A method and system for managing abnormal power data

Abstract

This application discloses an abnormal power data management method and system, comprising: collecting power data, generating a spatial location map, identifying abnormal data based on the power data, identifying the cause of the abnormality based on the spatial location map and the abnormal data, and repairing the abnormal data; using a monitor to monitor entropy increase in multiple domains, the multiple domains including a physical domain, an information domain, and a business domain, wherein the physical domain is the physical equipment of the power data center; constructing a cross-domain entropy flow model based on the monitoring data of the physical domain, information domain, and business domain; predicting the entropy increase trend based on the monitoring data of the physical domain, information domain, and business domain; adjusting the entropy increase based on the cross-domain entropy flow model; establishing a situational model based on the power system status; timely detection and location of anomalies through real-time data collection and analysis; accurately identifying the root cause and propagation path of the anomaly through three-dimensional mapping and root cause decoupling analysis; and improving the repair efficiency of operation and maintenance personnel through dynamic guidance functions.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a method and system for managing abnormal power data. Background Technology

[0002] In today's digital age, the stability and reliability of power data centers are crucial, serving as core infrastructure supporting numerous critical business operations. With the continuous expansion of power data centers and the increasing complexity of their equipment, power data management faces many challenges. Power data centers generate massive amounts of power data, encompassing multiple dimensions such as equipment operation data, communication data, and environmental data. This data is not only vast in quantity but also complex in type, containing rich information but also potentially concealing various anomalies. If abnormal power data is not detected and processed in a timely manner, it may lead to a series of serious consequences, such as equipment failure, communication interruptions, and business system crashes, thereby affecting the normal operation of the entire power data center and even causing huge economic losses.

[0003] Traditional power data management methods often rely on manual inspections and simple threshold alarms. This approach is not only inefficient, but also makes it difficult to accurately identify the root causes and propagation paths of abnormal data, failing to meet the needs of modern power data centers for managing abnormal power data. Summary of the Invention

[0004] This application provides a method and system for managing abnormal power data. By collecting and analyzing data in real time, it can promptly detect and locate anomalies. Through three-dimensional mapping and root cause decoupling analysis, it can accurately identify the root cause and propagation path of anomalies. Through dynamic guidance functions, it can improve the repair efficiency of operation and maintenance personnel, realize the real-time location and root cause decoupling of concurrent anomalies in the power data center, and improve the operational stability and reliability of the power data center.

[0005] This application provides a method for managing abnormal power data, including:

[0006] S101: Collect power data, generate spatial location map, identify abnormal data based on power data, identify the cause of abnormality based on spatial location map and abnormal data, and repair abnormal data;

[0007] S102, Use a monitor to monitor the entropy increase of multiple domains, the multiple domains including physical domain, information domain and business domain, the physical domain being the physical equipment of the power data center;

[0008] S103, Construct a cross-domain entropy flow model based on monitoring data from the physical domain, information domain, and business domain;

[0009] S104, predict the entropy increase trend based on monitoring data from the physical domain, information domain, and business domain;

[0010] S105, adjust the entropy increase according to the cross-domain entropy flow model, and establish a situation model according to the power system state.

[0011] Preferably, the multi-domain entropy increase refers to the process by which a system in multiple domains transforms from an ordered state to a disordered state, that is, the process by which the entropy value of the system continuously increases.

[0012] Preferably, an entropy increase trend map is generated based on the predicted entropy increase trend, and an algorithm is used to identify the entropy increase inflection point based on the entropy increase trend map. The entropy increase inflection point marks the transition of the system from a steady state to a latent unstable state.

[0013] Preferably, the central difference method is used to calculate the first and second derivatives of the entropy increase curve, and the inflection point of entropy increase is identified based on the first and second derivatives. The formula for the central difference method is as follows: ,in, The first derivative of the entropy increase curve is represented by... Indicates a point in time The entropy increase Indicates a point in time The entropy increase Indicates a time interval.

[0014] Preferably,

[0015] S201, acquire the temperature of the physical device, and generate a heat vector diagram based on the device surface temperature and spatial information;

[0016] S202, calculate the thermal anomaly index based on the surface temperature of the physical equipment, and use the thermal anomaly index to identify thermal anomaly vortex zones;

[0017] S203 dissipates heat from the equipment based on the temperature of the thermal anomaly vortex zone and the surface temperature of the equipment.

[0018] Preferably, the convection rate is calculated based on the surface temperature of the physical equipment, using the following formula: ,in, Indicates the rate of natural air convection. This is an empirical coefficient. Indicates the surface temperature of the equipment. Indicates ambient temperature. Indicates an index.

