Critical aspects in the application of acoustic emission sensing techniques for robust and reliable tool condition monitoring

E.M.Rubio^a, I.L. Baciu^b, R. Teti^b

^aDistance Learning University of Spain (UNED) Spain

^bUniversity of Naples, Italy

Introduction

Turning, drilling, milling and grinding are material removal processes widely used in many industrial sectors

In some of them, such as the car or the aerospace industry, the geometrical accuracy and surface finish requirements have been strongly increased in recent years

To meet these new needs, conventional machine tools have been replaced by numerical control machine tools (NCMT) to eliminate the variability introduced by the operator and to obtain better quality parts

In order to increase productivity and reduce manufacturing costs, NCMT have been associated with other components, mainly robots and computers, in flexible and computer-integrated manufacturing systems able to operate automatically during long periods of time

This advanced manufacturing systems demand an optimal performance at all machining stages

Are included just in time tool change when the tool has reached a certain wear level

Background - Friction and Tool Wear

• The main critical points associated with the definition of cutting tests for tool condition monitoring (TCM) have been analyzed

• During a cutting process, friction takes place at the workpiece-tool interface

• This is a very complex phenomenon and involves combined effects of adhesion, erosion, abrasion, fatigue, plastic deformation, diffusion and corrosion

• Friction produces a gradual wear of the cutting tools causing a negative influence on the quality of the machined surface and on the workpiece geometry as specified in the design

• If the machine tool operates with other machines tools and robots under the control of a computer, the breakdown of the machine generates a critical problem for the whole system

• To prevent such problems, various methods for tool wear monitoring have been proposed, but none of these has been universally accepted due to the complexity of this type of procedures

Classification of Tool Wear Monitoring Methods

• Direct methods (direct wear measurement using different types of sensors):

• optical sensors

• radioactive sensors

• electrical resistance sensors

• Indirect methods (wear evaluation on the basis of parameters measured during the cutting operation)

• cutting forces

• acoustic emission

• vibrations

• On-line methods (while the cutting process is being performed )

• Off-line methods (interrupting the process to carry out the control either on the machine or away from it)

Friction and Tool Wear

• The last trends in tool condition monitoring are focused on the development of on-line indirect methods

• There are several in-process indirect measurement systems based on different types of sensors or a combination of them

• The most commonly applied sensors are vibration and acceleration, load and power or acoustic emission and force

• Among them, acoustic emission and force sensors are the most widely employed due to their interesting performance

• However, there are different critical points in the set up of these systems that can be improved in order to obtain better results

• This work is focussed on the analysis of some of the most critical problems involved in the experimental set up of tool condition monitoring systems based on acoustic emission detection and analysis

Friction and Acoustic Emission (AE)

• The friction phenomenon produces elastic stress waves generated by the release of energy stored within a material, called acoustic emission (AE), that propagate through the material and can be detected by adequate sensors when they arrive at the material surface, provoking small displacements

• The mechanical signal can be turned into an electrical signal of low amplitude and high frequency by a piezoelectric AE sensor

• Electronically processed signals can be later analysed with mathematical and computational techniques

• AE signals can be classified into two types:

• continuous (signals are associated with plastic deformation in ductile materials)

• burst (signals are observed during crack growth, impact or breakage)

AE Stress Waves Sources

The main sources of AE stress waves during cutting are associated with:

• primary (1) and secondary (2) shear zones, particularly in the case of fresh tools;
• the tertiary shear zone or the frictional contact at the interface between tool flank and workpiece, in case of flank wear (3);
• the frictional contact between the tool rake face and the chip, in case of crater wear (2);
• the crack growth at the contact zone between tool tip and workpiece (4);
• The plastic deformation of the chip (5);
• The collision between chip and tool (6);
• The chip breakage and catastrophic tool fracture

AE Stress Waves Sources II

The main sources of AE stress waves during cutting are presented the figure:

AE Detection Steps

AE signals allow for the identification of tool wear state by means of signal parameter changes

Obtaining information on tool wear by means of AE detection and analysis generally includes the following steps:

• define the cutting tests;
• define the tool condition monitoring tests;
• select/propose the method(s) for signal processing;
• select/propose the process modeller(s);
• set-up the experimental layout;
• perform the cutting and tool condition monitoring tests;
• process the signals;
• model the tool wear process

Definition of the Cutting Tests

In the literature though, there are papers presenting an accurate description of the cutting tests, most of the authors do not provide, or at least not with the desired detail the following information:

• how they obtained the wear of the tools used in the tests
• the time of tool use for wear generation
• the used workpiece material for tool wear generation
• under what cutting conditions the wear was achieved

This is a critical issue since this data is needed to complete the information about the experiment and to have a better knowledge about the process, allowing for more reliable conclusions

