Electric power transaction data full-link standardization processing method and system

Abstract

The invention is suitable for the technical field of power transaction data processing, and provides a power transaction data full-link standardization processing method and system, and the method comprises the steps: determining a remaining time interval corresponding to a completion behavior characteristic index of a target completion execution main body when different sample sets correspond to different remaining time intervals, and when the target completion execution main body executes subsequent missing data completion processing consistent with the service type, preferentially distributing the data completion task of which the predicted remaining time falls into the remaining time interval to the target completion execution main body. According to the method, an existing completion mode and an auditing mechanism are not changed, and reasonable distribution of completion tasks is realized through accurate matching of the tasks and time conditions, so that the amplification effect of completion errors in an exception recognition stage is effectively reduced, and the stability and consistency of exception recognition results are improved. The method can be stably operated in a large-scale and batch data processing scene.

Description

A standardized processing method and system for the entire power trading data chain Technical Field

[0001] This invention belongs to the field of power transaction data processing technology, and in particular relates to a standardized processing method and system for the entire power transaction data chain. Background Technology

[0002] In the field of power data trading, the time series of power transaction declarations is a crucial foundational data supporting power market clearing, transaction settlement, and operational monitoring. In actual operation, due to objective reasons such as metering device malfunctions, communication interruptions, system switching, or delays in cross-system data exchange, data gaps at certain points in the power transaction declaration time series are inevitable. To address this issue, existing technologies typically employ missing data completion methods to estimate and fill in the missing data, ensuring the continuity and integrity of subsequent data processing.

[0003] Existing methods for missing data completion largely rely on estimations based on historical data from adjacent time points, statistical patterns, or business experience. The completed data is then validated by setting reasonable ranges or auditing rules. If the completed data meets a preset reasonable range, it can proceed to subsequent processing steps, including anomaly identification, risk analysis, and business decision-making. This type of technology can, to a certain extent, meet the basic data integrity requirements of power trading systems and has been widely applied in practical engineering projects.

[0004] However, with the continuous expansion of electricity trading data and the increasing sensitivity of anomaly detection algorithms to data, existing technologies have gradually revealed their shortcomings. On the one hand, missing data completion is essentially a predictive process, and different completion execution entities inevitably have differences in operating habits, experience judgments, and risk preferences during the completion process. On the other hand, existing technologies typically use random or balanced allocation methods to assign completion tasks to different completion execution entities, failing to distinguish and utilize these individual differences. Although such differences usually do not directly cause anomalies during the completion stage itself, after the completed data enters the anomaly detection stage, the subtle differences introduced in the completion stage may be further amplified, thereby affecting the stability and consistency of the anomaly detection results.

[0005] Furthermore, existing technologies do not consider the modulating effect of the time distance difference between the completion of the completion process and the triggering of anomaly detection on the degree of influence of the completion result. Different data entering the anomaly detection stage under different time conditions will have different amplification effects on their completion errors, but existing systems lack mechanisms to identify and utilize this pattern. Therefore, how to identify and utilize the individual differences of the completion execution subject without changing the existing completion methods and review mechanisms, and reduce the amplification effect of completion errors in the anomaly detection stage, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] The purpose of this invention is to provide a standardized processing method and system for the entire power trading data chain, aiming to solve the problems mentioned in the background art.

[0007] This invention is implemented as follows: a standardized processing method for the entire power transaction data chain. The method includes: when a target completion execution entity performs completion processing on missing data in the current power transaction declaration time series, determining the business type of the power transaction declaration time series, and based on a preset historical database, obtaining completion behavior characteristic indicators formed by the target completion execution entity when processing this business type; constructing several sample sets based on the preset historical database, each sample set corresponding to a different completion execution entity and its completion behavior characteristic indicators, and the business type of each sample set being consistent, wherein each sample set contains the same number of samples, and different samples correspond to different arrival anomalies. The remaining time in the identification phase; obtaining the feedback evaluation indicators generated by each sample in the anomaly identification phase, and sorting the samples in the sample set according to the remaining time, analyzing the trend of the feedback evaluation indicators with the remaining time to determine whether there is a remaining time interval that makes the feedback evaluation indicators optimal; when different sample sets correspond to different remaining time intervals, determining the remaining time interval corresponding to the completion behavior characteristic indicators of the target completion execution subject, and when the target completion execution subject performs subsequent missing data completion processing consistent with the business type, prioritizing the allocation of data completion tasks whose predicted remaining time falls into the remaining time interval to the target completion execution subject.

[0008] As a further limitation of the technical solution of this invention embodiment, the process of obtaining the completion behavior feature index includes: based on a preset historical database, extracting at least one of the following feature indices from the completion result data generated by the completion execution subject during the historical missing data completion process: completion value bias feature index, used to characterize the degree of bias of the completion result relative to the data at adjacent time points before and after the missing data; completion smoothness feature index, used to characterize the degree of suppression or amplification of time series fluctuations by the completion result; completion stability feature index, used to characterize the stability of the change amplitude of the completion result within a continuous time period; and performing comprehensive calculation based on the at least one feature index to obtain the completion behavior feature index characterizing the completion behavior characteristics of the completion execution subject.

[0009] As a further limitation of the technical solution of the present invention, when constructing the sample set, different completion execution entities process the data completion processing scenario in the corresponding sample set in the same way as the current one. The data completion processing scenario being consistent with the current one means that the power transaction declaration time series corresponding to each sample has the same time granularity and consistent anomaly identification processing logic, and the missing position and missing length of the missing data are within a preset comparable range.

[0010] As a further limitation of the technical solution of the present invention, the remaining time to reach the anomaly identification stage refers to the time interval from the time when the missing data completion processing is completed to the time when the corresponding power transaction declaration time series enters the anomaly identification processing process.