[0019] Preferably, the thermal anomaly vortex zone is used to identify areas where heat is abnormally accumulated among equipment clusters.

[0020] Preferably,

[0021] S301, monitor internal changes in physical equipment and identify misalignment and accumulation of metal connectors in physical equipment;

[0022] S302, correct the heat vector diagram and thermal anomaly vortex region based on the misalignment and accumulation of metal connectors;

[0023] S303 uses a heat dissipation method based on the staggered stacking of metal connectors.

[0024] Preferably, the misalignment of the metal connectors is caused by a thermal hysteresis effect, which describes the phenomenon that the thermal response of a substance lags behind the temperature change.

[0025] This application also provides an abnormal power data management system, including a thermal flow field modeling system, a thermal anomaly vortex identification system, and a dynamic heat dissipation strategy formulation system. The thermal flow field modeling system is used to generate a heat vector diagram. The thermal anomaly vortex identification system is used to calculate a thermal anomaly index and identify thermal anomaly vortices. The dynamic heat dissipation strategy formulation system is used to determine the heat dissipation requirements of power equipment. The thermal flow field modeling system and the thermal anomaly vortex identification system are electrically connected, and the thermal anomaly vortex identification system and the dynamic heat dissipation strategy formulation system are also electrically connected.

[0026] One or more technical solutions provided in this application have at least the following technical effects or advantages: Real-time data collection and analysis enable timely detection and location of anomalies; Three-dimensional mapping and root cause decoupling analysis accurately identify the root causes and propagation paths of anomalies; dynamic guidance improves the repair efficiency of maintenance personnel, achieving real-time location and root cause decoupling of concurrent anomalies in the power data center, thus improving the operational stability and reliability of the power data center; Cross-domain entropy flow balance is achieved through cross-domain entropy increase monitoring, entropy flow coupling analysis, entropy increase inflection point prediction, and cross-domain entropy flow scheduling; Situational models and anomaly suppression strategies enable early detection and suppression of anomalies, achieving cross-domain entropy balance in the power data center, maintaining the steady-state operation of the system, proactively intervening before system instability occurs, and preventing steady-state collapse; Comprehensive system monitoring and anomaly suppression are achieved by identifying cross-domain entropy increases in the physical, information, and business domains.

[0027] Through steps such as thermal vector mapping, thermal anomaly vortex identification, dynamic heat dissipation strategy formulation, and heat dissipation assessment and verification, precise, dynamic, and efficient heat dissipation management is achieved. The end result is effective control of the surface temperature of power equipment, a significant reduction or elimination of thermal anomaly vortex zones, and a significant improvement in the stability and reliability of the power data center. This solves the problem of abnormal heat accumulation among equipment clusters in the power data center, ensuring efficient heat dissipation of power equipment and stable system operation.

[0028] By monitoring changes in the internal materials of equipment and identifying misaligned accumulation of metal connectors, combined with heat dissipation optimization algorithms, more precise and efficient heat dissipation management can be achieved, improving the stability and heat dissipation efficiency of power data centers, reducing heat dissipation energy consumption, and improving overall energy efficiency. This enables high-precision monitoring of misaligned accumulation of metal connectors, intelligent formulation and dynamic adjustment of heat dissipation strategies, reducing heat dissipation energy consumption, and improving overall energy efficiency. Attached Figure Description

[0029] Figure 1 This is a flowchart illustrating an abnormal power data management method and system according to the present invention.

[0030] Figure 2 This is a schematic diagram of the process for forming the heat vector diagram and the thermal anomaly vortex region in this invention;

[0031] Figure 3 This is a schematic diagram illustrating the process of identifying misaligned and stacked metal connectors in physical equipment according to the present invention. Detailed Implementation

[0032] To facilitate understanding of the present invention, a more complete description of this application will be given below with reference to the accompanying drawings, which illustrate preferred embodiments of the invention. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to enable a more thorough and complete understanding of the disclosure of the present invention.

[0033] It should be noted that the terms "vertical," "horizontal," "up," "down," "left," "right," and similar expressions used in this article are for illustrative purposes only and do not represent the only possible implementation.

