Multi-dimensional quality evaluation method and device for power grid time series data

Abstract

The application provides a kind of multi-dimension quality evaluation method and device of power grid time series data, it is related to power grid data management technical field, the method includes: based on the time alignment of multiple-source power grid time series data to preset acquisition cycle, obtain benchmark data set;Benchmark data set includes the cross-source associated data corresponding to different time points;All time points in benchmark data set respectively corresponding cross-source associated data are carried out abnormal identification, obtain abnormal data;Based on abnormal data and benchmark data set carry out quality evaluation, obtain multi-dimension quality evaluation index;Based on abnormal data, multi-dimension quality evaluation index and multiple-source power grid time series data, determine multi-dimension quality evaluation result.The application can unify the originally dispersed multiple-source power grid time series data, avoid the analysis one-sidedness caused by only relying on single data source or single quality index, improve the integrity, accuracy and availability of master station side data management, abnormal discovery and quality analysis.

Description

Technical Field

[0001] This invention relates to the field of power grid data governance technology, and in particular to a method and apparatus for multi-dimensional quality assessment of power grid time-series data. Background Technology

[0002] The power grid master station continuously receives a large amount of time-series data from terminal equipment, electricity meters, and data acquisition links. This type of data is typically generated continuously at a fixed sampling period and is accessed, stored, queried, and the results pushed through a time-series database. Quality analysis also requires consideration of user capacity within the equipment, historical data from the same period across days, and the business meanings corresponding to different meter readings.

[0003] In existing technologies, collected data, equipment records, standard sampling periods, and historical data are typically scattered across different storage or logical modules. Quality analysis often processes only a single data table, making it difficult to establish a complete quality assessment context. Furthermore, quality analysis usually only outputs scattered anomalies or single-item proportions, failing to consider detailed anomaly information and daily-level comprehensive quality results. Anomaly identification results and multi-dimensional quality results are also difficult to directly support main site page display, downstream governance closure loops, and business linkages. Therefore, improving the completeness, accuracy, and usability of multi-dimensional quality assessment methods for power grid time-series data is a pressing technical problem that needs to be solved. Summary of the Invention

[0004] This invention provides a multi-dimensional quality assessment method and apparatus for power grid time series data, which solves the problems of low completeness, accuracy and usability of existing multi-dimensional quality assessment methods for power grid time series data.

[0005] This invention provides a multi-dimensional quality assessment method for power grid time-series data, comprising the following steps.

[0006] Time alignment is performed on multi-source power grid time-series data based on a preset acquisition cycle to obtain a benchmark dataset; the benchmark dataset includes cross-source correlation data corresponding to different time points; Anomaly identification is performed on the cross-source correlated data corresponding to each time point in the benchmark dataset to obtain abnormal data; Based on the abnormal data and the benchmark dataset, a quality assessment is performed to obtain multi-dimensional quality assessment indicators. Based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data, the multi-dimensional quality assessment results are determined.

[0007] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, the multi-source power grid time-series data includes electrical energy time-series data and equipment attribute data; The process of aligning multi-source power grid time-series data based on a preset acquisition period to obtain a benchmark dataset includes: A standard time axis is constructed using the preset acquisition period as the interval; The time series data to be analyzed in the electrical energy time series data is mapped to different time points of the standard time axis; The historical energy data in the energy time series data is time-aligned with the time series data to be analyzed to obtain aligned time series data corresponding to different time points; the historical energy data and the time series data to be analyzed both correspond to the same electricity meter; The baseline dataset is obtained by associating the aligned time-series data corresponding to different time points with the device capacity in the device attribute data.

[0008] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, the step of identifying anomalies in the cross-source correlated data corresponding to each time point in the benchmark dataset to obtain anomaly data includes: Based on the cross-source correlation data corresponding to all time points, the first anomaly identification result corresponding to all abnormal time points and the second anomaly identification result corresponding to the historical electrical energy data are determined. The abnormal data is determined based on the first and second anomaly identification results corresponding to all abnormal time points.

[0009] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, the step of determining the first anomaly identification result corresponding to all abnormal time points based on cross-source correlation data corresponding to all time points includes: Iterate through the cross-source correlation data corresponding to all time points. Based on the time series data to be analyzed corresponding to the current time point, the historical energy data corresponding to the previous time point, and the historical energy data corresponding to the same time point on the previous day, determine whether the current time point is an abnormal time point. If the current time point is an abnormal time point, determine the first anomaly identification result corresponding to the abnormal time point. Continue until all time points have been traversed to obtain the first anomaly identification results corresponding to all abnormal time points.

[0010] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, the first anomaly identification result includes null value anomaly, negative value anomaly, or data anomaly at the same time point; The process involves determining whether the current time point is an abnormal time point based on the time series data to be analyzed corresponding to the current time point, the historical energy data corresponding to the previous time point, and the historical energy data corresponding to the same time point on the previous day. If the current time point is an abnormal time point, the first anomaly identification result corresponding to the abnormal time point is determined, including: If the time series data to be analyzed corresponding to the current time point is null, the current time point is determined as an abnormal time point, and the first abnormal identification result corresponding to the current time point is determined as a null value abnormality. If the time series data to be analyzed at the current time point is less than the historical electrical energy data at the previous time point, the current time point is determined as an abnormal time point, and the first abnormal identification result corresponding to the current time point is determined as a negative value abnormality. If the time series data to be analyzed at the current time point is less than the historical electrical energy data at the same time point on the previous day, the current time point is identified as an abnormal time point, and the first anomaly identification result corresponding to the current time point is identified as data anomaly at the same time point.

[0011] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, determining the second anomaly identification result corresponding to the historical electrical energy data includes: Determine the energy difference between the time-series data to be analyzed at midnight of the day to be analyzed and the historical energy data at midnight of the previous day; The capacity product is determined based on the preset multiplier and the device capacity; If the difference in electrical energy is greater than or equal to the product of the capacities, the second anomaly identification result corresponding to the historical electrical energy data is determined to be an anomaly in all data from the previous day.

