Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The following describes a load sequence modeling method provided by an embodiment of the present invention with reference to the accompanying drawings.
Fig. 1 is a flowchart of an implementation of a load sequence modeling method in an embodiment of the present invention, and as shown in the figure, a load sequence may be constructed according to the following steps:
step S101: and acquiring historical load data in a preset time range according to a preset time resolution. Wherein the load data refers to an electrical load value of the electrical grid.
Step S102: and classifying the historical load data according to a preset load type, and calculating a load curve of each load type.
Step S103: randomly generating a load scene by adopting a Monte Carlo method, and judging whether the load scene meets the requirements of each load curve and the peak-valley difference probability distribution thereof: if the load scene is satisfied, the load scene is reserved, and if the load scene is not satisfied, the load scene is deleted. Wherein a load scenario refers to a time series of electrical loads.
In the embodiment, the load scene refers to a load sequence, namely a load-time curve, formed by load data and time, wherein various load scenes are generated according to various types of load curves after historical load data are classified according to preset load types, different load change conditions can be simulated, and the simulated result is guaranteed to be in accordance with the actual load change of a period of time in the future, so that the robustness of the new energy production simulation calculation result is improved, and the deviation of simulation calculation is reduced.
Further, step S101 in this embodiment may be implemented as follows.
1. And dividing the historical load data at equal time intervals.
In this embodiment, the time interval may be 24 hours, and the historical load data in each time period obtained by dividing the historical load data at equal time intervals is the daily load.
2. And respectively carrying out normalization processing on the historical load data of each time period in a preset time range.
In this embodiment, the maximum value of the historical load data in each time period is obtained, and the historical load data in each time period is normalized by using the maximum value as a base value.
3. And identifying and repairing historical load data of each time period.
(1) Counting historical load data in each time period, and identifying whether missing data exists or not;
(2) and analyzing the autocorrelation characteristic of the historical load number in each time period, and judging whether bad data which does not accord with the autocorrelation characteristic exists.
(3) And (3) eliminating bad data in each time period by adopting an autoregressive moving average model and/or recovering leaked data.
Further, in the present embodiment, the step S102 of classifying the historical load data according to the preset load type may be implemented according to the following steps.
(1) And (4) counting the peak-valley difference, the load rate and the maximum load probability distribution of the historical load data in each time period after the restoration in the step S101. Wherein the content of the first and second substances,
the peak-to-valley difference is the difference between the maximum value and the minimum value of the historical load data in each time period.
The load rate is the ratio of the average value to the maximum value of the historical load data in each time period.
The maximum load probability distribution is the probability distribution of the maximum value of the historical load data in each time period in a preset time range. The probability of the maximum value of the historical load data in each time period can be obtained by dividing the number of times of the maximum value of the historical load data in any time period in each time period by the total number of the time periods in a preset time range.
(2) And constructing an SOM (self-organizing map) model of unsupervised self-organizing mapping according to the peak-valley difference, the load rate and the maximum load probability distribution, and classifying the historical load data according to the SOM model and the preset load type.
The load types include daily load and holiday load: the daily load includes a sunday load, a tuesday load, a wednesday load, a thursday load, a friday load, a saturday load, and a sunday load, and the holiday load is a load within a time range in which a preset holiday is present.
The holiday loads in this embodiment can include a quintessenary holiday load, an eleven holiday load, and a spring festival holiday load. Wherein, the Wuyi holiday load refers to the load in the time range of the legal holiday of the Wuyi International labor festival, the eleven holiday load refers to the load in the time range of the legal holiday of the national celebration anniversary, and the spring holiday load refers to the load in the time range of the legal holiday of the spring festival.
Further, in the present embodiment, the calculation of the load curve of each load type in step S102 may be performed according to the following steps.
(1) All historical load data corresponding to each load type is obtained.
In this embodiment, if the load types include a day-of-week load, a day-of-five holiday load, an eleven holiday load, and a spring holiday load, then obtaining all historical load data corresponding to each load type includes:
the method comprises the steps of respectively obtaining historical load data of Monday, Tuesday, Wednesday, Thursday, Friday, Saturday and Sunday in a preset time range, and respectively obtaining loads in time ranges of legal holidays of ' five international labor festivals ', national commemorative holidays ' and ' spring festival ' legal holidays in the preset time range.
(2) And calculating the average value of the historical load data corresponding to each acquisition time in each load type according to all the historical load data. In this embodiment, the time resolution may be 15min or 1h, that is, the interval between the acquisition times of the historical load data in the preset time range is 15min or 1 h.
The following description will be made of calculating the average value of the historical load data corresponding to each collection time of the monday load, taking the first collection time as an example: firstly, according to the historical load data of all the loads in the week and day, counting all the data corresponding to the first collection time of the loads in the week and day, and then calculating the average value of the data to obtain the average value of the historical load data corresponding to the first collection time of the loads in the week and day.
(3) And constructing a load curve of each load type according to the average value of the historical load data.
(4) The autocorrelation characteristics of each load curve were analyzed: and if the autocorrelation coefficient of the load curve does not meet the statistical rule, correcting the load curve to enable the load curve to meet the statistical rule.
