Big data analysis and prediction method and system based on artificial intelligence

Abstract

The invention discloses a big data analysis and prediction method and system based on artificial intelligence, and belongs to the technical field of deep learning, and the method comprises the steps: S10, dividing a key monitoring time period of a preset power grid, including dividing the key monitoring time period into a morning peak time period and a evening peak time period; collecting current changes of a preset power grid in the morning peak period and the evening peak period, and generating a database; and S20, analyzing the change rule of the current in the database, intercepting a group of characteristic current data corresponding to the morning peak period and the evening peak period from the change rule, marking the two groups of characteristic current data as first reference data and second reference data, and obtaining the maximum value in the first reference data and the second reference data. According to the invention, accurate monitoring and early warning of the operation state of the power grid are realized through multiple measures such as monitoring time interval division, data acquisition and analysis, safety threshold setting, early warning mechanism and industrial electricity user monitoring.

Description

Technical Field

[0001] The present invention relates to the field of deep learning technology, and in particular to a big data analysis and prediction method and system based on artificial intelligence. Background Art

[0002] In recent years, artificial intelligence (AI) technologies, particularly machine learning and deep learning algorithms, have demonstrated advantages and potential in data processing and prediction. AI-based big data analysis and prediction have broad application prospects and value in the power sector. The stable operation of power systems relies on real-time monitoring of equipment status and rapid response to faults. By deploying sensor networks to collect equipment operating data (such as current, voltage, and temperature), AI algorithms can provide early warning of equipment status failures.

[0003] Among them, the application document with technical application number 202310592620.8 provides an intelligent data prediction and analysis system and method based on big data, including a data acquisition unit, a data preprocessing unit, a data dimensionality reduction unit, and a distributed server cluster; the data acquisition unit, whose output end is connected to the input end of the data preprocessing unit, is used to collect hourly power load, relative humidity, temperature, atmospheric pressure, rainfall, wind speed, and electricity price data; the data preprocessing unit, whose output end is connected to the input end of the data dimensionality reduction unit, is used to process the raw data; and the data dimensionality reduction unit, whose output end is connected to the input end of the distributed server cluster. This technical solution can improve the training speed of the model.

[0004] Another application, CN202410413347.2, provides an AI-based data analysis method and system, including the following steps: S1: receiving a storage request from a user node; S2: performing reliability verification based on the storage request and obtaining a reliability verification result. If the reliability verification result is reliable, the data to be stored is received and S3 is executed; if the reliability verification result is unreliable, the process ends and an alarm is generated; S3: automatically analyzing the data to be stored based on a pre-built storage framework to obtain a storage strategy. This technical solution can improve the accuracy and security of data analysis and data storage.

[0005] However, the operation optimization and fault handling of power systems require real-time decision-making, but the existing deep learning models have high computational complexity and slow inference speed, which makes it difficult to meet the real-time requirements of the power grid. When faced with new data or unknown scenarios, the prediction performance will be reduced, and the prediction of faults will be delayed. Summary of the Invention

[0006] In view of the above-mentioned problems existing in the existing field of deep learning technology, the present invention is proposed.

[0007] Therefore, one of the purposes of the present invention is to provide a big data analysis and prediction method and system based on artificial intelligence, which realizes accurate monitoring and early warning of the operation status of the power grid through multiple measures such as monitoring period division, data collection and analysis, safety threshold setting, early warning mechanism and monitoring of industrial electricity users.

