Papers by Keyword: Design of Experiment

Abstract: In this article, technological parameters such as cutting speed, feed rate and depth of cut and their relationships with average power are investigated and modelled using statistical design of experiments. This allows a quantitative and qualitative description of the relationship between average power and process parameters. The analytical knowledge gained from the tests provides the conditions for optimising the energy consumption of the turning process.

Abstract: Plasma Arc Machining is a metal cutting where conductor metal such as sheets metal are cut with plasma arc. Problem in plasma arc machining is the result of cutting has burr which is quite large due to the heat, resulting the surface roughness on the workpiece. This research aim to minimize the surface roughness of the stainless steel plate uses a design of experiment method with full factorial design. In this research, there are three factors, that are torch height, cutting speed, and electric current. Each factor has three levels. By using full factorial design, the number of treatments are 3³=27 trials. The results of the research on data processing analysis of variance show that the most influential factor on surface roughness is cutting speed with contribution value of 90.76% followed by two other factors, that is height torch with contribution value of 2.42% and electric current with contribution value of 0.23% and contribution value of noise by 6.59%. Then based on data processing robust design the optimum combination of parameters is obtained by using setting 1 mm torch height, 2400 mm/min cutting speed, and 30 A electric current. Based on the confirmation experiments, experiments with optimum parameter combinations can reach a gap noise of 2.283 dB. Therefore, the design of experiment for determining parameter setting plasma arc machining can determine the optimum combination of parameters to minimize the surface roughness.

Abstract: In this study, the influence of operating parameters on the relative density and microhardness property of a septenary equiatomic Ti-Al-Cr-Nb-Ni-Cu-Co high entropy alloy developed via spark plasma sintering (SPS) process was investigated at constant heating rate (100 °C/min), dwell time (5 min), pressure (50 MPa). Using response surface methodology (RSM) on the sintering temperature (ST) and milling time (MT) as the process variable parameters, a predictive model was established. The design of experiment approach was employed to minimize numbers of runs of experiment, which invariably eliminates trial by error associated with traditional experimental methods. MT and ST were taken as the variables towards the development of the design model. The optimum operating parameters were predicted using the user-defined design (UDD) under RSM and the result was validated through experiments. Observation from the results shows that MT and ST play a significant role in achieving high densification, which translates to high hardness. At 900 °C ST and MT of 10 hours, the highest hardness value of 580.1 HV, densification of 99.98%, and percentage porosity of 0.02% were recorded.

Abstract: The quality of the finished product is affected by a number of factors during the plastic injection molding process. Two crucial process variables in the creation of products are melted and mold temperature. The study uses the Design of Experiments tool in Autodesk Moldflow to look at the impact of melt and mold temperatures on injection molding technology. Analytical items are specifically made of polypropylene (PP) using kids' chair mold. According to simulation analysis results, there is a remarkable effect of the melt temperature on both time at end of packing as well as deflection in the range of the analytical temperature at t_mold of [40, 80]°C and t_melt of [180, 220]°C. Melt temperature also shows a notable influence not only on deflection but also on sink mark depth and volumetric shrinkage, along with the criteria to evaluate the expense of a product (time at end of packing, total part weight).

Abstract: Friction stir welding (FSW) has become very popular for joining similar or dissimilar aluminum alloys. The heat used for this welding process is caused by friction between the welding tool and the workpiece which the axial force, the main parameter for heat generation, plays a very important role. Insufficient heat during welding will result in defective workpieces. This research is aimed to predict the axial force from the relevant factors of the FSW of dissimilar materials (aluminum alloys AA2024-T3 and AA6061-T6). The 2³ full factorial design with center point was used for this experiment that consisted of 3 main factors: 1) rotation speed (rpm), 2) welding speed (mm/min), and 3) pin geometry each factor has 2 levels and 2 replications with the total of 20 experiments. The axial force data of each experiment were collected using a stationary dynamometer which obtained the data acquisition every 0.1 seconds (frequency of 10 Hz). The results from the design of experiment were analyzed by statistical method at the significance level of α = 0.05 which found that the significance and the optimum value of the main factors were rotation speed of 1500 rpm, welding speed of 35 mm/min, pin geometry of tri flat threaded, and the 2-way interaction between rotation speed and pin geometry. Furthermore, the prediction of the average axial force value on dissimilar aluminum alloys obtained from the specified parameters equals 478.91 N.

Abstract: The present work addresses the powder bed binder jet 3D printing as an additive manufacturing process for cement-based materials in the constructions industry. Features are created through the interaction among the droplets of the liquid binding agent and the layered powder bed. The printhead movement over the powder bed at a given feed rate forms voxels and single-lines from the coalesce of successive droplets and adjacent lines are consolidated to create the designed cross-section. Here, statistical models have been developed to study the effect of printing parameters (aggregate particle size, feed rate, velocity of powder spread, pressure of the fluid and nozzle diameter) on the resultant dimension of a single printed line, using a factorial design of experiment. The hardware of the 3D printer, the physical properties of the powder blend and binder are initial constraints for designing voxels. Linear regression models of significant parameters are presented. Pressure is one of the most significant factors, it has a profound effect on the granule formation mechanism. Cubic samples printed with higher pressure level are characterized by higher residual porosities from crater channels during the printing process. The results demonstrate a fundamental understanding of the binder–powder interaction for cementitious materials which can be leveraged to determine the minimum printable feature with required dimensional accuracy, based on the chosen process parameters.

