logo


Feb-2014

Enhanced control techniques for cryogenic turboexpander recompressors

Turboexpander compression trains are commonly used in petrochemical processes for extracting the heavier hydrocarbons from natural gas.

Wayne Jacobson
Compressor Controls Corporation

Viewed : 7228


Article Summary

These heavier hydrocarbon gases include but are not limited to ethane, propane, butane and pentane and are commonly called natural gas liquids or NGL. Figure 1 shows a typical gas processing plant design. In this example, the expander is used to cool the process gas stream before entering a distillation column or deethaniser where it is used as reflux for the distillation. Gas from the top of the column is routed through a cold box before being compressed by the compressor portion of the turboexpander compression train.

The turboexpander and recompressor parts of the unit operate on the same shaft. A J-T (Joule-Thompson) valve is used to bypass flow around the turboexpander when the expander is offline or is unable to meet the flow demands of the process. The J-T valve is also used for startup of the train. The recompressor uses a recycle valve for antisurge protection as well as startup and shutdown of the unit. There is a bypass line around the recompressor to divert flow when the turboexpander train is offline and the process is operating in J-T mode. See Figure 2.

Production is controlled by manipulating the expander guide vanes. Typically this is done by maintaining a constant pressure at the distillation column, however, other parts of the process such as the inlet separator pressure or recompressor discharge pressure may be controlled as well.

Conventional Control Methods for Turboexpander Recompressors
Traditionally controls for the turboexpander trains are supplied by the OEM in a package together with the machine. There are several variations of these conventional control schemes used but most of these controls are based on simplistic assumptions and use simplified control solutions. In general, the conventional control schemes for turboexpander trains are based on a split range control between the expander guide vanes and the J-T valve with an addition of a low signal selector to allow for independent closure of the expander guide vanes. The control of speed is typically not performed continuously, but maximum speed is sometimes limited, and as a proxy for speed control sometimes the power dissipated in the expander is limited by controlling maximum acceptable pressure differential across the expander. While speed of the expander is limited by the guide vanes, the J-T valve is opened to meet process demands. The centrifugal compressor of the unit is equipped with a recycle valve for antisurge protection, although frequently the cooler enabling effective and prolonged operation of the antisurge protection is omitted. The control scheme for the antisurge protection of the compressor is often a simplified scheme sometimes even without the use of inlet flow measurement in which case antisurge protection is based on preventing high pressure differential across the compressor.

This control scheme for the Turboexpander-recompressor unit is simple in use but it presents a few problems. First, the scheme does not provide for a seamless operation of the expander when the speed needs to be limited. Between the moment the guide vanes begin to close to protect the expander and the moment the J-T valve begins to open to compensate for the flow that has to be rerouted, a gap exists due to split level solution in the output signal. The J-T valve does not begin to open until the signal of the pressure controller reaches the position corresponding to the beginning of opening of the J-T valve. In the meantime the controlled variable previously controlled by opening of the guide vanes shifts out of control and only later is being caught up by opening of the J-T valve. The system has a similar defect when going in the opposite direction from the J-T valve closing and eventually switching back to the guide vanes. As a result, every time the operation of the J-T valve is used or terminated, the control variable suffers an upset. The upset results in lower quality of operation of the NGL separation.

Second, the scheme does not assure smooth operation when the turboexpander-recompressor unit shuts down. In this situation the J-T valve must open rapidly to compensate for the missing flow. The split level control scheme does not allow for the J-T valve to rapidly travel to a position in which flow would be reasonably close to that of the expander prior to shut down. As a result, the plant suffers a large upset in pressure/temperature profile that definitely reduces quality and may lead to plant shut down. Third, the antisurge control of recompressor is often primitive and may not be able to protect the recompressor adequately especially when gas composition is changing.

Increasing Production of NGL
Although turboexpander trains require an overspeed protection system, most do not benefit as much from continuous speed control. Speed of the train is determined by the difference in torque developed by the turboexpander and the recompressor. The greater the torque delivered by the expander compared to that of the recompressor, the greater the rotating speed. Conventional control systems limit speed by limiting the opening of the expander guide vanes which limits the torque delivered by the expander. The torque delivered by the compressor remains relatively constant.

 For most compressors, increasing the flow rate leads to increasing the power needed to drive the compressor. Since opening the recycle valve increases the flow rate through the compressor, the recycle valve can be used to vary the rotational speed of the turboexpander train. Therefore, the speed of the train can be first limited by opening the recompressor’s recycle valve. This will result in an increase in torque delivered by the compressor while the torque from the expander remains the same and in turn will slow down the turboexpander train. By using the recompressor recycle valve first, the expander guide vanes are allowed to remain in a more open position and this results in an increase in condensate production.

The increased condensate is due to the expansion of gas in the expander in an isentropic work producing process whereas the expansion across the J-T valve is an isenthalpic process. This means that both temperature and pressure will be lower after expansion through the expander than it would be if expanded across a J-T valve. Lower temperature and pressure allows for increased production in condensates. Figure 3 shows the difference between the expansions of gas through the expander versus expansion across the J-T valve. The isentropic work producing expansion through the expander occurs from point 1 to point 2. The adiabatic expansion across the J-T valve that occurs at constant enthalpy is the nearly vertical path from point 1 to point 3.

Separate control loops for guide vane and J-T valve
Split range control between the J-T valve and the expander guide vanes creates several challenges.  First, the turboexpander guide vane actuation and J-T valve behave differently which requires making compromises in tuning in order to accommodate both. Second, provisions are necessary to allow for the J-T valve to take over control of the process when the guide vane control is being overridden due to either a limiting action or manual override.

Having separate control loops for the expander guide vane and the J-T valve as shown in Figure 4 allow for better control because each loop will have dedicated tuning which is more straight forward to implement than split range control. Often times the expander guide vanes are difficult to control and can lead to process swings or even expander trip on overspeed. Using a separate control loop for the J-T valve allows for it to assist the expander control during significant and rapid disturbances by opening the J-T valve for short periods of time to help stabilise the process.


Add your rating:

Current Rating: 4


Your rate: