Some of the most perennially popular posts (say that fast 10 times!) here on the blog involve axial or centrifugal compressor surge control, or better stated—anti-surge control.
In a paper presented at this year’s AIChE Spring Meeting, Emerson’s Greg McMillan—a member of the Process Automation Hall of Fame—presented Compressor Surge Modeling and Control.
In his abstract, he highlights why this subject is so popular.
Axial and centrifugal compressor control is exceptionally challenging due to the extraordinary speed and severity of problems and the extreme consequences in terms of plant safety and performance. The fastest and most dangerous phenomenon is surge. An axial or centrifugal compressor in surge can reverse flow in 0.03 seconds going from a large positive flow to a large negative flow.
Greg defines compressor surge:
A surge point is the point on the characteristic curve where the slope goes to zero. The characteristic curve is a plot of pressure rise versus suction flow for a given speed or inlet guide vane position.
For axial compressors:
…the consequences of surge tend to be more disastrous because of the higher flows and the possibility of extreme acceleration. One very large compressor would hit a high speed shutdown in a fraction of a second potentially damaging the impeller that cost several million dollars causing production losses of many more millions of dollars a day.
Closed loop control in the PID [proportional-integral-derivative] controller monitors the suction flow to open surge valves. While this works for slower developing conditions, it is not suitable to address fast changes highlighted earlier in this post. Open loop backup is required:
…to preemptively position and hold the surge valves open… When triggered, the open loop backup puts the PID surge controller in remote output putting the control valves in a position large enough to prevent surge.
Greg explains how this works:
The key component for both recovery and prediction is a deadtime block to create a low noise reliable rate of change signal that can be updated as fast as the PID execution rate. For recovery from surge, a high rate of change triggers the backup that is the input minus the output of the deadtime block divided by the block deadtime that is typically just large enough to provide a good signal to noise ratio. For prevention of surge, a predicted flow one deadtime into the future that is too close to the surge curve triggers the open loop backup. The predicted flow is simply the input to a deadtime block plus the change that is the block input minus its output with the block deadtime set equal to the total loop deadtime.
These same types of control strategies can be used for recirculation valves around the axial or centrifugal compressor. Employing “Digital Twin” simulations can help test and refine these control strategies. Emerson’s Mimic simulation software v3.8 includes a compressor with surge object to model compressor surge and compute dynamic compressor response including hidden flow characteristics during surge.
As explained by Emerson’s Mart Berutti in an Emerson Exchange 365 post, Compressor Surge Modeling and Control in Mimic v3.8:
The knowledge gained from the model was critical to show the extreme jumps in flow caused by positive feedback on the hidden characteristic curve to the left of the surge point. The Mimic Compressor w/Surge object modeled this accurately and Mimic Future Value Block we used to detect the surge and identify the surge curve online.
Read the paper for much more depth on sources of automation dynamics, how to make valve response faster and common mistakes in selecting valves for this application.
Learn more in the Dynamic Simulation section on Emerson.com. Greg also has a newly updated 6th edition of the Process/Industrial Instruments and Controls Handbook which features contributions from 50 top technical experts. I won’t include myself in this esteemed list, but I did contribute a chapter on the Industrial Internet of Things and another on building your personal brand.