Both Generators Trip While Running in Parallel

Hi,

First find what is the load sharded by each generator. When one govenor fails , the generator will trip on reverse power. If the total load is above the capacity of remaning generator then this will trip on overload.

You didn't mention detail about your load and the generator capacity. If one of the generators fails, is the second generator able to supply the loads in your system? If not, is there load shedding system active so that a part of loads are disconnected and the second generator is able to supply rest of the loads? etc, etc.

You mentioned 'one generator was on reverse power and the other was on overload before blackout occurred' but didn't mention which one tripped first. It may happen that the first one tipped by reverse power by some reason and all the loads then move to the other generator and it tripped by overload. This may happen if the load shedding system is not properly designed or not active.

It also may happen that one of the generator trips by some reason and the second generator becomes unstable due to the large impact in the system (see http://cr4.globalspec.com/thread/51895/Power-Angle).

- MS

Thanks a lot for the replies.

The present arrangement is 2 gen. each of 1.5 MW rating, running at 80 -85%, i.e. about 1.3 MW. [Management did want to run them at 100% load but was declined because of our concerns - overloading and the effect of sudden impact load (instantaneously +100%=200% total load) to the diesel engine when the other generator trips. Consultant suggested that it should be OK though.]

Tests had been conducted when both generators were running at 85% load; one generator was tripped manually (open one gen. circuit breaker). The other generator could picked up the total load (2.6 MW=170%), load shedding took place within seconds and completed at in about 15 seconds (mechanical delay). Total load would be below 100%. No blackout occurred.

At issue is how about if a governor fails? Is it possible for the generator with a failed governor goes to full fuel position, causes the good one trip on reverse power, then it trip on over frequency or overspeed itself? In that case, it doesn't matter how good the protections settings are, they will not prevent a blackout but rather protect the individual equipment from suffering severe damage.

Hi, Dwong.

How is the load-sharing implemented? If the load-sharing module does not get alarm status from each unit, it simply will not know if a unit has tripped or has a governor problem. It might sense reduction in output from the defective unit and load up the the healthy one. What kind of governors? If they are mechanical are they smooth throughout the travel length? 'Sticky' governor response does funny things to controllers.

I can't say much unless you clarify in what sequence the tripping occurs. Is it the reverse power first and then over-load (more logical), or at the same time? If it is one after the other and it is always the same unit with repeating alarms, I would look at the load-share unit first. The governors seem okay if they are responding to load changes while working individually. Again, I have to ass(ume) this is a new installation. If the problem is new the start with the communications between the generators.

Rev power relay could act for governor malfunction as well as some types of avr mal operation.if exact sequence and settings are available it may be possible to determine exact cause.

I would like to add that reverse power relay would come in operation when the prime mover is starved of power and the generator remains connected to the grid. So if fuel supply system for the two generators is common, one generator prime mover may draw most of the fuel (may be due to Mal functioning of governor) resulting in starvation of other engine resulting in reverse power relay operation. Increasing the time delay of reverse power relay may help

[Management did want to run them at 100% load but was declined because of our concerns - overloading and the effect of sudden impact load (instantaneously +100%=200% total load) to the diesel engine when the other generator trips. Consultant suggested that it should be OK though.]

My first take on this is that you are tied in with a grid versus running in an island mode. Otherwise there are no choices to make, the load that comes into the system is whatever plant load that is available.

Regardless, the input from your mechanical/'electrical governor will be in the form speed. If you are tied into a system, the system will appear to control the speed of the prime mover. In actuality the system is having the same effect on the speed of your unit as if it were running in a island mode only the system is usually so large and the load is shared by so many units no noticeable effects are seen.

The way that the governor works in a system (large or small) is to work with some percent regulation, The governors are designed to control the fuel or valves so that if full load is applied to the governor the fuel or valves will travel their full travel to pick up the load which will cause the speed to decay by the system design regulation. 5% speed droop/regulation is normal in the US. So given this full load application and the consequent speed droop the speed has to be driven back up to 100% speed to get the frequency back where you want it. This leaves the governor setting at 105% of rated speed. If the breaker is opened the speed will increase to 105% and will have to be reset to get back to synchronizing/rated speed.

