Power grid load forecasting method and system based on big data analysis

Abstract

The present invention is applicable to the technical field of load forecasting, and in particular to a power grid load forecasting method and system based on big data analysis, the method comprising: S100: determining energy consumption units in a region, defining the power consumption of the energy consumption units the previous day as historical power consumption, inputting the historical power consumption into a pre-built prediction model, and outputting a prediction result; S200: splitting the corresponding relationship between the prediction result and time, defining the energy consumption units as child nodes, and transferring the corresponding relationship to the child nodes, integrating the child nodes based on the configured clustering rules, and constructing a balance tree. The present invention can define the normal load fluctuation range of the power grid by setting the fluctuation range, reduce unnecessary intervention, and also quickly identify anomalies, improve the accuracy of load forecasting, and quantify the load gap of the power grid by reading the inclination angle of the balance tree, thereby providing data support for the normal operation of the power grid.

Description

Technical Field

[0001] The present invention relates to the technical field of load forecasting, and in particular to a power grid load forecasting method and system based on big data analysis. Background Art

[0002] Power system load forecasting is an important part of the safety technical measures in the power system. Like relay protection, stability calculation and short-circuit calculation, it plays a very important role in the safe, economical and stable operation of the power system. Power load forecasting is the basis of power grid planning and operation. Accurate load forecasting helps to improve the safety and stability of power grid operation, reduce power generation costs and ensure the accuracy of planning. In addition, the use of big data technology to analyze and process massive historical data and real-time data can further improve the accuracy and real-time performance of power grid load forecasting.

[0003] When electricity demand is relatively stable and there are no major emergencies or extreme weather events, using the previous day's electricity consumption to predict the current day's electricity consumption can effectively capture short-term trends and provide data support for power supply load forecasting; therefore, "how to use the previous day's historical electricity consumption to predict the power supply load and correct the forecast results based on real-time electricity consumption" is the technical problem that the present invention needs to solve. Summary of the invention

[0004] The purpose of the present invention is to provide a power grid load forecasting method and system based on big data analysis to solve the problem raised in the above background technology of "how to use the historical power consumption of the previous day to forecast the power supply load and correct the forecasting results according to the real-time power consumption".

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A power grid load prediction method based on big data analysis, the method comprising:

[0007] S100: Determine energy-consuming units in the region, define the power consumption of the energy-consuming units on the previous day as historical power consumption, input the historical power consumption into a pre-built prediction model, and output a prediction result;

[0008] S200: splitting the corresponding relationship between the prediction result and the time, defining the energy consumption unit as a child node, transferring the corresponding relationship to the child node, integrating the child nodes based on the configured clustering rules, and constructing a balanced tree;

[0009] S300: placing the child node on the left side of the balancing tree, creating a point corresponding to the child node on the right side of the balancing tree, transferring the collected real-time power consumption of the energy-consuming unit to the point, building a balancing link between the child node and the point, determining the fluctuation range of the child node, and embedding the fluctuation range into the balancing link;

[0010] S400: insert a processing node in the middle of the balancing tree, build a data transmission channel between the processing node and the face recognition device, the face recognition device is set at the entrance and exit of the energy-consuming unit, obtains portrait recognition data through the data transmission channel, and when the user enters or leaves the child node, reads the change value of the historical power consumption corresponding to the child node, and establishes a mapping between the user and the change value. After collecting the real-time entry and exit data of the user, integrate the mapping, generate a prediction result, and send the prediction result to a preset terminal.

[0011] Furthermore, the S100 includes:

[0012] Determining a number of input variables, wherein the input variables include at least: temperature, wind speed, and social activity;

[0013] Collect real-time data from IoT devices within energy-consuming units and use the real-time data to update prediction results.

[0014] Furthermore, the S100 further includes:

[0015] Calculate the quartiles and interquartile ranges of historical electricity consumption, summarize and identify outliers, and delete the outliers;

[0016] The time is divided into segments with hours as the unit, a prediction result of each segment is calculated, and the corresponding relationship is synchronized to the segments and the prediction result.

