ADAPTIVE MODULE FOR MONITORING WELDING PROCESSES IN MANUAL WELDING WITH MOTION SENSOR

Abstract

The invention relates to an adaptive module for monitoring welding processes during manual welding using motion sensors, wherein the adaptive module comprises at least one energy unit, a computer unit with at least one sensor and a mounting unit. Furthermore, the invention relates to a use of the adaptive module.

Description

[0001] The invention relates to a device as an adaptive module for monitoring welding processes during manual welding using motion sensors. The device relates to an adaptive module that can be attached to a respective welding or cutting torch, which has at least one motion sensor system and, in particular, records the movement patterns in real time as processable process data to enable process mapping of the torch guidance. This forms the basis for monitoring process sequences, process deviations, welding defects, and the like.

[0002] Process parameter analyses using machine learning models are used to evaluate and analyze the data collected by the sensor system in the digital process mapping.

[0003] According to the invention, an adaptive module for monitoring welding processes during manual welding with motion sensors (AMB) is a device for recording the welder's movement patterns during manual welding / cutting or related thermal processes. It can be implemented independently of the welding torch as an additional, standalone device via a standardized mounting interface, or integrated into the welding torch.

[0004] Machine learning is understood to be a sub-area of ​​AI (artificial intelligence) that can carry out algorithmic processes using data processing models. The core processes here include the model's ability to learn from data, understand logical relationships, and self-correct through adaptation. This enables the digital generation of knowledge and experience. A model learns from examples using training models, enabling the identification and expanded interpretation of input data. For this purpose, machine learning algorithms generate approaches to relationships based on the training data, which are then tested against the resulting data. In this way, patterns and regularities are recognized in the learning data. This enables the system to evaluate even unknown data. The most striking feature of this method is the lack of provision of solution approaches.The system generates a solution based only on input and output data and checks it for accuracy.

[0005] One approach for processing serial process data is to mathematically classify, weight, evaluate and assign the recorded data sets to a result space using artificial neural networks (CNN - Convolutional Neural Network) with classifiers of numerical measurement results.

[0006] The adaptive module consists of connectable housing units that together form a separate system, whereby different components with different properties are considered as a common whole due to defined relationships with each other, which solve the given task.

[0007] In a first embodiment, the AMB comprises housing units that can be detachably connected to one another and which consist, for example, of an energy unit, a computer unit with sensor and / or actuator functions and a fastening unit.

[0008] The power unit comprises at least one housing housing a charger with a battery management system and accumulators. Power is transferred to the consuming units using spring contact pins (pogo pins). The connection to the computing unit is established via a magnetic connection for easy replacement.

[0009] The computer unit comprises at least one housing in which a motion sensor with a sufficient number of axes (e.g., for recording acceleration in 6 axes) and a sufficiently compact computer module for digitally processing the recorded data (e.g., a single-board computer) with shielding against process-related electromagnetic interference (e.g., ferrite foil) are installed. Furthermore, the computer unit includes external connections for communication with external devices, in particular IoT (Internet of Things)-capable devices. A module for communication via a standardized, wireless or wired data transmission, such as Bluetooth and Wi-Fi, is provided for this purpose. In the case of wireless transmission, an antenna designed to meet the requirements of the data transmission is also provided.

[0010] The design allows for the expansion of additional sensors and actuators, such as light-optical or haptic signals.

[0011] The simple design of the AMB enables the replacement of damaged individual components and / or housing units, such as the energy unit, computer unit or the individual components in the housing units, such as charger, batteries, switches, single-board computer, sensors, antenna and the fastening.

[0012] The housing units are preferably manufactured according to applicable industry standards and are designed to be heat-resistant, especially for welding conditions, so they can be used even in closed welding cells or exceptionally tight work spaces. To increase heat resistance, additives such as fillers or stabilizers are added to the housing materials, which are preferably made of plastic. The housing units are designed for additive manufacturing.

[0013] The attachment comprises at least one basic element for mounting on the computer unit, a clamping element (e.g. as a clamp or jaw), and the corresponding screw connection. The attachment is made to round tool areas, for example on the gooseneck as a variant with a defined diameter or diameter range. The geometry and type of attachment make the assembly suitable for standard tools (MIG, WSG, oxy-fuel torches). Great importance is attached to ensuring that the method of attachment does not influence the measurement or the welding or cutting process. This is due, among other things, to the lightweight construction of the AMB and the simple and flexible attachment. The high degree of compatibility allows all standard tools to be upgraded.

[0014] The processing of the relative measured values ​​recorded by the sensors enables, among other things, manual adjustment of the sensor by eye. Gravitational acceleration serves as a reference point for the calibration of the device.

[0015] The basic adjustment of the sensor is achieved by aligning and positioning it in the housing and by attaching it to the welding / cutting device using a form-fit, force-fit, or material-fit connection in a defined position as close to the process as possible. The attachment can be screwed. Adhesive bonding, riveting, and plug-in connections are also possible thanks to the modular design. State of the art

[0016] A number of publications are known from the state of the art that deal with the topic in general terms without showing concrete solutions: Journal of Manufacturing Systems; Volume 57, October 2020, Pages 1-18: Application of sensing techniques and artificial intelligence-based methods to laser welding real-time monitoring: A critical review of recent literature; Wang Cai, JianZhuang Wang, Ping Jiang , LongChao Cao, GaoYang Mi, Qi Zhou; https: / / doi.org / 10.1016 / j.jmsy.2020.07.021.

