Int J Adv Manuf Technol (2007) 34:1062–1071 DOI 10.1007/s00170-006-0675-3

ORIGINAL ARTICLE

Integrated rough machining methodology for centrifugal impeller manufacturing Li-Chang Chuang & Hong-Tsu Young

Received: 22 June 2005 / Accepted: 2 May 2006 / Published online: 18 August 2006 # Springer-Verlag London Limited 2006

Abstract In mechanical engineering, most products or components, especially those for aerospace applications, are designed to fit the requirements of free-form surface features. The impeller often required by 5-axis machine operations is a key component of the aerospace industry. When 3-axis CNC machining center is used to manufacture the impeller, great difficulties, i.e., collisions between the cutting tool and impeller, need to be overcome. Presently most commercial CAM systems for 5-axis control lack generality, and functions for the rough tool-path generation are far from sufficient. Although the rough machining is the most important procedure influencing the machining efficiency and the condition for the following finishing process, many difficulties arise in performing 5-axis rough machining. The main objective of the present study is to overcome this problem by integrating the state-of-art machining technology, and consequently effective rough tool-paths are to be generated. This study aims to implement the algorithm of the constant scallop height method to improve tool-path planning of rough machining. As a result CL data based on the geometry model of blade and hub are generated. The CL data are confirmed by comparing them with original CAD model through software simulations and later by machining experiments. The verification results show that the machining methodology and procedure adopted turn out to be a successful case. Keywords Centrifugal impeller . Tool-path planning . 5-axis Machine . 5-axis rough machining . 5-axis machining L.-C. Chuang (*) : H.-T. Young Department of Mechanical Engineering, National Taiwan University, Taipei, Taiwan e-mail: [email protected]

1 Introduction 5-axis computer numerical controlled (CNC) machines are widely used to manufacture aerospace parts, turbine impellers, and certain special dies. These parts usually have complex geometry and are represented by parametric or freeform surfaces. The centrifugal impeller gives a perfect demonstration of the efficient designing and manufacturing capabilities of 5-axis machining. The surface model of centrifugal impeller is designed with an extremely twisted surface and its blades highly overlap with each other. If traditional 3-axis machining is implemented, serious collisions between the cutting tool and the blades of centrifugal impeller are frequently encountered. For such geometric complexity, it is the general practice to adopt 5-axis machining. An improvement over 3-axis machining, 5-axis machining offers advantages such as higher productivity and better machining quality. The tool axis in 5-axis machining has two additional degrees of freedom, allowing the tool to take arbitrary posture as well as position. Furthermore, 5-axis machining can improve working precision, efficiency and quality of the machined surface to satisfy various requirement of product design. Due to the high cost of 5-axis machines and the trend of increasingly complicated relative positions between the cutting tool and workpiece, it is important to develop an algorithm that gives effective toolpaths and, subsequently, correct cutter location data (CL data). These are very important tasks for manufacturing an impeller and form major focuses for the present study. To shorten the product development process and make production more competitive, computer-aided design (CAD) and computer-aided manufacturing (CAM) are implemented in the manufacturing process. There are some CAM systems with 5-axis control capability to finish a centrifugal impeller; however, functions for efficient rough

Int J Adv Manuf Technol (2007) 34:1062–1071

machining are not available. Since the rough machining is the key process influencing the total machining time and accuracy of finished product, it is high on the agenda to discuss this process. In previous research [1], Young and Chuang propose a 5-axis rough machining approach for a centrifugal impeller. They use the tool-path of flank milling as the base tool-path, then rotate the tool-path around the central axis with a specific angle. But some tool-paths are redundant. The study tries to plan tool-path with constant cutting depth, and the algorithm of constant scallop height method is implemented to improve tool-path planning of rough machining. Meanwhile the purpose of this research is to develop a reformed 5-axis rough machining module dedicated for centrifugal impeller manufacturing. For impellers with similar geometric characteristics, the reformed module developed in this research can be used to plan the tool-path of rough machining effectively, which shortens the lead-time for manufacturing and eventually reduces the total manufacturing cost. This study has been validated ultimately for its usefulness.