[0034] 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 invention pertains; the terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to limit the invention; the term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0035] Example 1: Figure 1 This is a flowchart illustrating an abnormal power data management method according to an embodiment of the present invention, including:

[0036] S101: Collect power data, generate spatial location map, identify abnormal data based on power data, identify the cause of abnormality based on spatial location map and abnormal data, and repair abnormal data;

[0037] Furthermore, by deploying sensors in the power data center to collect real-time equipment operation data, communication data, and environmental data, the raw data is cleaned, standardized, and formatted using unified units. A detailed survey of the power data center's physical layout is conducted, resulting in a 3D model. The model labels the specific location of each device, including its cabinet, rack, row number, and column number. A unified naming convention and identifiers are used to identify the devices, and their adjacency and connection relationships are determined. Based on the device location labels and adjacency records, a spatial location map is generated. Network monitoring tools are used to analyze data transmission relationships between devices, identifying data senders and receivers, and recording the data transmission path from the source device to the target device, including intermediate devices and network nodes. A path diagram is used to represent the data transmission path, and a data flow graph is generated based on the data transmission relationships and paths. This data flow graph is used to visualize the flow of data within the power data center. A topology connection map is generated based on the device connectivity. According to the specific needs and historical data characteristics of the power data center, a deep learning algorithm (LSTM network) is used to monitor power data. Anomaly monitoring thresholds are set based on historical data and business requirements. Real-time monitored data is compared with these thresholds to determine if the data exceeds the normal range. When the monitored real-time data exceeds the preset threshold, it is identified as abnormal data, and relevant information such as occurrence time, device ID, anomaly type, and anomaly value is recorded. This triggers an anomaly alarm mechanism, notifying maintenance personnel or relevant systems to handle the abnormal data. Based on the abnormal data information, a spatial location model is used to pinpoint the specific spatial location of the anomaly, including the computer room, cabinets, and devices. A data flow model is used to analyze the flow path of the abnormal data within the power data center, including source devices, intermediate devices, and target devices. A topology connection model is used to confirm the connection relationships and connection strength between the abnormal device and other devices, determining whether the anomaly is caused by a connection problem.

[0038] Using a causal analysis algorithm, information such as the specific location of the anomaly, the data flow path, the involved devices, and their connections is input into the algorithm. The algorithm analyzes the propagation path of the anomaly between the physical, communication, and service layers, and sets a coupling strength threshold to quantify the coupling relationship between these layers. The coupling strength threshold can be set based on historical data and expert experience. Based on this threshold, the coupling strength between each layer is calculated. Coupling strength refers to the degree of correlation and influence between the anomaly and the various layers (physical, communication, and service layers) of the power data center. It can be measured by quantifying the anomaly's propagation capability, impact range, and interdependence between layers. The magnitude of the coupling strength reflects the correlation and propagation capability of the anomaly between layers. Coupling strength calculation is a current technology and will not be elaborated upon here. Based on the propagation path analysis and coupling strength calculation results, the ultimate root cause of the anomaly is determined.

[0039] The system acquires the physical layout diagram, equipment location information, and network topology diagram of the power data center, and analyzes the current location of maintenance personnel. Using the location information from the maintenance terminal, the system determines the current location of maintenance personnel. Combining the power data center layout information and the personnel's current location, it uses path planning algorithms such as A* to calculate the optimal route for maintenance personnel to reach the location of the anomaly. Based on the root cause of the anomaly, a specific repair strategy is formulated, including the equipment to be repaired, the repair methods, and the expected results. This strategy is then translated into specific repair suggestions, such as replacing faulty equipment, adjusting configuration parameters, and fixing software vulnerabilities. The generated repair suggestions and the optimal route are integrated into a guidance message, which is sent to the maintenance personnel's mobile device or handheld terminal via the communication module of the maintenance terminal. The maintenance personnel then perform repair operations at the location of the anomaly according to the repair suggestions. The system monitors anomaly-related indicators in real time to determine whether the anomaly has been successfully repaired.

[0040] S102, Use a monitor to monitor the entropy increase of multiple domains, including the physical domain, information domain and business domain;

[0041] Specifically, entropy is a concept in thermodynamics used to measure the degree of disorder or chaos in a system. Entropy increase refers to the process by which a system transforms from an ordered state to a disordered state, that is, the process by which the entropy value of the system continuously increases. Under natural conditions, isolated systems always tend towards entropy increase, meaning that the system will spontaneously evolve from a low-entropy (ordered) state to a high-entropy (disordered) state. Entropy increase manifests itself in the physical, information, and business domains as follows: In the physical equipment of a power data center, as the equipment is used and ages, its performance gradually declines, and the probability of failure increases. The process of disordering equipment performance is the entropy increase process in the physical domain. In communication networks, as data traffic increases and network equipment ages, the disorder of data transmission (such as packet loss rate and latency) increases. The process of disordering information transmission is the entropy increase process in the information domain. The entropy increase process in the information domain; In business systems, as business load increases and business processes become more complex, the system's load demand and power balance become difficult to control. The disordered process of business operation is the entropy increase process in the business domain. Monitoring physical domain entropy increase involves setting up sensors on physical devices to monitor their status in real time and transmitting the monitored data wirelessly to the data center monitoring platform. Monitoring the information domain involves setting up network monitors on network devices to monitor network bandwidth, latency, packet loss rate, and other indicators in real time, using network analysis tools like Wireshark to collect network monitoring data in real time, and assessing the entropy increase in the information domain. Load monitors are also set up on business system servers to monitor changes in load demand, power balance, and other business parameters in real time, identifying the entropy increase in the business domain.

