Power grid supply chain index management system

Abstract

The invention provides a power grid supply chain index management system, and relates to the technical field of power grid management. Comprising a historical data module used for collecting and storing historical index data of all levels, an index obtaining module used for obtaining index data of all levels, a first monitoring module used for conducting first-time anomaly detection and key index detection on the index data of all levels according to the historical data to obtain detection data, and a second monitoring module used for conducting second-time anomaly detection on the index data of all levels according to the detection data. The second monitoring module is used for monitoring whether new data exists in the detection data or not according to the historical data module, if yes, the new data are transmitted to the historical data module and updated, and the early warning module is used for conducting second-time anomaly detection on the detection data to obtain second-time detection data; and the sharing platform is used for publicizing the second detection data and carrying out early warning. According to the invention, the problem of low management efficiency caused by incapability of unified analysis of power grid material supply chain index fragmentation and incapability of timely response of supply chain data in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid management, and in particular to a power grid supply chain indicator management system. Background Art

[0002] The current power grid material supply chain is facing challenges such as complex indicators, inefficient management, data fragmentation, and difficulty in unified data analysis. Existing technologies often use manual statistics of operating indicators, which are reported and sorted out at the provincial, municipal and county levels, and are mostly focused on local links, lacking overall vision, system logic and intelligent systems. The power grid supply chain operating indicators involve performance appraisal, peer benchmarking, and professional evaluation systems, covering the ten major businesses of the supply chain. Existing technologies are difficult to cope with the complex and changeable management needs of power grid material supply chain indicators; at the same time, many existing monitoring systems cannot achieve real-time data analysis and cannot quickly respond to problems in the supply chain. In summary, the power grid supply chain indicator management system in the existing technology is inefficient due to the scattered indicator data, insufficient monitoring and lack of standardized indicator definitions. Summary of the invention

[0003] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a power grid supply chain indicator management system, which solves the problem in the prior art that the fragmentation of power grid material supply chain indicators cannot be uniformly analyzed and the supply chain data cannot be responded to in a timely manner, resulting in low management efficiency.

[0004] To achieve the above object, the present invention provides the following solutions:

[0005] A power grid supply chain indicator management system, comprising:

[0006] A historical data module and an indicator acquisition module, a first monitoring module, a second monitoring module, an early warning module and a sharing platform connected in sequence;

[0007] The historical data module is connected to the first monitoring module, the second monitoring module and the early warning module respectively;

[0008] The historical data module is used to collect and store historical indicator data at all levels, the indicator acquisition module is used to acquire indicator data at each level, the first monitoring module is used to perform a first anomaly detection and a key indicator detection on the indicator data at each level according to the historical data, and obtain first normal indicator data, first key indicator data and first anomaly data, the second monitoring module is used to monitor whether there is new data in the first normal indicator data, the first anomaly data and the first key indicator data according to the historical data module, and if so, the new data is transferred to the historical data module and updated, the early warning module is used to perform a second anomaly detection on the first normal indicator data, the first anomaly data and the first key indicator data, and obtain second key indicator data, second anomaly data and second normal indicator data, and the sharing platform is used to publicize the second key indicator data, the second anomaly data and the second normal indicator data and issue an early warning.

[0009] Preferably, the first monitoring module includes:

[0010] A model building submodule and a model training submodule and a model updating submodule, both of which are connected to the model building submodule;

[0011] The model building submodule is used to build a detection model based on historical data, the model training submodule is used to input the indicator data of each level into the detection model to obtain normal indicator data, key indicator data and abnormal data, and the model updating submodule is used to update the detection model in real time according to the new data.

[0012] Preferably, the model building submodule includes:

[0013] A data acquisition unit, used to acquire historical indicator data and use the historical indicator data as original training time series data samples;

[0014] A first preprocessing unit, used for performing data preprocessing on the original training time series data samples to obtain training data samples;

[0015] A first model building unit, used to build an initial LSTM-AE model using an LSTM deep model;

[0016] A second model construction unit is used to reconstruct the initial LSTM-AE model using the training data set, optimize the network weight value of the initial LSTM-AE model, and obtain the trained LSTM-AE model and the error sequence of the training data sample;

[0017] The third model building unit is used to obtain an adaptive threshold formula by using the error sequence of the training data set, and build a first detection model according to the adaptive threshold formula.

[0018] Preferably, the expression of the adaptive threshold formula is:

[0019] TH = μ + α · 0.8;

[0020] μ is the mean of the error sequence, which indicates the average level of error under normal conditions, and α is the standard deviation of the error sequence, which indicates the degree of error fluctuation.

