User power load prediction method

Abstract

The invention discloses a user power load prediction method. The method comprises the steps that historical power load data sets of a user in multiple sampling periods are acquired; according to a plurality of historical power load data in the historical power load data group, determining an average value, a minimum value and a maximum value of the plurality of historical power load data, and taking the average value of the plurality of historical power load data as a prediction reference value of a corresponding sampling period; calculating the average value of the predicted reference values, the average value of the predicted drop amplitude, the average value of the predicted rise amplitude and the average value of the fluctuation amplitude of the plurality of sampling periods; and taking the average value of the prediction reference values of the plurality of sampling periods as a power load prediction value at the prediction time point, and calculating a fourth difference value between the average value of the prediction reference values of the plurality of sampling periods and the average value of the prediction drop amplitudes of the plurality of sampling periods as a power load lower limit value at the prediction time point. According to the invention, the precision of user power load prediction can be improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric power monitoring, and in particular to a method for predicting electric power load of a user. Background Art

[0002] The user power load is used to reflect the user's power consumption in a specific time period. The purpose of user power load prediction is to understand and track the changes in the user's power demand by predicting the user's future power consumption. Power supply companies can better design power supply plans and make reasonable arrangements for power dispatching to achieve the purpose of improving power supply efficiency by predicting user power load. For example, when power supplying power, power supply companies supply power to users in a certain area according to a predetermined power supply plan. For example, in summer, the temperature rises and users increase their use of air conditioners. Therefore, it can be inferred that the total power demand of users in the area has increased. Power supply companies can design power dispatching plans to try to meet the power demand of users in the area. However, the above estimation of the total power demand of all users is only based on the total power consumption of all users in the area in previous summers. This estimation has low accuracy and can only provide a rough reference. For a single user, this method uses the total power demand of all users divided by the number of users to roughly estimate the power consumption of a single user. That is, this method lacks a comprehensive analysis of the user habits of a single user, and the user prediction accuracy for a single user is not enough. Therefore, the accuracy of the estimation of the total power demand of all users is even more insufficient, which affects the design of the power supply plan of the power supply company. Summary of the invention

[0003] An object of the present invention is to solve at least the above problems and / or disadvantages and to provide at least the advantages which will be described hereinafter.

[0004] An object of the present invention is to provide a method for predicting user power load, which can improve the prediction accuracy of user power load.

[0005] In order to achieve these purposes and other advantages according to the present invention, a method for predicting user power load is provided, comprising:

[0006] Step 1: Obtain a historical power load data group of the user in multiple sampling periods, wherein each sampling period includes multiple sampling time points, and the historical power load data group in each sampling period includes historical power load data at each sampling time point in each sampling period;

[0007] Step 2: Perform the following processing on the historical power load data group of each sampling period: determine the average value, minimum value and maximum value of the multiple historical power load data according to the multiple historical power load data in the historical power load data group, use the average value of the multiple historical power load data as the prediction reference value of the corresponding sampling period, calculate the first difference between the average value and the minimum value of the multiple historical power load data and the second difference between the maximum value and the average value of the multiple historical power load data, use the first difference as the predicted decrease amplitude of the corresponding sampling period, use the second difference as the predicted increase amplitude of the corresponding sampling period, calculate the third difference between the maximum value and the minimum value of the multiple historical power load data, and use the third difference as the fluctuation amplitude of the corresponding sampling period;

[0008] Step 3: Obtain the predicted reference values, predicted decrease ranges, predicted increase ranges, and fluctuation ranges of the multiple sampling periods, and calculate the average values ​​of the predicted reference values, the average values ​​of the predicted decrease ranges, the average values ​​of the predicted increase ranges, and the average values ​​of the fluctuation ranges of the multiple sampling periods;

[0009] Step 4: Use the average value of the predicted reference values ​​of the multiple sampling cycles as the predicted value of the power load at the prediction time point, calculate the fourth difference between the average value of the predicted reference values ​​of the multiple sampling cycles and the average value of the predicted decrease amplitudes of the multiple sampling cycles as the lower limit value of the power load at the prediction time point, calculate the sum of the average value of the predicted reference values ​​of the multiple sampling cycles and the average value of the predicted increase amplitudes of the multiple sampling cycles as the upper limit value of the power load at the prediction time point, and use the average value of the fluctuation amplitudes of the multiple sampling cycles as the power load fluctuation range at the prediction time point; wherein, the prediction time point is a time occurring after the multiple sampling cycles.

