Power load prediction method based on air temperature and power load correlation analysis

Abstract

The invention discloses a power load prediction method based on air temperature and power load correlation analysis, and relates to the technical field of power load prediction, and the method comprises the steps: obtaining the daily maximum power load data in a period of time, and the air temperature data in the period of time; determining the typical temperature of the electrical load not containing the air conditioning load; the daily mean value of the daily maximum load corresponding to the typical air temperature every year is determined, and the reference load value without the air conditioner load every year is obtained; subtracting the current-year reference load value from the daily maximum load to obtain a daily air-conditioning load value; according to the air temperature data and the air conditioner load value data, constructing a unary quadratic regression equation, wherein the unary quadratic regression equation takes the air temperature as an independent variable x and takes an air conditioner load predicted value as a dependent variable y; acquiring forecast air temperature data of the meteorological website; and substituting the forecast air temperature data into the unary quadratic regression equation to obtain an air conditioner load forecast value. And the future electrical load is predicted by establishing the correlation between the air temperature and the electrical load.

Description

Technical Field

[0001] The present invention relates to the technical field of electric load forecasting, and particularly relates to an electric load forecasting method based on the correlation analysis of temperature and electric load. Background Art

[0002] The centralized turning on of air-conditioning loads is a very important factor leading to the increase of peak-valley difference. The increase of peak-valley difference brings many hidden dangers to the safe and economic operation of the power grid. Therefore, strengthening the analysis of the maximum load of the power grid during peak periods is beneficial to improving the accuracy of load forecasting and laying a foundation for ensuring the safe, stable and economic operation of the power grid.

[0003] Based on the daily maximum load data and temperature data of Hefei City, the base load and air-conditioning load are quantified, and the correlation between the maximum load and temperature is analyzed based on a regression model, and a research conclusion on the correlation between the maximum load and temperature is obtained. Based on this, the present application proposes an electric load forecasting method based on the correlation analysis of temperature and electric load. Summary of the Invention

[0004] The present invention provides an electric load forecasting method based on the correlation analysis of temperature and electric load, and forecasts the future electric load by establishing the correlation between temperature and power consumption load.

[0005] According to one aspect of the present disclosure, there is provided an electric load forecasting method based on the correlation analysis of temperature and electric load, characterized in that the method includes: Obtain the daily maximum power consumption load data within a period of time, and the temperature data within the period of time; Determine the typical temperature when the power consumption load does not include the air-conditioning load; Determine the daily average value of the daily maximum load corresponding to the typical temperature every year, and obtain the base load value without air-conditioning load for each year; The daily air-conditioning load value can be obtained by subtracting the annual base load value from the daily maximum load; Construct a quadratic regression equation according to the temperature data and the air-conditioning load value data, where the quadratic regression equation takes the temperature as the independent variable x and the predicted air-conditioning load value as the dependent variable y; Obtain the predicted temperature data of the meteorological website; Substitute the predicted temperature data into the quadratic regression equation to obtain the predicted air-conditioning load value.

[0006] In a possible implementation manner, the power consumption load includes a base load insensitive to temperature and an air-conditioning load sensitive to temperature, and the air-conditioning load sensitive to temperature includes a heating load or a cooling load.

[0007] In a possible implementation manner, the calculation of the goodness of fit of the quadratic regression equation is as shown in formula (1): (1) Among them, R 2 represents the goodness of fit, which is used to represent the prediction accuracy of the quadratic regression equation. y represents the actual value of the air-conditioning load, represents the predicted value of the air-conditioning load, represents the average value of the actual air-conditioning load.

[0008] In a possible implementation manner, the quadratic regression equation is as follows: (2) Among them, the coefficients a, b, and c in the formula are coefficients fitted according to the temperature data and the air-conditioning load value data. In a possible implementation manner, a quadratic regression equation is constructed according to the temperature data and the air-conditioning load value data. The quadratic regression equation takes the temperature as the independent variable and the predicted value of the air-conditioning load as the dependent variable, including: Construct a quadratic regression equation according to the lowest temperature data and the heating load value data. The quadratic regression equation takes the lowest temperature as the independent variable and the predicted value of the heating load as the dependent variable.

