Power Factor Correction: How APFC Panels Cut an Industrial Electricity Bill
An APFC panel cuts an industrial electricity bill through three separate mechanisms, and only one of them is the power factor penalty most buyers focus on. It removes the low power factor surcharge, it reduces billed kVA demand, and it lowers the current in every cable and transformer between the incomer and the load, which cuts resistive losses. On a 300 kW plant running at 0.85 power factor in Karnataka, the surcharge alone is worth about Rs 3.6 lakh a year before either of the other two effects is counted.
TL;DR
-
Power factor is the ratio of working power to apparent power. Correcting it does not reduce the kW your machines need, it reduces the current drawn to deliver those kW.
-
The Karnataka tariff specifies 0.90 for HT consumers, with a surcharge of 3 paise per unit for every 0.01 below it. A plant at 0.85 is paying 15 paise on every unit consumed.
-
Three savings stack: surcharge removal, lower billed kVA demand, and reduced I squared R losses. Correcting 0.85 to 0.98 cuts current by about 13 percent and copper losses by about 25 percent.
-
CEA connectivity regulations require distribution systems and bulk consumers to hold power factor at not less than 0.95 and not to depend on the grid for reactive power support.
-
Sizing is a two-line calculation: kVAR equals kW multiplied by the difference in tan of the two angles. Taking 300 kW from 0.85 to 0.98 needs 125 kVAR.
-
Never connect capacitors to the motor side of a VFD or a soft starter, and never leave a bank fixed on a plant whose load varies. Over-correction is a real cost, not a theoretical one.
-
Capacitors bought in India must carry the ISI mark under IS 13340 (Part 1): 2012, and capacitor switching stresses the switchgear that does the switching.
What does an APFC panel actually do to your bill?
An APFC panel measures plant power factor continuously and switches capacitor stages in and out to hold it at a set target. The capacitors supply the reactive current that inductive loads need, locally, so that current no longer travels from the grid through your transformer and cables. Your machines draw the same kW. The supply sees less kVA.
Three distinct savings follow, and they are frequently confused. The surcharge disappears, which is a tariff penalty rather than a physics effect and is usually the largest single line. Billed demand falls, because where the tariff bills demand in kVA, reducing apparent power reduces the demand charge every month whether or not a penalty clause exists. And losses fall, because lower current through the same cables and transformer means lower I squared R heating. Only the third is a genuine energy saving rather than a billing change, and it is the smallest.
New India Electricals Ltd builds automatic power factor correction panels as part of its low voltage panel range, alongside power and motor control centres, distribution boards, generator AMF and synchronising panels, and bus duct systems.
How much is a low power factor actually costing you?
Work it from the tariff, not from a rule of thumb. The Karnataka Electricity Regulatory Commission Tariff Order 2025 sets the specified power factor for HT consumers at 0.90, and states that it is the responsibility of the HT consumer to determine the capacity of power factor correction apparatus and maintain an average power factor of not less than 0.90. Below that, a surcharge of 3 paise per unit consumed is levied for every reduction of 0.01.
For LT consumers the specified figure is 0.85, the surcharge is 2 paise per unit for every 0.01 below, capped at 30 paise per unit. There is a separate 30 paise per unit surcharge where an inspection finds installed capacitor capacity below what the Conditions of Supply require, on LT categories involving motive power.
The arithmetic on a 300 kW HT plant consuming 200,000 units a month:
| Power factor | Steps below 0.90 | Surcharge per unit | Monthly surcharge | Annual surcharge |
|---|---|---|---|---|
| 0.88 | 2 | 6 paise | Rs 12,000 | Rs 1.44 lakh |
| 0.85 | 5 | 15 paise | Rs 30,000 | Rs 3.60 lakh |
| 0.80 | 10 | 30 paise | Rs 60,000 | Rs 7.20 lakh |
| 0.75 | 15 | 45 paise | Rs 90,000 | Rs 10.80 lakh |
Add the demand charge effect. That same 300 kW at 0.85 presents 353 kVA to the meter; at 0.98 it presents 306 kVA. The 47 kVA difference, at a demand charge around Rs 250 per kVA per month, is roughly Rs 1.4 lakh a year on top of the surcharge.
These are Karnataka figures for the stated order year. Every state commission sets its own thresholds and steps, and revises them. Read your own tariff order and your last twelve bills before sizing anything.
