Startup of refurbished recip compressor - basic questions about recip operation 1

[ChEMatt]

So over the past two days we've commissioned and started up a refurbished recip compressor to recycle gas in our facility. It's a 75 hp two throw one stage taking a gas feed from 120 to 500 psig. It is oversized (with respect to capacity) for the service but was "left over" after another project so it was repurposed.

When we actually went to get it running, the VFD kept tripping on high amps. So we lowered the suction pressure from 120 to about 70 psig while keeping the discharge pressure constant (and the recycle valve was in manual at a fixed position) and the amps on the motor went down. I'm confused. I would have thought that as the compression ratio increased, the load on the motor would increase. But the opposite happened.

Would one of you kind folks be able to explain to me the relationship between power required on a recip and the inlet and outlet pressures. I have an idea in my head but it would be nice to have one of you smart folks confirm this relationship.

Thank you for your help!

-Matt

 

[BigInch]

Reducing the suction pressure made the fixed intake volume less dense. That was an apx. (70+15)/(120+15) = 63% reduction of the mass flowrate. Net result was less power required, even though the compression ratio was greater.

Independent events are seldomly independent.

 

[BigInch]

Whoops. Actually the compression ratio probably didn't change, unless you keep it constant. Discharge pressure probably dropped, because the volumetric compression ratio, based on volume drawn in to volume expelled, probably did not change.

Independent events are seldomly independent.

 

[JJPellin]

BigInch did a very good job of explaining why the motor amps went down. Mass flow rate is more important than compression ratio. I wanted to add a word of caution. Increasing the compression ratio can be dangerous on some double acting reciprocating compressors. As the compression ratio increases, the rod load increases. Depending on the particular design, it may be possible to fail a rod. Consult with the compressor manufacturer to determine if this is a potential problem with your compressor.

We have had rods fail on much larger machines. You do not want to see a broken rod. If rod load can exceed design limits, you should add appropriate alarms based on differential pressure.


Johnny Pellin

 

[ChEMatt]

Thanks for the feedback guys, stars for both of you.

BigInch, you were correct with your first answer. There's a pressure control valve on the suction as well as the discharge, so we kept the discharge pressure constant as we lowered the suction pressure.

Alarms were recommended by the vendor that refurbished the machine. I was not the process engineer who worked on the project so I can only hope that the setpoints came from a review of potential rod loading issues.

I've got another questions, but I need to look at something before I figure out how to phrase it exactly. I'd appreciate some additional feedback from you all if you have some time.

Thanks again.

-Matt

 

[zdas04]

BigInch's first answer was spot on.

Many compressor manufacturer's have programs to evaluate their frames. I use Ariel's Performance program a lot and it is hands down the best computer model I've ever seen. If the numbers on the ground are different from what Performance predicts, then there is a physical problem with the machine. Every time.

I often use Ariel's program or Cooper's Ajax program for other manufacturer's equipment and they usually get pretty close. What I'm trying to say is that if it is high speed compressor (even if its not Ariel) I'd take the data into Performance and pick a similar frame and similar cylinders and see if it approaches a rod load problem at your new conditions. The rod-load numbers in Performance are very conservative and I've never seen a rod part when Performance said it was fine. For low speed machines, I take them into Ajax, but I don't have as much faith in the results.

David Simpson, PE
MuleShoe Engineering

"Belief" is the acceptance of an hypotheses in the absence of data.
"Prejudice" is having an opinion not supported by the preponderance of the data.
"Knowledge" is only found through the accumulation and analysis of data.

 

[BigInch]

CheMatt, Your pressure control valves don't work exactly like that. It is important that you said PCV, not a PI connected to a PID controller, connected to compressor speed govenor. Step away from the process engineer and view from the mechanical engineer's perspective. For example, turning the dial on your discharge pressure control valve would not change the compressor's discharge pressure. It didn't change the cylinder's clearance and it did not change the driver's speed, nor the driver's power capacity. So, it did not change the compressor's volume-based compression ratio. The compressor takes at what is then 70 psig and compresses to (I calculate from performance figures at 120 psig suction, when the compression ratio was 515/135 = 3.81 and I keep that constant) 70 * 3.81 = 267 psig Neither does the discharge pressure control valve change the power capacity of the driver. The discharge pressure control valve can only regulate the pressure in the discharge pipe, if the pipe pressure is on the correct side of the valve's control set point. If the discharge pressure is on the opposite side of the set point, the valve is either full open or partially (maybe fully) closed, or v/v.

If the compressor's discharge piping's pressure was <= 267 psig, the compressor could discharge into the pipe and the discharge pipe's pressure would tend to increase, if evacuation of that pipe continued normally (ie. there were no control valves trying to regulate it to some other pressure). If there was a valve trying to regulate the discharge piping's pressure to a lower value, the discharge pressure upstream of that valve would increase.

