Power grid load prediction system and method

Abstract

The invention relates to the field of power grid load prediction, in particular to a power grid load prediction system and method, and the system comprises a data collection module, a data processing module, and a load prediction module. The data acquisition module is used for acquiring electric power system operation data Dsj, meteorological data Qsj and calendar information data Rsj; the data processing module is used for preprocessing the collected data, including data cleaning and data attribution standardization; by comprehensively considering the power consumption proportions of industrial, resident and commercial users, meteorological data and holiday and festival information in the operation of the power system, the factors influencing the power grid load can be more comprehensively captured, and the fusion of the multi-dimensional data enables the prediction model to better adapt to the complex and changeable power consumption environment. And prediction errors caused by insufficient consideration of a single factor are reduced, so that the load change of the power grid is predicted more accurately.

Description

Technical Field

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

[0002] With the rapid economic development and continuous social progress, the scale and complexity of the power system are continuing to grow. Today, the power network covers a wide range, connecting a large number of power generation, transmission, transformation, distribution and power consumption equipment. In such a huge system, the balance between power supply and demand has become a key factor in ensuring the stable operation of the power system. Therefore, predicting the grid load in advance can better manage the grid power system.

[0003] In the operation of the power system, different types of users have different electricity consumption behaviors. The production plans of industrial users, the daily living habits of residential users, and the business hours of commercial users will lead to changes in their electricity consumption ratios, which in turn affect the load of the power grid. At the same time, the impact of meteorological conditions on the load of the power grid cannot be ignored. Meteorological factors such as temperature, humidity, wind speed and light intensity are closely related to the power load. Extreme temperatures will prompt the extensive use of air conditioning, heating and other equipment, and humidity changes may affect the operation of dehumidification or humidification equipment; wind speed has a direct impact on the amount of wind power connected to the power grid, thereby changing the supply and demand balance of the power grid; light intensity affects photovoltaic power generation and lighting equipment electricity consumption. Holiday factors also play an important role in the prediction of power grid load. During statutory holidays and special holidays, the social electricity consumption pattern will change greatly, industrial enterprises may stop production or reduce production, and the time and intensity of commercial and residential activities will also change, all of which will cause the load of the power grid to be very different from that on weekdays;

[0004] Existing power grid load forecasting systems often fail to fully consider the combined impact of these factors, or lack effective integration and analysis methods when considering these factors, resulting in low accuracy of power grid load forecasting. For this reason, this case proposes a power grid load forecasting system and method. Summary of the invention

[0005] In order to overcome the above-mentioned technical problems, the purpose of the present invention is to provide a power grid load forecasting system and method: by comprehensively considering the proportion of electricity consumption of industrial, residential and commercial users in the operation of the power system, meteorological data and holiday information, it is possible to more comprehensively capture the factors affecting the power grid load. The fusion of such multi-dimensional data enables the prediction model to better adapt to the complex and changeable power consumption environment, reduce the prediction error caused by insufficient consideration of a single factor, and thus more accurately predict the changes in power grid load, solving the problem that the existing power grid load forecasting system often fails to fully consider the comprehensive impact of these factors, or lacks effective integration and analysis methods when considering these factors, resulting in low accuracy of power grid load forecasting.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A power grid load forecasting system, comprising: a data acquisition module, a data processing module, and a load forecasting module;

[0008] The data acquisition module is used to collect power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj;

[0009] The data processing module is used to pre-process the collected data, including data cleaning and data standardization;

[0010] The load prediction module is used to collect features for predicting power grid load in combination with processed power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj, and predict the power grid load through these features.

