Bitzer screw on VSD "stalling"

[garyb]

Hi All

We are commissioning two Bitzer screws driven by Siemens VSDs on an air to water heat pump system for a swimming pool complex, +8/50oC R22, speed range 30-60Hz, lead/lag, unloaded start, VSDs control to suction pressure. We initially had some difficulty getting the screws past 15-20Hz but got over that by applying starting boost in conjunction with the overload service factor.

Now that the complex is approaching design conditions we are having problems with the lag machine "stalling" at around 35hz and sending the current high. The lead machine operates fine but always the lag machine is the one that surges the current at around this speed. It makes no difference which of the machines is in the lead position. If the machine are run together in parallel the problem does not occur and both will run up to full speed. I don't want to leave it in parallel as it greatly increases the minimum turn down capacity - obviously the best efficencey is at the highest suction pressure allowable for the duty. I have got around similar problems on Danfoss drives with the torque control function but the Siemens doesn't appear to function quite the same.

Either way I'm open to suggestions!

Regards
Gary

[SteinarN]

I have installed some systems with Danfoss drives. I'm not familiar with Siemens drives, but the basics should be the same nevertheless.

When you speak of current, I assume it is the output current of the drive. I also assume the drives has higher continous ampere output than the compressor nameplate amperage.

When you says it's sending current high, is it higher than the nameplate amperage?

Important parameters:
The torque must be set to constant torque, not variable torque as would be normal for centrifugal fans. Ref. Danfoss parameter 101. Set voltage, Hz and amperage from nameplate, but be very careful when setting motor power. Many (most) compressors dont list the actual motor power on the nameplate, but instead an almost randomly selected number for motor power. Take the amperage, myltiply with the voltage (lowest value if it is listed as say 380-420V), multiply with 1,73, multiply with 0,85 and multiply with 0,8. This should be roughly the motor power for a motor of say 10-50kW. If this power deviates considerably from the nameplate power, then use this calculated motor power instead of the nameplate power. Run an auto tune.

One other important parameter regarding the start capability is parameter 110 on Danfoss drives. I set this parameter to 200% on my Danfoss drives.

[nike123]

Take the amperage, myltiply with the voltage (lowest value if it is listed as say 380-420V), multiply with 1,73, multiply with 0,85 and multiply with 0,8.

What is this last factor in your power calculation?
I presume that 0,85 is cosØ

[Electrocoolman]

Hi gary,

I think you need to post your compressor and drive model numbers so that we can help you further.

As my colleague Steinar says, you will need a drive suitable for Constant Torque load. These have a much greater torque capability. Drive manufacturers offer a range of drives thesedays, dependant upon the type of load they are expected to drive (as a means of reducing the price and staying competitive)...'simple' ones for fans and waterpumps are not suitable for compressors or hydraulic pumps...(Complex machines)

The vector control equations are different as well as the V/F curves.

I am not very familiar with screw compressors, except to say that they use slides to vary the load / compression ratio. Is the problem also linked to the control of these, that somehow the slide is loading the compressor at 35Hz, thus the motor is drawing more current, at which point you are trying to accelerate the motor. Are the two control systems (i.e slide and VSD) fighting one another rather than operating in unison?

ECM

[garyb]

The compressor name plate rating is 90A, the selection program suggests it should be 79A at design conditions (60Hz - the program won't let me change the speed) but it goes to 120-130A during the stall before the thermal model trips out. This will be approaching the drives current limit for normal running. Additional heat is being generated in the motor and if it doesn't recover then motor thermistors have tripped once.

I guess it has something to do with the lag trying to ramp up against the high discharge pressure generated by the already running lead machine. There is a "continuous boost" that we have experimented with but it only results in more motor heating.

[garyb]

The drives are Siemens type SED2 - I don't have the capacity details with me as they were supplied by our automation subcontractor. compressors are 2x HSK6451-50.

After the unloaded start at 30Hz and a delay for stabilization the machines are mechanically loaded to 100% with all subsequent capacity control being done by the drives.

One of the parameters call for a selection of the volts/Hz relationship as linear or parabolic - opinions as to which is appropriate?

[SteinarN]

The drives are Siemens type SED2 - I don't have the capacity details with me as they were supplied by our automation subcontractor. compressors are 2x HSK6451-50.

After the unloaded start at 30Hz and a delay for stabilization the machines are mechanically loaded to 100% with all subsequent capacity control being done by the drives.

One of the parameters call for a selection of the volts/Hz relationship as linear or parabolic - opinions as to which is appropriate?

V/Hz should be linear, that gives constant torque. Parabolic is used for sentrifugal fans and other equipment where the torque increases with the square of the Hz. This could be your fault. A parabolic V/Hz will cause the compressor to have hard to start and to overheat and draw excessive amps at Hz below nameplate value (60Hz).

Does the drive have any parameter for the torque "mode"? Like constant or variable?
.

[SteinarN]

What is this last factor in your power calculation?
I presume that 0,85 is cosØ

0,85 is motor efficiency, 0,8 is cosØ.

[garyb]

V/Hz is currently set as parabolic and I can't see any reference through the other parameters that specifically relates to torque control apart from the belt failure protection mode which obviously is not what we want.

