Multi-dimensional data collaborative analysis early warning method based on intelligent operation and maintenance of power transmission network

Abstract

The invention discloses a multi-dimensional data collaborative analysis early warning method based on intelligent operation and maintenance of a power transmission network, and relates to the technical field of power transmission network early warning, and the method comprises the steps: collecting target data of target equipment, and obtaining sudden change data of a power load in a system; performing correlation analysis on the instant characteristics of the electrical quantity and the dynamic change of the power grid structure in combination with the sudden change data of the power load to obtain correlation factors capable of predicting a fault; according to the power transmission network intelligent operation and maintenance early warning method, the problem that fault early warning is inaccurate and not timely due to insufficient data collaboration and early warning response lag in power transmission network operation and maintenance is solved, and the effect of improving the early warning precision and response speed of power transmission network intelligent operation and maintenance is achieved.

Description

Technical Field

[0001] The present invention belongs to the technical field of power transmission network early warning, and specifically is a method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission networks. Background Art

[0002] In practical applications, the collaborative analysis and early warning method based on multi-dimensional data of transmission network intelligent operation and maintenance often faces the problem of insufficient data collaborative application.

[0003] Taking the fault warning in the daily operation and maintenance of the transmission network as an example, after the system collects the target equipment data and the power load mutation data, it fails to fully integrate the power load mutation data into the correlation analysis of the instantaneous characteristics of the electrical quantity and the dynamic changes of the grid structure, resulting in difficulty in accurately obtaining the correlation factors that can predict faults. At the same time, when determining the way in which the various correlation factors influence each other, the lack of targeted analysis combined with the power load mutation amplitude makes the analysis results lack accuracy. In addition, the call of the warning model combination is not fully based on the real-time values ​​and influence methods of the correlation factors, the information exchange between models is not timely, and when the deviation exceeds the set value, the adjustment of the adaptation relationship between the data collection frequency and the load control parameters fails to be effectively combined with the model judgment results, resulting in insufficient real-time and effectiveness of the warning plan; This situation not only affects the efficiency of transmission network operation and maintenance, but may also pose potential risks to the stable operation of the power grid due to untimely or inaccurate early warning. The problem is particularly prominent when the power grid load fluctuates greatly. Summary of the Invention

[0004] The purpose of the present invention is to provide a collaborative analysis and early warning method based on multi-dimensional data of intelligent operation and maintenance of transmission network, which solves the technical problems of inaccurate and untimely fault warning caused by insufficient data collaboration and delayed early warning response in the operation and maintenance of transmission network in the prior art.

[0005] The early warning method based on the collaborative analysis of multi-dimensional data of intelligent operation and maintenance of the transmission network includes: Collect target data from target devices and obtain sudden changes in power load in the system; Combined with the sudden change data of power load, the instantaneous characteristics of electrical quantities and the dynamic changes of the grid structure are analyzed to obtain the correlation factors that can predict faults; According to the sudden change amplitude of power load, determine the mutual influence mode between various related factors; According to the real-time values ​​and impact modes of related factors, the corresponding early warning model combination is called, and the models communicate with each other in a timely manner as the factors change; When the deviation exceeds the set value, the linkage response of data collection frequency and load control parameters is taken as the core, combined with the judgment results of each model, the adaptation relationship between the two is dynamically adjusted to generate a real-time early warning plan.

[0006] Furthermore, correlation analysis is performed on the instantaneous characteristics of electrical quantities and the dynamic changes of the grid structure, including the following steps: S1: Convert the instantaneous characteristics of electrical quantities into corresponding energy signals, and then map them to relevant nodes in the power grid structure; S2: After receiving the signal, the node generates an initial state change; S3: Tracks the chain transmission path formed by this state change as the grid structure changes. Each node in the path will provide real-time feedback on the impact of its own state change on the upstream signal. S4: Add the effect back to the original corresponding node to form the first round of interactive cycle between features and structures; S5: taking the result of the cycle as a new starting point, repeating the above steps S1 to S4; S6: Obtain the cumulative deviation of node states in multiple cycles and extract related factors. In each cycle, the real-time load fluctuation of the power grid is introduced as a dynamic adjustment parameter to make the factor extraction process consistent with the actual situation of power grid operation.

[0007] Furthermore, when determining how each related factor influences each other: Based on the correspondence between electrical quantity characteristics and grid structure nodes, the initial causal relationship sequence is established according to the order in which factors appear; Refer to the acquired power load mutation data and verify the effectiveness of the triggering of each link in the sequence through real-time data. Replace and complete invalid links. The completion process needs to include relevant factors that may affect the subsequent model judgment information exchange; Based on the extraction of the correlation factors, the response delay differences and amplitude fluctuation ranges of other correlation factors under different grid load conditions after the change of a single correlation factor are recorded, and the possible impact of the recorded results on the subsequent deviation judgment link is simultaneously predicted; Based on the above records, a dynamically updateable multi-scenario impact relationship table is formed, and the content in the table is continuously revised as new collected data is added.

[0008] Furthermore, when collecting target data, include: First, the filter threshold is dynamically adjusted based on the real-time load fluctuation of the power grid, and then the algorithm is used to distinguish the interference information from the effective features in the data; The filtered valid data is bound to the key node parameters of the grid structure changes and time-series marked, so that when the archived data enters the association analysis link, the corresponding factor transmission path can be automatically activated according to the mark.

[0009] Furthermore, when distinguishing interference information from effective features through dynamic filtering algorithm: First, a reference signal generated by the current load fluctuation of the power grid is captured. Then, a differential analysis is performed between the electrical quantity data and the reference signal to extract abnormal segments where the difference exceeds the dynamically generated threshold. For abnormal fragments, the source is traced in combination with the real-time trajectory of the grid structure changes. If it can be associated with the state change of a specific node, it is marked as a valid feature; otherwise, it is judged as interference information.

[0010] Furthermore, when calling the early warning model combination, it includes: First, convert the mutual influence intensity of the correlation factors into a dynamic adjustment coefficient. The higher the influence intensity, the larger the coefficient value. The computing power weight is assigned to the corresponding model according to the coefficient value, and after each round of analysis, the coefficient value is reverse-calibrated according to the newly generated influence relationship table, so that the computing power allocation can adapt to the factor association rules in real time.

