Felix Abt

New CNN based algorithms for the full penetration hole extraction in laser welding processes: Experimental results.

In this paper the results obtained by the use of new CNN based visual algorithms for the control of welding processes are described. The growing number of laser welding applications from automobile production to micro mechanics requires fast systems to create closed loop control for error prevention and correction. Nowadays the image processing frame rates of conventional architectures [1] are not sufficient to control high speed laser welding processes due to the fast fluctuation of the full penetration hole [3]. This paper focuses the attention on new strategies obtained by the use of the Eye-RIS system v1.2 which includes a pixel parallel Cellular Neural Network (CNN) based architecture called Q-Eye [2]. In particular, new algorithms for the full penetration hole detection with frame rates up to 24 kHz will be presented. Finally, the results obtained performing real time control of welding processes by the use of these algorithms will be discussed.

Feature extraction in laser welding processes

There is a rapidly growing demand for laser welding in a wide variety of manufacturing processes ranging from automobile production to precision mechanics. Up to now, the high dynamics of the process has made it impossible to construct a camera based real time quality and process control. Since new pixel parallel architectures are existing, which are now available in systems such as the ACE16k, Q-Eye, and SCAMP-3 (P. Dudek et al., 2006), one has become able to implement a real time laser welding processing. In this paper we will propose a feature extraction algorithm, running at a frame rate of 10 kHz, for a laser welding process. The performance of the algorithm has been studied in detail. In particular, it has been implemented on an Eye-RIS v.1.1 system and has been applied to laser welding processes.

Cellular Neural Network (CNN) based control algorithms for omnidirectional laser welding processes: Experimental results

The high dynamics of laser beam welding (LBW) in several manufacturing processes ranging from automobile production to precision mechanics requires the introduction of new fast real time controls. In the last few years, algorithms for the control of constant-orientation LBW processes have been introduced. Nevertheless, some real life processes are also performed changing the welding orientation during the process. In this paper experimental results obtained by the use of a new CNN based strategy for the control of curved welding seams are discussed. It is based on the real time adjustment of the laser power by the detection of the full penetration hole in process images. The control algorithm has been implemented on the Eye-RIS system v1.2 leading to a visual closed loop control solution with controlling rates up to 6 kHz.

New CNN based algorithms for the full penetration hole extraction in laser welding processes

In this paper new CNN based visual algorithms for the control of welding processes are proposed. The high dynamics of laser welding in several manufacturing processes ranging from automobile production to precision mechanics requires the introduction of new fast real time controls. In the last few years, analogic circuits like Cellular Neural Networks (CNN) have obtained a primary place in the development of efficient electronic devices because of their real-time signal processing properties. Furthermore, several pixel parallel CNN based architectures are now included within devices like the family of EyeRis systems [1]. In particular, the algorithms proposed in the following have been implemented on the EyeRis system v1.2 with the aim to be run at frame rates up to 20 kHz.

Real-Time Control of Laser Beam Welding Processes: Reality

Cellular neural networks (CNN) are more and more attractive for closed-loop control systems based on image processing because they allow for the combination of high computational power and short feedback times. This combination enables new applications, which are not feasible for conventional image processing systems. Laser beam welding (LBW), which has been largely adopted in the industrial scenario, is an example for such processes. Concerning the latter, monitoring systems using conventional cameras are quite common, but they do a statistical postprocess evaluation of certain image features for quality control purposes. Earlier attempts to build closed-loop control systems failed due to the lack of computational power. In order to increase controlling rates and decrease false detections by a more robust evaluation of the image feature, strategies based on CNN operations have been implemented in a cellular architecture called Q-Eye. They allow enabling the first robust closed-loop control system adapting the laser power by observing the full penetration hole (FPH) in the melt. In this paper, the algorithms adopted for the FPH detection in process images are described and compared. Furthermore, experimental results obtained in real-time applications are also discussed.

Omnidirectional algorithm for the full penetration hole extraction in laser welding processes

In this paper a new Cellular Neural Network (CNN) based visual algorithm for welding processes is proposed. The idea described in [1] can be used in processes, whose welding direction has a constant orientation well known a priori. The algorithm proposed in the following is omnidirectional in the sense that it does not depend on the welding direction. This fact enables closed loop control systems for welding processes with curved seeds. On Eye-RIS systems [2] processing times of about 110 μs are achievable for both acquisition and evaluation of full frame images.

Online measurement and closed loop control of penetration depth in laser welding processes

The key quality factor for reliable deep penetration laser welding processes is a correct penetration depth. Today, this parameter is only accessible by offline quality control methods. Since the reliability of a butt-joint strongly depends on the penetration depth, it is desirable to have a possibility to measure and control the penetration depth online during the welding process.A sensor for measuring the penetration depth has to fulfill several requirements: Illumination and detection has to be coaxially to the high power laser, since the capillary can be very deep and slender. The measurement has to work despite of a very high dynamic range of the detected light and despite of strong background radiation from the welding process. The sensor should realize scan rates in the kilohertz range to enable real time closed loop control of the welding process with high control frequency. The penetration depth should be measured with a resolution in the range of 100 µm.Today no sensor exists that fulfills these requirements. Different approaches, like in [1] and [2], were discussed in the past, but none of them has emerged to a fast and reliable measuring product.The key quality factor for reliable deep penetration laser welding processes is a correct penetration depth. Today, this parameter is only accessible by offline quality control methods. Since the reliability of a butt-joint strongly depends on the penetration depth, it is desirable to have a possibility to measure and control the penetration depth online during the welding process.A sensor for measuring the penetration depth has to fulfill several requirements: Illumination and detection has to be coaxially to the high power laser, since the capillary can be very deep and slender. The measurement has to work despite of a very high dynamic range of the detected light and despite of strong background radiation from the welding process. The sensor should realize scan rates in the kilohertz range to enable real time closed loop control of the welding process with high control frequency. The penetration depth should be measured with a resolution in the range of 100 µm.Today no sensor exists that fulfills thes...

