PSI - Issue 41

N.A. Fountas et al. / Procedia Structural Integrity 41 (2022) 638–645 Author name / Structural Integrity Procedia 00 (2019) 000–000

643

6

A result less than 0.05 for p -value suggests that the corresponding independent variable is significant. When it comes to lack of fit, p -value has to be greater than 0.05. Note that an insignificant lack of fit is preferred which means that any term excluded by the model is insignificant and thus the model fits well. Anderson–Darling normality test is used to validate the generated model’s suitability referring to UTS response. In Anderson–Darling test, if p is lower than the selected significance level (c.i. = 0.05) the data fails to follow the normal distribution. In the present study ANOVA results, for the quadratic model generated, indicate that the model is suitable for predicting UTS of hybrid PLA/wood fabricated specimens. This outcome is verified by the regression model’s p value which is less than 0.05 (i.e. 0.011). With reference to p -value, it is shown that UTS for hybrid PLA/wood tensile parts is primarily influenced by linear parameters; whilst square terms and interactions follow next. Individual importance referring to each parameter is checked through t-test at 95% confidence level. Consequently if a term has p -value less than 0.05 is considered as significant. Coefficient of determination ( R 2 ) indicates the percentage of total variation in the response explained by the terms in the model. It this work it has has been found equal to 97.79%. Evidence was shown that raster angle deposition (RDA) and nozzle temperature (NT) had the strongest effect on UTS (MPa). Results for the regression model are shown in Fig.2. Fig.2a illustrates a graphical comparison between experimental and predicted results. From Fig.2b the model’s adequacy in predicting the response is evident. Since p -value of the normality plot is found to be far beyond 0.05 (i.e. 0.183) it is concluded that residuals follow a normal distribution and predictions by the regression model are in good agreement with experimental results.

(b) Probability plot of residuals for UTS (MPa) 95% CI

(a) Experimental vs predicted results for UTS (MPa)

17.5 15.0 12.5 10.0

99 95 90 80 70

65 00 43 00 20 10 5 1

UTS (MPa)

Mean -0.00005556 StDev 0.5357 N 18 AD 0.499 P-Value 0.183

Percent

7.5 5.0

UTS exp. UTS model

2 4 6 8 10

12

14

16

18

-2

-1

0

1

2

Number of experiment

Residuals

Fig. 2. (a) Experimental vs predicted results for UTS (MPa), (b) Probability plot of residuals at 95% c.i. (Anderson-Darling test).

Fig.3 illustrates the interaction effects corresponding to the number of process-related independent parameters, LT, NT, DA and PS in the form of contour plots. Fig.3a refers to the interaction effect among layer thickness and nozzle temperature. According to the graphical area where advantageous results are obtained for maximizing UTS (i.e. dark green color) layer thickness should be set between 0.25 and 0.3 mm provided that nozzle temperature is to be set between 210 o C and 220 o C. Fig.3b shows the interaction effect between layer thickness and deposition angle where beneficial values for maximizing UTS are found in lower setting for deposition angle while keeping layer thickness between 0.10 mm to 0.20 or 0.25 mm. Fig.3c shows the interaction effect among the settings of layer thickness (LT) and printing speed (PS). It is concluded that decision making concerning these two variables is quite complex. It is suggested that high UTS can be obtained either by setting layer thickness and printing speed at low levels or by setting both parameters at high levels, i.e., 0.30 mm layer thickness and 50mm/sec for printing speed. Fig.3d suggests that low levels for nozzle temperature (i.e. 180 o C -190 o C) and deposition angle (i.e. 0deg.-7deg.) should be determined to maximize UTS. Fig.3e refers to the interaction effect between nozzle temperature and printing speed. It is observed that low settings i.e. 180 o C-187 o C and 30 mm/sec-32 mm/sec should be determined for nozzle temperature and printing speed respectively to maintain high UTS results. Finally Fig.3f indicates that the entire range of values referring to printing speed may be advantageous for maximizing UTS. However this is valid for relatively low settings for deposition angle, i.e., 0deg to 20deg.