[0012] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, the step of determining the abnormal data based on the first anomaly identification result and the second anomaly identification result corresponding to all abnormal time points includes: Based on the timestamps corresponding to adjacent abnormal time points, at least one abnormal duration interval is determined; When all abnormal time intervals are greater than the preset collection period, the first abnormal identification result and the second abnormal identification result corresponding to all abnormal time points are determined as the abnormal data. If at least one target abnormal duration interval is equal to the preset collection period among all abnormal duration intervals, adjacent abnormal time points in at least one target abnormal duration interval are merged into an abnormal time window; based on the abnormal time window, the first abnormal identification result and the second abnormal identification result corresponding to all abnormal time points, the abnormal data is determined.

[0013] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, the multi-dimensional quality assessment indicators include consistency indicators, integrity indicators, validity indicators, and timeliness indicators; The quality assessment based on the abnormal data and the benchmark dataset yields multi-dimensional quality assessment indicators, including: The consistency index is determined based on the number of first time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of first time points is the number of time points that do not belong to the abnormal data at the same time point or the abnormal data of the previous day. The integrity index is determined based on the number of second time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of second time points is the number of time points that do not belong to null value anomalies. The validity index is determined based on the number of third time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of third time points is the number of time points that do not belong to null value anomalies and negative value anomalies. The timeliness index is determined based on the fourth time point number and the second time point number in the abnormal data; the fourth time point number is the number of time points in the second time point number that are saved in a timely manner according to the preset collection cycle.

[0014] According to the multi-dimensional quality assessment method for power grid time series data provided by the present invention, the step of determining the multi-dimensional quality assessment result based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data includes: Construct a quality analysis data model; the quality analysis data model is used to characterize the logical relationships between fields required for multi-dimensional quality assessment of power grid time-series data; The quality analysis data model is updated based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment results.

[0015] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, the step of updating the quality analysis data model based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data to obtain the multi-dimensional quality assessment result includes: The quality analysis data model is populated based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data to obtain the multi-dimensional quality assessment results; or... Based on the multi-dimensional quality assessment indicators, a comprehensive quality score is determined; the comprehensive quality score is used to characterize the quality assessment level corresponding to the multi-source power grid time series data; the quality analysis data model is filled based on the comprehensive quality score, the abnormal data, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment result.

[0016] According to the multi-dimensional quality assessment method for power grid time-series data provided by the present invention, the quality analysis data model includes: meter device identifier, date, time point, collected value, historical same period value, device capacity, anomaly type, quality dimension indicators, and result identifier.

[0017] The present invention also provides a multi-dimensional quality assessment device for power grid time series data, comprising the following modules.

[0018] The alignment module is used to perform time alignment on multi-source power grid time-series data based on a preset acquisition period to obtain a benchmark dataset; the benchmark dataset includes cross-source correlation data corresponding to different time points. The anomaly detection module is used to detect anomalies in the cross-source correlated data corresponding to each time point in the benchmark dataset, and obtain the abnormal data. The quality assessment module is used to perform quality assessment based on the abnormal data and the benchmark dataset to obtain multi-dimensional quality assessment indicators. The determination module is used to determine the multi-dimensional quality assessment result based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data.

[0019] The present invention also provides a server, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement a multi-dimensional quality assessment method for power grid time-series data as described above.

[0020] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a multi-dimensional quality assessment method for power grid time-series data as described above.

[0021] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements a multi-dimensional quality assessment method for power grid time series data as described above.

[0022] The present invention provides a method and apparatus for multi-dimensional quality assessment of power grid time-series data. This method performs cross-source time alignment on multi-source power grid time-series data through a preset acquisition cycle, obtaining a benchmark dataset including cross-source correlated data corresponding to different time points. Anomalies are identified in the cross-source correlated data at different time points to obtain anomalous data. Quality assessment is then performed using the anomalous data and the benchmark dataset to obtain multi-dimensional quality assessment indicators. Finally, the multi-dimensional quality assessment result is determined using the anomalous data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data. In this invention, by performing cross-source time alignment on multi-source power grid time-series data, the previously scattered multi-source power grid time-series data is unified and integrated, avoiding the analytical bias caused by relying on a single data source or a single quality indicator. Through fine-grained anomaly identification, cross-dimensional assessment, and unified output of power grid time-series data quality, the completeness, accuracy, and usability of data governance, anomaly detection, and quality analysis at the master station are improved, as well as the integrated processing capability of anomaly identification and quality assessment at the master station. Attached Figure Description

[0023] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0024] Figure 1 This is a flowchart illustrating the multi-dimensional quality assessment method for power grid time-series data provided in this embodiment of the invention.

[0025] Figure 2 This is a schematic diagram of the structure of the multi-dimensional quality assessment device for power grid time series data provided in an embodiment of the present invention.

[0026] Figure 3 This is a schematic diagram of the server structure provided in an embodiment of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0028] To address the issues of low completeness, accuracy, and usability in existing multi-dimensional quality assessment methods for power grid time-series data, this invention provides a multi-dimensional quality assessment method for power grid time-series data. Figure 1This is a flowchart illustrating the multi-dimensional quality assessment method for power grid time-series data provided in this embodiment of the invention, as shown below. Figure 1 As shown, the method includes steps 110 to 140.

[0029] Step 110: Time-align the multi-source power grid time-series data based on a preset acquisition period to obtain a benchmark dataset; the benchmark dataset includes cross-source correlation data corresponding to different time points.

[0030] Specifically, the preset acquisition period refers to the fixed time interval set by the power grid master station for periodically sampling electrical energy data. For example, the preset acquisition period can be 15 minutes, 30 minutes, or 1 hour, and this embodiment of the invention does not limit it.

