A multi-level management method for real-time data of new energy station power generation equipment

Abstract

The application discloses a kind of multi-level management methods for real-time data of new energy station power generation equipment, and specifically relates to the field of electric power data management, including according to global data management strategy, edge management rules and central processing logic are generated cooperatively, and edge rules are issued to edge node;At the same time, by monitoring data quality in real time, and the results are fed back for dynamic adjustment of global data management strategy, realize the closed-loop optimization of data management.A kind of multi-level management methods for real-time data of new energy station power generation equipment generates edge management rule set and central processing logic set by global data management strategy, and dynamically adjusts strategy based on the third management data, effectively solves the problem that edge node processing rule and central data processing logic lack cooperation and dynamic adaptation;By implementing cross-level timestamp synchronization calibration, the secondary alignment processing of center node to multi-source heterogeneous data is avoided, and the equipment state evaluation accuracy is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric power data governance, and more particularly, to a multi-level governance method for real-time data of power generation equipment in a new energy station. BACKGROUND

[0002] With the rapid development of new energy power generation technology and the continuous expansion of the station scale, efficient governance of real-time data of new energy station power generation equipment is particularly important. Traditional new energy station data processing collects raw data generated by various devices in the new energy station through data acquisition terminals, and then transmits the raw data to a central data center for unified storage, cleaning, calculation and analysis. However, the direct transmission of massive raw data causes a huge communication network bandwidth pressure. The central data processing center needs to process raw data from numerous stations, which are heterogeneous, redundant and even include noise, resulting in excessive computing load.

[0003] To alleviate the network congestion and central computing pressure in the traditional new energy station data processing technology, an edge computing node is introduced as an improvement scheme. An intelligent gateway with local computing capability is deployed at the substation level to pre-screen and compress the raw data, reduce the data throughput of the central node, reduce the bandwidth demand of the backbone communication network, and share part of the computing load of the central server.

[0004] However, in actual use, there are still some shortcomings, such as the lack of effective coordination and dynamic adaptation mechanism between the processing rules of the edge node and the central data processing logic, which makes it difficult to guarantee the global optimality and consistency of the data quality after multi-level processing. The central node still needs to perform secondary alignment processing on multi-source heterogeneous data, and the timestamp synchronization deviation of cross-level data directly affects the equipment state evaluation accuracy. The existing architecture lacks a closed-loop feedback mechanism for the data governance process, and manual intervention is still required for abnormal data tracing, resulting in long operation and maintenance response time. SUMMARY

[0005] In order to overcome the above-mentioned defects of the prior art, the present application provides a multi-level governance method for real-time data of power generation equipment in a new energy station, which solves the problems in the background art by the following scheme.

[0006] To achieve the above-mentioned purpose, the present application provides the following technical scheme:

[0007] A multi-level governance method for real-time data of power generation equipment in a new energy station, comprising:

[0008] S1: generating a set of edge governance rules and a set of central processing logic that are mutually coordinated according to a preset global data governance strategy, and sending the set of edge governance rules to the corresponding new energy station edge node;

[0009] S2: The edge node of the new energy power station receives the edge governance rule set and performs a first governance operation on the raw data acquired in real time by the power generation equipment. The first governance operation includes at least data cleaning and format standardization to generate first governance data.

[0010] S3: Based on the central processing logic set, perform a second governance operation on the first governance data received from the edge node. The second governance operation includes cross-level timestamp synchronization calibration and data fusion processing to generate the second governance data.

[0011] S4: Monitor the data quality indicators during the first and second governance operations in real time, compare the monitoring results with preset quality thresholds, and generate third governance data including data quality assessment results and abnormal data tracing information;

[0012] S5: Based on the third governance data, dynamically adjust the global data governance strategy and update the edge governance rule set and the central processing logic set accordingly. Based on the abnormal data tracing information, automatically trigger the preset abnormal data handling process.

[0013] Preferably, in step S1, the preset global data governance strategy includes a device type hierarchical mapping table, a data quality target matrix, and network status awareness rules.

