Riddle Number Two: Synchronization of two generators

BEWARE: This one is a toughy (unfortunately this one we haven't figured out) and requires a bit of basic electrical engineering! For those that are not interested in reading a bunch of technical mumbo jumbo I've interspersed pictures of Mpeketoni scenes for you instead.

Evening in Mpeketoni town

Since the big 150 kVA generator went down we have not had enough capacity to meet the load during the day and night. However, the overhaul was completed on the second small generator so we now have two operating gensets – one rated at 60 kVA and the other rated to 57 kVA. Unfortunately having two small gensets doesn’t allow us to meet more demand unless the two generators can be run in parallel.
To run two or more generators in parallel, you must first synchronize the generators so that they are operating at the same voltage, at the same frequency, and are both producing power in phase. If these conditions are not met one generator will act as a load for the other one and cause all sorts of problems, so it is very important that the process is done right.
The generators produce AC power on three separate lines or phases, here they call it the Red phase, the Yellow phase, and the Blue phase. Each phase has a similar sinusoidal pattern to the voltage and current, but the phase of the pattern is shifted by 120 degrees between the three phases.
The synchronization process ensures that the AC pattern on the Red phase of Gen1 is in phase with the Red phase of Gen2. If they are in phase then both the Yellow and Blue phases will be in phase between the generators.
MEP has a synchronization panel that is used to synchronize the generators. The main components of the panel are three switches: one switch for each generator and one for the load. If all switches are closed then the Red phase of Gen1 is electrically connected to the Red phase of Gen2 and the Red phase of the load. However, the switches for the two generators include a safety feature that doesn’t allow the switch to be closed if there is a large voltage difference across the switch.
To synchronize the generators the panel has three dials to help with synchronization. The first is a frequency meter for each generator (f). The setting on the governor of each generator is adjusted until the frequency (or rotational speed of the engine) is exactly the same as the other generator (f1=f2). The next dial indicates the voltage of each generator (V). The voltage is controlled by an accessory called the AVR or Automatic Voltage Regulator on each generator. You next adjust a setting on each generator’s AVR until the voltage on each generator is exactly the same (V1 = V2). The last dial is called the zero voltage dial. The dial indicates the voltage difference between the Red phase of each generator (V0). Only when the frequency and voltage of the two generators is the same and the lines are in phase will the zero voltage dial drop to zero (in phase only if V0 = 0).

Wiring diagram for synchronization process

Once the zero voltage dial drops to zero, you can close the switches between the generators (S1 and S2). The generators are now synchronized. When the load switch is open and the generators are synchronized, there should be no current flow on any of the lines (i1=i2=0).
If everything goes okay with the synchronization, you can now close the load switch (SL) and start providing power with the two generators (now i1 and i2 are set by the load). Even if you only use one generator at a time, you still use this same panel, you just leave the switch open to the generator that will not be in use. At this time we can use either generator on its own to run the load, so we know all of the switches and circuitry are in order.

Phanuel Mwaimbe, the manager of MEP

Here is the problem. We are able to match the frequency and the voltages between the generators and when the zero voltage dial drops to zero we are able to close both generator switches (S1 ­and S2). However, when we close the switches the current meters in each phase on both generators quickly climb to very high currents even when the load switch is still open. The frequencies remain equal and the zero voltage dial (V0) remains stable showing zero volts. After about 30 seconds or so the generators trip off and stop producing current. As soon as they trip off the synchronizing panel opens the two generator switches as they loose synchronization.
Normally with one generator, when you engage the load the currents quickly climb and you hear the diesel engine suddenly change sounds as it struggles to meet the new high load. However, when we close the two generator switches in the synchronization process and the currents quickly climb we do not hear the diesel engine change noises; it sounds as through the engines remain idling even with extremely high currents.

Young girl enjoying an evening in Mpeketoni

Since the line-to-line voltage is positive and the line currents are positive, the generator apparent power (kVA) is very large. However, since there is no load and we do not hear the diesel engine sound change we can infer that the real power (kW) is close to zero. The power factor is the ratio of the real power to the apparent power (kW/kVA), which in this case is zero. A zero power factor indicates the currents are 90 degrees out of phase with the line voltages and that all of the apparent power is made up of what is called reactive power (kVAR). I’m guessing that since there are only the two generators connected to each other and no other equipment, the reactive power is due to one generator acting as a purely inductive load to the other generator. In other words one generator is producing kVARs and the other is consuming kVARs.

Kulia, one of the MEP casual workers, installing a new section of three phase line

The question is why are they acting that way and how do we stop it? Unfortunately, the synchronizing panel doesn’t have any documentation or instructions and none of the current staff was around when the panel was first installed. At least one of the staff was around when they were still using the panel just over a year ago, so we are positive that it worked with these generators before one of them was taken off line for a major overhaul (which included rewinding the stator of the generator).
The only clues we were able to find are from past letters filed away in the MEP records. The first one is from the commissioning of the panel in 1998. The letter indicates that the panel was installed and appears to be working, but that it cannot be used to run the generators in parallel because of high “inter-circulating currents”. They blame these currents on differences in excitation of the rotors. The excitation can be adjusted by changing the “voltage droop” setting on the Automatic Voltage Regulator (AVR) which controls the excitation through a “droop current transformer” or droop CT. Adjusting the voltage droop allows you to minimize the inter-circulating currents. The letter goes on to state that the droop CTs were not installed so these currents could not be minimized until they are installed.
A second letter from 2000 indicates that the droop CTs were installed and the synchronization of the generators was successful. The installers trained the MEP technician on the synchronization process and passed along copies of the documentation to the technician. Unfortunately the technician mentioned in the letter is long gone and it is likely that when he left, the documentation left too.
So we found the voltage droop setting on the AVR, but adjusting the dials while the two gensets are synchronized has no visible effect on the magnitude of the circulating currents. We are not sure what a droop circuit transformer is or where it might be located, so we are not even sure that they are still installed.
So now what? We seem to be very close to having the problem solved, which would allow us to stop rationing power and continue normal operations, but we’re stuck. What is it that is causing these inter-circulating currents and how do we go about getting rid of them? How can we tell if the voltage droop settings are correct, or if changing the dial has any effect on the excitation of the rotor? Any ideas out there? See, I warned you this would be a tough one.