Tài liệu Cảm biến trong sản xuất P9 pdf

In addition, these devices need one of the following signals as a referencebasis for the monitoring or analysis: Acoustic Emission Sensors Short-term disturbances tool breakage or cracks

Metal Forming

E Doege, F Meiners, T Mende, W Strache, J W Yun, Institute for Metal Forming and Metal Forming Machine Tools, University of Hannover, Germany

The profitable use of advanced monitoring systems is more and more integrated

in modern mass manufacturing processes since reliable equipment is available.The idea is to improve the metal forming process due to the high availability oftools and machines by decreasing machine setup and failure times Therefore, it

is important to employ new sensor technologies in metal forming systems for theobservation of process signals

4.2.1

Sensors for the Punching Process

In the last 30 years, enormous improvements have been achieved in the stampingprocess concerning economic production, accuracy and possible shape of the parts[1] Today’s tools are more sophisticated and more expensive The costs of a mod-ern multi-stage tool can be more than $ US 100 000 and requires constant processmonitoring to achieve high availability of the tool This aspect is very importantfor the trend of just-in-time production Also, customer requirements for 100%quality control can be fulfilled with indirect quality control by the process signals.Therefore, the demand for tool safety devices and process control units is increas-ing constantly [2] Traditional limit switches [3] are not sufficient The manufac-turers’ expectations for modern process control systems are as follows:

· complete quality assurance and documentation (100% indirect product qualitycontrol);

· protection of expensive and complex multi-stage tools against breakage andsubsequent damage;

· machine overload protection;

· detection of feeding faults;

· extended production time with no supervision (ghost shifts);

· decrease of setup times and support with stored parameters;

· fewer production stoppages by premature recognition of process disturbances;

· permanent process monitoring to support the user with process information topermit optimal process setting;

· higher press speeds to increase productivity;

· control of existing tools;

· no sensor handling in the tooling room

To fulfil all these requirements, the process control system should have sensorswhich are sensitive enough to recognize the disturbances and they must guaran-tee easy handling in daily production (no cables in the tool room) The signal pro-cessing must also be very sophisticated to detect breakage, wear, and processtrends

Copyright © 2001 Wiley-VCH Verlag GmbH ISBNs: 3-527-29558-5 (Hardcover); 3-527-60002-7 (Electronic)

4.2.1.1 Sensors and Process Signals

The most common process signals for the monitoring of the punching processand the press load [5] are forces and acoustic emissions Both signals include pro-cess information, which can be controlled or analyzed by process monitoring de-vices In addition, these devices need one of the following signals as a referencebasis for the monitoring or analysis:

Acoustic Emission Sensors

Short-term disturbances (tool breakage or cracks in the product material) can beeasily detected by acoustic emission sensors [6–8] This sudden change in thepress load produces an acoustic emission signal up to 150 kHz, traveling throughthe tool and the machine Most importantly, acoustic emission sensors should beplaced as close as possible to the metal forming process to avoid disturbances (the

Fig 4.2-1 Typical signals of the punching process

machine’s vibrations) Each mechanical contact (gap) between the forming cess and the acoustic emission sensors filters the acoustic emission spectrum as alow pass Therefore, the acoustic emission sensors should be placed into the tool[6] or next to the tool In Figure 4.2-2 a piezoelectric sensor is shown with a verywide transmission band, which enables the sensor to measure acoustic emissionsignals in the 100 kHz range, because the piezoelectric element is mounted in adamping mass with no seismic mass (no resonance).

pro-Force Sensors

The most important process signals are the signals of the force sensors (see ure 4.2-3), which are placed in the structure or on the surface Piezoelectric forcesensors or piezoelectric transverse measuring pins are mostly used in the struc-ture On the surface the common devices are piezoelectric or resistive straingages A later calibration of all these sensors is necessary, because the strain andthe sensitivity of the sensors depend on the surrounded structure of the machine

Fig-or tool Existing monitFig-oring systems are mostly based on simple fFig-orce monitFig-or-ing The force signal is mainly used for process monitoring When the adjustedforce limit is exceeded, the machine will automatically be stopped by the emer-gency stop

monitor-4.2.1.2 Sensor Locations

In Figure 4.2-4 the most common sensor locations for the punching process trol are shown Acoustic emission sensors must be placed very close to the pro-cess Typical locations for the acoustic emission sensors are the upper and thelower tool or the slide and the table A greater distance to the process will in-crease the noise signal by the press

con-The force signal is normally measured by sensors which are placed in the pressframe, the connecting rods, the slide or directly in the tool [9] Some presses areequipped by the press manufacturer with sensors to protect the press against