[0014] Preferably, S2, the first treatment operation, specifically includes:

[0015] Executed through a first data processing channel that includes at least a data cleaning channel and a format standardization channel;

[0016] The data cleaning channel, based on the device-level data cleaning rule library in the edge governance rule set, performs differentiated processing on data of different types of power generation equipment. The differentiated processing algorithm includes at least one of the following: a sliding window-based filtering algorithm, a wavelet denoising coefficient dynamic adjustment method, and a dynamic baseline correction algorithm under equipment start-up and shutdown conditions.

[0017] Furthermore, the format standardization channel converts the data from the data cleaning channel into a predefined standard format based on the format conversion template in the edge governance rule set, and encapsulates it into a unified encoding structure containing device ID, high-precision timestamp, data value range and quality flag bits.

[0018] Preferably, the second treatment operation, S3, specifically includes:

[0019] This is performed through a second data processing channel that includes at least a timing alignment calibration channel and a multi-source data fusion channel.

[0020] The timing alignment calibration channel, based on the cross-level timing alignment rules in the central processing logic set, constructs a transmission delay topology map for the encoding structure of the first governance data and uses a hybrid clock synchronization mechanism to achieve data alignment. The hybrid clock synchronization mechanism includes at least one of NTP protocol, PTP protocol, and Kalman filter algorithm.

[0021] Furthermore, the multi-source data fusion channel performs multi-source data fusion processing on the data from the time-series alignment calibration channel according to the multi-source data fusion algorithm chain in the central processing logic set, and encapsulates it into second governance data containing quality assessment labels, processing metadata and semantic association information.

[0022] Preferably, in S3, the metadata includes at least a clock compensation coefficient, a redundancy elimination identifier, and a knowledge graph reasoning path.

[0023] Preferably, S3, multi-source data fusion processing, specifically includes:

[0024] A spatiotemporal correlation weighted model is constructed based on equipment topology correlation, where the correlation weights are calculated using an exponential decay function of electrical distance, specifically expressed as follows:

[0025]

[0026] Where δ represents the slope factor, d max Let d represent the maximum electrical distance between power generation equipment i and power generation equipment j. ij This is expressed as the electrical distance between power generation equipment i and power generation equipment j;

[0027] Clustering algorithms are used to eliminate redundant data;

[0028] Construct a triplet knowledge graph containing equipment entities, operating status, and power grid commands, apply graph neural networks for semantic association analysis, and perform consistency verification of equipment operation logic.

[0029] Preferably, in step S4, the abnormal data tracing information includes at least the power generation equipment ID, data flow path, and processing timeline information.

[0030] Preferably, S5, the preset abnormal data handling process, specifically includes:

[0031] For known abnormal data identified based on the abnormal data tracing information, the corresponding handling plan is retrieved from the preset system database to perform automated repair.

[0032] For new types of abnormal data identified based on the abnormal data source tracing information, a preset general handling plan is executed, the impact of the general handling plan is evaluated, and the plan with the best repair effect is selected for execution.

[0033] Furthermore, in cases where the abnormal data source information indicates a hardware-level fault, an equipment maintenance work order containing location information is automatically generated and synchronized to the maintenance personnel.

[0034] The technical effects and advantages of this invention are as follows:

[0035] 1. This invention generates an edge governance rule set and a central processing logic set through a global data governance strategy, and dynamically adjusts the strategy based on third-party governance data. This effectively solves the problem of lack of coordination and dynamic adaptation between edge node processing rules and central data processing logic, ensuring that the data quality after multi-level processing reaches global optimality and consistency, and improving the overall coordination of data governance.

[0036] 2. By implementing cross-level timestamp synchronization calibration, this invention significantly reduces the timestamp synchronization deviation of cross-level data, avoids secondary alignment processing of multi-source heterogeneous data by the central node, and effectively improves the accuracy of equipment status assessment.

[0037] 3. This invention generates third-party governance data containing abnormal data tracing information, and combines it with a preset abnormal data handling process to achieve automatic repair and strategy adjustment, which greatly shortens the operation and maintenance response time and improves the automation level of data governance. Attached Figure Description

[0038] Figure 1 This is a flowchart illustrating the steps of a multi-level management method for real-time data of power generation equipment in a new energy power plant, according to an embodiment of this application. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0041] Hereinafter, the terms "first," "second," and "third" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first," "second," and "third" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0042] As attached Figure 1 The method for multi-level governance of real-time data from power generation equipment in new energy power plants includes: collaboratively generating edge governance rules and central processing logic based on a global data governance strategy, and distributing the edge rules to edge nodes; simultaneously, by monitoring data quality in real time and feeding the results back to dynamically adjust the global data governance strategy, a closed-loop optimization of data governance is achieved. The specific steps are as follows:

[0043] S1: Based on the preset global data governance strategy, generate a set of mutually cooperating edge governance rules and a set of central processing logic, and send the set of edge governance rules to the corresponding edge nodes of the new energy power station.