[0031] The multi-source power grid time-series data refers to power grid operation data originating from different power grid equipment, different business systems, or different databases. In this embodiment of the invention, the multi-source power grid time-series data includes electrical energy time-series data and equipment attribute data. The electrical energy time-series data originates from the power grid time-series database on the power grid master station side and refers to the meter readings that accumulate over time and reflect user electricity consumption. For example, the electrical energy time-series data may include the time-series data to be analyzed corresponding to the date to be analyzed, as well as historical electrical energy data prior to the date to be analyzed. The date to be analyzed can be today. The equipment attribute data originates from the equipment file database and refers to the fixed static attribute data of the power grid equipment. For example, the equipment attribute data may include equipment capacity, meter code parameters, and preset collection cycle, etc. The meter code parameters may include meter equipment identifier, equipment type, and equipment configuration parameters, etc.

[0032] In actual power grid operation, due to the long data acquisition links and numerous nodes, electrical energy time-series data often suffers from time drift. Simultaneously, equipment attribute information is stored independently in equipment archive databases, leading to problems such as time-series disorder and scattered storage locations in the multi-source power grid time-series data. Therefore, in this embodiment of the invention, after acquiring the multi-source power grid time-series data, it is uniformly mapped to the same standard time axis according to a preset acquisition cycle to obtain a benchmark dataset. This benchmark dataset includes cross-source correlated data corresponding to different time points. This cross-source correlated data includes data from multiple dimensions of the multi-source power grid time-series data fused across sources at the same time point, providing a unified semantic foundation for subsequent anomaly identification and multi-dimensional quality assessment.

[0033] It should be noted that, in this embodiment of the invention, the multidimensionality of the multidimensional power grid time series data refers not only to the different storage locations of the data sources, such as the multidimensional power grid time series data coming from meter devices, the time series database on the power grid master station side, the equipment archive database, etc., but also to the different data dimensions, such as the electric energy time series data being dynamic electric energy acquisition values, and the equipment attribute data being fixed and static attribute data.

[0034] Step 120: Perform anomaly identification on the cross-source correlation data corresponding to each time point in the benchmark dataset to obtain abnormal data.

[0035] Specifically, after obtaining the baseline dataset through time alignment, the cross-source correlated data at each time point is traversed one by one for anomaly identification. Abnormal time points are identified from all time points, thus obtaining abnormal data. This abnormal data may include meter device indication, anomaly type, anomaly time, and statistical point count, etc.

[0036] Step 130: Perform a quality assessment based on the abnormal data and the benchmark dataset to obtain multi-dimensional quality assessment indicators.

[0037] Specifically, after identifying outlier data and the benchmark dataset, a unified quality assessment is conducted, which involves uniformly calculating the values ​​of multi-dimensional quality assessment indicators. These multi-dimensional quality assessment indicators include consistency indicators, completeness indicators, validity indicators, and timeliness indicators.

[0038] Step 140: Based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data, determine the multi-dimensional quality assessment result.

[0039] Specifically, after generating multi-dimensional quality assessment indicators, quality analysis results are generated based on these indicators. Combined with abnormal data and multi-source power grid time-series data, multi-dimensional quality assessment results are obtained. These results are then pushed to the business system and relational database for subsequent queries, report generation, and display on the power grid master station side page.

[0040] The multi-dimensional quality assessment method for power grid time-series data provided in this invention performs cross-source time alignment on multi-source power grid time-series data through a preset acquisition cycle, obtaining a benchmark dataset including cross-source correlated data corresponding to different time points. Anomalies are identified in the cross-source correlated data at different time points to obtain anomalous data. Quality assessment is then performed using the anomalous data and the benchmark dataset to obtain multi-dimensional quality assessment indicators. Finally, the multi-dimensional quality assessment result is determined using the anomalous data, multi-dimensional quality assessment indicators, and multi-source power grid time-series data. In this invention, by performing cross-source time alignment on multi-source power grid time-series data, the originally scattered multi-source power grid time-series data is unified and integrated, avoiding the analytical bias caused by relying on only a single data source or a single quality indicator. Through fine-grained anomaly identification, cross-dimensional assessment, and unified output of power grid time-series data quality, the completeness, accuracy, and usability of data governance, anomaly detection, and quality analysis on the master station side are improved, as well as the integrated processing capability of anomaly identification and quality assessment on the master station side is enhanced.

[0041] In one embodiment, the step of time-aligning multi-source power grid time-series data based on a preset acquisition period to obtain a benchmark dataset includes: A standard time axis is constructed using the preset acquisition period as the interval; The time series data to be analyzed in the electrical energy time series data is mapped to different time points of the standard time axis; The historical energy data in the energy time series data is time-aligned with the time series data to be analyzed to obtain aligned time series data corresponding to different time points; the historical energy data and the time series data to be analyzed both correspond to the same electricity meter; The baseline dataset is obtained by associating the aligned time-series data corresponding to different time points with the device capacity in the device attribute data.

[0042] Specifically, a standard timeline is constructed using natural days as boundaries and preset collection cycles as intervals. This standard timeline includes continuous time points covering the entire day. For example, with a preset collection cycle of 15 minutes, the constructed standard timeline includes 96 time points. After determining the day to be analyzed, the time-series data to be analyzed on that day is extracted from the electricity time-series data. The timestamps in this time-series data are then mapped to the corresponding time points on the standard timeline. If no time-series data exists at a certain time point, it is marked as a placeholder. Next, historical electricity data from the same meter on the day before the analysis date is extracted from the electricity time-series data. The timestamps in the historical electricity data are then time-aligned with the time-series data to obtain aligned time-series data for each time point. This aligned time-series data includes the time-series data to be analyzed and the historical electricity data collected from the same meter at the same time point. Finally, the aligned time-series data for each time point is correlated with the device capacity to obtain cross-source correlated data corresponding to each time point. This cross-source correlation data includes time-series data and historical energy data collected from the same meter at the same point in time, as well as equipment capacity. All cross-source correlation data constitute the benchmark dataset.