[0017] Further, the S200 includes:

[0018] Using the clustering rule, the sub-nodes are integrated into groups, and a balanced tree is constructed based on the groups;

[0019] The balanced tree is split into a left side, a middle side and a right side, the child nodes are transferred to the left side, and the point positions are transferred to the right side.

[0020] Further, the S300 includes:

[0021] Extracting the maximum and minimum values ​​in the segments, summarizing them to generate a fluctuation range, and using the input variables to offset the fluctuation range;

[0022] The balancing link is introduced into the balancing tree, and the child nodes and points are linked.

[0023] Further, the S400 includes:

[0024] A monitoring mechanism is embedded in the balancing tree, wherein when the real-time power consumption in the child node exceeds the segmented prediction result, the balancing tree tilts to the left;

[0025] The load gap is calculated using the tilt angle of the balancing tree, and a root node is inserted into the bottom of the balancing tree. The root node is used to superimpose the prediction results and load gaps of all segments and send them to a preset terminal.

[0026] Furthermore, the system comprises:

[0027] An output module is used to determine the energy consumption units in the area, define the power consumption of the energy consumption units on the previous day as historical power consumption, input the historical power consumption into a pre-built prediction model, and output the prediction results;

[0028] A construction module is used to split the corresponding relationship between the prediction result and time, define the energy consumption unit as a sub-node, transfer the corresponding relationship to the sub-node, integrate the sub-nodes based on the configured clustering rules, and construct a balanced tree;

[0029] An embedding module, used to classify the subnode into the left side of the balancing tree, create a point position corresponding to the subnode on the right side of the balancing tree, transfer the collected real-time power consumption of the energy-consuming unit into the point position, build a balancing link between the subnode and the point position, determine the fluctuation range of the subnode, and embed the fluctuation range into the balancing link;

[0030] The generation module is used to embed a monitoring mechanism into the balancing tree, read the tilt angle of the balancing tree, and generate a prediction result.

[0031] Furthermore, the output module includes:

[0032] A determination unit, configured to determine a plurality of input variables, wherein the input variables at least include: temperature, wind speed, and social activities;

[0033] An updating unit, used to collect real-time data of IoT devices in energy-consuming units, and update prediction results using the real-time data;

[0034] A deletion unit, used to calculate the quartiles and interquartile ranges of historical electricity consumption, summarize and identify abnormal values, and delete the abnormal values;

[0035] The synchronization unit is used to divide the time into segments with hours as the unit, calculate the prediction result of each segment, and synchronize the corresponding relationship into the segments and the prediction results.

[0036] Furthermore, the building blocks include:

[0037] A building unit, used to integrate the sub-nodes into groups by using the clustering rule, and build a balanced tree based on the groups;

[0038] The transfer unit is used to split the balanced tree into a left side, a middle part and a right side, transfer the child nodes to the left side, and transfer the points to the right side.

[0039] Furthermore, the embedded module includes:

[0040] An offset unit, used to extract the maximum value and the minimum value in the segment, summarize and generate a fluctuation range, and offset the fluctuation range using the input variable;

[0041] The linking unit is used to import the balancing link into the balancing tree and link the child node and the point.

[0042] Compared with the prior art, the present invention has the following beneficial effects:

[0043] By constructing a prediction model, it is possible to use the power consumption of the previous day to predict the power load, and at the same time capture short-term trends and reduce computing costs. By constructing a balancing tree, it is possible to achieve real-time monitoring of the power load, greatly improving the stability of the power grid and optimizing the allocation of power resources. By setting the fluctuation range, the normal load fluctuation range of the power grid can be defined to reduce unnecessary intervention. At the same time, it is also possible to quickly identify anomalies and improve the accuracy of load forecasting. By reading the inclination angle of the balancing tree, the load gap of the power grid can be quantified, thereby providing data support for the normal operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] Figure 1 A flowchart of a power grid load forecasting method based on big data analysis provided by an embodiment of the present invention;

