Power Generation Operation Analysis-based Electric Quantity Prediction Platform Management System

Abstract

The present invention discloses a power consumption prediction platform management system based on power generation operation analysis, which relates to the technical field of power dispatching. It uses a trained power consumption demand prediction model to predict the power consumption status in the power supply area, generates a power supply prediction data set. If there is an abnormal power supply prediction, a multi-level alarm mechanism is enabled in the power supply area to send an alarm instruction to the outside; according to the identified key factors, the power supply area is divided into several sub-areas, and according to the user group portraits in the sub-areas, the corresponding response strategies are output by the response strategy library; the supply abnormality degree and the abnormal threshold are constructed from the feedback data after executing the response strategy, and according to the relationship between the supply abnormality degree and the abnormal threshold, the corresponding feedback measures are selected for the power supply area. When there are abnormalities in power demand and supply in the power supply area, a targeted and personalized response can be quickly made to avoid the current power supply imbalance.

Description

Technical Field

[0001] The present invention relates to the technical field of power dispatching, and specifically to a power quantity prediction platform management system based on power generation operation analysis. Background Art

[0002] Power quantity prediction based on power generation operation analysis refers to comprehensively analyzing and predicting the power generation situation of generator sets, the operating characteristics of the power system, historical power quantity data, and various factors that may affect the power quantity (such as regional economy, policies, climate, etc.) to determine the power quantity value within a specific future time period. This helps to formulate more reasonable power generation plans and dispatching strategies, ensure the balance between power supply and demand, and also helps new energy enterprises optimize their business decisions and reduce market risks.

[0003] In the Chinese invention patent with the application publication number CN117318183, a power dispatching method and system are disclosed, which can be applied to the technical field of power system automation. The method includes: based on the electric load prediction data in the power dispatching plan data of the target scenario in the future time period and the historical electric load data set of the target scenario, determining the target historical electric load data that matches the electric load prediction data from the historical electric load data set; based on the target historical electric load data, determining the target bus load in the future time period; determining the target braking power and the reactive power output in the future time period; based on the target bus load, the target braking power, and the reactive power output, determining the scenario risk assessment result of the target scenario in the future time period; determining that the scenario risk assessment result is used to characterize that there are potential risks in the power dispatching plan, optimizing the power dispatching plan data to obtain the target power dispatching plan data, and performing power dispatching on the power system according to the target power dispatching plan data.

[0004] When power generation facilities, such as power plants, regional microgrids, and other power generation devices, supply power to the outside, it is necessary to consider the power supply and demand data within the power supply area, and use these data as feedback to adjust the power generation status and power dispatching status, so as to avoid and reduce the occurrence of over-generation or insufficient power supply, and reduce the frequency of power waste or power shortage.

[0005] In the existing power prediction management methods, based on the meteorological condition data and user power consumption data within the power supply area, the power consumption demand within the power supply area is predicted, and a response is made in a timely manner when there is an abnormality in the power supply, so as to prevent the current abnormal situation from further deteriorating. However, when there is a certain degree of imbalance in the power supply and demand situation within the power supply area, in this scenario, if a holistic response to the power supply abnormality in the power supply area is still made based on the prediction of the power supply, it is very difficult to alleviate the upcoming power supply abnormality problem.

[0006] To this end, the present invention provides a power prediction platform management system based on power generation operation analysis. Summary of the Invention

[0007] (1) Technical problems to be solved

[0008] In view of the deficiencies of the prior art, the present invention provides a power prediction platform management system based on power generation operation analysis. By using the trained power consumption demand prediction model to predict the power consumption status in the power supply area, a power supply prediction data set is generated. If there is an abnormal power supply prediction, a multi-level alarm mechanism is enabled in the power supply area to send an alarm instruction to the outside; according to the key factors identified and obtained, the power supply area is divided into several sub-areas, and according to the user group portraits in the sub-areas, the corresponding response strategies are output by the response strategy library; the supply abnormality degree and the abnormality threshold are constructed from the feedback data after executing the response strategy, and according to the relationship between the supply abnormality degree and the abnormality threshold, the corresponding feedback measures are selected for the power supply area. When there are abnormalities in power demand and supply in the power supply area, a targeted and personalized response is quickly made, thus solving the technical problems raised in the background art.