[0043] In this embodiment of the invention, time-series data of multi-source power grids are time-aligned using a standard time axis to construct a benchmark dataset for single-day electricity meters. This transforms the originally scattered, time-disordered, and single-dimensional underlying raw data into structured data with unified structure, strict time alignment, and rich contextual information. This eliminates deviations in data time sequence and provides a unified input for subsequent anomaly identification and calculation of multi-dimensional quality assessment indicators. It also avoids performance bottlenecks and analytical biases caused by repeated cross-database queries during the assessment process.

[0044] In one embodiment, the step of identifying anomalies in the cross-source correlated data corresponding to each time point in the benchmark dataset to obtain anomalous data includes: Based on the cross-source correlation data corresponding to all time points, the first anomaly identification result corresponding to all abnormal time points and the second anomaly identification result corresponding to the historical electrical energy data are determined. The abnormal data is determined based on the first and second anomaly identification results corresponding to all abnormal time points.

[0045] Specifically, after determining the cross-source correlation data corresponding to all time points, anomaly identification is performed from both individual time point and daily perspectives. If the cross-source correlation data corresponding to a given time point is abnormal, that time point is identified as an abnormal time point, and the anomaly type of that time point is determined as the first anomaly identification result for that time point; that is, the first anomaly identification result is used to characterize the anomaly type at the individual time point level. If the cross-source correlation data for a given day is abnormal, the anomaly type for that day is determined as the second anomaly identification result corresponding to the historical energy data; that is, the second anomaly identification result is used to characterize the anomaly type at the daily level. Then, based on the first and second anomaly identification results corresponding to each of the abnormal time points, the abnormal data is summarized.

[0046] In one embodiment, determining the first anomaly identification result corresponding to all abnormal time points based on cross-source correlation data corresponding to all time points includes: Iterate through the cross-source correlation data corresponding to all time points. Based on the time series data to be analyzed corresponding to the current time point, the historical energy data corresponding to the previous time point, and the historical energy data corresponding to the same time point on the previous day, determine whether the current time point is an abnormal time point. If the current time point is an abnormal time point, determine the first anomaly identification result corresponding to the abnormal time point. Continue until all time points have been traversed to obtain the first anomaly identification results corresponding to all abnormal time points.

[0047] Specifically, the cross-source correlation data corresponding to all time points is traversed. For the current time point, the anomaly type is determined point-level by analyzing the time-series data to be analyzed at the current time point, the historical energy data corresponding to the previous time point, and the historical energy data corresponding to the same time point on the previous day. If the cross-source correlation data corresponding to the current time point meets the anomaly type, the current time point is identified as an anomalous time point, and the satisfied anomaly type is determined as the first anomaly identification result for that anomalous time point. The above steps are repeated until all time points are traversed, thus identifying all anomalous time points across all time points and obtaining the first anomaly identification result for each anomalous time point. The point-level anomaly type can include null value anomaly, negative value anomaly, and data anomaly at the same time point. The first anomaly identification result is any one of the point-level anomaly types, that is, the first anomaly identification result includes null value anomaly, negative value anomaly, or data anomaly at the same time point.

[0048] In one embodiment, determining whether the current time point is an abnormal time point based on the time series data to be analyzed corresponding to the current time point, the historical energy data corresponding to the previous time point, and the historical energy data corresponding to the same time point on the previous day, and determining the first anomaly identification result corresponding to the abnormal time point if the current time point is an abnormal time point, includes: If the time series data to be analyzed corresponding to the current time point is null, the current time point is determined as an abnormal time point, and the first abnormal identification result corresponding to the current time point is determined as a null value abnormality. If the time series data to be analyzed at the current time point is less than the historical electrical energy data at the previous time point, the current time point is determined as an abnormal time point, and the first abnormal identification result corresponding to the current time point is determined as a negative value abnormality. If the time series data to be analyzed at the current time point is less than the historical electrical energy data at the same time point on the previous day, the current time point is identified as an abnormal time point, and the first anomaly identification result corresponding to the current time point is identified as data anomaly at the same time point.

[0049] Specifically, for the current time point, if the time series data to be analyzed corresponding to the current time point is null, i.e., data is missing, it indicates that the current time point meets the null value anomaly. In the scenario of electricity metering, the cumulative electricity collection value under normal circumstances should show a monotonically non-decreasing trend. If the collection value of the time series data to be analyzed corresponding to the current time point is less than the collection value of the historical electricity data corresponding to the previous time point, it indicates that the current time point meets the negative value anomaly. If the collection value of the time series data to be analyzed corresponding to the current time point is less than the collection value of the historical electricity data corresponding to the same time point on the previous day, it indicates that the current time point meets the same time point data anomaly. If the current time point meets any of the above point-level anomaly types, then the current time point is determined as an anomalous time point, and the satisfied point-level anomaly type is determined as the first anomaly identification result corresponding to the anomalous time point.

[0050] In this embodiment of the invention, all anomaly types at the point level of power grid time series data are comprehensively covered. By comparing time series monotonicity and historical co-occurrence, the invention reflects the physical characteristics of electrical energy time series data while accurately capturing anomalies at individual time points, significantly reducing the false alarm rate and providing reliable data support for subsequent fine-grained data governance.

[0051] In one embodiment, determining the second anomaly identification result corresponding to the historical electrical energy data includes: Determine the energy difference between the time-series data to be analyzed at midnight of the day to be analyzed and the historical energy data at midnight of the previous day; The capacity product is determined based on the preset multiplier and the device capacity; If the difference in electrical energy is greater than or equal to the product of the capacities, the second anomaly identification result corresponding to the historical electrical energy data is determined to be an anomaly in all data from the previous day.

