GB2340604A

GB2340604A - Acoustic crack detection in shaping by deep drawing

Info

Publication number: GB2340604A
Application number: GB9918170A
Authority: GB
Inventors: Norbert Bergs; Stefan Klatzer; Dieter Barschdorff; Hagen Haupt
Original assignee: Benteler Deutschland GmbH
Current assignee: Benteler Deutschland GmbH
Priority date: 1998-08-18
Filing date: 1999-08-02
Publication date: 2000-02-23
Anticipated expiration: 2019-08-02
Also published as: ES2159476B1; ITRM990524A0; FR2782464B1; GB2340604A8; IT1309043B1; DE19837369A1; FR2782464A1; GB9918170D0; ES2159476A1; CZ9902906A3; GB2340604B; ITRM990524A1

Abstract

An apparatus for crack detection in the shaping of metal workpieces (4) by deep drawing in a drawing press (1) by picking up the noises accompanying shaping using body sound sensors (7, 8). To this end, during the deep drawing process simultaneous and direct detection of the body sound is carried out on the workpiece both in the shaping zone and outside the shaping zone. Signals with the same amplitude and frequency are determined as noise signals by comparison of the detected body sound signals and are filtered out. The body sound signals picked up in the shaping zone (VZ) are adjusted by removal of the noise signals and subjected to crack detection analysis.

Description

2340604 "Method and Apparatus for Crack Detection in Shaping by Deep

Drawing" THIS INVENTION relates to a method of crack detection in the shaping of metal work-pieces, particularly tubes, by deep drawing, and to an apparatus for performing the method.

In the deep drawing of metal workpieces, for example of seamlessly drawn or welded tubes, which are used as axle brackets in motor vehicles, microscopic damage to the material of the workpieces can give rise to the formation of cracks, which can lead to surface cracking and in an extreme case to through cracking. Both crack formations imply a weakening of the crosssection of the workpieces and lead to scrapping of the damaged parts.

Because of the shaping operations in the deep drawing of work-pieces, in particular tubes, overloads of high tensile and compressive stresses occur in some regions of the workpiece overloads of high tensile and compressive stresses. Damage, for example in the form of longitudinally directed scoring, can lead to surface cracking or through cracking in the region of high tensile or C> compressive stresses. However, complete cracking through of the material does not result in every case. Often the material tears only in a manner such that it cannot be detected in the final pressure testing of the deep drawing process.

It is generally known that analysis of the emitted sound is suitable for detecting cracks in the shaping of materials. In this connection, DE 42 42 442 C2 discloses a prior art method for adjusting the clamping force of the jack of a drawing press. Here one finds the use of sound emissions analysis to the detection of cracks and folds in the deep drawing of sheets. The sound generated in the body during the drawing procedure is picked up and evaluated

2 by comparison with reference traces obtained with the same drawing tool before starting the production of the drawing press. Depending upon the results of the sound emissions analysis, the quality of the drawn part is then determined and the clamping force of the jack of the drawing press is controlled accordingly.

It is disadvantageous that for each type of drawn part on the same drawing press with the same drawing tool reference characteristics must be ascertained before production. This is expensive. Also differing properties of the material of the drawn part influence the sound emissions analysis.

Further, it works out to be disadvantageous that the body sound in the drawing tool is picked up indirectly. A body sound sensor is fixed to the drawing tool of the known arrangement. During the shaping process the vibrations are conducted through the drawing tool to the body sound sensor and are thereby damped or corrupted. This prejudices the quality of the measurement results and their analysis. Consequently only through cracks in the workpieces can be detected.

The present invention aims to provide an improved method and apparatus for detecting cracks capable of identifying the kind of crack during the deep drawing process.

Accordingly, the invention provides a method of crack detection in the shaping of metal workpieces by deep drawing in a drawing press by picking up 0 the noise generated during shaping using a body sound sensor, including the 0 0 following steps: using respective body sound sources to simultaneously and directly detect the body sound from a workpiece in the shaping zone (VZ) and outside the shaping zone (VZ) during the deep drawing process. identifying signal components with substantially the same amplitude and frequency as interfering noise by comparing the detected sound signals; processing the body sound signal detected in the shaping zone ('Z) by removing the interfering noise signals; and subjecting the processed signal to crack analysis.