How is power factor measured on your bill, and why does that matter?
Most Indian tariffs compute power factor as the ratio of kWh to kVAh recorded over the billing period, rather than as an instantaneous reading. The KERC order specifies that the ratio is determined to three decimals then rounded to the nearest second decimal, so 0.8449 becomes 0.84 and 0.8451 becomes 0.85.
Two consequences follow. First, it is an average and not a peak, so a plant sitting at 0.96 for most of the month and collapsing to 0.70 during a night shift with only a few motors running is billed on the blended figure. Correction has to work across the whole load profile, not just at rated production. Second, the rounding is a cliff edge: 0.8949 rounds to 0.89 and attracts a surcharge step, while 0.8951 rounds to 0.90 and does not. Targeting 0.90 exactly is a coin toss, so set the controller above the threshold with margin.
Where a DISCOM bills energy in kVAh rather than kWh, the mechanism changes again. There is no separate penalty clause because there does not need to be one. Poor power factor inflates the kVAh figure directly, and every unit of reactive current is billed as though it were working power. In those states, correcting power factor reduces the billed units, which is the most direct saving of all.
What power factor should you actually target?
Target between 0.95 and 0.99, not unity. The regulatory floor and the economics both point above 0.95, while the risk of over-correction points below unity, and the useful band sits between them.
The Central Electricity Authority (Technical Standards for Connectivity to the Grid) Regulations require that distribution licensees and bulk consumers provide adequate reactive compensation so that they do not depend upon the grid for reactive power support, and that the power factor of the distribution system and bulk consumer shall not be less than 0.95. That is the regulatory floor, and it sits above the Karnataka tariff threshold of 0.90.
Unity is a poor target for three reasons. Capacitor cost per kVAR rises steeply at the top of the range, because the reactive requirement to move from 0.98 to 1.00 is disproportionate. A bank sized for unity at full load will over-correct at part load unless the control is fast and finely stepped. And a leading power factor causes voltage rise on a lightly loaded feeder, which is penalised under some tariffs. A controller target of 0.98 with adequate step resolution holds the monthly average clear of both the 0.90 tariff threshold and the 0.95 regulatory floor.
How do you size the capacitor bank?
Capacitor rating in kVAR equals plant kW multiplied by the difference between the tangent of the existing power factor angle and the tangent of the target angle. It is one line of arithmetic, and the inputs are the two power factors and the real load.
| Existing PF | Target PF | kVAR per 300 kW | kVA before | kVA after |
|---|---|---|---|---|
| 0.85 | 0.95 | 87 | 353 | 316 |
| 0.85 | 0.98 | 125 | 353 | 306 |
| 0.80 | 0.98 | 164 | 375 | 306 |
| 0.75 | 0.95 | 166 | 400 | 316 |
Four things the formula does not tell you decide whether the panel works.
Use measured load, not connected load. Size against maximum demand and the load profile from twelve months of billing data or a logger, not the sum of motor nameplates, which produces a bank that over-corrects for most of the month.
Step size sets the accuracy. Five equal 25 kVAR steps cannot hold 0.98 on a plant whose reactive demand swings by 10 kVAR. Unequal binary stepping, for example 1:2:2:4, gives finer resolution from the same number of contactors.
Match switching speed to the load. Contactor-switched stages suit slow-varying loads. Welding sets, presses and rolling mill drives change faster than a contactor and its discharge interval can follow, and need thyristor switching.
Confirm the lowest load case. The controller must shed far enough not to over-correct on a weekend or a night shift.
Where should the capacitors physically sit?
Correct as close to the reactive load as the economics allow, because every metre of cable upstream of the capacitor carries the reactive current and every metre downstream does not. Most plants use a central APFC panel at the LT incomer, which is the right default but is a compromise.
Central correction fixes the billing problem completely and is simplest to control and maintain, but does nothing for losses or voltage drop inside the plant, because reactive current still flows through all the internal distribution. Group correction at a motor control centre relieves the feeder to that board as well as the incomer. Individual correction at a large motor removes reactive current from the whole path back to the source including the motor cable, and suits large motors running continuously at steady load.