[0044] Specifically, in S1, a collaborative governance architecture between the edge and the center is achieved through a dynamic rule generation system driven by a pre-built global data governance strategy.

[0045] In one possible implementation, a preset global data governance strategy is dynamically generated by a cloud-based strategy engine, including a device type hierarchical mapping table, a data quality target matrix, and network status awareness rules. The cloud-based strategy engine generates the global data governance strategy based on the topology of the target renewable energy power plant, historical operating data of the power generation equipment, and real-time network bandwidth status using a multi-objective optimization algorithm. Specifically, this includes defining data sampling frequency and accuracy requirements according to the type of power generation equipment, generating a device type hierarchical mapping table, where the power generation equipment types include, but are not limited to, wind turbines and photovoltaic inverters. In this embodiment, the vibration data of doubly-fed asynchronous wind turbines uses a 0.5Hz-80Hz bandpass filter threshold, and the current data of photovoltaic inverters is configured with a dynamic baseline adjustment outlier detection method. The data quality target matrix uses orthogonal experimental design to determine the dynamic weight allocation rules for integrity, consistency, and timeliness. In this embodiment, when the network latency exceeds 50ms, the timeliness weight is automatically increased to 0.6. The network status awareness rules collect the bandwidth utilization and transmission jitter values ​​of each edge node in real time and set the edge node data cache window size based on the network transmission jitter value threshold.

[0046] Furthermore, the global data governance strategy is decomposed into an edge governance rule set and a central processing logic set through a decoupling algorithm. The edge rule set includes a device-level data cleaning rule base, format conversion templates, and local caching strategies, while the central processing logic set constructs cross-level time-series alignment rules and a multi-source data fusion algorithm chain.

[0047] It should be noted that the device-level data cleaning rule library in the edge rule set includes, but is not limited to, outlier removal algorithms based on sliding windows, dynamic adjustment methods for wavelet denoising coefficients, and dynamic baseline correction algorithms under device start-up and shutdown conditions; the format conversion templates include, but are not limited to, mapping rules from multiple protocols such as Modbus and IEC 61850 to standard JSON Schema, and definitions including device ID, high-precision timestamps, data value ranges, and quality flag bits in TLV encoding structures; the local caching strategies include, but are not limited to, hierarchical storage mechanisms based on data timeliness, priority-weighted queue management algorithms, and dynamic cache window adjustment rules triggered by network congestion; the cross-level timing alignment rules in the central processing logic set adopt a hybrid NTP and PTP clock synchronization mechanism, which embeds hierarchical transmission path markers in data packets, including node ID, transmission hop count, and estimated delay, and applies Kalman filtering algorithms to compensate for network jitter, achieving clock deviation control within ±10ms across three levels of nodes; the multi-source data fusion algorithm chain includes, but is not limited to, spatiotemporal correlation weighting algorithms based on device topology correlation, redundant data elimination algorithms, and cross-regional semantic association analysis algorithms based on knowledge graphs.

[0048] This embodiment accesses the digital twin model of the target renewable energy power station, analyzes the device connection relationships in the topology, and associates device metadata, including the rated power of the doubly-fed asynchronous wind turbine and the MPPT curve characteristics of the photovoltaic inverter. Simultaneously, based on the raw data stream with a sampling period of 10ms stored in the historical operation database for the past 6 months, it extracts device type feature patterns. Through probe programs deployed on edge nodes, it collects the round-trip latency and jitter values ​​of each communication link in real time. When the network jitter value of the edge node is detected to exceed 15ms for 5 consecutive sampling periods, the attenuation factor β in the data cache window calculation formula is dynamically adjusted from 0.8 to 0.5, where the data cache window W... dc The calculation formula is as follows:

[0049] W dc =α×(1―RTT / 100)+β×J

[0050] Where α and β represent the attenuation factors of round-trip time delay and jitter value, respectively, RTT represents the round-trip time delay, and J represents the jitter value; the decoupling algorithm adopts a hierarchical decision tree algorithm, which splits the global data governance strategy into wind turbine-specific rule branches and photovoltaic-specific rule branches according to the equipment type hierarchical mapping table. For the wind turbine branch, the vibration data processing rule set is extracted, which includes three-level wavelet denoising and peak detection based on a 3-second sliding window; for the photovoltaic branch, a dynamic baseline model of current data is generated.