Given this same situation with a 50% application of full load, the fuel/valves will travel 50%, the speed will decay 2.5% and the load will be accepted at that speed until the governor is raised in speed setting to 100% rated or 102.5% actual. Within the confines of mechanical reality this will occur at any load versus speed.

The significance of all that is that if a 10 MW unit is running in parallel with a 1 MW unit and 50% load is applied to that system the 10 MW unit will pick up 5 MW and the 1 MW will pick up .5 MW and the system speed will be at 97.5% rated.

Throw a couple of 600 MW fossils and a few 1200 MW nukes in there and you start to get the feel for the effect of the 1 MW unit.

I think the reason that your consultant wasn't concerned with jump loading the unit to 200% was that the prime mover/governor can only go to 100% open, most prime movers are designed for valves wide open/fuel rack full travel, relative to load. I am not quite sure the generator would relish this operation.

At any rate relative to your questions if your connection is as an island configuration and the breaker does not open, the governor is going to satisfy the load until it can't within design spec's, then it is going to maintain valves wide open and the machine is going to slow down. With the kind of loads you are talking about I would suspect under frequency trip if you have one or lacking that reverse current.

"one generator was on reverse power and the other was on overload before blackout occurred" just a quick reaction - the one in reverse power trip I assume tripped first then the second one tripped on overload. the first tripped on reverse due to some reason - one reason - governor failure thus no more primary drive or a field failure turning the generator momentarily as a motor. when the first unit tripped due to reverse power, it overloaded the second unit thus the second unit tripped on overload.. i hope this will help.

then it appears the trip is correct.only the load shed has to be coordinated to prevent recurrence.on field fail alone the unit will run as an induction generator.unless the governor is also affected in some way or the line voltage is distorted to give illusory reverse power.

Hi EnergyConsultant,

Thanks for the input.

-The incident described is my observation only. You know when blackout occurs, time seems frozen, yet the mind turns faster than the computer trying to restore the power. It is difficult to keep track of everything. Even printed event log are too slow and probably cannot directly point out the cause. I think it needs some very accurate instruments to record ns electrical data to review the events and determine the cause. In this case, the plant does not have these instruments.

-Sometimes governor failure would give full fuel.

-If the reverse power one has tripped, shouldn't it be out of the picture and could not overload the second unit? I have no knowledge about the current and voltage vectors under these conditions.

-How about the settings of reverse power and overload trips?

Dwong,

Some informations is missed,

Did your generators individualy work fine? frequency stable at dynamic load changes, if so your governors are fine

Is Your fueling system tested? if you have a poor fueling system, at high load ( high fuel cosumption) one generator will be poorly fuel supplied and will go in reverse power condition, causing the other one to be overloaded..

What engine you are running? what is the governor brand? a wrong setting of governor parameters as gain and stability will cause a fluctuation on the load with dynamic change and may cause reverse power

What controller you are using? also setting of dynamic parameters is as important as governor setting....

Does your controller has an alarm log? you can check what is the fault that cause the trip and wich generator trip first and you can diagnose...

What type of controller you use? ( brand)

What type of load share your controlloers use? analogue? can bus? a communication fault ( lose wiring) may cause also this fault

I hope my reply is helpful

Dear Mr. dwong,

i have read the opinions and answers by forum members and would wish to sum that up with some additional experience.

1) For proper load sharing(KW) of DG sets(in proportion to their KVA rating) , The governor should have droop settings i.e they should be asynchronous. Isochronous generators will always try to keep the speed constant and hence frequency. For DG sets to be operated in parallel in a better manner, engines with electronic governor is recommended, which will respond to the dynamic changes in KW output and hence the speed of alternator.

Now coming to your problem, when the governor fails(as per your observation), the engine will not be able to maintain the nominal speed of alternator. when the speed drops and since the generator is connected to the synchronised system, the power will flow in a reverse direction, which will activate the reverse power relay and the first dg trips.( also please check the excitation)

when the first DG trips, entire load will come on to the 2nd DG. since the load exceed the capacity of 2 nd DG, and since DG sets take time to adjust to the dynamic and transient loads, the overload relay will sense and 2nd DG set trips.