[0008] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0009] In one aspect, the present invention provides a big data analysis and prediction method based on artificial intelligence, comprising the following steps:

[0010] Step S10: Dividing the key monitoring period of the preset power grid into a morning peak period and an evening peak period; collecting current changes of the preset power grid during the morning peak period and the evening peak period, and generating a database;

[0011] Step S20: analyzing a variation pattern of current in the database, extracting a set of characteristic current data corresponding to the morning peak period and the evening peak period from the variation pattern, marking the two sets of characteristic current data as first reference data and second reference data, and obtaining a maximum value between the first reference data and the second reference data;

[0012] Step S30: Preset a safety threshold according to the maximum value, collect at least 10 current data that are smaller than the safety threshold, the current data being the current data closest to the safety threshold, and divide the 10 current data into groups in the following manner:

[0013] Divide the first to third current data into initial data;

[0014] The 4th to 6th current data are divided into comparison data;

[0015] The 7th to 10th current data are divided into judgment data;

[0016] Step S40: Analyze the regular characteristics of the change from the initial data to the comparison data, and the analysis method includes giving an analysis period, analyzing the fluctuation changes of the current data in the analysis period, and the analysis period includes an analysis period of 20 seconds; in the analysis period, if the current data shows a stable trend and / or a decreasing trend, an early warning is issued for the operating status of the preset power grid.

[0017] As a preferred solution of the present invention, if the current data changes in an increasing trend, the difference between the current data and the largest current data in the judgment data is calculated, and the largest current data in the judgment data is marked as a critical value. When the current data exceeds the critical value in the process of changing in an increasing trend, it is determined that the current data exceeds the safety threshold, and at the same time, it is determined that the preset power grid is in an abnormal operating state; and an early warning is issued for the operating state of the preset power grid; otherwise, no early warning is issued.

[0018] As a preferred solution of the present invention, when an early warning is issued for the operating status of the preset power grid, a calculation period is preset based on the time point corresponding to the exceeding of the critical value, and the time point is marked as a reference time point; the calculation period is 15 to 20 minutes from the reference time point as one calculation period, and the change pattern of the difference in current data is calculated within the calculation period, and is calculated according to the following formula:

[0019] Among them, p j represents the p-th difference value collected at the j-th time;

[0020] In the above formula, m j represents the mth time point corresponding to the maximum difference value collected for the jth time, k represents the average difference value collected, l represents the interval between adjacent current data changes, and δ represents the number of times the same difference value appears.

[0021] As a preferred solution of the present invention, the total number of differences obtained in the calculated change law is counted, and a data set G=[G1, G2, G3, ..., G n ], wherein n represents the nth difference, and the first 8 to 10 differences are selected from the data set, and the differences are marked as comparison differences; if the change in the difference of the preset power grid current data in the future time period is the same as the comparison data, it is determined that the preset power grid is in an abnormal operating state; and an early warning is issued for the operating state of the preset power grid; otherwise, no determination is made.

[0022] As a preferred solution of the present invention, an analysis mechanism is given in the comparison difference, and the analysis mechanism includes a monitoring node for current data changes preset according to the comparison difference, and the monitoring node monitors based on the first 3 to 4 differences in the comparison difference, and monitors every two adjacent differences at intervals of 10 seconds, and counts the change characteristics of the current data during the interval, calculates the percentage of adjacent current data changes based on the change characteristics, and marks the percentage as a reference percentage; when the current data changes are monitored in the monitoring node in the future time period, if the adjacent current data changes are lower than the reference percentage, it is determined that the preset power grid is in a normal operating state; otherwise, it is not determined.

[0023] As a preferred solution of the present invention, the analysis mechanism further includes collecting the 10 to 20 industrial electricity users with the largest electricity consumption in the area where the preset power grid is located when issuing an early warning on the operating status of the preset power grid, analyzing the current data changes of the industrial electricity users 30 minutes to 40 minutes before the early warning is issued, and when monitoring the current data changes of the industrial electricity users in the future time period, if the change of adjacent current data is the same as the reference percentage and / or greater than the reference percentage, it is determined that the preset power grid will be in an abnormal operating state subsequently; and an early warning is issued on the operating status of the preset power grid; otherwise, no determination is made.

[0024] As a preferred solution of the present invention, the daily electricity consumption of each industrial electricity user is calculated among the industrial electricity users, and the industrial electricity users are sorted by size according to the daily electricity consumption. When analyzing the current data changes of the industrial electricity users 30 minutes to 40 minutes before the early warning is issued, the industrial electricity users whose current data changes first exceed the reference percentage and / or are the same as the reference percentage are obtained. When the industrial electricity users are obtained, it is determined that the preset power grid will be in an abnormal operating state subsequently; and an early warning is issued for the operating state of the preset power grid.