[0009] (2) Technical solutions

[0010] To achieve the above object, the present invention is realized through the following technical solutions: A power prediction platform management system based on power generation operation analysis, including a demand analysis unit, which generates an imbalance value Spm after collecting the power demand scenario data in the power supply area. If the imbalance value Spm exceeds the expectation, a data collection instruction is sent to the outside; among them, the imbalance value Spm is generated from the power data in the power demand scenario data set in the following manner:

[0011]

[0012] In the formula: N is the total number of regions, T is the total number of time periods, D i,t is the power demand of region i in time period t, S i,t is the power supply of region i in time period t, U i,k =D i,t -S i,t , is the power imbalance value of region i in time period t, α is the cumulative effect coefficient, and its value falls within [0,1], β is the time decay coefficient, and its value falls within [0,1]; k represents a past time period, ranging from 1 to t-1;

[0013] A prediction unit that uses the trained power consumption demand prediction model to predict the power consumption status in the power supply area and generates a power supply prediction data set;

[0014] An alarm unit, if it is predicted that there is an abnormal power supply, enables a multi-level alarm mechanism within the power supply area and issues an alarm instruction to the outside;

[0015] A policy response unit divides the power supply area into several sub-areas according to the key factors obtained by recognition, and outputs corresponding response policies from the response policy library according to the user group portraits in the sub-areas;

[0016] A feedback unit constructs a supply abnormality degree Soj and an abnormality threshold [Ba, Bb] from the feedback data after executing the response policy, and selects corresponding feedback measures for the power supply area according to the relationship between the supply abnormality degree Soj and the abnormality threshold [Ba, Bb].

[0017] Further, after collecting the power demand data of each time period in the power supply area, use the trained pattern recognition model for pattern recognition to obtain the peak power consumption period and the non-peak period; collect the power demand data and power supply data of the peak period and the non-peak period respectively, and summarize them to generate a power demand scenario data set.

[0018] Further, after receiving the data collection instruction, collect the historical power consumption data and meteorological data in the power supply area, and summarize them to generate a historical power consumption scenario data set in the power supply area; extract part of the data from the historical power consumption scenario data set as sample data, and train to obtain a power demand prediction model;

[0019] Use the trained power demand prediction model to predict the power demand during the peak period in the power supply area, obtain the predicted values of the power demand, duration and power supply load, and summarize the predicted data at multiple consecutive time nodes to generate a power supply prediction data set.

[0020] Further, if the predicted value of the recent power supply load exceeds the load threshold, or when there are currently multiple consecutive power supply load predicted values, arrange the power supply load predicted values along the time axis. If the power supply load is on an upward trend, issue a first-level alarm instruction to the outside; if the number of first-level alarm instructions received within the preset alarm period exceeds the expectation, generate a warning value D according to the issued state data of the first-level alarm instruction M (X t )。

[0021] Further, if the obtained warning value D M (X t ) exceeds the warning threshold beyond expectation, issue a second-level alarm instruction to the outside; the obtaining method of the warning value D M (X t ) is as follows:

[0022]

[0023] Where: X t The standardized multi-variable time series vector, including the standardized values of the alarm instruction time series and the power supply load time series; A i Is the regression coefficient matrix of the VAR model, used to describe the influence of the lag term on the current vector X t The influence, p is the lag order of the VAR model, The inverse matrix of the residual covariance matrix, μ e Is the residual mean.

[0024] Furthermore, after receiving the secondary alarm instruction, collect the power supply and demand data, meteorological data, electricity consumption scenario data, and electricity consumption data in the power supply area, and obtain the key factors affecting power supply and the influence degree of each key factor after performing principal component analysis on the power supply demand data.