Two hard rules apply regardless of location. Never connect power factor capacitors to the motor terminals of a variable frequency drive or a soft starter, because the drive output is not a sinusoidal supply and the capacitor will be damaged or will damage the drive. And never fit fixed capacitors sized for full load to a motor that runs unloaded for long periods, because self-excitation on coast-down is a genuine hazard. Where a plant is converting to variable speed, the reactive picture changes enough that the existing bank should be reassessed rather than reused, so the low voltage AC drives and the correction scheme belong in one conversation.
Why do capacitor banks fail, and what does the panel build have to do with it?
Capacitor banks fail through heat, through switching duty, and through harmonic loading, and the panel that houses them is as much a part of the failure mode as the capacitors themselves. A capacitor is a low impedance path to high frequency current, so any harmonic content on the bus concentrates in the bank rather than distributing evenly.
The switching duty is the part most often underestimated. Every stage energised draws a steep inrush and produces a bus transient. New India Electricals Ltd type tests its medium voltage panels for lightning impulse withstand at ERDA Vadodara under IEC 62271-200, and the reason that test matters is not only lightning. Capacitor switching produces the same class of fast transient, repeatedly, every working day, for the life of the installation.
The build quality point is the one that decides any LT panel. As Hemanshu Desai, CEO of New India Electricals Ltd, puts it, quality is what is inside the panel, because everything looks good coming off the shop floor and the difference only shows when it is wired up, fired up and loaded. On a capacitor panel that means busbar sizing carried through at the rated current rather than tapering after the incomer, properly crimped lugs, and cable specification that is not quietly downgraded. A badly crimped lug on a stage that switches forty times a day is a different proposition from one on a feeder energised once.
Three specification lines decide capacitor panel life. Capacitor duty contactors with pre-charging resistors are a different product from standard motor contactors, and substituting the latter destroys contacts quickly. The discharge arrangement must bring capacitors to a safe voltage before reconnection, and the controller interval must respect it, because reconnecting a charged capacitor is how contacts weld. And ventilation matters more than it looks, because capacitor life is strongly temperature dependent and a bank in an unventilated LT room in a Bengaluru summer is not operating in the conditions its rating assumes.
What do harmonics do to a capacitor bank?
Harmonics raise the current a capacitor bank draws and can create a resonance between the bank and the supply transformer inductance at or near a harmonic frequency present on the bus. Capacitive reactance falls as frequency rises, so a bank correctly sized at 50 Hz presents a progressively lower impedance to the fifth, seventh and eleventh harmonics, and is often the lowest impedance path available to them.
This matters more every year, because variable frequency drives, DC drives, induction furnaces and welding sets all inject harmonic current and all keep growing as a share of the industrial bus.
The CEA connectivity regulations set the limits at the connection point. Voltage total harmonic distortion must not exceed 5 percent with no individual harmonic higher than 3 percent, and current total harmonic distortion at the connection point must not exceed 8 percent.
The practical consequence is a sequencing rule: measure the harmonic spectrum before specifying a capacitor bank, not after the bank starts failing. Where significant harmonic content is present, detuned reactors in series with the capacitor stages shift the bank’s resonant frequency below the lowest significant harmonic. Selecting the detuning ratio is a study rather than a catalogue choice, and it needs the measured spectrum as an input.
New India Electricals Ltd has supplied capacitor banks with reactors at medium voltage on export projects, including a 2600 kVAR HT capacitor bank with associated reactors for a hot strip and cold rolling mill in Nepal, alongside the power transformers and control and relay panels on the same project.
What else improves when power factor improves?
Correcting power factor releases capacity you have already paid for, which is the benefit that never shows on the bill. A 500 kVA transformer at 0.85 power factor delivers 425 kW. The same transformer at 0.98 delivers 490 kW. That is 65 kW of additional real load on existing plant with no transformer, no cable and no civil work, which can defer a distribution transformer upgrade by years. Taking 0.85 to 0.98 also reduces current by about 13 percent, and resistive losses fall with the square, so copper losses drop by roughly 25 percent; from 0.75 to 0.95 the reduction is nearer 38 percent. Lower current also means lower volt drop along long feeders, which matters on sprawling sites where the furthest motor sees noticeably less than nominal voltage at starting.
An APFC panel specification checklist
Pull twelve months of bills and record the monthly average power factor, the kWh and kVAh figures, recorded maximum demand and any surcharge already levied.
Read your state tariff order clause on power factor, noting the specified threshold, the paise per unit step and whether billing is on kWh or kVAh.