[0025] As a preferred solution of the present invention, the current data of the industrial electricity users whose current data changes first exceed the reference percentage and / or are the same as the reference percentage are divided into several evaluation indicators, and the proportion of the current data corresponding to the evaluation indicators to the current data outside the morning peak period and the evening peak period is calculated; a risk threshold is preset according to the proportion, and when the current data of the preset power grid and the industrial electricity users are collected outside the morning peak period and the evening peak period in the future, if the current data of the industrial electricity users exceeds the risk threshold, it is determined that the preset power grid will be in an abnormal operating state subsequently; and an early warning is issued for the operating state of the preset power grid.

[0026] In another aspect, the present invention provides a system for applying the above-mentioned artificial intelligence-based big data analysis and prediction method, comprising:

[0027] a data partitioning module, configured to divide a key monitoring period of a preset power grid into a morning peak period and an evening peak period; collect current changes of the preset power grid during the morning peak period and the evening peak period, and generate a database;

[0028] a data analysis module, the data analysis module responding to the database, configured to analyze a variation pattern of current in the database, extract a set of characteristic current data corresponding to the morning peak period and the evening peak period from the variation pattern, mark the two sets of characteristic current data as first reference data and second reference data, and obtain a maximum value between the first reference data and the second reference data;

[0029] A fusion processing unit, configured to preset a safety threshold according to the maximum value and collect at least 10 pieces of current data that are less than the safety threshold; the fusion processing unit includes a distinguishing module, a processing module, and a determining module;

[0030] The differentiation module is used to divide the 10 current data in the following manner:

[0031] Divide the first to third current data into initial data;

[0032] The 4th to 6th current data are divided into comparison data;

[0033] The 7th to 10th current data are divided into judgment data;

[0034] The processing module is used to analyze and process the regular characteristics of the change from the initial data to the comparison data, and the analysis and processing method includes a given analysis period, and analyzing the fluctuation change of the current data in the analysis period, wherein the analysis period includes an analysis period of 20 seconds;

[0035] The determination module responds to the analysis period, and issues an early warning on the operating status of the preset power grid if the current data shows a stable trend and / or a decreasing trend during the analysis period.

[0036] Beneficial effects:

[0037] 1. By collecting current data during key monitoring periods and analyzing its changing patterns, the present invention can promptly detect abnormal current changes, issue early warnings, avoid overload operation of power grid equipment, and reduce the risk of equipment failure.

[0038] 2. The present invention uses artificial intelligence technology to analyze current data, which can quickly and accurately determine whether the power grid is in an abnormal operating state and issue early warning signals in a timely manner;

[0039] 3. The present invention monitors the changing trend of current data in real time through preset calculation cycles and analysis mechanisms, and can predict possible abnormal conditions in the power grid in advance, providing sufficient time for power grid operators to take measures;

[0040] 4. By analyzing and sorting the current data of industrial electricity users, the present invention can quickly locate key users that may cause power grid anomalies, provide accurate control targets for power grid operation management, and optimize power grid operation management strategies. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0042] Figure 1 Schematic diagram of the modular structure of a big data analysis and prediction system based on artificial intelligence according to an embodiment of the present invention;

[0043] Figure 2 Schematic diagram of a method flow in an embodiment of the present invention;

[0044] Figure 3 This is a schematic diagram of the process structure of an embodiment of the present invention;

[0045] Numbers in the figure: 110 - data division module; 120 - data analysis module; 130 - fusion processing unit; 1301 - differentiation module; 1302 - processing module; 1303 - determination module. DETAILED DESCRIPTION

[0046] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions of the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of 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 described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of the present invention.

[0047] Since the operation optimization and fault handling of the power system require real-time decision-making, the existing deep learning models have high computational complexity and slow inference speed, which makes it difficult to meet the real-time requirements of the power grid. When faced with new data or unknown scenarios, the prediction performance will be reduced, and there will be a lag in the prediction of faults.