The essence of the invention is based on the step of undertaking a measurement of the body sound directly from the workpiece during the deep drawing process, and moreover both in the shaping region, in which the drawing tool acts to shape the workpiece during the deep drawing process, as well as outside this shaping region. Signal components with substantially the same amplitude and frequency, which are present in both signals, are determined as interference noise by comparison of the picked up body sound signals. The body sound signal received from the shaping region is. then subjected to crack detection after removal of the interference noise signal.

By cracks in the context of the invention are to understood folds and surface cracks as well as through cracks.

The analysis of the signals indicates during the shaping clearly assessable frequency components, which can be interpreted as folds, surface cracks or through cracks of faulty workpieces, while the frequency components of the normal usual drawing noise or the interference noise resulting from the drawing press are eliminated. Based on the drawing noise when faulty workpieces are drawn, these can accordingly be distinguished from defect-free workpieces. Through cracks identify themselves in the drawing noise and in the signal derived therefrom specifically by spectrally spread amplitude changes, in contrast to which folds or surface cracks are detectable by a 4 significant change in the temporal or spatial development of the random and/or periodic sound components in comparison with the corresponding development in the drawing of defect-free parts.

The invention enables reliable crack detection during production in real time, whereby an identification of folds and surface cracks is just as possible as for through cracks.

The method of the invention is particularly suitable for crack detection in the shaping of seamlessly drawn or welded tubes.

A characteristic of shaping by deep drawing, particularly shaping of tubes is that a single or frequent change between compression and tensile requirements takes place during a transformation process. Cracks can arise in this way. With the appearance of such material changes during the shaping acoustical pulses are emitted as a result of displacements and stress loadings. In shaping, however, characteristic pulses also arise which are not correlated with a formation of cracks. The frequency and the amplitude of these pulses are dependent on the properties of the material and the degree of damage to the workpiece. The pulses of the acoustic emission consequent upon cracks arising are clearly recognisable in the body sound signal by virtue of their higher energy content and can be distinguished from the displacement pulses.

As already explained, during the signal reception a basic noise resulting from the drawing press occurs as background noise. As detection is carried out with two body sound sensors at the same time, this background noise appears in both signals with substantially the same amplitude and frequency. Thus, it can be removed from the signal subsequently or during processing using an adaptive filtering process, so that the sound emissions signal no longer carries the background noise.

The analysis of the signals to detect cracks is camed out using a computer. In this process the spectral power density and the energy of the processed pulse in the received frequency range is calculated. These quantities serve to distinguish between folds, surface cracks and through cracks.

The direct picking up of the transmitted sound from the workpiece avoids influencing of the sound signal by the tool. Additionally the simultaneous reception of the sound in the shaping zone and outside this zone enables an elimination of the background noise and the corresponding signals. In this way a reliable evaluation of the received signals by frequency analysis in order to recognise cracks is made possible. Accordingly, only process or crack-relevant signals are extracted for frequency analysis.

This makes possible real time processing which has a direct impact on the quality of the processed workplece. Workpieces suffering from cracks can be withdrawn from the production line without further ado.

During the deep drawing process, the probe pins of the body sound sensors are preferably resiliently pressed against the workpiece. In this way, the pick up quality of the signals is improved. The sensor coupling moreover enables a point measurement also for changes in the loading between tensile and compressive stresses.

6 The direct picking up of the body sound at the workpiece and the resilient contact of the body sound sensor with the workpiece makes possible a measurement up to the frequency region of 50Khz. In particular, the evaluation of the pulse with high frequency signal components between 1OkHz and 30kHz enables reliable crack detection and the identification of the kind of crack. - Practical tests have shown that, in contrast to cracks extending completely through the material, surface cracks are characterised by highfrequency body sound oscillations of low amplitude. Accordingly, in embodiments of the invention recourse is had to high frequency body sound signals of high amplitude for the detection of through cracks in the workpiece and recourse is had to body sound signals with low amplitude for the detection of surface cracks or folds in the workplece.

The signal evaluation accompanying the process is achieved using a data processing unit with corresponding central processor unit and signal processor. The arithmetic unit of the signal processor carries out all of the arithmetical and logical operations in combining the captured data. The control unit controls the system so that all operations are carried out in timely and logical sequence.

In another aspect, the invention provides apparatus for performing the method of the invention, comprising a drawing tool arranged in a press and having a counter die and a die, in which apparatus at least two body sound sensors are installed in the drawing tool with the probe pins thereof arranged to act on a workpiece at contact surfaces of the die and/or the counter die.

Suitably, the body sound sensor provided for contacting the workpiece C in the shaping zone is provided in the die, while the body sound sensor outside the shaping zone is installed in the counter die.

