Strategies to Enhance Agility and Machining Accuracy in Line Boring

IFAC

Copyright © IFAC Mechatronic Systems, California, USA, 2002

c:

0

[>

Publications www.elsevier.comllocate/ifac

STRATEGIES TO ENHANCE AGILITY AND MACHINING ACCURACY IN LINE BORING

z. J. Pasek, B.-K. Min, Y. Koren, and A. G. Ulsoy

P.Szuba

Department ofMechanical Engineering University of Michigan, Ann Arbor, MI

Lamb Technicon Warren, MI, USA

Abstract: This paper describes three strategies used in the development of an agile line boring station for precision machining of long bores. Careful structural design assures high overall stiffness. Additionally, two mechatronic devices assure high machining accuracy via on-line compensation schemes. One mechanism, a "smart" line boring tool with built-in instrumentation and control for real-time, facilitates correction of the boring process. Another, a dual-ballscrew system enables correction of tool pitch errors. Overall, the design integrates multiple new concepts for mechanical structure of the machine with an intelligent controller. Copyright © 2002 IFA C Keywords: Machine Tool, Cutting, Control

control with precise machine hardware. To achieve this goal, leading edge technologies in machine design, sensing, tooling, and advanced control were applied in this project in a coordinated systems approach.

I INTRODUCTION Increased global competitiveness and pressure to faster introduce new products while improving fuel economy, drive in the automotive market an urgent need for new powertrain machining technologies that provide flexibility at an affordable cost. Although FMS technology has been introduced over 20 years ago [Koren, 1983], only in the last few years CNC and FMS were started to be utilized in automotive applications . Introduction of CNC machines into powertrain machining systems increased the level of flexibility in engine production. But machining of long bores as needed in cam- and crankshafts of engine blocks (see Figure I) still remains an exception, requiring use of dedicated equipment, and hence impeding achievement of full system agility.

Fig. 1. View of a car engine blocks.

Precision line boring is a very demanding application in terms of both quality and production rate requirements . Increasing flexibility without compromising quality or production rate (in a twoshift operation mode) is difficult and challenging, and requires a systems approach integrating advanced

The goal of this project was to enable the implementation of fully agile machining systems for engine blocks, by designing an agile precision line boring station to machine the bores for crankshafts

563

Error mode ling has become an important part of machinc tool error compensation procedures. Initial methods originated from the error modeling of CMMs [Schultschik, 1972]. In most models a method relying on homogenous transformation was used to describe the quasistatic error [Kim, 1972; Yang, 1996]. Drawbacks of most of these models are due to the fact that many characteristics of machine tools were ignored (e.g., error behavior under cutting load, part accuracy issues, thermal effects).

and camshafts. However, some trade-offs between agility and machining precision are unavoidable . The process requires usage of a very long and heavy tool. Flexibility requires elimination of support bushings for the tool, thus the tool may become a subject to vibrations and deflections preventing achievement of the required precision of 10 !lm.

3 BORING STATION DESIGN The Agile Line Boring Station is a flexible, highspeed machine. Extel1lal support for the boring bar in the form of support bUs,hings incorporated into the fixture has been eliminated. Station is equipped with a tool magazine to hold multiple tools of different sizes. The station has the following specifications: I . Spindle: a conventional, motorized spindle with modified arbor, 7.5 HP (5 .5. Kw), 0-6,000 rpm. Optional rotary inductive power and data transfer system mounted around the arbor. Tool holder HSK 125B. High-pressure, through-tool cooling system. 2. Motion ranges: X-axis 350 mm, Y-axis 350 mm (plus additional 650 mm stroke for tool change), Z-axis 1250 mm 3. Positioning: accuracy 2 !lm; repeatability I !lm.

Fig.2. Line boring with guided tooling.