[0010] Preferably, the user power load prediction method further includes step five, in which a power load threshold is set, and the power load prediction value and the power load lower limit value are compared with the power load threshold value respectively; when the power load lower limit value is greater than the power load threshold, a first-level warning processing is performed; when the power load prediction value is greater than the power load threshold and the power load lower limit value is less than the power load threshold, a second-level warning processing is performed, wherein the level of the first-level warning processing is higher than the level of the second-level warning processing.

[0011] Preferably, in the user power load prediction method, in the step five, before comparing the power load prediction value and the power load lower limit value with the power load threshold value respectively, the power load upper limit value is first compared with the power load threshold value; when the power load upper limit value is less than the power load threshold value, no warning processing is performed, and the operation of comparing the power load prediction value and the power load lower limit value with the power load threshold value is no longer performed; when the power load upper limit value is greater than the power load threshold value, the power load prediction value and the power load lower limit value are compared with the power load threshold value respectively; when the power load upper limit value is greater than the power load threshold value and the power load prediction value is less than the power load threshold value, a third-level warning processing is performed, wherein the level of the third-level warning processing is lower than the second-level warning processing.

[0012] Preferably, in the user power load prediction method, in the step five, the power load threshold is obtained in the following manner: in the step one, a historical power load data group of multiple users in multiple sampling periods is obtained; in the step three, prediction reference values ​​of multiple sampling periods of the multiple users are obtained, and the average value of the prediction reference values ​​of multiple sampling periods of the multiple users is calculated; in the step five, the average value of the prediction reference values ​​of multiple sampling periods of the multiple users is set as the power load threshold.

[0013] Preferably, in the user power load forecasting method, the sampling period is one week or one month.

[0014] Preferably, in the user power load prediction method, when the sampling period is one week, the sampling time point is one day, and the prediction time point is any day occurring after the multiple sampling periods.

[0015] Preferably, in the user power load forecasting method, when the sampling period is one month, the sampling time point is one day, and the forecasting time point is any day occurring after the multiple sampling periods; when the sampling period is one month, the sampling time point is one week, and the forecasting time point is any week occurring after the multiple sampling periods.

[0016] The present invention has at least the following beneficial effects:

[0017] The present invention provides a method for predicting user power load, comprising: step one, obtaining a historical power load data group of the user in multiple sampling cycles, wherein each sampling cycle includes multiple sampling time points, and the historical power load data group in each sampling cycle includes the historical power load data at each sampling time point in each sampling cycle; step two, performing the following processing on the historical power load data group of each sampling cycle: determining the average value, minimum value and maximum value of the multiple historical power load data according to the multiple historical power load data in the historical power load data group, taking the average value of the multiple historical power load data as the prediction reference value of the corresponding sampling cycle, calculating a first difference between the average value and the minimum value of the multiple historical power load data and a second difference between the maximum value and the average value of the multiple historical power load data, taking the first difference as the predicted decrease amplitude of the corresponding sampling cycle, taking the second difference as the predicted increase amplitude of the corresponding sampling cycle, calculating the maximum value and minimum value of the multiple historical power load data The third difference between the values, and the third difference is used as the fluctuation amplitude of the corresponding sampling period; step three, obtain the predicted reference value, predicted decrease amplitude, predicted increase amplitude and fluctuation amplitude of the multiple sampling periods, and calculate the average value of the predicted reference value, the average value of the predicted decrease amplitude, the average value of the predicted increase amplitude and the average value of the fluctuation amplitude of the multiple sampling periods; step four, take the average value of the predicted reference value of the multiple sampling periods as the predicted value of the power load at the prediction time point, calculate the fourth difference between the average value of the predicted reference value of the multiple sampling periods and the average value of the predicted decrease amplitude of the multiple sampling periods as the lower limit value of the power load at the prediction time point, calculate the sum of the average value of the predicted reference value of the multiple sampling periods and the average value of the predicted increase amplitude of the multiple sampling periods as the upper limit value of the power load at the prediction time point, and take the average value of the fluctuation amplitude of the multiple sampling periods as the power load fluctuation range at the prediction time point; wherein, the prediction time point is a time occurring after the multiple sampling periods. The present invention obtains the user's historical power load data group in multiple sampling periods for analysis, obtains the user's predicted reference value, predicted increase range, predicted decrease range and fluctuation range in each sampling period, and further obtains the average value of the user's predicted reference value, predicted increase range, predicted decrease range and fluctuation range in multiple sampling periods. According to the above analysis results, the power load prediction value and other power load reference values ​​at the prediction time point are determined, thereby improving the prediction accuracy of the user's power load.