[0011] As a further solution of the present invention: the process of collecting the characteristics of the power system operation data Dsj is as follows:

[0012] Extract the grid load data of different time scales of the past day, week and month from the power system database, arrange them in chronological order, and calculate the maximum, minimum and average values ​​contained in the data; collect the power consumption data of different types of users, including industrial users Gy, residential users Jy, and commercial users Sy, record the identification and corresponding power consumption, and also arrange the data in chronological order to correspond to the historical load data; find out the maximum load of each day in the past week and use it as a feature, and further find out the maximum load and minimum load in the past month; calculate the power consumption ratio of different types of users, and the calculation formula is: power consumption ratio β = power consumption of a certain type of user / total power consumption, record the power consumption ratio of each user in real time, the power consumption ratio of industrial user Gy is marked as β1, the power consumption ratio of residential user Jy is marked as β2, and the power consumption ratio of commercial user Sy is marked as β3, and converted into a line graph by computer, and the line graph of power grid load changes is recorded in real time, and the relationship between them is analyzed by observing the synchronization or lag between the changes in the power consumption ratio β of different types of users and the overall load changes.

[0013] As a further solution of the present invention: by observing the synchronization or hysteresis of the change of the electricity consumption ratio β of different types of users and the change of the overall load, the relationship between them is analyzed, and the process is as follows:

[0014] If β1 increases significantly in a certain period of time, and the overall load also reaches a peak during this period of time, then it can be inferred that β1 has an important impact on the overall load change, otherwise the impact is small;

[0015] If β1 decreases significantly in a certain period of time, and the overall load also decreases during this period of time, it can be inferred that the electricity consumption of industrial users has a significant impact on the overall load change, otherwise the impact is small;

[0016] If β1 remains unchanged for a period of time, while the overall load changes significantly, it means that the change in the overall load is mainly caused by β2 or β3;

[0017] According to the same analysis method, β2 and β3 are substituted in turn for analysis, and finally the electricity consumption ratio of industrial user Gy β1, the electricity consumption ratio of residential user Jy β2, and the electricity consumption ratio of commercial user Sy β3 are obtained. The fluctuations in a certain period of time have a fluctuating impact on the electricity load.

[0018] As a further solution of the present invention: the characteristic collection process of the meteorological data Qsj is to record the temperature, humidity, wind speed, light intensity in the power supply area environment in real time, and simultaneously record the power grid load data.

[0019] As a further solution of the present invention: the calendar information data Rsj feature collection process is to create a holiday table based on historical data, which includes statutory holidays and special holidays, wherein statutory holidays are holidays stipulated by the state, and special holidays are holidays for special ethnic groups or special enterprises, with 0 representing non-holidays and 1 representing holidays, and at the same time, the grid load changes on holidays in historical data.

[0020] As a further solution of the present invention: the specific processing process of the load forecasting module is as follows:

[0021] Construct a decision model, starting with "Is it a holiday" as the root node. If it is a holiday, further divide the nodes according to the characteristics of the power grid load changes during historical holidays, including the differences in load changes during different types of holidays. If it is not a holiday, divide the nodes according to the user's electricity consumption ratio and meteorological data. At each leaf node, determine the corresponding power grid load forecast value based on historical data;

[0022] For new data, starting from the root node, according to the characteristics of the data, whether it is a holiday, the proportion of user electricity consumption, and meteorological data, judgment is made along the branches of the decision model until the leaf node is reached. The load forecast value corresponding to the leaf node is the forecast result of the power grid load. The changes in the power grid load are analyzed by comparing the forecast results at different times.

[0023] A power grid load forecasting method, the specific steps are as follows:

[0024] Step 1: Collect the power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj through the data acquisition module, and clean and standardize the data through the data processing module to improve the data;

[0025] Step 2: Combine the processed power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj, and collect the characteristics of the power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj respectively. Establish a decision model based on the obtained characteristics, and make judgments along the branches of the decision model until reaching the leaf node. The load prediction value corresponding to the leaf node is the prediction result of the power grid load.

[0026] Beneficial effects of the present invention:

[0027] By comprehensively considering the proportion of electricity consumption by industrial, residential and commercial users in the operation of the power system, meteorological data and holiday information, we can more comprehensively capture the factors affecting the power grid load. This fusion of multi-dimensional data enables the prediction model to better adapt to the complex and changeable power consumption environment, reduce the prediction error caused by insufficient consideration of a single factor, and thus more accurately predict the changes in power grid load, providing strong support for the stable operation of the power system. Through more accurate predictions of multi-dimensional data, more power generation resources can be dispatched online in advance, including starting standby generators and coordinating new energy power generation to meet power demand and avoid power outages caused by insufficient power supply. On the contrary, during low load periods, such as when industry is shut down during holidays and the weather is mild, the power generation capacity can be reasonably reduced to reduce power generation costs and improve energy efficiency.