[0009] In a possible implementation manner, a quadratic regression equation is constructed according to the temperature data and the air-conditioning load value data. The quadratic regression equation takes the temperature as the independent variable and the predicted value of the air-conditioning load as the dependent variable, including: Construct a quadratic regression equation according to the highest temperature data and the heating load value data. The quadratic regression equation takes the highest temperature as the independent variable and the predicted value of the heating load as the dependent variable.

[0010] In a possible implementation manner, a quadratic regression equation is constructed according to the temperature data and the air-conditioning load value data. The quadratic regression equation takes the temperature as the independent variable and the predicted value of the air-conditioning load as the dependent variable, including: Construct a quadratic regression equation according to the average temperature data and the heating load value data. The quadratic regression equation takes the average temperature as the independent variable and the predicted value of the heating load as the dependent variable.

[0011] In a possible implementation manner, a quadratic regression equation is constructed according to the temperature data and the air-conditioning load value data. The quadratic regression equation takes the temperature as the independent variable and the predicted value of the air-conditioning load as the dependent variable, including: Construct a quadratic regression equation according to the lowest temperature data and the cooling load value data. The quadratic regression equation takes the lowest temperature as the independent variable and the predicted value of the cooling load as the dependent variable.

[0012] In a possible implementation manner, a quadratic regression equation is constructed according to the temperature data and the air-conditioning load value data. The quadratic regression equation takes the temperature as the independent variable and the predicted value of the air-conditioning load as the dependent variable, including: Construct a quadratic regression equation based on the maximum temperature data and the cooling load value data. The quadratic regression equation uses the maximum temperature as the independent variable and the predicted cooling load value as the dependent variable.

[0013] In a possible implementation, constructing a quadratic regression equation based on the temperature data and the air-conditioning load value data, where the quadratic regression equation uses the temperature as the independent variable and the predicted air-conditioning load value as the dependent variable, includes: Construct a quadratic regression equation based on the average temperature data and the cooling load value data. The quadratic regression equation uses the average temperature as the independent variable and the predicted cooling load value as the dependent variable.

[0014] Compared with the prior art, the beneficial effects of the present invention are: A power load prediction method based on the correlation analysis of temperature and power load according to an embodiment of the present disclosure constructs quadratic regression equations of heating load, cooling load with respect to the maximum temperature, minimum temperature, and average temperature respectively according to the daily maximum power consumption load data within a period of time and the temperature data within the same period of time, and calculates the goodness of fit (R²) of the regression equations. The goodness of fit between the cooling load and the temperature is higher than that of the heating load, indicating that the sensitivity of the summer air-conditioning load to temperature is higher. By comparing the R² values of the three types of temperatures, it can be seen that the goodness of fit of the regression equation between the heating load, cooling load and the average temperature is higher, and the average temperature has a stronger explanatory power for the air-conditioning load. Next, we will conduct an in-depth analysis of the correlation between the air-conditioning load and the temperature based on the average temperature. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 A flowchart showing a power load prediction method based on the correlation analysis of temperature and power load according to an embodiment of the present disclosure.

[0016] Figure 2 A fitting curve graph showing the relationship between the air-conditioning heating load and the average temperature according to an embodiment of the present disclosure.

[0017] Figure 3 A fitting curve graph showing the relationship between the air-conditioning cooling load and the average temperature according to an embodiment of the present disclosure.

[0018] Figure 4 A curve graph showing the relationship between the predicted maximum load value and the true value based on the 15-day forecast temperature according to an embodiment of the present disclosure.

