key: cord-0324804-7nd6f9f8 authors: Olvera, D.; Artetxe, E.; Luo, M.; Urbikain, G. title: 5-axis milling of complex parts with barrel-shape cutter: cutting force model and experimental validation date: 2020-12-31 journal: Procedia Manufacturing DOI: 10.1016/j.promfg.2020.05.079 sha: 64d65424e61cc9080154f92efcce3f1746a7ee72 doc_id: 324804 cord_uid: 7nd6f9f8 Abstract The manufacturing of turbomachinery components for the aerospace sector is an important topic in machining research. Aeroengine manufacturing companies forecast increasing levels of design complexity and demand in the following years. Key elements are the Integral Bladed Rotor (IBR) parts, such as impellers and blade-disks. Despite advances in five-axis machining, manufacturing a reliable and efficient IBR component is still a challenge. In this work, the mechanics of the milling process with barrel-shape cutter is studied. These recently developed tools offer a potential advantage for a productive, chatter-free, high-performance machining of IBR parts. In this work, the cutting forces model is developed for this type of tools, then the model is verified in inclined milling operations. Main aircraft manufacturers predict to double the fleet of airplanes over the next 15-20 years [1] . This leads to a higher demand for elements used in turbine engine which need to be efficiently machined. Particularly, the milling of rotary blade rotors (IBRs) is an important topic of research. Production of IBRs is an intricate process where complex surfaces in 5 continuous axes need to be machined, often made of heat resistant super alloys (HRSA). Due to low machinability of HRSA, accelerated tool wear and poor surface quality are the usual trend. Furthermore, during finishing operations, thin and complex walls with variable vibration modes may lead to unstable conditions (chatter). Finding an optimal combination of all these factors is a challenge for process development engineers and NC programmers. The lack of a scientific approach to solve this problem leads to conservative and unproductive cutting parameters, increasing cycle time and reducing throughput. To be competitive, improving the efficiency of the manufacturing process of these high added value components is mandatory. The effective and accurate calculation of cutting forces in complex milling is a hot topic as these models give insight for the estimation of workpiece quality (surface roughness and surface location error), tool/workpiece relative vibrations and tool life, to mention a few. In a pioneer work, Bouzakis [2] exemplified the complete diagnosis of a force model, for the special case of milling with a ball-end mill. Ferry and Altintas [3, 4] presented the generalized milling model for a variety of mill-geometries and demonstrated its usefulness in roughing operations. They also proposed a feed optimization algorithm to minimize cutting forces. Budak et al. [5] investigated 5-axis milling operations with ball-end mills, being capable of stability analysis with different tool/workpiece orientations. Wojciechowski et al. [6] analyzed cutting forces and vibrations particularly for ball-end milling in hardened steels. They Main aircraft manufacturers predict to double the fleet of airplanes over the next 15-20 years [1] . This leads to a higher demand for elements used in turbine engine which need to be efficiently machined. Particularly, the milling of rotary blade rotors (IBRs) is an important topic of research. Production of IBRs is an intricate process where complex surfaces in 5 continuous axes need to be machined, often made of heat resistant super alloys (HRSA). Due to low machinability of HRSA, accelerated tool wear and poor surface quality are the usual trend. Furthermore, during finishing operations, thin and complex walls with variable vibration modes may lead to unstable conditions (chatter). Finding an optimal combination of all these factors is a challenge for process development engineers and NC programmers. The lack of a scientific approach to solve this problem leads to conservative and unproductive cutting parameters, increasing cycle time and reducing throughput. To be competitive, improving the efficiency of the manufacturing process of these high added value components is mandatory. The effective and accurate calculation of cutting forces in complex milling is a hot topic as these models give insight for the estimation of workpiece quality (surface roughness and surface location error), tool/workpiece relative vibrations and tool life, to mention a few. In a pioneer work, Bouzakis [2] exemplified the complete diagnosis of a force model, for the special case of milling with a ball-end mill. Ferry and Altintas [3, 4] presented the generalized milling model for a variety of mill-geometries and demonstrated its usefulness in roughing operations. They also proposed a feed optimization algorithm to minimize cutting forces. Budak et al. [5] investigated 5-axis milling operations with ball-end mills, being capable of stability analysis with different tool/workpiece orientations. Wojciechowski et al. [6] analyzed cutting forces and vibrations particularly for ball-end milling in hardened steels. They Main aircraft manufacturers predict to double the fleet of airplanes over the next 15-20 years [1]. This leads to a higher demand for elements used in turbine