7 Preferably, piezoelectric sound pickups are employed as the body sound sensors. The useable sound frequency range lies between a few Hertz and a few 100 Kilohertz.

A reliable coupling of the body sound sensors to the drawing tool, which is insensitive to interference, is achieved by mounting the body sound sensors in respective receptacles of the die and counter die, whereby the probe pins are guided in bores to the contact surfaces. This arrangement enables body sound to be picked up directly from the workplece to be shaped during the shaping process.

Preferably, the probe pins are pressed against the workpiece by the resiliently biasing force of a spring. In this manner, it is ensured that the contact between a probe pin of a body sound sensor and the workpiece is maintained during the shaping process. As an alternative to this, the probe pin of a sound sensor can also be hydraulically or pneumatically pressed against the workpiece. The high contact force achievable by hydraulically or pneumatically pressing the probe pin against the workpiece enables the quality of the signal picked up to be improved.

The mounting of a body sound sensor can also be carried out, so that this is carried by a lever arm arranged for limited pivoting movement. The lever arm can be spring-biased, or hydraulically or pneumatically pressed towards the workpiece, whereby the probe pin of the sound sensor comes into engagement with the workpieces.

Advantageously, in this arrangement a support of the lever arm is realised using an anti-friction bearing. This advantageously affects the sensitivity of the sensor unit and the signal quality. Static and dynamic friction 8 of the probe pin and sound sensor on the wall of a receptacle is obviated. The useable frequency region is larger. Also, the constructional height of the sensor unit can be reduced by these means.

In order that the invention may be more readily understood, embodiments thereof will now be described, by way of txample, with reference to the drawings, in which:

Figure I illustrates schematically an apparatus embodying the invention applied to the deep drawing of a tube:

Figure 2 is a vertical cross-section through a die of the Figure I apparatus provided with a slidably mounted body sound sensor; Figure 3 is a vertical cross-section through another embodiment provided with a rotatably supported sensor holder with anti-friction bearings, slideable body sound sensor and biasing spring; Figure 4 is a signal trace obtained when deep drawing a tube without surface or through cracks; Figure 5 is a typical signal trace for a through crack in the tube; and Figure 6 is the signal trace for a surface crack in the tube.

In Figures I to 3 corresponding components are indicated by the same references.

Figure I shows a portion of a drawing tool I installed in a press. The C drawing tool I comprises a counter die 2 and a die 3 3 moveable relative to the counter die 2 and in the form of a starnp. In the art, the die 3) Is termed a sword.

9 In Figures 1, 2 and 3, a tube is indicated by the reference 4. For use as an axle bracket in motor vehicles, the drawing tool I provides the tube 4 with a substantially V-shaped re-entrant double-wallqd middle section 5. The end sections 6 of the tube 4 are not shaped in the drawing process, so that the circular cross-section is retained here.

In the drawing process, the tube 4 is fixedly supported in the counter die 2, while the die 3 is driven upwardly into the tube 4 and thereby plastically deforms the middle section 5, so that this receives the desired V-shaped crosssection.

The shaping zone VZ in the longitudinal region of the middle section 5 is clearly shown in Figure 1. As already explained, the end sections 6 remain un-deformed.

To pick up the body sound developed during the deep drawing process, a piezo-electric sound sensor, 7, 8 is arranged both in the die 3 and also in the counter die 2. Each sound sensor 7, 8 comprises a sound detector unit 9 and an associated probe pin 10. The sound sensor 7 is slideable in a receptacle I I of the die 3 (see also Figure 2), while the sound sensor 8 is arranged in a corresponding receptacle 12 of the counter die 2. The probe pins 10 are guided in bores 13 and are pressed directly into contact with the tube 4 by the resilient biasing force of a spring 14. The probe pin 10 of the sound sensor 7 in the die 3 acts on the contact surface 15 in the shaping zone VZ. The sound sensor 8 in the counter die 2 is arranged outside the shaping zone VZ and accordingly acts on a contact surface 16 of the tube 4 in a region which is not shaped by the 0 deep drawing process.

Figure 2 makes it clear that the die 3 has a cross-sectional configuration corresponding to the desired V-shaped cross-section in the middle section 5 of the tube 4. The position of the sound sensor 7 ip the middle of the die 3 is seen as particularly suitable in practice.