[0042] S103, Construct a cross-domain entropy flow model based on monitoring data from the physical domain, information domain, and business domain;

[0043] Furthermore, monitoring data from the physical domain, information domain, and business domain are integrated through a data warehouse. The data warehouse includes a data source layer, a data storage layer, a data processing layer, and a data access layer. The data source layer contains various data sources from the physical, information, and business domains. The data storage layer uses a relational database as the storage for the data warehouse and sets up the table structure. The data processing layer is used to design data cleaning, transformation, and loading (ETL) processes. The data access layer provides a data access interface, cleans the integrated data, and uses Pearson correlation coefficients to calculate the correlation coefficients between the physical and information domains, the physical and business domains, and the information and business domains. The calculated correlation coefficients are then used to... The coefficients are integrated into a matrix, with rows and columns representing different domains. The elements in the matrix represent the correlation coefficients between corresponding domains. According to actual needs, a threshold for the correlation coefficient is set. Domain pairs with correlation coefficients exceeding the threshold are extracted from the correlation matrix. These domain pairs have a significant correlation in entropy increase. Domain pairs with correlation coefficients exceeding the threshold are regarded as key coupling factors for cross-domain entropy flow. Time series data of entropy increase in each domain are extracted from the data warehouse. A Bayesian network is used to train the extracted time series data. Based on the training results, a directed acyclic graph and a conditional probability table are obtained to form a causal relationship model. The constructed causal relationship model is validated using historical data. The generalization ability of the model is evaluated through cross-validation.

[0044] Based on matrix and causal relationship models, the association paths and causal chains of entropy increase between the physical domain, information domain, and business domain are identified. Nodes are created for each domain, and nodes with significant correlations are connected by dashed lines based on coefficients, with positive or negative correlations marked on the lines. According to the causal relationship model, nodes with causal relationships are connected by solid lines with arrows, the direction of the arrows indicating the causal direction, forming a conceptual map of cross-domain entropy flow. This map shows the flow direction and coupling points of entropy between different domains. The coupling points are the intersections of association paths or places with significant interactions; they are the convergence areas where entropy increase influences each other across multiple domains. A cross-domain entropy flow model is constructed using differential equation models, and the model is validated using historical data. The model's predictions are compared with actual data, and the values ​​of validation indicators are calculated. The initial framework of the model is built using the modeling tool AnyLogic, describing the coupling relationships between the domains, including direct and indirect coupling.

[0045] The coupling strength is represented by parameters or weights. The entropy increase data of the physical domain, information domain and business domain are used as input to the cross-domain entropy flow model. The output of the cross-domain entropy flow model is the state of the cross-domain entropy flow and the entropy value of each domain.

[0046] S104, predict the entropy increase trend based on monitoring data from the physical domain, information domain, and business domain;