[0048] Based on this, the present invention proposes an artificial intelligence-based big data analysis and prediction method and system, which realizes accurate monitoring and early warning of the power grid operation status through multiple measures such as monitoring period division, data collection and analysis, safety threshold setting, early warning mechanism and monitoring of industrial electricity users.

[0049] The present invention will be further described in detail below through embodiments and in conjunction with the accompanying drawings.

[0050] Reference Figures 1 to 3 , is an embodiment of the present invention, which provides a big data analysis and prediction method based on artificial intelligence, comprising the following steps:

[0051] Step S10: Dividing the key monitoring period of the preset power grid into a morning peak period and an evening peak period; collecting current changes of the preset power grid during the morning peak period and the evening peak period, and generating a database;

[0052] In this embodiment, by dividing the morning peak period into the evening peak period, focusing on the period with the most significant changes in grid load, it is possible to more efficiently monitor and analyze the current variation pattern; at the same time, it avoids indiscriminate monitoring of all time periods, saves data collection and processing resources, and improves monitoring efficiency;

[0053] It should be noted that, according to research and experience on power grid operation, the risk of power grid failure increases significantly during peak power consumption periods. This is mainly due to the fact that loads are concentrated during peak periods, and power grid equipment is in a high-load operation state for a long time, which increases the probability of equipment failure.

[0054] In addition, high-load operation will cause the grid voltage to drop and the voltage fluctuation range to increase, which will also affect the normal operation of the equipment;

[0055] The morning and evening peak hours are the key monitoring periods for the pre-set power grid because the electricity load characteristics and grid operation risks during these two periods are significantly different from those during other periods. For example, during the morning peak hour (7:00-9:00), people begin their daily activities, and household electrical appliances (such as electric water heaters, air conditioners, and ovens) are used in large quantities. At the same time, commercial and industrial electricity consumption also begins to gradually increase.

[0056] During the evening peak period (18:00-21:00), household electrical appliances (such as air conditioners, televisions, and water heaters) are used intensively again, while commercial electricity (such as shopping malls and office buildings) has not yet been completely shut down, resulting in a sharp increase in electricity load.

[0057] It also includes the relationship between traffic flow and electricity load. For example, urban traffic flow reaches its peak during the morning and evening rush hours, which increases the electricity demand for traffic lights and public transportation systems. In addition, traffic congestion may cause vehicles to start and stop frequently, increasing fuel consumption and exhaust emissions, indirectly affecting the city's energy consumption and grid load.

[0058] As the number of electric vehicles increases, the impact of their charging demand on the grid load during peak hours becomes increasingly significant.

[0059] Therefore, these two periods are critical for residential and commercial activities. By monitoring the electricity load during the morning and evening peak periods, the grid dispatching department can more accurately predict load changes and optimize power generation plans and grid operation methods.

[0060] Step S20: Analyzing the variation pattern of the current in the database, extracting a set of characteristic current data corresponding to the morning peak period and the evening peak period from the variation pattern, marking the two sets of characteristic current data as first reference data and second reference data, and obtaining the maximum value between the first reference data and the second reference data;

[0061] In this embodiment, by analyzing the current variation pattern, we can gain an in-depth understanding of the operating characteristics of the power grid during key periods, providing a scientific basis for subsequent early warning and decision-making;

[0062] At the same time, marking the reference data and obtaining the maximum value provides a clear benchmark for setting safety thresholds, ensuring the rationality of the early warning system. By extracting key feature data from a large amount of data, the amount of data to be processed later is reduced, thus improving analysis efficiency.

[0063] Step S30: Based on the preset safety threshold value of the maximum value, collect at least 10 current data that are less than the safety threshold value, the current data being the current data closest to the safety threshold value, and divide the 10 current data into the following ways:

[0064] Divide the first to third current data into initial data;

[0065] The 4th to 6th current data are divided into comparison data;

[0066] The 7th to 10th current data are divided into judgment data;

[0067] In this embodiment, the current data is divided into different levels, which facilitates subsequent targeted analysis and processing, and improves the accuracy of the early warning system. By dividing the data into different levels, hierarchical management of risks is achieved, which helps to formulate differentiated response measures.