Fig 4.2-2 Acoustic emission sensor (Kistler Instrumente AG)

force overload The distance between these sensors, which are placed in/on thepress frame or the connecting rod, and the forming process is too large to detectmore than the force overload The signals of these force sensors and the acousticemission sensor underlie many disturbances, eg, the press drive and vibrations.The best sensor signals can be obtained when the force sensors are directly placed

in the tool (see Figure 4.2-5) The second best solution for the signal quality is toplace the force sensors directly above or under the tool See the sensor plate andtable locations in Figure 4.2-4

Fig 4.2-3 Force and strain sensors for process control (Kistler Instrumente AG)

Fig 4.2-4 Possible sensor locations at a forming press [9]

4.2.1.3 Sensor Applications

In this section, sensor applications, which are close to the forming process, will

be described in detail The integration into the top plate of the upper tool isshown in Figure 4.2-5 With this application a single forming operation can beperfectly monitored The influence of a neighboring forming operation on themeasured force signal is very low Typical sensors for this application are piezo-electric transverse measuring pins or force rings, because the sensors are placed

in the structure The total or a part of the forming operation force is transmittedand measured by the sensors A disadvantage is the large number of expensivesensors in a tool and the bad tool handling in daily production The very roughenvironment in the tool shop also complicates the handling of the tools with sen-sitive sensor cables

Better tool handling and lower sensor costs can be achieved when the sensorsare integrated into machine parts or remain at the press structure One solution,which was presented by Terzyk et al [6], is the integration of force sensors intothe slots of the press table In Figure 4.2-6, two slot force sensors are shown,

Fig 4.2-5 Force sensors integrated into the upper tool [6]

Fig 4.2-6 Table slot force sensors [6]

which are placed under the lower tool The advantage of this solution is the highflexibility and the integration into existing processes, because the shape of the ta-ble slots is standardized.

On the other hand, the slots must be cleaned and must have straight surfaces.These sensors cannot be placed in the center of the tool, because there are holes

in the table in this area for scrap transportation

A good combination of process-sensitive signals and good handling is achieved

by a multi-sensor plate, which is placed between the slide and the upper tool InFigure 4.2-7 the scheme of the multi-sensor plate is presented The multi-sensorplate consists of a frame plate, which has the same shape as the slide, and severalsensor cassettes, which contain force and acoustic emission sensors The follow-ing requirements are the basis for the development of the system:

· easy handling in the production workshop;

· short distance to the process;

· integration of several force sensors for detailed process monitoring;

· connection devices for additional sensors;

· improved process control by a combined force/acoustic emission monitoring;

· modular design for high flexibility;

· integration into existing tool-press systems

Easy handling is solved by using a modular cassette system, which is fixed by aframe plate and two guiding rails to the press slide (Figure 4.2-7) During a toolchange the sensors will remain at the slide All cables between the cassettes andthe docking station are integrated in the frame plate Because of the modular de-sign, the multi-sensor plate can easily be adapted to the requirements of the user.The number and the locations of the standardized sensor cassettes can bechanged The docking station houses the charge amplifier and the connectors foradditional sensors and is mounted on the frame plate The frame plate has aheight of 25 mm and the same shape as the slide, so that the tools can be fixed tothe slide in the usual way

Fig 4.2-7 Scheme of the multi-sensor plate [9]

A multi-sensor plate with four cassettes and the docking station for a 500 kNpress is shown in Figure 4.2-8.

Some typical process signals measured with the multi-sensor plate are sented below A production tool with 11 forming operations (cutting, deep draw-ing, stamping) separated into four modules will be analyzed by means of the mul-ti-sensor plate The workpiece, the tool setup, the force and acoustic emission sig-nals are shown in Figure 4.2-9 The two force cassettes of the multi-sensor plateare placed above the first and above the last (fourth) module to demonstrate thelocal resolution of the system The acoustic emission cassette is placed in the mid-dle of the tool

pre-The measured signals contain information on the cutting/forming process, onthe blank holder and on the tool stop reaction The contact of the blank holder oc-curs at point A in the force diagram and at point 1 in the acoustic emission dia-gram Characteristic cutting operations can be identified at B/2 and C/3 The re-sulting cutting impact is very significant in the acoustic emission signal (peaks 2and 3) Owing to an incorrect slide height (too tight), the upper tool is running

on the stops of the lower tool (impact at D/4) The tight tool mounting causes anincrease in the force signals up to point E The force signal above the first mod-ule is higher than that above the last module, because the stops are in the firsttwo modules of the tool (four modules) The lower dead center is reached at point

E (highest force signal) The lift-off of the stops and of the blank holder occurs atthe moments F/5 and G/6 The force curve is evidence for the incorrect adjustedslide height (too tight)

The correlation between the force signals and the acoustic emissions in the grams is significant and the combination of the two signals permits the identifica-tion of different cutting/forming operations

dia-Fig 4.2-8 Multi-sensor plate for a 500 kN press

The sensor signals should be significant so that the user can ‘see and stand’ the complex forming operation Especially for the tool setup the stored sig-nals of previous setups can be very helpful by using the same setup and thereforesaving time and achieving the same product quality.