[0019] Figure 5 A curve graph showing the relationship between the predicted maximum load value and the true value based on the 15-day actual temperature according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0020] Various exemplary embodiments, features, and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. Identical reference numerals in the drawings denote functionally identical or similar elements. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0021] The term "exemplary" used herein means "serving as an example, embodiment, or illustration". Any embodiment described herein as "exemplary" is not necessarily to be construed as superior or better than other embodiments.

[0022] In addition, for a better description of the present disclosure, numerous specific details are given in the following detailed description. Those skilled in the art should understand that the present disclosure can be implemented without some specific details. In some instances, methods, means, elements, and circuits well known to those skilled in the art are not described in detail so as to highlight the gist of the present disclosure.

[0023] According to an aspect of the present disclosure, there is provided a power load forecasting method based on the correlation analysis of air temperature and power load, characterized in that the method includes: S01, obtaining the daily maximum power load data within a period of time and the air temperature data within the period of time; S02, determining the typical air temperature when the power load does not include the air-conditioning load; S03, determining the daily average value of the daily maximum load corresponding to the typical air temperature each year to obtain the annual benchmark load value without the air-conditioning load each year; S04, subtracting the annual benchmark load value from the daily maximum load to obtain the daily air-conditioning load value; S05, constructing a quadratic regression equation based on the air temperature data and the air-conditioning load value data, with the air temperature as the independent variable x and the predicted air-conditioning load value as the dependent variable y; S06, obtaining the predicted air temperature data from a meteorological website; S07, substituting the predicted air temperature data into the quadratic regression equation to obtain the predicted air-conditioning load value.

[0024] The power load is divided into two parts. One part is the temperature-insensitive load (benchmark load), and the other part is the temperature-sensitive load (air-conditioning load). Industrial loads and urban and rural basic power loads are temperature-insensitive loads, that is, the basic loads for production and life, which are not affected by seasonal weather; the temperature-sensitive load is the heating and cooling load, which is greatly affected by seasonal weather. Therefore, the calculation formula for the maximum load can be expressed as: y = e + f, where y is the maximum load, e is the benchmark load in the maximum load, and f is the air-conditioning load in the maximum load.

[0025] In a possible implementation, the electrical load includes a baseline load insensitive to temperature and an air-conditioning load sensitive to temperature. The air-conditioning load sensitive to temperature includes a heating load or a cooling load.

[0026] In a possible implementation, the calculation of the goodness of fit of the quadratic regression equation is as shown in Equation (1): (1) where R 2 represents the goodness of fit, and the goodness of fit is used to represent the prediction accuracy of the quadratic regression equation. y represents the actual value of the air-conditioning load, represents the predicted value of the air-conditioning load, represents the average value of the actual values of the air-conditioning load.

[0027] In a possible implementation, the quadratic regression equation is as follows: (2) where a, b, and c in the equation are coefficients fitted based on temperature data and air-conditioning load value data.

[0028] For example, based on the daily maximum load data and temperature data of Hefei City, the baseline load and the air-conditioning load are quantified. The correlation between the maximum load and temperature is analyzed based on the regression model, and the research conclusion on the correlation between the maximum load and temperature is obtained. The air-conditioning load and the maximum load for the next 15 days are predicted and evaluated to assist the energy supply guarantee work for peak summer in Hefei.

[0029] Based on the daily maximum load data of Hefei City from January 1, 2020 to May 15, 2023 of the Anhui Electric Power Dispatching and Control Center and the temperature data of Hefei City from January 1, 2020 to May 30, 2023 of the meteorological website.

[0030] Generally, it is considered that the daily maximum load is generally low in a certain temperature range in spring or autumn, and the electrical load does not include the air-conditioning load. Therefore, this temperature range is found as the typical temperature, and the daily average value of the daily maximum load corresponding to the typical temperature of each year is taken to obtain the baseline load value without the air-conditioning load for that year. After calculation, the typical temperature is about 17°C, and the baseline loads from 2020 to 2023 are 4.4927 million kilowatts, 5.0173 million kilowatts, 5.3418 million kilowatts, and 5.8353 million kilowatts respectively. Finally, the daily air-conditioning load value can be obtained by subtracting the baseline load of that year from the daily maximum load.