[0045] Figure 2 A first sub-process flowchart of a power grid load forecasting method based on big data analysis provided by an embodiment of the present invention;

[0046] Figure 3 A second sub-process flowchart of the power grid load forecasting method based on big data analysis provided by an embodiment of the present invention;

[0047] Figure 4 A third sub-process flowchart of the power grid load forecasting method based on big data analysis provided by an embodiment of the present invention;

[0048] Figure 5A fourth sub-process flowchart of the power grid load forecasting method based on big data analysis provided by an embodiment of the present invention;

[0049] Figure 6 A block diagram of a power grid load forecasting system based on big data analysis provided by an embodiment of the present invention;

[0050] Figure 7 A block diagram of the composition of an output module in a power grid load prediction system based on big data analysis provided by an embodiment of the present invention;

[0051] Figure 8 A block diagram of the components of the building blocks in the power grid load forecasting system based on big data analysis provided by an embodiment of the present invention;

[0052] Fig. 9 A block diagram of the components of the embedded modules in the power grid load forecasting system based on big data analysis provided by an embodiment of the present invention;

[0053] Fig.10 A block diagram of the composition of a generation module in a power grid load forecasting system based on big data analysis provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] In Example 1, Figure 1 The implementation process of the power grid load prediction method based on big data analysis provided by an embodiment of the present invention is shown and described in detail below:

[0056] S100: Determine energy-consuming units in the region, define the power consumption of the energy-consuming units on the previous day as historical power consumption, input the historical power consumption into a pre-built prediction model, and output a prediction result.

[0057] Determine the energy-consuming units in the area. The energy-consuming units here can be a household or a company, and determine the electricity consumption of the energy-consuming unit on the previous day as the historical electricity consumption. It should be noted that the temperature, wind speed, etc. of the previous day need to be the same. Only in this case, the electricity consumption of the previous day has a reference value. If the weather changes drastically and the temperature difference between two adjacent days is large, it has no reference value. In this case, you can continue to backtrack to find the most recent day with the same weather conditions, and transfer the historical electricity consumption to the prediction model to output the prediction results.

[0058] S200: Split out the corresponding relationship between the prediction result and time, define the energy consumption unit as a child node, and transfer the corresponding relationship to the child node, integrate the child nodes based on the configured clustering rules, and construct a balanced tree.

[0059] The prediction results are refined, and a corresponding relationship between the prediction results and the time points is established. The corresponding relationship can be 300w at 5 a.m. and 1500w at 10 a.m. The energy consumption units are defined as child nodes, and the above corresponding relationship is transferred to the child nodes, and then the child nodes are clustered. The clustering rule is to determine the specific location of each child node, and classify the child nodes based on each building, and create a corresponding balanced tree for each building.

[0060] Among them, the balanced tree is a self-balancing data structure, which is mainly used to efficiently compare data in dynamic sets (child nodes and points); the balanced tree is generally divided into a left and a right side. When the data on the left side is greater than the data on the right side, the balanced tree will tilt to the left.

[0061] S300: classify the sub-node into the left side of the balancing tree, create a point corresponding to the sub-node on the right side of the balancing tree, transfer the collected real-time power consumption of the energy-consuming unit into the point, build a balancing link between the sub-node and the point, determine the fluctuation range of the sub-node, and embed the fluctuation range into the balancing link.

[0062] Transfer the child node to the left side of the balancing tree, and create a point on the right side of the balancing tree, where each child node corresponds to a point. Collect the real-time power consumption of the energy unit and transfer the real-time power consumption to the point. At the same time, use the balancing link to link the child node and its corresponding point, and insert the fluctuation range into the balancing link. The fluctuation range is the normal load fluctuation, which is formulated according to the historical power consumption.