[0047] Specifically, monitoring data from the physical, information, and business domains are extracted from the data warehouse. This monitoring data is entropy increase data. The Statsmodels time series analysis tool is used to identify the long-term trend and short-term fluctuations of entropy increase. Fourier transform is used to convert the time series data from the time domain to the frequency domain, resulting in a spectrum. In the spectrum, the frequency axis represents the reciprocal of the period, and the amplitude axis represents the intensity of the corresponding frequency component. The frequency components with larger amplitudes are identified, and the periods corresponding to these frequency components are the periods of the entropy increase data. The time series data is divided into training and testing sets. The training set is used to train a Long Short-Term Memory (LSTM) network model, and the model parameters are adjusted to optimize prediction performance. The trained model is used to predict the test set, and the error between the predicted and actual values ​​is calculated. Based on the error, the model is adjusted, and the trained LSTM network model is used to predict entropy increase, generating an entropy increase trend graph. Based on the entropy increase trend graph, an algorithm is used to identify the entropy increase inflection point, which marks the transition of the system from a steady state to a latently unstable state. In power data centers, as entropy accumulates in the physical, information, and business domains, the system gradually deviates from its initial steady state. When entropy increases to a certain level, reaching an inflection point, the system faces the risk of latent instability. This algorithm combines first-order derivatives, second-order derivatives, and periodic verification to identify the inflection point. The time series data of entropy increase is S={ The first derivative of the entropy increase curve is calculated using the central difference method, and the formula is: ,in, The first derivative of the entropy increase curve is represented by... Indicates a point in time The entropy increase Indicates a point in time The entropy increase The first derivative represents the time interval and reflects the instantaneous rate of change of entropy. The second derivative of the entropy increase curve is calculated using the central difference method. For a sequence of first derivatives, the sequence is iterated through to find extreme points, including maxima and minima. These extreme points satisfy the following conditions: and The maximum point or and The minimum point, the candidate inflection point of the entropy increase curve corresponding to the extreme point, traverse the second derivative sequence to find the point of sign change, the point of sign change satisfies The sign change points correspond to changes in the concavity and convexity of the entropy increase curve. Based on the period of the entropy increase data, for each candidate inflection point, it is checked whether it matches the known periodic changes. If the candidate inflection point does not match the periodic analysis results, the inflection point is corrected or eliminated; otherwise, the candidate inflection point is the inflection point. The threshold for entropy increase is set based on historical data and expert experience. When the entropy increase data exceeds the warning threshold, the warning information generation mechanism is triggered. The warning information should include key information such as the occurrence time and location of the entropy increase inflection point and the predicted entropy increase rate.

[0048] S105, adjust the entropy increase according to the cross-domain entropy flow model, and establish a situation model according to the system state;

[0049] Furthermore, the cross-domain entropy flow model describes the flow patterns and mutual influences of entropy between different domains. Based on real-time monitored entropy increase data and trends, scheduling is dynamically adjusted. When a sudden increase in entropy is detected in a domain, and trend analysis indicates that the entropy increase will continue, the operating parameters of the equipment or network communication paths in that domain are immediately adjusted to suppress the entropy increase. Simultaneously, the adjusted strategies are fed back into the cross-domain entropy flow model to update model parameters and improve model accuracy. The system status is monitored in real-time. Combining historical system data, time series analysis is used to identify patterns and trends in the historical data. Real-time monitoring data is fused with historical data, and a situational awareness model is constructed using a dynamic Bayesian network method. When the model predicts that the equipment status in a domain is about to reach a fault threshold, backup equipment is activated in advance to ensure continuous system operation. When the model predicts that network traffic is about to become congested, communication paths are adjusted and network resource allocation is optimized. Priority for anomaly suppression strategies is set based on the severity and scope of the anomaly. This ensures that when multiple anomalies occur simultaneously, the anomaly with the greatest impact on the system can be addressed first.

[0050] The technical solutions described in the embodiments of this application above have at least the following technical effects or advantages: By collecting and analyzing data in real time, anomalies can be detected and located promptly; by using three-dimensional mapping and root cause decoupling analysis, the root causes and propagation paths of anomalies can be accurately identified; by using dynamic guidance functions, the repair efficiency of maintenance personnel can be improved; real-time location and root cause decoupling of concurrent anomalies in the power data center can be achieved, thereby improving the operational stability and reliability of the power data center; by using cross-domain entropy increase monitoring, entropy flow coupling analysis, entropy increase inflection point prediction, and cross-domain entropy flow scheduling, the balance of cross-domain entropy flow can be achieved; by using situational models and anomaly suppression strategies, the occurrence and propagation of anomalies can be detected and suppressed in advance, achieving cross-domain entropy balance in the power data center, maintaining the steady-state operation of the system, proactively intervening before the system becomes implicitly unstable, and preventing steady-state collapse; by identifying cross-domain entropy increase in the physical domain, information domain, and business domain, comprehensive system monitoring and anomaly suppression are achieved.

[0051] Example 2: Based on Example 1 above, this example uses overheat flow field modeling, thermal anomaly vortex identification, dynamic heat dissipation strategy formulation, and heat dissipation evaluation and verification to achieve efficient heat dissipation of the power equipment cluster and ensure the stable operation of the power data center. Figure 2 As shown.