[0031] From 2020 to 2022, the maximum load in Hefei showed a curvilinear growth trend. Among them, the maximum load in 2022 reached 10.3706 million kilowatts, a year-on-year increase of 22.18%. From 2020 to 2022, the overall maximum cooling load also showed an increasing trend, and in most years, the cooling load was significantly higher than the heating load. The cooling of air conditioners in summer became the main reason for the growth of the annual maximum load.

[0032] The base load is equal to the sum of the maximum loads per day at the typical temperature divided by the number of days. In 2020, there were 12 days with an average temperature of around 17°C (the typical temperature). The sum of the maximum loads on these 12 days divided by the number of days 12 is the base load in 2020.

[0033] Quadratic regression equations of heating load, cooling load with the highest temperature, the lowest temperature, and the average temperature were constructed respectively, and the goodness of fit (R²) of the regression equations was calculated. The results are shown in Table 1. The goodness of fit between the cooling load and temperature is higher than that of the heating load, and the sensitivity of the summer air-conditioning load to temperature is higher. By comparing the R² values of the three types of temperatures, it can be seen that the regression equation between the heating load, cooling load and the average temperature has a higher goodness of fit, and the average temperature has a stronger explanatory power for the air-conditioning load. Next, we will conduct an in-depth analysis of the correlation between the air-conditioning load and temperature based on the average temperature.

[0034] Table 1 Fitting of regression equations between air-conditioning load and different temperature types ; Analyze the correlation between the heating load and the average temperature. From the fitting Figure 2 it can be seen that there is a negative correlation between the heating load and the average temperature, and the heating load increases as the temperature drops.

[0035] To further clarify the relationship between the heating load and the average temperature, a segmented fitting analysis was carried out on the two. The results show that when the average temperature is lower than 17°C, the heating load begins to rise; but when the temperature is between 10 and 17°C, the heating load increases slowly. For every 1°C drop in temperature, the heating load increases by about 14,700 kilowatts; when the temperature is between 2 and 10°C, the growth rate of the heating load begins to rise. For every 1°C drop in temperature, the heating load increases by about 191,000 kilowatts; when the temperature is lower than 2°C, the growth rate of the heating load reaches the highest. For every 1°C drop in temperature, the heating load increases by about 204,400 kilowatts.

[0036] Analyze the correlation between the cooling load and the average temperature. From the fitting Figure 3 it can be seen that there is a positive correlation between the cooling load and the average temperature, and the cooling load increases as the temperature rises.

[0037] The segmented fitting analysis of the cooling load and the average temperature shows that when the average temperature is higher than 17°C, the cooling load begins to rise. However, when the temperature is between 17°C and 21°C, the increase in the cooling load is slow, and for every 1°C increase in temperature, the cooling load increases by approximately 50,400 kW. When the temperature is between 21°C and 25°C, the growth rate of the cooling load begins to rise, and for every 1°C increase in temperature, the cooling load increases by approximately 162,000 kW. When the temperature is higher than 30°C, the growth rate of the cooling load reaches the highest, and for every 1°C increase in temperature, the cooling load increases by approximately 492,500 kW.

[0038] To sum up, if the base load is regarded as a constant, the relationship between the maximum load and the temperature can be regarded as the curve relationship after the curve of the air-conditioning load and the temperature is translated upward along the vertical axis. The growth of the maximum load is closely related to the average temperature of the day. When the average temperature is less than 17°C, the maximum load climbs all the way with the decrease of the average temperature; when the average temperature is greater than or equal to 17°C, the maximum load shows a stronger growth trend with the increase of the temperature.