As mentioned by forum members in this thread, please mention whether the synchronisation method is manual/ or through PLC and genset controller.

Any opinions/information/advice additional than the above is highly appreciated.

Thanking you

For such small systems we are extending load control only to one machine .Although the load sharing may not be as fast, risk of mal operation is considerably reduced and very little hunting probability;hunting cycle if slower than reverse power relay time delay could lead to rp relay trip. With about 4-8% droop setting and with correct settings and periodic testing there will be very high operational reliability. For avr the droop ct connection is often found to be the cause.If the system is on load sharing and if for some reason it goes to droop mode with a ct not ok sure reverse power or field fail trip can be anticipated. we also give a plc based load shedding schemes which back up df/dt load shed.If the tripping is soon after synchronising the often found cause is controller settings.

Dear Mr. DWONG,

First Point:

You have not mentioned about the Power Factor. At the time of !st generator tripping, what was the Power Factor, was it Leading Power Factor or Lagging Power Factor.

If it is leading Power Factor it will result in higher terminal voltage at the alternator, and may lead to trip. Please check up this and if possible report further.

2nd Point:

It is possible that one Governor has a Steep Droop Character, combined with steep voltage droop from No-Load to Full - Load, resulting in cumulative impact which leads to tripping.

Generally the variation in DIFFERENCE for the droop character for the Governors,for putting in PARALLEL OPERATIONS, and Voltage droop in a combined manner, should not exceed 2.5% and theoritically it should be zero, but in practise there will be a minor difference.

To make it clear if one system the droop character is 6% the other set droop character should be with in 6% plus or minus 0.15% which means 6+0.15= 6.15% to 6.0-0.15=5.85%. Please check this also. If this limit exceeds there will be problem.

I hope my points are clear to you.

Let other CR4 MEMBERS ALSO give their opinion.

Thanks.

DHAYANANDHAN.S,

CR4 MEMBER,

INDIA.

in parallel running of gensets, When one genset fails because of defective governor controller, It will eventually driven by the other. The tripped genset generator will drive the prime mover and will act as additional load to the system. That's why you observed an overload on the existing genset. Also, sudden tripping of one would cause transient/surge current that send wrong signal to overcurrent relays thus tripping its mcb. You may have to let your EE consultant check the relays setting (Reverse power, overcurrent, etc.). I wonder what went wrong with your gov. controller?

Hi All,

Again, thank you for the replies. I am grateful to you sharing your knowledge and experiences with me.

More equipment info:

-Woodward electronic modules including: Load Sharing & Speed Control 2301A, SPM-A Synchronizer and Automatic Gen. Loading Control. Woodward EGB Governor.

-Basler Electric Over Current relay BE1-51/27, Generator Differential relay BE1-87G, Loss of Excitation relay BE1-40Q.

-Siemens PLC.

The only problem at hand is to convince my peer that doesn't matter how many protections and load managements, blackout out could still happen due to equipment failure.

I only had one blackout experience with this plant and it was not a good one. So it is hard to swallow my peer's comment.

Best regards.

The available devices appear to be basically sufficient to ensure safe parellel operation if tested and configured correctly.Especially ther plc should have surely prevented a blackout. wonder why it dint.Better have a specialist to set right.

Protection devices are to protect your equipments in case of a mulfuction or a fault, not to prevent a blackout, exactly as an air bag in your car , it protects you in case of an accident but can not prevent the accident.

You got one time blackout caused by a reverse power, a wrong setting cannot generate a false trip when you are running at 85% load, your RP relay did exactly what it is supposed to do, when you are running at 85% and suddenly you got a reverse power alarm that means you have had a real reverse power.