Re ECMs comment on the drive suitability I note that of the 8 application examples in the manual 7 are for fans and 1 is for a pump - there may be some truth there. I recall the Danfoss VLT6000 was for fans and the VLT5000 for screws with a respective difference in cost.

[SteinarN]

V/Hz is currently set as parabolic and I can't see any reference through the other parameters that specifically relates to torque control apart from the belt failure protection mode which obviously is not what we want.

Re ECMs comment on the drive suitability I note that of the 8 application examples in the manual 7 are for fans and 1 is for a pump - there may be some truth there. I recall the Danfoss VLT6000 was for fans and the VLT5000 for screws with a respective difference in cost.

Seems that your automation subcontractor might have suplied you with the wrong drives. You might contact your national Siemens suplier and get advice from them about your drives suitability for constant torque operation. But changing to linear V/Hz should make a big difference anyway.

I have used the VLT 5000 series on my compressors. I'm very pleased with the performance. It is able to accelerate from standstill and up to 60Hz in 0,2 sec if needed. But I may have selected unnecessarry large drives also.

[Electrocoolman]

Hi Gary,

The SED2 is designed for HVAC applications....namely pumps and fans. It is suitable for Variable Torque loads.

You will need (I think) a MICROMASTER 440 which is designed for Constant Torque loads (and will accelerate from startup under load).

Sorry for the bad news....
Back to your supplier then?

ECM

[Electrocoolman]

Hi Gary,

Any updates?

[Magoo]

similar problem with danfoss vsd, solved by setting vsd as 480 vac max input and not changing max amps, the vsd sends higher volts at lower hertz on start-up, more grunt so to speak, elliminates high current/temp situation at start up. Won't effect over-all control and protection systems inbuilt in vsd because vsd cannot apply higher voltage at full vac because not available. The break away hertz is around about what you have to start with, no grunt/torque at low hertz, try and monitor start. Learnt this the hard way, at night and with a hissy client after a system change, For you free advise.

[SteinarN]

similar problem with danfoss vsd, solved by setting vsd as 480 vac max input and not changing max amps, the vsd sends higher volts at lower hertz on start-up, more grunt so to speak, elliminates high current/temp situation at start up. Won't effect over-all control and protection systems inbuilt in vsd because vsd cannot apply higher voltage at full vac because not available. The break away hertz is around about what you have to start with, no grunt/torque at low hertz, try and monitor start. Learnt this the hard way, at night and with a hissy client after a system change, For you free advise.

What series VFD was it? What was your parameter 101 and 110 settings? That 480 V you are mentioning, is this parameter 103, motorvoltage? What was the real motor voltage?

[Greengrocer]

Hi All

We are commissioning two Bitzer screws driven by Siemens VSDs on an air to water heat pump system for a swimming pool complex, +8/50oC R22, speed range 30-60Hz, lead/lag, unloaded start, VSDs control to suction pressure. We initially had some difficulty getting the screws past 15-20Hz but got over that by applying starting boost in conjunction with the overload service factor.

Now that the complex is approaching design conditions we are having problems with the lag machine "stalling" at around 35hz and sending the current high. The lead machine operates fine but always the lag machine is the one that surges the current at around this speed. It makes no difference which of the machines is in the lead position. If the machine are run together in parallel the problem does not occur and both will run up to full speed. I don't want to leave it in parallel as it greatly increases the minimum turn down capacity - obviously the best efficencey is at the highest suction pressure allowable for the duty. I have got around similar problems on Danfoss drives with the torque control function but the Siemens doesn't appear to function quite the same.

Either way I'm open to suggestions!

Regards
Gary

Hi Gary. A little "off piste" but how come you are commissioning an R22 system? Are these new drives being retro-fitted to old R22 systems?

[Magoo]

Hi SteinarN, I apologize for late response.
The danfoss settings are for motor peramitors setup, cannot recall actual addresses. I was applying 400vac 50 Hertz, the minimum hertz reccomended by Bitzer was 26Hz for suitable oil feed. I set the ramp time real short and min Hz at 26, plus told VSD that max VAC was 480, the vsd gave higher input VAC and system ran a treat, still is.
Hope this helps. Magoo

[garyb]

Hi All

Apologies for the delay in coming back on this one. As suggested by several posters The drives were not the best selection however by changing from parabolic to linear output made all the difference and they got away without any problems after that. Thanks All for the support.

Regarding the use of R22 - I'll admit its a questionable call but there are a couple of factors. One is cost - do we stand on principles and not get the job? Times are tough and we need every job we can get. R22 is cheap and will be available for some time yet. Another is efficiency - R22 gives great COP, is a "cheap" install compared to ammonia and is excellent for this application. To my knowledge it performs better than any comparable refrigerant - though I am open to being corrected. (This para should all go into a new thread if you want to continue).

[SteinarN]

Thanks for the update garyb.

[Magoo]

garyb
You are letting the team down by even concidering R22.
Drop into the real world, and don't offer R22 as an option.
If client insists, ask him - does he dislike his children that much, to save a dollar and curse them with cancer.
magoo