[0052] S201, acquire the temperature of the physical device, and generate a heat vector diagram based on the device surface temperature and spatial information;

[0053] Specifically, the heat output of power equipment is obtained in the physical domain through the equipment's operation logs. Spatial information is acquired based on the physical layout of the power data center. Ambient temperature monitoring points are set up within the power data center to monitor the ambient temperature. A thermal flow field modeling system is built, and the acquired equipment heat output and spatial information are input into this system. Fiber optic temperature sensors are installed on the equipment surface to monitor the surface temperature in real time, and the data is transmitted to the thermal flow field modeling system. The difference between the equipment surface temperature and the ambient temperature is calculated, and the convection rate of the natural convection field is calculated using this difference. The formula for calculating the convection rate is: ,in, It represents the natural convection rate of air, reflecting how fast air moves under natural convection conditions. This is an empirical coefficient, obtained through experiments. This indicates the surface temperature of the equipment, which is the actual temperature reached by the equipment's surface during operation. This indicates the ambient temperature, which is the air temperature surrounding the equipment inside the power data center. The index represents the degree of influence of the temperature difference between the equipment surface and the ambient temperature on the natural convection rate of air. It is usually taken as 1 / 3 or 1 / 4, and in this embodiment, it is taken as 1 / 3.

[0054] Based on the convection rate and equipment space information, heat conduction and convection heat transfer models are used to calculate heat conduction and convection heat transfer. Through continuous iteration of the heat conduction and convection heat transfer models, heat diffusion paths are formed. Vector data is generated based on the heat diffusion paths. The magnitude of the vector represents the amount of heat, which is usually related to the intensity of heat diffusion. The direction of the vector represents the direction of heat diffusion, which is consistent with the direction of air flow. The generated vector data is drawn in the form of a vector diagram using drawing software. In the vector diagram, each vector is represented by an arrow. The length of the arrow represents the magnitude of the vector, and the direction of the arrow represents the direction of the vector. In the vector diagram, areas with small vectors and disordered directions are identified. Heat diffusion is hindered in these areas, forming heat stagnation. The identified heat stagnation areas are marked on the heat flow vector diagram.

[0055] S202, calculate the thermal anomaly index based on the surface temperature of the physical equipment, and use the thermal anomaly index to identify thermal anomaly vortex zones;

[0056] Furthermore, the temperature rise rate is calculated based on the equipment surface temperature measured in step S201, using the following formula: ,in, Indicates the rate of temperature rise. Indicates the final temperature of the equipment surface. This indicates the initial temperature of the device surface. To represent the time interval, airflow velocity sensors are placed around the device to calculate the airflow stagnation index, using the following formula: ,in, Indicates the airflow stagnation index. This indicates the duration during which the surrounding air velocity is less than 0.1 m / s within a time window. The total time is represented by the thermal anomaly index, which is calculated based on the rate of temperature rise and the airflow stagnation index. The formula is: ,in, This is the thermal anomaly index. For the rate of temperature rise, The airflow stagnation index is set according to the heat dissipation characteristics of the power equipment. Based on the calculated thermal anomaly index, the area where the thermal anomaly index is greater than the set threshold for a continuous period of time is the thermal anomaly vortex area. The thermal anomaly vortex area is used to identify the area where heat is abnormally accumulated between equipment clusters. These areas are prone to local overheating due to airflow organization failure, which affects the normal operation of the equipment. The thermal anomaly vortex area is marked in the heat flow vector diagram, and the location, size and severity of the thermal anomaly vortex area are recorded.

[0057] S203, dissipate heat from the equipment based on the temperature of the thermal anomaly vortex zone and the surface temperature of the equipment;

[0058] Specifically, safety thresholds are set based on the equipment type, material, and operating conditions. When a thermal anomaly vortex zone is detected or the equipment surface temperature approaches the safety threshold, the equipment is deemed to require heat dissipation. The severity of the thermal anomaly vortex zone is graded into three levels: mild, moderate, and severe. For mild to severe thermal anomaly vortex zones, a short-term heat dissipation period (10 minutes) is planned; for moderate to severe thermal anomaly vortex zones, a longer-term heat dissipation period (30-60 minutes) is planned. During the heat dissipation process, the changes in equipment surface temperature and the thermal anomaly vortex zone are monitored in real time. If the equipment surface temperature decreases rapidly or the thermal anomaly index of the thermal anomaly vortex zone decreases significantly, the heat dissipation time can be appropriately shortened; conversely, if the temperature decreases slowly or the thermal anomaly index does not change significantly, the heat dissipation time can be appropriately extended. Heat dissipation can be achieved through methods such as using fans, adjusting equipment layout density, and optimizing airflow channels.

[0059] The technical solutions described in the embodiments of this application have at least the following technical effects or advantages: through steps such as heat vector mapping, thermal anomaly vortex identification, dynamic heat dissipation strategy formulation, and heat dissipation evaluation and verification, accurate, dynamic, and efficient heat dissipation management is achieved. The final result is that the surface temperature of power equipment is effectively controlled, thermal anomaly vortex zones are significantly reduced or eliminated, the stability and reliability of the power data center are significantly improved, the problem of abnormal heat accumulation between equipment clusters in the power data center is solved, and efficient heat dissipation of power equipment and stable system operation are ensured.