[0051] S2: The edge node of the new energy power station receives the edge governance rule set and performs a first governance operation on the raw data acquired in real time by the power generation equipment. The first governance operation includes at least data cleaning and format standardization to generate first governance data.

[0052] Specifically, in S2, the edge node of the new energy power station receives the edge governance rule set sent from the cloud through a preset encrypted communication channel. At the same time, it obtains the raw data stream from each power generation device in the new energy power station in real time through the data acquisition interface. After receiving the edge governance rule set and obtaining the raw data, the edge node processes the raw data according to the edge governance rule set and performs the first governance operation to obtain the first governance data. The first governance data is the data after the edge node of the new energy power station performs the first governance operation on the raw data.

[0053] In one possible implementation, the first governance operation is performed through a first data processing channel that includes at least a data cleaning channel and a format standardization channel;

[0054] It should be noted that in the data cleaning channel, the edge nodes retrieve the device-level data cleaning rule library from the edge rule set to perform differentiated processing on the noise, outliers and drift problems in the original data for different types of power generation equipment.

[0055] In this embodiment, a sliding window-based filtering algorithm is applied to the wind turbine vibration data. The window size is dynamically adjusted between 20 and 50 sampling points based on the dynamic characteristics of the data, with a fixed step size of 5 sampling points. This is combined with a wavelet denoising coefficient dynamic adjustment method, using the db4 wavelet basis. The decomposition level is automatically selected between 3 and 5 levels based on the real-time signal-to-noise ratio. The denoising threshold DT is calculated using a formula, specifically expressed as follows:

[0056]

[0057] Where σ represents the estimated noise standard deviation, and N is the signal sequence length;

[0058] For the current data of the photovoltaic inverter, the dynamic baseline correction algorithm under the equipment start-up and shutdown conditions is enabled. The baseline adjustment period is adaptively set according to the power change rate. The period is no more than 1 second under the fast change condition and no more than 10 seconds under the steady state condition. At the same time, outlier detection is performed by combining the threshold dynamically floating according to ±3σ standard deviation.

[0059] In the format standardization channel, edge nodes convert data that has passed through the data cleaning channel into a predefined standard format based on the format conversion template in the edge governance rule set, and encapsulate it into a unified encoding structure that includes device ID, high-precision timestamp, data value range and quality flag bits;

[0060] In this embodiment, for Modbus RTU protocol data, register addresses are mapped to semantically meaningful tags, the data sampling rate is set to 10Hz, and the values ​​are normalized using a coefficient of 0.001; for IEC 61850MMS messages, they are parsed into a device status tree structure, the timestamp accuracy reaches 1 millisecond, and a 4-bit quality flag is encoded; for CAN bus data, the frame ID is parsed into a device operation instruction set, achieving a data compression rate of no less than 70%, while retaining the original check bits; all standardized data is encapsulated into a unified TLV encoding structure.

[0061] In addition, under certain circumstances, when the network connection between the edge node and the central layer is unstable or when data needs to be temporarily stored according to the preset edge governance rule set, the data processing pipeline will also include a local cache channel. In the local cache channel, the edge node implements hierarchical storage management according to the local cache strategy in the edge governance rule set.

[0062] In this embodiment, the real-time data buffer adopts a ring buffer structure. The dynamic window adjustment monitors the network round-trip latency with the central layer in real time with a period of 200 milliseconds. Once the round-trip latency value is detected to exceed 50 milliseconds, the buffer window size for data transmission is automatically reduced by 20% to 40% to adapt to network congestion.

[0063] S3: Based on the central processing logic set, perform a second governance operation on the first governance data received from the edge node. The second governance operation includes cross-level timestamp synchronization calibration and data fusion processing to generate the second governance data.

[0064] Specifically, in S3, the central node performs a second governance operation on the first governance data transmitted by each edge node based on the central processing logic set.

[0065] In one possible implementation, the second governance operation is performed via a second data processing channel that includes at least a timing alignment calibration channel and a multi-source data fusion channel;

[0066] It should be noted that in the timing alignment calibration channel, the coding structure in the first governance data is parsed according to the cross-level timing alignment rules in the central processing logic set, a transmission delay topology map is constructed, and a hybrid clock synchronization mechanism is used to achieve high-precision data alignment. The transmission delay topology map is a representation of the combination of delay data and transmission path. The hybrid clock synchronization mechanism includes at least one of NTP protocol, PTP protocol, and Kalman filter algorithm.

[0067] In this embodiment, the NTP protocol is used to achieve preliminary clock synchronization at the millisecond level, ensuring that the error range is controlled within ±50ms. The PTP protocol is introduced to record the time of data entry and exit at edge nodes, regional nodes, and central processing nodes through hardware timestamps, achieving microsecond-level precise clock synchronization. Based on the transmission delay topology map, a state-space model of the Kalman filter algorithm is constructed to achieve timing compensation. The compensation coefficient is dynamically adjusted according to the number of transmission hops, and the network jitter variance is statistically analyzed using a sliding window to dynamically update the parameters of the Kalman filter, thereby controlling the data timestamp deviation after alignment within ±10ms.

[0068] Furthermore, the state-space model constructed based on the Kalman filter algorithm is specifically represented as follows:

[0069]

[0070] in, Let u be the state vector at time k, including the time delay information at time k. Let A be the clock drift coefficient matrix, B be the network delay weight matrix, and u be the time delay weight matrix. k Represented as network disturbance, w k This is represented as Gaussian white noise.

[0071] In the multi-source data fusion channel, based on the multi-source data fusion algorithm chain in the central processing logic, the data from the time-series alignment calibration channel is processed by multi-source data fusion and encapsulated into second governance data containing quality assessment labels, processing metadata and semantic association information. The metadata includes at least clock compensation coefficients, redundancy elimination identifiers and knowledge graph reasoning paths.

[0072] Furthermore, the multi-source data fusion processing includes: constructing a spatiotemporal correlation weighted model based on equipment topology correlation, where the correlation weight is calculated using an exponential decay function of electrical distance; implementing redundant data elimination using a clustering algorithm, with dynamic density parameters dynamically set according to the mean and standard deviation of the data distribution; constructing a triplet knowledge graph containing equipment entities, operating status, and power grid commands, applying graph neural networks for semantic association analysis, and performing consistency verification of equipment operating logic.

[0073] Furthermore, a device topology correlation matrix is ​​constructed, and the electrical distance d between power generation device i and power generation device j is calculated. ij The calculation formula is specifically expressed as follows:

[0074] d ij =α d ×L l +β d ×|V i ―V j |

[0075] Among them, L l V represents the line length between power generation equipment i and power generation equipment j. i and V j Let α represent the voltages of power generation equipment i and power generation equipment j, respectively. d β d These are respectively represented as empirical coefficients used to balance the weights of line length and node voltage differences in electrical distance calculations. In this embodiment, j d =0.6, β d =0.4;

[0076] The correlation weight w between power generation equipment i and power generation equipment j ij The calculation formula is specifically expressed as follows:

[0077]

[0078] Where δ represents the slope factor, which is taken as 0.5 in this embodiment, d max Let d represent the maximum electrical distance between power generation equipment i and power generation equipment j. ij This is expressed as the electrical distance between power generation equipment i and power generation equipment j;

[0079] In this embodiment, an improved DBSCAN clustering algorithm is used, with dynamic density parameters set. For data repeatedly reported within a preset time window, the sample with the highest confidence is retained. A triplet knowledge graph containing equipment entities, operating status, and power grid commands is constructed. A graph neural network is applied for state reasoning, and real-time data is compared with the graph logic rules. When a contradiction is found, the quality flag bit is updated.

