Recently we had a big problem with a large compressor (27MW)start up using VFD. The issue is that the vendor sized VFD capacity is too small to start the centrifugal compressor from settle-out pressure. Several options were proposed including:
 1. Change to a larger VFD, which means a four-million dollars to be added and also a long time for delivery. Since the VFD is customized, the current VFD could not be used unless in the future there is a similar motor+compressor could match.
 2. Blowdown the process gas from settle-out to lower pressure so that the current undersized VFD could start up. It is estimated a two to four thousands of dollars to be lost on every start up since the vented the gas (refrigerant)is very expensive.
 3. Still use the current undersized VFD, the vendor will pay the cost of each start up for 10 years, which is roughly about 2-4 million dollars.
 4. Recovery the vented process gas from option 2 by using a recip compressor and a drum to the refrigeration loop. This is another two million dollar cost since of the recip compressor and drum and additional piping.
 5. Revise the current piping layout and adding drums etc to the compressor inner-connection piping volume so that the settle-out pressure could be reduced. This is not quite sure feasible, and still costly and time consuming since some part the piping has been done on site.
  
  Each option mentioned above looks ridiculous from the user's standpoint.

  I am thing a option 6. I don't know if it feasible.  

(Note the compressor is two-stage centrifugal compressor with inter-stage and post-stage air coolers )

My idea is:
  
  Add a control valve between the compressor 1st stage and 2nd stage. In this case, when the compressor is stopped, there will be two settle-out pressures before and after the control valve. The settle-out pressure in the first stage is lower and in the second stage is larger. I am guessing this may reduce the overall compressor start-up torque, and hence require lower VFD capacity. The problem is how I can estimate the start-up torque and power consumption in this scenario. Is there any problem if I add a control(block) valve between the 1st stage and 2nd stage?
 
  Does anyone have any ideas or insight?

MORE NEWS