[0018] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The present invention is a flowchart of a method for predicting user power load according to an embodiment of the present invention. DETAILED DESCRIPTION

[0020] The present invention is further described in detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0021] like Figure 1 As shown, an embodiment of the present invention provides a method for predicting user power load, including: step one, obtaining a historical power load data group of the user in multiple sampling periods, wherein each sampling period includes multiple sampling time points, and the historical power load data group in each sampling period includes the historical power load data at each sampling time point in each sampling period; step two, performing the following processing on the historical power load data group of each sampling period: determining the average value, minimum value and maximum value of the multiple historical power load data according to the multiple historical power load data in the historical power load data group, taking the average value of the multiple historical power load data as the prediction reference value of the corresponding sampling period, calculating a first difference between the average value and the minimum value of the multiple historical power load data and a second difference between the maximum value and the average value of the multiple historical power load data, taking the first difference as the predicted decrease amplitude of the corresponding sampling period, taking the second difference as the predicted increase amplitude of the corresponding sampling period, and calculating the maximum value of the multiple historical power load data and the third difference between the minimum value and the minimum value, and use the third difference as the fluctuation amplitude of the corresponding sampling period; step three, obtain the predicted reference value, predicted decrease amplitude, predicted increase amplitude and fluctuation amplitude of the multiple sampling periods, and calculate the average value of the predicted reference value, the average value of the predicted decrease amplitude, the average value of the predicted increase amplitude and the average value of the fluctuation amplitude of the multiple sampling periods; step four, use the average value of the predicted reference value of the multiple sampling periods as the predicted value of the power load at the prediction time point, calculate the fourth difference between the average value of the predicted reference value of the multiple sampling periods and the average value of the predicted decrease amplitude of the multiple sampling periods as the lower limit value of the power load at the prediction time point, calculate the sum of the average value of the predicted reference value of the multiple sampling periods and the average value of the predicted increase amplitude of the multiple sampling periods as the upper limit value of the power load at the prediction time point, and use the average value of the fluctuation amplitude of the multiple sampling periods as the power load fluctuation range at the prediction time point; wherein, the prediction time point is a time occurring after the multiple sampling periods.

[0022] Normally, users' electricity usage habits are relatively fixed. Therefore, the user's historical power load at multiple sampling time points in a sampling cycle is also relatively fixed, and the changes will not be too large. Generally, it fluctuates within a relatively limited range, but large fluctuations at individual sampling time points cannot be ruled out. In addition, users' electricity usage habits tend to be more stable in a longer historical cycle. The historical cycle may contain multiple sampling cycles, and the user's historical power load will also be relatively stable in these sampling cycles. However, it cannot be ruled out that the historical power load of individual sampling cycles will shift.