Figure 3 illustrates the installation in the die of a body sound sensor 9 having a rotatable mounting. The body sound sensor 9 is carried by a lever arm 17 mounted for limited rotational movement. The bearing of the lever arm 17 is achieved using an antifriction bearing 18. A biasing spring 19 presses the lever arm 17 in the direction towards the tube 4. The probe pin 10 is thus brought into contact with the tube 4 in the shaping zone. The rotatable mounting obviates static and dynamic friction between the probe pin 10 of the sound sensor 9 against the wall of a receptacle or easing. The sensitivity of the sensor unit and thereby the signal quality are distinctly improved. The useable frequency range is greater.

Sound emissions occur if, in the course of the deformation of the tube 4 deep drawing, predetermined local stress thresholds are overstepped or if it suddenly adjusts to a new state of equilibrium at another energy level. This arises in the shaping of the tube by changes taking place between compression and tensile loads. The development of cracks can then arise. The energy freed in this manner is transmitted in the form of elastic waves and can be measured as a sound pulse. The zone of the strongest crack formation is shown crosshatched in Figure 1 and is designated with the reference 20.

The arrangement of the body sound sensors 7 and 8 enables simultaneous direct reception of the sound in tile shaping zone W and outside this directly from the workpiece 4. The acoustic pulses resulting from crack formation are distinguished by their high energy content from body sound 11 signals which result from the normal drawing noise resulting from displacement pulses. Consequently, sound pulses resulting from crack formation can be distinctly recognised and distinguished from displacement pulses.

During the capture of signals in the deep drawing shaping a background noise is also encountered, the so-called "hum" of the drawing press, as interfering noise. As this is simultaneously detected by the two sound detectors 7, 8, the signals resulting from this affects the two sound sensors 7, 8, with the same amplitude and frequency. The interfering noise can be removed from the signal by means of an adaptive filtering process. The parallel measurement"of the body sound using the sound sensors 7 and 8 thus enables a direct determination of the interfering noise, which by signal processing steps can be filtered out of the signals received from the shaping zone W through the sensor 7. The relevant crack signals can then be subjected directly to crack detection.

The diagram of Figure 4 shows by way of example the signal trace obtained with shaping without surface or through cracks. It shows that significant pulses in the body sound signal can only be detected at the beginning and end of shaping.

A typical signal trace for a through crack is illustrated in Figure 5. Characteristic of this is a high frequency sound oscillation with high amplitude.

In contrast, the evaluated signal trace for a surface crack is marked by a high frequency body sound oscillation with low amplitude, as appears from Figure 6.

12 By the simultaneous direct detection of the body sound at the workpiece 4 in the shaping zone and outside the shaping zone the detection of crack formation and the identification of the particular form of crack (fold, surface crack or through crack) during the drawing process is possible in real time. in this way surface cracks in the tube 4 can already be determined in the deep drawing process. Damaged tubes 4 can then be withdrawn from the production line without further ado.

Claims

13 CLAIMS

1. A method of crack detection in the shaping of metal workpieces by deep drawing in a drawing press by picking up the noise generated during shaping using a body sound sensor, including the following steps: using respective body sound sources to simultaneously and directly detect the body sound from a workpiece in the shaping zone (VZ) and outside the shaping zone (VZ) during the deep drawing process; identifying signal components with substantially the same amplitude and frequency as interfering noise by comparing the detected sound signals; processing the body sound signal detected in the shaping zone (VZ) by removing the interfering noise signals; and subjecting the processed signal to crack analysis.

2. Method according to the Claim 1, including resiliently pressing probe pins of the sound sensors against the workpiece during the deep drawing process.

3. Method according to Claim I or 2 including detecting high frequency body sound signals of high amplitude to detect through cracks in the workpiece.

4. Method according to Claim lor 2, including detecting high frequency body sound signals of small amplitude to detect cracks or folds in the workpiece.

5. Apparatus for performing the method of any one of Claims I to 4, comprising a drawing tool arranged in a press and having a counter die and a C> die, in which apparatus at least two body sound sensors are installed in the drawing tool with the probe pins thereof arranged to act on a workpiece at contact surfaces of the die and/or the counter die.

14

6. Apparatus according to Claim 5, in which the sound sensors are disposed in respective receptacles of the die "d the counter die, whereby the probe pins are guided in bores to the contact surfaces.

7. Apparatus according to Claim 5 or 6, in which the probe pins are pressed against the workpiece by the resilient biasing force of a spring.

8. Apparatus according to Claim 5 or 6, in which the probe pins are hydraulically or pneumatically pressed against the workpiece.

9. Apparatus according to any one of Claims 5 to 8, in which a body sound sensor is carried by a lever arm arranged for limited pivoting movement.