[0011] As a further limitation of the technical solution of this embodiment of the invention, the step of obtaining the feedback evaluation index generated by each sample in the anomaly identification stage, sorting the samples in the sample set according to the remaining time, and analyzing the trend of the feedback evaluation index changing with the remaining time to determine whether there is a remaining time interval that makes the feedback evaluation index reach the optimal level includes: obtaining the feedback evaluation index generated by the samples in each sample set in the anomaly identification stage based on a preset historical database, and arranging the samples in each sample set in an ordered sequence according to the remaining time to form an ordered sequence of samples; analyzing the trend of the feedback evaluation index changing with the remaining time in the ordered sequence of samples in each sample set, and determining whether there is a feedback evaluation index interval that changes from rising to falling in the trend; when it is determined that the feedback evaluation index interval exists, the feedback evaluation index interval is determined as the optimal feedback evaluation index interval, and the corresponding remaining time interval is obtained.

[0012] As a further limitation of the technical solution of this embodiment of the invention, the process of obtaining the feedback evaluation index includes: obtaining the processing result of each sample in the anomaly identification stage based on a preset historical database; calculating at least one of the following values ​​in the anomaly judgment consistency value, anomaly trigger stability value, and anomaly characteristic fluctuation suppression value generated in the anomaly identification stage based on the processing result; and using the at least one value or its weighted result as the feedback evaluation index, wherein the larger the value of the feedback evaluation index, the higher the stability of the corresponding completion result for anomaly identification processing.

[0013] As a further limitation of the technical solution of this embodiment of the invention, when different sample sets correspond to different remaining time intervals, the step of determining the remaining time interval corresponding to the completion behavior feature index of the target completion execution subject, and prioritizing the allocation of data completion tasks whose predicted remaining time falls within the remaining time interval to the target completion execution subject when the target completion execution subject performs subsequent missing data completion processing consistent with the business type, includes: when it is determined that sample sets belonging to different completion behavior feature indices correspond to different remaining time intervals that optimize the feedback evaluation index, a matching relationship is determined between the completion behavior feature index and the remaining time interval; based on the matching relationship, a sample set that is the same as or within the preset allowable error range of the completion behavior feature index of the target completion execution subject is determined, and the remaining time interval corresponding to the sample set is determined as the matching remaining time interval of the target completion execution subject; when the target completion execution subject performs subsequent missing data completion processing consistent with the business type, the remaining time of the data completion task to be processed is predicted, and when the predicted remaining time falls within the matching remaining time interval, the data completion task is prioritized and allocated to the target completion execution subject.

[0014] A standardized processing system for the entire power transaction data chain, comprising: a behavior feature acquisition module, used to determine the business type of the power transaction declaration time series when the target completion execution entity performs completion processing on the missing data of the current power transaction declaration time series, and to acquire completion behavior feature indicators formed by the target completion execution entity when processing the business type based on a preset historical database; and a sample set construction module, used to construct several sample sets based on the preset historical database, each sample set corresponding to a different completion execution entity and its completion behavior feature indicators, and the business type of each sample set being consistent, wherein each sample set contains the same number of samples, and different samples correspond to different arrival anomaly identification stages. The remaining time of the segment; the interval analysis module is used to obtain the feedback evaluation indicators generated by each sample in the anomaly identification stage, and sort the samples in the sample set according to the remaining time, analyze the trend of the feedback evaluation indicators with the remaining time, and determine whether there is a remaining time interval that makes the feedback evaluation indicators reach the optimal level; the task allocation module is used to determine the remaining time interval corresponding to the completion behavior characteristic indicators of the target completion execution subject when different sample sets correspond to different remaining time intervals, and when the target completion execution subject performs subsequent missing data completion processing consistent with the business type, the data completion tasks whose predicted remaining time falls into the remaining time interval are preferentially allocated to the target completion execution subject.

[0015] As a further limitation of the technical solution of this invention embodiment, the process of obtaining the completion behavior feature index includes: based on a preset historical database, extracting at least one of the following feature indices from the completion result data generated by the completion execution subject during the historical missing data completion process: completion value bias feature index, used to characterize the degree of bias of the completion result relative to the data at adjacent time points before and after the missing data; completion smoothness feature index, used to characterize the degree of suppression or amplification of time series fluctuations by the completion result; completion stability feature index, used to characterize the stability of the change amplitude of the completion result within a continuous time period; and performing comprehensive calculation based on the at least one feature index to obtain the completion behavior feature index characterizing the completion behavior characteristics of the completion execution subject.

[0016] As a further limitation of the technical solution of the present invention, when constructing the sample set, different completion execution entities process the data completion processing scenario in the corresponding sample set in the same way as the current one. The data completion processing scenario being consistent with the current one means that the power transaction declaration time series corresponding to each sample has the same time granularity and consistent anomaly identification processing logic, and the missing position and missing length of the missing data are within a preset comparable range.

[0017] Compared with existing technologies, this invention has the following advantages: Addressing the problem that missing data completion results in power transaction declaration time series are easily amplified in the subsequent anomaly identification stage, this invention proposes a data end-to-end standardized processing method based on the adaptation of completion behavior characteristics to the anomaly identification time distance. By extracting completion behavior characteristic indicators that can characterize the individual differences of the completion execution subject from historical completion results, and introducing the remaining time between completion completion and anomaly identification trigger as a key variable, the optimal adaptation interval for different individual differences under different time conditions is systematically identified. This invention does not change the existing completion method and review mechanism, but achieves reasonable allocation of completion tasks through precise matching of tasks and time conditions, thereby effectively reducing the amplification effect of completion errors in the anomaly identification stage and improving the stability and consistency of anomaly identification results. This method can run stably in large-scale, batch data processing scenarios and has good engineering feasibility and promotional value. Attached Figure Description

[0018] Figure 1 is a flowchart of the method provided in the embodiment of the present invention; Figure 2 is a flowchart of identifying the remaining time interval in the method provided in the embodiment of the present invention; Figure 3 is a flowchart of prioritizing the allocation of data completion tasks in the method provided in the embodiment of the present invention; Figure 4 is an application architecture diagram of the system provided in the embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] Figure 1 shows a flowchart of the method provided by an embodiment of the present invention.