The reason in such a case is only a fuel supply problem, either caused by a malfunction of the governor or the fuel supply itself

as it happened once and never repeated, I assume the governor is doing well, it is something in fuel supply, a fuel filter can cause this,

Your boss should understand that blackouts happen in real life and protection devices are not to prevent them, they are to protect your equipments in case of faults

just watch your installation for the next blackout to see what exactly go on when it will occur

Regards

Hi Le Nobel,

Agree with you 100%; especially that protections are mainly for the individual generator only. I don't want to have another blackout, not in my watch, please. Too bad you are not my boss.

My first post stated that "Defective governor or load sharing control were the causes of some of these failures (others - human error!)." In this case, it was the EGB governor mechanical failure. It happened a few years ago.

A better design of the system, to increase the reliability for this kind of operation, is a split-bus design. The chance of both systems fail at the same time due to equipment failure is extremely low. Of course, some common factors like human error or contaminated fuel may still cause a blackout. Owners may not like it as it means higher capital cost, unless required by regulatory bodies. I had worked on a split-bus plant; it is much more reliable than a parallel one.

We have a stand by generator; my "boss" wanted me to take it out during operation to carry out some minor minor repair which could easily be done during the daily non-operational hours. His argument was "the plant will not blackout because of the protections, load shedding ...". (I like my other peer's comment - 'Only a canoe engineer will say that.') I had declined his request. Now I am in real trouble!

I am still having some problems with how you can pick and choose your load and not have a system tie.

At any rate, if you overload the generator the prime mover will slow down, if it takes 15 seconds to shed a reasonable amount of load to get the system back to normal, no one, not you, your boss and your associate is going to predict what is going to happen when you have the other generator fail.

If you can not or will not run reserve power or tie into the system I would suggest a bank of batteries and emergency lighting.

hi all ,

in place of these equipment like SPM-A , 2301 etc try to use new synchronizer of woodward like easYgen 3200/easaygen 2500/ GCP-32. these controller are very advance & fast & you can make load shed logic in that.

for any query contact me dev22_abes@yahoo.co.in

with all the suggestions and opinions on the cause of the blackout, you may want to design a system that provides capacity redundancy when equipment failure happens. In any power generation system, there will ALWAYS be failure and depending on your system configuration, you either survive such an equipment failure or end up in a black out condition. In base load power generation systems, we design the configuration of the units such that if a unit is lost for any reason, all the remaining units should be able to take up the demand automatically. depending on how much money you can spend (and that is directly proportional to the criticalness of your demand), the more units sharing a demand the better. The basic principle is having a unit as a spinning reserve such that if one unit fails, the remaining healthy units can take up the total demand with out failure. If the total demand is kWT, kW1 is the capacity of one unit, the minimum number of unit is N+1 = kW1 + kW1. This is what we call as 2 x100% redundancy. If we use 3 units, the number of units to carry the total demand is 3 x kWT/3 = x kW3. but since we the rule is N+1, the minimum number of operating units then is 4 x kWT/3. Clearly, as we increase the number of operating units, we reduce the capacity per operating unit, but reducing the possibility of a blackout. The myth by non power generation managers is such that by providing "ENOUGH" protection and metering, equipment failure or in this case blackout is prevented from happening. In reality, that is never true. Such that if you think that even if your equipment is fully functional blackouts WILL HAPPEN in your system simply because you got no reserve capacity. And what I describe above s ONLY FOR SPINNING RESERVE. YOu still to ADD UP what I call as stanbdy reserve - a unit that is equal in rating with the other units to take the place of one failed or unavailable unit due to forced outage or normal maintenance. The other end of the mitigation measure is to provide CORRECT load shedding system. In many power systems I'd studied and advised, the common load shedding trigger is a frequency setting. System frequency decays as a result when capacity is lower than demand. if your system is using analog systems, it places you in a deeper hole. I suggest that you install a frequency relay that has a df/dt function, rate of change frequency setting relay or a voltage relay with dv/dt, rate of voltage change setting to trigger load shedding actions that uses frequency rate of decay or voltage rate of decay as a result of capacity loss to trigger the isolation of demand to save your remaining unit on line. but check how fast your demand side circuit breakers - if you are using ordinary molded case circuit breakers - 8 to 16 cycle to trip - these are slow acting thus you need to consider the time delay of the MCCB to react in your time setting calculations as it may introduce delay too much to matter and brings the system down anyway. if your circuit breaker is electronically set and tripped - 3 to 5 cycles - then you have a better chance. Bottom line is the solution to your situation is not simply to know why it happened but rather to understand the peripheral issues of system failure and attendant mitigating measures to prevent it happening again. at the end of the day, it will require more money to prevent a blackout. Thus depending on how critical your demand is, maybe you can ask your boss to invest more...