[0060] Example 3: Building upon Example 2, this example addresses the issue of thermal hysteresis caused by accumulated thermal fatigue by dynamically modeling the heat flow field and thermal anomaly vortices, thus ensuring the stable operation and equipment safety of the power data center. Figure 3 As shown.

[0061] S301, monitor internal changes in physical equipment and identify misalignment and accumulation of metal connectors in physical equipment;

[0062] Furthermore, the internal changes of the equipment material are microscopic changes. X-ray diffraction (XRD) technology is used to identify the dislocation accumulation of metal connectors. If the diffraction peaks of the metal connectors shift and the peak intensity decreases, it indicates that dislocation accumulation has occurred in the metal connectors. Dislocation accumulation in metal connectors increases the scattering probability of electrons in the metal, leading to an increase in resistivity. When electrons move in the crystal, they collide with defects such as dislocations, hindering the free flow of electrons, thereby increasing the resistance. A four-wire resistance meter is used to monitor the resistance of the metal connectors in real time. By measuring the change in resistance value, combined with parameters such as the size and temperature of the connectors, the change in resistivity is calculated. When the resistivity exceeds the threshold, dislocation accumulation exists in the connectors. The dislocation accumulation of metal connectors is formed by the thermal hysteresis effect. The entropy hysteresis effect describes the phenomenon that the thermal response of a substance lags behind the temperature change. In metal connectors, when the temperature changes, due to the thermal hysteresis effect, the atomic motion state inside the metal does not immediately adjust to a new stable state with the temperature change, thus forming dislocation accumulation in the metal connectors.

[0063] S302, correct the heat vector diagram and thermal anomaly vortex region based on the misalignment and accumulation of metal connectors;

[0064] Specifically, misalignment and accumulation of metal connectors can cause frictional heating, resulting in an increase in total heating power, including both the heating power during normal operation and the power generated by friction. Simultaneously, thermal resistance also increases. The corrected heating power and thermal resistance parameters are input into the numerical simulation software ANSYS Fluent. In the simulation, boundary conditions are set in the misaligned accumulation area, treating it as an obstacle and increasing the local drag coefficient to simulate the formation of vortices due to obstructed airflow. ANSYS Fluent outputs the natural convection field distribution. The heat flow vector is calculated using the heat flow calculation module in the numerical simulation software. By solving the coupled equations of heat conduction and heat convection, the heat flow distribution inside and on the surface of the connector is obtained. The changing trend of the heat flow vector is analyzed, and a heat flow vector distribution map is plotted. By observing the magnitude and direction of the heat flow vector, areas of concentrated heat flow and areas of weak heat flow are identified. Areas of concentrated heat flow typically correspond to high-temperature regions of the connector or near heat sources, while areas of weak heat flow may appear in locations with high thermal resistance or poor airflow.

[0065] S303, heat dissipation method is set according to the staggered stacking of metal connectors;

[0066] Furthermore, by monitoring the rate of change of temperature, pressure, or heat flux, when these rates exceed preset thresholds, it is determined that thermal hysteresis has caused misalignment and accumulation in the metal connectors. If a region exhibits turbulent local heat flow, a large temperature gradient, and obstructed heat transfer, this region is identified as a thermally abnormal vortex zone. When misalignment and thermally abnormal vortex zones are detected in the metal connectors, heat dissipation is required. The severity of the misalignment and accumulation is assessed by measuring the area and volume of the thermally abnormal vortex zone and its distance from critical parts of the connectors. Based on the severity of the misalignment and the characteristics of the thermally abnormal vortex zone, a heat dissipation time is planned. For equipment with severe misalignment or large thermally abnormal vortex zones, the heat dissipation time is increased. For example, if the overall score for misalignment or accumulation is high or the vortex zone area is large, the heat dissipation time is set to 1.5-2 hours as normal. The heat dissipation intensity is measured by parameters such as the power, wind speed, or water flow rate of the heat dissipation equipment. Based on the severity of the misalignment and accumulation of metal connectors and the characteristics of the thermal anomaly vortex zone, the parameters of the heat dissipation equipment can be adjusted to increase the heat dissipation intensity. For example, for areas with severe misalignment and accumulation of metal connectors, the fan speed can be increased or the water flow rate of the water cooling system can be increased.

