Detection of stiction in flow control loops based on fuzzy clustering

• Flow control loops form significant control loops in industry.
• The applications of shape-based techniques of stiction detection are limited.
• The proposed technique does not rely on the shape of the stiction pattern.

In the presence of stiction, the control valves may present an oscillatory behaviour that affects the regulatory control performance, thereby causing a loss of product quality and increasing energy consumption. Detection of stiction in the early phase is a crucial key for process control to avoid major disruptions to the plant operations. In this paper, a novel technique based on a well-developed fuzzy clustering approach is proposed. Based on a dramatic change of the slope of the lines obtained from successive cluster centres in the presence of stiction, a new performance index to distinguish the cause of oscillation is proposed. The simulation, experimental and industrial results are provided.