The Use of Taguchi Method in Design of Dies Propeller Production Using CNC Machine

⎯propeller is a rounded blade that rotates in a circle, helping to move a vehicle by pushing against water or air. A speed boat's propeller moves it forward by spinning against the water. Propellers use the laws of physics to propel, or drive forward, an aircraft or a boat. The basic structure of a propeller is a spinning or rotating shaft with wide, curved blades attached to it. The very simplest propellers were first used in ancient Greece, where the inventor Archimedes invented a "screw propeller" that moved water for irrigating crops. A method for making propeller is to use a casting method with sand mold. In operating, the sand mold can only be used for one use, and in the molding process must be reconstructed, which of course, requires cost and additional time, so it is less efficient. Therefore, making a propeller with a permanent mold is the right solution. In this study, the design of the dies propeller production process with CNC (Computerized Numerical Control) machine was designed using the MasterCam software simulation method. In addition, variations in machining parameters feed rate, retract rate, and depth of cut were also carried out using the Taguchi method. The research has known the step of dies propeller production process and the time for cutting that used to calculate the most efficient variable variation. Based on the simulation that has been done, for making dies propeller requires ten steps of processing, there are facing, pocket, contour (top die), rough surface pocket, surface finish contour (top die), surface finish parallel, contour (bottom die), and surface finish contour (bottom die). From a series of work processes designed, the total processing time was produced for top die 729,17 minutes, and cutting power was 9,734 kWh, and for the bottom, die is about 1329,65 minutes, and cutting power was 14,543 kWh. Keywords⎯CNC machines, machining parameters, mastercam, permanent mold, propeller, simulation, taguchi method.


INTRODUCTION
anufacturing industry activities often deal with increasing customer demand. Manufacturing activities are required to develop production methods continue to be more competitive in making products such as sustainable change and increasing the flexibility of the production system. These developments have the main goal of getting a shorter production time, cheaper products, and the best performance of the product. An example of the manufacturing industry is the metal casting industry, and the products produced are propeller.
In the process of casting metal, there are several kinds of molds. Molds are often used are permanent mold (metal molds) and non-permanent mold (sand molds). In general, the metal casting industry uses sand mold (sand casting). In operating sand molds are made manually so that the dimensions are less precise and can only be used for one use and in the moulding process must be reconstructed which of course, requires cost and extra time [1].
Permanent mold (metal mold) is a type of mold that can be used repeatedly, and this mold is usually made of metal material. Using metal molds will produce cast products that are more economical both in quality and quantity. With this permanent mold, even though the initial investment is relatively more expensive, for more production, the use of metal molds will be profitable, and this mold can be used repeatedly. The technology that can be used to produce permanent molds for the propeller is CNC (Computerized Numerical Control) machines [2].
Software used for CNC programming is the MasterCam software. By using MasterCam software, a programmer only needs to draw an object image and plan the machining parameters to be used. After that, the instructions for the process of working the object will be made by the software. These commands will be converted into the numerical language (G-Code), which can be understood by CNC machines and then transferred to CNC machines for execution. But before executing the command directly on a CNC machine, MasterCam software can execute the command in the form of a simulation aluminum so that if there are errors, it can be identified. The simulation aims to reduce the risk of errors and work accidents when the G-code command is executed directly on a CNC machine [3].
Production costs are accumulated costs that occur during the production process of a material. Production cost are influenced by material costs, production processes, and other indirect costs that occur during the production process. This production cost calculation is used in the manufacturing industry to find out the true cost of making a machine component product so that it can be estimated how much profit will be obtained [4].
International Journal of Marine Engineering Innovation and Research, Vol. 5 (1), Mar. 2020. 20-33 (pISSN:2541-5972, eISSN: 2548-1479 21 combined with screws, clamps, plates or other removable tools after the product has hardened. In general, permanent molds are made from close-grain and clamped to one another. This mold is usually coated with heat-resisting wet mixture and soot, which will keep the mold from sticking and reduce the effect of cold on the metal [5].

B. Propeller
Propeller is a rounded blade that rotates in a circle, helping to move a vehicle by pushing against water or air. A speed boat's propeller moves it forward by spinning against the water. Propellers use the laws of physics to propel, or drive forward, an aircraft or a boat. The basic structure of a propeller is a spinning or rotating shaft with wide, curved blades attached to it. The very simplest propellers were first used in ancient Greece, where the inventor Archimedes invented a "screw propeller" that moved water for irrigating crops [6].