[0039] Observing the average temperature data forecasted by the meteorological website for the period from May 16th to 30th (hereinafter referred to as the next 15 days), we found that the temperature in the next 15 days is all greater than 17°C. Therefore, it is reasonable to use the regression equation of the cooling load and the average temperature (y = 2.0592x² - 75.812x + 693.31) to predict the cooling load for the next 15 days. After analysis, the mean absolute percentage error (MAPE) of this regression equation is 16.44%, and the R² of the fitting sequence and the actual sequence is 0.79, which is suitable for carrying out predictions. Then, based on the above predicted cooling load values and the base load values in 2023, the calculation formula for the maximum load is used to achieve the prediction of the maximum power load. The prediction results are shown in Table 2: Table 2 Prediction Results of the Daily Maximum Load in Hefei City for the Next 15 Days ; Evaluation of the Prediction Results of the Maximum Load in Hefei City for the Next 15 Days In the previous prediction, the forecasted temperature data was substituted into the regression equation to obtain the predicted values of the daily maximum load in Hefei City from May 16th to 30th. Now, the predicted values for these 15 days are compared with the actual values to calculate the mean absolute percentage error (hereinafter referred to as the deviation degree) and the accuracy. The specific comparison results are shown in Table 3 below: Table 3 Predicted and Actual Values of the Maximum Load in Hefei City from May 16th to 30th Based on the Forecasted Temperature ; After calculation, the deviation degree between the predicted values of the daily maximum load in Hefei City for the next 15 days predicted based on the forecasted temperature data of the meteorological website and the actual values is 6.27%, and the accuracy is 93.73%. The prediction results are relatively ideal. Combining Figure 4, it can be seen that the predicted maximum load values and the true values generally show a synchronous change relationship. The predicted values are generally larger than the true values. Preliminary analysis indicates that this may be due to the predicted temperature being higher than the actual temperature.

[0040] The Mean Absolute Percentage Error (MAPE) is a measure of relative error and can be used to compare predictions of different scales. Theoretically, the smaller the value of MAPE, the better the fitting effect of the prediction model and the higher the accuracy. Generally, when MAPE is less than 10%, the prediction accuracy is relatively high. Accuracy = 1 - Mean Absolute Percentage Error (MAPE).

[0041] Considering that the daily maximum load predictions for the next 15 days in Hefei are based on the temperature forecast data on the meteorological website for the next 15 days, and the temperature forecast data also belongs to predicted values, the accuracy of which may affect the model prediction effect. Therefore, it is necessary to first evaluate the accuracy of the temperature forecast values on the meteorological website and then explore the impact of the deviation of the forecast temperature on the model prediction effect.

[0042] First, compare and calculate the predicted temperature values from May 16th to 30th with the actual temperature values. The overall deviation degree between the predicted temperature and the actual temperature is 15.27%, and the accuracy is 84.73%. The overall accuracy of the predicted temperature is relatively low.

[0043] Then, substitute the actual temperature data from May 16th to 30th into the prediction model to predict the daily maximum load for the next 15 days, and compare the accuracy changes in the two predictions. After calculation, the deviation degree between the predicted daily maximum load values for the next 15 days in Hefei based on the actual temperature data and the actual values is 4.29%, and the accuracy is 95.71%. The accuracy of the maximum load prediction for the next 15 days based on the actual temperature is 2.11% higher than that based on the predicted temperature.

[0044] The specific comparison results are shown in Table 4 below: Table 4 Evaluation Results of Maximum Load Prediction Accuracy Based on Different Temperature Data ; Combined with Figure 5 , it can be seen that the overall situation of the predicted maximum load values being larger than the true values has also improved, reflecting that the deviation of the predicted temperature data has reduced the accuracy of the maximum load prediction to a certain extent.

[0045] The embodiments of the present disclosure have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, the practical application, or the improvement of technologies in the market, or to enable other ordinary technicians in the art to understand the embodiments disclosed herein.