[0063] S400: insert a processing node in the middle of the balancing tree, build a data transmission channel between the processing node and the face recognition device, the face recognition device is set at the entrance and exit of the energy-consuming unit, obtains portrait recognition data through the data transmission channel, and when the user enters or leaves the child node, reads the change value of the historical power consumption corresponding to the child node, and establishes a mapping between the user and the change value. After collecting the real-time entry and exit data of the user, integrate the mapping, generate a prediction result, and send the prediction result to a preset terminal.

[0064] Insert a processing node in the middle of the balancing tree, where the processing node is mainly used to process face recognition data. There is a corresponding relationship between face recognition data and child nodes. By collecting face recognition data at the import and export, the change in the number of users in the child node can be determined, thereby providing data support for power grid load forecasting. When the number of users in a child node changes, the change in the historical power consumption of the child node at this time is read, and a mapping between this change and the user is established. When forecasting the power grid load, when it is monitored that the user enters or leaves the energy-consuming unit, the scheduling of the power grid load can be optimized in advance according to the mapping.

[0065] In Example 2, Figure 2 The implementation process of the power grid load prediction method based on big data analysis provided by an embodiment of the present invention is shown, and S100 is described in detail below, as follows:

[0066] S101: Determine a number of input variables, wherein the input variables at least include: temperature, wind speed, and social activities.

[0067] Determine the input variables. The input variables here can be temperature, wind speed, social activities, etc. The common feature of these input variables is that they will have a significant impact on the grid load.

[0068] S102: Collect real-time data of IoT devices in energy-consuming units, and use the real-time data to update prediction results.

[0069] Obtain real-time data from IoT devices, including smart refrigerators, smart air conditioners, etc. During operation, IoT devices will record their own electricity and power usage. The recorded data can be used to update the prediction results, thereby further improving the accuracy of the prediction results.

[0070] In Example 3, Figure 2 The implementation process of the power grid load prediction method based on big data analysis provided by an embodiment of the present invention is shown, and S100 is further described in detail below, as follows:

[0071] S103: Calculate the quartiles and interquartile ranges of historical electricity consumption, summarize and identify outliers, and delete the outliers.

[0072] According to the historical electricity consumption, calculate the hourly electricity consumption of the energy-consuming unit, and sort the electricity consumption from large to small, and find the quartile from the sorting result, recorded as Q 1 and Q 3 , and Q 3 and Q 1 The difference is defined as the interquartile range, denoted as IQR, which is less than Q 1 -1.5IQR and greater than Q 3The historical electricity consumption of +1.5IQR was determined as an outlier and was deleted.

[0073] S104: Split the time into segments with hours as the unit, calculate the prediction result of each segment, and synchronize the corresponding relationship into the segments and the prediction results.

[0074] The time is divided into several segments according to the hour, and the prediction model and prediction results are used to determine the prediction results for each hour, and the corresponding relationship between the hours and the prediction results is established.

[0075] In Example 4, Figure 3 The implementation process of the power grid load prediction method based on big data analysis provided by an embodiment of the present invention is shown, and S200 is described in detail below, as follows:

[0076] S201: using the clustering rule, integrating the child nodes into groups, and building a balanced tree with the groups as units.

[0077] The above clustering rules are the classification and integration methods of sub-nodes. According to the specific locations of sub-nodes, sub-nodes are divided into several groups with each building as the unit. The group here can be understood as a building or an industrial park. A balanced tree is established for each group.

[0078] S202: Split the balanced tree into a left side, a middle side and a right side, transfer the child nodes to the left side, and transfer the points to the right side.

[0079] The balanced tree is split into a left side, a middle side and a right side, the child nodes are transferred to the left side, the points are transferred to the right side, and balanced links are established between the child nodes and the points.

[0080] In Example 5, Figure 4 The implementation process of the power grid load prediction method based on big data analysis provided by an embodiment of the present invention is shown, and S300 is described in detail below, as follows:

[0081] S301: extracting the maximum value and the minimum value in the segment, summarizing them to generate a fluctuation range, and using the input variable to offset the fluctuation range.

