambient temperature measuring point. S5. Based on the rated power parameters of the equipment, scientifically set a reasonable power step size (e.g., 10% of the rated power) to perform binning processing on the active power of all box-type substation equipment in the entire station. Divide the equipment operation data into several power ranges with similar load levels or operating conditions. Data with zero power is uniformly classified into the "static" range to distinguish whether the energy storage system is in standby or shutdown state. S6. Within each power range, perform statistical analysis on the temperature rise data of similar equipment (such as transformers of the same model), calculate the mean and standard deviation of the temperature rise, and based on the classic 3σ (three sigma) statistical rule, evaluate the deviation of the temperature rise performance of each equipment in this range, and identify abnormal individuals that deviate significantly from the group mean. S7. Aggregate the results by device to obtain the average deviation of temperature rise for each device in the corresponding power range; S8. Further, the comprehensive deviation value is converted into a standardized equipment health score of 0–100 points through a preset nonlinear mapping rule. This score intuitively reflects the risk level of the current temperature of the equipment. S9. Finally, based on the score range, the equipment is divided into different risk levels to achieve closed-loop management from quantitative assessment to visual early warning.

[0032] The specific implementation details of each of the above steps are as follows: Step 1: Statistical correlation analysis uses methods such as Pearson correlation coefficient, Spearman rank correlation coefficient, and Kendall rank coefficient. In this embodiment, Pearson correlation coefficient is used to calculate the pairwise correlation of each variable in the historical operating data. The results are visualized through a heat map, and the load variables (such as the active power measurement point of the equipment) with the highest correlation to the target temperature measurement point (such as the temperature of the A phase winding of the transformer) are selected as the key input features of the subsequent model.

[0033] Step 2: The system queries the station equipment hierarchy structure (e.g., station → transformer substation → PCS → PCS branch / module) through a PostgreSQL relational database and obtains a list of measurement point IDs associated with each device. Simultaneously, it efficiently reads the second-level time series data of the corresponding measurement points using time-series databases such as TaosDB. The measurement points related to the transformer substation include: A / B / C three-phase winding temperature, core temperature, corresponding active power, and ambient temperature; the measurement points related to the PCS cover the three-phase IGBT temperature, active power, cabinet temperature, and ambient temperature of each module or branch.

[0034] Step 3: In the data cleaning stage, the charging and discharging states are first distinguished according to the positive and negative active power, and "dead value" data under non-zero power (i.e., the value is constant for 300 consecutive seconds) is identified, marked as abnormal and removed to avoid misjudgment caused by sensor failure or communication interruption.

[0035] Step 4: For missing values ​​resulting from outlier removal, fill them with the average of a sliding window from the same device and measurement point within a similar time period; at the same time, average and aggregate the original second-level data in a 5-minute window to generate a low-frequency feature dataset for subsequent modeling, balancing data accuracy and model performance.

[0036] Step 5: Construct key temperature rise characteristics based on the temperature to be analyzed and the ambient temperature measurement points, thereby eliminating the possible influence of the ambient temperature on the AC side equipment temperature, avoiding the influence of external factors, and improving the accuracy of temperature early warning analysis; Step 6: Based on the rated power of the equipment, select an appropriate power size as the step size, divide the active power of all the box-type substation equipment in the whole station into boxes, and distinguish the time series data of the equipment under similar load or operating conditions (among which the data with 0 power is divided into the same interval as the static condition of energy storage). Step 7: Within each power range, calculate the mean and standard deviation (std) of the temperature rise for the same type of equipment; for each equipment record, calculate the deviation of its temperature rise from the mean, and convert it to a multiple of the standard deviation (i.e., temperature rise). (mean) / std); then aggregate by device. If a device has records with a deviation >3σ in a certain interval, then only the records with high deviation are averaged; otherwise, all records are averaged to obtain a more representative equipment-level temperature rise anomaly index.

[0037] Step 8: According to the custom mapping rules, when the average deviation is within 3 sigma, a linear mapping of 100-(std_value / 3)×10 is performed according to 0 standard deviation = 100 points and 3 standard deviation = 90 points; if it exceeds 3 sigma, a linear decreasing calculation is performed according to 3 standard deviation = 90 points, 6 standard deviation = 60 points and 9 standard deviation = 30 points. Finally, the average deviation of the device is mapped to the final score of the device from 0 to 100, which facilitates the overall comprehensive evaluation of the system and the push of early warning information.

