|
Before I continue with harmonic filters per se, I would like to concentrate on the whole aspect of harmonic filtering - well, it is this word "filtering" that can lead to confusion as to what is actually being done. I prefer to think of it as 'harmonic control' as opposed to 'harmonic filtering', and I use this from now on. Resonant (LC) arrangements are designed primarily for single frequency (harmonic) control. At best, it is possible to do two frequencies, but it involves complex C arrangements on a very specially tapped L - not something I enjoy (even at RF it is tricky!). The other side of the coin is one has to be careful to also not have an LC arrangement promote an unwanted frequency while controlling the wanted harmonic (parasitic resonance is not just confined to higher frequencies). I prefer to think of "Total harmonic control" i.e. active harmonic filters, as "distortion management and mitigation". Harmonics are only one aspect of distortion, and has caused many an engineer to walk up the garden path and have the door slammed in his (or her, lest someone think I am sexist) face.
"If there is a load that can A simple circuit would regulate a load such that the current presented to the source resembles that of a sinewave i.e. while the non-linear load is not drawing current i.e. ahead and after the peak, the active control draws current on the load's behalf (in the correct proportion) such that the current draw on the source appears 'linear'. However, this would result in a lot of heat loss for no true benefit. With "distortion management and mitigation", the oldest scientific principle comes into play. "For every action there is an equal and opposite reaction". A derivative of this is "an action can be nulled by an equal and opposite action", and this is exactly how 'active harmonic control' operates.
Cast your mind back to the principles of power factor correction. The initial problem is the inductive load draws current at the wrong time on the cycle (if purely inductive, the highest current occurs at the zero crossing after the peak voltage!). To correct this, a capacitor is placed across the load which imports energy while the inductor is exporting, and exports energy while the inductor imports. The result; An apparently linear load. With active control, the 'filter' does exactly the same thing. It's actually quite simple; It uses the time the non-linear load is not drawing current to charge a capacitor. The moment the load starts drawing current, the capacitor is used to augment the supply. The result is the load presented to the supply begins to resemble a linear load - the process being done with high-speed switching techniques. I want to draw attention to one small point made in the above paragraph; I said the active harmonic filter augments the supply when the non-linear load demands current - and it is this, together with the principle that an action can be nulled by an equal and opposite action, that makes the active harmonic filter a harmonic generator. Normal non-linear (hi-tech) loads demand energy not just at the fundamental, but at a number of harmonics too. The active harmonic filter, by generating the harmonic energy in the correct proportions, relieves the source of this burden leaving it to supply the needed energy at one frequency only - the fundamental. The great thing about active filters is they are based on phase locked loops and therefore track the incoming frequency (the failing of many an LC filter). They also adjust the conversion (harmonic generation) to the ever changing load so the supply always thinks the load is nothing but nearly resistive (well, so says the blurb! - see closing paragraph). This apparently perfect solution does come with a few warnings though. They are built with complicated technology and can, therefore, break down. And if not looked at, can stay this way for many years! They do need a little energy to run, and the conversion process is about 97% efficient. The operating energy varies between makes, but is usually about 200-500W, and the conversion process will usually dissipate about 3% of the total harmonic energy being combated e.g. if the harmonic content is, say, 100A @ 230V, the device will emit approx 700W heat over and above its running energy. I have heard of active filters injecting high doses of noise (from all the switching) onto the mains supply. I must say though, I have only heard of this and not seen it myself. Manufacturers of active filters are a tad reluctant to actually let me near their kit (in the way I want to get near) as they know I will unearth some little secret they would prefer the unsuspecting clients not know.
© 07.05.05 / 06.07.06 |
SOLUTIONS: