Please use this identifier to cite or link to this item: http://dx.doi.org/10.18419/opus-4158
|Title:||Online-Qualitätssicherung beim Laserstrahlbeschichten und -umschmelzen|
|Other Titles:||Quality control for laser beam cladding and laser beam remelting|
|Abstract:||Oberflächenveredelungen in Form von laserbeschichteten und umgeschmolzenen Randschichten werden meist in thermisch und mechanisch stark beanspruchten Bereichen eingesetzt. Da kleinste lokale Beschädigungen der Oberfläche verheerende Auswirkungen auf das gesamte Bauteil haben können ist die Kontrolle der Fertigungsqualität unerlässlich. Für die Überwachung von Qualitätsmerkmalen unterhalb der bearbeiteten Oberfläche wie die Spurtiefe und Spuranhaftung sind Sensoren zur Detektion der Prozessstrahlung besonders geeignet. Ebenso wird die Wirtschaftlichkeit der schnellen Laserbearbeitung durch ein Online-Prozessüberwachungssystem verstärkt, weil so teure manuelle Prüfungen oder aufwendige nachgeschaltete Prüfstationen entfallen.
Da Laseroberflächenveredelungen eine Vielzahl an Qualitätskriterien erfüllen müssen gilt es, die Schmelzbadeigenschaften aber auch verschiedene Prozesseinflussgrößen und Spurdefekte zu überwachen. Dem hohen Überwachungsaufwand wird in dieser Arbeit eine Kombination aus integralen und ortsauflösenden Sensoren entgegen gesetzt. Die Verarbeitung der von einer CMOS-Kamera empfangenen Emissionen zu Signalkurven beruht auf Bildauswertealgorithmen deren Entwicklung Teil dieser Arbeit ist. Dazu werden Bereiche der Wechselwirkungszone erfasst, in ihren Eigenschaften analysiert und zum Teil miteinander kombiniert zu Signalverläufen verarbeitet. Solche kamerabasierende aber auch aus integralen Sensoren direkt ableitbare Signalverläufe werden auf ihre Korrelation zu den Qualitätsgrößen geprüft.
Sowohl für das Laserumschmelzen als auch -beschichten wird ein Model vorgestellt, das Zusammenhänge zwischen den Vorgängen in der Wechselwirkungszone und der Prozessstrahlung ableitet. Davon ausgehend werden Algorithmen zur Überwachung von Bearbeitungsdimensionen wie Spurbreite und Spurhöhe aber auch die Erkennung von qualitätsmindernden Fehlern entwickelt. Die vorliegenden Algorithmen sind dabei großteils in der Lage, industrielle Anforderungen zu erfüllen und Qualitätsmerkmale unabhängig von der Fehlerursache und Parametereinstellung zu erkennen.
Um ein Überwachungssystem zu erhalten, das automatisiert und möglichst störungsfrei arbeitet, müssen Signalfilterungen und Signalkorrekturen abhängig von der Abtastrate des Sensors und der Prozessdynamik und –stabilität vor der eigentlichen Prozessüberwachung durchgeführt werden. Die vorgestellte Messstrategie berücksichtigt einerseits Fehlerentstehungszeitpunkte und definiert andererseits eine Bewertungsreihenfolge weil vereinzelte Qualitätsmerkmale durch Veränderung der Prozessemissionen aufgrund von Qualitätsfehlern nicht mehr klar interpretierbar sind.|
Laser beam processing is winning an ever broader area of application due to its flexibility. Included in these are surface treatements such as laser cladding and laser-remelting, which are implemented in thermally- and mechanically-stressed applications. Surfaces can be optimally adapted depending on the the type of surface procedure and the potential to use filler materials with the required properties. In this way, the surfaces of components with simple and inexpensive base materials can be processed thermally to yield mechanical stability, wear resistance, specific tribology or corrosion-resistance. The laser as a local limited energy source with associated flexibility represents thereby an ideal processing tool. Increasing quality requirements and the fact that the smallest incidence of damage on a specific surface may have wasteful effects on the entire component make inspections of the manufacturing quality a necessity. Amongst available quality assurance systems, currently in use, if monitoring of the laser and equipment is insufficient to ensure the processing quality, then an on-line process diagnostic system offers the most advantages. In addition to requiring less space and lower investment costs than a subsequent dedicated testing station in the line, the on-line process diagnostic system offers the possibility, by evaluation of process emissions directly from the interaction area, to detect additional quality criteria below the component surface. Quality characteristics such as for example the processing depth or adhesions remain hidden for offline light-assisted inspections or other controls. Manual inspections include additional disadvantages of a slower test speed and the risk of a less constant and less objective inspection in relation to an automated on-line process control. Despite considerable progress in sensor development, there is according to a literature and patent research no control system for laser remelting and -cladding available, which can demonstrate the necessary functionality, measuring accuracy and reliability for an industrial application. In this work a quality assurance system is developed which evaluates processing dimensions and detects quality defects on basis of process emissions. The investigations were carried out with three different CMOS cameras, an integral measuring temperature- and YAG-reflexsensor as well as a spectrometer and an acceleration sensor. The processing of the signals received by the CMOS camera into signal curves is based on image processing algorithms whose development is part of this doctoral work. In addition ranges of the interaction zone were measured, their characteristics analyzed and converted to signal curves. The camera based signal curves as well ascurves which are directly derivable from integral sensors are investigated on the basis of their correlation to the quality features. The investigations show that integral sensors react very well to relative changes within the process sequence. Thus, inegral sensors are better qualified to detect anomalies during the process than to estimate absolute values such as processing dimensions. In contrast, camera-sensors are capable of diagnosing not only the intensity but additionally the absolute position of the process emissions. Therefore cameras offer a comprehensive information content, which is reflected in the cognitive diversity of both measuring the absolute dimensions and recognizing short time span fluctuations generated by process disturbances. A model is presented for both laser remelting and -cladding that deduces connections between the processing in the interaction zone and the process radiation based on highspeed film data in addition to film data being recorded by the process control camera. On this basis, algorithms were developed that monitor the processing dimensions such as track width and trace height and recognize the occurance of minor process instances that effect component quality. The presented algorithms are largely able to follow industrial requirements and to detect quality features independently of the error cause and parameter setting. Indicators were developed and qualified to evaluate the processing quality for the laser remelting process. The width of the remelted trace corresponds to the width of an isotherm located in the interaction zone. The depth of the trace can be evaluated by the shift of a crescent-shaped bright area in the interaction zone on the feed motion axis. Different coating types cause changes in the intensity of the emissions and damages in the coating evoke disturbances within the respective areas of the interaction zone. The chemical change of the meltpool surface by an unsatisfactory inert gas atmosphere leads to increased beam absorption and can be measured by the proportional increase in process emissions. Porosity or gas inclusions in the basic material shift the equilibrium between laser energy and material mass, which leads to heating-up of the meltpool. The irregular temperature rises of the meltpool in this region depending on the number and size of blowholes can be clearly differentiated from the normal process. A separate evaluation area in the CMOS image with the principle of a level indicator recognises deviations in the given processing position on the basis of reference outlines. On laser cladding the width of a single cladding trace is evaluated in a fashion similar to the laser remelting by the observation of the width of an isotherme in the interaction zone. Changes in the height of the trace cause a shift of the meltpool on the feed axis in the plan view perspective of the monitoring camera. The properties of the fusion are approximated using a ratio relating the cross-section area of the fusion zone to the total cross-section area of the trace. The height and depth of the trace required are independently determined and derived by using empirical constants. The clad symmetry, which essentially depends on the relative placement of powder and nozzle, is controlled by observing the deviation of the laser beam center to the point of powder flow. An insufficient supply of inert gas can induce the formation of pores. Pores themselves cannot be recognized, however a connection between the process emissions and the supplied inert gas quantity ensure that no process conditions appear with reduced inert gas quantity, which may cause the formation of pores. Cracks, which can occur up to a certain time after process end, are detected by monitoring the impact sound waves spreading in the material. In order to aquire a monitoring system which works automatically and as trouble free as possible, signals must be filtered and corrected depending on the scanning rate, the sensor, the process dynamics and stability of the process before the actual process control be carried out. The presented measuring strategy considers the point in time at which the error arises on the one hand, whilst on the other hand it defines a measuring sequence because few quality criteria are not clearly interpretable when process defects alter the process emissions. In the case of laser remelting, the measurement begins with the process start and ends with process end, because no necessary information before and after the process is required. The first emissions at the process beginning are used to examine whether a workpiece with the correct coating is present, before quality assurance is carried out for the rest. Measurement for laser cladding processes begins before the actual process with a confirmation check of the powder nozzle position by the process control camera. During laser cladding, the information for control of the main quality criteria are collected and the process control ends with the impact sound measurement, which is continued beyond the end of process until the risk no longer exists that cracks can be generated at the cooling. Evaluation is not coupled to the absolute times at which emissions arose in order to allow a direct correlation to the quality, but controls the order in which the quality criteria can be analyzed. In the case of both remelting and cladding the evaluation sequence first checks errors that decrease the component quality, before validating the dimensions of the processed area. From a production viewpoint, the processing dimensions of a defective component are no longer of interest, as it must be scrapped anyway. From the view of process diagnostic an evaluation of the processing dimensions are no longer feasible if a process error disturbs the emissions of the meltpool in such a way that the different areas of the interaction zone cannot be identificated and analysed sufficiently. A sensor system, consisting of a CMOS camera, an acceleration sensor and a real time evaluation unit with potential to supplement further with an infrared and a YAG-reflex sensor, in combination with a measuring and an evaluation strategy is presented. It was demonstrated within the context of the system engineering and materials in this doctoral work, that this process diagnostic system is suitable for monitoring different quality criteria for the laser remelting and laser cladding processes. Based on this, it is possible to realise a transfer to real components and to pilot the real time prototype to a standard device for series production applications.
|Appears in Collections:||07 Fakultät Konstruktions-, Produktions- und Fahrzeugtechnik|
Files in This Item:
|Dissertation_Thomas_Ohmle_2009.pdf||22,06 MB||Adobe PDF||View/Open|
Items in OPUS are protected by copyright, with all rights reserved, unless otherwise indicated.