[0011] Furthermore, the judgment information exchanged between models includes factor feature matching results, trend change predictions, and risk probability assessments: When the real-time numerical variation of the associated factors exceeds the set ratio, the relevant models will synchronously send updated judgment information, and the receiving models will cross-check with their own analysis results; The consistent contents of the verification results will be used as the basis for comprehensive judgment, and the inconsistent contents will be subject to secondary analysis.

[0012] Furthermore, when the deviation exceeds the set value, it also includes: Focusing on the linkage response of data collection frequency and load control parameters, the judgment results of each model are summarized to obtain consistent conclusions on the causes of deviations. According to the conclusion, the relationship between data acquisition frequency and load control parameters is adjusted. The greater the load control range, the higher the acquisition frequency. Combining the adjusted coordination relationship and the results of the model supplementary analysis, a real-time early warning plan is generated that includes risk level, impact scope and regulatory recommendations.

[0013] Furthermore, when a sudden load shock occurs in the power grid, it also includes: Suspend the conventional factor correlation analysis process and extract the electrical quantity mutation characteristics before the shock occurs; The above characteristics are quickly compared with the typical characteristics in the historical impact case library, and the reference plan is generated after sorting by matching degree; At the same time, each model calculates the tolerance of the current power grid structure in parallel, and combines the calculation results with the reference scheme to form a temporary control strategy; Once the shock is alleviated, regular analysis will be conducted and temporary strategies will be calibrated.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention conducts multi-dimensional correlation analysis based on the mutation data of power load, determines the mutual influence mode of correlation factors according to the mutation amplitude, calls the early warning model combination and realizes information exchange between models, and generates an early warning plan by dynamically adjusting the adaptation relationship between data acquisition frequency and load control parameters. It solves the problems of inaccurate and untimely fault warning caused by insufficient data coordination and delayed early warning response in the operation and maintenance of the transmission network, and achieves the effect of improving the early warning accuracy and response speed of the intelligent operation and maintenance of the transmission network. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 Schematic diagram of the framework structure of the method of the present invention. DETAILED DESCRIPTION

[0016] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0017] See also Figure 1 This application provides a collaborative analysis and early warning method for multi-dimensional data based on intelligent operation and maintenance of power transmission networks, including: Collect target data from target devices and obtain sudden changes in power load in the system; It should be noted that target data refers to various types of data related to the operating status of target equipment in the transmission network. It can be achieved through real-time collection by sensors, uploading by the equipment's own monitoring module, or manual regular entry. It is mainly used to reflect the basic operating conditions of the target equipment and provide basic information for subsequent analysis.

[0018] Combined with power load mutation data, correlation analysis is performed on the instantaneous characteristics of electrical quantities and the dynamic changes in the grid structure to obtain correlation factors that can predict faults. It should be noted that power load mutation data is data indicating that the load in the power system has changed significantly in a short period of time. This can be achieved by using real-time recording of power monitoring systems, collection of data by load management terminals, or comparison and identification of historical load data through data analysis platforms. Furthermore, the associated factors that can predict failures refer to various factors that are potentially related to power grid failures, which are obtained through association analysis. They can be implemented by using big data association rule mining, machine learning classification algorithms, or expert experience screening. For example, the correspondence between instantaneous fluctuations in electrical quantities and weak links in the power grid structure is mainly used to discover potential hidden dangers that may lead to failures in advance and provide a basis for early warning.

[0019] Based on the sudden change amplitude of power load, determine the mutual influence mode between various related factors; it should be understood that the mutual influence mode between related factors refers to the functional relationship between different related factors in the process of affecting the operating status of the power grid, which can be achieved by using causal analysis, correlation calculation or system dynamics model, for example, a positive correlation relationship in which the enhancement of a certain related factor will lead to a significant change in another related factor. It is mainly used to clarify the mechanism of action between various factors and improve the accuracy of early warning analysis.

[0020] According to the real-time values ​​and impact methods of the correlation factors, the corresponding early warning model combination is called, and the models communicate with each other in a timely manner as the factors change. It should be noted that the early warning model combination refers to a collection of multiple models selected to achieve transmission network fault early warning, such as a neural network-based fault prediction model, a decision tree-based risk assessment model, etc., which corresponds to the early warning strategies under different combinations of correlation factors (such as warning thresholds, response levels, processing procedures, etc.).

[0021] When the deviation exceeds a set value, the system dynamically adjusts the matching relationship between data collection frequency and load control parameters, focusing on the coordinated response between data collection frequency and load control parameters, combined with the judgment results of each model, to generate a real-time early warning solution. For example, the coordinated response between data collection frequency and load control parameters refers to the process of mutual coordination and synergy between the data collection time interval and the various parameters used to regulate power loads. This can be achieved using preset linkage rules, dynamic response algorithms, or intelligent control strategies.

[0022] Specifically, the present invention conducts multi-dimensional correlation analysis based on the mutation data of the power load, determines the mutual influence mode of the correlation factors according to the mutation amplitude, calls the early warning model combination and realizes information exchange between models, and generates an early warning plan by dynamically adjusting the adaptation relationship between the data acquisition frequency and the load control parameters, thereby solving the problems of inaccurate and untimely fault warnings caused by insufficient data coordination and delayed early warning responses in the operation and maintenance of the transmission network, and achieving the effect of improving the early warning accuracy and response speed of the intelligent operation and maintenance of the transmission network.

[0023] It should be understood that in some examples of this embodiment, the intelligent operation and maintenance system of the transmission network may include one or more data acquisition devices, such as smart meters for collecting electrical quantities, and a data processing center, such as a high-performance server cluster.

