hrvatski

Uporaba backpropagation neuronske mreže za procjenu uklanjanja Cu2+ iona prirodnim i modificiranim zeolitom

Rapid industrialization and the increase in world population have all contributed to heavy metal population in ecosystems due their high toxicities. Elevated environmental levels of Cu2+ come from a variety of sources. Cu2+ ions are found in the water of many industrial processes, such as electroplating, metal industry, production of pigments, paper, fibres, paints, photographic industry, pesticide production. Various methods exist for the removal of toxic metal ions from aqueous solutions, ion exchange, precipitation, adsorption, and reverse osmossis, among others. Some naturally occurring zeolites minerals may serve as cost-effective sorbents for the removal of heavy metals. While their sorption capacity is usually less than those of synthetic sorbents, these materials could provide an inexpensive substitute for the treatment of heavy metal wastewaters. The low price of zeolite and the fact that exchangeable cations (sodium, potassium, calcium and magnesium) from zeolites are not toxic make zeolite an attractive alternative material for removal of undesirable toxic ions from wastewater. After wastewater treatment with natural zeolite, purified water may be reused in the production process or discharged into watercourses, which results in substantial economic saving along with protection of watercourses and the environment. The aim of the study is to explore the potentials of using natural zeolite-clinoptilolite from the Donje Jesenje deposit in the process of Cu2+ ion removal. The natural zeolite contains exchangeable sodium, potassium, magnesium, and calcium ions. In the examination of the ion exchanger should be homoionic. Samples of natural zeolite enriched with Na+, K+, Ca2+ and Mg2+ by conditioning with 2 mol L-1 NaCl, KCl, CaCl2 and MgCl2 at 25 oC. Results showed that removal of examined Cu2+ ions increases in the following order: Mg-form < Ca-form < K-form < natural zeolite < Na-form. Short description of work has been provided, as well as capabilities of neural networks. For the subject issue, using experimental data, feed-forward backpropagation neural network for estimation of removal Cu2+ ions using natural and modified zeolites has been modelled. Coparison of experimental data and data obtained by estimation of neural network shows that applied network model provides very good estimation of the quantity of bound Cu2+ ions, for the data sets which haven&#777 ; ; `t been used for network training.

neuronske mreže; zeoliti; Cu2+ ioni

nije evidentirano