[0052] Specifically, the maximum daily power consumption of an electricity meter is limited by its capacity. Therefore, in this embodiment of the invention, the difference in power consumption between the time-series data to be analyzed at midnight of the day to be analyzed and the historical power consumption data at midnight of the previous day is calculated. A product of a preset multiplier and the device capacity is also calculated. This power consumption difference and the capacity product are compared. If the power consumption difference is greater than or equal to the capacity product, it indicates that the daily power consumption has exceeded the limit that the electricity meter itself can bear. Therefore, the historical power consumption data of the previous day must be incorrect. Thus, the historical power consumption data of the previous day meets the daily-level anomaly type or the second anomaly identification result, indicating that all data of the previous day is abnormal.

[0053] In this embodiment of the invention, equipment capacity is introduced to impose boundary constraints on historical electrical energy data, and anomalies are identified from the perspective of the objective capabilities of physical equipment, so as to avoid systemic data disasters on the main station side and significantly improve the operational safety of the power grid.

[0054] In one embodiment, determining the anomalous data based on the first anomaly identification result and the second anomaly identification result corresponding to all anomalous time points includes: Based on the timestamps corresponding to adjacent abnormal time points, at least one abnormal duration interval is determined; When all abnormal time intervals are greater than the preset collection period, the first abnormal identification result and the second abnormal identification result corresponding to all abnormal time points are determined as the abnormal data. If at least one target abnormal duration interval is equal to the preset collection period among all abnormal duration intervals, adjacent abnormal time points in at least one target abnormal duration interval are merged into an abnormal time window; based on the abnormal time window, the first abnormal identification result and the second abnormal identification result corresponding to all abnormal time points, the abnormal data is determined.

[0055] Specifically, after identifying all abnormal times, the abnormal duration interval between adjacent abnormal time points is calculated based on their timestamps. This abnormal duration interval characterizes the time span between adjacent abnormal time points. If each abnormal duration interval is greater than the preset collection period, it indicates that at least one normal time point exists between adjacent abnormal time points. That is, multiple independent abnormal events may occur between adjacent abnormal time points. In this case, the first and second abnormal identification results corresponding to all abnormal time points can be identified as abnormal data. If at least one target abnormal duration interval among all abnormal duration intervals is equal to the preset collection period, then there are no other normal time points between adjacent abnormal time points corresponding to each target abnormal duration interval. That is, a continuous abnormal event of the electricity meter occurs between adjacent abnormal time points, and it lasts for at least two abnormal duration intervals corresponding to two abnormal time points. In this case, these two abnormal time points or multiple abnormal time points corresponding to multiple consecutive target abnormal duration intervals can be merged into an abnormal time window. The start and end times corresponding to this abnormal time window and the number of statistical points of the included abnormal time points are determined. In addition, for the second abnormal identification result, the corresponding full-day window and the number of statistical points corresponding to 96 time points covering the entire day can be directly output. Subsequently, the start and end times, number of statistical points, anomaly type, non-continuous first anomaly identification results, and the corresponding full-day window and the number of statistical points corresponding to the 96 time points covering the entire day corresponding to the second anomaly identification results are identified as abnormal data.

[0056] In one embodiment, the multi-dimensional quality assessment indicators include consistency indicators, integrity indicators, effectiveness indicators, and timeliness indicators; The quality assessment based on the abnormal data and the benchmark dataset yields multi-dimensional quality assessment indicators, including: The consistency index is determined based on the number of first time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of first time points is the number of time points that do not belong to the abnormal data at the same time point or the abnormal data of the previous day. The integrity index is determined based on the number of second time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of second time points is the number of time points that do not belong to null value anomalies. The validity index is determined based on the number of third time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of third time points is the number of time points that do not belong to null value anomalies and negative value anomalies. The timeliness index is determined based on the fourth time point number and the second time point number in the abnormal data; the fourth time point number is the number of time points in the second time point number that are saved in a timely manner according to the preset collection cycle.

[0057] Specifically, after identifying anomalous data, the following metrics are calculated: First, the proportion of time points representing data anomalies not belonging to the same time point data anomalies or all data anomalies from the previous day to the total number of time points corresponding to the baseline dataset; this proportion is the consistency index. Second, the proportion of time points representing data anomalies not belonging to null value anomalies to the total number of time points; this proportion is the completeness index. Third, the proportion of time points representing data anomalies not belonging to null value anomalies or negative value anomalies to the total number of time points; this proportion is the validity index. Fourth, the proportion of time points representing non-null data that are promptly saved according to a preset collection cycle (e.g., 15 minutes) to the total number of time points; this proportion is the timeliness index.

[0058] In one embodiment, determining the multi-dimensional quality assessment result based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data includes: Construct a quality analysis data model; the quality analysis data model is used to characterize the logical relationships between fields required for multi-dimensional quality assessment of power grid time-series data; The quality analysis data model is updated based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment results.

[0059] Specifically, a quality analysis data model is constructed to characterize the logical relationships between fields required for multi-dimensional quality assessment of power grid time-series data. This quality analysis data model can be understood as a statistical model, and its fields may include: meter equipment identifier, date, time point, collected value, historical period value, equipment capacity, anomaly type, quality dimension indicators, and result identifier. Then, the values ​​of each field in the quality analysis data model are updated based on anomaly data, multi-dimensional quality assessment indicators, and multi-source power grid time-series data to obtain the multi-dimensional quality assessment results. For example, the values ​​of multiple fields such as meter equipment identifier, date, time point, collected value, historical period value, and equipment capacity in the quality analysis data model are updated based on multi-source power grid time-series data; the value of the anomaly type field in the quality analysis data model is updated based on anomaly data; and the values ​​of fields such as quality dimension indicators and result identifiers in the quality analysis data model are updated based on multi-dimensional quality assessment indicators.