[0024] Specifically, the data processing center receives target data and power load mutation data from the data acquisition device, analyzes correlation factors, and determines how they influence each other. The data processing center stores preset early warning model combinations and corresponding processing strategies for different correlation factors. When a real-time numerical deviation of a correlation factor is detected that exceeds a set value, the data processing center invokes the corresponding early warning model combination. The models communicate their judgment information in real time and, based on their judgment results, dynamically adjust the matching relationship between data collection frequency and load control parameters. The system generates a real-time early warning plan based on the adjusted matching relationship. For example, when the load mutation amplitude is large and the correlation factors indicate high risk, the data collection frequency is increased to the maximum level and the load control parameters are adjusted to limit the rate of load growth. The data processing center sends the generated early warning plan to the operation and maintenance terminal, prompting personnel to take appropriate measures. During the intelligent operation and maintenance of the power transmission network, through multi-dimensional data collaborative analysis and dynamic parameter adjustment, timely and accurate fault warnings can be issued. This solves the problem of poor early warning effectiveness caused by isolated data and rigid responses in existing technologies, reducing the probability of power transmission network failures. It has improved the stability and reliability of the transmission network operation, reduced operation and maintenance costs, and enhanced the intelligent management level of the power grid.

[0025] In one embodiment, correlation analysis of instantaneous characteristics of electrical quantities and dynamic changes in power grid structure includes the following steps: S1: Convert the instantaneous characteristics of electrical quantities into corresponding energy signals, and then map them to relevant nodes of the power grid structure. It should be noted that energy signals refer to quantifiable energy representations obtained by mapping the instantaneous characteristics of electrical quantities such as voltage and current (such as peak value and mutation rate) through energy conversion algorithms (such as Fourier transform and wavelet analysis). These energy signals can be converted through the signal processing module of the power grid monitoring terminal.

[0026] S2: After receiving the signal, the node generates an initial state change; S3: Track the chain transmission path formed by this state change as the grid structure changes. Each node in the path will provide real-time feedback on the impact of its own state change on the upstream signal. It should be noted that the chain transmission path refers to the path formed by the node state change in the grid structure being transmitted in sequence according to the topological connection relationship. It can be dynamically tracked and determined based on the grid GIS topology map and real-time status monitoring data.

[0027] S4: Add the effect back to the original corresponding node to form the first round of interactive cycle between features and structures; S5: taking the result of the cycle as a new starting point, repeating the above steps S1 to S4; S6: Obtain the cumulative deviation of node states over multiple cycles and extract correlation factors. In each cycle, the real-time load fluctuations of the power grid are introduced as dynamic adjustment parameters to ensure that the factor extraction process is consistent with the actual operation of the power grid. It should be noted that the cumulative deviation refers to the sum of the deviations between the actual state of each node and the theoretical stable state over multiple cycles. This can be obtained by continuously recording node state data and calculating the deviations after each cycle. Dynamic adjustment parameters are quantitative indicators that reflect the degree of real-time load fluctuations in the power grid (such as load fluctuation rate and load mutation amplitude). These can be collected in real time by the load management system and transmitted to the correlation analysis module.

[0028] For example, in existing technologies, if only electrical quantity characteristics or grid structure are analyzed separately, it is difficult to capture the potential fault correlation under the dynamic interaction of the two. Electrical quantity changes may be caused by grid structure adjustments, and structural state changes may also react to electrical quantity characteristics. Single-dimensional analysis tends to ignore this mutual influence. This embodiment, by converting instantaneous electrical quantity characteristics into energy signals and associating them with grid nodes, establishes a cyclical analysis method for the interaction between characteristics and structure. Furthermore, by introducing energy signal conversion, abstract electrical quantity characteristics can be converted into input quantities that grid nodes can perceive, establishing a bridge between the two. Furthermore, tracing the chain transmission path can clearly show the transmission process of state changes, avoiding the omission of associations caused by local analysis. Ultimately, through multiple rounds of interactive cycles, the cumulative effect of subtle associations can be amplified, making previously hidden fault-related factors gradually apparent. Furthermore, real-time load fluctuations are incorporated as adjustment parameters in each cycle, ensuring that the analysis process consistently reflects the actual operating conditions of the power grid. Specifically, load fluctuations are the core variable influencing the relationship between electrical quantities and grid structure; ignoring their influence can lead to unrealistic correlation analysis. This multi-cycle approach, from features to structure to load linkage, ensures that the extracted correlation factors are closer to the actual causes of faults, resolving the distortion of correlation factors caused by static or single-dimensional analysis and providing a more reliable foundation for subsequent early warnings.

[0029] For ease of understanding, the specific implementation provided in this example is as follows: Assume that the target device is a 110kV substation outgoing line switch, and the factors associated with its short-circuit fault need to be analyzed. When a transient current peak (transient electrical characteristic) occurs at the switch, the system first converts the peak into a corresponding energy signal and maps it to the busbar node where the switch is located (S1). After receiving the signal, the busbar node undergoes an initial state change, indicating a voltage sag (S2). This change is transmitted along the outgoing line topology to the line tower node and the user transformer node, forming a chain transmission path. Each node provides real-time feedback on the attenuation effect of its own impedance change on the upstream voltage signal (S3). The system reversely superimposes these attenuation effects on the busbar node, completing the first round of interaction (S4). Starting from the state of the busbar node after this cycle, the above conversion, transmission, and feedback process are repeated (S5). After three cycles, the system calculates the cumulative deviation of each node state. For example, if the cumulative impedance deviation of the line tower node reaches 5%, the system also introduces real-time load fluctuation parameters (such as a 20% load surge during a certain period) to adjust the cycle weight in each cycle. Finally, the bus voltage sag amplitude, line impedance cumulative deviation, and node response delay during load mutation were extracted as correlated factors that can predict short-circuit faults. These factors are consistent with the actual logic that load mutations in actual power grid operation cause line impedance changes, thereby exacerbating the impact of switching current peaks.

[0030] In one embodiment, when determining how the various correlation factors influence each other: Based on the correspondence between the characteristics of electrical quantities and the nodes of the power grid structure, an initial causal relationship sequence is established according to the order in which the factors appear. It should be noted that the initial causal relationship sequence refers to a chain relationship model constructed according to the time order and logical association of the related factors, which can be generated through timestamp comparison and topological association algorithm.

