Capacitor bank selection

When the cheapest solution turns out to be the most expensive

Any technician with minimum electrical knowledge can determine or calculate reactive power compensation. The most common practice is using “a single” electricity bill. The emphasis here is on the “single” electricity bill as it is precisely here that a series of errors can start, which can often end up, with higher costs than those involved when a capacitor bank is correctly determined.
The calculation of the reactive power to be compensated using electricity bills provides us with a relatively correct approximation about which order of magnitude we are dealing with; our starting point. In these cases it is important to ensure that these calculations are carried out with the maximum number of invoices, as they may be heavily influenced by seasonality that we may have ignored (Example: offices or hotels with totally different consumptions in summer than in winter).
As we have mentioned before this must be our starting point, but we must bear in mind other factors which are not reflected in the electricity bill, and are of vital importance for correct compensation:

• Demand fluctuation speed
• System balance
• Harmonic distortion levels

If we focus on the latter, as it is becoming more and more common to find networks with harmonic distortion.

When we carry out inductive reactive power compensation, the incorporation of a parallel capacitor bank is logical to attenuate this demand in order to bring the demanded apparent power (kVA) nearer to the active power (kW) which is really used to carry out the purpose it is designed for. This simple concept can be summarized as a parallel circuit with inductance (L – Transformer and Grid) and capacity (C- Capacitor bank).

If we observe the frequency response of the system we see that for a frequency fR the impedance of the system is much greater than its normal behaviour.

As has been previously stated today’s installations contain loads with demands which are not linear thus provoking greater distortion in harmonic current in the installation, and at the same time in the voltage.

The existence of currents with frequencies higher than the fundamental frequency at 50 or 60 Hz, mean that the resonance conditions previously described are complied with. This would basically cause:

• Amplification of the distortion in voltage for the entire installation (this could affect the equipment and sensitive electrical elements).
• Greater absorption of current by the capacitors, with their consequential overheating, reduction of their capacity and useful life, and in some cases the destruction of the capacitor.

With all these arguments and effects in mind we are going to illustrate a REAL EXAMPLE:

Installation located in Spain, whose activity is set within the metallurgical sector (treatment of metal pieces). This installation comprises a 1 000 kVA transformer, different sub-switchboards with rotary machines (lathes, conveyors belt, elevators, etc.) and services (offices, dispatch warehouse, changing rooms, etc.).

The maintenance technician in charge of this company, having checked that the surcharge level due to reactive energy consumption was significant, calculated, using a single electricity bill, which capacitor bank needed to be installed without taking into consideration any other factors.

He then opted to purchase a conventional capacitor with 150 kvar switchgear.

After connecting the capacitor, a few weeks later, he observed that the capacitor was smoking; the outcome was that two capacitors were now unusable, in addition to the alarms caused in the nearby work centres. The capacitors were replaced after a few weeks, with the same effect being produced a short time later, together with the tripping of some lesser circuit breakers on smaller switchboards such as changing rooms, auxiliary machines and dispatch warehouse. The broken capacitors were replaced again, this time with capacitors strengthened up to 460 V and a short time later the same thing happened again. Finally they opted to disconnect the capacitor bank, meaning a return to paying the reactive energy surcharge.

The maintenance technician from the company asked CIRCUTOR, leading company in reactive energy compensation, to attempt to find out what had happened with this capacitor battery. Basic measurements were then carried out at the head of the installation. These measurements consist simply of measuring with and without the battery connected (always with the installation on full load).

THD(U)% and THD(I)% schematics indicating the capacitor bank connected and disconnected

Although the system denoted relatively low level current distortion (7-8% THD(I)% with 400 A), on the other hand the voltage level did not go unnoticed ( 3.3%  THD(U)% ). Based on empirical experience, the risk of the system entering into resonance is around 15% of the THD(I)% and 2% of THD(U)% (there is nothing stipulated to this effect).

We manually entered each one of the capacitors and we observed how the increase of the THD(U)% was substantial. This is an evident indicator that parallel resonance is being produced. With the capacitor bank connected, values of 80% of the THD(I)% were reached at full load in the factory and 23% THD(U)% (graphic 1). To get an idea, the limit which the supply quality on voltage establishes (UNE EN-50160) is 8%.

Finally we can evaluate the expenses generated by this bad choice:

Here we can see how an apparently cheaper solution turned out to be really more expensive. If a correct technical investment had been carried out with a FR type detuned capacitor bank, the final price would have been reduced by 60%.

 Download this article in PDF format