| The re-manufacturing is very important for improving resource utilization rates, constructing revolving economy and economical society and achieving sustainable development of economy and society. Micro arc oxidation as a method to achieve the re-manufacturing is a novel electrochemical surface treatment to generate ceramic layers on light metal and its alloys with low energy consumption. The MAO (Micro arc oxidation) process can be typically carried in alkaline electrolytes, which is environmental friendly, and that results in the continuous generation of short-lived, fine microdischarges across the coating surface. The aluminum alloy treated by MAO is of great interest, due to its high micro-hardness, high anti-corrosion and anti-wear and good adhesion with the substrate. In this research, the Micro Arc Oxidation technique was reseached, and introduced Neural Network into MAO to achieve prodiction of coating performance indicators. And then the main researching contents and results in the text following:1)The three-factor test for three-level orthogonal was desiged to optimize the electrical parameters(current density, pulse frequency and duty cycle) on the test indicators(thicknessã€surface roughness and hardness). The test selected current density of 30A/dm2, pulse frequency of 1000Hz and duty cycle of 8% for optimal process plan. The optimal process plan was used on 6061 aluminum alloy oxidation treatment to obtain the thickness of 15.98μm, surface roughness of 0.843μm and hardness of 1427.8HV of micro-arc oxidation coating.2) The five-factor test for four-level orthogonal design of micro arc oxidation process select Na2SiO3 concentration of lOg/L, NaOH concentration of 2.5g/L, (NaPO3)6 concentration of 6g/L, NaF concentration of 1.5g/L, Na2EDTA concentration of 1g/L for optimal process parameters. And thickness of 26.87μm, surface roughness of 2.413μm and hardness of 1700HV of micro-arc oxidation coating was prepared by micro arc oxidation on 6061 aluminum alloy for optimal process parameters. Moreover, the effects of parameters on test indicators were researched and the results show all factors on three indicators were significant, the Na2SiO3 concentration had the greatest impact on three indicators, and the second impact NaOH concentration, (NaPO3)6 concentration of influence again and NaF concentration again and Na2EDTA concentration impact is minimal.3) The effects of all factors on coating performance (surface morphology, phase composition and corrosion resistance) were observed and analyzed using SEM, X-Ray diffraction (XRD) and immersed test. The results show that the coating is porous and loose, with a large number of lamellar spherical pieces like volcano top. and some micro-cracks appear on the coating surface; XRD patterns indicate that the coating is composed of α-Al2O3 and γ-Al2O3, the treated alumiun alloys were immersed in 3.5% NaCl solution in 200h, the corrosion rate of samples is much lower than untreated smaples.4) Based on the data of orthogonal experiment, the two BP neural networks were designed to predict the thickness, roughness and hardness of coating through new inputing process parameters. It is proved that the network model has a 10% error, and variation of performance indicatiors causing by process parameters is consistent with variation of actual experimental values. Therefore, the established model can be used to qualitative and semi-quantitative analysis of micro-arc oxidation technique. |