[0021] Specifically, a standardized processing method for the entire power transaction data chain includes the following steps: Step S100, when the target completion execution entity performs completion processing on the missing data of the current power transaction declaration time series, the business type of the power transaction declaration time series is determined, and based on a preset historical database, the completion behavior characteristic indicators formed by the target completion execution entity when processing the business type are obtained.

[0022] The process of obtaining the completion behavior feature indicators includes: based on a preset historical database, extracting at least one of the following feature indicators from the completion result data generated by the completion execution subject during the historical missing data completion process: completion value bias feature indicator, used to characterize the degree of bias of the completion result relative to the data at adjacent time points before and after the missing data; completion smoothness feature indicator, used to characterize the degree of suppression or amplification of time series fluctuations by the completion result; completion stability feature indicator, used to characterize the stability of the change amplitude of the completion result within a continuous time period; and performing comprehensive calculation based on the at least one feature indicator to obtain the completion behavior feature indicators characterizing the completion behavior features of the completion execution subject.

[0023] In this embodiment of the invention, the method is applied to the operation and management of the power market, and is particularly suitable for data completion and subsequent processing when power transaction declaration data is incomplete during collection, transmission, and processing. The so-called completion execution entity refers to the processing port that actually undertakes the task of completing missing data. This can be a human processor, a human-participating processing station, or a semi-automatic processing node configured with human decision-making rules. In the power trading system, the power transaction declaration time series refers to time-series data formed according to a preset time granularity (e.g., 15 minutes, 30 minutes, or 1 hour), used to describe the declared electricity volume, price, or combined declaration information of market participants. This type of data is crucial foundational data for market clearing, deviation assessment, and risk control.

[0024] In existing technologies, due to objective reasons such as metering equipment malfunctions, communication interruptions, system switching, or delays in cross-system data integration, data gaps frequently occur in the time series of electricity trading declarations. Such missing data is often difficult to accurately retrieve through the original channels in actual operation; it can only be estimated and supplemented based on historical data before and after the missing location, data from similar time periods, or statistical patterns. Therefore, missing data supplementation is essentially a predictive and estimating process, rather than a precise restoration of the objective true value. For example, in the day-ahead market, if a power generation entity loses its declared electricity data for the period 10:00–10:15, the system will typically refer to the entity's declaration trends for the periods 9:45–10:00 and 10:15–10:30, combined with historical data from the same time period on the same day, to estimate and supplement the missing declaration value.

[0025] In step S100, the business type of the current electricity transaction declaration time series needs to be determined first. The business type distinguishes data generated under different market mechanisms or different transaction stages. Examples include day-ahead electricity transaction declarations, real-time electricity transaction declarations, ancillary service declarations, and deviation settlement-related declarations. Different business types differ in time granularity, volatility characteristics, and risk sensitivity. Therefore, when completing missing data, analysis and comparison within the same business type are necessary to ensure comparability in subsequent processing. The business type can be directly obtained through the transaction identifier field, business code, or information from the relevant market module; this is a standard processing step in existing electricity trading systems.

[0026] This invention introduces the concept of a completion behavior characteristic index. This index comprehensively characterizes the stable completion behavior characteristics formed by the completion execution entity during long-term missing data completion operations. These characteristics are not temporary parameter settings, but rather the result of the combined effects of factors such as the individual cognitive habits, professional experience, risk preferences, and familiarity with business rules of the completion execution entity. Because different completion execution entities inevitably differ in their experience accumulation, operational habits, and judgment methods, even if they receive the same business training and operational standards, their completion methods when faced with specific missing data will still exhibit stable but different tendencies. Therefore, the completion behavior characteristic indices corresponding to different completion execution entities are usually not the same.

[0027] To ensure the computability and feasibility of the completion behavior characteristic indicators, this embodiment of the invention selects three types of characteristic indicators as its foundation. The completion value bias characteristic indicator is used to characterize the numerical orientation of the completion result. It reflects whether the completion entity, when completing missing data, tends to refer to data from the previous time point, the next time point, or the middle value or trend extrapolation value between the two. For example, some completion entities, when faced with short-term missing data, tend to adopt a conservative completion method of "using the previous time point," while others tend to extrapolate according to the overall trend. This difference can be quantified by statistically analyzing the direction and magnitude of the deviation of the completed value relative to adjacent time point data.

[0028] The smoothing feature index for completion reflects the completion entity's preference for handling time series volatility. It characterizes the extent to which the completion result suppresses or amplifies the volatility of the original time series. Some completion entities prioritize the overall smoothness of the time series during the completion process, tending to weaken local abrupt changes, while others emphasize maintaining the original volatility characteristics, preserving even large jumps. This feature can be calculated by comparing the changes in the volatility amplitude of the time series before and after completion, thus reflecting the completion entity's trade-off between "smoothing" and "fidelity preservation."

[0029] The completion stability metric is used to characterize the stability of completion results over a continuous time period. It reflects whether the completion results exhibit consistency when dealing with adjacent missing data points or similar business scenarios. Some completion entities show relatively stable completion results when processing consecutive missing data, while others may frequently adjust the completion values, leading to significant fluctuations in the completion results within a short period. By statistically analyzing the magnitude of changes in completion results within a continuous time window, this type of stability characteristic can be effectively characterized.

[0030] The reason for choosing the above three types of feature indicators is that they characterize the completion behavior of the completion execution subject from three complementary dimensions: numerical orientation, fluctuation handling, and continuity consistency. These indicators can reflect individual differences without relying on specific completion algorithm forms, thus possessing good universality. By comprehensively calculating at least one of the above feature indicators, a completion behavior feature indicator can be formed to characterize the completion behavior of the completion execution subject, thereby reflecting the overall differences between different completion execution subjects.

[0031] The core research point of this invention lies in identifying and utilizing the impact of the aforementioned individual differences in subsequent processing links. Those skilled in the art recognize that the processing of power transaction declaration time series data is typically characterized by batch processing and high concurrency. In actual operation, a large number of power transaction declaration time series need to complete missing data completion within a limited time. To improve processing efficiency, ensure processing diversity, and avoid risk concentration at a single processing port, existing systems typically allocate the power transaction declaration time series to be completed to multiple completion execution entities for parallel processing.

