Update: Greg has published a post, Checklist for Compressor Surge Control Tips which should give you some things to consider for your centrifugal and axial compressors.
Orginal post: It’s great to questions from automation professionals from some of my posts built up over the years. A question came about one of my centrifugal compressor surge posts. I removed the proprietary elements of the question.
I turned to Greg McMillan, who wrote a book many years ago, Centrifugal and Axial Compressor Control, which is now available as a free eBook. Greg responded to the automation professional:
For air separation we use compressor and distillation column.
We have aaa brand centrifugal pressure whose capacity is suck the atmospheric air and build the pressure of bbb of the flow ccc Nm3/Hr .The compressor is consist of total 5 stages on single shaft.
Now we have used ddd anti surge controller which is supplied by the aaa. In which input is only outlet pressure of Compressor. As per my knowledge surge can be detected by Inlet pressure, Outlet pressure (By which calculate DP) and flow.
So I want to know that how they can detect the surge by only knowing the outlet pressure of compressor?
It will be of great help if you reply to me, as that will be root cause of some of our problem.
With just an outlet pressure measurement, you can detect a severe surge from the rapid oscillations in pressure. However, this is too late in that the compressor is in surge and the flow reversals and high axial thrust and radial vibration has already started possibly damaging the compressor and upsetting downstream operation. In order to prevent surge, you need a suction flow measurement besides a differential head pressure measurement to know the operating point on the compressor characteristic curve and the proximity to the surge curve.
You want to take preventative action that opens the surge valve(s) sufficiently before the operating point crosses the surge curve. If you have a multistage compressor, individual stages may be go into surge. In this case, you need differential pressure measurements across each stage. If you are not worried about detecting individual stage surge and the suction pressure is negligible compared to the discharge pressure, a discharge pressure measurement and suction flow measurement would suffice to locate the operating point on the compressor curve for surge prevention.
If you have a discharge instead of a suction flow measurement, you need to pressure and temperature compensate the flow reading to get back to suction flow that is the X-axis of the compressor map. The Y-axis is differential pressure.
Once a compressor goes into surge, a feedback surge controller usually gets too confused by the severe fast oscillations to get the compressor out of surge. An open loop back up that pops open and holds open the surge valve(s) is triggered by a precipitous drop in flow or severe oscillations. The valve(s) are held open and don’t go back on feedback control until the system has recovered and is stable at an operating point far to the right of the surge curve.
Greg noted that he will have a checklist post over at his Control Talk blog that you’ll want to check out, if you would like more ideas on managing compressor surge conditions. You can also follow his series of tips in the Emerson Exchange 365 community.
Also, if you’ll be joining us next week in Anaheim, here’s how to catch up with Greg. He noted:
I am teaching 2 short courses twice (“Adventures in pH Control ” with Dave Joseph and “”Effective Use of PID Features for Loop Perfromance and Optimization” with Hector Torres) and am part of a “Ask the Experts Panel on PID” on Friday at Emerson Exchange. I will be asking end users in my sessions to see me after the session if they want to be part of the ISA Mentor program. I will be giving out copies of my new book 101 Tips for a Successful Automation Career for those end users who want to be part of the program.