[0025] Furthermore, after obtaining the time, location information, and influence degree of the key factors, use the pre-trained clustering algorithm to classify several key factors, and then divide the power supply area into several sub-areas; construct an electronic map covering the power supply area, and mark each sub-area and the user information inside it on the electronic map.

[0026] Furthermore, after collecting user-related information, perform user portrait and behavior analysis to obtain the user group portrait; construct personalized response strategies for different user groups, including dynamic electricity price strategies, electricity consumption reward strategies, and electricity saving strategies: according to the correspondence between the user group portrait and the response strategy, output the corresponding response strategy from the response strategy library.

[0027] Furthermore, after the user groups in the power supply area execute the response strategy, collect the power demand scenario data in each sub-area during the pre-set observation period to generate the imbalance value Spm, construct the supply anomaly degree Soj from several imbalance values, and after continuously obtaining several supply anomaly degrees Soj along the time axis, construct the anomaly threshold [Ba, Bb] from the continuous supply anomaly degrees Soj.

[0028] Furthermore, if the supply anomaly degree Soj exceeds the anomaly threshold [Ba, Bb], select several sub-areas with the most serious power supply gaps as the power compensation areas, and start the standby generator sets to supply power to the power compensation areas;

[0029] If the supply anomaly degree Soj is within the anomaly threshold [Ba, Bb], take reducing the supply anomaly degree Soj in the power supply area as the scheduling target, take the balance of power supply and demand as the constraint condition, output the optimal power scheduling strategy by the pre-trained particle swarm optimization algorithm, execute the power scheduling strategy, and dispatch the power in the non-power compensation area to the power compensation area;

[0030] When the supply abnormality degree Soj is lower than the abnormality threshold [Ba, Bb], no additional processing is performed on the power supply within the sub-region.

[0031] (III) Beneficial effects

[0032] The present invention provides a power quantity prediction platform management system based on power generation operation analysis, having the following beneficial effects:

[0033] 1. By constructing the imbalance value Spm, the imbalance degree between power supply and demand within the current power supply area can be judged and evaluated. If the imbalance degree of power supply is relatively high, it indicates that there may be a large gap in power supply in some areas, and timely adjustment is required.

[0034] 2. By issuing alarm instructions externally according to the judgment of the power supply load degree, the power supply early warning process is completed, and timely handling can be carried out when the power supply is unbalanced.

[0035] 3. Construct a warning value based on the release status data of the first-level alarm instruction, build a second-level alarm mechanism, avoid being ignored by a single alarm or static alarm, ensure the reliability of the alarm mechanism, and can also switch and replace the release channels of the alarm instruction to make the alarm process smoother.

[0036] 4. Formulate corresponding feedback or adjustment plans according to key factors, and divide the power supply area into several sub-regions according to the differences in key factors in each region or location. When there are power supply problems, corresponding treatment methods can be taken respectively to improve the pertinence of treatment.

[0037] 5. By profiling users and obtaining corresponding user characteristics, match targeted and personalized response strategies for the same user group. When there are problems with the power supply within the power supply area, targeted handling can be carried out to prevent the existing power supply imbalance from further deteriorating.

[0038] 6. According to the supply abnormality degree Soj, the effectiveness of the given response strategy can be judged as a whole, and through the dynamic changes of the supply abnormality degree Soj and the abnormality threshold [Ba, Bb], the dynamic assessment of the current power supply state is realized, maintaining consistency with the actual state.

[0039] 7. Select corresponding feedback measures for the power supply area according to the relationship between the supply abnormality degree Soj and the abnormality threshold [Ba, Bb], use the feedback data of the first response measure as feedback, re-match feedback measures for the power supply area, and handle the current continuous power supply imbalance to relieve the existing power imbalance state.

