Developing a Generalised Neural-Fuzzy Hydrocyclone Model for Particle Separation
نویسنده
چکیده
Development of a neural-fuzzy model for an operational hydrocyclone is reported in this paper. The model integrates the benefits of the Artijkial Neural Network (ANN) and the fuzzy-logic techniques. It preserves the generalisation capability of an ANN while expressing the final model in fuzzy rules. These rules can be modiJied and examined by the user. This will in turn control the interpretation ability of the system. Results from a case study have shown that the new proposed neural-fuzzy hydrocyclone model produces comparable results as those from the ANN model but with an added advantage of the use of linguistic ficzzy rules.
منابع مشابه
Particle Size Effects on Hydro-Cyclone Performance
The hydrocyclone has a very important roll in industrial separation. The consideration of its behavior is very important for design. In this investigation, behavior of water flow and particles trajectory inside a hydrocyclone has been considered by means of numerical and experimental methods, and results have been compared together. To have a numerical simulation, a CFD software was used, and f...
متن کاملKaolin Processing Using Fuzzy Hydrocyclone Control
In many cases, the hydrocyclone control is desirable for efficient separation especially in case of changing feed conditions. Control concept based on exact mathematical models is not satisfactory due to the complexity of the physical process. In this case, fuzzy control is a suitable solution. The data are primarily derived from an optical sensor that monitors the outside view of the free unde...
متن کاملNumerical Simulation of Solid-liquid Flow in Hydrocyclone
Hydrocyclone is widely used as the centrifugal separation equipment to separate, classify and concentrate the product. In this paper, the multiphase flow models of mixture and Euler-Euler are used to simulate the internal three-dimensional flow field of hydrocyclone. It is found that compared to the experiment, the mixture model is shown to have the best performance among the models of mixture,...
متن کاملPSO-ANFIS and ANN Modeling of Propane/Propylene Separation using Cu-BTC Adsorbent
In this work, an artificial neural network (ANN) model along with a combination of adaptive neuro-fuzzy inference system (ANFIS) and particle swarm optimization (PSO) i.e. (PSO-ANFIS) are proposed for modeling and prediction of the propylene/propane adsorption under various conditions. Using these computational intelligence (CI) approaches, the input parameters such as adsorbent shape (S<su...
متن کاملA CFD Simulation of the Parameters Affecting the Performance of Downhole De-oiling Hydrocyclone
Among the all parameters affecting the performance of a downhole de-oiling hydrocyclone, the investigation of internal flow field deserves more attempts especially in the petroleum industry. In this study, the effects of inlet flow rate, inlet oil volume fraction, and oil droplet diameter on the separation efficiency and pressure drop ratio have been investigated along the hydrocyclone body. Al...
متن کامل