The path of the fluids in corrugated channels between circular plates is an arc of varying length, i.e. the flow direction firstly disperses and then converges. However, there is a shortage of data on the flow structures in these channels. In the present study, we experimentally investigate the flow patterns and pressure drop of two-phase flow in corrugated channels between circular platess. According to the observation, bubbly flow, slug flow, film flow and churn flow are sequentially defined. By analyzing the power spectral density of the flow patterns, the RBF neural network model is built for the flow pattern identification.