[0023] Therefore, in order to predict the power load at the prediction time point, the historical power load of the user in a sampling period is first considered, that is, the historical power load data of multiple sampling time points are calculated to obtain the average value as the prediction reference value of the sampling period, and then the first difference between the above average value and the minimum value and the second difference between the maximum value and the above average value are respectively used as the predicted decline and predicted increase, and the third difference between the maximum value and the minimum value is used as the fluctuation range. The above prediction reference value can reflect the average level of power consumption of the user in the sampling period, and this value excludes the interference of individual sampling time points. Although the minimum and maximum values ​​may deviate from the average level of power consumption of the user in the sampling period, they can still reflect the power demand of the user under specific circumstances. Therefore, the minimum and maximum values ​​can still reflect the real power demand of the user. Based on this, the average and minimum values ​​are used to determine the predicted decline, the average and maximum values ​​are used to determine the predicted increase, and the maximum and minimum values ​​are used to determine the fluctuation range. Further, the predicted reference values, predicted declines, and predicted increases of the user in multiple sampling periods are averaged to determine the power consumption of the user in a longer historical period, so as to more accurately reflect the power consumption habits of the user. On this basis, the average value of the predicted reference values ​​of multiple sampling cycles is used as the predicted power load at the predicted time point. At the same time, the fourth difference between the average value of the predicted reference value and the average value of the predicted decline is used as the lower limit of the power load, the sum between the average value of the predicted reference value and the average value of the predicted rise is used as the upper limit of the power load, and the average value of the fluctuation range is used as the fluctuation range of the power load. The predicted power load value reflects the possible power load of the user at the predicted time point, the upper limit of the power load and the lower limit of the power load reflect the upper and lower limits of the power consumption fluctuation of the user at the predicted time point, and the power load fluctuation range reflects the amplitude of the power consumption fluctuation of the user at the predicted time point, so as to more comprehensively and fully reflect the power consumption of the user at the predicted time point.

[0024] According to the power consumption of users at the predicted time point, the predicted power load values ​​of all users in a certain area are added together to obtain the predicted power load value of all users in the area, which reflects the predicted power consumption of all users in the area. The lower limit values ​​of the power load of all users in the area are added together to obtain a value that reflects the lower limit of the predicted power consumption of all users in the area. The upper limit values ​​of the power load of all users in the area are added together to obtain a value that reflects the upper limit of the predicted power consumption of all users in the area. Power supply companies can design power supply plans based on the above predictions.

[0025] To summarize, the present invention obtains the user's historical power load data group in multiple sampling periods for analysis, and obtains the user's predicted baseline value, predicted increase range, predicted decrease range and fluctuation range in each sampling period, and further obtains the average value of the user's predicted baseline value, predicted increase range, predicted decrease range and fluctuation range in multiple sampling periods. According to the above analysis results, the power load prediction value and other power load reference values ​​at the prediction time point are determined, thereby improving the prediction accuracy of the user's power load.

[0026] In a preferred embodiment, the user power load prediction method also includes step five, in which a power load threshold is set, and the power load prediction value and the power load lower limit value are compared with the power load threshold respectively; when the power load lower limit value is greater than the power load threshold, a first-level warning processing is performed; when the power load prediction value is greater than the power load threshold and the power load lower limit value is less than the power load threshold, a second-level warning processing is performed, wherein the level of the first-level warning processing is higher than the level of the second-level warning processing.

[0027] According to the user's electricity consumption at the predicted time point, an early warning can be further provided for the user's electricity consumption. That is, if the lower limit of the power load is already greater than the power load threshold, it means that even if the user's predicted power consumption at the predicted time point deviates from the predicted power load value and reaches the lower limit of the power load, it still exceeds the power load threshold. The user has a greater probability of exceeding the power consumption standard, and a first-level early warning is performed to remind the power supply company to pay attention to the user's electricity consumption. When the lower limit of the power load is less than the power load threshold, but the power load prediction value is greater than the power load threshold, it means that the average level of the user's predicted power consumption exceeds the specified power load threshold, and there is a certain probability that it exceeds the power consumption standard. A second-level early warning can be performed to remind the power supply company to pay attention to the user's power consumption.