[0040] 8. By constructing a multi-level alarm and early warning mechanism and a multi-level response and feedback mechanism, when there are abnormalities in power demand and supply in the power supply area, a targeted and personalized response can be quickly made, which can alleviate and avoid the current power supply imbalance problem that may exist or continue to exist. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 It is a schematic flowchart of the management method of the power quantity prediction platform based on power generation operation analysis of the present invention;

[0042] Figure 2 It is a schematic structural diagram of the management system of the power quantity prediction platform based on power generation operation analysis of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0043] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0044] Please refer to Figure 1 , the present invention provides a management method for a power quantity prediction platform based on power generation operation analysis, including,

[0045] Step 1: After collecting the power demand scenario data in the power supply area, generate an imbalance value Spm. If the imbalance value Spm exceeds the expectation, send a data collection instruction to the outside;

[0046] The said Step 1 includes the following contents:

[0047] Step 101: Train a convolutional neural network with the labeled data to obtain a trained pattern recognition model; after collecting the power demand data at each time period in the power supply area, use the trained pattern recognition model for pattern recognition to obtain the peak power consumption time period and the non-peak time period therein;

[0048] Collect the power demand data, power supply data, and supply-demand imbalance data at the peak time period and the non-peak time period respectively, and summarize them to generate a power demand scenario data set;

[0049] When in use, collect the power demand data at each time period, and through data recognition, the peak power consumption time period and the non-peak time period of electricity consumption can be distinguished, and different power supply feedback strategies can be adopted for different time periods;

[0050] Step 102: Generate an imbalance value Spm from the power data in the power demand scenario data set in the following manner:

[0051]

[0052] Where: N is the total number of regions, T is the total number of time periods, D i,t is the power demand of region i in time period t, S i,t is the power supply of region i in time period t, U i,k = D i,t - S i,t , which is the power imbalance value of region i in time period t, α is the cumulative effect coefficient, with a value falling within [0, 1], β is the time decay coefficient, with a value falling within [0, 1]; k represents a past time period, ranging from 1 to t - 1;

[0053] According to historical data and management expectations for power supply and demand imbalance, set the imbalance threshold in advance;

[0054] If the imbalance value Spm exceeds the imbalance threshold, it indicates that the current power supply imbalance degree is relatively high, which may lead to ineffective power supply in some regions, while there may be a large surplus of power supply in some regions. Therefore, timely adjustment is required. At this time, send a data collection instruction to the outside;

[0055] When in use, combine the content in steps 101 and 102:

[0056] After determining the power supply data and demand data for each time period, through the constructed imbalance value Spm, the imbalance degree between power supply and demand in the current power supply region can be judged and evaluated. If the power supply imbalance degree is relatively high, it indicates that there may be a large gap in power supply in some regions, and timely adjustment is required.

[0057] In the existing power prediction management method, based on meteorological condition data and user power consumption data in the power supply region, the power consumption demand in the power supply region is predicted, and a timely response is made when there is an abnormality in power supply to prevent the current abnormal situation from further deteriorating. However, when there is already a certain degree of imbalance between power supply and demand in the power supply region, in this scenario, if a holistic response to the power supply abnormality in the power supply region is still made based on power supply prediction, it is very difficult to alleviate the upcoming power supply abnormality problem.

[0058] Step 2: Use the trained power consumption demand prediction model to predict the power consumption status in the power supply region and generate a power supply prediction data set;

[0059] The said step 2 includes the following content:

[0060] Step 201: After receiving the data collection instruction, collect historical electricity consumption data and meteorological data within the power supply area. For example, historical electricity consumption data, which is segmented by industry, scale, and region, including monthly, quarterly, and annual electricity consumption data, and electricity consumption data during special events (such as high temperatures in summer, cold in winter, large-scale events, etc.); electricity consumption data for various user segments (industrial, commercial, residential, agricultural); meteorological data: temperature (including daily maximum, minimum, and average temperatures), humidity, wind speed, precipitation, sunshine duration, atmospheric pressure, historical meteorological data, real-time meteorological monitoring data, and meteorological forecast data. Summarize and generate a set of historical electricity consumption scenario data within the power supply area.

[0061] Step 202: Extract some data from the set of historical electricity consumption scenario data as sample data, and train a machine learning algorithm with the sample data to obtain a trained electricity demand prediction model.