2 Theoretical model Figure 1 illustrates a typical centrifugal impeller and its main parts. The centrifugal impeller itself is a circular revolving entity. It is composed of 15 identical blades and a hub. The angle variation between two blades is 24°. Usually, after a preliminary calculation is performed with aerodynamics and fluid mechanics, the hub curve is obtained by using the optimization algorithm. Rotating the hub curve around the central axis forms the hub surface. The geometric model of the blade is made up of the suction surface, pressure surface, leading edge and trailing Edge. While the shroud surface defines the outer boundary of the blade, it is formed by rotating the shroud curve around the central axis. There are a few related studies on the rough machining of centrifugal impellers. Morishige [2] shows that collision-free rough machining is generated on the basis of a 3-dimensional geometric model along with a proposed machining strategy, as shown in Fig. 2. The 2-dimensional configuration space (C-Space) defined by two parameters [3] is applied to obtain the relationship between all tool postures and the existence of collision, and the optimum tool posture at any point on the surface to be machined is found. If the above method is employed in the case of the centrifugal impeller, as shown in Fig. 1, serious collisions are not totally avoided due to the high overlap between the cutting tool and the extremely twisted blade surface. As a result, more interference checks are then necessary to correct the final tool axis. Chuang [4] presents a rugged rough cutting method for B-Spline surfaces. Several original B-Spline surfaces are first constructed, and each one is then decomposed

1063

Fig. 1 Surface model of centrifugal impeller

individually into Bezier surfaces. All of the Bezier convex hull boxes are united together to form an approximate model. A rough planning algorithm is applied to the approximate model by slicing the stock into non-uniform layers. In this way, rough tool-paths are generated automatically by treating each layer as a 2D pocket die cavity. Other related rough machining researches have already been reported [5–7] with a similar approach. The iso-parametric method is frequently used to plan toolpath. It means that each tool-path is generated along the direction of each parameter. More than often the planned tool-path will be highly concentrated in the narrow area of the curve surface. It is obvious that the effectiveness of machining is to be improved in this case. To reduce the amount of tool-path data, Suresh and Yang [8] calculate the tool-path by controlling the residual scallop height, and

Fig. 2 Morishige’s rough machining model

1064

the interval between the tool-path is obtained and expressed by the cutter diameter, scallop height and radius of curvature. Suresh and Yang find that this method can reduce the amount of tool-path data, but there are still problems to be resolved in real applications. Lin and Koren [9] propose a planning method for effective tool-path. They use an offset of the tool-path, which guarantees that the cutter will move in an unmoved area of the part surface, without redundant machining. Their research also comprises the investigation of the accurate tool-path interval and its conversion into the parametric interval. The derivation of the maximum CC path interval is based on the ball-mill cutter in the 3-axis machining. There will be errors on curve surface caused by the residual scallop-shaped area between each tool-path. Also, there will be a chord error along the moving direction of cutter. The chord error will affect the amount of cutting points needed for each tool-path. To have the error under control, it is necessary to evaluate the distance between the straight-line portion of tool-path and the curve of the workpiece. In dealing with the issue, Loney and Ozsoy [10] utilize numerical analysis method to calculate the parametric value and base on the result to determine if it is necessary to add extra cutting points. Bohez [11] discloses a scheme for machining a centrifugal impeller. This research concerns flank milling to machine the blade, represented by a ruled surface. The desired tool-path on the hub surface is generated by rotating the tool-path of flank milling around the central axis with an appropriate angular interval. But, rough machining, a crucial part in the machining of centrifugal impellers, is not covered in this research. You and Chu [12, 13], with focus on avoiding interference problem on 5-axis machining, describe utilizing checking surface representing the interfering surface to correct the tool-axis. This present study adopts the method to generate the collision-free tool-path. The above research mostly focuses on an individual machining issue. No provision is given as an integrated 5-axis rough machining methodology on the centrifugal impeller. This research focuses on the rough tool-path planning for narrow and deep machining areas, which resemble a deep die cavity. The rough machining technology for related centrifugal impellers, a special feature of this research, is integrated for 5-axis CNC machining.