Claims

1. A power load forecasting method based on the correlation analysis of air temperature and power load, characterized in that The method includes: Obtaining the daily maximum power consumption load data within a period of time, and the temperature data within the period of time; Determining the typical temperature when the power consumption load does not include the air-conditioning load; Determining the daily average value of the daily maximum load corresponding to the typical temperature each year, and obtaining the benchmark load value without the air-conditioning load for each year; Subtracting the benchmark load value of the current year from the daily maximum load to obtain the daily air-conditioning load value; Constructing a quadratic regression equation based on the temperature data and the air-conditioning load value data, where the quadratic regression equation has the temperature as the independent variable x and the predicted air-conditioning load value as the dependent variable y; Obtaining the predicted temperature data from a meteorological website; Substituting the predicted temperature data into the quadratic regression equation to obtain the predicted air-conditioning load value.

2. The power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, wherein The power consumption load includes a benchmark load insensitive to temperature and an air-conditioning load sensitive to temperature, and the air-conditioning load sensitive to temperature includes a heating load or a cooling load.

3. A power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, characterized in that, The calculation of the goodness of fit of the quadratic regression equation is as shown in Equation (1): （1） Among them, R 2 represents the goodness of fit, which is used to indicate the prediction accuracy of the quadratic regression equation. y represents the actual value of the air-conditioning load, represents the predicted value of the air-conditioning load, represents the average value of the actual air-conditioning load.

4. A power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, characterized in that, The unary quadratic regression equation is as follows: (2) Among them, a, b, and c in the equation are coefficients fitted based on the temperature data and the air-conditioning load value data.

5. A power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, characterized in that, Constructing a quadratic regression equation based on the temperature data and the air-conditioning load value data, where the quadratic regression equation has the temperature as the independent variable and the predicted air-conditioning load value as the dependent variable, includes: Constructing a quadratic regression equation based on the minimum temperature data and the heating load value data, where the quadratic regression equation has the minimum temperature as the independent variable and the predicted heating load value as the dependent variable.

6. A power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, characterized in that, Constructing a quadratic regression equation based on the temperature data and the air-conditioning load value data, where the quadratic regression equation has the temperature as the independent variable and the predicted air-conditioning load value as the dependent variable, includes: Constructing a quadratic regression equation based on the maximum temperature data and the heating load value data, where the quadratic regression equation has the maximum temperature as the independent variable and the predicted heating load value as the dependent variable.

7. A power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, characterized in that, Constructing a quadratic regression equation based on the temperature data and the air-conditioning load value data, where the quadratic regression equation has the temperature as the independent variable and the predicted air-conditioning load value as the dependent variable, includes: Constructing a quadratic regression equation based on the average temperature data and the heating load value data, where the quadratic regression equation has the average temperature as the independent variable and the predicted heating load value as the dependent variable.

8. A power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, characterized in that, Constructing a quadratic regression equation based on the temperature data and the air-conditioning load value data, where the quadratic regression equation has the temperature as the independent variable and the predicted air-conditioning load value as the dependent variable, includes: Constructing a quadratic regression equation based on the minimum temperature data and the cooling load value data, where the quadratic regression equation has the minimum temperature as the independent variable and the predicted cooling load value as the dependent variable.

9. A power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, characterized in that, Constructing a quadratic regression equation based on the temperature data and the air-conditioning load value data, where the quadratic regression equation has the temperature as the independent variable and the predicted air-conditioning load value as the dependent variable, includes: Constructing a quadratic regression equation based on the maximum temperature data and the cooling load value data, where the quadratic regression equation has the maximum temperature as the independent variable and the predicted cooling load value as the dependent variable.

10. A power load forecasting method based on the correlation analysis of air temperature and power load according to claim 1, characterized in that Constructing a quadratic regression equation based on the temperature data and the air-conditioning load value data, where the quadratic regression equation has the temperature as the independent variable and the predicted air-conditioning load value as the dependent variable, includes: A unary quadratic regression equation is constructed based on the average temperature data and the cooling load value data. The unary quadratic regression equation uses the average temperature as the independent variable and the predicted cooling load as the dependent variable.

Images