[0031] Referring to the acquired power load mutation data, the effectiveness of the triggering of each link in the sequence is verified through real-time data, and invalid links are replaced and supplemented. During the supplementation process, relevant factors that may affect the subsequent model judgment and information exchange need to be included; among them, invalid links refer to causal nodes that cannot be verified by actual operation data in real-time data verification. Their judgment can be achieved by setting a trigger threshold (for example, after factor A changes, factor B does not produce the expected fluctuation within the preset time window).

[0032] Based on the extraction of correlation factors, the response delay differences and amplitude fluctuation ranges of other correlation factors under different grid load conditions after the change of a single correlation factor are recorded, and the possible impact of the recorded results on the subsequent deviation judgment link is predicted simultaneously; it should be understood that the response delay difference refers to the difference in the time intervals for corresponding changes in other correlation factors under different load conditions after a certain factor changes, which can be calculated using timestamp tags and time series analysis tools.

[0033] Based on these records, a dynamically updateable multi-scenario impact relationship table is formed, and its contents are continuously revised as new collected data is added. It should be understood that the multi-scenario impact relationship table is a dynamic data table that integrates the interaction patterns of various factors under different load levels and grid topologies. It can be stored in a relational database or a time series database, and supports dynamic updates based on the frequency of new data collection.

[0034] Exemplarily, the present invention adds a process of checking and dynamically correcting power load mutation data on the basis of establishing an initial causal relationship sequence.

[0035] It should be understood that constructing a sequence based solely on the order in which factors appear may not accurately reflect the true impact of these factors. However, by introducing real-time verification of power load mutation data, the system can determine whether there is an actual triggering relationship between each link in the sequence, thereby preliminarily verifying the rationality of the impact method.

[0036] Furthermore, by replacing and completing invalid links and recording response characteristics under different loads, the system can eliminate false or one-sided correlations, ensuring that the acquired mutual influence patterns conform to the actual operating logic of the power grid. Only after these steps are completed can a dynamically updated multi-scenario influence relationship table be finally formed. This combination of initial sequence construction, load data verification and correction, and dynamic updating makes the determination of the mutual influence patterns of related factors more accurate and realistic.

[0037] Specifically, the initial causal sequence is used to establish a basic correlation framework, while the verification and correction of power load mutation data serves as a key calibration method to ensure that the impact pattern is consistent with the actual operation of the power grid. This solves the problem of correlation distortion caused by relying solely on the order of factor occurrence, avoids subsequent early warning models from making misjudgments based on incorrect impact relationships, and improves the reliability of early warnings for intelligent operation and maintenance of the transmission network.

[0038] To facilitate understanding of the above technical solution, let's use an example to illustrate: Consider a 220kV transmission network where the correlated factors are line voltage fluctuation, transformer operating temperature, and regional power load fluctuation. An initial sequence is established in order of occurrence: regional power load fluctuation - line voltage fluctuation - transformer operating temperature. Verification of power load fluctuation data reveals that when the load fluctuation amplitude is less than 5%, the link from line voltage fluctuation to transformer operating temperature is ineffective. Therefore, line current fluctuation is added, adjusting the sequence to regional power load fluctuation - line voltage fluctuation - line current fluctuation - transformer operating temperature. The responses under different loads are recorded: for light loads (<30%), the voltage fluctuation delay is 1.2 seconds, the amplitude is ±2%, and the current change delay is 1.5 seconds, the amplitude is ±3%. For heavy loads (>70%), the voltage fluctuation delay is 0.8 seconds, the amplitude is ±5%, and the current change delay is 1.0 seconds, the amplitude is ±6%. Predicting these data will affect the load adaptability of the subsequent deviation judgment threshold. Finally, a multi-scenario impact relationship table is formed. After collecting new rainstorm weather data, features such as the voltage fluctuation amplitude increasing by ±1% under the same load mutation are added, and the content of the table is continuously improved.

[0039] In one embodiment, collecting target data includes: First, the filter threshold is dynamically adjusted based on the real-time grid load fluctuations. Then, an algorithm is used to distinguish interference information from valid features in the data. It should be noted that dynamically adjusting the filter threshold refers to changing the critical value used in data filtering in real time based on the amplitude and frequency of real-time grid load fluctuations. This is achieved by establishing a mapping model between load fluctuation amplitude and threshold. Interference information refers to signals in the data that are unrelated to the operating status of grid equipment and fault warnings, such as random noise from measurement equipment. This interference information can be identified by comparing it with baseline data during normal operation.

[0040] The filtered valid data is bound to key node parameters that reflect changes in the power grid structure and time-stamped. This allows the archived data to automatically activate the corresponding factor transmission path based on the tag when entering the correlation analysis phase. It should be noted that key node parameters refer to core parameters that reflect changes in the power grid structure, such as node connectivity and impedance values, and can be retrieved in real time from the power grid topology database. Time-stamping refers to the timestamp and corresponding power grid operation time information added to the valid data. High-precision clock synchronization technology is used to ensure the accuracy of the tagging.

[0041] With reference to the above, further, in another embodiment of the present application, when distinguishing interference information from valid features by a dynamic filtering algorithm: First, a reference signal generated by the current grid load fluctuation is captured. Then, a differential analysis is performed between the electrical quantity data and this reference signal to extract abnormal segments where the difference exceeds a dynamic generation threshold. It should be noted that the reference signal refers to the characteristic electrical quantity signal (such as the mean voltage and the effective current value) that reflects the current stable operating state of the grid load. This signal is generated by smoothing load data over five consecutive sampling cycles. Abnormal segments are continuous data segments where the difference between the electrical quantity data and the reference signal exceeds the dynamic threshold. These segments must last for at least two sampling cycles to eliminate transient noise.

[0042] For abnormal fragments, the source is traced in combination with the real-time trajectory of the grid structure changes. If it can be associated with the state change of a specific node, it is marked as a valid feature; otherwise, it is judged as interference information.

[0043] Exemplarily, this embodiment constructs a complete mechanism from data collection to effective feature extraction through the progressive processing of load fluctuation benchmark-dynamic threshold adjustment-difference analysis-traceability verification.