[0038] Step 9: Finally, the equipment score is mapped to different risk levels. ≥90 points indicates no fault (risk level 0); 85-90 points indicates minor abnormality or low probability risk (risk level 1); 80-85 points indicates moderate risk (risk level 2); and <80 points indicates severe overheating risk (risk level 3). Based on this, the system automatically triggers the corresponding level of early warning information push and maintenance suggestions to achieve intelligent and hierarchical equipment health management.

[0039] This embodiment can sense the evolution trend of equipment thermal status in real time, and achieve accurate and proactive risk identification and alarm in the early stage of temperature anomalies. This significantly improves the thermal safety management level of key AC-side equipment of electrochemical energy storage systems, especially PCS and transformer substations, and provides solid technical support for the safe, reliable and efficient operation of large-scale energy storage systems.

[0040] See Figure 6 Temperature warning system for AC side equipment of energy storage system, including: First, the data acquisition module is used to acquire the relationships between energy storage system equipment, relevant parameter measurement points, and time-series data of the corresponding measurement points, supporting the analysis and execution of subsequent algorithm models; Second, the data processing module includes data cleaning and data aggregation. It removes dead / missing data and refills it according to anomaly identification rules, and aggregates the acquired second-level data to ensure the accuracy of analysis while optimizing model performance. Third, the key feature construction module; the power load data is divided into bins to distinguish different operating conditions of the energy storage system (such as charging, discharging, and idling). The specific binning conditions are related to the rated power of the corresponding equipment.

[0041] Fourth, the temperature anomaly detection module calculates the average temperature and standard deviation of all similar devices in each power range by grouping them into power ranges. Then, based on the 3σ anomaly detection method, it aggregates the average deviation of all records for each device to achieve temperature anomaly detection for each device. Fifth, the result mapping and conversion module converts the average standard value deviation of specific devices into corresponding scores and warning levels through mapping logic, which facilitates the overall comprehensive evaluation of the system and the push of warning information.

[0042] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0043] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0044] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0045] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the functions specified in one or more boxes. Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the claims of the present invention.

Claims

1. A method for early warning of temperature of AC side equipment in an energy storage system, characterized in that, Includes the following steps: Acquire multi-dimensional operational data of AC-side equipment in the energy storage system. The multi-dimensional operational data includes equipment temperature measurement point data, ambient temperature data, load power data, and historical operational data. Based on statistical correlation analysis, load measurement points with the highest correlation to the temperature measurement points to be analyzed are selected from the multi-dimensional operating data. The topological relationship, the list of load measurement points and the list of temperature measurement points to be analyzed are obtained through a relational database. The time series raw data of the load measurement points and the temperature measurement points to be analyzed are extracted by combining the time series database. The original time series data is preprocessed to obtain a low-frequency feature dataset; based on the low-frequency feature dataset, the equipment temperature rise is obtained by calculating the difference between the temperature measurement point to be analyzed and the ambient temperature data. The power step size is set according to the rated power of the AC side equipment of the energy storage system. Based on the load measurement point, the power range of similar loads or operating conditions is divided according to the active power of the AC side equipment of the energy storage system, and the multi-dimensional operating data is divided into each power range. For AC-side devices of the same type of energy storage system within each power range, calculate the group mean and standard deviation of temperature rise based on the temperature rise of the devices, and evaluate the degree of deviation of the temperature rise of a single device relative to the group mean. The degree of deviation is converted into a standardized equipment health score through a preset nonlinear mapping rule, and the equipment temperature risk level is divided according to the standardized health score to complete the graded early warning.

2. The method for early warning of AC side equipment temperature in an energy storage system according to claim 1, characterized in that, The statistical correlation analysis uses the Pearson correlation coefficient to calculate the pairwise correlation of various types of data in the historical operating data. The load power data with the highest correlation to the temperature measurement point to be analyzed is selected as the load measurement point through visualization results.

