Search

Based on genetic algorithm (GA) and fuzzy neural network, a new method for the study of sputtering process is proposed in this paper. Diamond-like carbon (DLC) coatings were deposited on SKD11 steel by magnetron sputtering. An orthogonal array design is implemented and the effects of control factors on surface properties of the coatings were systematically analyzed. The coating properties were investigated by scanning electron microscopy and Raman spectroscopy, and wear volume surface performance of the Zr-doped DLC coatings was evaluated by a wear tests pin-on-disk tribometer. The Raman analyses showed that, at lower ID/IG ratio, a lower wear volume of the Zr-doped DLC coatings can be obtained. Scratch tests as well as Rockwell indentation tests revealed that the graded Zr-doped DLC structures efficiently provide better adhesive strength of DLC coatings. The results show that the wear behaviors of the DLC coatings can be improved by Zr-doping, which the Zr-doped DLC coatings exhibited promising tribological properties. Also, the predictive ability of the GA-ANFIS computations for the tribological behaviors of the Zr-DLC coatings within the experimental domains proved to be reliably obtained, where the forecasted values and experimental results are close.

The surface properties of plasma-sprayed partial Y2O3 stabilized ZrO2 coatings with different travel speed and powder-feeder rate of sprayed gun on aluminum substrates have been reported. Several coatings analytical parameters were studied and these factors affecting coating properties were carried out in the plasma-sprayed experiments. Experimental results have shown that the structure of the sprayed coatings that makes relatively diverse in texture for most of the coatings was examined by a SEM. The hardened strength with a low level of a slant of powder-feeder rate became significantly weaker as the travel speed was decreased. Interestingly, the structure of the specimens for the travel speed of 30mm/s is more homogenous than that of the others, indicating a coating with good homogenous structures, while the interfaced structure of the specimens for the powder-feeder rate of 25 mm/s is more homogenous than that of the others, indicating a coating with good adhesive structures. The tendency of the effect of travel speed of sprayed gun on the hardened surface of coatings at powder feeder rate of sprayed gun varied to 30g/min from 20g/min seems to be apparent.

This study reports the statistical optimization of experiments, designed by Taguchi, to improve the wear-resistance of the tungsten carbide/cobalt-based coatings applied using high-velocity oxygen-fuel (HVOF) spraying processes. In this study, a group of needlelike textures appeared in the HVOF WC-Co coatings. The structures were evenly distributed and compacted closely, showing a more dense and homogenous configuration. The coating that formed with a large amount of pores, microcracks and unmolten particles was much looser, and the interface was locally prone to defects. The coatings were obviously not homogenous. Experimental results show that the wear-resistant coatings are highly dense, with an elastic lamellar structure. The tribological properties were significantly improved and the surface textures showed uniformity, in a molten droplet flattened formation. A homogenous, morphological structure, with local small aggregation, was more obvious after HVOF WC-Co coating. Accordingly, it is clear that the surface properties determined proved favorable to the mechanical properties of the WC/Co coatings. In addition, we compared the linear models with the quadric models, using RSM methods, in all the experiments. Statistically, the linear models produced an average error of 8.285％, while the quadric models generated an average error of 1.399％. It is clear that the RSM models successfully fitted the HVOF WC-Co coating process and the quadric models yielded more accurate predictions than the linear models, in most cases. The proposed procedure was applied to HVOF spraying of WC-CO coatings. The results demonstrated its feasibility and its effectiveness in improving the HVOF WC-Co coating.

This paper reports the effects of varying magnetic field strength on CrN films, deposited by a magnetic sputtering process. The strength of magnetic field in unbalanced magnetic sputtering processes is controlled by adjusting the gap distance between the magnet set and the target surface (GDMT). An improvement in overall intensity, at low GDMT, was observed by adjustable magnetic field distributions. In the chamber, it was readily noticeable that varying the magnetic field strength has an influence on the CrN structures. In experiments, at low GDMT, a high hardness value and lower wear rate become visible in the CrN films. In addition, the CrN films formed have a smooth surface with a dense tiny structure and display preferential orientation in the Cr2N(111) and Cr2N(002) planes, whereas CrN films prepared at higher GDMT exhibit more roughness and the CrN (200) plane is evident. Furthermore, the Cr2N (111) (002) plane possessed better tribological properties than that of the CrN(200) plane, where the wear scars show little failures on the coating surface.