[0068] Step S40: Analyzing regular characteristics of changes from the initial data to the comparison data. The analysis method includes analyzing fluctuations in the current data during a given analysis period, wherein each analysis period is 20 seconds. If the current data shows a stable trend and / or a decreasing trend during the analysis period, an early warning is issued regarding the operating status of the preset power grid.

[0069] It should be noted that in this embodiment, the current data changes in a steady trend, indicating that the current change remains within ±10% of the preset grid rated current;

[0070] On the basis of the above, if the current data changes in an increasing trend, the difference between the current data and the largest current data in the judgment data is calculated, and the largest current data in the judgment data is marked as the critical value. When the current data exceeds the critical value during the process of changing in an increasing trend, it is determined that the current data exceeds the safety threshold, and at the same time, it is determined that the preset power grid is in an abnormal operating state; and an early warning is issued for the operating state of the preset power grid; otherwise, no early warning is issued;

[0071] Furthermore, in this embodiment, when an early warning is issued for the operation status of a preset power grid, a calculation period is preset based on the time point corresponding to the exceeding of the critical value, and the time point is marked as a reference time point; the calculation period is 15 to 20 minutes from the reference time point as one calculation period, and the change pattern of the difference in current data is calculated within the calculation period, and is calculated according to the following formula:

[0072] Among them, p j represents the p-th difference value collected at the j-th time;

[0073] In the above formula, m j represents the mth time point corresponding to the maximum difference value collected for the jth time, k represents the average difference value collected, l represents the interval between adjacent current data changes, and δ represents the number of times the same difference value appears;

[0074] In this embodiment, by calculating the variation pattern of the difference, the potential patterns in the current data are deeply explored, providing a reliable basis for early warning. By selecting and comparing the difference for early warning judgment, the accuracy and reliability of the early warning are improved, and misjudgment is avoided. In addition, dynamic early warning based on the variation pattern of the difference can respond to dynamic changes in power grid operation in a timely manner.

[0075] On the basis of the above, this embodiment counts the total number of differences obtained in the calculated change rule, and generates a data set G=[G1, G2, G3, ..., G n ], where n represents the nth difference, and the first 8 to 10 differences are selected from the data set and marked as comparison differences; if the change in the difference of the preset power grid current data in the future period is the same as the comparison data, it is determined that the preset power grid is in an abnormal operating state; and an early warning is issued for the operating state of the preset power grid; otherwise, no determination is made;

[0076] Furthermore, an analysis mechanism is provided in the comparison difference, the analysis mechanism including presetting a monitoring node for current data changes based on the comparison difference, the monitoring node monitoring the first three to four differences in the comparison difference, and monitoring each two adjacent differences at intervals of 10 seconds, statistically analyzing the change characteristics of the current data during the interval, calculating the percentage of adjacent current data changes based on the change characteristics, and marking the percentage as a reference percentage; when the current data changes are monitored in the monitoring node in a future time period, if the adjacent current data changes are lower than the reference percentage, then the preset power grid is determined to be in a normal operating state; otherwise, no determination is made;

[0077] In this embodiment, by analyzing the current data changes of industrial electricity users, key users that may cause power grid abnormalities can be accurately located; setting monitoring nodes enables real-time monitoring of current data changes, improving the timeliness and accuracy of early warnings;

[0078] Furthermore, in this embodiment, the analysis mechanism further includes collecting data from the 10 to 20 industrial electricity users with the largest electricity consumption in the area where the preset power grid is located when issuing an early warning for the operation status of the preset power grid, analyzing the current data changes of the industrial electricity users 30 minutes to 40 minutes before the early warning is issued, and when monitoring the current data changes of the industrial electricity users in the future time period, if the change of adjacent current data is the same as and / or greater than the reference percentage, it is determined that the preset power grid will be in an abnormal operation state in the future; and an early warning is issued for the operation status of the preset power grid; otherwise, no determination is made;