[0028] Accurate load forecasting enables grid dispatchers to prepare in advance to deal with possible grid overload or voltage instability. When it is predicted that the grid load will exceed the safe operating range, timely measures can be taken, such as adjusting the grid topology and implementing load control to prevent grid failures, ensure the safe and reliable operation of the grid, and reduce the losses caused by grid failures to society and the economy.

[0029] In the electricity market environment, power generation companies can optimize their bidding strategies based on predicted load conditions and improve their market competitiveness; power sales companies can better plan their electricity purchasing plans and reduce electricity purchasing costs; and power regulatory agencies can also formulate more scientific market supervision policies based on load forecast information to promote the healthy development of the electricity market. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] The present invention will be further described below in conjunction with the accompanying drawings.

[0031] Figure 1 It is a principle block diagram of a power grid load forecasting system in the present invention. DETAILED DESCRIPTION

[0032] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0033] Embodiment 1:

[0034] See also Figure 1 As shown, this embodiment is a power grid load forecasting system, including: a data acquisition module, a data processing module, and a load forecasting module;

[0035] Data acquisition module, used to collect power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj;

[0036] The data processing module is used to pre-process the collected data, including data cleaning and data standardization;

[0037] The load forecasting module is used to collect the features used to forecast the grid load by combining the processed power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj, and forecast the grid load by using these features.

[0038] The process of collecting the characteristics of power system operation data Dsj is as follows:

[0039] Extract the grid load data of different time scales of the past day, week and month from the power system database, arrange them in chronological order, and calculate the maximum, minimum and average values ​​contained in the data; collect the power consumption data of different types of users, including industrial users Gy, residential users Jy, and commercial users Sy, record the identification and corresponding power consumption, and also arrange the data in chronological order to correspond to the historical load data; find out the maximum load of each day in the past week and use it as a feature, and further find out the maximum load and minimum load in the past month; calculate the power consumption ratio of different types of users, and the calculation formula is: power consumption ratio β = power consumption of a certain type of user / total power consumption, record the power consumption ratio of each user in real time, the power consumption ratio of industrial user Gy is marked as β1, the power consumption ratio of residential user Jy is marked as β2, and the power consumption ratio of commercial user Sy is marked as β3, and converted into a line graph by computer, and the line graph of power grid load changes is recorded in real time, and the relationship between them is analyzed by observing the synchronization or lag between the changes in the power consumption ratio β of different types of users and the overall load changes.

[0040] By observing the synchronization or lag between the change in the electricity consumption ratio β of different types of users and the change in the overall load, the relationship between them is analyzed. The process is as follows:

[0041] If β1 increases significantly in a certain period of time, and the overall load also reaches a peak during this period of time, then it can be inferred that β1 has an important impact on the overall load change, otherwise the impact is small;

[0042] If β1 decreases significantly in a certain period of time, and the overall load also decreases during this period of time, it can be inferred that the electricity consumption of industrial users has a significant impact on the overall load change, otherwise the impact is small;

[0043] If β1 remains unchanged for a period of time, while the overall load changes significantly, it means that the change in the overall load is mainly caused by β2 or β3;

[0044] According to the same analysis method, β2 and β3 are substituted in turn for analysis, and finally the electricity consumption ratio of industrial user Gy β1, the electricity consumption ratio of residential user Jy β2, and the electricity consumption ratio of commercial user Sy β3 are obtained. The fluctuations in a certain period of time have a fluctuating impact on the electricity load.

[0045] The feature collection process of meteorological data Qsj is to record the temperature, humidity, wind speed, and light intensity in the power supply area in real time, and record the power grid load data at the same time.

[0046] The process of collecting calendar information data Rsj features is to create a holiday table based on historical data, which includes statutory holidays and special holidays. Statutory holidays are holidays stipulated by the state, and special holidays are holidays for special ethnic groups or special enterprises. 0 represents non-holidays and 1 represents holidays. At the same time, the grid load changes on holidays in historical data.