If the discharge piping's pressure was > 267 psig, say 515 psig (assume it did remain constant), the compressor would discharge nothing, and actually the pipe would supply gas to the cylinder through the compressor's discharge valve and "top off" the cylinder in an attempt to reverse flow. The compressor discharge valve closes, now containing 515 psig. Then the cylinder would continue its cycle to the "draw inlet supply" position. At that position, the pressure in the cylinder would be 315/3.81 = 120 psig. 120 psig is > 70 psig in the suction line, so the cylinder gas is expelled into the suction. REVERSE FLOW.

So what your control valves do is attempt to regulate the pressure already in the piping, not what's going on in the compressor. The suction pressure control valve tries to regulate suction pressure, and it can, provided that the suction pressure is on the correct side of its set point. Let's say you want to keep AT LEAST 100 psig in the suction piping. The CV is upstream of the conpressor suction flange. If Pressure moves to 60, the valve opens allowing the suction pressure to increase, but only if the suction piping system can still supply enough gas to increase the pressure. If the increased pressure puts too much resistance on that supply of gas, no pressure incease occurs in the suction piping, even if the CV moves to full open, which it will. On the discharge side, if the discharge PCV's set point is 515 psig, but the compressor supplies only 267 psig, that valve closes as it attempts to increase the pressure in the discharge piping between it and the compressor. But, we've already seen that the compressor cannot reach that pressure, without changing the clearance pocket settings, making the compressor volume smaller (thereby increasing discharge pressure range). It can only reach 267 psig. Now, as the prevous paragraph explained, gas trapped in the discharge piping at 267 (remember, the discharge valve closed) backflows to suction and is reduced to 125 psig as the compressor returns to inlet draw position and the gas flows out into the suction piping.

Independent events are seldomly independent.

 

[BigInch]

Note the increase in discharge pressure accomplished by reducing head clarance, decreases compressor flow rate, given holding the same RPM.

And BTW, The rod load JJPellin mentioned is roughly discharge pressure x cylinder cross-sectional area, and often (very generally) falls into the range of 30,000 to 40,000 lbs. Obviuosly you need to check the exact value.

Independent events are seldomly independent.

 

[JJPellin]

The OP stated that the compressor discharge pressure was held constant. I will take this as true. All of our compressors would behave that way. Actually, the rod load would be related to differential pressure times piston area, not discharge pressure.

Johnny Pellin

 

[BigInch]

Maybe he thought it was held constant, as in he didn't change the set point. I guess in that case he probably did bend the rod.
Yes, rod load is differential pressure; my mistake. Thank you.

Independent events are seldomly independent.

 

[JJPellin]

BigInch, I am not familiar with the sort of process compressor you describe. In all of reciprocating process compressors (approximately 30 machines) your analysis would not be valid.

1. Back flow is impossible except in the case of discharge valve failure.
2. Compression ratio is not constant. A typical piston stroke might be 10 inches with a minimum head clearance of 1/8 inch.
3. Discharge pressure is controlled by the downstream process and is constant.
4. Compression ratio is dictated by differential pressure not by the volumetric ratio of the piston stroke within the cylinder.

I am away from the office or I might be able to recommend a suitable reference book.

Johnny Pellin

 

[petrochemical87]

This is how I would analyze this issue if it were my plant. For any positive displacement compressor power=adiabatic head*mass flow rate. Now these machines are capable of producing variable adiabatic heads (within certain limits ofcourse) and when the suction pressure is reduced keeping the discharge pressure constant the adiabatic head increases due to increase in the compression ratio but that is offset by the reduction in mass flow rate(reduction in suction pressure). So the net effect is a reduction in power. Punch in the numbers in the power equation and you shall see.

 

[JJPellin]

BigInch, I should have more specifically asked what type of compressors you are referring to. I am referring to our large double acting reciprocating compressors. All of my points are true for our compressors.

1. I have never seen a compressor of this type that did not have check valves at the suction and discharge. These are internal, built integral to the cylinder. I don't know how any recip compressor could function without them. R
2. Our compressors use fixed pocket uunloaders, variable pocket unloaders, discharge valve unloaders, partial valve unloaders and plug unloaders. I am very familiar with them. I am referring to a loaded machine. Refer to typical P-V diagrams for recip process compressors.
3. Our compressors all discharge into systems that are pressure controlled. The compressor does not control the pressure.
4. Refer to #1 above. The pressure in the cylinder increases until the cylinder pressure is greater than discharge pressure. Then the discharge valves open and the rest of the stroke is used to push the gas out of the cylinder. (P-V diagram again)

As usual, you must be referring to some pipeline equipment I have never heard of. The OP asked about a compressor in a process plant.

Johnny Pellin