Patent Application
20170052149
The invention relates to detection of defects and faults within manufacturing processes of elongated metal materials.
Manufacturing of elongated metal materials such as slabs, billets, strands strips, sheets or bars etc. typically starts so that molten metal is solidified to have a preformed cross section. The resulted preformed state may be a discontinuous billet or slab or the process may produce an endless slab, like a continuous casting process does. After the molten metal is solidified, it is typically subjected to forming the material mechanically to reduce the cross section of it. First forming of the solidified metal typically happens in a hot state when the material is soft and ductile and the cross section of it can be extensively changed.
In cold state the material has to be ductile enough to be able to keep its integrity during the following forming stages. Between cold forming stages, the material may be reheated to recrystallize the structure in order to make it possible to form the elongated material further. To maximize the production rate and to minimize the number of any forming stages, the deformation rate is adjusted to be close to the limits of ductility of the material. Typically the metal strip is cut into shorter pieces to make the handling and later processing possible. The mostly used process for reducing the cross section of the elongated metal material is forming between rolls i.e. rolling. Drawing may also be used when producing bar or tubular like forms in which the width and height of the formed cross section are quite equal. In cold rolling and especially in a drawing process a remarkable drawing force will create tensile stress and increase risks of initiating and advancing of cracks during the process. In a rolling process, compressive feeding force can be added by driven forming rolls.
As the thickness of a metal strip is reduced during rolling, the length of it is increased. The thinner the strip is, the less the width of the strip will increase. Especially in the cold rolling stages, edges of the strip are most prone to cracking as the metal sheet is not supported in the lateral direction. In any forming processes also for example impurities or local inhomogeneity of alloying elements can induce cracking. The greater the deformation rate is and the harder and less ductile the formed metal is, the more susceptible it is to cracking. The cracks will mostly be in the transverse direction of a strip if the crack is not caused by metallurgical reasons. If there will be later forming phases, the crack will grow bigger. If the strip is under tension during forming, the crack can grow through the cross section and the strip may be broken transversally. The breakage of the material will cause damage i.e. to the rolls and also the surrounding periphery may get damaged. This leads to a quite long downtime and replacement or repairs of the expensive tooling.
Another stage prone to cracking is in the solidification phase, especially when the metal is highly alloyed. Differently cooled areas may then cause overstressing of the material which has fragile alloying elements concentrated during the solidification phase to the middle section of the material. Impurities can also concentrate equally and induce cracking in any later forming stage. Such cracking will often occur in the planar direction. Ragged surface and also internal defects like voids may also be formed during the casting process. Any cracking or other defects should be detected as early as possible to avoid further processing of the damaged area of the material and to avoid the growth of the defect and further increase of costs related to quality and incurred by repairing or scrapping processed material.
Non-destructing testing (NDT) such as for example using ultrasonic sensors, which receive signals or echoes which ultrasonic transmitters have emitted, have been tried for detecting cracking online during production of elongated materials and offline between process stages. During hot forming and directly after continuous casting, the high temperature creates the main problem when using ultrasonic testing which need a good contact to the metal to be able to transmit and receive the signals. Especially during cold forming the processing speed of the material excessively hinders getting the contacts in a stable state to deliver the sound waves and the speed also limits the time needed to detect the fast moving faults. Ultrasonic NDT methods are good at detecting faults that have a substantial planar component when the thickness of the strip is several millimeters. The thinner the strip is such as in cold forming stages, the more unreliable the method is for detecting typical defects that are transverse to the plane of the strip and formed especially when the feeding force is a drawing force. Some examples of the ultrasonic NDT of rolling processes are presented in publications U.S. Pat. No. 7,987,719 and U.S. Pat. No. 4,160,387.
Acoustic emission (AE) is widely used for monitoring the condition of highly stressed components such as pressure vessels and rotating objects. The method can detect abnormalities such as excess friction, damaged bearings or gears, unwanted contacts to other components and crack formation. An example of using AE in monitoring of faults and defects of rotary machinery is disclosed in publication U.S. Pat. No. 4,669,315.