[0044] It should be understood that if only a fixed filtering threshold is used, it will be difficult to adapt to the dynamic changes of the grid load. When the load is light, the fixed threshold may misjudge minor anomalies as interference, and when the load is heavy, it may miss significant anomalies, resulting in the loss of valid data or the mixing of interference information.

[0045] This embodiment adjusts the threshold value based on real-time load fluctuations, allowing the filtering process to adapt synchronously with the grid operation status. For example, when the load fluctuation amplitude increases from 5% to 10%, the filtering threshold value is correspondingly increased from 2V to 4V, avoiding the disconnection between the threshold value and the actual working conditions.

[0046] Next, by introducing differential analysis of reference signals, abnormal data that deviates from normal operating conditions can be accurately identified. Combined with tracing the trajectory of grid structure changes, this approach addresses the core question of whether abnormal signals truly reflect the grid state. Further elaboration reveals that some anomalies may be caused by sensor failure or electromagnetic interference (such as electromagnetic pulses during thunderstorms), which cannot be distinguished simply by numerical differences. Their validity can only be confirmed by linking them to specific node state changes (such as switch opening and closing, tower tilt). Binding valid data to node parameters and time-stamping them establishes a connection from data to nodes and finally to paths for subsequent correlation analysis. This avoids repeated searches for the factor transmission paths corresponding to the data during correlation analysis and improves analysis efficiency.

[0047] Through the above process, the accuracy of effective feature extraction can be guaranteed, and precise guidance can be provided for the flow of data to the association analysis link, solving the problem of missing effective information or misjudgment of interference information caused by traditional fixed filtering, making the basic data entering the association analysis more reliable and more efficient.

[0048] To further illustrate this technical solution, let's use a specific example: Assume that line voltage data from a 35kV distribution network is collected and analyzed, and the real-time load fluctuation amplitude is 8%. The system first dynamically adjusts the filter threshold to 10% (8% × 1.25). Simultaneously, a voltage reference signal (e.g., 10kV ± 0.2kV) is captured when the load is operating stably. The real-time collected voltage data is compared with this reference signal. If the voltage drops to 9.5kV for three consecutive sampling periods during a certain period, and the difference of 0.5kV exceeds the dynamic threshold, this segment is identified as an abnormal segment.

[0049] Subsequently, combined with the traceability of the grid structure change trajectory, it was discovered that the tower in the middle section of the corresponding line during this period had tilted due to external force collision (a node state change). Therefore, this abnormal segment was marked as a valid feature. If the traceability did not detect any node state change, it was determined to be interference information (such as signal drift caused by loose sensor wiring). Finally, the valid data was associated with key parameters such as the tower ID "G35-012" and the time of the structural change "14:32:15", and then archived with a time stamp. When this data enters the correlation analysis phase, the system automatically activates the factor transfer path of "tower node line voltage load distribution" based on the tag, without the need to re-match the correlation relationship.

[0050] Furthermore, in one embodiment, calling the early warning model combination includes: First, convert the mutual influence strength of the associated factors into a dynamic adjustment coefficient. The higher the influence strength, the larger the coefficient value. It should be understood that the dynamic adjustment coefficient refers to a numerical parameter obtained by mapping the mutual influence strength between the associated factors and is used to adjust the model computing power allocation. Its value range can be set to 0-1. It should be understood that the intensity of the impact is positively and linearly correlated with the coefficient value.

[0051] Computing power weights are assigned to corresponding models based on their coefficient values. After each round of analysis, the coefficient values ​​are reverse-calibrated based on the newly generated impact relationship table, allowing computing power allocation to adapt to the correlation patterns of factors in real time. It should be understood that computing power weights refer to the proportion of total computing power resources allocated to each early warning model, with the sum being 1. This can be determined by dynamically adjusting the coefficients and converting them proportionally.

[0052] It should be understood that distributing computing power evenly across all models will result in insufficient resources for models corresponding to key factors, while models for non-critical factors will occupy too many resources. For example, in a fault scenario dominated by sudden load changes, the model responsible for load correlation analysis should be allocated more computing power; otherwise, insufficient computing power may lead to delayed warnings.

[0053] In this embodiment, the impact intensity is converted into an adjustment coefficient, which can tilt the computing power toward the model with more significant impact. For example, when the impact intensity of line current and temperature is 0.8, the computing power weight obtained by the corresponding model is higher than that of the environmental humidity and insulation performance model with an impact intensity of 0.3.

[0054] It should be further understood that the reverse calibration mechanism after each round of analysis solves the coefficient mismatch problem caused by changes in the factor correlation pattern. The intensity of the factors affecting the power grid operation will change with conditions such as load and weather. For example, the intensity of the impact of ambient temperature and equipment temperature will increase in hot weather. If the coefficient is not adjusted in time, it will cause the computing power allocation to lag behind actual demand. Calibration coefficients based on the newly generated influence relationship table can ensure that the adjustment coefficient is always synchronized with the current factor correlation pattern. For example, when the relationship table shows that the intensity of the impact of a sudden load change on voltage fluctuations increases from 0.6 to 0.9, the adjustment coefficient of the corresponding model is adjusted upward, and the computing power weight increases accordingly.

[0055] This approach, combining dynamic calibration with the conversion of impact intensity into adjustment coefficients and allocating computing power based on these coefficients, enables the early warning model combination to flexibly allocate computing resources based on real-time changes in factor relationships. This ensures the computational efficiency of key models while avoiding resource waste. This addresses the inadequate adaptability of fixed computing power allocation in scenarios with dynamically changing factor relationships, improving the real-time and accuracy of early warning analysis. For example, consider a 110kV transmission network early warning model combination involving a line overload warning model, an equipment temperature warning model, and a voltage stability warning model. Analysis shows that the mutual impact strength between line overload and equipment temperature is 0.8, and the mutual impact strength between line overload and voltage stability is 0.5. The system converts these strengths into dynamic adjustment coefficients of 0.8 and 0.5, respectively.

[0056] Furthermore, according to the aforementioned coefficient values, the computing power weights allocated to the line overload warning model and the equipment temperature warning model received a higher proportion of computing power, while the voltage stability warning model received a relatively lower proportion of computing power. After a round of analysis, the newly generated impact relationship table showed that the mutual influence strength between line overload and voltage stability had reached 0.7. Based on this, the system reversed the dynamic adjustment coefficient, adjusting the corresponding coefficient to 0.7, and then redistributed the computing power weights, giving the voltage stability warning model more computing power to adapt to the new factor correlation pattern.