Definition of the Cutting Tests II

It is critical to know the time of tool use in order to carry out a comparison of results when cutting conditions, workpiece material or type of tool are changed

It is important to have information about the workpiece material used for previous tool wear generation

In general, nothing is specified about the used workpiece material for tool wear generation: if the material is the same as the one for the AE tests or if the tool wear was achieved by machining different types of workpiece materials

Depending on the used material and the cutting conditions, different variations of the tool geometry can occur

The main cause of wear development is the adhesion mechanism, that consists basically in the transfer of small particles from the tool to the chips

Definition of the Cutting Tests III

The incorporation of macroscopic fragments from the workpiece material to the tool surface occurs because of work material adhesion on the tool rake face in two different mechanically unstable forms:

• Built-Up Edge (BUE)
• Built-Up Layer (BUL)

The existence of these small amounts of material that are going to partially or totally adhere to the tool rake face during the machining process can disturb the AE signal detection during the tests and can be removed from the tool surface by the action of the cutting forces

Main Elements to Establish in Cutting Test Definition

The main elements to establish in cutting test definition are reported in the table:

Tool Wear Clasification

During tool wear monitoring, only two parameters are generally  considered for tool wear characterisation:

• flank wear (VB)
• crater wear (KT)

Two wear levels are usually taken into account: fresh and worn

Lee and Dornfeld* proposed a finer distinction of wear levels measured by flank wear:

• fresh
• medium
• worn

* Lee, C.S. and Dornfeld, D.A., Design and Implementation of Sensor-Based Tool-Wear Monitoring Systems, Mechanical Systems and Signal Processing, 10 (4),1996, 328-347.

tool condition monitoring tests

The definition of tool condition monitoring tests consists basically in the selection of the most appropriate elements for the measurement chain, defining its better physical location for each cutting test, and establishing the necessary links for AE signal detection

Before studying the AE signal from a cutting process, it is necessary to condition it adequately

A measurement chain is usually composed of the following elements:

• piezoelectric AE sensor
• pre-amplifier
• low/high pass filters
• amplifier
• signal processing and recording

Tool Condition Monitoring Configuration

The TCM measurement chain presents the following elements

Measurement in AE System based TCM

The AE signal usually needs to be initially pre-amplified and then bandpass filtered

A high bandpass filter is used to reduce the influence of the low frequency noise, considered to be uncorrelated with the tool state

The AE signal is fed through a low bandpass filter to eliminate the high frequency noise components due to electric sparks and/or to avoid signal aliasing

The amplitude spectra consist of components only from the frequency range of interest

Among all the elements composing the measurement chain, the most important is the piezoelectric AE sensor

AE Sensors - Characteristics

The most critical aspects regarding the AE sensor are its selection and its location on the machine tool

There are many different AE sensor brands on the market but only some of them can be used in a real machining process

AE sensors for tool wear monitoring must be resistant to dirt, lubricants, chips and wear. Besides, it would be desirable that they need no maintenance, are easily replaced, and have a low cost

The AE sensor should be placed as close to the machining point as possible without decreasing the working space reducing the cutting parameter range or the static and dynamic stiffness of the machine

AE Sensors - Positioning

If aspects concerning the sliding between tool flank and workpiece or the deformation/fracture along the shear plane are to be studied, the workpiece is the best sensor location

If information concerning the sliding between tool flank and workpiece or between chip and rake face is required, the tool is the best sensor location

The tool will be used as sensor location also if the chip fracture phenomenon is to be analysed

When the sensor is placed on the workpiece, the main problem that can occur is the onset of cracks or defects that can intercept the AE signal before it correctly reaches the workpiece surface, loosing part of the information

This could be solved by using other types of sensors simultaneously as, for example, a sound analyser utilizing air as signal transmission medium

AE Sensors - Positioning II

Mounting the AE sensor on a tool that has a rotating motion, as in milling or grinding, can generate serious difficulties

This type of mounting is needed because if the sensor is located on the side of the workpiece (e.g. the machine tool table) the detected signal amplitude is conditioned by the distance between cutting point and sensor location that varies as a consequence of the relative movement between table and spindle head.

To solve the difficulty in detecting AE signals from rotating parts, special methods have been developed (radio or optical methods)

These techniques are in general not yet practically viable, except for an interesting solution where the AE signal is transmitted through the cutting fluids

Conclusions

The main problems involved in the design of cutting tests to carry out tool condition monitoring based on AE detection and analysis have been analysed

The study has revealed the need to pay attention to the way the previous tool wear level is achieved

The establishment of a methodology that could help comparing the results obtained from different tool condition monitoring tests and different laboratories has been proposed

The main elements and characteristics that should be provided for by such methodology have been established

Finally, it has been suggested that two or more types of sensors should be used simultaneously to prevent possible loss of information generated by internal defects in the workpiece or the tool