Modeling and analysis of control valves applied in compressor plants with consideration of time-variable factors

In process industries, compressor plants provide process fluids with certain flows and pressures. The plant behavior may change over the entire life cycle due to component-specific time-variable factors. Control valves applied in compressor plants are affected by time-variable factors. Their modeling with consideration of specific time-variable factors is described in this paper. Control valves ensure functional reliability of the plant, as they are opened to blow off excessive compressed process fluid in order to prevent surging and thus to protect the compressor from damage. The influences of the time-variable factors fouling, leakage, changing friction force and loss of supply energy are investigated and integrated into the valve model. Fouling of the valve affects the overall flow and control behavior. Further on, leakage can occur in all flooded components and may jeopardize the functional capability, as required pressure values cannot be reached. In order to avoid leakage at the actuator, stuffing boxes are used for sealing the piston rod. These boxes cause friction forces, which depend on tightening force, temperature and wear of the box. Changing of friction force leads to deviating travel paths or even stick-slip effects. Furthermore, the provided pneumatic energy to run the control valve might be variable. Energy sources with insufficient performance lead to significant changes in movement time, because the provided pressure collapses once the control valve is opened. It can be stated that all investigated factors affect the positioning behavior and lead to deviating movement times, travel paths or end positions. The developed modeling approach depicts all of these changes and the gained knowledge can be used for developing diagnosis methods. Several model parts, especially the modeled variable factors, are reusable for other components, like non-return and throttle valves or even the compressor itself. Hence, a diagnosis system for the whole compressor plant, based on the modeled influences, can be developed.