[0032] Although different entities performing data completion typically receive the same business training and follow unified completion standards, their basic principles and reasonable scope for completion operations are consistent. Therefore, during the data completion phase, the completion results usually do not directly cause obvious anomalies, and existing technologies generally have review mechanisms in place. As long as the completion results fall within a preset reasonable range, they are allowed to proceed to subsequent processing steps. However, those skilled in the art have discovered through long-term operational analysis that the completed data itself has estimable and uncertain aspects, and this subtle uncertainty is not entirely neutral in subsequent processing links.

[0033] Especially in the anomaly identification stage, which is typically used to identify abnormal fluctuations, abnormal declarations, or potential risks in the power transaction declaration time series, the identification results often depend on the time series' changing trends, fluctuation amplitudes, and statistical characteristics. Because anomaly identification algorithms are highly sensitive to input data, subtle differences introduced in the completion stage may be amplified in the anomaly identification stage, thus affecting the anomaly determination results. Furthermore, the time distance between the completion of different power transaction declaration time series and the anomaly identification stage varies. For example, different scheduling cycles or different processing queues can lead to differences between the completion time and the anomaly identification trigger time. This difference in time distance further affects the performance of the completed data in the anomaly identification stage, causing the amplification effect of completion differences to vary. This is because the different time distances between the completion processing and the anomaly identification trigger cause differences in the way and frequency of the completed data's participation in statistical analysis and feature calculations in subsequent processing, thereby changing its influence weight in the anomaly identification stage and resulting in varying degrees to which subtle differences formed in the completion stage are amplified in the anomaly identification results.

[0034] Through analysis of a large number of historical samples, this invention has found that, all other things being equal, different completion execution entities exhibit varying adaptability of their completion results to the anomaly detection stage due to differences in their completion behavior characteristic indicators. Specifically, different completion behavior characteristic indicators are more suitable for different time distances between completion completion and anomaly detection triggering. Existing technologies fail to identify and utilize this relationship between individual differences and time distances. They typically employ random or balanced allocation methods to assign completion tasks to various completion execution entities, failing to reasonably match completion tasks with different anomaly detection time distances based on the completion behavior characteristics of different completion execution entities.

[0035] The preset historical database originates from historical business data accumulated by the power trading system during its long-term operation. This data may include, but is not limited to, historical power trading declaration time series data, records of the location and length of missing historical data, corresponding completion results, completion execution entity identification information, processing results from the anomaly identification phase, and feedback evaluation indicators. By organizing and storing this historical data, a reliable data foundation can be provided for extracting completion behavior characteristic indicators, constructing sample sets, and subsequent analysis.

[0036] Furthermore, the power transaction data end-to-end standardization processing method also includes the following steps: Step S200, constructing several sample sets based on a preset historical database. Each sample set corresponds to a different completion execution entity and its completion behavior characteristic indicators, and the business types of each sample set are consistent. Each sample set contains the same number of samples, and different samples correspond to different remaining times to reach the anomaly identification stage. The remaining time to reach the anomaly identification stage refers to the time interval from the completion of the missing data completion processing to the trigger time when the corresponding power transaction declaration time series enters the anomaly identification processing flow.

[0037] When constructing the sample set, different completion execution entities process the data completion scenarios of the corresponding sample sets in the same way as the current ones. The data completion scenarios being consistent with the current ones means that the power transaction declaration time series corresponding to each sample has the same time granularity and consistent anomaly identification and processing logic, and the missing location and missing length of the missing data are within a preset comparable range.

[0038] In this embodiment of the invention, step S200 aims to construct a set of comparable and representative historical samples for subsequent analysis of the adaptation relationship between the feature indicators of data completion behavior and the remaining time in the anomaly identification stage. This step directly echoes the core research point raised in step S100, namely, that the individual differences exhibited by different data completion subjects in the missing data completion stage do not immediately trigger anomalies during the completion stage, but rather exhibit varying degrees of amplification effects in the subsequent anomaly identification stage due to the influence of time distance. Therefore, it is necessary to conduct a quantitative analysis of the relationship between this difference and time distance through the systematic construction of historical samples.

[0039] In step S200, several sample sets are constructed based on a preset historical database. Each sample set corresponds to a completion execution entity and its completion behavior characteristic indicators. The difference between different sample sets lies only in the difference between the completion execution entity and its completion behavior characteristic indicators, while the business types corresponding to each sample set remain consistent. In this way, horizontal comparative analysis of the completion behavior characteristics of different completion execution entities can be performed without interference from differences in business types.

[0040] To ensure the comparability and statistical significance of the analysis results, each sample set contains the same number of samples, with different samples corresponding to different remaining times to reach the anomaly identification stage. It should be noted that in different sample sets, samples with the same sequence number have consistent remaining times, meaning that each sample set forms a one-to-one correspondence in the remaining time dimension. For example, each sample set includes a group of samples with remaining times of T1, T2, T3, etc., ensuring that the comparison objects are completely consistent in the time distance dimension when analyzing different completion behavior characteristic indicators. This approach effectively avoids analytical bias caused by inconsistent sample distribution, making the trends of feedback evaluation indicators directly comparable between different sample sets.

[0041] When constructing the sample set, it is also necessary to further ensure that the data completion processing scenario remains consistent with the current processing scenario. This limitation is not intended to narrow the scope of application, but rather to minimize the interference of non-core factors on the results during the analysis phase. Specifically, requiring that the power transaction declaration time series corresponding to each sample have the same time granularity is to avoid changes in the fluctuation characteristics of the time series and the sensitivity of anomaly identification due to different time resolutions; requiring that the anomaly identification processing logic be consistent is to ensure that the same identification rules, feature extraction methods, and judgment mechanisms are used in the anomaly identification stage, thereby giving the feedback evaluation indicators a unified physical meaning; at the same time, limiting the missing location and missing length of the missing data to a preset comparable range can avoid uncontrollable impacts on the difficulty of completion and the stability of the completion results due to excessive differences in the missing data form.