3. The method for early warning of AC side equipment temperature in an energy storage system according to claim 1, characterized in that, The preprocessing includes cleaning, missing value completion, and time-granular aggregation; The cleaning process distinguishes between charging and discharging states based on the positive and negative values ​​of active power, and identifies and eliminates dead value data that remain constant for a continuous preset time under non-zero power. The missing value is filled by the average value of a sliding window within a similar time period at the same measurement point on the same device. The time-granularity aggregation is performed by averaging according to a preset time window.

4. The method for early warning of AC side equipment temperature in an energy storage system according to claim 1, characterized in that, The power range includes a static range and several load power ranges. The operating condition of equipment with zero active power is classified into the static range.

5. The method for early warning of AC side equipment temperature in an energy storage system according to claim 1, characterized in that, The calculation of the temperature rise group mean and standard deviation based on the equipment temperature rise includes calculating the mean and standard deviation of the equipment temperature rise of the same type of equipment within each power range, and converting the equipment temperature rise data of a single equipment into a standard deviation multiple; When merging equipment, if there are records with a standard deviation multiple > 3σ within the current power range, only the records with a standard deviation multiple > 3σ are averaged; otherwise, all records are averaged to obtain the equipment-level temperature rise anomaly deviation.

6. The method for early warning of AC side equipment temperature in an energy storage system according to claim 5, characterized in that, The preset nonlinear mapping rules include: When the abnormal temperature rise deviation of the equipment level is ≤3σ, linear mapping is performed according to the rule of preset upper limit of score - (deviation value / 3)×10; When the abnormal temperature rise deviation of the equipment is greater than 3σ, the calculation is performed linearly decreasing according to the gradient of 3σ corresponding to the first score, 6σ corresponding to the second score, and 9σ corresponding to the third score, and finally a standardized equipment health score between the preset lower limit and the preset upper limit is obtained.

7. The method for early warning of AC side equipment temperature in an energy storage system according to claim 1, characterized in that, The criteria for classifying temperature risk levels include: When the standardized equipment health score is greater than or equal to the first preset score, the temperature risk level is a fault-free risk level. When the second preset score ≤ the standardized equipment health score < the first preset score, the temperature risk level is a slight abnormality or a low probability risk level; When the third preset score is less than or equal to the standardized equipment health score and less than the second preset score, the temperature risk level is classified as a medium risk level. When the standardized equipment health score is less than the third preset score, the temperature risk level is classified as a severe overheating risk level. The system automatically triggers early warning information and maintenance suggestions based on the temperature risk level.

8. A temperature early warning system for AC side equipment of an energy storage system, characterized in that, include: The multi-dimensional data acquisition module is used to acquire multi-dimensional operating data of the AC side equipment of the energy storage system. The multi-dimensional operating data includes equipment temperature measurement point data, ambient temperature data, load power data, and historical operating data. The raw data extraction module is used to filter out the load measurement points with the highest correlation to the temperature measurement points to be analyzed from the multi-dimensional operational data based on statistical correlation analysis, and to obtain the topological relationship, the list of load measurement points and the list of temperature measurement points to be analyzed through a relational database, and to extract the time series raw data of the load measurement points and the temperature measurement points to be analyzed by combining the time series database. The equipment temperature rise calculation module is used to preprocess the raw time series data to obtain a low-frequency feature dataset; based on the low-frequency feature dataset, the equipment temperature rise is obtained by calculating the difference between the temperature measurement point to be analyzed and the ambient temperature data. The power range division module is used to set the power step size according to the rated power of the AC side equipment of the energy storage system, and divide the power range of similar loads or operating conditions according to the active power of the AC side equipment of the energy storage system based on the load measurement points, so as to distinguish the multi-dimensional operating data into each of the power ranges. The temperature rise deviation assessment module is used to calculate the group mean and standard deviation of the temperature rise of AC side equipment of the same type of energy storage system in each power range, and to assess the degree of deviation of the temperature rise of a single device from the group mean. The temperature grading and early warning module is used to convert the degree of deviation into a standardized equipment health score through a preset nonlinear mapping rule, and to classify the equipment temperature risk level according to the standardized health score to complete the grading and early warning.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the method for early warning of AC side equipment temperature of an energy storage system as described in any one of claims 1-7.

10. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method for early warning of AC side equipment temperature in an energy storage system as described in any one of claims 1-7.

Images