Influence mechanisms and control of complex clearance flow on hydraulic instability in a pump turbine using optimization

Pumped storage is vital for low-carbon energy systems. The crown labyrinth seal of a pump–turbine affects leakage, efficiency, and stability. Traditional seal designs rely on empirical rules and cannot balance multiple objectives. The physical role of clearance flow in hydraulic instability remains unclear. This study optimizes the crown seal of a 300–MW pump–turbine using multi–objective optimization. It combines computational fluid dynamics, a genetic–algorithm surrogate model, and entropy production analysis. The study shows that optimized seal reduces leakage by 17.10% and axial thrust by 7.20%. The S–shaped curve shifts only slightly. In contrast, the hump characteristic improves markedly: the hump margin rises from 2.66% to 3.21%. The study reveals the underlying flow physics. Reducing the clearance inlet size and labyrinth throttling lowers the jet momentum. The weaker jet enters the vaneless space with less lateral shear. This reduces circumferential velocity distortion, delays guide–vane separation, and alleviates blockage. The loss reduction is a spatial redistribution, not a uniform drop. The seal acts as a primary control for the hump but only a secondary modulator for the S–shape. A general design principle emerges: suppressing clearance jet momentum reduces vaneless–space shear, delays separation, and enhances stability. This principle is transferable to other high–head reversible machines.