3 Process planning for rough machining Process planning of 5-axis rough machining is described in the following. First to be discussed are the problems encountered in rough machining, and then the algorithm is given for planning the tool-path on the hub surface. The process of rough machining is to divide the narrow and deep openings between two blades into several machining sections, which are similar to the hub surface. Next it is necessary is to obtain

Int J Adv Manuf Technol (2007) 34:1062–1071

the cutter contact points, the cutting tool-path on each machining section, and finally connect all the rough tool-paths. The result that gives a complete manufacturing process for rough machining of impeller is then obtained by a designated manner. What follows are the details of each machining step and the result of associated machining data calculation. 3.1 Strategies for rough machining Rough machining is the main process that removes the major part of material from the blank and carves the preform into the rugged profile of an impeller. An increase in the material removal rate to raise the overall working efficiency is a key issue in designing the rough machining module. It influences not only the total machining time but also the accuracy of the resulting impeller in the finish process. Furthermore, the residual thickness and surface conditions after rough machining will affect the final finish machining. If the residual thickness is too large, it will certainly lower working efficiency and result in excessive tool wear. The residual surface conditions also influence the tool life. If some burrs are left on the workpiece, they might need further semi-finishing and even cause tool breakage. To solve the above problems, this study provides a machining module with which users could input machining parameters, so as to control the scallop height of the toolpath, cutting depth and the residual, for impeller production with high efficiency and high accuracy. The centrifugal impeller is typically designed into complicated shapes with overlapping parts that require 5-axis machining. In 5-axis machining, the tool is allowed to take an itrary posture as well as the position to form the required surface. Consequently in order to make good use of the special geometric features, this study proposes a new machining methodology allowing the cutter to remove the materials along the ruling line of blades from the shroud surface to the hub surface. As a result the residual materials always remain on the bottom between two neighboring blades. It is seen that the residuals provide extra support to the blades and, hence, reduce the probability of chatter during machining. This study tries to carve the narrow and deep opening efficiently, the cutting depth in each machining sections is constant in the proposed methodology, and the pocket die cavity is divided into several machining sections according to its relative dimension. In planning the tool-path for each machining section, this study uses the algorithm of tool axis determination adopted in flank milling on blades, the rough tool-paths in each machining sections are generated by constant scallop height method. By this method, the generation of all rough tool-paths is achieved, and this study could prevent collisions from happening. The following is a detailed description of the rough machining methodology proposed in this research.

Int J Adv Manuf Technol (2007) 34:1062–1071

1065

3.2 Constant scallop height tool-path Suresh and Yang [8] provide a systematic approach for the tool-path generation of free-form surfaces, maximizing the side step along the entire tool-path by constant scallop height. The cutter contact (CC) path is the tangential trajectory between the ball-end cutter and machining surface, and is shifted from the tool-path by a distance equal to the cutter radius. The CC path interval, the distance between the parallel trajectories, depends on the local curvature of the surface, the radius of the cutter, and the allowable scallop height remaining on the surface. By demanding that the scallop height remains as a constant, the CC path interval can be determined as the distance that the cutter can slide without exceeding the allowable surface finish value. The tool-path interval can be obtained by offsetting the CC path in the surface normal direction by a distance equal to the cutter radius. Lin and Koren [9] calculate the geometric relationships when machining a flat plane, a convex surface, and a concave surface with a ballend cutter. For a given allowable scallop height the CC path interval can be obtained: Machining a convex surface sffiffiffiffiffiffiffiffiffiffiffiffiffi 8hrRc P
Rc þ r

ð4–1Þ

Machining a flat plane qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffi P ¼ 2 r2  ðr  hÞ2
2 2rh

ð4–2Þ

ð4–3Þ

Where P denotes the tool-path interval, r denotes the radius of a ball-end cutter, and h denotes the allowable scallop height. The assumption in the approximation is that h<
P sin θ

Where Pp is the path interval in the parametric direction and P is the path interval in the orthogonal direction is the angle between the tool-path and the parametric curve. Next, in order to place the tool on the calculated path, a conversion from Pp to the parametric domain Δu is needed. This conversion, which uses the Taylor expansion, is ρ presented below. Given a parametric curve C ðuÞ, O≥u≥1, which is used to determined the CC path interval, the Taylor series expansion of the parametric curve is ρ ρ ρ 1ρ C ðuÞ ¼ C ðu0 Þ þ C 0 ðu0 ÞΔu þ C 00 ðu0 ÞΔu2 2