2 STATE-OF-THE-ART In powertrain components, where bearing support is necessary for cam and crank shafts, a number of precision bearing nests are located in-line. When the distance between two such bores is not too large (less than 5-10 bore diameters), line boring is used. Line boring ensures that the holes are of the same diameter and concentricity error is minimized. In typical automotive applications the range of machined bore diameters is D = 30-75 mm; the required cylindricity of the bore lies within 10-25 !lm, and the surface finish is Ra = 2-50 !lm. In extreme cases boring bars can be as long as 1350 mm and weigh up to 120 kg. The state-of-the-art in flexible line boring technology review [Pasek, 1993] provided relatively limited results, most likely because proprietary information is not made public by the manufacturers, and academic research is almost non-existent in this area. Reviews of the recent developments in the general-purpose boring technology [Aronson, 1996; Mason, 1997; Hanson, 1994] reveal new trends in the area of traditional boring applications, the most prominent one of which is an increasing presence of instrumentation and control in the boring tools.

Fig. 3. Line boring machine tool prototype.

Since requirements for precision and accuracy are critical for manufacturing, machine tool error analysis have been a very important research area. Machine tool error analysis and compensation is heavily based on the Abbe Principle [Bryan, 1979; Weck et ai., 1996; Sartori, 1995]. Error induced by the machining process are usually classified into quasistatic errors (e.g., geometric, thermal, and alignment errors), and non-quasi static (e.g., tool deflection [Rivin, 1992] and vibrations, workpiece deflections , controller tracking errors) [Hocken, 1980]. Quasistatic errors usually account for more than 70 percent of total error budget.

Fig. 4. Tool magazine A number of different concepts were studied during the design process, and the concept based on closed machine frame was selected [Pasek, 1998, Beecherl,

564

2000]. In this non-traditional concept, the spindle is mounted on thc undersidc of a cross-slide and suspended from a robust machine frame (Figure 3). The tool changer (see Figure 4) is located at the bottom of the frame and positioned directly under the spindle. In the machine development, symmetry of the design was sought to ensure even thermal growth (i.e., minimum distortion) of the machine. The vertical Y-axis has two ballscrews that have independent, but coordinated motion control.

utilizing sensing and intelligence built into the tool itself to compensate for the increased compliance of the tool without supports. This sensorized tool IS called a "smart" tool [O'Neal, 1998, Szuba, 2000]. The boring bar is located in the motorized, highprecision spindle and is connected with it through the standard taper mechanism enhanced with additional electrical connectors. The tool body (see Figure 7) contains a number of sensors and the actuation mechanism . The instrumentation supporting the sensors and the actuation mechanism is located in the package attached to the rear end of the spindle and rotating with it. The instrumentation package (including on-board computer) and the piezoelectric actuator are receiving power from an external source via an inductive device . The housing for the instrumentation can work with multiple boring bars thus eliminating the need for frequent disconnections.

Through iterative analysis and design modification, the design was modified multiple times to increase the static stiffness (from initial 10 N/!1m to 50 N/llm) and the lowest natural frequency (from 15Hz to 38 Hz). This was accomplished by careful consideration of the Finite Element Analysis and adding strength/increasing inertia without adding substantial mass, thereby decreasing the natural frequency. By making modifications in the design parameters it was possible to eliminate most of the undesirable modes of vibration.

Flexure

Mechanism

4 TOOLING

A typical cutting tool used in the station is a solid boring bar, equipped with two single-point cutters, one for roughing, and the other for the finish machining. Only one of the cutters operates during a cutting pass. The boring tool has multiple guiding pads supporting it in the already machined part of the workpiece (see Figure 5). The initial threading of the tool into the uncut bore and its support while machining the first hole (when tool is completely unsupported) is facilitated by a support bushing integrated with the tool. During the tool change both the tool and the bushing are replaced as a joint unit.

"

-,~.......

Beam Sptiner

... Mirror

...... Capadtance Sensor "'- Plezoelectlk: Stack

,

~ Gulde Pads

Fig. 6. The instrumented (smart) boring bar. Functionally, the bar contains three subsystems: a) sensors, b) actuation mechanism, and c) cutting inserts. Details of the Smart Tool design, sensing, and control are elaborated on in [O'Neal, 1998; Koren, 1999; O'Neal, 2000, Szuba, 2000). The purpose of the tool tip controller is to isolate the tool tip from any erroneous non-cylindrical motion of the boring bar, holding the tool tip still relative to the spindle under external disturbances. A single-laser, double detector system provides the controller with the displacements of the tool tip and the boring bar relative to the spindle. For the purpose of tracking the reference signal, a digital linear quadratic optimal control law has been designed and implemented on a single-board PC/l 04 computer running at 150 llS sampling period. A full feedback control law was designed to create a stiff tool tip with ability to reject the dynamic cutting forces. Such control law requires knowledge of all the current states of the system. A state space realization was chosen in such a way that the first state is current tool tip position. A Kalman filter