[0028] In a preferred embodiment, in the user power load prediction method, in the step five, before comparing the power load prediction value and the power load lower limit value with the power load threshold value respectively, the power load upper limit value is first compared with the power load threshold value; when the power load upper limit value is less than the power load threshold value, no warning processing is performed, and the operation of comparing the power load prediction value and the power load lower limit value with the power load threshold value is no longer performed; when the power load upper limit value is greater than the power load threshold value, the power load prediction value and the power load lower limit value are compared with the power load threshold value respectively; when the power load upper limit value is greater than the power load threshold value and the power load prediction value is less than the power load threshold value, a third-level warning processing is performed, wherein the level of the third-level warning processing is lower than the second-level warning processing.

[0029] Before analyzing the lower limit of the power load and the predicted power load, the upper limit of the power load can be analyzed first. That is, the upper limit of the power load is first compared with the power load threshold. When the upper limit of the power load is less than the power load threshold, it means that the lower limit of the power load and the predicted power load are both less than the power load threshold, and the user's predicted power consumption does not exceed the power consumption standard, and no warning processing is performed. When the upper limit of the power load exceeds the power load threshold, but the predicted power load does not exceed the power load threshold, it means that there is a small probability that the user's predicted power consumption exceeds the power consumption standard, and a third-level warning processing is performed.

[0030] In a preferred embodiment, in the user power load prediction method, in the step five, the power load threshold is obtained in the following manner: in the step one, a historical power load data group of multiple users in multiple sampling periods is obtained; in the step three, prediction reference values ​​of multiple sampling periods of the multiple users are obtained, and the average value of the prediction reference values ​​of multiple sampling periods of the multiple users is calculated; in the step five, the average value of the prediction reference values ​​of multiple sampling periods of the multiple users is set as the power load threshold.

[0031] The power load threshold is determined by the predicted reference values ​​of multiple users in multiple sampling periods. The power load threshold reflects the average power consumption level of all users.

[0032] In a preferred embodiment, in the user power load forecasting method, the sampling period is one week or one month.

[0033] In a preferred embodiment, in the user power load prediction method, when the sampling period is one week, the sampling time point is one day, and the prediction time point is any day that occurs after the multiple sampling periods. When the sampling period is one week, the sampling time points are Monday, Tuesday, Wednesday, Thursday, Friday, Saturday and Sunday, and the prediction time point is a day of a week after the sampling period. Users may have different electricity usage habits on weekdays and weekends. Therefore, the sampling period can be set to a weekday in a week, that is, Monday to Friday, and the sampling time points are Monday, Tuesday, Wednesday, Thursday and Friday. The prediction time point can be any day from Monday to Friday of any following week.

[0034] In a preferred embodiment, in the user power load prediction method, when the sampling period is one month, the sampling time point is one day, and the prediction time point is any day occurring after the multiple sampling periods; when the sampling period is one month, the sampling time point is one week, and the prediction time point is any week occurring after the multiple sampling periods. When the sampling period is one month, the sampling time point is every day of the month, and the prediction time point is a certain day of a certain month after the sampling period.

[0035] To summarize, the present invention obtains the user's historical power load data group in multiple sampling periods for analysis, and obtains the user's predicted baseline value, predicted increase range, predicted decrease range and fluctuation range in each sampling period, and further obtains the average value of the user's predicted baseline value, predicted increase range, predicted decrease range and fluctuation range in multiple sampling periods. According to the above analysis results, the power load prediction value and other power load reference values ​​at the prediction time point are determined, thereby improving the prediction accuracy of the user's power load.

[0036] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and the embodiments. It can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily realized. Therefore, without departing from the general concept defined by the claims and equivalent scope, the present invention is not limited to the specific details and the illustrations shown and described here.