This study presented the application of Taguchi designed experiments to optimize the tungsten carbide/cobalt (WC-Co) deposited quality properties of the high-velocity oxygen-fuel (HVOF) process. The eight control factors were designed in a L18 factorial orthogonal array experiment, and the effects of control parameters on the surface properties were critically evaluated in the experiments. The surface anti-wear properties by optimal design have been studied comparatively. By optimization from Taguchi S/N ratios and analysis of variance, the surface wear volume values are decreased by 37%. It was clear that anti-wear performance estimations of HVOF sprayed WC-Co coatings are greatly improved by way of the optimal experiment, and these findings achieved the desired values in the WC-CO coatings. The experimental results indicated that the proposal method offers an effective and efficient approach for developing a robust and highly quality HVOF sprayed WC-Co coating.

This study reported on Taguchi-Fuzzy logic model with statistical experiments to characterize the tungsten carbide/cobalt (WC-Co) deposited quality properties of the high-velocity oxygen-fuel (HVOF) process. The eight control factors were designed in a factorial orthogonal array experiment, and the effects of control parameters on the integrated surface properties were critically evaluated in the Taguchi designed experiments. A fuzzy logic system with knowledge bases in the inference engine was used to explore the relationships between variables and responses in the modified structures. In addition, the structure and properties of HVOF sprayed WC-Co coating were reported. Furthermore, a confirmation experiment of the optimal process was conducted. It was clear that HVOF sprayed WC-Co coating estimations are greatly improved through the Taguchi-Fuzzy logic operation, and these findings achieved the desired values in the modified properties.

The paper describes response surface methodology (RSM) based on design of experiments and analysis of variance (ANOVA) as a statistical design while developing a robust plasma transfer arc (PTA)coating process. Based on ANOVA, The relative important parameters with respect to surface at hardness values were identified in the Taguchi design, where they were further used in predictors. In addition, we applied three-dimensional graphs in RSM to develop a robust PTA response surface yielding the desired-better area of a treated layer. In this study, a quadratic polynomial with a Box-Behnken design is utilized. The results reveal that RSM provides the effective methods as compared to the traditional trial-and-error method for exploring the effects of controlled factors on response. A very good agreement was observed, as evidenced by R-squared value, 90%, between the predicted and the experimental data, and its error percent is found to be approximately 3.801% in the PTA-coating process. It is clear that RSM model demonstrated better accuracy in predicting surface hardness for PTA-coating process. Accordingly, RSM based on design of experiments was used as statistical PTA-coating design tools combined with the hardness model. Device zone optimization and yield enhancement have been demonstrated.

This study optimizes the thermal dissipation ability of aluminum nitride (AlN) ceramics to increase the thermal performance of light-emitting diode (LED) modulus. AlN powders are deposited on heat sink as a heat interface material, using an electrostatic spraying process. The junction temperature of the heat sink is developed by response surface methodology based on Taguchi methods. In addition, the structure and properties of the AlN coating are examined using X-ray photoelectron spectroscopy (XPS). In the XPS analysis, the AlN sub-peaks are observed at 72.79 eV for Al2p and 398.88 eV for N1s, and an N1s sub-peak is assigned to N-O at 398.60eV and Al–N bonding at 395.95eV, which allows good thermal properties. The results have shown that the use of AlN ceramic material on a heat sink can enhance the thermal performance of LED modules. In addition, the percentage error between the predicted and experimental results compared the quadric model with between the linear and he interaction models was found to be within 7.89%, indicating that it was a good predictor. Accordingly, RSM can effectively enhance the thermal performance of an LED, and the beneficial heat dissipation effects for AlN are improved by electrostatic spraying.

This paper proposes a design for an adaptive system by modeling the relationship between coating surface roughnesses and the controling factors in plasma spray coating processes. A statistical design was used to obtain sufficient experimental information with the least number of experiments. Analysis of variance was then used to select significant control factors for reinforced coatings, and these identified factors used to construct an adaptive fuzzy logic control model. In order to model the process, a fuzzy logic controller (FLC) was utilized. A genetic algorithm (GA) was applied as a tool to optimize rule bases from traditional FLCs. Therefore, with the use of a GA-optimized FLC, robust reinforced deposition for coatings in the plasma spraying process can be obtained. The experimental results show that the obtained optimal rule base for FLC is capable of achieving the desired results. That is to say, the proposed design, which combines a statistical method and a GA-optimized FLC, is efficient and robust for the investigation of reinforced coatings in a plasma spraying process.