under-A tight slide position causes unnecessary high press forces in the lower deadcenter and product defects This load decreases the tool and the machine lifetimeand increases the energy consumption The signals in Figure 4.2-10 were mea-sured in a press shop with a production tool At the normal slide height a forcesignal of a cutting operation is measured before the lower dead center At thetight setup of the tool (0.6 mm lower) a significant second peak occurs at the low-

er dead center The stored signals of the force cassettes enable the user to setupthe tool properly with less load for the machine and the tool

Another important aspect for the monitoring of the punching process is the tection of tool breakage For this detection acoustic emission sensors should beused, because the reaction of the tool on overload and breakage is more signifi-cant in the acoustic emission signal than in the force signal The force com-presses the punch and energy will be stored in the punch After the breakage(overload), the stored energy is released as acoustic emissions to the environmentlike a compressed spring These acoustic emissions have significant amplitudesand can easily be detected in a ‘silent moment’ of the process

de-In Figure 4.2-11 the signals of force and acoustic emissions of a normal punchand of a breaking punch are shown There are only slight differences in the forcesignals Especially the ‘small valley’ around 60 ms cannot be found in the signal

Fig 4.2-10 Force signals of normal and incorrect tool setup

of the breaking punch This is the moment when the punch moves upwards inthe mold The acoustic emission signal is more significant The punch causes asecond peak at the moment of breakage This event can easily be detected with anarrow tolerance band [6] around the ‘normal’ curve.

4.2.2

Sensors for the Sheet Metal Forming Process

Sheet metal forming is a complex process which is affected by a manifold of fluences The high demands to quality and cost efficiency at the production ofsheet metal components are increasing continuously These high requirementscan only be met with optimum designed and faultless manufacturing processes.Hence it is necessary to have fundamental knowledge about the behavior of theused materials and machines as well as the possibilities for the control of the ac-tual process parameters Furthermore it is of great importance to control thecourse of events during the forming operation because the process affecting pa-rameters cannot be kept constant for any space of time Material and tool proper-ties as well as machine parameters are subjected to variations which are affectingthe process stability adversely Improvements can be achieved by the on-line mea-surement of indirect and direct process describing parameters and their transfer

in-to a process moniin-toring system

Fig 4.2-11 Force and acoustic emission signals of a breaking punch

4.2.2.1 Deep Drawing Process and Signals

Sheet metal forming processes are affected by a manifold of influences ure 4.2-12 shows the different succeeding stages of the deep drawing process.Examples of monitorable signals are material properties such as tensilestrength, anisotropy, ductility, lubrication dose, wear of tools, and adjustment offorming machines Changes in these parameters cause several failure modes such

Fig-as cracks, wrinkles, etc., and also long process starting times, production ity and deviations from the required quality [11] Differences in material chargesoften lead to a change of ductility, formability, and surface properties The periph-ery affects the forming operation by variations within the straightening process,the accuracy of blanking, and the blank position in the drawing die Also, changes

insecur-in lubrication, tool wear, and different tool positions with respect to the press areunavoidable The forming press affects the drawing result by changes in ram tilt,deflection of the press table, frame deformation, and deviations in the adjustment

of punch speed and die cushion force

4.2.2.2 Material Properties

In sheet metal forming, the working accuracy depends on mechanical propertiessuch as tensile strength, normal anisotropy, and hardening exponent These pa-rameters fluctuate from coil to coil and charge in the ranges of the specified toler-ances The increasing standard quality requirements for the process and productsdemand a testing method which is capable of monitoring these material proper-ties on-line and prior to the deep drawing process Applying the magnetoinductive

Fig 4.2-12 Deep drawing process chain and monitorable signals

testing method, a sensor is inserted into the process chain after reeling off theblank from a coil The sensor head is held at a defined distance above the sheetmaterial and an exciting signal is brought to a magnetic coil to induce a magneticfield in the material (Figure 4.2-13).