Claims

1. A method for predicting user power load, characterized in that: include: Step 1: Obtain a historical power load data group of the user in multiple sampling periods, wherein each sampling period includes multiple sampling time points, and the historical power load data group in each sampling period includes historical power load data at each sampling time point in each sampling period; Step 2: Perform the following processing on the historical power load data group of each sampling period: determine the average value, minimum value and maximum value of the multiple historical power load data according to the multiple historical power load data in the historical power load data group, use the average value of the multiple historical power load data as the prediction reference value of the corresponding sampling period, calculate the first difference between the average value and the minimum value of the multiple historical power load data and the second difference between the maximum value and the average value of the multiple historical power load data, use the first difference as the predicted decrease amplitude of the corresponding sampling period, use the second difference as the predicted increase amplitude of the corresponding sampling period, calculate the third difference between the maximum value and the minimum value of the multiple historical power load data, and use the third difference as the fluctuation amplitude of the corresponding sampling period; Step 3: Obtain the predicted reference values, predicted decrease ranges, predicted increase ranges, and fluctuation ranges of the multiple sampling periods, and calculate the average values ​​of the predicted reference values, the average values ​​of the predicted decrease ranges, the average values ​​of the predicted increase ranges, and the average values ​​of the fluctuation ranges of the multiple sampling periods; Step 4: Use the average value of the predicted reference values ​​of the multiple sampling cycles as the predicted value of the power load at the prediction time point, calculate the fourth difference between the average value of the predicted reference values ​​of the multiple sampling cycles and the average value of the predicted decrease amplitudes of the multiple sampling cycles as the lower limit value of the power load at the prediction time point, calculate the sum of the average value of the predicted reference values ​​of the multiple sampling cycles and the average value of the predicted increase amplitudes of the multiple sampling cycles as the upper limit value of the power load at the prediction time point, and use the average value of the fluctuation amplitudes of the multiple sampling cycles as the power load fluctuation range at the prediction time point; wherein, the prediction time point is a time occurring after the multiple sampling cycles.

2. The user power load prediction method according to claim 1, characterized in that: The method further includes step five, in which a power load threshold is set, and the power load prediction value and the power load lower limit value are compared with the power load threshold value respectively; when the power load lower limit value is greater than the power load threshold value, a first-level warning process is performed; and when the power load prediction value is greater than the power load threshold value and the power load lower limit value is less than the power load threshold value, a second-level warning process is performed, wherein the level of the first-level warning process is higher than the level of the second-level warning process.

3. The user power load prediction method according to claim 2, characterized in that: In the step five, before comparing the power load prediction value and the power load lower limit value with the power load threshold value respectively, the power load upper limit value is first compared with the power load threshold value; when the power load upper limit value is less than the power load threshold value, no warning processing is performed, and the operation of comparing the power load prediction value and the power load lower limit value with the power load threshold value is no longer performed; when the power load upper limit value is greater than the power load threshold value, the power load prediction value and the power load lower limit value are compared with the power load threshold value respectively; when the power load upper limit value is greater than the power load threshold value and the power load prediction value is less than the power load threshold value, a third-level warning processing is performed, wherein the level of the third-level warning processing is lower than the level of the second-level warning processing.

4. The user power load prediction method according to claim 3, characterized in that: In the step five, the power load threshold is obtained in the following manner: in the step one, a historical power load data group of multiple users in multiple sampling periods is obtained; in the step three, predicted reference values ​​of multiple sampling periods of the multiple users are obtained, and the average value of the predicted reference values ​​of multiple sampling periods of the multiple users is calculated; in the step five, the average value of the predicted reference values ​​of multiple sampling periods of the multiple users is set as the power load threshold.

5. The user power load prediction method according to claim 1, characterized in that: The sampling period is one week or one month.

6. The user power load prediction method according to claim 5, characterized in that: When the sampling period is one week, the sampling time point is one day, and the predicted time point is any day occurring after the plurality of sampling periods.

7. The user power load prediction method according to claim 5, characterized in that: When the sampling period is one month, the sampling time point is one day, and the predicted time point is any day occurring after the multiple sampling periods; when the sampling period is one month, the sampling time point is one week, and the predicted time point is any week occurring after the multiple sampling periods.

Images