[0079] It should be noted that according to research experience on power grid operation, during peak hours of power grid operation, some enterprises with the largest industrial electricity consumption will have an impact on the safety of power grid operation. This impact includes load concentration and power grid overload. Due to the sharp increase in industrial electricity consumption during peak hours, especially the concentrated electricity consumption of high-energy-consuming enterprises (such as steel, cement, and chemical industries), local power grid load overload will be caused. The long-term high-load operation of power transmission lines, transformers and other equipment can easily lead to equipment overheating, insulation aging or even damage, which in turn causes failures.

[0080] At the same time, in enterprises that use industrial electricity, the simultaneous start-up or shutdown of a large number of industrial equipment (such as high-power motors) can also cause grid voltage fluctuations, which in turn affects the normal operation of other user equipment.

[0081] Therefore, this embodiment takes into account the actual situation and collects the current data changes of the 10 to 20 industrial electricity users with the largest electricity consumption in the area where the preset power grid is located when issuing an early warning on the operating status of the preset power grid, which has practical significance;

[0082] Based on the above, this embodiment calculates the daily electricity consumption of each industrial electricity user and sorts them by size based on their daily electricity consumption. When analyzing the changes in the current data of the industrial electricity users 30 to 40 minutes before the issuance of the warning, the embodiment obtains the industrial electricity user whose current data change first exceeds and / or is the same as the reference percentage. When the industrial electricity user is obtained, it is determined that the preset power grid will subsequently be in an abnormal operating state, and a warning is issued regarding the operating state of the preset power grid.

[0083] Furthermore, this embodiment divides the current data of industrial electricity users whose current data changes first exceed a reference percentage and / or are the same as the reference percentage into several evaluation indicators, and calculates the proportion of the current data corresponding to the evaluation indicators to the current data outside the morning peak period and the evening peak period; a risk threshold is preset based on the proportion, and when current data of the preset power grid and industrial electricity users are collected outside the morning peak period and the evening peak period in the future, if the current data of the industrial electricity users exceeds the risk threshold, it is determined that the preset power grid will subsequently be in an abnormal operating state; and an early warning of the operating state of the preset power grid is issued;

[0084] In this embodiment, by calculating evaluation indicators, a risk assessment is conducted on the current data of industrial electricity users, providing a scientific basis for grid operation management; the preset risk thresholds clarify the standards for risk warnings, and improve the scientificity and rationality of the warning system; and by incorporating industrial electricity users into the monitoring scope, comprehensive monitoring of grid operation is achieved, thereby improving the overall safety of grid operation.

[0085] Based on the above, this application realizes accurate monitoring and early warning of the power grid operation status through multiple measures such as monitoring period division, data collection and analysis, safety threshold setting, early warning mechanism and monitoring of industrial electricity users; its advantages are to improve the safety, reliability and intelligence level of power grid operation, optimize the power grid operation management efficiency, reduce operating costs, and at the same time enhance the power grid's ability to cope with complex working conditions.

[0086] This embodiment combines the above-mentioned big data analysis and prediction method based on artificial intelligence and also proposes a working system of the method, as follows:

[0087] The data partitioning module 110 is configured to divide the key monitoring period of the preset power grid into a morning peak period and an evening peak period; collect current changes of the preset power grid during the morning peak period and the evening peak period, and generate a database;

[0088] A data analysis module 120, which responds to a database and is configured to analyze a variation pattern of current in the database, extract a set of characteristic current data corresponding to a morning peak period and an evening peak period from the variation pattern, mark the two sets of characteristic current data as first reference data and second reference data, and obtain a maximum value between the first reference data and the second reference data;

[0089] The fusion processing unit 130 is used to preset a safety threshold according to the maximum value and collect at least 10 current data that are less than the safety threshold; the fusion processing unit 130 includes a distinguishing module 1301, a processing module 1302 and a determining module 1303;

[0090] The distinguishing module 1301 is used to divide the 10 current data in the following manner:

[0091] Divide the first to third current data into initial data;

[0092] The 4th to 6th current data are divided into comparison data;

[0093] The 7th to 10th current data are divided into judgment data;

[0094] The processing module 1302 is used to analyze and process the regular characteristics of the change from the initial data to the comparison data. The analysis and processing method includes analyzing the fluctuation changes of the current data in a given analysis period, and the analysis period includes an analysis period of 20 seconds.

