Study On Separation Characteristics And Software Design Of Hydrocyclones For Oil-water Separation

The hydrocyclone for oil-water separation is a new kind of centrifugal separator in petroleum industry. The separation characteristics of hydrocyclones with different physical property were studied in this paper. Based on the results of the separation experiment, the artificial neural network, Monte Carlo method and some other mathematics models were applied to simulate the oil-water separation processing in the software design of hydrocyclones. By means of the software, the geometric parameters and operational parameters were optimized according to the different physical property and separation requirement. Moreover, it will output the grade efficiency of the optimized hydrocyclone.The effect of oil droplet size on the separation efficiencies of each part of the hydrocyclone and the whole hydrocyclone was studied systematically by using a pressure buffer sample device designed by ourselves to obtain the sample at the wall of hydrocyclone at the first time. The author drawn a conclusion firstly that the separation efficiencies of the reducing section and the tapered section increased with the increasing of inlet droplet size, whereas the separation efficiency of the parallel section did not vary very much with the change of mean inlet droplet size. The oil droplet size distribution at the wall of the hydrocyclone along the axial direction and its influencing factors were also studied firstly in this paper. The result suggested that the mean oil droplet size decreased gradually along the axial direction on the condition of proper separation without droplet break up, and itshowed that the separation was happened in each part of the hydrocyclone if thehydrocyclone worked on a proper condition. The split ratio does not affect very much on the mean oil droplet size at the wall of the hydrocyclone. The variation of the mean oil droplet size at the wall is mainly dependent on the inlet flowrate of the hydrocyclone. When inlet flowrates exceed a certain value, the mean oil droplet size will decrease with the inlet flowrate increasing. If the inlet flowrate is too small, the stable center full of oil cannot be formed in the hydrocyclone, and the mean oil droplet size at the wall will not change with the variation of the inlet flowrate which will result a low separation efficiency of the hydrocyclone. It is beneficial to improve the separation efficiency of hydrocyclones by increasing the mean oil droplet size, optimizing the structure of the reducing section and the tapered section and operating at a proper inlet flowrate.The author also have made a systematic study on the pressure drop and pressure drop ratio in the hydrocyclone. It can be concluded that the pressure is in a linear relationship with the split ratio, and it will rise with the increasing of the split ratio. The pressure drops of overflow outlet and the sample point along the axial direction increase distinctly with the split ratio increasing, whereas the pressure drop of underflow outlet decreases slightly with the increasing of the split ratio. The pressure drops of each part and the whole of the hydrocyclone change with the variation of inlet flowrates according to a index law. The proportion of pressure loss was also discussed in this paper. The result showed that the proportion of pressure of inlet, the cylindrical section and the reducing section is the largest, and it does not vary very much with the change of inlet flowrate. The proportion of pressure loss of the tapered section is less than that of inlet, the cylindrical section and the reducing section, and it will increase with the increasing inlet flowrate. The proportion of pressure loss of the parallel section is the least, and it will decrease with the rise of inlet flowrate. Decreasing the split ratio and inlet flowrate without decreasing the separation efficiency of the hydrocyclone distinctly, increasing the section area of inlet properly will help us to decrease the energy loss of hydrocyclones.The improved Back Propagation Neural Network (BPNN) was used tos...