engine which need to be efficiently machined. Particularly, the milling of rotary blade rotors (IBRs) is an important topic of research. Production of IBRs is an intricate process where complex surfaces in 5 continuous axes need to be machined, often made of heat resistant super alloys (HRSA). Due to low machinability of HRSA, accelerated tool wear and poor surface quality are the usual trend. Furthermore, during finishing operations, thin and complex walls with variable vibration modes may lead to unstable conditions (chatter). Finding an optimal combination of all these factors is a challenge for process development engineers and NC programmers. The lack of a scientific approach to solve this problem leads to conservative and unproductive cutting parameters, increasing cycle time and reducing throughput. To be competitive, improving the efficiency of the manufacturing process of these high added value components is mandatory. The effective and accurate calculation of cutting forces in complex milling is a hot topic as these models give insight for the estimation of workpiece quality (surface roughness and surface location error), tool/workpiece relative vibrations and tool life, to mention a few. In a pioneer work, Bouzakis [2] exemplified the complete diagnosis of a force model, for the special case of milling with a ball-end mill. Ferry and Altintas [3, 4] presented the generalized milling model for a variety of mill-geometries and demonstrated its usefulness in roughing operations. They also proposed a feed optimization algorithm to minimize cutting forces. Budak et al. [5] investigated 5-axis milling operations with ball-end mills, being capable of stability analysis with different tool/workpiece orientations. Wojciechowski et al. [6] analyzed cutting forces and vibrations particularly for ball-end milling in hardened steels. They Main aircraft manufacturers predict to double the fleet of airplanes over the next 15-20 years [1]. This leads to a higher demand for elements used in turbine engine which need to be efficiently machined. Particularly, the milling of rotary blade rotors (IBRs) is an important topic of research. Production of IBRs is an intricate process where complex surfaces in 5 continuous axes need to be machined, often made of heat resistant super alloys (HRSA). Due to low machinability of HRSA, accelerated tool wear and poor surface quality are the usual trend. Furthermore, during finishing operations, thin and complex walls with variable vibration modes may lead to unstable conditions (chatter). Finding an optimal combination of all these factors is a challenge for process development engineers and NC programmers. The lack of a scientific approach to solve this problem leads to conservative and unproductive cutting parameters, increasing cycle time and reducing throughput. To be competitive, improving the efficiency of the manufacturing process of these high added value components is mandatory. The effective and accurate calculation of cutting forces in complex milling is a hot topic as these models give insight for the estimation of workpiece quality (surface roughness and surface location error), tool/workpiece relative vibrations and tool life, to mention a few. In a pioneer work, Bouzakis [2] exemplified the complete diagnosis of a force model, for the special case of milling with a ball-end mill. Ferry and Altintas [3, 4] presented the generalized milling model for a variety of mill-geometries and demonstrated its usefulness in roughing operations. They also proposed a feed optimization algorithm to minimize cutting forces. Budak et al. [5] investigated 5-axis milling operations with ball-end mills, being capable of stability analysis with different tool/workpiece orientations. Wojciechowski et al. [6] analyzed cutting forces and vibrations particularly for ball-end milling in hardened steels. They 48th SME North American Manufacturing Research Conference, NAMRC 48 (Cancelled due to applied their model to optimize milling parameters for an inclined operation [7] . Ozkirimli et al. [8] developed a generalized model for variable pitch and helix cutters. Their solution was based on the frequency domain approach, including the process-damping effect. Urbikain et al. [9] [10] [11] developed the static and dynamic force models for a circlesegment tool intended for complex cutting operations. Experimental validation for the stability analysis was performed using a thin-walled workpiece, showing very good agreement in terms of cutting forces [9, 10] and surface roughness [11] . However, these authors used linear paths and fixed orientations during the tests. Monitoring of the milling process is also a useful technique to analyze tool wear and breakage, surface and dimensional quality, as well as machine behavior. Cutting forces, power consumption and tool/workpiece vibrations are the most appropriate signals for monitoring. There have been many attempts summarized in [12] to introduce sensors for data analysis during machining processes. Kuljanic and Sortino [13] proposed the concept of tool wear indicators that can be determined by analyzing cutting force signals. Lately, Postel et al. [14] proposed a low-cost method for a reliable estimation of cutting forces and vibrations using an accelerometer attached to the stationary spindle housing. This technique