[0080] S4: Monitor the data quality indicators during the first and second governance operations in real time, compare the monitoring results with preset quality thresholds, and generate third governance data including data quality assessment results and abnormal data tracing information;

[0081] Specifically, in S4, for the first governance operation performed at the edge layer, the system deploys a lightweight quality probe. The quality probe collects three types of core data quality indicators in real time, including integrity indicators, accuracy indicators, and consistency indicators. For the second governance operation performed at the center layer, a hierarchical quality assessment mechanism is implemented. For data quality assessment results that do not meet the quality assessment mechanism, abnormal data records with confidence level labels are generated.

[0082] In this embodiment, integrity indicators are evaluated by verifying the missing rate, null value ratio, and data acquisition cycle fluctuation of data fields; verification is performed based on a preset rule base that reflects the physical characteristics of the equipment, and the data is predicted by combining a long short-term memory network model, and the deviation rate between the actual value and the predicted value is calculated to evaluate the accuracy of the data; a distributed time window algorithm is used to verify the consistency between data from multiple logically related equipment within the same site.

[0083] In one possible implementation, the hierarchical quality assessment mechanism includes: in the timing alignment calibration channel, an NTP-enhanced clock synchronization service is deployed to continuously monitor the time offset between each edge node and the central node, and to force the offset to be controlled within a preset offset. In this embodiment, the preset offset is set to ±10 milliseconds. If a synchronization failure or deviation exceeds the limit, a resynchronization process with precise time-stamp compensation is automatically triggered to restore clock accuracy. In the multi-source data fusion channel, a metadata mapping model based on a knowledge graph is constructed to calculate the pattern matching degree between the first-governed data from different sources that have undergone preliminary governance in real time. The score is required to be no less than a preset pattern matching degree. In this embodiment, the preset pattern matching degree is set to 0.85. Semantic consistency is also calculated, and the similarity score is required to be no less than a preset similarity score. In this embodiment, the preset similarity score is set to 0.9.

[0084] In one possible implementation, generating third-party governance data includes: using sliding window statistical techniques to integrate the monitored data and generate a quality matrix of quantitative indicators. The quality matrix of quantitative indicators includes, but is not limited to, field fill rate, rule violation count, time series continuity score, data freshness, etc. If any indicator in the quality matrix exceeds a preset threshold, a response mechanism is activated.

[0085] It should be noted that the response mechanism includes: automatically marking abnormal data segments and generating a source traceability information containing the power generation equipment ID, data flow path, and processing timeline information; locating the specific link that caused the quality problem; and generating a structured quality report, which includes a visual heatmap of the severity level of the anomaly, the scope of the problem's impact, and specific remediation suggestions.

[0086] S5: Based on the third governance data, dynamically adjust the global data governance strategy and update the edge governance rule set and the central processing logic set accordingly. Based on the abnormal data tracing information, automatically trigger the preset abnormal data handling process.

[0087] Specifically, in S5, a trial run is conducted on selected edge nodes based on the quality matrix and traceability information in the third governance data. The difference in data governance efficiency between the repair suggestions in the third governance data and the original data governance is compared. When the data governance efficiency reaches the preset target and the operation is stable, the updated edge governance rule set and central processing logic set are distributed to all nodes in batches through the differential upgrade protocol.

[0088] In one possible implementation, the preset abnormal data handling process includes: judging known abnormal data based on the source information of S4, and calling the corresponding handling plan from the preset system database to perform automated repair; in this embodiment, for the detected timestamp out-of-synchronization anomaly, the operation process of "timestamp compensation-data resampling-historical window repair" is automatically executed; when new abnormal data is identified based on the source information of S4, in the virtualized new energy power station operating environment, the preset general anomaly handling plan is executed, the potential impact path of the general anomaly handling plan on data quality and system stability is evaluated, and the plan with the least expected negative impact and the best repair effect is selected for actual execution; if the source information of S4 points to a hardware-level fault, an equipment maintenance work order containing precise three-dimensional positioning coordinates is automatically generated and synchronized to the power station operation and maintenance personnel in real time.

[0089] Secondly: The accompanying drawings of the embodiments disclosed in this invention only involve the structures involved in the embodiments disclosed in this invention. Other structures can refer to the general design. In the absence of conflict, the same embodiment and different embodiments of this invention can be combined with each other.

