|
Author(s): |
Shaohui Xu, School of Petrochemical Engineering, Liaoning Shihua University, Fushun, China,113001
Xiaoning Wang, School of Petrochemical Engineering, Liaoning Shihua University, Fushun, China,113001
Junfeng Cao, School of Petrochemical Engineering, Liaoning Shihua University, Fushun, China,113001
Shougen Hu, University of Shanghai for Science and Technology, Shanghai 200093, P.R.China |
|
Abstract: |
The pneumatic conveying experiments for the millet and hollow glass bead which had the same average particle diameter and different densities were carried out by using compressed air in horizontal Tee branch pipe. The paper used the simulated annealing algorithm of BP neural network to simulate and forecast solid phase flow distribution characteristics of the pipe network system. The results show : when the operation pressure was const, and two branch pipes had the same control valve opening, the solid-phase mass fraction of the two collection container were almost equal. When the opening value of the flow metering valves was different，with the control valve opening of the branch B decreases, the distributive solid mass fraction decreases in the collection container B. When the opening angles of the flow metering valves was the 0o-60o, the millet particle was earlier to get the critical point than the hollow glass bead, and the mass fraction of the collection container B for the millet is greater than for the glass bead. With the comparison between the simulation results of the simulated annealing algorithm of BP neural network and the experimental data, it is shown that the simulation values agree well with the experimental data. This illuminates that the simulated annealing algorithm had a better adjustability to the simulation and prediction of the flow distribution characteristics of solid phase for tee branch pipe．
|