[0060] In one embodiment, updating the quality analysis data model based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data to obtain the multi-dimensional quality assessment result includes: The quality analysis data model is populated based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data to obtain the multi-dimensional quality assessment results; or... Based on the multi-dimensional quality assessment indicators, a comprehensive quality score is determined; the comprehensive quality score is used to characterize the quality assessment level corresponding to the multi-source power grid time series data; the quality analysis data model is filled based on the comprehensive quality score, the abnormal data, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment result.

[0061] Specifically, when determining the multi-dimensional quality assessment results, the quality analysis data model can be directly populated based on abnormal data, multi-dimensional quality assessment indicators, and multi-source power grid time-series data. That is, the server directly retains the values ​​of the four quality assessment indicators and outputs them separately for multi-dimensional display on the power grid master station's side page. Furthermore, the values ​​of the four quality assessment indicators can be normalized, and a comprehensive quality score can be calculated according to preset weights or hierarchical rules to form a unified quality result. This unified quality score, abnormal data, and multi-source power grid time-series data are then populated into the quality analysis data model to obtain the multi-dimensional quality assessment results.

[0062] The following describes the multi-dimensional quality assessment device for power grid time series data provided by the present invention. The multi-dimensional quality assessment device for power grid time series data described below and the multi-dimensional quality assessment method for power grid time series data described above can be referred to in correspondence.

[0063] This invention also provides a multi-dimensional quality assessment device for power grid time-series data. Figure 2 This is a schematic diagram of the structure of the multi-dimensional quality assessment device for power grid time series data provided in an embodiment of the present invention, as shown below. Figure 2 As shown, the multi-dimensional quality assessment device 200 for power grid time series data includes: an alignment module 210, an anomaly identification module 220, a quality assessment module 230, and a determination module 240.

[0064] Alignment module 210 is used to perform time alignment on multi-source power grid time series data based on a preset acquisition period to obtain a benchmark dataset; the benchmark dataset includes cross-source correlation data corresponding to different time points.

[0065] The anomaly identification module 220 is used to identify anomalies in the cross-source associated data corresponding to each time point in the benchmark dataset, and obtain the abnormal data.

[0066] The quality assessment module 230 is used to perform quality assessment based on the abnormal data and the benchmark dataset to obtain multi-dimensional quality assessment indicators.

[0067] The determination module 240 is used to determine the multi-dimensional quality assessment result based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data.

[0068] The multi-dimensional quality assessment device for power grid time-series data provided in this invention performs cross-source time alignment on multi-source power grid time-series data through a preset acquisition cycle, obtaining a benchmark dataset including cross-source correlated data corresponding to different time points. Anomalies are identified in the cross-source correlated data at different time points to obtain anomalous data. Quality assessment is then performed using the anomalous data and the benchmark dataset to obtain multi-dimensional quality assessment indicators. Finally, the multi-dimensional quality assessment result is determined using the anomalous data, multi-dimensional quality assessment indicators, and multi-source power grid time-series data. In this invention, by performing cross-source time alignment on multi-source power grid time-series data, the originally scattered multi-source power grid time-series data is unified and integrated, avoiding the analytical bias caused by relying on only a single data source or a single quality indicator. Through fine-grained anomaly identification, cross-dimensional assessment, and unified output of power grid time-series data quality, the completeness, accuracy, and usability of data governance, anomaly detection, and quality analysis on the master station side are improved, as well as the integrated processing capability of anomaly identification and quality assessment on the master station side is enhanced.

[0069] Optionally, the multi-source power grid time-series data includes electrical energy time-series data and equipment attribute data.

[0070] Optionally, the alignment module 210 is specifically used for: A standard time axis is constructed using the preset acquisition period as the interval; The time series data to be analyzed in the electrical energy time series data is mapped to different time points of the standard time axis; The historical energy data in the energy time series data is time-aligned with the time series data to be analyzed to obtain aligned time series data corresponding to different time points; the historical energy data and the time series data to be analyzed both correspond to the same electricity meter; By associating the aligned time-series data corresponding to different time points with the device capacity in the device attribute data, cross-source associated data corresponding to different time points in the benchmark dataset is obtained.

[0071] Optionally, the anomaly detection module 220 is specifically used for: Based on the cross-source correlation data corresponding to all time points, the first anomaly identification result corresponding to all abnormal time points and the second anomaly identification result corresponding to the historical electrical energy data are determined. The abnormal data is determined based on the first and second anomaly identification results corresponding to all abnormal time points.

[0072] Optionally, the anomaly detection module 220 is specifically used for: Iterate through the cross-source correlation data corresponding to all time points. Based on the time series data to be analyzed corresponding to the current time point, the historical energy data corresponding to the previous time point, and the historical energy data corresponding to the same time point on the previous day, determine whether the current time point is an abnormal time point. If the current time point is an abnormal time point, determine the first anomaly identification result corresponding to the abnormal time point. Continue until all time points have been traversed to obtain the first anomaly identification results corresponding to all abnormal time points.

[0073] Optionally, the first anomaly identification result includes null value anomaly, negative value anomaly, or data anomaly at the same time point.

[0074] Optionally, the anomaly detection module 220 is specifically used for: If the time series data to be analyzed corresponding to the current time point is null, the current time point is determined as an abnormal time point, and the first abnormal identification result corresponding to the current time point is determined as a null value abnormality. If the time series data to be analyzed at the current time point is less than the historical electrical energy data at the previous time point, the current time point is determined as an abnormal time point, and the first abnormal identification result corresponding to the current time point is determined as a negative value abnormality. If the time series data to be analyzed at the current time point is less than the historical electrical energy data at the same time point on the previous day, the current time point is identified as an abnormal time point, and the first anomaly identification result corresponding to the current time point is identified as data anomaly at the same time point.

[0075] Optionally, the anomaly detection module 220 is specifically used for: Determine the energy difference between the time-series data to be analyzed at midnight of the day to be analyzed and the historical energy data at midnight of the previous day; The capacity product is determined based on the preset multiplier and the device capacity; If the difference in electrical energy is greater than or equal to the product of the capacities, the second anomaly identification result corresponding to the historical electrical energy data is determined to be an anomaly in all data from the previous day.

