In the event of an Emergency Shut Down (ESD) initiation, the fuel supply to the gas turbine, or the electric power to the electric motor are instantaneously shut off to protect the equipment, and the compressor station. Upon fuel shutoff, the inertia of the rotor system (including the compressor rotor, the power turbine rotor or the rotor of the motor, the inertia of the coupling, and, if present, of a gearbox) keeps the compressor running. Since the compressor still compresses gas, it has to provide the power necessary. Therefore, the rotor will slow down.
On the other hand, the check valve on the discharge side will close automatically, so there is a trapped volume of gas in the piping between the compressor discharge, the check valve, and the recycle valve. To understand the dynamic behavior of compressors it is important to understand that the operating point of the compressor is determined by the head imposed by the suction and discharge pressure of the system. The compressor, at a given running speed, reacts to the imposed head with a certain amount of flow. The compressor consumes power based on said head and flow (and its efficiency at the operating point).
The pressure at the discharge side depends on the amount of gas trapped in the fixed volume. The compressor adds gas to this volume, while the recycle valve, depending on its instantaneous flow capacity, will reduce the amount of gas in the volume. There is a chance for surge to be initiated.
Because the avoidance of surge is important for the safety and reliability of a compressor station, the dynamic behavior of the system during emergency shutdowns is often studied in the planning and design phase of the compressor station. This allows to verify in particular whether the valve size, opening speed and the arrangement of a single or multiple valves, is appropriate.
In this phase of the project, it may still be possible to modify the piping layout. One of the key uncertainties in these simulations is the speed decay of the compressor. To some extent, this is due to the fact that de-energizing the driver does not lead to an instant loss of power to the compressor. In the case of a gas turbine driver, there is a significant amount of power supplied for about 200 to 300 ms after the fuel valve is closed. This is due to the thermal mass of the driver.
More importantly, even the estimate of the compressor speed decay in the absence of driver power is often a problem. The frequently used approach is to use a fixed decay rate for the compressor. Obviously, the faster the speed decay, the more difficult is the task for the recycle valve to keep the compressor from surging. A decay rate of 30% speed loss per second is often used in simulations. This rate has proven to be rather conservative.
While this is generally comforting, it may also force systems to be more complex and expensive than necessary. Another issue that also should be considered comes from the assumption of the compressor operating point at the moment the initiation of the shutdown. If that point is close to the surge line, the system has to react very fast to avoid surge. It is thus important to select the starting points for the study so that they so that they reflect realistic, actually achievable operating conditions.
The governing equations consider the following:
Mass conservation in the piping system
Torque balance of the rotor system
Friction losses in the piping -wave propagation in the pipe
Flow through the valve (based on the instantaneous valve flow area, and the pressure ratio over the valve)
The equations indicate that at a given speed the speed decay rate will be slower for a large compressor consuming a small amount of power, than for a small compressor consuming a high amount of toque. For applications , especially for compressors directly driven by gas turbine, the dominant inertia comes from the power turbine. For a typical 15,000 hp gas turbine driving a 2 stage pipeline compressor, the power turbine inertia may be 10 to 15 times higher than the compressor inertia.