A camera based closed loop control system for keyhole welding processes: Algorithm comparison

Real time monitoring of laser welding has a more and more importance in several manufacturing processes ranging from automobile production to precision mechanics. Despite the huge improvement in welding technology, sophisticated image based closed loop control systems have not been integrated in commercially available equipments yet. Due to the high dynamics of laser beam welding (LBW) processes, robust closed loop control systems require fast real time image processing with frame rates in the multi kilo Hertz range. In the last few years, some new high speed Cellular Neural Network (CNN) based algorithms for the full penetration hole detection in keyhole welding processes have been introduced. In particular, they can be distinguished in two categories: Orientation dependent and orientation independent algorithms. The former can be used only for the welding of straight lines, while the latter has been implemented for the control of curved weld seams. Both algorithms have been used to build up a real time closed loop control system for LBW processes. An algorithm comparison by the description of some experimental results is addressed in this paper.

Formation mechanism of process instabilities and strategies to improve welding quality

The deep penetration welding process with CO2 lasers has been employed successfully for many years in industry. It generates little spatter on the work piece surface and therefore produces excellent seam quality at high process speed with good process reliability over a wide range of parameters. With a wavelength around 1 µm, solid-state lasers are being increasingly used in industrial production thanks to simple beam guidance by means of laser light cable and their high electrical efficiency. Disk and fiber lasers have advanced into the domain of CO2 lasers by way of power and beam parameter product. However the seam quality is highly dependent on the focusing conditions used, whereby the mechanisms that cause process instabilities are still not properly understood. Also, at high intensities with high feed rates, considerable spatter is generated on the work piece surface, reducing productivity in applications where the demands on surface quality are high. In this publication, based on online X-ray observation and high-speed imaging, the suitability for welding and the formation mechanism of process instabilities like spattering and humping were compared at different feed rates. Based on a better process understanding strategies which can improve welding quality are presented.The deep penetration welding process with CO2 lasers has been employed successfully for many years in industry. It generates little spatter on the work piece surface and therefore produces excellent seam quality at high process speed with good process reliability over a wide range of parameters. With a wavelength around 1 µm, solid-state lasers are being increasingly used in industrial production thanks to simple beam guidance by means of laser light cable and their high electrical efficiency. Disk and fiber lasers have advanced into the domain of CO2 lasers by way of power and beam parameter product. However the seam quality is highly dependent on the focusing conditions used, whereby the mechanisms that cause process instabilities are still not properly understood. Also, at high intensities with high feed rates, considerable spatter is generated on the work piece surface, reducing productivity in applications where the demands on surface quality are high. In this publication, based on online X-ray obser...

X-ray videography for investigation of capillary and melt pool dynamics in different materials

The investigation of capillary and melt pool dynamics in laser welding processes with conventional diagnostic equipment is very challenging. High-speed cameras in the visual and infrared spectrum are offering brilliant image quality and outstanding frame rates but they only access the behavior of the process surface. With these conventional techniques it is thus not possible to observe the key mechanisms inside the volume of the material which essentially determine the process.To gain insight into the process dynamics phenomena, such as the shape and movement of the capillary or the actual melt flow in the weld bead, X-ray videography is an ideal instrument. However, state of the art high-brightness laser processes require an X-ray system which is capable of imaging capillary diameters of about 100 µm at frame rates between 1,000 fps and 5,000 fps.This paper describes different experiments and their results, realized with the high-speed micro focus X-ray system of the IFSW that was presented and characterized in recent publications like [1] and [2]. The experiments include different materials like steel, aluminum and bronze, which were welded with focal diameters between 150 µm and 600 µm at feed rates of up to 30 m/min. For the visualization of the fluid dynamics, tracer materials were used in order to visualize melt flows with minimum falsification of the welding process.The investigation of capillary and melt pool dynamics in laser welding processes with conventional diagnostic equipment is very challenging. High-speed cameras in the visual and infrared spectrum are offering brilliant image quality and outstanding frame rates but they only access the behavior of the process surface. With these conventional techniques it is thus not possible to observe the key mechanisms inside the volume of the material which essentially determine the process.To gain insight into the process dynamics phenomena, such as the shape and movement of the capillary or the actual melt flow in the weld bead, X-ray videography is an ideal instrument. However, state of the art high-brightness laser processes require an X-ray system which is capable of imaging capillary diameters of about 100 µm at frame rates between 1,000 fps and 5,000 fps.This paper describes different experiments and their results, realized with the high-speed micro focus X-ray system of the IFSW that was presented and character...