[0042] It should be noted that this invention is not based on empirical judgment under a small sample size or lenient conditions, but rather on an analytical method built upon large-scale historical data. Because power trading systems accumulate a large amount of historical data under different completion execution entities, business types, and processing time sequences during long-term operation, even after applying the aforementioned strict screening conditions, a sufficient and reasonably distributed sample set can still be obtained. By constructing the sample set under strict comparability conditions, the subsequent analysis results can be made more stable and reliable, thus providing solid data support for identifying the fit between completion behavior characteristic indicators and remaining time. This method of fine-tuning and comparative analysis based on big data is also one of the important reasons why this invention can effectively discover patterns that have not been identified in existing technologies.

[0043] Furthermore, the power transaction data end-to-end standardization processing method also includes the following steps: Step S300, obtaining the feedback evaluation indicators generated by each sample in the anomaly identification stage, sorting the samples in the sample set according to the remaining time, analyzing the trend of the feedback evaluation indicators changing with the remaining time, so as to determine whether there is a remaining time interval that makes the feedback evaluation indicators reach the optimal level.

[0044] Specifically, Figure 2 shows a flowchart for identifying the remaining time interval.

[0045] The process of obtaining feedback evaluation indicators generated by each sample during the anomaly identification phase, sorting the samples in the sample set according to the remaining time, and analyzing the trend of the feedback evaluation indicators with the remaining time to determine whether there is a remaining time interval that makes the feedback evaluation indicators optimal includes the following steps: Step S301, based on a preset historical database, obtain the feedback evaluation indicators generated by the samples in each sample set during the anomaly identification phase, and arrange the samples in each sample set in order according to the remaining time to form an ordered sample sequence; Step S302, analyze the trend of the feedback evaluation indicators in the ordered sample sequence of each sample set with the remaining time, and determine whether there is a feedback evaluation indicator interval that changes from rising to falling in the trend; Step S303, when it is determined that the feedback evaluation indicator interval exists, the feedback evaluation indicator interval is determined as the optimal feedback evaluation indicator interval, and the corresponding remaining time interval is obtained.

[0046] The process of obtaining the feedback evaluation index includes: based on a preset historical database, obtaining the processing results of each sample in the anomaly identification stage; calculating at least one of the following values ​​in the anomaly judgment consistency value, anomaly trigger stability value, and anomaly characteristic fluctuation suppression value generated in the anomaly identification stage based on the processing results; and using the at least one value or its weighted result as the feedback evaluation index, wherein the larger the value of the feedback evaluation index, the higher the stability of the corresponding completion result for anomaly identification processing.

[0047] In this embodiment of the invention, the core function of step S300 is to introduce a quantifiable and comparable feedback evaluation index, based on the completion of the sample set construction and the identification of the completion behavior characteristic indicators corresponding to different completion execution subjects. This index provides a unified measurement of the performance of the completion results in the anomaly identification stage, thereby identifying the degree of adaptation of different completion behavior characteristics under different remaining time conditions. It should be noted that the feedback evaluation index itself is not a completely new concept, but rather originates from commonly used evaluation systems in the prior art for the stability, consistency, and volatility of anomaly identification results. The innovation of this invention lies not in proposing the evaluation index, but in jointly analyzing it with the completion behavior characteristic indicators and the variable of remaining time.

[0048] Step S300 directly addresses the core research objective of this invention: given a fixed completion behavior characteristic index, it examines how the feedback performance of the completion result changes during the anomaly identification stage when the variable of remaining time changes, thereby identifying the time condition most suitable for this type of individual difference expression. In other words, for each sample set, the corresponding completion execution subject and completion behavior characteristic index are fixed, while the only difference between different samples is the remaining time. By comparing the changes in the feedback evaluation index under different remaining times, it can be identified within which remaining time range the completion result corresponding to the completion behavior characteristic index performs optimally in the anomaly identification stage. The remaining time interval corresponding to this range is the time interval most suitable for the completion execution subject. This analytical process directly echoes the core research point proposed in step S100: "different individual differences adapt to different anomaly identification time distances."

[0049] In step S301, the specific implementation method is as follows: Based on a preset historical database, the actual processing results of the corresponding samples in each sample set during the anomaly identification stage are read, and the corresponding feedback evaluation index is calculated according to predefined evaluation rules. Subsequently, the samples in the same sample set are arranged in an ordered manner based on the remaining time corresponding to each sample, forming an ordered sequence of samples that monotonically changes along the direction of remaining time. Through this ordered sequence, a one-to-one correspondence can be established between the feedback evaluation index and the remaining time, providing basic data for subsequent trend analysis. This process can be implemented through database query, timestamp matching, and sorting algorithms, which are conventional methods in existing data processing technologies.

[0050] In step S302, the trend of the feedback evaluation index in the ordered sequence of samples as the remaining time changes is analyzed, with a focus on determining whether there is an interval where the index changes from rising to falling. Using "rising to falling" as the criterion for determining the optimal interval has clear rationality. On the one hand, when the remaining time is short, the completed data almost immediately enters the anomaly identification stage, and its impact on anomaly identification is not yet fully reflected, and the stability advantage brought by the completion behavior feature is not fully realized. As the remaining time increases, the completed data participates more fully in subsequent processing for statistical and feature calculations, and the advantage of adapting to the completion behavior feature gradually emerges, resulting in an upward trend in the feedback evaluation index. On the other hand, when the remaining time further increases, the cumulative effect of the prediction error of the completed data, environmental changes, or other time-related factors begin to dominate, causing the subtle differences introduced in the completion stage to be excessively amplified, thereby weakening the stability of the anomaly identification stage, and the feedback evaluation index subsequently declines. Therefore, from the underlying logic, the feedback evaluation index shows a change pattern of first rising and then falling with the remaining time, reflecting the objective process of the advantages of the completion behavior characteristics from "not fully utilized" to "optimal utilization" and then to "gradually being offset by other uncertainties", rather than a simple monotonic change.