[0021] Preferably, the second monitoring module includes:

[0022] An indicator classification submodule, used to classify the first normal indicator data, the first key indicator data and the first abnormal data to obtain classified data, wherein the classified data includes: indicator name, indicator definition, calculation logic, indicator category, related major, indicator level, release frequency, responsible department, indicator unit, and early warning rules;

[0023] A feature extraction submodule is used to extract features from the classified data to obtain features of various types of indicator data;

[0024] The importance assessment submodule is used to perform importance assessment on the characteristics of the various indicator data according to the LIME algorithm to obtain importance data;

[0025] A clustering submodule, used to cluster the historical indicator data to obtain cluster data;

[0026] The judgment submodule is used to judge whether new data is generated according to the clustering data and the importance data, and if so, obtain new data.

[0027] Preferably, the judgment submodule includes:

[0028] A second preprocessing unit, used for performing standardization processing on the importance data and the clustering data to obtain preprocessed importance data and clustering data;

[0029] A fusion unit, used for performing feature fusion on the pre-processed importance and clustering data to obtain comprehensive data;

[0030] A threshold setting unit, used to set a new data generation threshold;

[0031] A multi-model building unit, used to build multiple discriminant models;

[0032] The comprehensive judgment unit is used to use the multiple discrimination models to judge the comprehensive data according to the threshold value generated by the new data to obtain a judgment result.

[0033] Preferably, the sharing platform includes:

[0034] A wireless communication submodule and several multi-level display submodules;

[0035] The wireless communication submodule is used to transmit the second key indicator data, the second abnormal data and the second normal indicator data to the multi-level display value module, and the multi-level display submodule is used to publicize the second key indicator data, the second abnormal data and the second normal indicator data and mark the second abnormal data.

[0036] The present invention discloses the following technical effects:

[0037] The present invention provides an indicator management system for a power supply chain, comprising: a historical data module and an indicator acquisition module, a first monitoring module, a second monitoring module, an early warning module and a sharing platform connected in sequence; the historical data module is connected to the first monitoring module, the second monitoring module and the early warning module respectively; the historical data module is used to collect and store historical indicator data of each level, the indicator acquisition module is used to acquire indicator data of each level, the first monitoring module is used to perform a first abnormality detection and a key indicator detection on the indicator data of each level according to the historical data, and obtain first normal indicator data, first key indicator data and first abnormal data, the second monitoring module is used to monitor whether there is new data in the first normal indicator data, the first abnormal data and the first key indicator data according to the historical data module, and if so, the new data is transmitted to the historical data module and updated, the early warning module is used to perform a second abnormality detection on the first normal indicator data, the first abnormal data and the first key indicator data, and obtain second key indicator data, second abnormal data and second normal indicator data, and the sharing platform is used to publicize the second key indicator data, the second abnormal data and the second normal indicator data and issue an early warning. The systematic design, hierarchical monitoring mechanism and timely early warning mechanism of the present invention use the monitoring model to self-update historical data, which not only significantly improves the abnormality detection capability and data management level of the power grid supply chain, but also provides strong support for decision-making, fully demonstrating its security and through centralized storage, standardized format, real-time integration, hierarchical monitoring and data quality improvement, it not only effectively reduces the degree of data fragmentation, but also improves the overall data utilization efficiency, providing reliable information support for decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0039] Figure 1 A schematic diagram of the structure of a power grid supply chain indicator management system provided by an embodiment of the present invention.

[0040] Description of reference numerals:

[0041] 1-historical data module, 2-indicator acquisition module, 3-first monitoring module, 4-second monitoring module, 5-early warning module, 6-sharing platform. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] like Figure 1 As shown, the present invention provides a power grid supply chain indicator management system, comprising:

[0045] A historical data module 1 and an indicator acquisition mode 2, a first monitoring module 3, a second monitoring module 4, an early warning module 5 and a sharing platform 6 connected in sequence;

[0046] The historical data module 1 is connected to the first monitoring module 3, the second monitoring module 4 and the early warning module 5 respectively;

[0047] The historical data module 1 is used to collect and store historical indicator data at all levels, the indicator acquisition mode 2 is used to acquire indicator data at each level, the first monitoring module 3 is used to perform a first anomaly detection and a key indicator detection on the indicator data at each level according to the historical data, and obtain first normal indicator data, first key indicator data and first anomaly data, the second monitoring module 4 is used to monitor whether there is new data in the first normal indicator data, the first anomaly data and the first key indicator data according to the historical data module 1, and if so, the new data is transferred to the historical data module 1 and updated, the early warning module 5 is used to perform a second anomaly detection on the first normal indicator data, the first anomaly data and the first key indicator data, and obtain second key indicator data, second anomaly data and second normal indicator data, and the sharing platform 6 is used to publicize the second key indicator data, the second anomaly data and the second normal indicator data and issue an early warning.