[0082] Determine the maximum and minimum values ​​in each segment. Since the outliers in the segment have been removed, the range composed of the maximum and minimum values ​​is the fluctuation range of the prediction results.

[0083] In real life, after predicting the electricity consumption of energy-consuming units, the energy-consuming units may experience fluctuations in electricity consumption. As long as the electricity consumption of the energy-consuming units is within the normal load fluctuation range, there is no need to make major adjustments to the grid load.

[0084] S302: Import the balancing link into the balancing tree, and link the child nodes and points.

[0085] The balancing links are used to link the sub-nodes and points, and the fluctuation range is embedded in the balancing links; in other words, as long as the real-time power consumption of the energy-consuming unit does not exceed the predicted power consumption and the fluctuation range, the balancing tree will not tilt.

[0086] In Example 6, Figure 5 The implementation process of the power grid load prediction method based on big data analysis provided by an embodiment of the present invention is shown, and S400 is described in detail below, as follows:

[0087] S401: embedding a monitoring mechanism into the balancing tree, wherein the monitoring mechanism is: when the real-time power consumption in the child node exceeds the segmented prediction result, the balancing tree tilts to the left.

[0088] A monitoring mechanism is embedded in the balancing tree, where the monitoring mechanism is: if the real-time power consumption of the energy-consuming unit exceeds the segmented prediction result, the balancing tree tilts to the left.

[0089] S402: Calculate the load gap using the tilt angle of the balancing tree, insert a root node into the bottom of the balancing tree, and use the root node to superimpose the prediction results and load gaps of all segments and send them to a preset terminal.

[0090] According to the inclination angle of the balancing tree, the load gap is calculated. The grid load needs to be adjusted according to this load gap. The root node is inserted into the bottom of the balancing tree, where the root node is mainly used for data processing and the output of the balancing tree; using the root node, the prediction results and load gaps of all segments are calculated, and the calculation results are sent to the preset terminal, where the preset terminal can be the power grid system management terminal.

[0091] Figure 6 The block diagram of the composition structure of the power grid load prediction system based on big data analysis provided by an embodiment of the present invention is shown. The power grid load prediction system based on big data analysis 1 includes:

[0092] The output module 11 is used to determine the energy consumption units in the area, define the power consumption of the energy consumption units on the previous day as the historical power consumption, input the historical power consumption into the pre-built prediction model, and output the prediction result;

[0093] A construction module 12 is used to split the corresponding relationship between the prediction result and time, define the energy consumption unit as a sub-node, transfer the corresponding relationship to the sub-node, integrate the sub-nodes based on the configured clustering rules, and construct a balanced tree;

[0094] The embedding module 13 is used to classify the subnode into the left side of the balancing tree, create a point corresponding to the subnode on the right side of the balancing tree, transfer the collected real-time power consumption of the energy-consuming unit into the point, build a balancing link between the subnode and the point, determine the fluctuation range of the subnode, and embed the fluctuation range into the balancing link;

[0095] The generating module 14 is used to embed a monitoring mechanism into the balancing tree, read the tilt angle of the balancing tree, and generate a prediction result.

[0096] Figure 7 The structure block diagram of the power grid load prediction system based on big data analysis provided by an embodiment of the present invention is shown, and the output module 11 includes:

[0097] A determination unit 111 is used to determine a number of input variables, wherein the input variables at least include: temperature, wind speed and social activities;

[0098] An updating unit 112 is used to collect real-time data of IoT devices in the energy-consuming unit and update the prediction result using the real-time data;

[0099] A deletion unit 113 is used to calculate the quartiles and interquartile ranges of the historical power consumption, summarize and identify abnormal values, and delete the abnormal values;

[0100] The synchronization unit 114 is used to divide the time into segments with hours as the unit, calculate the prediction result of each segment, and synchronize the corresponding relationship into the segments and the prediction results.

