Font Size: a A A

Study On The Artificial Neural Network Analysis Method In Double Components

Posted on:2011-07-26Degree:MasterType:Thesis
Country:ChinaCandidate:S LiFull Text:PDF
GTID:2120330332964756Subject:Marine Chemistry
Abstract/Summary:PDF Full Text Request
Calcium and magnesium are major elements in seawater, are the key parameters of examining water hardness. So it is significant to determine simultaneously the two elements. Iron and manganese are the trace metal nutrients in seawater and play important roles in the growth of marine life. Therefore it is also meaningful to study the determination method of iron and manganese.In this paper, we combine the artificial neural network (ANN) in chemometrics with spectrophotometric method or with the catalytic kinetic spectrophotometric method so as to determine calcium and magnesium simultaneously and determine iron and manganese simultaneously. Two new ANN methods are established. Appling the two methods to actual water system, the satisfying results are obtained.The main achievements of this paper are as follows:1. A new ANN-spectrophotometric method which can determine simultaneously calcium and magnesium is established.(1) The ascertained reaction system is arsenazo (Ⅲ) as indicator, borax-hydrochloric as buffer solution and phenanthroline as masking agent.(2) The optimal determination conditions are:w[arsenazo(Ⅲ)]= 0.010%, pH= 9.10, w (phenanthroline)= 0.40%,λ=610~660 nm and the reaction time at t=10 min.14 kinds of common cations and anions do not interfere with the determination under this analytical condition.(3) The network structure and arithmetic in data process using ANN are determined. Back-Propagation (BP) network and Levenberg-Marquardt (L-M) arithmetic are chosen. The topological structure of BP ANN network is chosen as 13-14-2 (node), the initial gradient coefficientμis fixed at 0.001. The data are processed by computer with our own programs written in MATLAB 7.0.(4) The new method is successfully applied in actual water system.The precision for calcium and magnesium is 2.31% and 2.14% respectively with this new method. In the tap water media, the recoveries of calcium and magnesium are in the range of 101.1%~103.6% and 100.8%~102.4%, the RSD is in the range of 0.082%~1.43%. In the Yellow River water media, the recoveries of calcium and magnesium are in the range of 100.7%~103.2% and 96.6%~101.3%, the RSD is in the range of 0.14%~0.33%. In the seawater medium, the recoveries of calcium and magnesium are in the range of 98.8%~103.3% and 98.4%~103.4%, the RSD is in the range of 0.22%~2.21%.2. A new ANN-catalytic spectrophotometric method which can determine simultaneously iron and manganese is established.(1) The ascertained reaction system is rhodamine B as indicator, potassium periodate as oxidant, nitrilotriacetic acid and phenanthroline as activation reagents and HAc-NaAc as buffer solution.(2) The optimal determination conditions are:p (rhodamine B)= 4.00×10-3 g/L, pH= 3.85, c (potassium periodate)= 2.00×10-4mol/L, c (phenanthroline)= 2.00×10-3mol/L, c (nitrilotriacetic acid)= 1.00×10-5mol/L,λ= 555 nm, the testing time in 6 min and the initial temperature at 35℃.21 kinds of common cations and anions do not interfere with the determination under this analytical condition.(3) The network structure and arithmetic in data process using ANN are determined. Back-Propagation (BP) network and Levenberg-Marquardt (L-M) arithmetic are chosen. The topological structure of BP ANN network is chosen as 15-16-2 (node), the initial gradient coefficientμis fixed at 0.001. The data are processed by computer with our own programs written in MATLAB 7.0.(4) The new method is successfully applied in actual water system.The precision for iron and manganese is 2.30% and 2.67% respectively with this new method. In the tap water media, the recoveries of iron and manganese are in the range of 98.0%~104.3% and 96.5%~104.5%, the RSD is in the range of 0.23%~0.98%. In the Yellow River water media, the recoveries of iron and manganese are in the range of 96.0%~101.0% and 98.7%~104.2%, the RSD is in the range of 0.13%~2.52%. In the seawater medium, the recoveries of iron and manganese are in the range of 95.3%~104.8% and 95.3%~104.7%, the RSD is in the range of 0.14%-2.66%.3. Using the established ANN- catalytic spectrophotometric method, the iron and manganese concentrations in the surface seawater of 11 sites in Qingdao offshore are determined and the level distribution map of iron and manganese are drawn.
Keywords/Search Tags:artificial neuron network, simultaneous determination, catalytic kinetic spectrophotometric method, calcium,magnesium, iron,manganese
PDF Full Text Request
Related items