The signals received mainly depend on the microstructural composition such asgrain size, grain orientation, alloying elements, and dislocation density Further,the resulting electromagnetic properties are correlated with mechanical parame-ters which were determined previously in tensile tests The dependences on theproperties are shown in Figure 4.2-14

By using correlation statistics, a multiple regression equation allows the tion of mechanical properties directly from the magnetoinductive measurements

predic-Fig 4.2-13 Determination of the magneto-inductive signal parameters [12]

Fig 4.2-14 Dependences on magnetic, material, and mechanical properties

and nondestructive testing method to avoid wrinkles and cracks caused by ance deviations in the sheet quality.

toler-4.2.2.3 Lubrication

The lubrication properties affect the formability during the deep drawing process.The importance of control and analysis of the lubrication properties has signifi-cantly increased in pressing processes owing to the introduction of new genera-tions of automatic transfer presses The control of incoming material gives possi-bilities to reject the material before further processing them The yield of thepressing process will increase, giving savings of material and production costs[13]

Pressforming processes require uniformity of oil films on the metal surface.During deep drawing the oil film separates the sheet metal from the die to allowthe material to flow constantly between blankholder and die The use of oil avoidscold welding of the steel on the active tool surfaces which can cause galling andpassing of the friction force limit As a result, the deep drawing process fails ow-ing to cracks in the material Galling means the formation of cold welds betweenblank sheet material, especially stainless steel and aluminium alloys with diematerial at high local pressures During sliding these welds shear off and causescratches in the material Another important process parameter is the necessaryblankholder force This force is affected directly by the friction coefficient whichdepends on the quantity of the lubricant as shown in Figure 4.2-15

In deep drawing processes, the blankholder force will be kept at a defined level

to reach a defined surface pressing Differences in the amount of lubricationcause deviations from the acceptable tolerance zone For an increased amount of

Fig 4.2-15 3D tolerance zone with dependence of lubrication, blankholder force, and drawing distance [14]

inter-lubrication the process will fail owing to wrinkles, whereas for reduced amountsthe deep drawing part will tear off.

A portable sensor based on high-resolution infrared spectrometry has been veloped for the measurement of oil film thickness on metal surfaces This light-weight hand-held device is intended for use in rolling mills and sheet metal engi-neering workshops The sensor permits the measurement of thin oil films and isuseful for optimizing the thickness of oil coatings or pressforming lubricants Fig-ure 4.2-16 shows a schematic diagram of the analyzer, a two-part system consist-ing of a measurement head and data collection unit The analytical measurementprinciple is based on the absorption of infrared radiation by hydrocarbons, thecommon constituent in all oils

de-The optical measurement head includes a compact multichannel infrared zer, electronics, and an LCD display The optics, mechanical parts, light source,and multichannel detector electronics are integrated into the measurement head

analy-to provide stable, high-resolution analyses in a production environment A controlunit includes a data processing unit, LCD display, keypad, and PC interface Thecontrol unit collects measurement data and calculates oil amounts in terms ofweight per unit area by comparing measured data with pre-calibration curvesstored in memory

The measurement is performed by placing the sensor head on the metal face (Figure 4.2-17) Spacer pins at the measurement head stabilize the fixed dis-tance between the measuring head and the surface After triggering, the mea-sured amount of oil is displayed on both the measurement and control units Theactual result is derived as an average value from several sub-results measured atdifferent points on the surface

sur-Owing to optical differences in the surface texture of materials such as rolled and hot-rolled steel, copper, and aluminium, the analyzer is calibrated forthe type of surface to be measured Calibration also eliminates effects caused bypossible differences in oil quality It is recommended that each calibration ismade using the same type of surface and oil as is expected in actual measure-ment The repeatability of the analyzer, which can be expressed by the standarddeviation of readings in a single-point measurement, depends on the oil filmthickness In the case of cleaned cold-rolled steel the standard deviation is of the

cold-Fig 4.2-16 Schematic diagram of the infrared analyzer

order of 1 g/m2 The influence of surface textures increases the standard deviationwhen measurements are performed at separate points on the surface.

4.2.2.4 In-Process Control for the Deep Drawing Process

Nowadays, deep drawing processes are controlled on the basis of predeterminedstatic values Considering the heavy demands on the quality of deep drawn com-ponents and low production costs, it is necessary to observe process-influencingparameters In a first step, higher process security can be obtained with the practi-cal operation of multiple sensors located directly in the drawing process

For process monitoring, direct process-based and time-dependent informationfor the characterization of the process course have to be available This is very dif-ficult because the forming takes place in closed tools at high forces Therefore, it

is not possible to react automatically to parameter changes which occur, eg, withthe use of another coil with different forming or surface properties [15]

Flange Insertion Sensor

For the consideration of the deep drawing process, the measurement of the flangeinsertion offers information which contains a reliable prediction of the progress

of the deep drawing process and further of the part quality A flange insertionsensor has been developed to measure the flange insertion distance and to drawconclusions regarding stress and strain [16]

The sensor consists of an inductive position sensor with a thin metal tongue atthe top The tongue has a thickness of 0.5 mm and is brought into the gap be-tween blankholder and die to touch the outer edge of the blank sheet from the be-ginning to the end of the deep drawing process (Figure 4.2-18) The flange inser-tion is measured over the drawing distance and can be used to detect a deviationfrom the tolerance field describing the non-failure area Deviations can lead towrinkles and cracks in the drawing part

Fig 4.2-17 Infrared sensor head placed on the metal surface