1 ρ000 C ðu0 ÞΔu3 þ Λ ð4–5Þ 3! ρ where C ðuÞ ¼ xðuÞbi þ yðuÞbj þ zðuÞb k, and Δu ¼ u  u0 . If we neglect the higher order terms, the path interval can be derived below, p  ρ   Pp ¼ C ðuÞ  C ðu0 Þ ρ   0  1 ρ 00 2  ð4–6Þ ¼ C ðu0 ÞΔu þ C ðu0 ÞΔu  2 þ

Machining a concave surface sffiffiffiffiffiffiffiffiffiffiffiffiffi 8hrRc P
Rc  r

Fig. 3 The angular difference between the tool_path and the parametric direction

ð4–4Þ

or

Pp2 ¼ AΔu4 þ BΔu3 þ DΔu2 where " 2  2 2  2 2 # 1 d2 d y d z A¼ þ þ 2 2 4 du du du2 

2

2

2



u¼u0

dx d x dy d y dz d z    þ þ du du2 du du2 du du2 u¼u0 "   2  2 # dx 2 dy dz D¼ þ þ du du du B¼

ð4–7Þ

u¼u0

To speed up the process, we introduce an error-compensation method to solve Δu in Eq. (4–7). The first order Pp approximation of Eq. (4–7) is Δua ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 2 2 ðdudxÞ þðdudyÞ þðdudzÞ u¼u0

1066

Int J Adv Manuf Technol (2007) 34:1062–1071

The above equation is valid if Δu is very small, an error indicator can be defined as " ¼ Δu  Δua ¼ AΔu4 þ BΔu3

ð4–8Þ

Substituting Eq. (4–8) back into Eq. (4–7), a more accurate conversion from CC path interval to the parametric interval is obtained. qffiffiffiffiffiffiffiffiffiffiffiffiffi Pp2  " Δu ¼ rhffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð4–9Þ

dx 2 dy 2 dz 2 i þ du þ du du u¼u0

Figure 4 illustrates the generating of the next tool-path based on accurate tool-path intervals. Along the first toolpath, a non-constant offset tool-path can be found by calculating the path interval according to the constant scallop height. 3.3 Determination of tool axis Before carrying out rough tool-path planning, an algorithm for flank milling needs to be clearly outlined to obtain the tool axis. The algorithm for flank milling on the blade surface is the basis of planning the rough tool-path. The initial tool axis is selected to make it parallel to the ruling line on the blade surface. In essence the cutter needs to be offset at a distance of the cutter radius. In considering the cutter strength, the tapered cutter is often used for enhancing to the required degree. Before the final tool axis is determined, the cutter is offset along the same direction at both the inner and outer edge of the ruled surface. The respective displacements will not be the same due to the tapered cutter dimensions. Since the initial tool axis is parallel to the ruling line, there will be overcut on the ruled surface. The maximum overcut occurs at the outer edge of the ruled surface because the angle between normal at the other edge of the ruled surface and the direction of the cutter offset is the maximum.

In order to investigate the maximum cut error, the initial tool axis is used as the normal of the examination plane. Meanwhile the ruled surface is properly divided into a collection of ruling lines. The intersections between the examination plane and each ruling line are made and the distance from each intersection to the tool axis is evaluated. After comparing all the values, the location marking the shortest distance between the tool axis and intersection is then identified. If the distance is smaller than the cutter radius, overcut is generated. Otherwise, there is undercut where the distance is greater than the cutter radius. Subsequently, the cutter is moved to tangent the ruled surface on the outer edge and the final tool axis is obtained, as shown in Fig. 5. 3.4 Algorithm of rough machining The tool axis is determined from flank milling and is adopted to plan rough machining. The next step is to determine the machining sections in the narrow opening between two blades. Meanwhile the machining sections are similar to the hub surface. Furthermore, the machining sections separate the narrow opening analogously and individually. Before determining the machining sections, the boundary curve on the blade surface should be decided at first. Consequently, a boundary curve on the blade surface could be decided by taking an equal interval from the hub surface, and other boundary curves could be decided by the same way. As referred to Fig. 6, it provides a more detailed description for generating the boundary curves on blade. Generally, a surface can be represented by u-v parameters. In order to facilitate boundary curve calculation, the u-v parameters are utilized to define the boundary points, which form the boundary curve. Figure 6 shows the way to determine the boundary points on the blade surface. The boundary curve 0 is the intersection of the blade and the hub, which is the bottom of the blade. The first boundary points are obtained from the boundary curve 0 according to the cutting depth that the user input. The boundary points are defined from the u-parameter of constant interval, and those points form the boundary curve 1. Then the boundary curve 2 is obtained in the same way.