[0051] Based on the above analysis, in step S303, when a feedback evaluation index interval that changes from rising to falling is detected, this interval is determined as the optimal feedback evaluation index interval, and the corresponding remaining time interval is further obtained. This remaining time interval indicates that within this time range, the completion behavior feature index corresponding to the current sample set can achieve the most stable and consistent feedback performance in the anomaly identification stage.

[0052] The rationality of the process for obtaining the feedback evaluation indicators lies in the high degree of consistency between the evaluation object and the evaluation objective. The anomaly judgment consistency value reflects whether the anomaly identification results remain consistent under similar time or conditions, directly reflecting the impact of supplementary data on the stability of the anomaly identification results. The anomaly triggering stability value describes the stability of the anomaly identification triggering frequency over time, reflecting whether the supplementary data introduces unnecessary noise. The anomaly feature fluctuation suppression value measures the fluctuation changes of anomaly identification features after the participation of supplementary data, reflecting the strength of the impact of the supplementary results on the anomaly identification features. Using at least one of the above values ​​or their weighted result as feedback evaluation indicators allows for a comprehensive characterization of the supplementary results' performance in the anomaly identification stage from multiple complementary perspectives. Furthermore, a larger value for the feedback evaluation indicator indicates higher stability, facilitating ranking and trend analysis.

[0053] It should be noted that, apart from the specific implementation methods described above, this invention does not limit the specific composition of the feedback evaluation indicators. In other embodiments, methods such as the false alarm rate of anomaly identification, the consistency ratio of anomaly identification results, and the reciprocal of the variance of anomaly scores can also be used to quantify the stability of the anomaly identification stage. As long as the evaluation method used can reflect the impact of the completion results on the stability of anomaly identification processing, it falls within the protection scope of this invention.

[0054] Furthermore, the power transaction data end-to-end standardization processing method also includes the following steps: Step S400, when different sample sets correspond to different remaining time intervals, determine the remaining time interval corresponding to the completion behavior characteristic index of the target completion execution subject, and when the target completion execution subject performs subsequent missing data completion processing consistent with the business type, prioritize the allocation of data completion tasks whose predicted remaining time falls into the remaining time interval to the target completion execution subject.

[0055] Specifically, Figure 3 shows a flowchart of the priority allocation of data completion tasks.

[0056] Specifically, when different sample sets correspond to different remaining time intervals, the remaining time interval corresponding to the completion behavior characteristic index of the target completion execution subject is determined. When the target completion execution subject performs subsequent missing data completion processing consistent with the business type, the data completion task whose predicted remaining time falls within that remaining time interval is preferentially assigned to the target completion execution subject. This includes the following steps: Step S401, when it is determined that sample sets belonging to different completion behavior characteristic indicators correspond to different remaining time intervals that optimize the feedback evaluation index, the relationship between the completion behavior characteristic indicator and the remaining time interval is determined. In the adaptation relationship; step S402, based on the adaptation relationship, determine a sample set that is the same as or within the preset allowable error range of the completion behavior characteristic index of the target completion execution subject, and determine the remaining time interval corresponding to the sample set as the adaptation type remaining time interval of the target completion execution subject; step S403, when the target completion execution subject performs subsequent missing data completion processing consistent with the business type, predict the remaining time of the data completion task to be processed, and when the predicted remaining time falls into the adaptation type remaining time interval, prioritize the allocation of the data completion task to the target completion execution subject.

[0057] In this embodiment of the invention, the purpose of step S400 is to apply the conclusions obtained from the historical sample analysis to the actual data completion task allocation process, thereby utilizing the individual differences of different completion execution subjects, rather than merely remaining at the analysis level. Through the processing of steps S200 and S300, the optimal remaining time intervals corresponding to different sample sets can be obtained based on historical data. Since the number of historical samples is sufficient and the coverage of scenarios is wide, in practical applications, sample sets that are highly similar to or basically consistent with the completion behavior characteristic indicators of the target completion execution subject can usually be found in the sample set, thereby determining an adaptive remaining time interval with practical guiding significance for the target completion execution subject.

[0058] Step S400 directly echoes the core research point raised in step S100, namely: the individual differences formed by different completion execution subjects during the completion stage are not manifested as significant superiority or inferiority in the completion stage itself, but rather as differences in their adaptability to different anomaly identification time distances. By introducing the dimension of remaining time in the task allocation stage and matching it with completion behavior characteristic indicators, the completion execution subjects can complete the completion task under the time conditions that are more advantageous, thereby improving the overall stability of the subsequent anomaly identification stage.

[0059] In step S401, a crucial determination is first made: whether there exists a situation where sample sets belonging to different completion behavior characteristic indicators correspond to different remaining time intervals that optimize the feedback evaluation indicator. This determination is significant. On the one hand, if different sample sets correspond to the same optimal remaining time interval despite having different completion behavior characteristic indicators, it indicates that the impact of the remaining time on the feedback evaluation indicator is unrelated to the individual differences of the completion execution subject, and the setting of the completion behavior characteristic indicators would lose its distinguishing meaning. On the other hand, when the determination result shows that different completion behavior characteristic indicators do indeed correspond to different optimal remaining time intervals, it indirectly verifies the core finding in step S100, namely, that individual differences of the completion execution subject objectively exist, and these differences are amplified and manifested through interaction with the time distance in the anomaly identification stage. Therefore, step S401 not only establishes subsequent allocation rules but also effectively verifies the core research conclusions.

[0060] In step S402, based on the established adaptation relationship between the completion behavior feature indicators and the remaining time interval, a sample set that is identical to or within a preset allowable error range of the completion behavior feature indicators of the target completion execution subject is determined using feature similarity matching. This process can be implemented through feature distance calculation, threshold judgment, or similarity matching algorithms, and is a conventional method in existing data analysis techniques. Since the completion behavior feature indicators are stable features formed based on long-term historical completion results statistics, when the number of sample sets is sufficient, a sample set that is highly similar to the characteristics of the target completion execution subject can usually be found, and the remaining time interval corresponding to this sample set is determined as the adapted remaining time interval of the target completion execution subject.