[0090] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A multi-level management method for real-time data of power generation equipment in new energy power plants, characterized in that, include: S1: Based on the preset global data governance strategy, generate a set of mutually cooperating edge governance rules and a set of central processing logic, and send the set of edge governance rules to the corresponding edge nodes of the new energy power station; S2: The edge node of the new energy power station receives the edge governance rule set and performs a first governance operation on the raw data acquired in real time by the power generation equipment. The first governance operation includes at least data cleaning and format standardization to generate first governance data. S3: Based on the central processing logic set, perform a second governance operation on the first governance data received from the edge nodes. The second governance operation includes cross-level timestamp synchronization calibration and data fusion processing to generate second governance data, including: This is performed through a second data processing channel that includes at least a timing alignment calibration channel and a multi-source data fusion channel. The timing alignment calibration channel, based on the cross-level timing alignment rules in the central processing logic set, constructs a transmission delay topology map for the encoding structure of the first governance data and uses a hybrid clock synchronization mechanism to achieve data alignment. The hybrid clock synchronization mechanism includes at least one of NTP protocol, PTP protocol, and Kalman filter algorithm. Furthermore, the multi-source data fusion channel performs multi-source data fusion processing on the data from the time-series alignment calibration channel based on the multi-source data fusion algorithm chain in the central processing logic set, and encapsulates it into second governance data containing quality assessment labels, processing metadata and semantic association information. The metadata includes at least clock compensation coefficients, redundancy elimination identifiers and knowledge graph reasoning paths. S4: Monitor the data quality indicators during the first and second governance operations in real time, compare the monitoring results with preset quality thresholds, and generate third governance data including data quality assessment results and abnormal data tracing information; S5: Based on the third governance data, dynamically adjust the global data governance strategy and update the edge governance rule set and the central processing logic set accordingly. Based on the abnormal data tracing information, automatically trigger the preset abnormal data handling process.

2. The multi-level management method for real-time data of power generation equipment in a new energy power plant according to claim 1, characterized in that: The preset global data governance strategy in S1 includes a device type hierarchical mapping table, a data quality target matrix, and network status awareness rules.

3. The multi-level management method for real-time data of power generation equipment in a new energy power plant according to claim 1, characterized in that: The S2, the first governance operation, specifically includes: Executed through a first data processing channel that includes at least a data cleaning channel and a format standardization channel; The data cleaning channel, based on the device-level data cleaning rule library in the edge governance rule set, performs differentiated processing on data of different types of power generation equipment. The differentiated processing algorithm includes at least one of the following: a sliding window-based filtering algorithm, a wavelet denoising coefficient dynamic adjustment method, and a dynamic baseline correction algorithm under equipment start-up and shutdown conditions. Furthermore, the format standardization channel converts the data from the data cleaning channel into a predefined standard format based on the format conversion template in the edge governance rule set, and encapsulates it into a unified encoding structure containing device ID, high-precision timestamp, data value range and quality flag bits.

4. The multi-level management method for real-time data of power generation equipment in a new energy power plant according to claim 1, characterized in that: The S3, multi-source data fusion processing, specifically includes: A spatiotemporal correlation weighted model is constructed based on equipment topology correlation, where the correlation weights are calculated using an exponential decay function of electrical distance, specifically expressed as follows: in, Represented as slope factor, Represented as power generation equipment and power generation equipment Maximum electrical distance between Represented as power generation equipment and power generation equipment Electrical distance between them; Clustering algorithms are used to eliminate redundant data; Construct a triplet knowledge graph containing equipment entities, operating status, and power grid commands, apply graph neural networks for semantic association analysis, and perform consistency verification of equipment operation logic.

5. A multi-level management method for real-time data of power generation equipment in a new energy power plant according to claim 1, characterized in that: The abnormal data tracing information in S4 includes at least the power generation equipment ID, data flow path, and processing timeline information.

6. A multi-level management method for real-time data of power generation equipment in a new energy power plant according to claim 1, characterized in that: The pre-defined abnormal data handling process in S5 specifically includes: For known abnormal data identified based on the abnormal data tracing information, the corresponding handling plan is retrieved from the preset system database to perform automated repair. For new types of abnormal data identified based on the abnormal data source tracing information, a preset general handling plan is executed, the impact of the general handling plan is evaluated, and the plan with the best repair effect is selected for execution. Furthermore, in cases where the abnormal data source information indicates a hardware-level fault, an equipment maintenance work order containing location information is automatically generated and synchronized to the maintenance personnel.

Images