[0062] Use the trained electricity demand prediction model to predict the electricity demand during peak hours within the power supply area, obtain the predicted values of electricity demand volume, duration, and power supply load, and summarize the predicted data at multiple consecutive time nodes to generate a set of power supply prediction data.

[0063] When in use, combine the content in Steps 201 and 202:

[0064] Based on the electricity demand prediction model trained on the basis of the collected data, for the electricity demand data in the next several time periods, such as peak hours and off-peak hours, when there is a shortage in power supply, it can respond in advance, for example, start standby power generation equipment, etc.

[0065] Step 3: If it is predicted that there is an abnormal power supply, enable a multi-level alarm mechanism within the power supply area and send an alarm instruction to the outside.

[0066] The said Step 3 includes the following content:

[0067] Step 301: Set a load threshold for peak hours according to historical data. If the predicted value of the recent power supply load exceeds the load threshold, or when there are currently multiple consecutive predicted values of power supply load, arrange the predicted values of power supply load along the time axis. If the power supply load is on an upward trend, send a first-level alarm instruction to the outside.

[0068] When in use, after obtaining the load data within the power supply area, based on the relationship between the obtained load data and the load threshold, judge whether the power supply load is gradually increasing, and send an alarm instruction to the outside according to the judgment of the degree of power supply load, so as to complete the power supply early warning process and deal with it in time when the power supply is unbalanced.

[0069] Step 302: If the number of first-level alarm instructions received within the preset alarm period exceeds the expectation, after standardization, generate a warning value D based on the sending status data of the first-level alarm instructions M (X t ), the method is as follows:

[0070]

[0071] In the formula: X t The standardized multi-variable time series vector, including the standardized values of the alarm instruction time series and the power supply load time series; A i Is the regression coefficient matrix of the VAR model, used to describe the influence of the lag term on the current vector x t , p is the lag order of the VAR model, The inverse matrix of the residual covariance matrix, μ e Is the residual mean;

[0072] According to historical data and the management expectation of the degree of power supply imbalance, preset a warning threshold;

[0073] If the obtained warning value D M (X t ) exceeds the warning threshold beyond expectation, it indicates that the current degree of power supply imbalance is relatively serious, and a second-level alarm instruction is sent to the outside;

[0074] When in use, combine the content in Steps 301 and 302:

[0075] Construct a warning value based on the release status data of the first-level alarm instruction, build a second-level alarm mechanism, avoid being ignored by a single alarm or a static alarm, ensure the reliability of the alarm mechanism, and can also switch and replace the release channels of the alarm instruction to prevent the neglect of the first-level alarm instruction due to a malfunction in the current notification channel, making the alarm process smoother.

[0076] Step Four: Divide the power supply area into several sub-areas according to the identified key factors, and output corresponding response strategies from the response strategy library according to the user group portraits in the sub-areas;

[0077] The said Step Four includes the following content:

[0078] Step 401: After receiving the second-level alarm instruction, collect power supply and demand data, meteorological data, power consumption scenario data, power consumption data, etc. in the power supply area, and after performing principal component analysis on the power supply demand data, obtain the key factors affecting power supply and the influence degree of each key factor;

[0079] In step 402, after obtaining the time, location information and influence degree of the key factors, use the pre-trained clustering algorithm to classify several key factors, and then divide the power supply area into several sub-areas; construct an electronic map covering the power supply area, and mark each sub-area and the user information inside it on the electronic map;

[0080] When in use, after enabling the multi-level alarm mechanism, enable the principal component analysis. For the main factors or key factors that will affect the power supply load in the power supply area, in scenarios where the conditions are met, the current power load level and the dilemma of power supply imbalance can be improved by adjusting the key factors. Correspondingly, corresponding feedback or adjustment plans can also be formulated according to the key factors, and the power supply area is divided into several sub-areas according to the differences in key factors in each area or location. When there are power supply problems, corresponding treatment methods can be taken respectively to improve the pertinence of the treatment.

