Checklist for Compressor Surge Control Tips
Surge control valves are often quite large. Boosters are required to reduce the pre-stroke deadtime to less than 0.2 seconds and the stroking time to less than 1 second. A booster should not be used instead of a positioner because the booster high outlet port sensitivity causes positive feedback that has been shown to cause diaphragm actuated butterfly valves to slam shut. The booster should be on the output of the positioner as shown in slides 91 and 92 in the ISA Edmonton short course "Effective Use of Measurements Valves and PID Controllers". The booster bypass valve must be slightly opened so the positioner sees some of the volume of the actuator to prevent fast cycling of the valve.
The flow measurement must not introduce more than a 0.2 second delay or attenuate the oscillations or precipitous drop in flow to the degree the oscillations are not recognizable. Slide 33 shows the effect of various transmitter damping settings. Note that the 1.7 sec default damping setting of some differential pressure transmitters will seriously delay the correction and hide the recognition of surge.
The PID controller execution time should be 0.1 seconds or less. A deadtime block should be used to compute the rate of change in flow. If a precipitous drop in flow or rapid oscillations is detected, an open loop backup as exemplified in slide 118 should be triggered to pop open the surge valve enough to recover from surge. The valve should be held open long enough for the system to stabilize and then released for a bumpless transition to PID feedback control. The PID output tracking function is typically used.
To provide fast opening and slow closing, external-reset feedback should be enabled and a setpoint rate limit (velocity limit) added to the analog output (AO) block in the direction of closing the valve. The AO setpoint should be used for external-reset feedback so the PID output does not change faster than permitted by the AO in the direction of closing the surge valve.
The PID controller should open the surge valve(s) well before the compressor flow drops below the minimum flow to prevent surge, which is point of zero slope on the compressor characteristic curve. If integral action is not suppressed, the surge valve(s) will not open until the suction flow has crossed the PID setpoint. Since the full response is not seen until the compressor flow drops below the setpoint and the loop response is not instantaneous, surge will occur if the PID setpoint is too close to the minimum flow. An adaptation of the reset setting per Tip #92 in the new ISA book 101 Tips for a Successful Automation Career based on the rate of approach to surge can help you get the surge valve(s) open at the right time by making integral action smarter. The setting of the anti-reset windup (ARW) limits relative to the output limits will also affect the point at which the surge valve(s) open. The ARW algorithm varies with DCS software supplier and version number.
Redundant parallel surge valves are often used. If the compressor feeds many downstream users, there may be frequent period of time where the compressor must smoothly operate with a partially open surge valve. Here a smaller surge valve is added in parallel to provide more precise control. The small valve should be able to handle the shutdown of at least one downstream user. An enhanced PID for valve position control (VPC) can optimize the position of the large valve to keep the small valve in a good throttle position suppressing unnecessary movement, interaction, and limit cycles in the large valve. External reset feedback and directional setpoint rate limits are used in the VPC to provide fast opening and slow closing. Feedforward action should be added to the VPC to handle large fast disturbances.
If the operating point reaches the zero slope point the flow jumps in 0.05 seconds to a negative flow characteristic curve. After the reversal in flow, the operating point walks along the negative flow characteristic curve until it reaches a zero slop at which point it jumps in 0.05 seconds to a large positive flow. The surge cycle repeats until the surge valve stays open long enough to stabilize the operating point to the right of the surge curve. Compressor maps typically do not show the negative flow characteristic curve of the compressor. In some cases the compressor may stall with the oscillations attenuating and flow converging on an operating point near zero flow.
The compressor map shows the characteristic curves for various speeds plotted with a Y axis of pressure rise across the compressor and X axis of suction volumetric flow. A discharge flow measurement must pressure and temperature compensated back to volumetric suction flow. The surge curve will change with suction temperature, pressure, molecular weight, and specific heat ratio and with wear and tear.
The closing of downstream feed valves should be slowed down as much as possible. However, the benefit may not be as much as expected because of the flatness of installed characteristic of rotary valves for large openings. If the user valves cannot be slowed down enough, add a feedforward signal based on installed characteristics of the user and feed valves to preemptively position the surge valve.
For a deeper understanding of the dynamics, modeling, and control of compressors, my book Centrifugal and Axial Compressor Control has been put back into print by Momentum Press.
•(1) Are boosters on the outlet port(s) of positioner needed for a pre-stroke deadtime of less than 0.2 seconds and a stroking time of less than 1 second to prevent surge?
•(2) Is the booster bypass valve opened just enough to prevent valve oscillations?
•(3) Is the transmitter damping setting 0.2 seconds or less?
•(4) Is the PID execution time 0.1 seconds or less?
•(5) Is the Analog Input (AI) and PID block signal filter time zero?
•(6) Is a deadtime block used to create a flow rate of change and rate of approach signal?
•(7) Is the surge curve corrected for suction temperature, pressure, molecular weight, and specific heat ratio?
•(8) Do the anti-reset windup limits and output limits enable the surge valve to open soon enough?
•(9) Is smart reset action needed based on rate of approach to get the surge valve sufficiently open before the operating point reaches the surge point?
•(10)Is an open loop backup triggered from oscillations or precipitous drop in flow used to ensure rapid sustained recovery from surge?
•(11) Are external-reset feedback and an Analog Output (AO) block setpoint rate limit in the closing direction used to provide a slow closing and fast opening surge valve?
•(12) Is a feedforward signal needed to preemptively open the surge valve?