[0061] In step S403, when the target completion execution entity performs subsequent missing data completion processing consistent with the aforementioned business type, the remaining time for the data completion task to be processed is first predicted. It should be noted that this remaining time can be accurately predicted because the processing chain of power transaction data is usually transparent and pre-defined. After completion processing, the trigger time for the data to enter the anomaly identification processing flow can be directly determined through system scheduling rules, processing queue status, or time plan. Therefore, the remaining time between the completion completion time and the anomaly identification trigger time can be accurately known during the completion processing stage. After the remaining time prediction is completed, when the predicted remaining time falls within the adaptive remaining time interval corresponding to the target completion execution entity, the system will prioritize assigning the data completion task to the target completion execution entity, thereby enabling it to complete the completion operation under time conditions more suitable for its own completion behavior characteristics.

[0062] Through the above steps, this invention forms a complete closed loop from individual difference identification and historical pattern mining to actual task scheduling. Compared with the existing data completion task allocation methods based on balanced or random allocation, this invention does not require changing the operating habits or completion methods of the completion execution subjects. Instead, it fully utilizes the individual advantages of different completion execution subjects through reasonable matching of tasks and time conditions, thereby improving the stability and consistency of the anomaly identification stage without increasing additional labor costs and system complexity.

[0063] From an overall beneficial perspective, this invention effectively addresses the core research problem raised in step S100: the amplification of subtle individual differences introduced during the completion stage in the subsequent anomaly identification stage, with the degree of amplification influenced by time distance. By identifying and utilizing the adaptation relationship between completion behavior characteristic indicators and remaining time, the instability and misjudgment risk in the anomaly identification stage can be significantly reduced. This method is suitable for large-scale, batch data completion scenarios in power trading systems, possesses good scalability and engineering feasibility, and has broad application prospects in areas such as power market operation monitoring, transaction risk control, and automated data processing.

[0064] Furthermore, Figure 3 shows the application architecture diagram of the system provided in the embodiment of the present invention.

[0065] In another preferred embodiment of the present invention, a standardized processing system for the entire power transaction data chain is provided. The system includes a behavior feature acquisition module 100, which is used to determine the business type of the power transaction declaration time series when the target completion execution entity performs completion processing on the missing data of the current power transaction declaration time series, and to acquire the completion behavior feature indicators formed by the target completion execution entity when processing the business type based on a preset historical database.

[0066] The process of obtaining the completion behavior feature indicators includes: based on a preset historical database, extracting at least one of the following feature indicators from the completion result data generated by the completion execution subject during the historical missing data completion process: completion value bias feature indicator, used to characterize the degree of bias of the completion result relative to the data at adjacent time points before and after the missing data; completion smoothness feature indicator, used to characterize the degree of suppression or amplification of time series fluctuations by the completion result; completion stability feature indicator, used to characterize the stability of the change amplitude of the completion result within a continuous time period; and performing comprehensive calculation based on the at least one feature indicator to obtain the completion behavior feature indicators characterizing the completion behavior features of the completion execution subject.

[0067] The sample set construction module 200 is used to construct several sample sets based on a preset historical database. Each sample set corresponds to a different completion execution entity and its completion behavior characteristic indicators, and the business types of each sample set are consistent. Each sample set contains the same number of samples, and different samples correspond to different remaining times to reach the anomaly identification stage. When constructing the sample sets, different completion execution entities process samples in the corresponding sample sets using the same data completion processing scenario as the current one. This consistency means that the power transaction declaration time series corresponding to each sample has the same time granularity and consistent anomaly identification processing logic, and the missing location and length of the missing data are within a preset comparable range.

[0068] The interval analysis module 300 is used to obtain the feedback evaluation index generated by each sample in the anomaly identification stage, sort the samples in the sample set according to the remaining time, analyze the trend of the feedback evaluation index with the remaining time, and determine whether there is a remaining time interval that makes the feedback evaluation index reach the optimal level.

[0069] The task allocation module 400 is used to determine the remaining time interval corresponding to the completion behavior feature index of the target completion execution subject when different sample sets correspond to different remaining time intervals. When the target completion execution subject performs subsequent missing data completion processing consistent with the business type, the data completion task whose predicted remaining time falls into the remaining time interval is preferentially allocated to the target completion execution subject.

[0070] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0071] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, storage, preset historical databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0072] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0073] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0074] 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, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A standardized processing method for the entire power trading data chain, characterized in that, The method includes: when the target completion execution entity performs completion processing on the missing data of the current power transaction declaration time series, determining the business type of the power transaction declaration time series, and obtaining the completion behavior characteristic indicators formed by the target completion execution entity when processing the business type based on a preset historical database; constructing several sample sets based on the preset historical database, each sample set corresponding to a different completion execution entity and its completion behavior characteristic indicators, and the business types of each sample set being consistent, wherein each sample set contains the same number of samples, and different samples correspond to different remaining times to reach the anomaly identification stage; obtaining the feedback evaluation indicators generated by each sample in the anomaly identification stage, and sorting the samples in the sample set according to the remaining time, analyzing the trend of the feedback evaluation indicators changing with the remaining time to determine whether there is a remaining time interval that makes the feedback evaluation indicators reach the optimal level; when different sample sets correspond to different remaining time intervals, determining the remaining time interval corresponding to the completion behavior characteristic indicators of the target completion execution entity, and when the target completion execution entity performs subsequent missing data completion processing consistent with the business type, prioritizing the allocation of data completion tasks whose predicted remaining time falls into the remaining time interval to the target completion execution entity.

2. The method for standardized processing of end-to-end power trading data according to claim 1, characterized in that, The process of obtaining the completion behavior feature indicators includes: based on a preset historical database, extracting at least one of the following feature indicators from the completion result data generated by the completion execution subject during the historical missing data completion process: completion value bias feature indicator, used to characterize the degree of bias of the completion result relative to the data at adjacent time points before and after the missing data; completion smoothness feature indicator, used to characterize the degree of suppression or amplification of time series fluctuations by the completion result; completion stability feature indicator, used to characterize the stability of the change amplitude of the completion result within a continuous time period; and performing comprehensive calculation based on the at least one feature indicator to obtain the completion behavior feature indicators characterizing the completion behavior features of the completion execution subject.