[0048] Specifically, indicator acquisition mode 2 obtains basic data through data middle platforms at different levels to support the calculation and analysis of operational indicators. Specifically, the data middle platform provides basic data support such as indicator access, calculation, decomposition, and traceability. If it involves cross-level data, it is obtained through vertical exchange between two-level data middle platforms. The low level receives high-level indicator warnings and supervision orders, and feedbacks on problem verification and disposal.

[0049] Furthermore, the first monitoring module 3 includes:

[0050] A model building submodule and a model training submodule and a model updating submodule, both of which are connected to the model building submodule;

[0051] The model building submodule is used to build a detection model based on historical data, the model training submodule is used to input the indicator data of each level into the detection model to obtain normal indicator data, key indicator data and abnormal data, and the model updating submodule is used to update the detection model in real time according to the new data.

[0052] Furthermore, the model building submodule includes:

[0053] A data acquisition unit, used to acquire historical indicator data and use the historical indicator data as original training time series data samples;

[0054] A first preprocessing unit, used for performing data preprocessing on the original training time series data samples to obtain training data samples;

[0055] Specifically, before building the model, a data cleaning process is set up to deduplicate historical data, fill in missing values, remove outliers, etc. to ensure the quality of the training data.

[0056] A first model building unit, used to build an initial LSTM-AE model using an LSTM deep model;

[0057] A second model construction unit is used to reconstruct the initial LSTM-AE model using the training data set, optimize the network weight value of the initial LSTM-AE model, and obtain the trained LSTM-AE model and the error sequence of the training data sample;

[0058] The third model building unit is used to obtain an adaptive threshold formula by using the error sequence of the training data set, and build a first detection model according to the adaptive threshold formula.

[0059] Furthermore, the expression of the adaptive threshold formula is:

[0060] TH = μ + α · 0.8;

[0061] μ is the mean of the error sequence, which indicates the average level of error under normal conditions, and α is the standard deviation of the error sequence, which indicates the degree of error fluctuation.

[0062] The formula includes a training data reconstruction error parameter. By assigning a numerical value to the training data reconstruction error parameter, the value of the adaptive threshold can be obtained. The value of the adaptive threshold is used to judge the outlier value and complete the construction of the anomaly detection model.

[0063] Specifically, indicator definition and threshold setting: Based on the analysis results of historical data, this module will determine key indicators and their thresholds in order to monitor their performance.

[0064] Real-time data monitoring: Monitor the real-time data of key indicators, compare the real-time data with the preset thresholds, and detect abnormal situations in time.

[0065] Specifically, based on the frequency of indicator calculation, monitor the status of each link of indicator data access, aggregation and result calculation, compare the consistency of data before and after calculation, and ensure the correctness of the results: display the statistical results of each operating indicator, compare and verify with the intermediate summary data, and use automatic verification and manual verification modes to verify the indicator calculation results.

[0066] By analyzing the distribution and historical data of supply chain operation indicators at each level, abnormal indicators are found. For indicators composed of multiple levels, the weight of the impact of the indicator composition on the final result is evaluated, the importance of the impact is ranked, and the difference between abnormal indicators and normal indicators of users at the same level is analyzed to find the key factors that cause abnormal indicators.

[0067] Based on the historical data of supply chain operation indicators, the indicator change trend is predicted, and the fluctuation range of a certain period in the future is determined. According to the different levels and professions of supply chain operation indicators, the indicator analysis report template customization function is provided, and the template upload, download and update functions are provided. The generation and release of the above-mentioned operation analysis report includes: according to the format, operation indicators and statistical time range defined by the template, the indicator calculation results are automatically summarized, the analysis report is generated, and the operation analysis report review, release and online viewing functions are provided.

[0068] Specifically, the business objectives of the power grid supply chain should be clearly defined, such as improving operational efficiency, reducing downtime, reducing operating costs, and improving customer satisfaction. Key indicators should be directly related to these objectives.

[0069] Stakeholder needs: Understand the needs of different stakeholders (such as management, operation team, etc.) for data and results, and ensure that the selected indicators can provide value to them; use the historical indicator data module to conduct data mining and analysis to identify which factors have had a significant impact on efficiency and effectiveness in past performance; evaluate data from different time periods (such as years, quarters, months), identify daily and seasonal fluctuations, and mine possible important indicators.

