Rethinking power factor correction

Fuseco Power Solutions Pty Ltd

By David Gale
Monday, 11 January, 2016


Rethinking power factor correction

Low switching loss, high temperature semiconductors combined with pulse width modulation (PWM) and control algorithms based on Park-Clark transformations, providing a simple orthogonal vector presentation of three-phase vectors as the basis for control, are the backbone of active filters. The same technology is now being applied to displacement power factor correction giving rise to a smaller physical package, extended longevity, minimal maintenance requirements and superior power factor control.

It is surprising, given the limitations of capacitor-based power factor correction, that there has not, to date, been a wider adoption of solid-state technology given that its use in HV and MV VAR compensation is well established. Self-commutated VAR compensators, in particular three-level compensated systems, are increasingly used in MV networks. For HV networks, the switching levels are increased to four and beyond. In flexible AC transmission system (FACTS) controllers, the technology has been well established.

The Sinexel SVG three-level compensator, using PWM inverter control technology, is designed for LV power factor correction and offers some advantages over capacitor-based systems, chiefly: infinitely fine control of compensating reactive current; per phase compensation; leading as well as lagging compensation; rapid dynamic response, eg, regenerative drives (eg, overhauling elevators, winders, etc); line voltage independent compensating reactive current.

Capacitive power factor correction for LV reticulation grew out of its use in MV and HV transmission and distribution where it was often a more economic and simpler alternative to synchronous condensers for the control of voltage. The application to LV is obvious, but what is less obvious are the engineering considerations to provide reliability and safety, speed of response and overcoming the limitations on correction dynamic range as well as fineness of control. As will be seen, providing for all these factors is, practically speaking, an impossibility.

The use of PWM inverter control technology (see Figure 1) to provide lagging or leading correction of load current from -90° to +90° provides a unified method which can only be approximately simulated by a switched capacitor–inductor paralleled system. Although this combination is seen in transmission line compensation, it is not applied in LV reticulation where therefore only lagging power factor can be controlled. It might be argued that unlike the case in transmission lines, leading power factors are not encountered in LV reticulation. However, that is not the case: consider, for example, the use of passive filters in server room UPS systems exhibiting leading power factor during ‘walk-in’.

Figure 1: Self-commutated shunt VAR compensator and simulated leading compensation and lagging compensation.

Speed of response is important and in practice a limitation for capacitive systems. Loads such as motors in load-no load or light load cycle operation oscillate between relatively high power factor and low power factor when essentially only magnetising current is drawn. Generally the contactor switching technology employed in capacitive correction is lagging a significant number of cycles therefore, in effect, controlling on the basis of a ‘quasi-average power factor’.

Under or overcompensation because of slow response is avoided using the inverter-based topology described here. The superior dynamic capability in comparison to capacitor systems is critically important in reticulation systems with large regenerative loads. Four-quadrant operation, for example, overhauling lifts, braking of high inertial loads, etc requires the use of highly responsive correction. In practice, the use of PWM inverter control technology is the only practicable solution.

Historically, the prevalence of stable linear loads made capacitive correction viable. It could be applied directly on fixed loads such as squirrel cage and wound rotor motors, or provided with control steps where necessary. In any event, the leading compensation current supplied by a capacitive system is line voltage dependent. Because of the undesirability of leading power factor, compensation is kept below 1.00. No such limit is required with the solid-state equipment described here.

The prevalence of nonlinear loads in modern electrical installations provides operational hazards for capacitor banks. The rule of thumb is to assume the 5th harmonic as leading culprit for parallel resonance and therefore an inductor in series with the capacitor bank is used to bring the resonant frequency to 4.7 equivalent. Although this measure is largely effective for harmonic patterns such as those from 6-pulse converters, other loads such as burst-fired and phase-fired ones can cause problems including overheating of the capacitor banks.

Traditionally, power factor correction capacitors were over-engineered and physically large, which for transmission and distribution was not critical. In LV reticulation on the other hand, capacitors are required to have a much larger kVAr/volume ratio so that power factor correction equipment can be conveniently housed, usually at the incomer switchboard. Oil-filled capacitors have been replaced by metallised polypropylene (MPP) types. Basically formed by a spirally wound sheet of metallised polypropylene film, these capacitors are compact and have as a benefit a self-healing feature which permits them to function after a momentary short has occurred (the metal film in that area is vaporised).

Further developments of pin holes through the occurrence of shorts alter the value of capacitance and can change resonance conditions. Temperature degradation both in terms of component value and reduced lifetime are serious problems if not taken account of by regular preventive maintenance testing including capacitance measurement. In Figure 2, the lifetime of MPP capacitors is shown as a function of the rating at 70°C. For this reason alone, twice-yearly check-ups are recommended, but this is rarely done in practice even though desirable for the continued effective operation of the equipment.

Mostly contactor switched capacitor banks are employed. Problems occur with frequently exercised capacitor banks, particularly in regard to inrush current which can be several orders of magnitude larger than under steady state conditions causing welding of contacts. Specialised contactors are sometimes employed with early make contacts in series with a current limiting resistor to help ‘form’ the capacitor, followed by the later main contact closure. Some contactors employ the use of small, air core inductors to limit inrush current. This form of technology, however, has the effect of increasing the response time of the power factor correction equipment. The use of zero voltage cross-over thyristor (triac) control diminishes inrush current and improves response time.

In many installations, if not most, capacitive power factor correction equipment is placed into the main switchboard. It is a ready solution but a bad one in that a short circuit in capacitor banks can bring about a massive switchboard fire. High temperature in the switchboard tends to compromise capacitor life. There have been many instances of fire caused by capacitor banks and forensic specialists strongly advise that the power factor correction equipment should be housed separately and at some distance from the main switchboard.

The above problem area which can add significantly to the expense of the installation if remedied by separately housing the power factor correction equipment can be obviated by the Sinexel SVG three-level compensator. The advantages of using opposite (anti) phase PWM created current to sink reactive lagging or reactive leading load current are too significant to relegate the technique as simply an alternative to capacitor-based methods.

Its modest footprint per kVAr and its temperature environmental ruggedness make it suitable for incorporation within switchboards, or in close proximity.

In conclusion, the use of three-level PWM inverter technology power factor compensation provides a more responsive, wide range of control (-1 to +1 power factor range), with immunity from harmonics, far safer operation and a more cost-effective solution on whole-of-life basis.

Figure 2: Practical example of power factor correction in a fast food outlet. The screenshot shows the trend of kVA, kW and kVAr for a fast food outlet over three days following installation of the Sinexel SVG VAR compensator. The red trace indicates kVA, green for kW and blue for kVAr. The compensator was switched in at 11 am on Friday, and red trace from then on is identical to the green one, kVAr having been brought down to an average value of 3.5 kVAr. As can be seen from midnight to 4 am, consumption is at its lowest.

Top image credit: ©iStockphoto.com/isgaby

Related Articles

How can we manage our grid workload efficiently and sustainably?

As national demand for electricity continues to grow, there are still questions surrounding our...

How multifaceted tech can help tame Australia's bushfire threat

Following a warmer than expected winter and high fuel loads in many areas, Australian authorities...

Gearboxes for wind-harvesting kite ships

To support production of wind-powered hydrogen, startup OCEANERGY opted to use WITTENSTEIN...


  • All content Copyright © 2024 Westwick-Farrow Pty Ltd