[0047] The specific processing process of the load forecasting module is as follows:

[0048] Construct a decision model, starting with "Is it a holiday" as the root node. If it is a holiday, further divide the nodes according to the characteristics of the power grid load changes during historical holidays, including the differences in load changes during different types of holidays. If it is not a holiday, divide the nodes according to the user's electricity consumption ratio and meteorological data. At each leaf node, determine the corresponding power grid load forecast value based on historical data;

[0049] For new data, starting from the root node, according to the characteristics of the data, whether it is a holiday, the proportion of user electricity consumption, and meteorological data, judgment is made along the branches of the decision model until the leaf node is reached. The load forecast value corresponding to the leaf node is the forecast result of the power grid load. By comparing the forecast results at different times, the calculation process of analyzing the change of power grid load is as follows:

[0050] Decision model: Is it a holiday?

[0051] Root node: whether it is a holiday;

[0052] It’s a holiday;

[0053] Node 1: Holiday type;

[0054] Subnode 1.1: statutory holidays;

[0055] Subnode 1.1.1: The load increases significantly;

[0056] Strategy: Increase grid capacity in advance and optimize power supply strategy;

[0057] Subnode 1.1.2: The load is slightly increased;

[0058] Strategy: Monitor grid load and prepare for emergencies;

[0059] Subnode 1.2: Illegal holidays;

[0060] Subnode 1.2.1: The load increased slightly;

[0061] Strategy: Maintain conventional power grid operation strategy;

[0062] Subnode 1.2.2: No significant change in load;

[0063] Strategy: Continue to monitor grid load;

[0064] It’s not a holiday;

[0065] Node 2: User electricity consumption ratio;

[0066] Subnode 2.1: High electricity consumption ratio;

[0067] Subnode 2.1.1: meteorological data;

[0068] Subnode 2.1.1.1: high temperature;

[0069] Strategy: Strengthen heat dissipation measures to ensure stable operation of equipment;

[0070] Subnode 2.1.1.2: low temperature;

[0071] Strategy: Check heating equipment and ensure power supply safety;

[0072] Subnode 2.1.2: meteorological data;

[0073] Subnode 2.1.2.1: Sunny day;

[0074] Strategy: Maintaining conventional power grid operation strategy;

[0075] Subnode 2.1.2.2: Rainy day;

[0076] Strategy: Strengthen line inspections to prevent failures;

[0077] Subnode 2.2: Low power consumption ratio;

[0078] Subnode 2.2.1: meteorological data;

[0079] Subnode 2.2.1.1: high temperature;

[0080] Strategy: Monitor grid load and prepare for emergencies;

[0081] Subnode 2.2.1.2: low temperature;

[0082] Strategy: Maintaining conventional power grid operation strategy;

[0083] Subnode 2.2.2: meteorological data;

[0084] Subnode 2.2.2.1: Sunny day;

[0085] Strategy: Continue to monitor grid load;

[0086] Subnode 2.2.2.2: Rainy day;

[0087] Strategy: Strengthen line inspections to prevent failures. . . .

[0091] A power grid load forecasting method, the specific steps are as follows:

[0092] Step 1: Collect the power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj through the data acquisition module, and clean and standardize the data through the data processing module to improve the data;

[0093] Step 2: Combine the processed power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj, and collect the characteristics of the power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj respectively. Establish a decision model based on the obtained characteristics, and make judgments along the branches of the decision model until reaching the leaf node. The load prediction value corresponding to the leaf node is the prediction result of the power grid load.

[0094] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0095] The above contents are merely examples and explanations of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the invention or exceed the scope defined by the claims, they shall all fall within the protection scope of the present invention.

Claims

1. A power grid load forecasting system, characterized in that: include: Data acquisition module, data processing module, load forecasting module; The data acquisition module is used to collect power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj; The data processing module is used to pre-process the collected data, including data cleaning and data standardization; The load prediction module is used to collect features for predicting power grid load in combination with processed power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj, and predict the power grid load through these features.

2. A power grid load forecasting system according to claim 1, characterized in that: The process of collecting the characteristics of the power system operation data Dsj is as follows: Extract the grid load data of different time scales of the past day, week and month from the power system database, arrange them in chronological order, and calculate the maximum, minimum and average values ​​contained in the data; collect the power consumption data of different types of users, including industrial users Gy, residential users Jy, and commercial users Sy, record the identification and corresponding power consumption, and also arrange the data in chronological order to correspond to the historical load data; find out the maximum load of each day in the past week and use it as a feature, and further find out the maximum load and minimum load in the past month; calculate the power consumption ratio of different types of users, and the calculation formula is: power consumption ratio β = power consumption of a certain type of user / total power consumption, record the power consumption ratio of each user in real time, the power consumption ratio of industrial user Gy is marked as β1, the power consumption ratio of residential user Jy is marked as β2, and the power consumption ratio of commercial user Sy is marked as β3, and converted into a line graph by computer, and the line graph of power grid load changes is recorded in real time, and the relationship between them is analyzed by observing the synchronization or lag between the changes in the power consumption ratio β of different types of users and the overall load changes.

3. A power grid load forecasting system according to claim 2, characterized in that: By observing the synchronization or lag between the change in the electricity consumption ratio β of different types of users and the change in the overall load, the relationship between them is analyzed. The process is as follows: If β1 increases significantly in a certain period of time, and the overall load also reaches a peak during this period of time, then it can be inferred that β1 has an important impact on the overall load change, otherwise the impact is small; If β1 decreases significantly in a certain period of time, and the overall load also decreases during this period of time, it can be inferred that the electricity consumption of industrial users has a significant impact on the overall load change, otherwise the impact is small; If β1 remains unchanged for a period of time, while the overall load changes significantly, it means that the change in the overall load is mainly caused by β2 or β3; According to the same analysis method, β2 and β3 are substituted in turn for analysis, and finally the electricity consumption ratio of industrial user Gy β1, the electricity consumption ratio of residential user Jy β2, and the electricity consumption ratio of commercial user Sy β3 are obtained. The fluctuations in a certain period of time have a fluctuating impact on the electricity load.

4. A power grid load forecasting system according to claim 1, characterized in that: The meteorological data Qsj feature collection process is to record the temperature, humidity, wind speed, and light intensity in the power supply area environment in real time, and simultaneously record the power grid load data.

5. A power grid load forecasting system according to claim 1, characterized in that: The calendar information data Rsj feature collection process is to create a holiday table based on historical data, which includes statutory holidays and special holidays, where statutory holidays are holidays stipulated by the state, and special holidays are holidays for special ethnic groups or special enterprises. 0 represents non-holidays and 1 represents holidays. At the same time, the grid load changes on holidays in historical data.

6. A power grid load forecasting system according to claim 1, characterized in that: The specific processing process of the load forecasting module is as follows: Construct a decision model, starting with "Is it a holiday" as the root node. If it is a holiday, further divide the nodes according to the characteristics of the power grid load changes during historical holidays, including the differences in load changes during different types of holidays. If it is not a holiday, divide the nodes according to the user's electricity consumption ratio and meteorological data. At each leaf node, determine the corresponding power grid load forecast value based on historical data; For new data, starting from the root node, according to the characteristics of the data, whether it is a holiday, the proportion of user electricity consumption, and meteorological data, judgment is made along the branches of the decision model until the leaf node is reached. The load forecast value corresponding to the leaf node is the forecast result of the power grid load. The changes in the power grid load are analyzed by comparing the forecast results at different times.

7. A method for predicting power grid load, characterized in that: The method is based on the power grid load forecasting system according to any one of claims 1 to 6, and the specific steps are as follows: Step 1: Collect the power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj through the data acquisition module, and clean and standardize the data through the data processing module to improve the data; Step 2: Combine the processed power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj, and collect the characteristics of the power system operation data Dsj, meteorological data Qsj, and calendar information data Rsj respectively. Establish a decision model based on the obtained characteristics, and make judgments along the branches of the decision model until reaching the leaf node. The load prediction value corresponding to the leaf node is the prediction result of the power grid load.

Images