Surface Roughness Analysis and Optimization for ST 37 Steel in Lathe Operation using Design of Experiments Method

The machining process is one of the processes used to process steel materials. The value or level of material roughness is one of the values that is considered in the machining process, according to the desired roughness value. One example is in the turning process, spindle speed, depth of feed, and feed motion are factors that can affect the level of roughness of a material. The level of roughness that is not according to demand is one of the problems in the manufacturing industry, especially in the results of the turning process. One of the causes of the discrepancy in the roughness value is because the parameters used are not suitable. Therefore, this study aims to determine the optimal parameters in the turning process to obtain the optimal roughness value, and to determine the parameters that most influence the roughness of the turning process. In this study, ST 37 steel was used. This study used an experimental design method with Orthogonal Array L8. The experiment was carried out with 5 control factors and 3 noise factors. Each factor has 2 levels and is repeated 8 times. In this study using a conventional lathe and for roughness testing using a Surface Roughness Tester. Based on the results of data analysis, the movement factor has the highest influence on the surface roughness value. The ranking factors that affect the roughness value are as follows: spindle speed of 345 rpm, feed motion of 0.051 mm/rad, depth of cut 1 mm, dry cooling, and automatic sledding. The Design of Experiments (Taguchi) method use quality characteristics to determine the results of the analysis. In this study, the quality characteristic is smaller the better. Smaller the better characteristics is where the smaller value, then better the quality. In variations of the spindle rotation speed and feed motion, we get where if the spindle speed is high and the feed motion is low it will produce a low level of roughness, whereas if the spindle speed is low and the feed movement is high it will produce a high level of roughness.


Introduction
The development of technology and manufacturing science is growing rapidly, this can be seen from the increase in production output [1].The increase in production output is directly proportional to the increase in production quality [2].The machining process is a process that is most widely used to make a finished product with metal as the main material [3].In the machining process, one of the things to pay attention to is the level of surface roughness [4].The value of surface roughness plays an important role in the manufacture of machine components.In the turning process, things that affect the quality include the spindle rotation speed, feeding depth, and feeding movement [5].In the manufacturing industry, especially in the turning process, the cutting energy is converted into heat through the friction process between the chisel and the work piece [6].In the turning process, the surface rough-ness of the work piece that is not suitable is a problem that is often encountered [7].Inappropriate parameters result in the roughness of the work piece not being as de-sired [8].With the background of these problems, ideas arise to analyze the value or level of roughness of turning results using the Taguchi method.Data analysis in this study used the Taguchi method and statistical analysis (Analysis of Variance or ANOVA) to determine the most influential parameter on the level of surface rough-ness.

Taguchi Method
The Taguchi method is one of the methods in which its application is to minimize defects.The Taguchi method was first introduced by Dr. Genichi Taguchi 1949 [9].This method has often been used in industry.The Taguchi method is a method that aims to improve product quality and production processes simultaneously to reduce costs to a minimum [10].The Taguchi method is one of the Off-line methods where this method is carried out before and after production, this method is very good for improving quality and reducing costs.In the Taguchi experiment there are three main stages that include an experimental approach, namely experimental planning, experimental implementation, and analysis, as for the meaning as follows [11]: 1. Experiment Planning Experimental planning is the most important stage, such as determining goals, formulating experimental problem-determining, selecting de-pendent variables, knowing factors, differentiating control factors and confounding factors, selecting the number of levels and factor values.To find out the influencing factors, you can use a cause-and-effect diagram or a fishbone diagram [12].2. Stages of Experiment Implementation The stage of implementing this experiment is an important stage in collecting test results.If an experiment is implemented according to a well-designed one, it will make it easier for researchers to conduct analysis and can get more accurate results [1]. 3. Stages of Experimental Analysis This experimental analysis stage is a data processing stage, namely data collection, data calculation, data management, and delivery of experimental data, where this analysis uses ANOVA [13].To find out the influence of control factors with uncontrollable factors, statistical tests can be carried out [4].
1.2.Quality Characteristic: Smaller is Better.Smaller is better is a quality characteristic with a non-negative value that has a value starting from zero.So that the one with the smallest value or close to zero is the better value [8].The equation to calculate the value of smaller is better are as follows [1]: 1.3.Design of Experiment.Design of Experiment (DOE) is a statistically method that we can use to conduct experiment.DoE is one of many techniques used in practice of quality improvement [14].In DOE there are two main types of variables, the first is the independent variable (factor) and the second is the dependent variable (response).Design of experiment is an experimental strategy in which effects of multiple factors studied simultaneously by running test at various levels of factors [9].In general, in carrying out the DoE procedure there are six main steps, such as deter-mining the purpose of the experiment, selecting the related variables, determining the experimental design, collect data, checking whether the data taken is in accordance with the experimental assumptions, and analyzing the experimental data [15].

Research Methodology
The experiment was conducted using the conventional lathe machine.Mild steel ST 37 is chosen as the work material of this experiment.The Taguchi method is used to analyzing and identify the optimize parameters in the lathe process.Table 1 shows the control factors that used in this experiment.Table 2 shows the experiment design using L8.The range of roughness () in the turning process is used as a standard value for assessing the level of material surface roughness.In this experiment, the DRT320 surface roughness tester is used to measure roughness.The setup of the device is measuring standard multi-indent structure, reading the Ra value, and com-pare the value with the roughness standard value for turning process.The range of roughness () is between 0.5 -6 .

Result and Discussion
The ANOVA results are as follows.Table II and IV show the sum square value for determining a significant factor that affect the surface roughness.According to Table 3 it can be seen that factors E and A are significantly influencing the surface roughness of the ST 37. Based on Table 4 shown the SST of control fac-tors or inner array is 177.8896.In this experiment, the confidence level 95% is re-quired, so the risk or  is 5%.Based on the result, the value of  >  i.e. for factor E 175.76 > 4.26 and for factor A 8.30 > 4.26.Based on the statically analysis, factors E and A are the large significant factor that influence the average value for the surface roughness with at least 95% confidence level.Table 5 summarizes the results of surface roughness.Figure 1 shows the relationship between the carriage traversing wheels with the SN Ratio (signal to noise ratio).To measure the performance of the experiment can use SN-ratio method.The desirable effect to the response is called signal.Based on Figure 1, shows that with carriage auto motion (1) the SN ratio value is on the standard range of turning process.Figure 2 shows that the surface roughness increase as spindle speed decrease and feed rate increase.According to figure 3, the interaction between depth of cut and feed rate is relatively effect the surface roughness.In the other hand, at low spindle speed and auto motion carriage traversing wheel, surface roughness is on the standard range of turning process (0.5 -6 ).

Conclusion
The research (experimental) process was successfully done, and the data was collected through lathe turning process.Based on the analysis data was found that the carriage traversing wheel is the main factor that influence the surface roughness.Increasing spindle speed has a negative effect on the surface roughness.In contrast the feed rate does not affect the surface roughness while the small depth of cut is relatively affected.

Figure 2 .Figure 3 .
Figure 2. Effect of Spindle Speed and Feed Rate on Surface Roughness.

Figure 4 .
Figure 4. Effect of Spindle Speed and Carriage Traversing Wheel on Surface Roughness In this experiment, 3D contour plot were generated using MATLAB software.Based on the figure 2, 3, and 4 the standard value of turning process is shown by color red, purple, and blue.

Table 2 .
Anova result unpooled for inner array.

Table 3 .
ANOVA Result with MS-Excel.