[0095] The determination module 1303 responds to the analysis period, and issues a warning on the operation status of the preset power grid if the current data shows a stable trend and / or a decreasing trend during the analysis period.

[0096] In summary, the present invention significantly improves the operational safety and reliability of the power grid during peak hours through intelligent big data analysis and prediction methods, enhances early warning capabilities, optimizes operational management efficiency, and reduces operating costs.

[0097] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A big data analysis and prediction method based on artificial intelligence, characterized in that: The following steps are involved: step S10: Dividing the preset key monitoring period of the power grid, including dividing the key monitoring period into a morning peak period and an evening peak period; Collecting current changes of the preset power grid during the morning peak period and the evening peak period, and generating a database; Step S20: analyzing a variation pattern of current in the database, extracting a set of characteristic current data corresponding to the morning peak period and the evening peak period from the variation pattern, marking the two sets of characteristic current data as first reference data and second reference data, and obtaining a maximum value between the first reference data and the second reference data; Step S30: Preset a safety threshold according to the maximum value, collect at least 10 current data that are smaller than the safety threshold, the current data being the current data closest to the safety threshold, and divide the 10 current data into groups in the following manner: Divide the first to third current data into initial data; The 4th to 6th current data are divided into comparison data; The 7th to 10th current data are divided into judgment data; Step S40: analyzing the regular characteristics of the change from the initial data to the comparison data, wherein the analysis method includes analyzing the fluctuation of the current data in a given analysis period, wherein the analysis period includes 20 seconds as one analysis period; During the analysis period, if the current data shows a stable trend and / or a decreasing trend, an early warning is issued for the operating status of the preset power grid.

2. The big data analysis and prediction method based on artificial intelligence according to claim 1, characterized in that: If the current data changes in an increasing trend, a difference between the current data and a largest current data in the determination data is calculated, and the largest current data in the determination data is marked as a critical value. When the current data exceeds the critical value while changing in an increasing trend, it is determined that the current data exceeds the safety threshold, and the preset power grid is determined to be in an abnormal operating state; and an early warning is issued for the operating state of the preset power grid; Otherwise, no warning will be issued.

3. The big data analysis and prediction method based on artificial intelligence according to claim 2, characterized in that: When an early warning is issued for the operation status of the preset power grid, a calculation period is preset based on the time point corresponding to the exceeding of the critical value, and the time point is marked as a reference time point; the calculation period is 15 to 20 minutes from the reference time point as a calculation period, and the change pattern of the difference of the current data is calculated within the calculation period, and the calculation result is calculated according to the following formula: Among them, p j represents the p-th difference value collected at the j-th time; In the above formula, mj represents the mth time point corresponding to the maximum difference value collected for the jth time, k represents the average difference value collected, l represents the interval between adjacent current data changes, and δ represents the number of times the same difference value appears.

4. The big data analysis and prediction method based on artificial intelligence according to claim 3, characterized in that: The total number of differences obtained in the calculated change rule is counted, and a data set G=[G1, G2, G3, ..., G n ], wherein n represents the nth difference, and the first 8 to 10 differences are selected from the data set, and the differences are marked as comparison differences; if the change in the difference of the preset power grid current data in the future time period is the same as the comparison data, it is determined that the preset power grid is in an abnormal operating state; and an early warning is issued for the operating state of the preset power grid; otherwise, no determination is made.