[0057] Furthermore, in another embodiment of the present application, when the judgment information communicated between models includes factor feature matching results, trend change predictions and risk probability assessments, the following steps are included: It should be understood that the factor feature matching results refer to the degree of consistency judgment obtained by comparing the characteristics of the associated factors with the preset feature library by different models, which can be calculated by the similarity of feature vectors. Trend change prediction refers to the model's speculation on the development direction of the associated factors in the future based on the real-time values ​​and historical change patterns of the associated factors, which can be achieved using a time series prediction algorithm. Risk probability assessment refers to the model's estimation of the probability of possible failure risks based on the current status of the associated factors and their mutual influence, usually presented as a percentage.

[0058] When the real-time numerical variation of the correlation factors exceeds the set ratio, the relevant models will synchronously send the updated judgment information, and the models receiving the information will cross-check with their own analysis results; the contents with consistent verification results will be used as the basis for comprehensive judgment, and those with inconsistent results will be subject to secondary analysis. It should be understood that the set ratio refers to the pre-set threshold value of the correlation factor numerical variation amplitude used to trigger the synchronous transmission of information between models, which can be adjusted according to different operating scenarios of the power grid and the importance of the correlation factors. Secondary analysis refers to the process of re-analyzing and verifying the differences by introducing more relevant data or using more sophisticated algorithms when the cross-check results between models are inconsistent.

[0059] This embodiment ensures the accuracy and reliability of the early warning model combination analysis by clarifying the judgment information types and information interaction and verification methods communicated between models.

[0060] It should be understood that when the real-time numerical fluctuation of the associated factors exceeds the set ratio, the updated judgment information is synchronously transmitted to ensure that each model has the latest analysis basis in a timely manner, avoiding judgments based on outdated information. The receiving model cross-checks the information with its own analysis results. The consistent results are used as the basis for the comprehensive judgment, which can enhance the credibility of the judgment results. In the event of inconsistency, a secondary analysis is carried out to help identify the cause of the discrepancy and reduce misjudgments.

[0061] This inter-model information exchange and verification method can resolve the problem of analytical result deviations caused by the inability of individual models to work together, enabling the early warning model combination to fully leverage its synergy and improve the accuracy of collaborative analysis and early warning of multi-dimensional data in intelligent transmission network operation and maintenance. Simply put, assume that the early warning model combination for a 110kV transmission network includes a line status analysis model, a load forecasting model, and a fault diagnosis model. When the real-time value fluctuation of the associated factor line current exceeds a set ratio of 10%, the line status analysis model will simultaneously send its factor feature matching results (85% matching with the line overload feature), trend change prediction (likely to increase by 15% in the next 5 minutes), and risk probability assessment (60% probability of line overload failure) to the load forecasting model and fault diagnosis model.

[0062] After receiving this information, the load forecasting model cross-checked it with its own load trend analysis and found that their assessments of line current trends were consistent. This information served as the basis for a comprehensive assessment. The fault diagnosis model then checked this information against its own analysis of line insulation status and found a discrepancy in its assessment of the fault risk probability (its own assessment was 40%). Therefore, a secondary analysis was initiated, incorporating more relevant data such as line temperature and humidity for a recalculation, ultimately resulting in a more accurate risk probability assessment.

[0063] In one embodiment, when the deviation exceeds a set value, the method further includes: Focusing on the coordinated response of data collection frequency and load control parameters, the judgment results of each model are summarized to obtain a consistent conclusion on the cause of the deviation. It should be noted that the data collection frequency refers to the number of times relevant data of the transmission network is collected per unit time. Its high or low level directly affects the real-time and integrity of the data and can be dynamically set according to actual needs. It's important to reiterate that load control parameters refer to the various parameters used to adjust the distribution and size of grid loads, such as transformer tap positions and reactive power compensation device switching capacity. A coordinated response approach involves the interconnected and coordinated changes between data collection frequency and load control parameters, ensuring they work together to adapt to grid deviations.

[0064] It should be further explained that consistent conclusions refer to the same or similar judgment results reached by each model after analyzing the causes of deviation, which is an important basis for subsequent adjustments and decision-making.

[0065] According to the conclusion, the relationship between data acquisition frequency and load control parameters is adjusted. The greater the load control range, the higher the acquisition frequency. Combining the adjusted coordination relationship with the results of the model's supplementary analysis, a real-time early warning plan is generated, including the risk level, impact scope, and control recommendations. It should be noted that the risk level refers to the classification of potential risks based on factors such as the severity and development trend of the deviation, and is generally classified as minor, general, severe, and urgent. The impact scope refers to the area or equipment range where the deviation may affect the transmission network. Control recommendations are specific operational suggestions for resolving deviations and reducing risks, based on the analysis and judgment of the deviation.

[0066] For example, when dealing with deviations, if only one parameter is adjusted, the deviation may not be properly handled due to inaccurate understanding of the grid status; Therefore, by aggregating the results of each model to reach a consistent conclusion on the cause of the deviation, we can avoid the one-sidedness of individual model analysis and provide a reliable basis for subsequent adjustments. Based on this conclusion, we adjust the coordination between the two models, increasing the data collection frequency when the load regulation amplitude is large. This can more accurately capture the changes in the power grid during large-scale regulation, providing more sufficient data for analysis and decision-making.

[0067] Finally, combining the adjusted coordination relationship with the model's supplementary analysis results, a real-time early warning plan is generated. This plan better aligns with grid realities, improving the accuracy and practicality of early warnings. This approach differs from traditional solutions, which can be ineffective due to poor parameter coordination or incomplete analysis. Specifically, consider a 220kV transmission network where voltage deviation exceeds the set value by 5%. Focusing on the coordinated response of data acquisition frequency and load control parameters, the system summarizes the results of the line state analysis model, load forecasting model, and voltage stability model, concluding that the deviation is caused by a sudden increase in load in a specific area.