Even if frequency and voltage load shedding is not present the plc SHOULD have prevented the black out. This looks like the simplest first priority.

Mate bottomline is how much is a facility owner is willing to spend to ensure availability of supply the plc maybe there but set with wrong assumptions is as good as it was not there. All I am saying is with all our good intentions to help the guy I would defer to the facility owner and manager what's best for them it's not about the simplest or most complex solutions it's about what the owners are willing to do and spend for And that decision is directly proportional to the quality of service the owner wants I believe we all said the good things to help the owner keep participating ...

What i was trying to say is that with correct plc blackout in this situation i s very less probable. The fist point may be to check why the plc did not act.probably it. may be as simple as a fouled relay contact. If a big investment is sought and afterwards if some one diagnoses the member will feel even more bad.d

I dont disagree with the PLC function - but remember the plc is NOT the ONLY quipment to effect the load shedding. remember that the output of the plc will have to be communicated to the circuit breaker that isolates the load to be shed. If the reaction and operating time from the PLC sensing/processing and sending a trip signal to the circuit breaker, is added to the time for the circuit breaker to react and actually open its contact, that will be in the order from 15 to 30 cycles depending on the type of circuit breaker being used for the load shedding system. If the load shedding circuit breaker are the slow type, then you got no chance because the 32 device is set on instantaneous trip to open the generator circuit breaker. Then when the 1st unit tripped by its 32, then the second unit is suddenly loaded twice as much as it was loaded prior to the 32 trip on the fist unit. The second unit 50/51 device took this sudden overload as a "short circuit" on its external system, thus the second unit tripped on overload (51) or even on SC condition (50) - which again is set instantaneously. No amount of PLC programming can solve that mate!! - because you got no extra capacity to ride out the outage of one unit!!! Thus is this case the 2 x 50% configuration we have here does NOT HAVE A CHANCE to survive a loss of a unit!!!. Thus the Owner must think of installing a redundant unit if he wants to survive a similar situation in the future - to provide a spinning reserve to the generation capacity only to give time for the PLC driven load shedding system to work.

Perhaps the member may resolve simpler issues like device co ordination testing etc before asking the owners for investments. we have seen up to 220kv 100 mva systems running successfully on plc load shedding.only rider is if remote terminal units are unavoidable more care may be exercised.incidentally non unit protections like reverse power etc are not usually set for instantaneous operation unless interlocked with say valve limit switch etc to avoid unintentional trip.

in this case if reverse power relay is set for fast operation that may be a cause.Since during minor power swings the relay is sure to trip.the reverse power relay pick up need not be less than 10% unless the engine is very very well run.The point is all such simpler things may be resolved before fresh investments are requested.

I don't see with the equipment list Dwong provided where the PLC interfer in load shedding,

it is most likely performed by Basler over current relay BE1 51/27

it could be used either to shedd the load, in this case output delay should be decreased

or to trip the generator breaker, in this case delay should be increased a little to give more time to load shedding circuit to act

regards

Updates.

Recently, while running in //, one of the generator overspeed tripped (mechanical overspeed setting problem). Basically it did not affect the other running generator; of course, load shedding occurred.

Same equipment, each running at about 80% load but it did not cause a blackout this time. I am not sure if it is because of the settings/PLC had been changed or the nature of the failure was different. I believe what have happened this time: somehow the mechanical overspeed trip mechanism was activated, the engine fuel rack was pushed to zero fuel then; the generator had reverse power and generator circuit breaker opened, PLC activated load-shed. No overload alarm on the remaining gen. Mechanical os mechanism adjusted; it seems problem solved.