C. CNC (Computerized Numerical Control)
Computer Numerical Control, abbreviated as CNC is a machine tool automation system operated by abstractly programmed and stored commands on storage media, this is in contrast to a previous habit in which machine tools are usually controlled by hand spin or simple automation using cam [7]. The word NC itself is an English abbreviation of the word Numerical Control which means "numerical control". The first NC machines were first invented in the 1940s and 1950s, by modifying ordinary machine tools. In this case, ordinary machine tools are added with motors that will move the control wheels include the points entered into the system by the paper recorder. The blend of motor and mechanical servo was soon replaced by analog systems and then digital computers, creating a modern Machine Tool called the CNC Machine (computer numerical control) which later revolutionized the design process. Currently, CNC machines have a very close relationship with the CAD program. CNC machines are built to address the challenges in the modern manufacturing world. With a CNC machine, the accuracy of a product can be guaranteed up to 1/100 mm more, mass product workmanship with exact same results molding and fast machining time [8].

D. CAD/CAM Programming
Computer-Aided Manufacturing (CAM) is a function of computer software to control the machine and its equipment in the work process of the object. CAM aims to make the production process faster, with more precise and consistent dimensions. CAM software is used in computer systems to facilitate programmers in producing complex contour and surface forms (machining of artistic products) without using complicated mathematical formula calculations. In the manufacturing process, a product generally begins with the design concept of the product, then proceeds with making a 3D model using a Computer-Aided Design (CAD) system. After modeling using a CAD system, the product model is transferred to the CAM system [9]. With the CAM system, a path will be made that the cutter (toolpath strategy) will pass through the 3D product that we made. The cutter path in the CAM system will later be used as a cutter path in the actual machining process in a CNC machine with the help of NC; post-processing is done on the cutter location (Cutter Location / CL data) where it defines the cutter path. With a post-processor, the CL data can be converted into a G-Code that can be read by CNC machines [10].

E. Taguchi Method
Taguchi method used in offline quality control activities at the design process stage of production, what is meant by offline quality control, is quality control activities in product design. In the Taguchi method, a matrix called orthogonal array is used to determine the number of minimal experiments that can give information as much as possible of all factors that influence the parameters. The most important part of orthogonal arrays lies in the selection of combinations level of input variables for each experiment [11].

F. Analysis of Variance (ANOVA)
Analysis of variance or ANOVA is a multivariate analysis technique that serves to differentiate the mean of more than two groups of data by comparing the variance. Analysis of variance is included in the statistical parametric category. As a parametric statistical tool, to be able to use the ANOVA formula must first be tested for assumptions including normality, heterogeneityand random sampling [13].
International Journal of Marine Engineering Innovation and Research, Vol. 5 (1), Mar. 2020. 20-33 (pISSN:2541-5972, eISSN: 2548-1479 22 Analysis of variance can be used to analyze data from various types and research designs. Analysis of variance is widely used in many studies involving comparative testing, namely testing the dependent variable by comparing it to the independent sample groups observed [14].

III. RESULT AND DISCUSSION
In doing this thesis, a process must be prepared. There must be planning in the work so that it feels more focused and easier. In the discussion of this research will be discussed step by step, which will be carried out in the execution of the thesis.

A. Model Design
Making dies propeller models using existing data propeller. Modeling is done using SolidWorks software, which is then simulated using the Mastercam software to determine the total production time and cutting power requirement.

B. Machining Process Design (CAM) and Simulation
It was selecting and calculating CNC machining parameters needed, such as toolpath, tools, spindle speed, feed rate, and depth of cutfor the production ofDies Propeller. After determining the machining parameters of the process, a simulation (running) the MasterCam program is performed. It aims to look at the work results in each groove on each toolpath.

C. Calculation of Total Time& Power Requirement
The total time of the production process is the overall processing time on the variations of the tool with the most optimal results and also the time of non-productive machinery. After obtaining the optimal time, we calculated also the power requirement for the work of the object.

D. Analysis of Varian (ANOVA)
Analysis of variance (ANOVA) statistics used to determine the effect of each factor variable (feed rate, retrack rate, and depth of cut) on the response variable (machining process power). The variances analysis process uses Minitab software 18.

A. Model Design
Workpieces that will be processed by CNC milling are dies propeller models. Below is the model of the dies propeller:

B. Mastercam Design Process
The design of the machining process at MasterCam is the stage for determining the type of machining, the selection of boundaries or working area boundaries, and the selection of the right type of work for making dies propeller. In the MasterCam software, each type of work is called the toolpath. For dies propeller production process there are ten tool paths that is facing, pocket, contour (top die), rough surface pocket, surface finish contour (top die), multiaxis roughing, surface finish parallel, contour (bottom die), high-speed surface waterline, and surface finish contour (bottom die).

C. Determination of Machining Parameters
All machining parameters are calculated to get the engine optimization in the form of smoother finishes, more durable tools, and more optimal processing time and power. Machining parameters that need to be determined include: a The results of setting machining parameters can be seen in Table 2 to Table 5.