[0068] Based on the above conclusions, the system adjusted the coordination between the data collection frequency and the load control parameters. Since the load control amplitude of this area needed to be set to 20%, which was a relatively large amplitude control, the data collection frequency was increased from 2 times per second to 5 times per second.

[0069] Afterwards, the system combines the adjusted coordination relationship and the results of the supplementary analysis of each model to generate a real-time early warning plan, in which the risk level is judged to be general, the impact range is the power grid equipment in the area and the surrounding area within 10km, and the control recommendation is to appropriately invest in reactive power compensation devices to stabilize the voltage.

[0070] Furthermore, in one embodiment, when a sudden load shock occurs in the power grid: The conventional factor correlation analysis process is suspended to extract the sudden change characteristics of electrical quantities before the shock occurs; the above characteristics are quickly compared with the typical characteristics in the historical shock case library, and a reference plan is generated after sorting by the degree of matching; at the same time, each model calculates the tolerance capacity of the current power grid structure in parallel, and combines the calculation results with the reference plan to form a temporary control strategy; after the shock is alleviated, conventional analysis is carried out and the temporary strategy is calibrated.

[0071] It should be noted that methods for extracting electrical quantity mutation characteristics before an impact occur, such as those based on the singular value decomposition of the Hankel matrix, use this algorithm to detect mutation points in the recorded signal and compare the absolute values ​​of the elements in the component signal with a set threshold value to determine the moment of the electrical quantity mutation. There are also methods based on wavelet transforms, which use multi-scale analysis to extract transient characteristics of the current waveform, such as spikes and mutation points. Furthermore, there are also methods that use catastrophe theory to extract multiple characteristic quantities that characterize arc current mutations, such as time-domain integral value characteristics and difference-root mean square characteristics. These characteristics are then evaluated using catastrophe theory indicators to comprehensively calculate the characteristic quantities and determine the mutation membership value of the line current.

[0072] Sudden load shocks refer to a large, rapid increase or decrease in grid load over a short period of time, potentially caused by unexpected factors such as extreme weather or the start-up or shutdown of large equipment. Sudden electrical quantity changes refer to abnormal changes in electrical quantities such as voltage, current, and power, such as sudden voltage drops or current surges, that occur before a load shock occurs.

[0073] Among them, the historical impact case library is built using a relational database (such as MySQL), in which each case contains the time of impact occurrence, load change amplitude, electrical quantity characteristic parameters corresponding to the extracted features, grid topology, treatment measures and implementation effects.

[0074] The grid structure tolerance refers to the maximum pressure that the current grid equipment, lines and other structures can withstand under load impact, reflecting the grid's ability to resist impact.

[0075] Among them, the temporary control strategy can be generated in JSON format but is not limited to it, and includes feedback verification indicators such as the control object ID, operation amount, execution timing accurate to milliseconds, and the allowable range of voltage fluctuation.

[0076] It should be further explained that parallel computing is implemented based on a distributed computing architecture (such as the Apache Spark framework), which means that multiple models calculate the tolerance capacity of different structural parts of the power grid at the same time.

[0077] Furthermore, when the power grid encounters a sudden load shock, by suspending the normal process and prioritizing the extraction of the sudden change characteristics of the electrical quantity before the shock occurs, the key to the problem can be quickly identified.

[0078] Furthermore, the extracted features are quickly compared with typical features in a historical impact case database, and reference solutions are generated based on the degree of match. This is equivalent to drawing on past successful experiences, allowing targeted response methods to be formed in a short period of time. Because historical cases contain treatment methods for various impact scenarios, it can speed up the development of solutions.

[0079] It should be understood that each model simultaneously calculates the current grid structure's withstand capacity. Specifically, independent computation threads are assigned to each model. The line withstand capacity model focuses on calculating the maximum current and temperature tolerance of each line segment under the current load shock; the transformer withstand capacity model focuses on the load factor upper limit and insulation tolerance of each transformer; and the switchgear model calculates the switching capacity and withstand voltage of various switches. These models each utilize corresponding data modules, such as the line parameter library, transformer nameplate data, and switchgear performance parameters, and perform their calculations simultaneously without interfering with each other. This allows for a comprehensive understanding of the withstand capacity of each key grid structure within a short period of time, accurately assessing the grid's current withstand capacity and preventing a disconnect between the reference solution and actual conditions. Combining the calculated results with the reference solution to form a temporary control strategy draws on historical experience while aligning with current grid conditions, making the strategy more feasible and secure.

[0080] It should be noted that the execution of the temporary control strategy adopts a closed-loop control mode: after the control instruction is issued, the change curve of the target parameters such as the node voltage is monitored in real time, and the deviation is compared with the expected curve preset by the strategy. If it exceeds the range, the secondary fine-tuning is immediately triggered.

[0081] When the impact is alleviated, the system will start the regular analysis module, trace the root cause of the impact using the cause-and-effect diagram analysis method, and correct the parameters of the temporary strategy based on the actual data. The corrected values ​​will be automatically updated to the historical case library, forming a closed loop of strategy optimization.

[0082] To better illustrate this solution: Assume that a 110kV power grid experiences a sudden load surge due to the sudden startup of high-power equipment in a nearby factory, with the load increasing by 30% within 10 seconds.

[0083] The system immediately suspended the regular factor correlation analysis process and extracted the sudden change characteristics of electrical quantities before the impact. It was found that the voltage dropped from 110kV to 102kV within 2 seconds before the impact, and the current suddenly increased from 500A to 680A.

[0084] By comparing these characteristics with the historical impact case library, it was found that the match degree with the load impact characteristics caused by the startup of equipment in a factory three years ago was 85%. Based on this, a reference plan was generated, including temporarily reducing non-critical loads in the surrounding area and adjusting transformer taps.

[0085] Simultaneously, parallel computations were initiated for the line withstand model and transformer load model. The line withstand model, using parameters such as the material, cross-sectional area, and installation environment of each section of the 110kV power grid, calculated the maximum current the line could withstand to be 700A. The transformer load model, based on the rated capacity, operating age, and cooling method of each transformer in the area, concluded that the maximum transformer load factor was 90%. Combined with the reference plan, a temporary control strategy was developed: reducing non-critical loads in the surrounding area by 10% and adjusting transformer taps to appropriate positions.