[0081] In step 403, after collecting user-related information, such as electricity consumption, electricity consumption period, electricity consumption type and purpose, etc., conduct user portrait and behavior analysis to obtain the user group portrait;

[0082] Construct personalized response strategies for different user groups, including dynamic electricity price strategies: adjust electricity prices according to peak load conditions, electricity consumption reward strategies, integral rewards, cash back, etc., to encourage users to reduce electricity consumption during peak hours; energy-saving strategies: provide energy-saving suggestions and electricity consumption optimization plans, and summarize the above response strategies to generate a response strategy library; according to the correspondence between the user group portrait and the response strategies, output the corresponding response strategies from the response strategy library;

[0083] When in use, combine the content in steps 401 to 403:

[0084] When in use, predict the power demand and supply in the power supply area, and on the basis of obtaining several sub-areas, by drawing portraits of users and obtaining corresponding user characteristics, personalized response strategies targeted to the same user group can be matched. When there are problems with the power supply in the power supply area, targeted treatment can be carried out to prevent the existing power supply imbalance from further deteriorating.

[0085] Step Five: Construct the supply anomaly degree Soj and the anomaly threshold [Ba, Bb] from the feedback data after executing the response strategy, and select corresponding feedback measures for the power supply area according to the relationship between the supply anomaly degree Soj and the anomaly threshold [Ba, Bb];

[0086] The said Step Five includes the following content:

[0087] Step 501: After the user groups in the power supply area execute the response strategy, collect the power demand scenario data in each sub-area during a pre-set observation period to generate an imbalance value Spm. After marking the corresponding sub-areas with the imbalance value, construct a supply anomaly degree Soj from a number of imbalance values. Among them, after linearly normalizing the imbalance value Spm, according to the following method:

[0088]

[0089] Weight coefficient: 0 ≤ V1 ≤ 1, 0 ≤ V2 ≤ 1 and V2 + V1 = 1; i = 1, 2, …, k, where k is the number of sub-areas, Spm i is the imbalance value of the i-th sub-area, Spm a is its mean value, is the qualified reference value of the imbalance value;

[0090] After continuously obtaining a number of supply anomaly degrees Soj along the time axis, construct an anomaly threshold [Ba, Bb] from the continuous supply anomaly degrees Soj in the following way:

[0091]

[0092] Among them, Soj max is the maximum value of the supply anomaly degree, Soj min is the minimum value of the supply anomaly degree, Soj a is the mean value of the supply anomaly degree, Soj i is the supply anomaly degree in the i-th sub-period;

[0093] When in use, after taking corresponding feedback data after taking corresponding response strategies in each sub-area respectively, and construct a supply anomaly degree Soj from the corresponding feedback data. According to the supply anomaly degree Soj, the current power supply imbalance degree in the power supply area can be judged as a whole, and the effectiveness of the previous response strategy can be judged; and through the dynamic changes of the supply anomaly degree Soj and the anomaly threshold [Ba, Bb], the dynamic evaluation of the current power supply state is realized, and the consistency with the actual state is maintained.

[0094] Step 502: According to the relationship between the supply anomaly degree Soj and the anomaly threshold [Ba, Bb], select corresponding feedback measures for the power supply area, where:

[0095] If the supply anomaly degree Soj exceeds the anomaly threshold [Ba, Bb], after selecting several sub-areas with the most serious power supply gaps, use them as power compensation areas, and start standby generator sets to supply power to the power compensation areas;

[0096] If the supply abnormality degree Soj is within the abnormality threshold [Ba, Bb], taking reducing the supply abnormality degree Soj within the power supply area as the scheduling target and the power supply and demand balance as the constraint condition, the optimal power scheduling strategy is output by the pre-trained particle swarm optimization algorithm, and the power scheduling strategy is executed to schedule the power in the non-power compensation area to the power compensation area;

[0097] When the supply abnormality degree Soj is lower than the abnormality threshold [Ba, Bb], no additional processing is performed on the power supply within the sub-region;

[0098] When in use, combine the content in steps 501 to 502:

[0099] After taking personalized response solutions for each sub-region respectively, by constructing the supply abnormality degree Soj as feedback and based on the relationship between the supply abnormality degree Soj and the abnormality threshold [Ba, Bb], select corresponding feedback measures for the power supply area, start from the entire power supply area again, use the feedback data of a response measure as feedback, re-match feedback measures for the power supply area, and handle the currently existing power supply imbalance to relieve the currently existing power imbalance state.

