Method and device to predict a state of a power system in the time domain

1. A method to predict a state of a power system with a computing device, comprising the steps of using the computing device to perform a Fast Fourier Transform on measured samples of a periodic waveform signal in order for the computing device to determine a frequency spectrum of the periodic waveform signal, where the computing device derives the periodic waveform signal from at least one measurable electric quantity of the power system and is representative of the state of the power system, the computing device performing the following steps:

determining frequencies and related complex amplitudes at selected peaks in the frequency spectrum and storing the frequencies and amplitudes together with a time stamp in a storage device connected to the computing device,

determining and storing one or more predicted values of the periodic waveform signal (φ) in the time domain at one or more prediction times by calculating a sum of sinusoidal signals at the corresponding prediction times, each of the sinusoidal signals based on one of the determined frequencies and its related complex amplitude and the time stamp,

comparing each of the one or more predicted values with a measured value derived from a measurement of the at least one electric quantity of the power system, where the one or more prediction times each equal to a point in time when the corresponding measurement was taken, and

generating a notification based on a change in the state of the power system if a select number of differences between the one or more predicted values and the corresponding measured values each exceed a predetermined limit.

2. The method according to claim 1 , where the predicted value is determined based on the equation

ϕ

p

=

Re

[

∑

i

⁢

(

A

i

·

ⅇ

2

⁢

π

⁢

⁢

f

i

⁡

(

t

p

-

t

0

)

)

]

where φ p is the predicted value, i is an index to number consecutively the determined frequencies f i , and their corresponding complex amplitudes A i , t 0 is the time stamp when the frequencies f i , and amplitudes A i , were determined and t p is the prediction time.

3. The method according to claim 1 , where the frequencies and related complex amplitudes are determined from peaks in the frequency spectrum whose amplitude exceeds a select amplitude limit.

4. The method according to claim 1 , where the frequencies and related complex amplitudes are determined after the frequency spectrum was filtered with a frequency filter.

5. The method according to claim 1 , where said computing device comprises a real-time operating system and where the step of determining the predicted value is performed simultaneously with the steps of performing a Fast Fourier Transform and determining the frequencies and related complex amplitudes.

6. The method according to claim 1 , further comprising a network connection and wherein the computing device is connected to the network connection such that the measured values are compared to the corresponding of the multiple predicted values substantially simultaneous with when the measurements are performed.

7. The method according to claim 1 , where the state is the absence of short-circuits in the power system.

8. The method according to claim 1 , where the power system is a power grid and where the state is the availability of a power transmission line in the grid.

9. The method according to claim 1 , where the power system comprises a power transform with a tap changer and where the state is the quality of operation of the tap changer.

10. The method according to claim 2 , where the frequencies and related complex amplitudes are determined from peaks in the frequency spectrum whose amplitude exceeds a select amplitude limit.