Log the load profile across a full production cycle including night shifts, weekends and shutdowns, so the minimum reactive demand is known rather than assumed.
Measure the harmonic spectrum at the LT bus before sizing, and state the measured THD in the enquiry.
State the target power factor with margin above the tariff threshold, typically 0.98, rather than the threshold itself.
Specify the number of stages and the stepping arrangement, and check the smallest step against the minimum reactive demand from the load log.
Specify capacitor duty contactors, discharge arrangement and reconnection interval, not just capacitor kVAR.
Confirm capacitors carry the ISI mark to IS 13340 (Part 1): 2012, mandatory for shunt power capacitors up to 1000 V.
State the ambient temperature and ventilation of the panel room, and the enclosure IP rating required.
Exclude any motor fed by a VFD or soft starter from the terminal correction scheme.
Frequently asked questions
- How much can an APFC panel save on an industrial electricity bill?
- It depends on the starting power factor, the tariff and the consumption, so calculate it from your own bills rather than from a percentage claim. On the Karnataka tariff, an HT plant consuming 200,000 units a month at 0.85 power factor pays 15 paise per unit in surcharge, which is Rs 30,000 a month or Rs 3.6 lakh a year, because the specified figure is 0.90 and the surcharge is 3 paise per unit for every 0.01 below it. Reducing apparent demand from 353 kVA to 306 kVA saves a further sum on the kVA demand charge.
- What power factor should an industrial plant maintain?
- Between 0.95 and 0.99 as a monthly average. The CEA (Technical Standards for Connectivity to the Grid) Regulations require distribution systems and bulk consumers to hold power factor at not less than 0.95 and not to depend on the grid for reactive power support, which sets the regulatory floor. State tariff thresholds are often lower, with Karnataka specifying 0.90 for HT and 0.85 for LT, but meeting the tariff threshold is not the same as meeting the connectivity requirement. Unity is not recommended, because capacitor cost per kVAR rises steeply at the top of the range and over-correction at part load becomes likely.
- How do you calculate the kVAR rating needed for a capacitor bank?
- Multiply the real load in kW by the difference between the tangent of the existing power factor angle and the tangent of the target angle. Taking 300 kW from 0.85 to 0.98 requires 125 kVAR; from 0.80 to 0.98 requires 164 kVAR. Use measured maximum demand rather than connected load, because sizing on motor nameplates produces a bank that over-corrects for most of the billing period. Then check against the minimum load condition, because a bank that cannot shed far enough pushes the plant into leading power factor on a night shift.
- Can capacitors be connected to a motor running on a VFD?
- No. Power factor capacitors must never be connected to the motor side of a variable frequency drive or a soft starter. The drive output is a switched waveform rather than a sinusoidal supply, and connecting capacitance to it can damage both the capacitor and the drive output stage. A VFD-fed motor does not require terminal correction in any case. Where a plant is converting drives to variable speed, reassess the existing correction scheme rather than carrying it over, because the reactive profile at the incomer changes.
- Why do capacitor banks fail early?
- Heat, switching duty and harmonic loading, usually in combination. Capacitor life is strongly temperature dependent, so an unventilated panel room shortens it directly. Every stage energisation draws a steep inrush and produces a bus transient, and standard motor contactors substituted for capacitor duty contactors fail quickly under it. Harmonic current concentrates in capacitors because capacitive reactance falls as frequency rises, so a bank on a bus carrying significant drive or furnace load runs hotter than its 50 Hz rating suggests. Measuring the spectrum before specifying, and confirming ventilation and contactor type, prevents most early failures.
- Do I need detuned reactors with my capacitor bank?
- Only the measured harmonic spectrum on your bus can answer that. Detuned reactors sit in series with capacitor stages and shift the bank’s resonant frequency below the lowest significant harmonic present, so the bank cannot resonate with the supply transformer inductance. They are necessary where variable frequency drives, DC drives, induction furnaces or welding sets form a meaningful share of the bus load, and unnecessary where they do not. The CEA connectivity regulations cap voltage THD at 5 percent with no individual harmonic above 3 percent, and current THD at the connection point at 8 percent. Measure first, then specify.
Bring twelve months of bills, a load log covering the lowest load condition and a harmonic measurement to New India Electricals Ltd before a kVAR figure is fixed. Those three inputs decide the panel, and all three are cheaper to gather at enquiry stage than to discover after commissioning.
