Virtually all industrial power factor correction is done to correct lagging power factor.
Industrial users are typically charged a penalty for a lagging power factor.
There are two main approaches to penalties:
Old school, the penalty was based on a monthly average of PF based on monthly KWHrs and monthly KVARHrs.
With the arrival of electronic revenue meters, penalties may be based on realtime values rather than averages.
Old school, bulk power factor correction and load switched correction was used.
With penalties based on real time values, real time correction must be used.
When a customer came to us with an issue with power factor penalties, I would follow the following steps:
1. A visit to the customers accounting department. A polite request for copies of the power bills for at least the last year and for the last two years if possible.
2. When penalties are charged for poor power factor the following information will be found on the power bill;
Monthly KWrs and Monthly KVARHRS.
The KWHrs and KVARHrs for each month would be entered into a spread sheet.
3. PF penalties generally started at a PF of 0.9 A column would be added to the spread sheet to calculate the KVARHrs needed to correct the PF to 0.9
4. Evaluation; Look at the values for each month. If one month was extremely low, it may be excluded from future calculations.
If one month was extremely high inquiries would be made as to why it was high.
Another column may be added to the spread sheet; The KVARHrs to correct the worst month to 0.9 would be calculated and then a column would be generated to determine what effect that would have on the average PF of the other months.
5. A target for a corrected average PF would be selected. Possibly right on 0.9, possibly 0.92 or perhaps 0.95.
A spread sheet column would be generated with the data for any extremely low months or extremely high months excluded.
A calculation would be done to determine the penalty that would be charged for the extra KVARHrs in the worst month.
Was this worth spending money to correct or can we live with this once or twice a year?
6. Decision time: What will be our target corrected average PF? How many KVARHrs will that take?
How may we apply those KVARHrs?
a) Bulk correction. During times of light loading, capacitors tend to drive the voltage up. Too much correction during periods of light loads may result in the voltage going so high as to damage equipment and burn out lights.
A value based on experience will be chosen. A value that results in a capacitive current of twice the magnetizing current of all the transformers on the system is safe for unswitched capacitors.
Multiply the KVARs by the hours in a month and compare with the needed KVARHrs.
b) Look for large motors or a group of small motors that work together 24/7 and switch capacitors in with the motors. Refer back to item "a)" and beware of system over correction if you over correct the motors.
Multiply the KVARs by the hours per month that the motor(s) run and compare with the needed KVARHrs.
c) Look for large motors that run continually during working hours. Again use judgement. Many large motors are part of a process with a number of smaller motors that run when the large motor is running. Considering this, it is often feasible to use switch twice the KVARs that would be required to correct the large motor to unity.
d) Continue adding protection to progressively smaller motors.
At each step multiply the KVARS by the hours per month that the motor(s) run and compare with the needed KVARHrs.
Power factor correction when real time KVAR consumption is monitored and penalized.
Look at the figures for the worst month.
Divide the KVARHrs by the number of hours in a month to get a value in KVARs.
Buy that many KVARs of PF correction capacitors and a Power Factor Correction Controller.
I have only once ever had to do a calculation to determine the actual capacity in Farads or mfd.
That story also answers Gunnar's question as to why bother for 50 vars.
I had a customer with a small plant and a small PF penalty.
Only a few cents a month, but he wanted to know why, and how to get rid of it.
I forget the actual VARs involved, but a run capacitor intended for a fractional HP motor hidden in one of the panels and connected to one phase as enough to avoid the penalty and make the customer happy. The payback period was less than one year.
Memory fails but I will guess less than 1000 VARs.
The industrial power factor correction capacitors that I used were always rated in KVARs at a stated voltage and frequency.
Bill
--------------------
"Why not the best?"
Jimmy Carter