D. Calculation of Total Time & Power Requirement
The calculation of time and power uses the Taguchi method with a variation of three-parameter combinations that is feed rate, retrack rate and depth of cut with three different levels then simulated in the Mastercam software to analyze the combination that produces the most optimum power response which is then used as a guideline for planning machining parameters.   Table 6 shows the parameter value of the facing process. From table 7 it can be concluded that the optimal power response value in the facing process is in the 5 th combination with feed rate value of 592 mm/minute (level II), re-track rate 7500 mm/minute (level II), and depth of cut 2 mm (level III) with power response value 3,474 Wh and time 20 minutes.  Table 8 shows the parameter value of the rough surface pocket. From table 9 it can be concluded that the optimal power response value in the surface rough pocket process is in the 6 th combination with feed rate value 1184 mm/minute (level II), re-track rate 10000 mm/minute(level III), and depth of cut 1 mm (level I) with power response value 7,9214 kWh and processing time 440,92 minutes.   Table 10 shows the parameter value of the contour process. From table 11 it can be concluded that the optimal power response value in the contour process is in the 5 th combination with feed rate value 1184 mm /minute (level II), re-track rate 7500 mm/minute (level II), and depth of cut 3 mm (level III) with power response value 0,1697 kWh and time 75,56 minutes.  Table 12 shows the parameter value of the surface finish contour process. From table 13 it can be concluded that the optimal power response value in the surface finish contour process in the 3 rd combination with feed rate value 1110 mm/minute (level I), re-track rate 10000 mm/ minute (level III), and depth of cut 1 mm ( level I).
International Journal of Marine Engineering Innovation and Research, Vol. 5(1), Mar. 2020. 20-33  (pISSN:2541-5972, eISSN: 2548-1479) Table 14 shows the parameter value of the surface to finish the parallel process. From table 15 it can be concluded that the optimal power response value in the surface finish parallel process is in the 6 th combination with feed rate value of 711 mm/minute (level II), re-track rate 10000 mm/minute (level III), and depth of cut 1 mm (level 1I) with a power response value of 0,17294 kWh and processing time 16,03 minutes.  Table 16 shows the parameter value of the facing process. From table 17 it can be concluded that the optimal power response value in the facing process is in the 8 th combination with feed rate value 790 mm/minute (level III), re-track rate 7500 mm/minute (level II), and depth of cut 1 mm (level I) with power response value 0,32092 kWh and processing time 44,62 minutes.   Table 18 shows the parameter value of the contour process. From table 19 it can be concluded that the optimal power response value in contour process is in the 8 th combination with feed rate value 1776 mm/minute (level III), re-track rate 7500 mm/minute (level II), and depth of cut 1mm (level I) with power response value of 0,10517 kWh and processing time 156,1 minutes  Table 20 shows the parameter value of the surface rough pocket process. From table 21 it can be concluded that the optimal power response value in the surface rough pocket process is in the 8 th combination with feed rate value 1776mm/minute (level III), re-track rate 7500 mm/minute (level II), and depth of cut 1 mm (level I) with power response value 12,2241 kWh and processing time 756,02 minutes. Vol. 5(1), Mar. 2020. 20-33  (pISSN:2541-5972, eISSN: 2548-1479) Table 22 shows the parameter value of the surface finish contour process. From table 23 it can be concluded that the optimal power response values in the process of surface finish contour are in the 3 rd combination with feed rate value of 1110 mm/minute (level I), re-track rate 10000 mm/minute (level III), and depth of cut 1 mm (level I) with power response value 0.07826 kWh and time 185,86 minutes.  Table 24 shows the parameter value of the surface finish the parallel process. From table 25 it can be concluded that the optimal power response value in the surface finish parallel process is in the 8 th combination with feedrate value 1066 mm/minute (level III), ret-rack rate 7500 mm/minute (level II), and depth of cut 1 mm (level I) with power response value 1,81533 kWh and processing time 187,05 minutes.

