A kind of power equipment warning information trend forecasting method based on exponential smoothing
Technical field
The present invention relates to defect of transformer equipment trend analysis technical field, and in particular to a kind of electricity based on exponential smoothing
Power equipment alarm information trend forecasting method.
Background technology
With the rapid development of economy, the year-on-year rapid growth of society's electricity consumption amount, power network scale Rapid Expansion, electric network composition day
It is beneficial complicated;The power equipment quantity rapid growth such as bus, transformer, disconnecting link, the warning information that power equipment operation is produced also is got over
Come more.
The normal operation for ensureing power equipment is the important component for ensureing power network safety operation, as power network is advised
The development of mould and structure, reliability requirement more and more higher to power equipment, it is necessary in operation of power networks, find in time, anticipation
It has the equipment deficiency that will occur and failure, formulates Response project and eliminate defect among rudiment, so can not only reduce
The number of times of grid collapses, can also reduce processing time when breaking down, so as to not only increase the reliability of power network
And stability, operation of power networks expense is also reduced, high-quality service is provided for power consumer, economic loss is reduced.
Lack the ability that trend prediction is alerted to future device in current electric grid operation, be merely capable of obtaining after alarm occurs
Warning information is taken, so, is often caught unprepared for operation of power networks accident, it is impossible to which precognition makes corresponding place in advance
Put prediction scheme.
The content of the invention
The purpose of the present invention is exactly in order to solve the above-mentioned technical problem, and to provide a kind of electric power based on exponential smoothing and set
Standby warning information trend forecasting method.
Specific method of the present invention is:Each transformer station's history alarm information content is counted, in units of day, statistics is each
Transformer station's warning information quantity, as observation, the initial value that number of days determines exponential smoothing is observed according to historical information, according to
Historical data trend chooses smoothing factor, sets up the forecast model of single exponential smoothing, to following transformer station's warning information number
Amount possible values is made prediction.
The determination method of the initial value is:Consider from the item number of time series, if the observation period n of time series is more than
When 15, initial value is used as using first phase observation;If observation period n is less than 15, the average value of first three observation is taken as initial
Value.
It is described selection smoothing factor specific method be:When time series be in maintenance level trend when, smoothing factor should take compared with
Small value;When time train wave moves larger, α should take median;When time series, which has, significantly rises or falls trend, α should
Take higher value.
The calculation formula of the forecast model of the single exponential smoothing is:
St (1)-- the smooth value of t phases, subscript (1) represents single exponential smoothing;
St-1 (1)-- the smooth value of t-1 phases;
α -- smoothing factor, value is between 0 to 1;
-- the predicted value of t+1 phases.
The present invention has advantages below:1st, monitor's burden is alleviated.By predicting alarm trend, to imminent big
Measure alarm situation to do sth. in advance to know, Response project is formulated in advance.2nd, history alarm data are better profited from.By to history alarm number
According to mining analysis, constantly calculate the prediction of following alarm trend, give full play to the effect of history alarm data.3rd, enrich
The content of application is monitored, the practicality of monitoring application is improved.Lifting means operation trend pre-alerting ability, realize security perimeter from
The leap of " ex-post analysis " to " pre-control in advance ".
Brief description of the drawings
Fig. 1 is transformer station's alarm data reality of the present invention and anticipation trend comparison diagram.
Embodiment
The present invention will be further described below in conjunction with the accompanying drawings.
As shown in figure 1, specific method of the present invention is:Each transformer station's history alarm information content is counted, using day to be single
Position, counts each transformer station's warning information quantity, as observation, observes number of days according to historical information and determines exponential smoothing
Initial value, chooses smoothing factor according to historical data trend, sets up the forecast model of single exponential smoothing, to the following transformer station
Warning information quantity possible values is made prediction.
The determination method of the initial value is:Consider from the item number of time series, if the observation period n of time series is more than
When 15, initial value is used as using first phase observation;If observation period n is less than 15, the average value of first three observation is taken as initial
Value.
It is described selection smoothing factor specific method be:When time series be in maintenance level trend when, smoothing factor should take compared with
Small value;When time train wave moves larger, α should take median;When time series, which has, significantly rises or falls trend, α should
Take higher value.
The calculation formula of the forecast model of the single exponential smoothing is:
St (1)-- the smooth value of t phases, subscript (1) represents single exponential smoothing;
St-1 (1)-- the smooth value of t-1 phases;
α -- smoothing factor, value is between 0 to 1;
-- the predicted value of t+1 phases.
Citing:The existing one group observation data that have counted:Certain transformer station 1 predicts 10 to September alarm amount according to available data
Month alarm amount.
1) determination of initial value:
Because n=9<15, so initial value selects first three monthly average value:
2) smoothing factor is selected:
Here α=0.9 is selected.
3) predicted according to historical data
According to formula:
S1 (1)=0.9 × 4000+ (1-0.9) × 4566.67=4506.67
The rest may be inferred draws:
So prediction alarm in October amount is 5984.26.
Embodiment of above is merely to illustrate the present invention, and not limitation of the present invention, about the common of technical field
Technical staff, without departing from the spirit and scope of the present invention, can also make a variety of changes and modification, therefore all
Equivalent technical scheme falls within scope of the invention, and scope of patent protection of the invention should be defined by the claims.