Specific embodiment
To make purpose, technical scheme and the advantage of the embodiment of the present invention clearer, below in conjunction with the embodiment of the present invention
In accompanying drawing, the technical scheme in the embodiment of the present invention is clearly and completely illustrated, it is clear that described embodiment is
The a part of embodiment of the present invention, rather than the embodiment of whole.Based on the embodiment in the present invention, those of ordinary skill in the art
The every other embodiment obtained under the premise of creative work is not made, belongs to the scope of protection of the invention.
Accompanying drawing is combined separately below, and a kind of load Series Modeling method provided in an embodiment of the present invention is illustrated.
Fig. 1 is a kind of load Series Modeling method implementing procedure figure in the embodiment of the present invention, as illustrated, in the present embodiment
Load sequence is built as steps described below can, specially:
Step S101:The historical load data in default time range is obtained according to default temporal resolution.Its
In, load data refers to the power load charge values of electrical network.
Step S102:Historical load data is classified according to default load type, and calculate each load type
Load curve.
Step S103:Load scenarios are generated at random using Monte Carlo method, and judges whether the load scenarios meet institute
State the requirement of each load curve and its peak-valley difference probability distribution:Retain the load scenarios if meeting, this is deleted if being unsatisfactory for
Load scenarios.Wherein, load scenarios refer to the time serieses of electric load.
In the present embodiment, load scenarios refer to the load sequence i.e. load-time curve that load data and time are constituted,
Generate according to all types of load curves after wherein classifying to historical load data according to default load type various negative
Lotus scene, can simulate different load variations situations, it is ensured that the result of simulation meets the actual load of following a period of time and becomes
Change, so as to improving the robustness of new forms of energy production simulation result of calculation and reducing simulating the deviation for calculating.
Further, in the present embodiment, step S101 can be implemented as steps described below.
1st, constant duration division is carried out to historical load data.
In the present embodiment, time interval can be 24 hours, then historical load data is carried out obtaining after constant duration division
To within each time period, historical load data is daily load.
2nd, respectively the historical load data of each time period in default time range is normalized.
The maximum of the historical load data within each time period is obtained in the present embodiment, with the maximum as base value to it is each when
Between historical load data in section be normalized calculating.
3rd, recognize and repair the historical load data of each time period.
(1) historical load data in each time period is counted, recognizes whether there are leakage data;
(2) autocorrelation performance of historical load number in each time period is analyzed, judges whether not meeting autocorrelation performance
Bad data.
(3) bad data in each time period is rejected using ARMA model, and/or recovers leakage data.
Further, historical load data is classified according to default load type in step S102 in the present embodiment
Can implement as steps described below.
(1) in each time period after statistic procedure S101 is repaired, the peak-valley difference of historical load data, rate of load condensate and maximum are negative
Lotus probability distribution.Wherein,
Peak-valley difference is the difference of the maxima and minima of historical load data in each time period.
Rate of load condensate is the ratio of the meansigma methodss of historical load data and maximum in each time period.
Peak load probability distribution is the maximum of historical load data in default time range within each time period
Probability distribution.Wherein, probability of the maximum of historical load data within each time period can occur in each time period by which
Number of times in interior any time period is obtained divided by the sum of time period in default time range.
(2) the SOM models of non-supervisory Self-organizing Maps are built according to peak-valley difference, rate of load condensate and peak load probability distribution,
Historical load data is classified according to the SOM models and according to default load type.
Load type includes daily load and vacation load:Daily load includes that Monday daily load, Tuesday daily load, day Wednesday are born
Lotus, Thursday daily load, Friday daily load, Saturday daily load and Sunday daily load, vacation load be located default vacation when
Between in the range of load.
In the present embodiment, vacation load can include vacation on May Day load, 11 vacation loads and Spring Festival holiday load.Its
In, vacation on May Day load refer to " May Day " official holiday be located time range in load, 11 is false
Phase load refers to the load in the time range that is located the official holiday of " commemoration day on National Day ", and Spring Festival holiday load refers to
The load in time range that the official holiday in " Spring Festival " is located.
Further, the load curve for calculating each load type in the present embodiment in step S102 can be as steps described below
Implement.
(1) all historical load datas corresponding with each load type are obtained.
Load type Monday daily load, Tuesday daily load, Wednesday daily load, Thursday daily load, day Friday in the present embodiment
Load, Saturday daily load, Sunday daily load, vacation on May Day load, 11 vacation loads and Spring Festival holiday load, then obtain with respectively
The corresponding all historical load datas of load type include:
Obtain all in Monday, Tuesday, Wednesday, Thursday, Friday, Saturday, Sunday in default time range respectively
Historical load data, and obtain the official holiday of " May Day ", " state in the default time range respectively
The load in time range that the official holiday of the official holiday and " Spring Festival " of celebrating commemoration day " is located.
(2) historical load data corresponding with each acquisition time in each load type is calculated according to all historical load datas
Meansigma methodss.In the present embodiment temporal resolution can be 15min or 1h, i.e., historical load data in default time range
Each acquisition time at intervals of 15min or 1h.
Below to calculating the corresponding historical load of its each acquisition time by taking first acquisition time of Monday daily load as an example
The meansigma methodss of data are illustrated:The historical load data of all of Monday daily load is first depending on, its first collection is counted
Time corresponding all data, the meansigma methodss for then calculating these data obtain first acquisition time pair of Monday daily load
The meansigma methodss of the historical load data answered.
