Hi, For reactive power measuremt,we need to shift one signal by 90 degree.Now my question is can we not do simple digital Integration/differentiation to get the 90deg. phase shift. Why we require processes like HILBERT transform.
HILBERT TRANSFORM
Started by ●July 19, 2011
Reply by ●July 19, 20112011-07-19
>Hi, > >For reactive power measuremt,we need to shift one signal by 90 degree.Now >my question is can we not do simple digital Integration/differentiationto>get the 90deg. phase shift. Why we require processes like HILBERT >transform.It depends whether you care about what happens with the harmonics. Any form of 90 degree shift will be fine for the fundamental. Doing a Hilbert for this job is not easy, though. If you are sampling at a reasonable rate, to catch the key harmonics (I assume this is mains power monitoring) the fundamental is very low in the band, and you probably want it accurate to a small fraction of a percent. That can be a *lot* of terms in your Hilbert transform. Steve
Reply by ●July 19, 20112011-07-19
On Jul 19, 8:55�am, "er_pawan82" <pawan.kumar@n_o_s_p_a_m.landisgyr.com> wrote:> Hi, > > For reactive power measuremt,we need to shift one signal by 90 degree.Now > my question is can we not do simple digital Integration/differentiation to > get the 90deg. phase shift. Why we require processes like HILBERT > transform.If you just do an integral/derivative operation you get the phase shift, but each sinusoidal component's amplitude of your signal will scale with its frequency The reason for using a hilbert transform is to shift the phase of each sinusoidal component without scaling their amplitudes based on frequency. Of course a practical hilbert transformer will have to be some sort of compromise between computational complexity and fidelity to an ideal one. Clay
Reply by ●July 19, 20112011-07-19
On Jul 19, 12:55�pm, Clay <c...@claysturner.com> wrote:> On Jul 19, 8:55�am, "er_pawan82" > > <pawan.kumar@n_o_s_p_a_m.landisgyr.com> wrote: > > Hi, > > > For reactive power measuremt,we need to shift one signal by 90 degree.Now > > my question is can we not do simple digital Integration/differentiation to > > get the 90deg. phase shift. Why we require processes like HILBERT > > transform. > > If you just do an integral/derivative operation you get the phase > shift, but each sinusoidal component's amplitude of your signal will > scale with its frequency �The reason for using a hilbert transform is > to shift the phase of each sinusoidal component without scaling their > amplitudes based on frequency. Of course a practical hilbert > transformer will have to be some sort of compromise between > computational complexity and fidelity to an ideal one.Accurately shifting the voltage harmonics is fruitless unless the current is known with the same precision. You don't actually need a full Hilbert transform on voltage or current. A pair of filters whose outputs are in relative quadrature (with no restriction on the actual phases) is easier to implement. Jerry -- Engineering is the art of making what you want from things you can get.
Reply by ●July 19, 20112011-07-19
Jerry Avins wrote:> On Jul 19, 12:55 pm, Clay <c...@claysturner.com> wrote: > >>>For reactive power measuremt,we need to shift one signal by 90 degree.> You don't actually need a > full Hilbert transform on voltage or current. A pair of filters whose > outputs are in relative quadrature (with no restriction on the actual > phases) is easier to implement.Without doing a Hilbert transform, you either have to synchronize to the fundamental or integrate through many periods to get accurate results. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
Reply by ●July 19, 20112011-07-19
On Jul 19, 3:23�pm, Vladimir Vassilevsky <nos...@nowhere.com> wrote:> Jerry Avins wrote: > > On Jul 19, 12:55 pm, Clay <c...@claysturner.com> wrote: > > >>>For reactive power measuremt,we need to shift one signal by 90 degree. > > You don't actually need a > > full Hilbert transform on voltage or current. A pair of filters whose > > outputs are in relative quadrature (with no restriction on the actual > > phases) is easier to implement. > > Without doing a Hilbert transform, you either have to synchronize to the > fundamental or integrate through many periods to get accurate results.How so? You get an accurate VAR indication in one fundamental period if the amplitudes of the harmonics are retained and each voltage harmonic is in quadrature with its current counterpart. Jerry -- Engineering is the art of making what you want from things you can get.
Reply by ●July 19, 20112011-07-19
On Jul 19, 4:05�pm, Jerry Avins <j...@ieee.org> wrote:> On Jul 19, 3:23�pm, Vladimir Vassilevsky <nos...@nowhere.com> wrote: > > > Jerry Avins wrote: > > > On Jul 19, 12:55 pm, Clay <c...@claysturner.com> wrote: > > > >>>For reactive power measuremt,we need to shift one signal by 90 degree. > > > You don't actually need a > > > full Hilbert transform on voltage or current. A pair of filters whose > > > outputs are in relative quadrature (with no restriction on the actual > > > phases) is easier to implement. > > > Without doing a Hilbert transform, you either have to synchronize to the > > fundamental or integrate through many periods to get accurate results. > > How so? You get an accurate VAR indication in one fundamental period > if the amplitudes of the harmonics are retained and each voltage > harmonic is in quadrature with its current counterpart. > > Jerry > -- > Engineering is the art of making what you want from things you can get.
Reply by ●July 20, 20112011-07-20
On Tue, 19 Jul 2011 09:55:37 -0700 (PDT), Clay <clay@claysturner.com> wrote:>On Jul 19, 8:55�am, "er_pawan82" ><pawan.kumar@n_o_s_p_a_m.landisgyr.com> wrote: >> Hi, >> >> For reactive power measuremt,we need to shift one signal by 90 degree.Now >> my question is can we not do simple digital Integration/differentiation to >> get the 90deg. phase shift. Why we require processes like HILBERT >> transform. > >If you just do an integral/derivative operation you get the phase >shift, but each sinusoidal component's amplitude of your signal will >scale with its frequency The reason for using a hilbert transform is >to shift the phase of each sinusoidal component without scaling their >amplitudes based on frequency. Of course a practical hilbert >transformer will have to be some sort of compromise between >computational complexity and fidelity to an ideal one. > >ClayGood answer Clay! [-Rick-]