[0070] Alarm and notification: When the indicator exceeds the set threshold, the alarm mechanism is automatically triggered and the relevant personnel are notified to handle it to ensure the efficient operation of the supply chain.

[0071] Furthermore, the second monitoring module 4 includes:

[0072] An indicator classification submodule, used to classify the first normal indicator data, the first key indicator data and the first abnormal data to obtain classified data, wherein the classified data includes: indicator name, indicator definition, calculation logic, indicator category, related major, indicator level, release frequency, responsible department, indicator unit, and early warning rules;

[0073] A feature extraction submodule is used to extract features from the classified data to obtain features of various types of indicator data;

[0074] The importance assessment submodule is used to perform importance assessment on the characteristics of the various indicator data according to the LIME algorithm to obtain importance data;

[0075] A clustering submodule, used to cluster the historical indicator data to obtain cluster data;

[0076] The judgment submodule is used to judge whether new data is generated according to the clustering data and the importance data, and if so, obtain new data.

[0077] Furthermore, the judgment submodule includes:

[0078] A second preprocessing unit, used for performing standardization processing on the importance data and the clustering data to obtain preprocessed importance data and clustering data;

[0079] A fusion unit, used for performing feature fusion on the pre-processed importance and clustering data to obtain comprehensive data;

[0080] A threshold setting unit, used to set a new data generation threshold;

[0081] A multi-model building unit, used to build multiple discriminant models;

[0082] The comprehensive judgment unit is used to use the multiple discrimination models to judge the comprehensive data according to the threshold value generated by the new data to obtain a judgment result.

[0083] Specifically, the weighted decision method is used to comprehensively consider the output results of multiple discriminant models. Rules can be set, such as increasing the credibility of the judgment when the judgment results of multiple models are consistent.

[0084] At the same time, an anomaly detection mechanism is introduced to analyze abnormal signals in the judgment process and enhance the reliability of judgment.

[0085] More specifically, the judgment results of all models are collected to form a result list. For example, assuming there are four models, the results may be:

[0086] Model 1: With new data

[0087] Model 2: With new data

[0088] Model 3: No new data

[0089] Model 4: No new data

[0090] Implement a simple voting (majority voting) mechanism. Count the votes for "new data" and "no new data":

[0091] Number of votes for "new data" = 2

[0092] Votes for "No new data" = 2

[0093] In this example, since the two are evenly matched, we can proceed to the next step.

[0094] If each model has a different weight (based on historical performance, model complexity, etc.), the voting results can be weighted. In this case, assume that the weights of each model are as follows:

[0095] Model 1 (high confidence): weight = 0.4

[0096] Model 2 (high confidence): weight = 0.4

[0097] Model 3 (medium credibility): weight = 0.1

[0098] Model 4 (medium credibility): weight = 0.1

[0099] Weighted result calculation:

[0100] Weighted score of "new data" = (0.4 + 0.4) = 0.8

[0101] Weighted score for "no new data" = (0.1 + 0.1) = 0.2

[0102] In this case, based on the weighted voting results, the judgment that "there is new data" is more credible.

[0103] Furthermore, the sharing platform 6 includes:

[0104] A wireless communication submodule and several multi-level display submodules;

[0105] The wireless communication submodule is used to transmit the second key indicator data, the second abnormal data and the second normal indicator data to the multi-level display value module, and the multi-level display submodule is used to publicize the second key indicator data, the second abnormal data and the second normal indicator data and mark the second abnormal data.

[0106] Specifically, visualization tools such as dashboards and charts are introduced into the sharing platform 6, so that users can intuitively view indicators and monitoring results, thereby improving user experience.

[0107] Provides flexible data filtering and custom view functions, allowing users to view different indicator combinations as needed.

[0108] More specifically, the sharing platform 6 is also equipped with an early warning mechanism, which issues special annotations for abnormal data so that people can see it at a glance.

[0109] More specifically, the overview of supply chain operation indicators in all links supports multi-dimensional evaluation points and ranking operations, which directly reflects the shortcomings of the supply chain. The visualization interface of each professional focus indicator focuses on the scores and change trends of related indicators in the profession, analyzes the factors of indicator changes, and locates the problem to the responsible person. The user interface of each responsible person can view the scores and change trends of supply chain operation indicators, analyze the reasons for changes, and guide users to optimize business work arrangements;

[0110] Based on the alarm rules of operational indicators, SMS, email or contact form will be automatically sent to the person responsible for the indicator to give an alarm reminder;

[0111] A work order is generated based on the indicator warning information and sent to relevant units for confirmation, rectification and feedback. For warning work orders with higher importance levels, the superior unit or the responsible business department will upgrade them through supervision orders to complete the closed loop.