The purpose of the invention is to reliably detect defects initiated and advanced during manufacturing processes of elongated metal materials. Preferably the defects are detected in as early a stage as possible and the position of the defect will also be located for earliest possible corrective actions with least extent. The purpose is achieved such that the method and/or apparatus defined in the preamble of the independent claims are implemented as defined in the characterizing part of the claims. Preferred embodiments of the invention may correspond to the dependent claims.
The invention is based on detection of acoustic emission, which is emitted during crack formation as a crack is originated or advanced when the elongated material is solidified or deformed. Also other surface or internal defects and process faults may cause distinguishable abnormal acoustic emission during a manufacturing process. The type of defect can mostly be recognized by the characteristic properties of the acoustic emission and usually it is useful to know that in a located place some kind of defect may exist. At least one acoustic emission (AE) sensor is used to receive the acoustic emission from the originating or advancing defect and to transduce it to indicative electric signals. Filtering, frequency analysis and other digital signal analyzing and processing means may be used to separate background noise generated by bearings, other process conditions or devices for detecting the indications of the originating or advancing defects.
A defect of an elongated metal material will usually be in form of a crack. Other material defects such as grooves, notches or other surface defects as well as internal defects such as voids may also be detected especially when the material is deformed. Different types of defects usually give different indications so that the frequency is in a certain range and over a threshold amplitude or the duration of an occasional signal may have a characterizing and/or quantitative indication. For detecting certain types of defects, an AE sensor tuned to be sensitive to a certain narrow frequency area may be used to filter out disturbing signals. Mostly the monitored processes are recurrent and therefore the different indications can be examined and learned and also information about the extent of the defect can be calculated based on the frequency, amplitude and/or duration of the indicating signal.
The AE sensor needs a path with good contact to the monitored material to function. The generated acoustic emission from the defect travels well quite a long way through metals and liquids to the sensor. Therefore the AE sensor does not need to be very close to the originating or advancing defect, but the closer and the better the contact is the better is the selectivity to get the indication from the defect. Also any boundaries between participative parts should be avoided.
If only one forming stage is in the production line, only one AE sensor is needed to indicate and locate accurately a defect in the longitudinal direction as the defect will originate or advance in the forming position. If the corrective action is to split transversely the material in that point to remove the defect, that information is enough to direct the cutting operation to the calculated or metered lengthwise position. Also, if the defect will be analyzed by other NDT methods before correcting operations, the lengthwise information is mostly adequate for fast finding of the defect. Transversely cutting of cold formed coiled strip is often not the best corrective action as it leads to more discontinuities in the production processes. Usually longitudinal trimming of the coiled strip is a preferred and more cost effective correcting action if there is a crack of limited length at an edge of a strip.
Preferably indications of the indicated defects and the indicated sizes of them are stored so that the trend of indicated defects can be used to adjust the process parameters for preventing origination of defects. Excessive existence of defects may also be caused by inadequate pretreatment or metallurgical properties of the processed metal being out of tolerance. Halting a process or alarming an operator of it may also be needed to perform, if an oversized defect is determined by calculation or an indicative amplitude exceeds a predetermined indicative threshold value of an excessive defect.
A monitoring system for detecting and reporting defects comprises of at least one AE sensor 9 which is in an acoustic path contact to a processed elongated material 1 and between different manufacturing stages and lines may be used to further process and report the indications and perform corrective actions.
If the need is to locate the defect in the lateral direction of a strip, the task can be solved by positioning at least a pair of AE sensors to different transverse distances from an edge of the strip. The longitudinal position should preferably be substantially equal. Then there will be difference in the receiving time and typically in the amplitude of the indication of the defect. If the strip to be rolled is thin, the location of the defect is presumably on one of the two edges of the strip. Then comparing timing of the indications received by the pair of AE sensors will lead to a conclusion on which side the defect is located. The conclusion based on the timing is not relevant if the time difference does not correspond to the timing difference of acoustic paths from an edge of the strip to the AE sensors. The stronger and/or the sooner indication will be at the side of the defect. Even if the amplitudes are substantially at the same or slightly conflicting level, the earlier receiving time is a stronger indication of the side of the defect. If the strength of the amplitude and the receiving time are in a stronger conflict against these principles, the indications are supposed to represent two different defects at different sides of the sheet. By knowing the side of the defect and the level of the amplitudes, the monitoring system can calculate a report on which side of ?? the strip should be trimmed, how much and at which longitudinal location. If the whole length of a defected strip will be trimmed, at least from one damaged side, the information of the longitudinal position of the defect is irrelevant. If there is a need to find and examine the defect more accurately for comparing it to the calculated extent of it, the longitudinal location is still needed.