P2 P1

Outer Edge of Surface Fig. 4 Illustration of generating the nelling

Inner Edge of Surface Modified Vector

Fig. 5 Tool axis planning in flank milling

Int J Adv Manuf Technol (2007) 34:1062–1071

Fig. 6 Determination the boundary curves and boundary points on blade surface

Centrifugal Impeller itself is a circular revolving entity. It is composed of fifteen identical blades and a hub. Therefore, the actual surfaces of machining sections are obtained by rotating the boundary curves around the central axis. The surface A of the machining section, shown in Fig. 7, is obtained by the boundary curve 1. The surfaces of the machining sections look like the hub surfaces with similar blades. In each surface of the

Fig. 7 Surface of machining section

1067

machining section, the tool-path of the flank milling on the two adjacent blades is adopted as the two remaining boundaries of the machining section. The tool-path of flank milling is a new boundary on the surface of the machining section. This will reduce the redundant toolpath between blade machining and rough machining, and prevent overcut. In this case, however, the tool-path of flank milling on the blade is a prerequisite. The machining sections trimmed by the tool-path of blade machining are rather irregular. There is an extremely narrow area in the middle and upper part of the bottom machining section. The narrow opening gives the whole machining section a funnel shape. The lower portion shows an area with an extremely wide difference in width between the upper and lower part. An uneven surface with low cut efficiency is cased when using iso-parametric method. Therefore, the non-constant parametric method is a better choice for toolpath planning, which is described in Chap. 4-2. A one-way cut strategy is adopted in this research for tool-path planning on each machining section. The working boundaries of the machining section have been previously defined. It takes the working boundaries as the first toolpath. This study chooses the flank milling tool-path on the right blade as the initial base-path, shown in Fig. 8. The rough tool-path is generated from the initial base-path, based on the constant scallop height and the allowable

Fig. 8 Tool_path in a machining section

1068

residual, and the next tool-path is calculated with the same method until this path runs into the working boundary on the opposite side. Due to the irregular shape of the machining section, once the tool-path is available, it is necessary to check if it exceeds the working boundary. Bad tool-paths should be eliminated at the end of the process. The tool-paths generated in each machining sections are then linked in a zig-zag style. For the planning of the tool axis, an interpolation scheme is adopted. The thickness of the blade bottom is greater than that of the blade top. If a plane is used to cut two consecutive blades, the cross sections of the blades and hub surface is found close to the “V” shape. Since the blade itself is a revolving component, it should be geometrically feasible if the interpolated method for the planning of tool axis is adopted. Therefore, the initial tool axis for rough machining is given by using the prior flank milling on the blades at the two sides with linear interpolation. An interference check is then made to check the final tool axis. Finally, the complete rough tool-paths are achieved efficiently by connecting each cutting tool-path on different machining sections, from the shroud surface to the hub surface. This study provides an integrated method with approximate constant cutting depths for rough machining on the centrifugal impeller. All machining methodology is formularized to a rough machining module. It is found that there is no overlap between each of the tool-paths, and better machining efficiency is achieved as a result.

