A Symmetric breaking hysteresis in a plane wake

We have observed a new phenomenon called symmetric breaking hysteresis. The flow is a confined plane wake in a pipe, i.e. the inlet of the pipe is a plane wake with adjustable initial mean flow velocity U$_{1}$ and U$_{2}$ respectively on each side of the splitter plate. When the forcing level is sufficiently high at low frequency, the wake becomes asymmetric to the trailing edge. Randomly one side has much higher turbulence than the other. When U$_{1}$ is at the side with lower turbulence and is gradually increased up to a critical value (thus the flow is no more a wake, but mixing layer), there is a sudden switch of the asymmetric flow: the side with lower turbulence and higher mean velocity now has higher turbulence than the other side (i.e. U$_{2})$. If U$_{1}$ is now reduced to the same value as U$_{2}$, no change in the flow can be observed. However, when U$_{1}$ is reduced to a critical value lower than U$_{2}$, the asymmetric flow suddenly switches again. In the whole process, no symmetric wake can be observed. This phenomenon of symmetric breaking hysteresis becomes weaker with the increase of the Reynolds number of the wake. Quantitative scalar measurement confirmed the visualization.