Surge Control Considerations in Centrifugal Compressors

Pete Sharpe Director, Global Industry Solutions, Refining & Chemical

Pete Sharpe
Director, Global Industry Solutions, Refining & Chemical

Any process manufacturer that has centrifugal compressors knows about a condition known as surge. Emerson’s Pete Sharpe described it to me:

Surge is a condition that occurs on compressors when the amount of gas they are trying to compress is insufficient for the speed of the compressor and the turbine blades lose their forward thrust, causing a reverse movement in the shaft. This condition can have catastrophic effects on the machine, so compressor manufacturers include anti-surge valves that recycle gas from the discharge to the suction when a low flow is detected. Usually these valves are designed to be only open on startup or under reduced rates.

This surge condition is analogous to an aircraft “stall” condition, or pump cavitation.

Compressor Antisurge Control Schematic
Given the possible harm to the equipment, specialized anti-surge controllers exist that have very high-speed execution and special surge detection algorithms. It also requires fast acting antisurge valves and differential pressure transmitters. The required control response is a function of the response time of the surge valve, measurement devices and control algorithm relative to the surge dynamics of the compressor. In an earlier surge control post, ModelingAndControl.com’s Greg McMillan noted that surge valves typically have response times from 1 to 5 seconds.

The system speed-of-response becomes more critical the closer you operate to the surge limit. Since recycling gas from the discharge back to the suction of a compressor wastes energy, the most economic operation is to keep the recycle flows as low as possible, thus operating close to the surge line.

This analog may not be a perfect fit, but it reminds me of the Grand Canyon. The closer you are to the edge the better the view down, but it is also increases your risk.

Pete was recently asked about his thoughts on doing anti-surge compressor controls within an automation system, in this case, the DeltaV system. His first point is that surge control is only needed when the gas flow is low, so most compressors are designed to have their surge valves closed for the majority of the time.

Centrifugal Compressor Surge Curve
If you are trying to operate right on the surge line, Pete agrees that these specialized, dedicated anti-surge controllers may be necessary. However, if you choose to set the surge flow controller setpoints more conservatively to include a safety margin, anti-surge control functions can be safely implemented within the DeltaV system, or in a Foundation fieldbus-based arrangement within the fieldbus devices.

These more conservative settings mean that the compressor would go into recycle or blow-off condition earlier than is absolutely necessary. Energy is wasted only when the recycle valve opens. So, to evaluate whether an expensive, dedicated surge control system is required, plant engineers need to answer the following questions:

  • How often is the surge valve open?
  • Under what conditions does it need to open?
  • How quickly do changes in gas rates, composition, and pressure occur?
  • How much energy is wasted by having surge controllers set 5% higher than the surge point?

If the answers to these questions provide for this more conservative operation, Pete notes that he has worked with process manufacturers to execute 1-second anti-surge control within a DeltaV controller. These strategies can run at up to 100msec, which is faster than the valve response time. In the ModelingAndControl post referenced earlier, Greg suggested a control strategy.

A recent DeltaV whitepaper, Fast Control Response Requirement describes 300msec control by having the control algorithms run in the Foundation fieldbus devices–a practice known as control in the field. In this case, the digital valve controller on the surge control valve would run the analog output block and the PID control block.

The advances in technology allow the operating target to safely move closer to the surge curve. The advantages of incorporating this control into your automation system is an integrated operator interface, advanced measurement diagnostics, alarming, and reporting information on the compressor. All good stuff to keep your process operating effectively.

GreenPodcast.gif Surge-Control-Considerations-in-Centrifugal-Compressors.mp3 | iTunes

Audio clip: Adobe Flash Player (version 9 or above) is required to play this audio clip. Download the latest version here. You also need to have JavaScript enabled in your browser.

16 comments

  1. Shijinp Puthalath says:

    thanks..this is good experiences for me
    can you give me,which is the detecting options are compressor surge?