[0101] Figure 8 The structure block diagram of the power grid load prediction system based on big data analysis provided by an embodiment of the present invention is shown, and the building module 12 includes:

[0102] A building unit 121 is used to integrate the sub-nodes into groups by using the clustering rule, and build a balanced tree based on the groups;

[0103] The transfer unit 122 is used to split the balanced tree into a left side, a middle part and a right side, transfer the child nodes to the left side, and transfer the points to the right side.

[0104] Fig. 9 The structure block diagram of the power grid load prediction system based on big data analysis provided by an embodiment of the present invention is shown, and the embedded module 13 includes:

[0105] An offset unit 131 is used to extract the maximum value and the minimum value in the segment, summarize them to generate a fluctuation range, and use the input variable to offset the fluctuation range;

[0106] The linking unit 132 is used to import the balancing link into the balancing tree and link the child node and the point.

[0107] Fig.10 The structure block diagram of the power grid load prediction system based on big data analysis provided by an embodiment of the present invention is shown, and the generation module 14 includes:

[0108] The tilting unit 141 is used to embed a monitoring mechanism into the balancing tree, wherein the monitoring mechanism is: when the real-time power consumption in the child node exceeds the segmented prediction result, the balancing tree tilts to the left;

[0109] The sending unit 142 is used to calculate the load gap using the tilt angle of the balancing tree, insert a root node into the bottom of the balancing tree, and use the root node to superimpose the prediction results and load gaps of all segments and send them to a preset terminal.

[0110] The output module 11 is mainly used to complete step S100, the construction module 12 is mainly used to complete step S200, the embedding module 13 is mainly used to complete step S300, and the generation module 14 is mainly used to complete step S400;

[0111] The determining unit 111 is mainly used to complete step S101, the updating unit 112 is mainly used to complete step S102, the deleting unit 113 is mainly used to complete step S103, and the synchronizing unit 114 is mainly used to complete step S104;

[0112] The building unit 121 is mainly used to complete step S201, and the transfer unit 122 is mainly used to complete step S202;

[0113] The offset unit 131 is mainly used to complete step S301, and the link unit 132 is mainly used to complete step S302;

[0114] The tilting unit 141 is mainly used to complete step S401, and the sending unit 142 is mainly used to complete step S402.

[0115] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0116] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0117] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A power grid load forecasting method based on big data analysis, characterized in that: The method comprises: S100: Determine energy-consuming units in the region, define the power consumption of the energy-consuming units on the previous day as historical power consumption, input the historical power consumption into a pre-built prediction model, and output a prediction result; S200: splitting the corresponding relationship between the prediction result and the time, defining the energy consumption unit as a child node, transferring the corresponding relationship to the child node, integrating the child nodes based on the configured clustering rules, and constructing a balanced tree; S300: placing the child node on the left side of the balancing tree, creating a point corresponding to the child node on the right side of the balancing tree, transferring the collected real-time power consumption of the energy-consuming unit to the point, building a balancing link between the child node and the point, determining the fluctuation range of the child node, and embedding the fluctuation range into the balancing link; S400: insert a processing node in the middle of the balancing tree, build a data transmission channel between the processing node and the face recognition device, the face recognition device is set at the entrance and exit of the energy-consuming unit, obtains portrait recognition data through the data transmission channel, and when the user enters or leaves the child node, reads the change value of the historical power consumption corresponding to the child node, and establishes a mapping between the user and the change value. After collecting the real-time entry and exit data of the user, integrate the mapping, generate a prediction result, and send the prediction result to a preset terminal.

2. The power grid load forecasting method based on big data analysis according to claim 1 is characterized in that: The S100 includes: Determining a number of input variables, wherein the input variables include at least: temperature, wind speed, and social activity; Collect real-time data from IoT devices within energy-consuming units and use the real-time data to update prediction results.