Fig. 5. The regular (dumb) boring bar. While elimination of support bushings facilitates agility by enabling automated tool change of the long boring bars, the more traditional design of the tool is still recognized as a major bottleneck in increasing flexibility and precision of the line boring processes. A new concept of the line boring tool have been proposed to support agility at the boring station level. It relies on an on-line compensation mechanism

565

observer has been designed to estimate the states in real-time for the full state feedback (see Figure 7).

Fig. 8. Dual ball screw error compensation mechanism To carry out an active error compensation scheme, additional dcgrees of frecdom (DOF) in the machine structure need to be introduced. This additional DOF for this purpose imple~ented by use of the dual linear actuator system ' in the Y direction for correction of large pitch errors . Concept of this mechanism is shown in Figure 8.

Toof tip position

Cutting force monitoring

··········1

Machine controller

Fig. 7. Block diagram of the Smart Tool control The implemented control scheme requires full knowledge of all the states of the control system, which are used in a full-state feedback. However, since the Smart Tool uses sensors to measure only the end positions of the tool tip and the bar, an observer must be used to estimate remaining states. Furthermore, there is a need to measure the cutting force and compensate for it to improve the system performance. For this purpose, an observer model implemented includes both cutting force dynamics and the tool dynamics. The estimated cutting force is then used in two ways: for tool tip position control and for process diagnostics [M in, 2002].

The dual ball screw system can be effectively used for correcting pitch error, as well as linear errors in the Y direction. In the presence of an angular pitch error, the platen carrying the tool can be rotated in the opposite direction due to a differential motion between the two ball screws. The resolution of the differential motion depends directly on the resolution of the axis feedback devices (e.g., servomotormounted encoder). When two ball screws are used to move a common axis, incremental differences induce an angular motion to the moving platen (see Figure 8). Based on the known machine geometry, the angular displacement is

5 ERROR ANALYSIS AND COMPENSA nON The six geometric errors affecting each axis of motion can be described in two different ways. The errors are either defined with reference to successive translatcd and rotated axes (succesive approach) or with respect to a fixed set of coordinate axes where the orientation does not change. The second approach is more experimentally viable, since the error is measured with respect to the same set of fixed reference axes.

g=atan[(B2 -B,)/S]

(I)

and the resultant motion at the tool tip is equal to: Y=(B2 -B,)(S+O+L)/S

(2)

Using two ball screws with 20 mm pitch and encoder resolution of 64,000 divisions gives the resolution for pitch correction Dg = 2.308 arc-sec and linear error correction at the tool tip D Y = 0.00052 mm. Further improvement of the resolution can be obtained through use of ball screws with different pitches and corresponding modification of motion control algorithms for each of them.

With the resultant error motions known for all six geometric errors, the total resultant error equation was found by the superposition of component errors. The results of the calculations implies that with the maximum angular error of 10.0 arc-sec or less, the error correction scheme has to be capable of correcting errors in excess of 50 Jlm [Szuba, 1998]. It was also important to determine how each geometric error, i.e. linear displacement error, roll, pitch, yaw, oX, aY, contributes to the overall tool tip error. Clearly, the pitch and yaw errors have the highest contributions, which is due to the fact that these angular errors are amplified with the excessively long cantilevered boring bar. The roll error has the smallest effect.

6 PC-BASED, OPEN CONTROLLER The line boring station is equipped with a PC-based, intelligent controller that utilizes several control algorithms [Koren, 1992]. It includes a combination of traditional algorithms [Ulsoy, 1993; Mehrabi, 2002] and feedforward algorithm as well as the (optional) integration of the smart tool controller with the main machine controller. The tasks of the

566

intelligent boring controller is : (I) to tune the controller and the machining parameters to their optimum conditions despite changing environment conditions, such as temperature and load, (2) to provide the system with the ability to autonomously respond to irregular, unanticipated events which may occur during the boring process, and (3) to provide a platform for integration of multiple, non-standard control and monitoring functions (e.g., thermal and geometric error compensation). A block diagram of the controller is shown in Figure 9.

narrower error band can then be further reduced by application of "smart" tool, down to ± 1/lm range. In case when a "dumb" tool is used, the accuracy of the process can be kept within standard process range.