E. Analysis of Variance (ANOVA)
From the results of simulations that have been carried out then proceed with analysis using variance analysis according to parameter factors, level, and processing of data based on ANOVA statistics to determine the effect of each factor variable (feed rate, re-track rate and depth of cut) on the response variable (machining process power). The variance analysis process software 18.
Results of Analysis of Variance:  Table 27, it can conclude several things. If the value of P>0.05 means that there are no significant factors that influence the power response. In the facing feed rate process, and the depth of cut has a value of P <0.05 means that it significantly affects the power response. In the facing process the greatest F value of the three factors in the experiment is feed rate, which is 41,81 which means that feed rate has the greatest influence on the power response, then the second depth of cut with F = 40,85 and the last re-track rate with F = 2,25.  Table 28it can conclude several things. If the value of P >0.05 means that there are no factors that most influence the power response. In the contour, all factors have a value of P >0.05, which means there are no factors that significantly affect the power response. In the contour process, the greatest F value of the three factors in the experiment is a depth of cut, which is 10.36, which means that depth of cut has the greatest influence on the power response, then the second is feed rate with F = 9.71 and the last re-track rate with F = 1.09.  Based on table 29, it can conclude several things. If the value of P >0.05 means that there are no significant factors that influence the power response. In the surface rough pocket process, all factors have a value of P >0.05, which means there are no factors that significantly influence the power response. In the surface rough pocket process the greatest F value of the three factors in the experiment is the feed rate which is 9.61 meaning that the feed rate has the greatest influence on the power response, then the second is the depth of cut with the value F = 7.51 and the last re-track rate with the value F = 1.27.  Table 30it can conclude several things. If the value of P >0.05 means that there are no significant factors that influence the power response. In the surface finish contour process, the depth of cut has a value of P <0.05 means that it significantly affects the power response. In the surface finish contour process, the greatest F value of the three factors in the experiment is the depth of cut, which is 37,55 which means that depth of cut has the greatest influence on the power response, then the second feed rate with F = 7,60 and the last retrack rate with F = 1,27.  Table 31, it can conclude several things. If the value of P >0.05 means that there are no significant factors that influence the power response. In the process of surface finish parallel feed rate has a value of P <0.05, which means that it significantly affects the power response. In the process of surface finish parallel, the highest value of F from the three factors in the experiment is feed rate which is 11.31, meaning that feed rate has the greatest influence on power response, then the second depth of cut with F = 8.74 and the last retrack rate with F = 3.63.
International Journal of Marine Engineering Innovation and Research, Vol. 5 (1), Mar. 2020. 20-33 (pISSN:2541-5972, eISSN: 2548-1479 Table 32, it can conclude several things. If the value of P >0.05 means that there are no factors that most influence the power response. In the facing, all factors have a value of P >0.05, which means there are no factors that significantly affect the power response. In the facing process, the greatest F value of the three factors in the experiment is feed rate, which is 11.31, which means that feed rate has the greatest influence on the power response, then the second is a depth of cut with F = 8.74 and the last re-track rate with F = 3.63.  Table 33, it can conclude several of things. If the value of P >0.05 means that there are no significant factors that influence the power response. In the contour process, all factors have a value of P >0.05, which means there are no factors that significantly affect the power response. In the contour process, the highest value of the three factors in the experiment is the feed rate, which is 7.13. It means that feed rate has the most influence on the power response, then the second depth of cut with F = 6.88 and the last re-track rate with F = 0.81.

31
Based on Table 34, it can conclude several of things. If the value of P >0.05 means that there are no significant factors that influence the power response. In the surface rough pocket process, all factors have a value of P >0.05, which means there are no factors that significantly influence the power response. In the surface rough pocket process, the highest F value of the three factors in the experiment is feed rate, which is 10.36, which means that feed rate has the most influence on the power response, then the second depth of cut with F = 9.18 and the last re-track rate with F = 0.98.   Table 35 it can conclude several of things. If the value of P >0.05 means that there are no significant factors that influence the power response. In the process of surface finish contour depth of cut has a value of P <0.05 which means that it significantly affects the power response. In the process of surface finish contour, thehighest value of F from the three factors in the experiment is depth of cut which is 37.23, meaning that depth of cut has the greatest influence on power response, then the second feed rate with F = 9.39 and the last re-track rate with F = 1.00.  Table 36, it can conclude a number of things. If the value of P >0.05 means that there are no significant factors that influence the power response. In the process of surface finish parallel, all factors have a value of P >0.05, meaning that there are no factors that significantly affect the power response. In the surface finish parallel process, the highest F value of the three factors in the experiment is the depth of cut, which is 7.45. It means that the depth of cut has the greatest influence on the power response, then the second is feed rate with the value F = 7.37 and re-track rate with F = 0.85.

A. Conclusion
Based on the simulation results of the production of dies propeller in the MasterCam software with several combinations of machining parameters, obtained several conclusions to answer the formulation of the problem on this paper, as follows: 1. In the dies propeller production simulation, five steps are performed; there are facing, contour, rough surface pocket, surface finish contour and, surface finish parallel. That 5 processes are used for both top and bottom dies.
2. From the simulation results that have been carried out, the total production time of dies propeller is obtained, which is 729,17 minutes with power consumption 9.734 kWh for top dies and 1329,65 minutes with power consumption 14,54 kWh for bottom dies.