    • Shijinp, Thanks for your question! Here’s what I found out,

       
      “Yes, surge is easily detected from the pressures on the outlet of each compressor stage.  Problem is, it occurs very quickly and can damage the compressor, so you really don’t want it to happen.  The Cadillac surge control system with the largest market share is Compressor Control Corp (CCC) which uses dedicated equipment with very high-speed execution to detect insipient surge and slam the recycle valves open.  But they also use a surge curve rather than the surge detection to adjust the valves under normal operating conditions.  In a DCS system, we would also program the surge curve, but would then add more contingency (than CCC) to deal with the worse expected upset.  Depending on the service of the compressor, a CCC system would make sense (e.g. ethylene cracked gas compressors).  For others, like refrigeration compressors, that don’t see significant changes in composition or system pressure swings, a more simple surge control is justified. 
       
      Also, if the compressor system is designed right, the surge valves should be closed under normal operation.  They are only used when the demand is lower than design, like start up.  So if the surge valves are open a bit more than they really need to be during those periods, it’s usually not that big of a cost.”

       

    • Shijinp, I’m not sure why I didn’t see the notification on your comment… sorry for the long delay! The basic measurement is differential pressure across the compressor. There is a good write up at:
      http://www.mechanicalengineering4u.com/technical-stuff/127

  2. N.A SHAIKH says:

    have you any presentation on surge system for centrifugal compressors can i have it??? it would be great kind honor

  3. dear jim

    i have a question? the PV equation of woodward A/S contoler have firstly, the flow (orifice flow) from second to third stage . the transferencial function of this flow transmiter is cuadratic or lineal? for the analog output.?

    very thanks for your time.

  4. iman saleh says:

    Dear Sir,
    Is it possible to detect compressor surge by measuring motor current?

    • Hi Iman, Thanks for your question. I’m checking around with some of our consultants to see what thoughts they have.

      • Iman, I did hear back. It’s gonna depend on how fast the current condition is sensed and action taken. Here’s the note I received, “Theoretically, the answer is yes. As a compressor approaches the surge condition, the compressor begins to unload. This will cause a corresponding decrease in electric motor current. Whether the instrumentation monitoring the motor current is fast enough to indicate a surge condition depends on the monitoring devices.”

        • I did get one additional response looking at the problem from a pressure diagnostic perspective, “For an Ethylene cracker compressor, a 3051S pressure transmitter with advanced diagnostics was installed to help prevent unreliable pressure measurements from giving false lows that instigate (unnecessary) surge control. Surge controls recirculate the flow to protect the compressor, but cause a drop in downstream flow, introducing slowdowns as well as variation in the process.”

          • Joseito, it has been a while since you posted your question but unfortunately I got to read this interesting discussion just now. If you still need help on Woodward surge control system I would be happy to to talk to you. Please send be an email at alex.benim@woodward.com with the phone number and best day/time to reach you.

  5. chris wehr says:

    Would any one have a CCC control system description you could send me chris_wehr@Hotmail.com

  6. Suresh Mangalore says:

    Our sister concern has 2 nos. Elliott 330DA3 air compressors. The pressure ans capacity control are monitored by an inlet guide vane controlled valve and a Fisher (Emerson) blow off valve. These 2 units are
    controlled by means of 2 nos. Fairchild Transducers (0 to 5 VDC).The Engineer serviced the Fisher control thru’ the Emerson Facility in Jakarta.
    The inlet guide vanes control was repaired in a good workshop.
    The Fairchild transmitters were replaced with “Elliott” transducers.
    However, the system does not operate at all.
    Would it be possible to replace the “Elliott” transducers with JUMO pressure sensors and control the Fisher valve and the Inlet guide vane valve by adding 2 PID digital Jumo controllers (0 to 5 VDC)?
    Would this modification affect the surge control system ?
    Awaiting your valued reply.
    Kind Regards

  7. Jim Cahill says:

    Hi Suresh, Thank you for your comment and question. I’m checking with one of our experts in your region and will have them connect back with you through the email you provided.

Leave a Reply