Thiết lập mô hình mạng Nơ-ron nhân tạo (ANN) tính toán độ sâu sau nước nhảy trong kênh lăng trụ mặt cắt chữ nhật
Abstract
The hydraulic jump is the result of an abrupt reduction in flow velocity, converting a high-velocity supercritical flow into a low-velocity subcritical flow. The sequent depth of the hydraulic jump is an important characteristic that needs to be determined to calculate the vortex length of the hydraulic jump and the size of the stilling basin or water canal. Neglecting the frictional force, the ratio of conjugate depths of the hydraulic jump can be determined according to the Belanger formula for a horizontal rectangular prismatic channel. However, when the friction force is present in practice, the sequent depth will have a smaller value than that calculated by the Belanger formula. Therefore, this paper presents the development of an artificial neural network (ANN) model to calculate the conjugate depth ratio of the hydraulic jump. This model considers the surface roughness of the channel and the fluid viscosity. The proposed ANN model has very high accuracy calculation results; the coefficient R2 is approximately 1 in the validating and testing phases. The application scope of this model is quite large, so it can be applied in practice to calculate the sequent depth of the jump in a horizontal prismatic channel with a rectangular cross-section.