5. The big data analysis and prediction method based on artificial intelligence according to claim 4, characterized in that: An analysis mechanism is given in the comparison difference, and the analysis mechanism includes a monitoring node for current data changes preset according to the comparison difference. The monitoring node monitors the first 3 to 4 differences in the comparison difference based on the comparison difference, and monitors each adjacent difference at an interval of 10 seconds. The change characteristics of the current data during the interval are statistically analyzed, and the percentage of adjacent current data changes is calculated based on the change characteristics, and the percentage is marked as a reference percentage. When the current data changes are monitored in the monitoring node in a future time period, if the adjacent current data changes are lower than the reference percentage, the preset power grid is determined to be in a normal operating state; otherwise, no determination is made.

6. The big data analysis and prediction method based on artificial intelligence according to claim 5, characterized in that: The analysis mechanism further includes collecting data of the 10 to 20 industrial electricity users with the largest electricity consumption in the area where the preset grid is located when issuing an early warning on the operating status of the preset grid, analyzing the changes in current data of the industrial electricity users 30 minutes to 40 minutes before the early warning is issued, and when monitoring the changes in current data of the industrial electricity users in future time periods, if the changes in adjacent current data are the same as and / or greater than the reference percentage, determining that the preset grid will subsequently be in an abnormal operating state; and issuing an early warning on the operating status of the preset grid; Otherwise, no judgment is made.

7. The big data analysis and prediction method based on artificial intelligence according to claim 6, characterized in that: The daily electricity consumption of each industrial electricity user is calculated among the industrial electricity users, and the industrial electricity users are sorted by size according to the daily electricity consumption. When analyzing the current data changes of the industrial electricity users 30 minutes to 40 minutes before the early warning is issued, the industrial electricity user whose current data change first exceeds the reference percentage and / or is the same as the reference percentage is obtained. When the industrial electricity user is obtained, it is determined that the preset power grid will be in an abnormal operating state subsequently; and an early warning is issued on the operating state of the preset power grid.

8. The big data analysis and prediction method based on artificial intelligence according to claim 7, characterized in that: Classifying the current data of the industrial electricity users whose current data changes first exceed the reference percentage and / or are the same as the reference percentage into a plurality of evaluation indicators, and calculating the proportion of the current data corresponding to the evaluation indicators to the current data outside the morning peak period and the evening peak period; According to the preset risk threshold of the proportion, when the current data of the preset power grid and the industrial electricity users are collected in the future time period outside the morning peak period and the evening peak period, if the current data of the industrial electricity users exceeds the risk threshold, it is determined that the preset power grid will be in an abnormal operating state subsequently; and an early warning is issued for the operating state of the preset power grid.

9. A system for the big data analysis and prediction method based on artificial intelligence as claimed in claim 1, characterized in that: include: A data division module is used to divide the key monitoring period of the preset power grid, including dividing the key monitoring period into a morning peak period and an evening peak period; Collecting current changes of the preset power grid during the morning peak period and the evening peak period, and generating a database; a data analysis module, the data analysis module responding to the database, configured to analyze a variation pattern of current in the database, extract a set of characteristic current data corresponding to the morning peak period and the evening peak period from the variation pattern, mark the two sets of characteristic current data as first reference data and second reference data, and obtain a maximum value between the first reference data and the second reference data; A fusion processing unit, configured to preset a safety threshold according to the maximum value and collect at least 10 pieces of current data that are less than the safety threshold; the fusion processing unit includes a distinguishing module, a processing module, and a determining module; The differentiation module is used to divide the 10 current data in the following manner: Divide the first to third current data into initial data; The 4th to 6th current data are divided into comparison data; The 7th to 10th current data are divided into judgment data; The processing module is used to analyze and process the regular characteristics of the change from the initial data to the comparison data, and the analysis and processing method includes a given analysis period, and analyzing the fluctuation change of the current data in the analysis period, wherein the analysis period includes 20 seconds as one analysis period; The determination module responds to the analysis period, and issues an early warning on the operating status of the preset power grid if the current data shows a stable trend and / or a decreasing trend during the analysis period.

Images