experimentally was proven both in low and high-speed milling operations. Barrel cutting tools can be an efficient alternative but there is a lack of knowledge about the way these tools behave. Here, a mechanistic model was developed for barrel-shape cutters with the objective of being subsequently generalized for the estimation of cutting forces in free toolpaths. The main purpose of the work is to estimate the maximum and mean total cutting forces during the milling of complex surfaces. The most common option for machining blisk and impellers is to use solid carbide coated milling tools. These tools cover the usual range of diameters from 6 to 25 mm. The indexable insert milling cutters can be an economic option in roughing operations, although their use in the finishing of IBR components is not very extended. The selection of the right coating (TiAlN, TiN, AlCrN , AlCr or CrN) could improve the performance of the cutter by increasing its resistance to abrasion, adhesion, diffusion and oxidation wear. The most used geometries in carbide milling tools are: straight-bull-sinusoidal, tapered, ball and circle-segment (see Fig. 1 and Table 1 ). The straight-bull milling cutter is the most suitable for roughing since the cutting speed, depending on the diameter of the tool, does not change over the whole edge. However, the edges of the tip can be excessively stressed for CNC trajectories with sharp corners in 5-axis operations. An option to reduce this problem is to use bull-nose geometries that have a roundness at the tip. This type of milling cutter can also be a very good option for semi-finish milling operations for blades. The main applications of tapered milling tools are flank milling operations. These mills not only perform finishing with marks in a single direction but also can be redesigned (helix angle, number of teeth, chip breaker) to perform semi-finishing operations, making it possible to achieve uniform over-material and fewer levels compared with a spherical milling tool. Similarly, the ball milling tool is preferred for finishing operations. For concave blade geometries with reduced radii of curvature, spherical geometries are the only option to perform semi-finishing and finishing operations. New circle-segment milling geometries are a potential alternative in semi-finishing and finishing operations for curved blade parts. A high radius of curvature in the profile of the milling tool allows a reduction in the number of passes and therefore, productivity increases significantly with reduced machine cycle time. The tool-holder choice is also critical for the machining process because it determines precision, clamping force and cooling performance. Machining operations near to the base of the blades may need a very slender tool holder to have access between blades, avoiding collisions. The main difference between a circle-segment cutter and a straight end-mill is the smoothness in the tool/workpiece engagement. In this case, the axial and radial cutting depths are coupled parameters as described by Eq. (1). (1) In Fig. 2 , setting a radial immersion results in a fixed axial depth of cut. The basic parameters for tool geometry are tool diameter d1, the radius of the barrel envelope r2, helix angle β and nose radius at the tooltip r1. The latter will be neglected throughout the study as tools are often intended for peripheral milling operations using whole flank, not for point milling operations. Table 2 list the geometrical features of the barrelshape cutter. Assuming that the bottom end of one flute is selected as the reference immersion angle and the cutter spins accordingly to the spindle speed n, ( ) = (2 60 ⁄ )( ). The bottom end points of the flutes (j = 1, 2, …, Nz) are at angles φj (0) and the immersion angles for flute j at axial depth of cut is: The classical mechanistic approach considers cutting forces over differential edge elements using a local coordinate system, described by Eq. (3) . Here, db = ds/sinκ, a differential element from the chip width and ds a differential element from the edge length. The side cutting edge angle κ(z) depends on the axial position of the cutting edge. The cartesian milling forces are calculated as shown in Eq. (4), where c and s stand for cosine and sine trigonometric functions respectively. To compute the total forces along the edge, it is essential to define the differential edge element ds(z) and the phase shift (z), both depending on tooth geometry. While straight cutters lead to constant helix angle β along the cutting edge, tools with no cylindrical profile project this angle over a variable profile r(z). This gives a z-varying angle that needs to be calculated: For a barrel end mill, the value for r(z), referred to the tool axis, leads to: Finally, the side cutting edge angle κ is defined as A circle-segment cutter is a tool intended for 5-axis milling operations using additional rotations, tilt θt and lead θl angles, as shown in Fig. 3 . These angles are needed to relate F-C-N and X-Y-Z coordinate systems with a rotation matrix. The process coordinate system (F-C-N) consists of the feed direction (F), the cross-feed direction (C) and the surface normal direction (N). 