[0100] Combined with the above application, when there is an imbalance between the current power supply and demand in the power supply area, based on the prediction of power supply and demand data, by constructing a multi-level alarm and early warning mechanism and a multi-level response feedback mechanism, when there are abnormalities in power demand and supply in the power supply area, a targeted and personalized response can be made quickly, which can relieve and avoid the current possible or persistent power supply imbalance problem.

[0101] Please refer to Figure 2 , the present invention provides a power consumption prediction platform management system based on power generation operation analysis, including,

[0102] A demand analysis unit, which generates an imbalance value Spm after collecting the power demand scenario data in the power supply area. If the imbalance value Spm exceeds the expectation, it sends a data collection instruction to the outside;

[0103] A prediction unit, which predicts the power consumption status in the power supply area using the trained power consumption demand prediction model and generates a power supply prediction data set;

[0104] An alarm unit, if there is a predicted power supply abnormality, enables a multi-level alarm mechanism in the power supply area and sends an alarm instruction to the outside;

[0105] A strategy response unit, which divides the power supply area into several sub-regions according to the identified key factors, and outputs corresponding response strategies from the response strategy library according to the user group portraits in the sub-regions;

[0106] A feedback unit constructs a supply abnormality degree Soj and an abnormality threshold [Ba, Bb] from the feedback data after implementing a response policy, and selects a corresponding feedback measure for the power supply area according to the relationship between the supply abnormality degree Soj and the abnormality threshold [Ba, Bb].

[0107] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0108] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0109] In several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only for some logical function divisions, and there can be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0110] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or can be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0111] As described above, this is only the specific implementation manner of this application, but the protection scope of this application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in this application, and all should be covered by the protection scope of this application. Therefore, the protection scope of this application should be subject to the protection scope of the claims.

Claims

1. The power forecasting platform management system based on power generation operation analysis is characterized by: include, The demand analysis unit generates an imbalance value Spm after collecting the power demand scenario data in the power supply area. If the imbalance value Spm exceeds the expectation, a data collection instruction is issued to the outside. The imbalance value Spm is generated by the power data in the power demand scenario data set in the following manner: Where: N is the total number of regions, T is the total number of time periods, D i,t is the electricity demand of region i in time period t, S i,t is the power supply of region i in time period t, U i,k =D i,t -S i,t , is the power imbalance value of region i in time period t, α is the cumulative effect coefficient, and its value falls within [0,1], β is the time attenuation coefficient, and its value falls within [0,1]; k represents a past time period, ranging from 1 to t-1; A prediction unit, which uses the trained power demand prediction model to predict the power consumption status in the power supply area and generates a power supply prediction data set; The alarm unit, if it is predicted that there is an abnormality in the power supply, activates a multi-level alarm mechanism in the power supply area and issues an alarm command to the outside; The strategy response unit divides the power supply area into several sub-areas based on the key factors identified and obtained, and outputs the corresponding response strategy from the response strategy library based on the user group portraits in the sub-areas; The feedback unit constructs the supply abnormality Soj and the abnormality threshold [Ba, Bb] from the feedback data after executing the response strategy, and selects corresponding feedback measures for the power supply area according to the relationship between the supply abnormality Soj and the abnormality threshold [Ba, Bb].

2. The power forecasting platform management system based on power generation operation analysis according to claim 1, characterized in that: After collecting the power demand data of each period in the power supply area, the trained pattern recognition model is used to perform pattern recognition to obtain the peak and non-peak power consumption periods; The power demand data and power supply data during peak and non-peak periods are collected separately, and summarized to generate a power demand scenario data set.