4 Machining simulation and verification The machining module was developed with appropriate user interfaces using C++ language as the programming implement. The flow chart of the C++ machining module is shown in Fig. 9. The impeller surface model is shown in Fig. 1, where the hub and blades are displayed. The angle between two blades is 24° (there are 15 blades for a complete impeller). The maximum tool diameter is deter-

Fig. 9 Flow chart of the C++ machining module

Int J Adv Manuf Technol (2007) 34:1062–1071

mined by the space between two adjacent blades. Larger tool diameters will result in a smaller number of tool-paths. If the tool diameter is too small, the tool will require higher spindle speeds and has a greater tendency to break. In comparison tests, it is seen that a taper ball-mill of 2.5 mm, with a free length of 30 mm and the taper angle of 3°, is a better choice in practice. The modules proposed in this study generate several kinds of rough tool-paths with different sets of machining parameters. The machining parameters selected in this study are as follows: the desired chord error and scallop height of tool-path is 0.1 mm, the cutting depth is 0.8 mm and the allowance is 0.3 mm. The following machining simulation is made with the software package Anvil Verify [14]. The rough machining simulation between two neighboring blades is shown in Fig. 10. No interference is found between the tool and the two blades during machining. This also confirms the feasibility of interpolating the tool axis and dividing the narrow and deep opening into several machining sections. The complete rough machining of the impeller is obtained by revolving tool-paths at an appropriate angular interval around the central axis. 4.1 Verification of machining data After making the rough machining simulation, it is essential to verify the machining results in detail. The work comprises the profile and surface condition of residues on the blade after rough machining, and the analysis of residual thickness. This present study applies certain tools to simulate and analyze the tool-path repeatedly, so as to make sure that there is no interference on the impeller. Respective work includes: (1) Investigation of the profile and surface condition of residues. After machining simulation, the residue data are transformed to STL format. Both STL file and original surface model are input into Pro/E software package. Figure 11 shows that the surface topography is very uniform and the residual profile corresponds well with the impeller. The rugged marks on the hub surface are caused by the step-over of tool-path. Consequently, the machining methodology applied is deemed feasible. The semi-finishing process is found redundant and finishing machining can be implemented directly as a results of fine surface condition received from the rough machining. (2) Analysis of residual thickness. This study uses the VeriCut software package to further verify the machining data. The residue and original surface models are compared numerically, and typical residual thickness is shown in Fig. 12. When the data are anaylzed, it is found that the materials are cut at a uniform

Int J Adv Manuf Technol (2007) 34:1062–1071

1069

roughness, and that the residual thickness on the blade is between 0.2 and 0.3 mm, and its corresponding size on the hub is 0.3 mm. The result of residual thickness corresponds well with the preset finish cut allowance, which shows that the tool-paths generated by this machining module can achieve the goal with satisfactory precision.

Fig. 11 Profile and surface condition of residue

4.2 Feasibility of machining modules The tool-path generated in this study could be compared with the tool-path developed in the previous research [1]. A ball-mill is selected in the previous study, the radius is 1.5 mm, with a shank length of 30 mm, step-over of 1.2 mm, cutting depth of 0.8 mm and residual of 0.3 mm. The length of the previous tool-path between two blades is 9241 mm. It is found that the length of tool-path is

Fig. 10 Machining simulation made with the software package Anvil Verify

Fig. 12 Analysis of residual thickness

1070

shortened to 5367 mm by the taper cutter in this study. Meanwhile, one of the most significant concepts is that the same taper cutter could be adopted in the fine machining of the centrifugal impeller, and the taper cutter provides better toughness. The tool-path developed in the previous study is shown in Fig. 13. The figure only shows parts of the tool-path. While the rough machining area is divided into three portions, it is obvious that the tool-path is irregular and some tool-paths overlap with other tool-paths. The previous study uses the tool-path of flank milling as the base toolpath, then rotates the tool-path around the central axis with a specific angle. If the rough machining area is not similar to the parallelogram, some residue would not be machined or even the tool-path would overcut the blades. The toolpath generated in the previous study needs to be inspected carefully. It results in inefficient tool-path planning and longer machining time. Referring to Fig. 8, the tool-path is smoother and more efficient. Therefore, the method implemented in this study provides a great improvement in rough machining for centrifugal impeller. The rough machining module combines the module of blade machining and hub machining, and generates the rough tool-path and finish tool-path at the same time. The tool-path of rough machining is similar to the tool-path in the finishing. The roughing is integrated with the finishing.