By having the first indication about a crack as soon as possible, the loss of material in trimming will be less than if the sheet is rolled thinner in the next production line or next pass in the same line because normally the crack will grow. If the time values of the indications do not differ as much as they should according to the time differences in acoustic paths, the defect may be within the inner area of the sheet and the location should be inspected or be cut away and scrapped. In this case there might also be two separate defects which have given indications.
Acoustic emission may not travel well through a forming process as the forming itself creates acoustic emission and the waves may be otherwise damped. If there is more than one rolling stage in the forming process line, each stage may be needed to be monitored separately as the indication may not reach the AE sensors through another rolling stage. There should be at least one AE sensing point for two forming stages. An AE sensor in contact with the elongated metal material between two forming stages can detect indication of a defect from both directions. If a crack is formed in the first stage, an indication of the defect will happen also at the second stage as the crack will eventually grow then. If there are two indications of a defect which have equal time difference that is the time difference between the two stages there supposedly is only one defect which has given two indications. If there will be only one indication, the crack has initiated at the second stage.
If the size of the defect needs to be evaluated, all forming stages and their corresponding AE sensors and analyzing units should be calibrated to correspond to the size of the indication. By examining the detected defects, an equation between the detected indicative amplitude and the size of the defect can be created. The equation will be affected by different combinations of process parameters and processed materials.
Continuous casting is typically a vertical process. After solidification of at least the surfaces of the metal, the process is turned to be a horizontal forming process as the material is in a soft state. The turn is made by rolls that typically do not deform the metal thinner but they just direct it to move horizontally. During solidification and at later cooling and hot deforming phases with especially heavily alloyed or otherwise less ductile alloys may occur hot cracking or deformation cracking in most intensive hot deforming stages. A moving hot rough surface in the solidification stage and hot forming stages is not a good contact surface for an AE sensor as the contact itself will create noise. As the material is hot, the AE sensors cannot contact directly the processed material. Also there will typically be slag on the surfaces of especially steels or the surface is too rough to have direct and reliable contact. Preferably the AE sensors are connected indirectly via rolls or their supporting structure or rollers to suppress conduction of heat. The sensors may be of a type which resists elevated temperatures.
Locating of an originating crack or other defect in the solidification phase is not as accurate as in the forming stages as the point where the cracking will occur can be in a longer area. As the crack is later advanced in a deforming stage, the advancing crack or other defect can be more accurately positioned. Then the locating procedure is quite similar to the prior described. As the continuous casting process and the hot forming stages after solidification are continuous processes, it is difficult to make corrective actions during the process before the metal is cut into shorter pieces. If the defects are only local and occasional, the process needs not be cancelled but maybe it needs to be adjusted to eliminate the formation of defects. When accomplishing corrective actions to continuously casted billets or slabs, the location information of the defects is critical in finding them fast. At the same time of localizing the defect, the originating stage of it can be detected too. That information helps to analyze which corrective actions are needed to adjust the parameters of the defect-prone process or stage in the often long continuous casting process line with several successive forming stages.
AE sensors can be arrangedin their detecting positions in many ways. The simplest solution is to attach the AE sensor to the supporting structure or frame of the processing device or component which is in contact with the processed material. An acoustic emission path to the AE sensor may be via a forming roll, supporting roll or other roll or roller. In most solutions this is an adequate way. The elongated metal material has a good contact to the roll or roller and the acoustic path will go through it and its bearings to the bearing housing and the path continues also to the other support structure of the roll such as the frame of the device. Acoustic emission travels very fast and it does not damp aggressively in metals. Any junction or boundary that the acoustic emission needs to travel through can dampen it and can cause echoing. Also any other moving elements such as the supporting rolls and the bearings will create background acoustic emission of different levels and frequencies. Therefore the acoustic path from the originating or advancing crack to the AE sensor should be as short as possible for maximum resolution of indications.