4 Author name / Procedia Manufacturing 00 (2019) 000-000 The experimental setup consists of an Al7075T6 workpiece mounted on a Kistler dynamometer model 9255B. Linear cuts along the length of the workpiece where conducted and the recorded forces from the dynamometer were used to calibrate cutting coefficient at different radial immersion values, feed and cutting speeds. The window of machining parameters (listed in Table 3 ) is set from tool suppliers' recommendations with some imposed limitations. For instance, cutting speed is limited by a fifth of the natural frequency of the Kistler dynamometer to avoid resonance effects in the measurements and ensure experimental data accuracy and reliability. The computed cutting coefficients were obtained by performing a linear regression of the mean forces versus feed values. No variations were found between both cutting speeds. Circle-segment tools lead to different subtypes of cutters (according to Emuge tool maker): taper, oval, lens and barrel. There are two considerations with respect to oval forms. First, these cutters can be used with positive or negative tilt angles. Secondly, the tilt angle has more limitations for barrel shapes compared to oval-form and should not exceed an inclination of more than 4 degrees (while tilt angle can be of 8-10º for the oval form). As a rule of thumb, the following criterium was found for the maximum (allowable) radial immersions ae. For positive tilt angles, ae,max = 0.6 mm when θt = 2º and ae,max = 0.2 mm when θt = 6º; for negative tilt values, ae,max = 0.5 mm when θt = 2º and ae,max = 0.1 mm when θt = 6º. Under this consideration, a campaign test was designed to analyze the model's predictions in operations with different tool axis orientations with respect to the machined surface. Table 4 compares the experimental and simulated resultant cutting forces acting on the barrelshaped tool which is: In both cases, the maximum force ( = ( )) and mean force ( = ( ) ) were computed during 10 revolutions for comparison purposes. The set of experimental tests progressively increases the level of uncertainty between the measured experimental forces and simulations. Tests #1 to #4 study the influence of the θt on the cutting forces while interpolating ae in the characterization window (0.1-0.7). Feed and cutting speed are not interpolated. Test #5 interpolated ae, f and θt, Test #6 included an interpolated cutting speed S, Test #7 interpolated ae and fz for a zero θt. Finally, test #8 is intended for the analysis of a negative θt. Fig. 4 presents the validation for some of the cases presented in Table 4 . Although the runout was not taken into consideration, a close agreement was found between experimental and simulated forces. This work presents a new comprehensive cutting force model that accurately estimates the forces produced by barrelshape cutter used to machine complex geometries such as aerospace rotary components. The study has validated the proposed model through experimental cutting forces of straight trajectories as the first step towards model validation in 5-axis toolpaths. The cutting force model uses the main geometrical parameters (d1, r2, β) , and orientation tilt (θt)-lead (θl) angles. • A general good agreement between experimentally measured milling forces was found. Among the three components, Fz component tends to be slightly overestimated by the simulation results thus giving relatively the highest deviation between experimental and predicted results. The predicted maximum total force was lower than the experimental one. One the contrary, the predicted mean total force was found to be slightly higher than the measured one. • The first step to control maximum milling forces in 5axis tool trajectories was more than satisfactorily completed without considering runout effects. The next stage is the preparation of complex surfaces to extend and validate the model in general milling operations in a simplified manner. Determination of the chip geometry, cutting force and roughness in free form surfaces finishing milling, with ball end tools Virtual five-axis flank milling of jet engine impellers-Part I: mechanics of five-axis flank milling Virtual five-axis flank milling of jet engine impellers Part II: feed rate optimization of five-axis flank milling Modeling and simulation of 5-axis milling processes The estimation of cutting forces and specific force coefficients during finishing ball end milling of inclined surfaces Application of signal to noise ratio and grey relational analysis to minimize forces and vibrations during precise ball end milling, Precision Engineering Generalized model for dynamics and stability of multi-axis milling with complex tool geometries Stability charts with large curve-flute end-mills for thin-walled workpieces Modelling of static and dynamic milling forces in inclined operations with circle-segment end mills, Precision Engineering Modelling of surface roughness in inclined milling operations with circle-segment end mills, Simulation Modelling Practice and Theory Advanced Monitoring of Machining Operations TWEM, a method based on cutting forcesmonitoring tool wear in face milling Monitoring of vibrations and cutting forces with spindle mounted vibration sensors This research was funded by Tecnológico de Monterrey through the Research Group of Nanotechnology for Devices Design, and by the Consejo Nacional de Ciencia y Tecnología (CONACYT).Acknowledgments are addressed to Basque Country University Excellence Group IT1337-19