3. The power grid load forecasting method based on big data analysis according to claim 1 is characterized in that: The S100 further includes: Calculate the quartiles and interquartile ranges of historical electricity consumption, summarize and identify outliers, and delete the outliers; The time is divided into segments with hours as the unit, a prediction result of each segment is calculated, and the corresponding relationship is synchronized to the segments and the prediction result.

4. The power grid load forecasting method based on big data analysis according to claim 1 is characterized in that: The S200 includes: Using the clustering rule, the child nodes are integrated into groups, and a balanced tree is constructed based on the groups; The balanced tree is split into a left side, a middle side and a right side, the child nodes are transferred to the left side, and the point positions are transferred to the right side.

5. The power grid load forecasting method based on big data analysis according to claim 3 is characterized in that: The S300 includes: Extracting the maximum and minimum values ​​in the segments, summarizing them to generate a fluctuation range, and using the input variables to offset the fluctuation range; The balancing link is introduced into the balancing tree, and the child nodes and points are linked.

6. The power grid load forecasting method based on big data analysis according to claim 5 is characterized in that: The S400 includes: A monitoring mechanism is embedded in the balancing tree, wherein when the real-time power consumption in the child node exceeds the segmented prediction result, the balancing tree tilts to the left; The load gap is calculated using the tilt angle of the balancing tree, and a root node is inserted into the bottom of the balancing tree. The root node is used to superimpose the prediction results and load gaps of all segments and send them to a preset terminal.

7. The power grid load forecasting system based on big data analysis is characterized by: The system comprises: An output module is used to determine the energy consumption units in the area, define the power consumption of the energy consumption units on the previous day as historical power consumption, input the historical power consumption into a pre-built prediction model, and output the prediction results; A construction module is used to split the corresponding relationship between the prediction result and time, define the energy consumption unit as a sub-node, transfer the corresponding relationship to the sub-node, integrate the sub-nodes based on the configured clustering rules, and construct a balanced tree; An embedding module, used to classify the subnode into the left side of the balancing tree, create a point position corresponding to the subnode on the right side of the balancing tree, transfer the collected real-time power consumption of the energy-consuming unit into the point position, build a balancing link between the subnode and the point position, determine the fluctuation range of the subnode, and embed the fluctuation range into the balancing link; A generation module is used to insert a processing node in the middle of the balancing tree, build a data transmission channel between the processing node and the face recognition device, the face recognition device is set at the entrance and exit of the energy-consuming unit, and obtains portrait recognition data through the data transmission channel. When a user enters or leaves a subnode, the change value of the historical power consumption corresponding to the subnode is read out, and a mapping between the user and the change value is established. After collecting the real-time entry and exit data of the user, the mapping is integrated to generate a prediction result, and the prediction result is sent to a preset terminal.

8. According to the power grid load forecasting system based on big data analysis according to claim 7, the output module comprises: A determination unit, configured to determine a plurality of input variables, wherein the input variables at least include: temperature, wind speed, and social activities; An updating unit, used to collect real-time data of IoT devices in energy-consuming units, and update prediction results using the real-time data; A deletion unit, used to calculate the quartiles and interquartile ranges of historical electricity consumption, summarize and identify abnormal values, and delete the abnormal values; The synchronization unit is used to divide the time into segments with hours as the unit, calculate the prediction result of each segment, and synchronize the corresponding relationship into the segments and the prediction results.

9. According to the power grid load forecasting system based on big data analysis according to claim 7, the building module comprises: A building unit, used to integrate the sub-nodes into groups by using the clustering rule, and build a balanced tree based on the groups; The transfer unit is used to split the balanced tree into a left side, a middle part and a right side, transfer the child nodes to the left side, and transfer the points to the right side.

10. The power grid load forecasting system based on big data analysis according to claim 8, wherein the embedded module comprises: An offset unit, used to extract the maximum value and the minimum value in the segment, summarize and generate a fluctuation range, and offset the fluctuation range using the input variable; The linking unit is used to import the balancing link into the balancing tree and link the child node and the point.

Images