A deforming metal will create continuous acoustic emission due to dislocations and other ductile forms of deformation. This continuous acoustic emission needs to be filtered out to extract data concerning the occasional short high amplitude acoustic emission bursts created by defects.
The surface of a processed strip is usually smooth and cool enough in cold forming stages for having a reliable direct contact to the strip. The AE sensor can be most directly contacted to the moving surface via a rolling means or via a sliding means. Both types of direct contacts may be lubricated by oil or other liquid to get a better signal path and on sliding contact to dampen or eliminate acoustic emission originating from friction. A liquid will conduct the acoustic emission well to the AE sensor. The directly contacted AE sensor can easily be isolated from the background signals generated by surrounding machinery and mainly acoustic emission from the processed target is received.
Due to temperature, surface properties and lubrication problems, a rolling contact AE sensor is a preferable direct contact solution for monitoring of hot processes. Using contacting intermediate material that is liquid at elevated temperatures may cause more trouble, such as contamination, than give advantage in the sensing task. When a direct contact to the metal material is used, an AE sensor may be isolated and/or cooled to prevent overheating.
When the AE sensors are attached to the supporting structure of a rolling machine, the same collected data can be filtered and examined also by other means for monitoring the whole process to find out also abnormalities or unwanted functions of the attached components or process parameters. For example worn out bearings and gears, lack of lubrication of the process or moving parts, defected surfaces of the rolls, loose parts, dirt like burr stuck on moving surfaces or fractured components can be detected and reported by the same monitoring system.
In the accompanying drawings:
After the casted metal is solidified to a strand 1, it is either hot formed thinner by one or several rolling stages or it is cut into defined material lengths. This feed stock can be further processed in hot or cold forming processes into sheets 1 or coils 8 (
A pair of AE sensors 9 are arranged to be in contact with the strip 1 via a roller 11 after the earlier forming stage 5 to monitor the forming process. Any other forming stage 5 may be equally monitored. A sliding contact could be used instead of the rolling contact. The two AE sensors 9 are located close to the both edges of the strip 1. The closer the AE sensors 9 are to the edges and work roll 6 of the forming stage 5, the bigger is the time difference and difference in amplitude between AE signals of the two AE sensors 9, which signals are indications of a single defect originated or advanced at an edge of the strip under the forming stage 5. The time differences in an acoustic path from one edge to the AE sensors of different sides can be calculated or determined for further use from time coded indicative signals of earlier recorded and confirmed cracks.
Typically the same defect that has existed in the earlier forming stage will advance in the successive forming stage 5. That emission from the advance of the defect may also be received by the same AE sensors 9. Due to different dampening properties equation for calculating the size of the defect should be adapted to fit to the changed circumstantial factors. Also due to dampening and tiny time difference due to long distances, locating the side of the originating defect in the successive forming stage 5 will be less reliable or impossible.
Due to applied strip 1 tension, the defect more typically originates and advances on the last moments of the forming stage 5 and it can still advance after passing the work roll 6. Therefore the AE sensor 9 should be positioned in the direction of movement of the strip 1 after the forming stage 5 which is monitored by the AE sensor 9.
An advantage of arranging an AE sensor to the bearing housings or other support structure of the work roll 6 or to the frame of the process machine is that from the same signal data other abnormal process conditions and faults in the components can be detected and monitored simultaneously. Since the characterizing frequencies, durations and/or levels of other acoustic emissions of faults are known by research or other sources, the same signals can be filtered and/or otherwise analyzed. Then any abnormalities are detected and reported or at least a report is created of a need to analyze the process and/or devices more deeply before any product quality issues or mechanical faults will happen.
A threshold value may be defined for an excess size of a determined crack which can cause breakage of the strip 1 and wider damage to process devices. If an excess size of a crack is determined and reported by analyzing unit 21 to process control 22, the process control can halt the production line to avoid potential damages. The analyzing unit may only report the level and the time code of the peak values and/or durations of the indicated or suspected cracks and the process control may be programmed to analyze and calculate the extent of the cracks and the overall extent or length of the advanced cracks.
If the time difference between these indications corresponds to the time difference between the paths from a side of a strip 1 to the pair of AE sensors it can be reported to the process control 22 that a crack occurred to the side on which the AE sensor received the acoustic emission first.