(3) load curve of each load type is built according to the meansigma methodss of historical load data.
(4) analyze the autocorrelation performance of each load curve:If the autocorrelation coefficient of load curve is unsatisfactory for statistical law,
Load curve is modified makes which meet statistical law.
Further, can also comprise the steps in step S103 in the present embodiment.
1st, compare the autocorrelation performance of the historical load data in load scenarios and default time range, judge load field
Whether the peak-valley difference of scape is consistent with the peak-valley difference of historical load data:If inconsistent, foundation regenerates load scenarios, if one
Cause then carries out proportional zoom to load scenarios and obtains final load scenarios.
2nd, the minima of historical load data in load scenarios is obtained, the history at each acquisition time in load scenarios is born
Lotus data are subtracted each other with the minima, obtain load fluctuation time serieses.
3rd, with default peak-valley difference as ratio value, scaled down is carried out to each data in load fluctuation time serieses or is put
Greatly.
4th, in the load fluctuation time serieses after zooming in or out, each data are added with the minima of historical load data,
Obtain final load scenarios.
Present invention also offers a kind of load Series Modeling system, and provide specific embodiment.
In the present embodiment, load Series Modeling system includes that historical load data statistical module, historical load data calculate mould
Block and load scenarios generation module.Wherein,
Historical load data statistical module, for obtaining in default time range according to default temporal resolution
Historical load data.
Historical load data computing module, for classifying to historical load data according to default load type, and
Calculate the load curve of each load type.
Load scenarios generation module, for generating load field at random according to each load curve and its peak-valley difference probability distribution
Scape.
In the present embodiment, load Series Modeling system includes that historical load data statistical module, historical load data calculate mould
Block and load scenarios generation module, it is bent that its historical load data computing module calculates load corresponding with each default load type
Line, load scenarios generation module can generate various load scenarios at random according to the load curve, can simulate different loads
Situation of change, it is ensured that the result of simulation meets the actual load change of following a period of time, so as to improve new forms of energy production simulation
The robustness of result of calculation simultaneously reduces simulating the deviation for calculating.
Further, in the present embodiment, historical load data statistical module can also include following structures.
In the present embodiment, historical load data statistical module includes that data dividing unit, normalization computing unit, data are distinguished
Know unit and data repair unit.Wherein,
Data dividing unit, for carrying out constant duration division to the historical load data in default time range.
Normalization computing unit, for obtaining the maximum of the historical load data within each time period in default time range
Value, is normalized calculating to the historical load data in each time period by base value of maximum.
Data identification unit, for counting the historical load data in each time period, recognizes whether there are leakage data;Analysis
The autocorrelation performance of historical load number in each time period, judges whether not meeting the bad data of the autocorrelation performance.
Data repair unit, for rejecting the bad data in each time period using ARMA model, and/
Or recover the leakage data.
Further, in the present embodiment, historical load data computing module can also include following structures.
In the present embodiment, historical load data computing module includes data sorting unit and load curve computing unit.Its
In,
Data sorting unit, repairs each time period unit repair after in historical load data through data for foundation
Peak-valley difference, rate of load condensate and peak load probability distribution build the SOM models of non-supervisory Self-organizing Maps, and according to the SOM models
The historical load data is classified.
Load curve computing unit, for calculating the load curve of each load type.
Further, in the present embodiment, load curve computing unit can also include following structures.
In the present embodiment, load curve computing unit includes that data acquisition subelement, data computation subunit, curve build
Subelement and tracing analysiss subelement.Wherein,
Data acquisition subelement, for obtaining all historical load datas corresponding with each load type.
Data computation subunit, for calculating and each acquisition time pair in each load type according to all historical load datas
The meansigma methodss of the historical load data answered.
Curve builds subelement, builds the load curve of each load type for the meansigma methodss according to historical load data.
Tracing analysiss subelement, for analyzing the autocorrelation performance of each load curve:If the autocorrelation coefficient of load curve
Statistical law is unsatisfactory for, then load curve is modified makes which meet statistical law.
Further, in the present embodiment, modeling can also include following structures.
In the present embodiment, modeling also includes load scenarios amending unit.Wherein,
Load scenarios amending unit includes load scenarios comparing subunit and load scenarios scaling subelement, load scenarios contracting
Putting subelement includes the first computation subunit, the second computation subunit and the 3rd computation subunit.
Load scenarios comparing subunit, for comparing the historical load data in load scenarios and default time range
Autocorrelation performance, and judge whether the peak-valley difference of load scenarios is consistent with the peak-valley difference of historical load data:Control if inconsistent
Load scenarios generation module processed regenerates load scenarios;Scape scaling subelement is created by load if consistent to enter load scenarios
Row proportional zoom obtains final load scenarios.
First computation subunit, for obtaining the minima of historical load data in load scenarios, will be each in load scenarios
Historical load data at acquisition time is subtracted each other with the minima, obtains load fluctuation time serieses.
Second computation subunit, for default peak-valley difference as ratio value, to each data in load fluctuation time serieses
Carry out scaled down or amplification.
3rd computation subunit, for each data and historical load in the load fluctuation time serieses after zooming in or out
The minima of data is added, and obtains final load scenarios.
Obviously, those skilled in the art can carry out the essence of various changes and modification without deviating from the present invention to the present invention
God and scope.So, if these modifications of the present invention and modification belong to the scope of the claims in the present invention and its equivalent technologies
Within, then the present invention is also intended to comprising these changes and modification.