3. The method for standardized processing of end-to-end power trading data according to claim 1, characterized in that, When constructing the sample set, different completion execution entities process the data completion scenarios of the corresponding sample sets in the same way as the current ones. The data completion scenarios being consistent with the current ones means that the power transaction declaration time series corresponding to each sample has the same time granularity and consistent anomaly identification and processing logic, and the missing location and missing length of the missing data are within a preset comparable range.

4. The method for standardized processing of end-to-end power trading data according to claim 1, characterized in that, The remaining time to reach the anomaly identification stage refers to the time interval from the completion of the missing data completion process to the trigger time when the corresponding power transaction declaration time series enters the anomaly identification process.

5. The method for standardized processing of end-to-end power trading data according to claim 1, characterized in that, The steps of obtaining feedback evaluation indicators generated by each sample during the anomaly identification stage, sorting the samples in the sample set according to the remaining time, and analyzing the trend of the feedback evaluation indicators with the remaining time to determine whether there is a remaining time interval that makes the feedback evaluation indicators reach the optimal level include: obtaining the feedback evaluation indicators generated by the samples in each sample set during the anomaly identification stage based on a preset historical database, and arranging the samples in each sample set in an ordered sequence according to the remaining time to form an ordered sequence of samples; analyzing the trend of the feedback evaluation indicators in the ordered sequence of samples in each sample set with the remaining time, and determining whether there is a feedback evaluation indicator interval that changes from rising to falling in the trend; when it is determined that the feedback evaluation indicator interval exists, the feedback evaluation indicator interval is determined as the optimal feedback evaluation indicator interval, and the corresponding remaining time interval is obtained.

6. The method for standardized processing of end-to-end power trading data according to claim 5, characterized in that, The process of obtaining the feedback evaluation index includes: based on a preset historical database, obtaining the processing results of each sample in the anomaly identification stage; calculating at least one of the following values ​​in the anomaly judgment consistency value, anomaly trigger stability value, and anomaly characteristic fluctuation suppression value generated in the anomaly identification stage based on the processing results; and using the at least one value or its weighted result as the feedback evaluation index, wherein the larger the value of the feedback evaluation index, the higher the stability of the corresponding completion result for anomaly identification processing.

7. The method for standardized processing of end-to-end power trading data according to claim 6, characterized in that, When different sample sets correspond to different remaining time intervals, the steps of determining the remaining time interval corresponding to the completion behavior characteristic index of the target completion execution entity, and prioritizing the allocation of data completion tasks whose predicted remaining time falls within the remaining time interval to the target completion execution entity when the target completion execution entity performs subsequent missing data completion processing consistent with the business type, include: when it is determined that sample sets belonging to different completion behavior characteristic indices correspond to different remaining time intervals that optimize the feedback evaluation index, a matching relationship is determined between the completion behavior characteristic indices and the remaining time intervals; based on the matching relationship, a sample set that is the same as or within the preset allowable error range of the completion behavior characteristic index of the target completion execution entity is determined, and the remaining time interval corresponding to the sample set is determined as the matching remaining time interval of the target completion execution entity; when the target completion execution entity performs subsequent missing data completion processing consistent with the business type, the remaining time of the data completion task to be processed is predicted, and when the predicted remaining time falls within the matching remaining time interval, the data completion task is prioritized and allocated to the target completion execution entity.

8. A standardized processing system for the entire power trading data chain, characterized in that, The system includes: a behavior feature acquisition module, used to determine the business type of the power transaction declaration time series when the target completion execution entity performs completion processing on the missing data of the current power transaction declaration time series, and to acquire completion behavior feature indicators formed by the target completion execution entity when processing the business type based on a preset historical database; a sample set construction module, used to construct several sample sets based on the preset historical database, each sample set corresponding to a different completion execution entity and its completion behavior feature indicators, and the business type of each sample set is consistent, wherein each sample set contains the same number of samples, and different samples correspond to different remaining times to reach the anomaly identification stage; interval division. The analysis module is used to obtain the feedback evaluation indicators generated by each sample during the anomaly identification stage, sort the samples in the sample set according to the remaining time, analyze the trend of the feedback evaluation indicators with the remaining time, and determine whether there is a remaining time interval that makes the feedback evaluation indicators optimal. The task allocation module is used to determine the remaining time interval corresponding to the completion behavior characteristic indicators of the target completion execution subject when different sample sets correspond to different remaining time intervals, and when the target completion execution subject performs subsequent missing data completion processing consistent with the business type, the data completion tasks whose predicted remaining time falls into the remaining time interval are preferentially allocated to the target completion execution subject.

9. The power transaction data end-to-end standardized processing system according to claim 8, characterized in that, The process of obtaining the completion behavior feature indicators includes: based on a preset historical database, extracting at least one of the following feature indicators from the completion result data generated by the completion execution subject during the historical missing data completion process: completion value bias feature indicator, used to characterize the degree of bias of the completion result relative to the data at adjacent time points before and after the missing data; completion smoothness feature indicator, used to characterize the degree of suppression or amplification of time series fluctuations by the completion result; completion stability feature indicator, used to characterize the stability of the change amplitude of the completion result within a continuous time period; and performing comprehensive calculation based on the at least one feature indicator to obtain the completion behavior feature indicators characterizing the completion behavior features of the completion execution subject.

10. The power trading data end-to-end standardized processing system according to claim 9, characterized in that, When constructing the sample set, different completion execution entities process the data completion scenarios of the corresponding sample sets in the same way as the current ones. The data completion scenarios being consistent with the current ones means that the power transaction declaration time series corresponding to each sample has the same time granularity and consistent anomaly identification and processing logic, and the missing location and missing length of the missing data are within a preset comparable range.

Images