[0086] After 15 minutes, the load shock eased and the system conducted routine analysis. It was found that the temporary strategy effectively stabilized the power grid, but it could be further optimized. Therefore, the temporary strategy was calibrated, such as adjusting the proportion of non-critical loads to 8%, to make the strategy more reasonable.

[0087] The above is only a specific implementation method of the present application, but the scope of protection of the present application is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed in this application, which should be covered by the scope of protection of the present application.

Claims

1. A collaborative analysis and early warning method based on multi-dimensional data of intelligent operation and maintenance of power transmission network, characterized by: include: Collect target data from target devices and obtain sudden changes in power load in the system; Combined with the sudden change data of power load, the instantaneous characteristics of electrical quantities and the dynamic changes of the grid structure are analyzed to obtain the correlation factors that can predict faults; According to the sudden change amplitude of power load, determine the mutual influence mode between various related factors; According to the real-time values ​​and impact modes of related factors, the corresponding early warning model combination is called, and the models communicate with each other in a timely manner as the factors change; When the deviation exceeds the set value, the linkage response of data collection frequency and load control parameters is taken as the core, combined with the judgment results of each model, the adaptation relationship between the two is dynamically adjusted to generate a real-time early warning plan.

2. The method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission network according to claim 1 is characterized in that: Correlation analysis of the instantaneous characteristics of electrical quantities and the dynamic changes of the grid structure includes the following steps: S1: Convert the instantaneous characteristics of electrical quantities into corresponding energy signals, and then map them to relevant nodes in the power grid structure; S2: After receiving the signal, the node generates an initial state change; S3: Tracks the chain transmission path formed by this state change as the grid structure changes. Each node in the path will provide real-time feedback on the impact of its own state change on the upstream signal. S4: Add the effect back to the original corresponding node to form the first round of interactive cycle between features and structures; S5: taking the result of the cycle as a new starting point, repeating the above steps S1 to S4; S6: Obtain the cumulative deviation of node states in multiple cycles and extract related factors. In each cycle, the real-time load fluctuation of the power grid is introduced as a dynamic adjustment parameter to make the factor extraction process consistent with the actual situation of power grid operation.

3. The method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission network according to claim 2 is characterized in that: When determining how the various correlates influence each other: Based on the correspondence between electrical quantity characteristics and grid structure nodes, the initial causal relationship sequence is established according to the order in which factors appear; Refer to the acquired power load mutation data and verify the effectiveness of the triggering of each link in the sequence through real-time data. Replace and complete invalid links. The completion process needs to include relevant factors that may affect the subsequent model judgment information exchange; Based on the extraction of the correlation factors, the response delay differences and amplitude fluctuation ranges of other correlation factors under different grid load conditions after the change of a single correlation factor are recorded, and the possible impact of the recorded results on the subsequent deviation judgment link is simultaneously predicted; Based on the above records, a dynamically updateable multi-scenario impact relationship table is formed, and the content in the table is continuously revised as new collected data is added.

4. The method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission network according to claim 3 is characterized in that: When collecting target data, include: First, the filter threshold is dynamically adjusted based on the real-time load fluctuation of the power grid, and then the algorithm is used to distinguish the interference information from the effective features in the data; The filtered valid data is bound to the key node parameters of the grid structure changes and time-series marked, so that when the archived data enters the association analysis link, the corresponding factor transmission path can be automatically activated according to the mark.

5. The method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission network according to claim 4 is characterized in that: When distinguishing interference information from valid features through dynamic filtering algorithms: First, a reference signal generated by the current load fluctuation of the power grid is captured. Then, a differential analysis is performed between the electrical quantity data and the reference signal to extract abnormal segments where the difference exceeds the dynamically generated threshold. For abnormal fragments, the source is traced in combination with the real-time trajectory of the grid structure changes. If it can be associated with the state change of a specific node, it is marked as a valid feature; otherwise, it is judged as interference information.

6. The method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission network according to claim 1 is characterized in that: When calling the early warning model combination, it includes: First, convert the mutual influence intensity of the correlation factors into a dynamic adjustment coefficient. The higher the influence intensity, the larger the coefficient value. The computing power weight is assigned to the corresponding model according to the coefficient value, and after each round of analysis, the coefficient value is reverse-calibrated according to the newly generated influence relationship table, so that the computing power allocation can adapt to the factor association rules in real time.

7. The method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission network according to claim 1 is characterized in that: The judgment information exchanged between models includes factor feature matching results, trend change predictions, and risk probability assessments: When the real-time numerical variation of the associated factors exceeds the set ratio, the relevant models will synchronously send updated judgment information, and the receiving models will cross-check with their own analysis results; The consistent contents of the verification results will be used as the basis for comprehensive judgment, and the inconsistent contents will be subject to secondary analysis.

8. The method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission network according to claim 7 is characterized in that: When the deviation exceeds the set value, it also includes: Focusing on the linkage response of data collection frequency and load control parameters, the judgment results of each model are summarized to obtain consistent conclusions on the causes of deviations. According to the conclusion, the relationship between data acquisition frequency and load control parameters is adjusted. The greater the load control range, the higher the acquisition frequency. Combining the adjusted coordination relationship and the results of the model supplementary analysis, a real-time early warning plan is generated that includes risk level, impact scope and regulatory recommendations.

9. The method for collaborative analysis and early warning of multi-dimensional data based on intelligent operation and maintenance of power transmission network according to claim 1 is characterized in that: When a sudden load shock occurs to the power grid, it also includes: Suspend the conventional factor correlation analysis process and extract the electrical quantity mutation characteristics before the shock occurs; The above characteristics are quickly compared with the typical characteristics in the historical impact case library, and the reference plan is generated after sorting by matching degree; At the same time, each model calculates the tolerance of the current power grid structure in parallel, and combines the calculation results with the reference scheme to form a temporary control strategy; Once the shock is alleviated, regular analysis will be conducted and temporary strategies will be calibrated.

Images