Abstract: This paper discuss the use of Box Behnken design (BBD) to optimize parameters used in conducting experiment for radiation induced grafting (RIG) experiment of graft cinnamaldehyde (antimicrobial agent) to polyvinyl alcohol/sago starch (PVA/SS) film in order to develop antimicrobial film for food packaging. BBD is having the maximum efficiency with objective to have maximum value of grafting yield (GY). This experiment involving three parameters which is absorbed dose (kGy), temperature (°C), and reaction time (min), all in three levels. The proposed BBD requires 15 runs of experiment for data acquisition and modeling the response surface. Three regression models were developed, and their adequacies were verified to predict the output values at nearly all conditions. This work resulted in identifying the optimized set parameters values for RIG experiment, which is absorbed dose at 102.67 kGy, reaction time at 51.67 minutes and reaction temperature 44.68°C in order to achieve maximum value of grafting yield at 20.79%. Afterwards, the models were validated by performing actual experiments, taking three sets of random input values. The output parameters (actual value) measured through experiments are in good consistency with the predicted values, where the actual value of GY is 18.7% as compared to predicted value of GY of 20.79%. The deviation value 2.09% prove success of developed model in predicting grafting yield in RIG using limited number of experiments.

Abstract: Medium chain triglycerides (MCT) are important substrates of the energy metabolism and anabolic processes in mammals. In this study, MCT-rich oil was encapsulated in the mixing ratios of maltodextrin and protein by spray drying to produce spray-dried MCT-rich oil (SMCT). Spray-dried conditions were an inlet temperature of 200 °C, an outlet temperature of 90 °C, and a flow rate of 0.70 L/h. Box–Behnken experimental design and response surface methodology were applied for modeling the influence of formulation variables on powder recovery of SMCT. The key variables were concentration of maltodextrin (10-30% W/W), total protein (5–15% w/w), and MCT-rich oil (5–15% w/w). The microparticles were characterized in terms of particle morphology, yield, Carr's index, moisture content, flowability, hygroscopicity, and powder diffraction. The highest yield of SMCT was 41.19% obtained under the optimized conditions (maltodextrin concentration of 15% w/w, total protein concentration of 8% w/w, MCT-rich oil concentration of 15%). Experimentally obtained values were consistent with those predicted by the model, indicating the suitability of the employed model and the success of the model in optimizing the formulation.

Abstract: Colonic drug delivery systems (CDDS) show several advantages for treatment of inflammatory bowel disease such as improving the clinical outcomes and minimizing side effects of corticosteroids. However, variation of the patient's gastrointestinal tract (GIT) in terms of transit time and pH causes the fluctuation of the drug releasing site in the GIT resulting in low therapeutic efficiency. Consequently, 3D-printing techniques have been applied for preparation of personalized colonic drug delivery systems in this study. Prednisolone was selected as a model drug and prepared in the form of a core tablet. Polylactic acid (PLA) and polyvinyl alcohol (PVA) were printed as a tablet housing and a drug control release plug, respectively. A two-factor full factorial model was utilized to design the experiment and partial least square regression (PLS) models were generated to reveal and predict drug release time of the system. From the results, only thickness of the PVA plug significantly affected the drug release at sampling times of 5, 6, 10, and 24 h. The relations between thickness of the plug and drug releases at 5, 6, and 10 h are quadratic but that of 24 h is linear. The relation between thickness of the plug and drug releases is quadratic. The drug could not be completely released in 24 h because the drug was entrapped by PVA gel. The search results show the possibility to utilize the PLS models to modify drug release time for individual patients. However, alteration of plug polymer is a suggestion to obtain complete drug release.

Abstract: The prototyping of dryer design and performance by application of the trial-and-error technique in one-factor-at-a-time (OFAT) testing is completely arbitrary, expensive and time consuming. Reducing product development lead-time and cost while concurrently improving customer satisfaction for a good manufacturer enhance rapid response to market demand which is a highly effective way of improving returns on investment. In this study a numerical model for the digital prototyping of the rectangular passive greenhouse dryer design and the optimization of the batch process in the solar dryer was developed. An interactive, user-friendly computer package ANSYS 14.0 was used to develop an empirical model. The package was used among others, for the response surface methodology (RSM) optimization to specify the dryer parameters that maximize the dryer mean temperature. The factorial experiments in a central composite design (CCD) revealed that only the inlet vent dimensions influence the mean temperature within the greenhouse dryer. The parametric analysis for robust design yielded the inlet vent height of 0.27m and inlet vent width of 0.45m as the optimum design variables that maximize the mean temperature of the drying air as 320.48K (47.30 °C). The numerical approach established facilitated the prototyping and optimization of the batch process in the passive greenhouse dryer.The prototyping of dryer design and performance by application of the trial-and-error technique in one-factor-at-a-time (OFAT) testing is completely arbitrary, expensive and time consuming. Reducing product development lead-time and cost while concurrently improving customer satisfaction for a good manufacturer enhance rapid response to market demand which is a highly effective way of improving returns on investment. In this study a numerical model for the digital prototyping of a rectangular passive greenhouse dryer design and the optimization of the batch process in the solar dryer was developed. Multiple regression was used as the data-analytic system for the factorial experiment to develop an empirical model, predict the response variable and then test hypothesis in an interactive, user-friendly computer package ANSYS 14.0. The package was further used for the response surface methodology (RSM) optimization to specify the dryer parameters that maximize the dryer mean temperature. The factorial experiments in a central composite design (CCD) revealed that only the inlet vent dimensions influence the mean temperature within the dryer. Appraisal of the model through the coefficient of determination ( =0.99973) showed that the model can account for 99.973% variability observed in the dryer mean temperature consequently, the suitability of RSM for the analysis of the dryer variables. The parametric analysis for robust design yielded the inlet vent height of 0.27m and inlet vent width of 0.45m as the optimum design variables that maximize the mean temperature of the drying air as 320.48K (47.30°C). The numerical approach established facilitated the prototyping and optimization of the batch process in the passive dryer.