In this paper, we present the application of the fuzzy logic analysis to a Taguchi orthogonal experiment for developing a robust model with high efficiency in multiple performance characteristics (MPCs) of the plasma transfer arc welding (PTAW) hardfacing process. It eliminates uncertain information and is a simple, effective, and efficient approach. A fuzzy logic system is used to simultaneously investigate relationships between various MPCs and to determine the efficiency of each trial of the Taguchi experiments. From the fuzzy inference process, we are able to determine the optimal setting of factor-levels for the MPCs. In addition, we also use the analysis of variance (ANOVA) to identify the significant factors, which coincide with findings from the fuzzy logic analysis and are found to account for about 79% of the total variance. Furthermore, a confirmation experiment of the optimal process is conducted, and it verifies that both individual performance characteristics and MPCs are successfully optimized and satisfy our desired levels of MPCs.

We propose the application of an artificial neural network to a Taguchi orthogonal experiment to develop a robust and efficient method of depositing alloys with a favorable surface morphology by a specific microwelding hardfacing process. An artificial neural network model performs self-learning by updating weightings and repeated learning epochs. The artificial neural network construct can be developed based on data obtained from experiments. The root of mean squares (RMS) error can be minimized by applying results obtained from training and testing samples, such that the predicted and experimental values exhibit a good linear relationship. An analysis of variance indicates that the significant factors explain approximately 70% of the total variance. Consequently, the Taguchi-based neural network model is experimentally confirmed to estimate accurately the hardfacing roughness performance. The experimental results reveal the hardfacing roughness performance of the product of PTA coating is greatly improved by optimizing the coating conditions and is accurately predicted by the artificial neural network model. The combination of the neural network model with Taguchi-based experiments is demonstrated as an effective and intelligent method for developing a robust, efficient, high-quality coating process.

This study applies the Taguchi method to elucidate the quality characteristic of the melt depth in electron beam surface treatment (EBST) of cast iron. The-nominal and the-better was a measured quality characteristic to evaluate the experimental results by computing the signal-to-noise ratios (SNRs) of the EBST melt depth after surface-hardening tests. It includes two steps to determine the optimum parameters or conditions. The first step is to maximize SNR to minimize the variance, and the second step is to adjust the EBST melt depth to a target value using the sensitivity. Results from the experiment indicate that the optimal process parameters are a substrate matrix of gray cast iron, a travel speed of 39 mm/s, an accelerating voltage of 10 kV, an electrical current of 15 mA, a melt width of 20 mm, an elliptical beam oscillation, and a post-process heat treatment at 150 °C. The best EBST melt depth obtained in the study was 0.747 mm, corresponding to the maximum SNR. The microstructure is composed of fine dendrites and martensite phases. The obtained EBST depth is confirmed by its falling into the predicted 95% confidence interval. The prediction of the optimal setting is close to the actual target value. The reproducibility of the optimal EBST parameters was experimentally established with data from the confirmation run. EBST wear-resistant surface hardening can be performed using Taguchi experiments.

This study optimizes aluminum nitride (AlN) ceramics, in order to enhance the thermal performance of light-emitting diode (LED) packages. AlN coatings are grown on copper/ aluminum substrates as a heat interface material, using an electrostatic spraying process. The effect of the deposition parameters on the coatings is determined. The thermal performance of AlN coated Cu/Al substrates is evaluated in terms of the heat dissipated and compared by measuring the LED case temperature. The structure and properties of the coating are also examined a scanning electron microscopy (SEM). In sum, the thermal performance of the LED is increased and good heat resistance characteristics are obtained. The results show that using AlN ceramic coating on a copper/aluminum substrate increases the thermal performance.

The applicability of friction stir welding to 6061 aluminum alloy with a cylinder-screw-shaped S45C steel spin has been investigated. Three different tilted angles of spin were designed in the friction stir welding experiments. Experimental results have shown that surface defects at spin tilted angle were not observed at the weld zone such as face, toe and root of the weld, and a concentric onion ring pattern with increasing tilted angle of spin appears to be faintly visible. The tensile strength of the butt joints of tilted angles of spin are found to be higher that that of the normal. Also, butt joints through tilted angles of spin are higher than that of substrate matrix, while the tensile strength of the normal was less that that of the substrate matrix. The proposed procedure was applied at butt joints of friction stir welding, and the implementation results demonstrated its feasibility and effectiveness to improve tensile tested experiments for developing robust friction stir welding