[0067] The technical solutions described in the above embodiments of this application have at least the following technical effects or advantages: by monitoring the internal changes of equipment materials, identifying the misalignment and accumulation of equipment metal connectors, and combining with heat dissipation optimization algorithms, more accurate and efficient heat dissipation management can be achieved, thereby improving the stability and heat dissipation efficiency of power data centers, reducing heat dissipation energy consumption, improving overall energy efficiency, realizing high-precision monitoring of misalignment and accumulation of equipment metal connectors, realizing intelligent formulation and dynamic adjustment of heat dissipation strategies, reducing heat dissipation energy consumption, and improving overall energy efficiency.

[0068] Example 4: This example provides an abnormal power data management system, including a thermal flow field modeling system, a thermal anomaly vortex identification system, and a dynamic heat dissipation strategy formulation system. The thermal flow field modeling system includes a data input interface, a model calculation module, and a result output module. The data input interface receives various data from the power equipment monitoring system, including temperature data, current data, voltage data, and equipment structural parameters. The model calculation module performs numerical calculations, and the result output module outputs the calculation results and generates a heat vector diagram. The thermal anomaly vortex identification system integrates a thermal anomaly index calculation module and a black hole determination rule. The thermal anomaly index calculation module calculates the thermal anomaly index, and the black hole determination rule is... A set of judgment criteria based on thermal entropy theory and practical experience is used to identify whether there are thermal anomaly vortex regions in power equipment. The dynamic heat dissipation strategy formulation system includes a heat dissipation demand judgment module, a heat dissipation time planning module, and a heat dissipation method selection module. The heat dissipation demand judgment module is used to dynamically judge the heat dissipation demand of the power equipment based on the judgment results of the thermal anomaly vortex identification system. The heat dissipation time planning module is used to plan the heat dissipation time arrangement based on the results of the heat dissipation demand judgment module. The heat dissipation method selection module is used to select heat dissipation methods based on the heat dissipation demand and the heat dissipation time planning. The thermal flow field modeling system is electrically connected to the thermal anomaly vortex identification system, and the thermal anomaly vortex identification system is electrically connected to the dynamic heat dissipation strategy formulation system.

[0069] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. For those skilled in the art, the present invention can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. An abnormal power data management method characterized by, Comprise: S101, collect power data, generate spatial position atlas, identify abnormal data according to power data, identify abnormal cause based on spatial position atlas and abnormal data, repair abnormal data; S102, monitor multi-domain entropy increase using a monitor, the multi-domain includes physical domain, information domain and business domain, the physical domain is the physical equipment of the power data center; S103, construct cross-domain entropy flow model according to the monitoring data of the physical domain, information domain and business domain; S104, predict entropy increase trend according to the monitoring data of the physical domain, information domain and business domain; S105, adjust entropy increase according to the cross-domain entropy flow model, and establish a situation model according to the state of the power system; the S105 specifically comprises: Real-time monitoring of the state of the system, combining historical data of the system, identifying patterns and trends in historical data using time series analysis, fusing real-time monitoring data with historical data, using dynamic Bayesian network method to construct situation model, when the model predicts that the state of a device in a certain domain will reach the failure threshold, start the standby device in advance to ensure the continuous operation of the system; when the model predicts that the network traffic will be congested, adjust the communication path and optimize the network resource configuration, set the priority of the abnormal suppression strategy according to the severity and influence range of the abnormality.

2. The method of claim 1, wherein, The multi-domain entropy increase refers to the process of the system in multiple fields from an ordered state to a disordered state, that is, the process of increasing the entropy value of the system.

3. The method of claim 1, wherein, Generate entropy increase trend graph according to the predicted entropy increase trend, use algorithm to identify entropy increase inflection point, which marks the transition of the system from steady state to implicit instability state.

4. The method of claim 3, wherein, The first derivative and the second derivative of the entropy increase curve are calculated using the central difference method, and the entropy increase inflection point is identified according to the first derivative and the second derivative, and the central difference method calculation formula is: wherein, represents the first derivative of the entropy increase curve, represents the entropy increase value at the time point , represents the entropy increase value at the time point , represents the time interval.

5. An abnormal power data management system applied to the abnormal power data management method according to any one of claims 1 to 4, characterized by, Comprise a thermal flow field modeling system, the thermal flow field modeling system comprises a data input interface, a model calculation module and a result output module, the data input interface receives various data from the power equipment monitoring system, including current data, voltage data and equipment structure parameters of the power equipment, the model calculation module is used for numerical calculation, and the result output module is used for outputting the result calculated by the calculation module.

Images