Prediction of Gas-Liquid Stratified Flow through Small Break at Bottom of Horizontal Pipe

The leak of a small break located at the bottom of a horizontal pipe with gas-liquid stratified flow was investigated experimentally and theoretically.A specially designed leak measurement device was developed to measure the quality and mass flux.A 2.5mm id holes placed at the pipe wall was used to simulate the small break instead of using a branch for conventional break experimental apparatus.A radical basis function（RBF） neural network was constructed for prediction of outlet quality.The gas-liquid level and pres-sure difference between the main pipe and the break were identified as two input parameters of the RBF neural network.The mass flux was calculated using a modified two-phase orifice equation.It was found the predicted values of quality and mass flow rate agreed well with the experimental data.