3. The power forecasting platform management system based on power generation operation analysis according to claim 2, characterized in that: After receiving the data collection instruction, the historical electricity consumption data and meteorological data in the power supply area are collected, and the historical electricity consumption scene data set in the power supply area is generated; Extract some data from the historical electricity consumption scenario data set as sample data to train and obtain the electricity demand prediction model; The trained electricity demand forecasting model is used to forecast the electricity demand during peak hours in the power supply area, and the forecast values ​​of electricity demand, duration and power supply load are obtained. The forecast data at multiple time nodes obtained continuously are aggregated to generate a power supply forecast data set.

4. The power forecasting platform management system based on power generation operation analysis according to claim 3 is characterized in that: If the most recent power supply load forecast value exceeds the load threshold, or there are multiple consecutive power supply load forecast values, the power supply load forecast values ​​are arranged along the time axis. If the power supply load is on an upward trend, a first-level alarm instruction is issued to the outside; if the number of first-level alarm instructions received within the preset alarm cycle exceeds the expectation, a warning value D is generated based on the issuance status data of the first-level alarm instruction. M (X t ).

5. The power forecasting platform management system based on power generation operation analysis according to claim 4, characterized in that: If the warning value D M (X t ) exceeds the warning threshold beyond expectations, and issues a secondary alarm command to the outside; among which, the warning value D M (X t ) is obtained as follows: Where: X t The standardized multivariate time series vector includes the standardized values ​​of the alarm instruction time series and the power supply load time series; A i is the regression coefficient matrix of the VAR model, which is used to describe the relationship between the lag term and the vector X t The influence of, p is the lag order of the VAR model, The inverse of the residual covariance matrix, μ e is the residual mean.

6. The power forecasting platform management system based on power generation operation analysis according to claim 5, characterized in that: After receiving the second-level alarm command, the power supply and demand data, meteorological data, power usage scenario data and power usage data in the power supply area are collected, and the principal component analysis is performed on the power supply demand data to obtain the key factors affecting power supply and the degree of influence of each key factor.

7. The power forecasting platform management system based on power generation operation analysis according to claim 6 is characterized in that: After obtaining the time, location information and impact degree of the key factors, a pre-trained clustering algorithm is used to classify several key factors, and then the power supply area is divided into several sub-areas; an electronic map covering the power supply area is constructed, and each sub-area and the user information within it are marked on the electronic map.

8. The power forecasting platform management system based on power generation operation analysis according to claim 7, characterized in that: After collecting user-related information, conduct user profiling and behavior analysis to obtain user group profiles; Build personalized response strategies for different user groups, including dynamic electricity price strategies, electricity consumption reward strategies and electricity saving strategies. According to the correspondence between user group portraits and response strategies, the response strategy library outputs the corresponding response strategies.

9. The power forecasting platform management system based on power generation operation analysis according to claim 8, characterized in that: After the user group in the power supply area executes the response strategy, the power demand scenario data in each sub-area is collected within the preset observation period to generate an imbalance value Spm, and the supply anomaly Soj is constructed by several imbalance values ​​Spm. After continuously obtaining several supply anomaly degrees Soj along the time axis, the anomaly threshold [Ba, Bb] is constructed by the continuous supply anomaly degrees Soj.

10. The power forecasting platform management system based on power generation operation analysis according to claim 9, characterized in that: If the supply anomaly degree Soj is lower than the anomaly threshold [Ba, Bb], no additional processing is performed on the power supply in the sub-area; If the supply anomaly Soj is within the anomaly threshold [Ba, Bb], the dispatching target is to reduce the supply anomaly Soj in the power supply area, and the balance of power supply and demand is used as a constraint. The pre-trained particle swarm optimization algorithm outputs the optimal power dispatching strategy, executes the power dispatching strategy, and dispatches the power in the non-power compensation area to the power compensation area; If the supply anomaly Soj exceeds the anomaly threshold [Ba, Bb], several sub-areas with the most serious power supply gap are selected as power compensation areas, and the backup generator sets are enabled to supply power to the power compensation areas.

Images