Controller Comput,r

The capabilities and effectiveness of each compensation mechanism were verified experimentally. In the first test, linear error long the Z-axis were recorded over the operating range using regular (dumb) tool. The guiding pads in this tool were not extended over its full length (meaning that the covered tool stroke was shorter than the length of the workpiece) . Such tool design caused a sudden sagging of the tool (over 300 /lm) when it slipped out of the guiding pads (see Figure 10). Recorded data were then used to curve-fit two-piece compensation function, which in turn was used to control relative motion of two ball screws. When the compensation mechanisms was activated, the linear error range was reduced to ±5 /lm band. 300

Fig. 9. Block Diagram of the Main Controller

- - lIVithout compensation ............ IlVith compensation

200

E'

The CNC controll er is based on the Cranfield Precision "CUPROC 6400" CNC Controller. It controls and coordinates all machine 110 functions and motions . The controller architecture is very generic and open, thus allowing the control of the Agile Line Boring station to be customized for particular user needs.

3- 100 c: 0

:-2

0

V>

8. (5

~ -2

The controller utilizes a dual ISA computer bus configuration, each with a single board Pentium computer and ethernet adapter card to connect the two platforms. Such architecture provides: (I) an open Windows NT environment for Human-Machine Interface (HMI) and third party software, and (2) deterministic hard real-time environment for trajectory planning, interpolation and transmitting position commands to the intelligent digital drives over the SERCOS drive network with 1 kHz frequency .

200

300

400

500

600

700

800

900

Z-axis (mm)

Fig. 10. Compensation of linear error in Z-axis using dual screw mechanism Another set of tests was performed to investigate effectiveness of control approach for "smart" tool [O'Neal, 1998; Min, 2002] . Figure II shows step response of the tool and its excellent tracking abilities. The high dynamics and sensitivity of the "smart" tool allow also to use the tool as a process tracking and diagnostic instrument [Min, 2002] .

7 EXPERIMENTAL RESULTS

8.--------------, The experimental modal analysis of the prototype machine has conformed closely to the results of FEA analysis. In the frequency range from 0 to 100 Hz seven structural modes were identified and the lowest significant one was at 37 Hz. That compares well with the analytical prediction of 38 Hz (typically, the first mode for a 3-axis CNC machining center is in the 30-50 Hz range).

6

E'4

~2 o

~ 0

8. -2

.g.

--4

~

The error compensation approaches described in previous sections are the basis for a two-stage accuracy enhancement strategy implementation. The dual ball screw compensation mechanism is capable of reducing tool tip errors induced by the structural inaccuracy of the machine to the ±5/lm range . This

-6

t-~....".......