Further, step S103 in this embodiment may further include the following steps.
1. Comparing the self-correlation characteristics of the load scene and the historical load data in a preset time range, and judging whether the peak-valley difference of the load scene is consistent with the peak-valley difference of the historical load data: if not, the load scene is regenerated again, and if the load scene is consistent, the load scene is scaled to obtain a final load scene.
2. And obtaining the minimum value of the historical load data in the load scene, and subtracting the historical load data at each acquisition time in the load scene from the minimum value to obtain a load fluctuation time sequence.
3. And (4) carrying out equal-scale reduction or amplification on each data in the load fluctuation time sequence by taking the preset peak-valley difference as a proportional value.
4. And adding the minimum values of the data and the historical load data in the reduced or amplified load fluctuation time sequence to obtain a final load scene.
The invention also provides a load sequence modeling system and provides a specific embodiment.
The load sequence modeling system in the embodiment comprises a historical load data statistics module, a historical load data calculation module and a load scene generation module. Wherein the content of the first and second substances,
and the historical load data counting module is used for acquiring the historical load data within a preset time range according to a preset time resolution.
And the historical load data calculation module is used for classifying the historical load data according to the preset load types and calculating the load curves of the load types.
And the load scene generation module is used for randomly generating load scenes according to each load curve and the peak-valley difference probability distribution thereof.
The load sequence modeling system in the embodiment comprises a historical load data counting module, a historical load data calculating module and a load scene generating module, wherein the historical load data calculating module calculates load curves corresponding to preset load types, and the load scene generating module can randomly generate various load scenes according to the load curves, can simulate different load change conditions, and ensures that a simulation result conforms to the actual load change in a period of time in the future, so that the robustness of a new energy production simulation calculation result is improved, and the deviation of simulation calculation is reduced.
Further, the historical load data statistics module in this embodiment may further include the following structure.
The historical load data statistics module in this embodiment includes a data dividing unit, a normalization calculating unit, a data identifying unit, and a data repairing unit. Wherein the content of the first and second substances,
and the data dividing unit is used for dividing the historical load data in a preset time range at equal time intervals.
And the normalization calculation unit is used for acquiring the maximum value of the historical load data in each time period in a preset time range, and performing normalization calculation on the historical load data in each time period by taking the maximum value as a base value.
The data identification unit is used for counting historical load data in each time period and identifying whether missing data exists or not; and analyzing the autocorrelation characteristic of the historical load number in each time period, and judging whether bad data which do not accord with the autocorrelation characteristic exist.
And the data restoration unit is used for eliminating the bad data in each time period by adopting an autoregressive moving average model and/or restoring the leakage data.
Further, the historical load data calculation module in this embodiment may further include the following structure.
The historical load data calculation module in the embodiment comprises a data classification unit and a load curve calculation unit. Wherein the content of the first and second substances,
and the data classification unit is used for constructing an SOM (self-organizing map) model of unsupervised self-organizing mapping according to the peak-valley difference, the load rate and the maximum load probability distribution of the historical load data in each time period after the data is repaired by the data repair unit, and classifying the historical load data according to the SOM model.
And the load curve calculation unit is used for calculating the load curve of each load type.
Further, the load curve calculating unit in this embodiment may further include the following structure.
The load curve calculation unit in the embodiment comprises a data acquisition subunit, a data calculation subunit, a curve construction subunit and a curve analysis subunit. Wherein the content of the first and second substances,
and the data acquisition subunit is used for acquiring all historical load data corresponding to each load type.
And the data calculation subunit is used for calculating the average value of the historical load data corresponding to each acquisition time in each load type according to all the historical load data.
And the curve construction subunit is used for constructing the load curve of each load type according to the average value of the historical load data.
A curve analysis subunit, configured to analyze an autocorrelation characteristic of each load curve: and if the autocorrelation coefficient of the load curve does not meet the statistical rule, correcting the load curve to enable the load curve to meet the statistical rule.
Further, the modeling system in this embodiment may further include the following structure.
The modeling system in this embodiment further includes a load scenario correction unit. Wherein the content of the first and second substances,
the load scene correction unit comprises a load scene comparison subunit and a load scene scaling subunit, wherein the load scene scaling subunit comprises a first calculation subunit, a second calculation subunit and a third calculation subunit.
The load scene comparison subunit is used for comparing the load scene with the autocorrelation characteristics of the historical load data within a preset time range, and judging whether the peak-to-valley difference of the load scene is consistent with the peak-to-valley difference of the historical load data: if not, controlling the load scene generation module to regenerate the load scene; and if the load scene is consistent with the preset load scene, scaling the load scene through the load scene creation scaling subunit to obtain a final load scene.
And the first calculating subunit is used for acquiring the minimum value of the historical load data in the load scene, and subtracting the minimum value from the historical load data at each acquisition time in the load scene to obtain a load fluctuation time sequence.
And the second calculating subunit is used for carrying out equal-scale reduction or amplification on each data in the load fluctuation time sequence by taking the preset peak-valley difference as a proportional value.
And the third calculating subunit is used for adding the minimum value of each data in the reduced or amplified load fluctuation time sequence and the historical load data to obtain a final load scene.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.