[0076] Optionally, the anomaly detection module 220 is specifically used for: Based on the timestamps corresponding to adjacent abnormal time points, at least one abnormal duration interval is determined; When all abnormal time intervals are greater than the preset collection period, the first abnormal identification result and the second abnormal identification result corresponding to all abnormal time points are determined as the abnormal data. If at least one target abnormal duration interval is equal to the preset collection period among all abnormal duration intervals, adjacent abnormal time points in at least one target abnormal duration interval are merged into an abnormal time window; based on the abnormal time window, the first abnormal identification result and the second abnormal identification result corresponding to all abnormal time points, the abnormal data is determined.

[0077] Optionally, the multi-dimensional quality assessment indicators include consistency indicators, integrity indicators, effectiveness indicators, and timeliness indicators.

[0078] Optionally, the quality assessment module 230 is specifically used for: The consistency index is determined based on the number of first time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of first time points is the number of time points that do not belong to the abnormal data at the same time point or the abnormal data of the previous day. The integrity index is determined based on the number of second time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of second time points is the number of time points that do not belong to null value anomalies. The validity index is determined based on the number of third time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of third time points is the number of time points that do not belong to null value anomalies and negative value anomalies. The timeliness index is determined based on the fourth time point number and the second time point number in the abnormal data; the fourth time point number is the number of time points in the second time point number that are saved in a timely manner according to the preset collection cycle.

[0079] Optionally, module 240 is defined, specifically for: Construct a quality analysis data model; the quality analysis data model is used to characterize the logical relationships between fields required for multi-dimensional quality assessment of power grid time-series data; The quality analysis data model is updated based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment results.

[0080] Optionally, module 240 is defined, specifically for: The quality analysis data model is populated based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data to obtain the multi-dimensional quality assessment results; or... Based on the multi-dimensional quality assessment indicators, a comprehensive quality score is determined; the comprehensive quality score is used to characterize the quality assessment level corresponding to the multi-source power grid time series data; the quality analysis data model is filled based on the comprehensive quality score, the abnormal data, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment result.

[0081] Optionally, the quality analysis data model includes: meter device identifier, date, time point, collected value, historical same period value, device capacity, anomaly type, quality dimension indicators, and result identifier.

[0082] Figure 3 This is a schematic diagram of the server structure provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the server may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340, wherein the processor 310, communications interface 320, and memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute a multi-dimensional quality assessment method for power grid time-series data. This method includes: time-aligning multi-source power grid time-series data based on a preset acquisition period to obtain a benchmark dataset; the benchmark dataset includes cross-source correlation data corresponding to different time points; identifying anomalies in the cross-source correlation data corresponding to each time point in the benchmark dataset to obtain anomalous data; performing quality assessment based on the anomalous data and the benchmark dataset to obtain multi-dimensional quality assessment indicators; and determining the multi-dimensional quality assessment result based on the anomalous data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data.

[0083] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0084] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the multi-dimensional quality assessment method for power grid time series data provided by the above methods. The method includes: performing time alignment on multi-source power grid time series data based on a preset acquisition period to obtain a benchmark dataset; the benchmark dataset includes cross-source correlation data corresponding to different time points; identifying anomalies in the cross-source correlation data corresponding to each time point in the benchmark dataset to obtain abnormal data; performing quality assessment based on the abnormal data and the benchmark dataset to obtain multi-dimensional quality assessment indicators; and determining the multi-dimensional quality assessment result based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data.

[0085] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon. When executed by a processor, the computer program implements a multi-dimensional quality assessment method for power grid time-series data provided by the methods described above. The method includes: performing time alignment on multi-source power grid time-series data based on a preset acquisition period to obtain a benchmark dataset; the benchmark dataset includes cross-source correlation data corresponding to different time points; identifying anomalies in the cross-source correlation data corresponding to each time point in the benchmark dataset to obtain anomalous data; performing quality assessment based on the anomalous data and the benchmark dataset to obtain multi-dimensional quality assessment indicators; and determining the multi-dimensional quality assessment result based on the anomalous data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data.

[0086] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0087] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0088] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-dimensional quality assessment method for power grid time-series data, characterized in that, include: Time alignment is performed on multi-source power grid time-series data based on a preset acquisition cycle to obtain a benchmark dataset; The benchmark dataset includes cross-source correlated data corresponding to different time points; Anomaly identification is performed on the cross-source correlated data corresponding to each time point in the benchmark dataset to obtain abnormal data; Based on the abnormal data and the benchmark dataset, a quality assessment is performed to obtain multi-dimensional quality assessment indicators. Based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data, the multi-dimensional quality assessment results are determined.

2. The multi-dimensional quality assessment method for power grid time-series data according to claim 1, characterized in that, The multi-source power grid time-series data includes electrical energy time-series data and equipment attribute data; The process of aligning multi-source power grid time-series data based on a preset acquisition period to obtain a benchmark dataset includes: A standard time axis is constructed using the preset acquisition period as the interval; The time series data to be analyzed in the electrical energy time series data is mapped to different time points of the standard time axis; The historical energy data in the energy time series data is time-aligned with the time series data to be analyzed to obtain aligned time series data corresponding to different time points; the historical energy data and the time series data to be analyzed both correspond to the same electricity meter; The baseline dataset is obtained by associating the aligned time-series data corresponding to different time points with the device capacity in the device attribute data.

3. The multi-dimensional quality assessment method for power grid time series data according to claim 2, characterized in that, The step involves identifying anomalies in the cross-source correlated data corresponding to each time point in the benchmark dataset, resulting in anomalous data, including: Based on the cross-source correlation data corresponding to all time points, the first anomaly identification result corresponding to all abnormal time points and the second anomaly identification result corresponding to the historical electrical energy data are determined. The abnormal data is determined based on the first and second anomaly identification results corresponding to all abnormal time points.