- - Reference

~~~--~~~~==~~

o

25

50

75

Time (msec)

Fig. 11 . Step response of the "smart" boring tool

567

8 SUMMARY

Min, B.-K., G. O'Neal, Z. Pasek, and Y. Koren (2002) Cutting Process Diagnostics Utilizing a Smart Cutting Tool. In: Mechanical Systems

A radically new machine tool concept for precision line boring has been developed. In this new design static stiffness was improved approximately 5 times (over existing designs), which allowed the increase of the first natural frequency of the machine to 38 Hz (75% improvement) . Two-stage scheme for compensation of geometric errors was developed : coarse one at the machine level, and more precise one at the cutting tool level. Machine control and compensation algorithms were implemented on an open-architecture, PC-based controller platform.

and Signal Processing (to appear) O'Neal G., Pasek Z., Min B.-K., and Koren Y .. (2000). Integrated Structure/Control Design of Micro-Positioner for Boring Bar Tool Insert. In:

Proceeding of SPIE's 7th International Symposium on Smart Structures and Materials, Conference on Smart Structures and Integrated Systems O'Neal, G., B.-K. Min, Li, C.-I. Pasek Z. J., Y. Koren, and P. Szuba. (l998). Precision Piezoelectric Micro-Positioner for Line Boring Bar Tool Insert. In: Symp. On Active Control of Vibration and Nfise. ASME IMECE, DE-Vol. 97IDCS-Vol. 65, pp. 99-106 Pasek, Z.1., G. Ulsoy, Y. Koren (1993). Flexible Line Boring Project: Benchmarking Study, Univ. of Michigan, Ann Arbor Pasek, Z. J., Szuba, P. (1998). Intelligent Agile Line Boring Station. In: Proc. of Dynamic Systems

ACKNOWLEDGEMENTS Research presented in this paper was sponsored in part by NIST Advanced Technology Program grant # 70NANB5H1158.

REFERENCES Aronson, R. B. (1996). The Whole Boring Business. In: Manufacturing Engineering, 05/96, pp. 5968 Beecherl; P., Pagels, D., Saeedy, A ., Szuba, P. (2000). U. S. Patent # 6,149,561 : Machine and Method for Flexible Line Boring. Bryan, J. B. (1979). Abbe Principle Revisited. In : Precision Engineering, 113, pp. 129-132 Hanson, D. R., Tsao, T.-COo (1994). Development ofa Fast Tool Servo for Variable-Depth-of-Cut Machining. In: ASME, DSC Vol. 55-2, pp. 863871 Hocken, R. J. (1980). Technology of Machine Tools, In: Machine Tool Accuracy, Vol. 5, Lawrence Livermore Lab, Univeristy of California, Livermore, CA Kim, K., Eman, K. F., and Wu, S. M .. (1987). Inprocess Control of Cylindricity in Boring Operations. In: Trans. ASME J. of Eng. for Industry, 109, pp. 291-296 Koren, Y . (1983) . Computer Control of Manufacturing Systems, McGraw-Hill Book Co., New York Koren, Y. and C. Lo. (1992). Advanced Controllers for Feed Drives (keynote paper) In: Annals of the CIRP, 4112, pp. 689-698 Koren, Y., Pasek, Z., Szuba, POo (1999). Design of a Precision, Agile Line Boring Station In: Annals of the CIRP, 4811, pp. 313-316 Mason, F. (1997). Smart Tools for Boring and Tapping. In: Manufacturing Engineering, 05/97, pp. 84-93 Mehrabi. M. G., Szuba,P., G. O'Neal , B. Min, Pasek Z ., and Koren Y . (2002) Geometric Error Compensation in Line Boring Process. In: Journal of Intelligent Manufacturing, 13/5, pp. 379-389

568

and Control Division, ASME International Mechanical Engineering Congress and Exposition, DSC-Vol. 64, pp. 439-446, Anaheim,CA Rivin, E., H. Kang. (1992) . Enhancement of Dynamic Stability of Cantilever Tooling Structures. In: Inll. 1. of Machine Tools & Manufacture, 32/4, pp. 539-556 Sartori, S. and G. X. Zhang. (1995). Geometric Error Measurement and Compensation of Machines. In: Annals of the CIRP, 44/2, pp. 599-609 Schultschik, R. (1972). Components of Volumetric Accuracy. In: Annals of the CIRP, 25/1. pp. 223-226 Szuba, P. (1998) . Improving Part Accuracy in

Machining Operations that Employ Cantilevered Boring Tools. PhD Thesis. Oakland University Szuba, P., O'Neal, G., Min, B.-K., Koren, Y., Pasek, Z. (2000). U. S. Patent # 6,062,778: Precision Positioner for a Cutting Tool Insert Szuba, P., Pasek, Z. (2001). U. S. Patent # 6,325,578: Method of Error Compensation for Angular Errors in Machining (Droop Compensation). U1soy, A. G., and Y. Koren. (1993). Control of Machining Processes. In: ASME Trans. J. of

Dynamic Systems. Measurement. and Control, 115/28. pp. 301-308 Weck, M., P. A. McKeown, and R. Bonse. (1995). Reduction and Compensation of Thermal Errors in Machine Tools. In: Annals of the CIRP, 44/2, pp. 589-598 Yang, S., 1. Yuan, and J. Ni. (1996) . Accuracy Enhancement of a Horizontal Machining Center. In: 1. of Manufacturing Systems, 15/2