Int J Adv Manuf Technol (2007) 34:1062–1071

98 mm in diameter and 48 mm in height was used as the blank. An external feature that represents the shroud surface of the impeller was first machined from the blank on a lathe. Then the workpiece was cut on a Chiron of model FZ 15 S 5-axis CNC milling machine. The maximum spindle speed is 20000 rpm and the feed rate is 1600ipm. Figure 14 a–c show

4.3 Machining experiments Finally the CL data generated by the machining module were tested through machining experiments. Aluminum of

Fig. 13 Tool_path calculated in the previous study

Fig. 14 Rough machining experiment machined with aluminum

Int J Adv Manuf Technol (2007) 34:1062–1071

respectively the actual rough machining in progress and its final the shape after machining. In the experiments, no collisions were ever detected and impellers with fine surface finish were produced.

5 Conclusion The objective of this study is to propose a method to effectively improve the rough machining of impellers by 5-axis machining, and a comparison is made with the previous research [1]. A main module developed with C++ language automatically generates a rough tool-path for centrifugal impellers and geometrically similar models. Various considerations have been taken to avoid interference, to shorten the lead-time of manufacturing, and to reduce the total manufacturing cost with high accuracy. The geometrical characteristics of centrifugal impellers are studied in this research and a suitable machining methodology introduced accordingly. Rough machining technique and 5-axis machining technology related to centrifugal impellers are integrated as a complete production process planning. Meanwhile, the CL data generated in this research are confirmed through software simulation. Finally, the results of verification in actual machining prove that the machining methodology and procedure applied are both useful and successful.

References 1. Young HT, Chuang LC, Gerschwiler K, Kamps S (2004) “The 5-axis rough machining approach for centrifugal impeller”. Int J Adv Manuf Technol 23(3&4):233–239

1071 2. Morishige K, Takeuchi Y (1997) “5-Axis control rough cutting of an impeller with efficiency and accuracy”. Proceedings of the 1997 IEEE International Conference on Robotics and Automation, pp 1241–1246 3. Lozano-Perez T (1983) “Spatial planning: a configuration space approach”. IEEE Trans Comput 32(2):108–120 4. Chuang SH, Pan CC (1998) “Rough cut tool path planning for Bspline surface using convex hull boxes”. Int J Adv Manuf Technol 14:85–92 5. Dong Z, Li H, Vicker GW (1993) “Optimal rough machining of sculptured parts on a CNC milling machine”. Transactions of the ASME 115 6. Trochu F, Abong Y, Balazinski M, Larbrisseau P (1996) “Rough machining of free-from surfaces defined by dual kriging”. INT J Prod Res 34(6):1603–1623 7. Huang YS, Webster PD, Dean TA (1996) “An image detection approach to NC rough-cut milling from solid models”. INT J Mach Tools Manufacture 36(12):1321–1333 8. Suresh K, Yang DCH “Constant scallop-height machining of free-form surfaces. Transactions of the ASME Journal of Engineering for Industry 116:253–259 May. 1 9. Lin RS, Koren Y (1996) “Efficient tool-path planning for machining free-form surfaces”. Transactions of the ASME Journal of Engineering for Industry 118:20–28 Feb 10. Loney GC, Ozsoy TM (1987) “NC machining of free form surfaces”. Comput aided des 19(2):85–90 11. Bohez ELJ, Senadhera SDR, Pole K, Duflou JR, Tar T (1997) “A geometric modeling and five-axis machining algorithm for centrifugal impellers”. J manuf syst 16 (6):422–436 12. You CF, Chu CH (1996) “Automatic Correction of Tool Interference in Five-Axis NC Machining of Multiple Surfaces”. Journal of the Chinese Society of Mechanical Engineers 17:435–442 13. You CF, Chu CH (1997) “Tool-path verification in five-axis machining of sculptured surface”. Int J Adv Manuf Technol 13:248–255 14. MCS Inc., “ANVIL EXPRESS 5.0,” 1999 15. CGTech, “VERICUT Version 4.1,” 1999 16. Young HT, Chuang LC (2003) “An integrated machining approach for a centrifugal impeller”. Int J adv manuf technol 21 (8):556–563