If a pair of AE sensors 9 are in successive positions in the longitudinal direction of the process, the corresponding analysis can report at least approximately the location of the defect. This kind of arrangement can be used for example for locating the position of a defect in a continuous casting process or in the successive hot forming processes. If the time difference is less than the time difference between the advance of the elongated metal material 1 in the process between the positions of the AE sensors 9, the location can be defined quite accurately. If the time difference is equal, there is only one defect before the earlier AE sensor 9 or after the later AE sensor 9. As the AE sensor 9 which receives first the indication is closer, it can be determined which one of the two situations is correct. If the time difference is longer than the time difference between the positions, there should be two separate defects for which only approximate locations can be reported. If the determination of the location is only approximate, other NDT methods are needed to find the exact place of the defect from the reported approximated location.
The monitoring system according to an example of
For example a Vallen VS-150-M sensor is tuned to 150 KHz and will receive frequencies between 50 and 500 kHz. This type of AE sensor 9 will itself filter out most of operational noises and can give selective indications from crack formations in cold processing of carbon steels. Different metals and metal alloys may need differently tuned or wideband AE sensors 9 for indications of certain defects. Also different types of defects and faults, temperatures, drawing conditions thicknesses and other process parameters may lead to using an AE sensor 9 of certain tuning characteristics. A person skilled in the art can use a frequency analyzer or a set of different sensors to find the best indicative characteristics.
An analyzing unit 21 will log and time code the signals received from connected AE sensors 9. The analyzing unit 21 can filter out the general background noise so that indications from defects and the amplitudes of them will be clear. The analyzing unit 21 may report to a process controller 22 time codes and amplitudes of indications which exceed a predetermined indicative threshold value. An indication may be categorized and reported to represent a suspected defect if for a suspected defect is predetermined a threshold value which is smaller than the threshold value which indicates a detected defect. The analyzing unit 21 or the process controller 22 may deliver the report to the operator by for example visual report on a monitor. An audible alert may also be generated as an indication of a defect or a suspected defect. Preferably both the operator and the process controller 22 will receive a report.
The analyzing unit 21 may be programmed to analyze a location of a defect by comparing indications of defects from two or more AE sensors. For example it can be programmed to follow the earlier described methods to report on which side the defect is or if there are two independent defects. Built-in features for locating indicated defects are also commercially available in for example Vallen AMSY-6 software. The analyzing unit 21 may be an independent device based on a PC with a multichannel PCI-card or the analyzing unit 21 may even be integrated into a process controller 22. Also modern AE-sensors 9 with integrated analyzing units 21 can be used, these sensors have a capacity to process AE-signals and send individual analyzing results with accurate time code to the process controller 22 or other analyzing unit 21 for further processing and reporting.
The process controller 22 controls the process parameters and operations of connected production stages of a production line. It will also receive and keep up information about the longitudinal position and processing speed of the processed elongated material 1. The position information may be obtained for example from a speed and distance metering roller or from a support roll 3. As the speed and longitudinal position of the processed material, positions of the AE sensors 9, and the time codes of indicated defects are known, the longitudinal location can be quite accurately calculated. The process controller 22 may be programmed to perform spraying a color or other visual marking of the elongated metal material at the calculated location of the defect to speed up finding the defect manually.
The analyzing unit 21 or process controller may be programmed to calculate an estimated size of a detected defect. This information can be transmitted and reported to an operator of a trimming stage 7. In
The process controller 22 may log and create a report of all of the indicated defects. This log can be used to determine if there is a need to adjust parameters of the controlled processes. If a certain amount of indications of defects is accumulated within a predetermined time interval, the operator can be alarmed to consider modifying the process parameters. Also if an amplitude or calculated size of an indicated individual defect will be over a predetermined threshold level, the process controller 22 may be programmed to halt the production line to avoid a potential breakage of the processed elongated material 1 and to avert wider damages.
| Number | Date | Country | Kind |
|---|---|---|---|
| 20145157 | Feb 2014 | FI | national |
| Filing Document | Filing Date | Country | Kind |
|---|---|---|---|
| PCT/EP2015/053199 | 2/16/2015 | WO | 00 |