4. The multi-dimensional quality assessment method for power grid time series data according to claim 3, characterized in that, The determination of the first anomaly identification result corresponding to all abnormal time points based on cross-source correlation data at all time points includes: Iterate through the cross-source correlation data corresponding to all time points. Based on the time series data to be analyzed corresponding to the current time point, the historical energy data corresponding to the previous time point, and the historical energy data corresponding to the same time point on the previous day, determine whether the current time point is an abnormal time point. If the current time point is an abnormal time point, determine the first anomaly identification result corresponding to the abnormal time point. Continue until all time points have been traversed to obtain the first anomaly identification results corresponding to all abnormal time points.

5. The multi-dimensional quality assessment method for power grid time series data according to claim 4, characterized in that, The first anomaly identification result includes null value anomaly, negative value anomaly, or data anomaly at the same time point; The process involves determining whether the current time point is an abnormal time point based on the time series data to be analyzed corresponding to the current time point, the historical energy data corresponding to the previous time point, and the historical energy data corresponding to the same time point on the previous day. If the current time point is an abnormal time point, the first anomaly identification result corresponding to the abnormal time point is determined, including: If the time series data to be analyzed corresponding to the current time point is null, the current time point is determined as an abnormal time point, and the first abnormal identification result corresponding to the current time point is determined as a null value abnormality. If the time series data to be analyzed at the current time point is less than the historical electrical energy data at the previous time point, the current time point is determined as an abnormal time point, and the first abnormal identification result corresponding to the current time point is determined as a negative value abnormality. If the time series data to be analyzed at the current time point is less than the historical electrical energy data at the same time point on the previous day, the current time point is identified as an abnormal time point, and the first anomaly identification result corresponding to the current time point is identified as data anomaly at the same time point.

6. The multi-dimensional quality assessment method for power grid time series data according to claim 5, characterized in that, Determining the second anomaly identification result corresponding to the historical electrical energy data includes: Determine the energy difference between the time-series data to be analyzed at midnight of the day to be analyzed and the historical energy data at midnight of the previous day; The capacity product is determined based on the preset multiplier and the device capacity; If the difference in electrical energy is greater than or equal to the product of the capacities, the second anomaly identification result corresponding to the historical electrical energy data is determined to be an anomaly in all data from the previous day.

7. The multi-dimensional quality assessment method for power grid time series data according to claim 3, characterized in that, The process of determining the abnormal data based on the first and second anomaly identification results corresponding to all abnormal time points includes: Based on the timestamps corresponding to adjacent abnormal time points, at least one abnormal duration interval is determined; When all abnormal time intervals are greater than the preset collection period, the first abnormal identification result and the second abnormal identification result corresponding to all abnormal time points are determined as the abnormal data. If at least one target abnormal duration interval is equal to the preset collection period among all abnormal duration intervals, adjacent abnormal time points in at least one target abnormal duration interval are merged into an abnormal time window; based on the abnormal time window, the first abnormal identification result and the second abnormal identification result corresponding to all abnormal time points, the abnormal data is determined.

8. The multi-dimensional quality assessment method for power grid time series data according to claim 6, characterized in that, The multi-dimensional quality assessment indicators include consistency indicators, completeness indicators, effectiveness indicators, and timeliness indicators; The quality assessment based on the abnormal data and the benchmark dataset yields multi-dimensional quality assessment indicators, including: The consistency index is determined based on the number of first time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of first time points is the number of time points that do not belong to the abnormal data at the same time point or the abnormal data of the previous day. The integrity index is determined based on the number of second time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of second time points is the number of time points that do not belong to null value anomalies. The validity index is determined based on the number of third time points in the abnormal data and the total number of time points corresponding to the benchmark dataset; the number of third time points is the number of time points that do not belong to null value anomalies and negative value anomalies. The timeliness index is determined based on the fourth time point number and the second time point number in the abnormal data; the fourth time point number is the number of time points in the second time point number that are saved in a timely manner according to the preset collection cycle.

9. The multi-dimensional quality assessment method for power grid time series data according to claim 6, characterized in that, The determination of the multi-dimensional quality assessment result based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data includes: Construct a quality analysis data model; the quality analysis data model is used to characterize the logical relationships between fields required for multi-dimensional quality assessment of power grid time-series data; The quality analysis data model is updated based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment results.

10. The multi-dimensional quality assessment method for power grid time series data according to claim 9, characterized in that, The process of updating the quality analysis data model based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment results includes: The quality analysis data model is populated based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time-series data to obtain the multi-dimensional quality assessment results; or... Based on the multi-dimensional quality assessment indicators, a comprehensive quality score is determined; the comprehensive quality score is used to characterize the quality assessment level corresponding to the multi-source power grid time series data; the quality analysis data model is filled based on the comprehensive quality score, the abnormal data, and the multi-source power grid time series data to obtain the multi-dimensional quality assessment result.

11. The multi-dimensional quality assessment method for power grid time series data according to claim 9, characterized in that, The quality analysis data model includes: meter device identifier, date, time point, collected value, historical values ​​for the same period, device capacity, anomaly type, quality dimension indicators, and result identifier.

12. A multi-dimensional quality assessment device for power grid time-series data, characterized in that, include: The alignment module is used to perform time alignment on multi-source power grid time-series data based on a preset acquisition period to obtain a benchmark dataset. The benchmark dataset includes cross-source correlated data corresponding to different time points; The anomaly detection module is used to detect anomalies in the cross-source correlated data corresponding to each time point in the benchmark dataset, and obtain the abnormal data. The quality assessment module is used to perform quality assessment based on the abnormal data and the benchmark dataset to obtain multi-dimensional quality assessment indicators. The determination module is used to determine the multi-dimensional quality assessment result based on the abnormal data, the multi-dimensional quality assessment indicators, and the multi-source power grid time series data.

Images