[0112] The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.

[0113] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A power supply chain indicator management system, characterized in that: include: A historical data module and an indicator acquisition module, a first monitoring module, a second monitoring module, an early warning module and a sharing platform connected in sequence; The historical data module is connected to the first monitoring module, the second monitoring module and the early warning module respectively; The historical data module is used to collect and store historical indicator data at all levels, the indicator acquisition module is used to acquire indicator data at each level, the first monitoring module is used to perform a first anomaly detection and a key indicator detection on the indicator data at each level according to the historical data, and obtain first normal indicator data, first key indicator data and first anomaly data, the second monitoring module is used to monitor whether there is new data in the first normal indicator data, the first anomaly data and the first key indicator data according to the historical data module, and if so, the new data is transferred to the historical data module and updated, the early warning module is used to perform a second anomaly detection on the first normal indicator data, the first anomaly data and the first key indicator data, and obtain second key indicator data, second anomaly data and second normal indicator data, and the sharing platform is used to publicize the second key indicator data, the second anomaly data and the second normal indicator data and issue an early warning.

2. A power supply chain indicator management system according to claim 1, characterized in that: The first monitoring module includes: A model building submodule and a model training submodule and a model updating submodule, both of which are connected to the model building submodule; The model building submodule is used to build a detection model based on historical data, the model training submodule is used to input the indicator data of each level into the detection model to obtain normal indicator data, key indicator data and abnormal data, and the model updating submodule is used to update the detection model in real time according to the new data.

3. A power supply chain indicator management system according to claim 2, characterized in that: The model building submodule includes: A data acquisition unit, used to acquire historical indicator data and use the historical indicator data as original training time series data samples; A first preprocessing unit, used for performing data preprocessing on the original training time series data samples to obtain training data samples; A first model building unit, used to build an initial LSTM-AE model using an LSTM deep model; A second model construction unit is used to reconstruct the initial LSTM-AE model using the training data set, optimize the network weight value of the initial LSTM-AE model, and obtain the trained LSTM-AE model and the error sequence of the training data sample; The third model building unit is used to obtain an adaptive threshold formula by using the error sequence of the training data set, and build a first detection model according to the adaptive threshold formula.

4. A power supply chain indicator management system according to claim 1, characterized in that: The expression of the adaptive threshold formula is: TH = μ + α · 0.8; μ is the mean of the error sequence, which indicates the average level of error under normal conditions, and α is the standard deviation of the error sequence, which indicates the degree of error fluctuation.

5. A power supply chain indicator management system according to claim 1, characterized in that: The second monitoring module includes: An indicator classification submodule, used to classify the first normal indicator data, the first key indicator data and the first abnormal data to obtain classified data, wherein the classified data includes: indicator name, indicator definition, calculation logic, indicator category, related major, indicator level, release frequency, responsible department, indicator unit, and early warning rules; A feature extraction submodule is used to extract features from the classified data to obtain features of various types of indicator data; The importance assessment submodule is used to perform importance assessment on the characteristics of the various indicator data according to the LIME algorithm to obtain importance data; A clustering submodule, used to cluster the historical indicator data to obtain cluster data; The judgment submodule is used to judge whether new data is generated according to the clustering data and the importance data, and if so, obtain new data.

6. A power supply chain indicator management system according to claim 5, characterized in that: The judgment submodule includes: A second preprocessing unit, used for performing standardization processing on the importance data and the clustering data to obtain preprocessed importance data and clustering data; A fusion unit, used for performing feature fusion on the pre-processed importance and clustering data to obtain comprehensive data; A threshold setting unit, used to set a new data generation threshold; A multi-model building unit, used to build multiple discriminant models; The comprehensive judgment unit is used to use the multiple discrimination models to judge the comprehensive data according to the threshold value generated by the new data to obtain a judgment result.

7. A power supply chain indicator management system according to claim 1, characterized in that: The sharing platform includes: A wireless communication submodule and several multi-level display submodules; The wireless communication submodule is used to transmit the second key indicator data, the second abnormal data and the second normal indicator data to the multi-level display value module, and the multi-level display submodule is used to publicize the second key indicator data, the second abnormal data and the second normal indicator data and mark the second abnormal data.

Images