[0017] US 9 672 757 B2 shows a welding training system and includes welding training software with three or more modes. The three or more modes include a live arc mode, a simulation mode, a virtual reality mode, an augmented reality mode, or a combination thereof. The live arc mode is configured to enable training with a live welding arc, the simulation mode is configured to enable training with a welding simulation, the virtual reality mode is configured to enable training using a virtual reality simulation, and the augmented reality mode is configured to enable training using an augmented reality simulation.

[0018] EP 3 010 677 B1 discloses a system and method for determining a welding speed, while EP 2 986 419 B1 discloses a power tool with, among other things, integrated sensors and evaluation devices. Both documents are examples of the wide range of integrated systems.

[0019] This is where the invention comes in, as the inventive AMB eliminates the deficiencies of integrated systems by: - The possibility of easily upgrading all commercially available welding systems from the field of forward processes. This is not possible with integrated solutions, as specific solutions must be developed for each system type. - easy installation on the device without any further specialist knowledge; - modularity of all components of the AMB and the fastening, which means that minimal effort is required for necessary repairs / replacement of damaged components (standard components); - easy service and maintenance; - system-independent handling, i.e. in the event of a failure of the AMB, the welding / cutting process is not affected or interrupted, as there is no direct system integration and no interface to the welding system; - an independent power supply; - an adjustment that can be carried out easily and requires only minimal effort and - non-slip installation by clamping during fastening. Tasks

[0020] A first object of the invention is to provide a technically simple device that can be manufactured as a modular product and is easily adaptable to known systems for digitally mapping torch guidance during manual welding processes, making the data sets generated from it processable for machine learning models. This enables the integration of basic process parameters that cannot be captured by conventional welding system interfaces.

[0021] Another task is to clean / prepare the recorded data. This is done by normalizing and scaling the measurement data into data sets and identifying atypical "outliers." The data is then formatted into a format usable for the corresponding analysis model.

[0022] Preferred exemplary embodiments include use as a standalone system with a computing unit and energy unit, as well as a variant of the mounting unit with a fixed diameter. The sensors used here record the movement patterns using 6-axis acceleration and make them available to the analysis model in the form of processed data sets.

[0023] Another embodiment, for example, provides for expanded sensor technology; for example, temperatures can be recorded using a corresponding sensor as an additional process variable in addition to the movement patterns. This enables monitoring of the process ambient temperature and could, for example, determine the necessary break times for the welder.

[0024] This problem is solved with the inventive adaptive module according to claim 1.

[0025] The adaptive module for monitoring welding processes during manual welding with motion sensors is designed such that the adaptive module has at least one energy unit, a computer unit with at least one sensor and a fastening unit.

[0026] Furthermore, the respective energy unit can comprise at least a housing, a charger with a battery management system and an accumulator.

[0027] In a preferred embodiment, the respective computer unit can comprise at least a housing, a computer module and a sensor.

[0028] Furthermore, the respective fastening means can comprise at least one base element and a clamping element or a clamping fastening.

[0029] Particularly preferably, the respective sensor can record the 6-axis acceleration in sufficiently accurate resolution as a combination of axial acceleration and rotatotic gyroscopy.

[0030] In particular, the adaptive module according to the invention can be used with a welding / cutting torch or for torches of the metal inert gas welding, tungsten inert gas welding and oxyacetylene welding / cutting processes.

[0031] The invention is described below with reference to the enclosed figures in the figure description, which are intended to explain the invention and are not necessarily to be considered limiting: Fig. shows an exemplary adaptive module for monitoring welding processes in manual welding in a perspective view. Fig. shows an exemplary adaptive module for monitoring welding processes during manual welding in a front view. Fig. shows an exemplary adaptive module for monitoring welding processes during manual welding in a side view. Fig. shows an exemplary adaptive module for monitoring welding processes during manual welding on a welding torch in a perspective view. Fig. shows an exemplary adaptive module for monitoring welding processes during manual welding on a welding torch in a side view. Fig. shows an exemplary adaptive module for monitoring welding processes during manual welding on a welding torch in a front view. Fig. shows an example energy unit in a perspective view. Fig. shows an example energy unit in a side view. Fig. shows an example energy unit in a top view. Fig. shows an example energy unit in a bottom view. Fig. shows an example computer unit in a perspective view. Fig. shows an example computer unit in a front view. Fig. shows an example computer unit in a bottom view. Fig. shows an example of an assembled fastening unit for fixed diameters in a perspective view. Fig. shows an example of an assembled fastening unit for fixed diameters in a front view. Fig. shows an example mounted fastening unit for a fixed diameter in a side view. Fig. shows an example of an assembled fastening unit for fixed diameters in a bottom view. Fig. shows an example of an assembled fastening unit for variable diameters in a perspective view. Fig. shows an example of an assembled fastening unit for variable diameters in a front view. Fig. shows an example of an assembled fastening unit for variable diameters in a side view. Fig. shows an example of an assembled fastening unit for variable diameters in a bottom view. Fig. shows an example mounting position for standard MIG / MAG welding torches. Fig. shows an example mounting position for standard oxyacetylene welding torches. Fig. shows an example mounting position for standard oxy-fuel cutting torches. Fig. shows an example mounting position for standard TIG welding torches. Fig. shows an example mounting position for stretched TIG welding torches. Fig. shows an example mounting position for short TIG welding torches. Fig. shows an example flowchart for data acquisition in battery and wired operation. This illustrates the working principles of the basic modes. They enable the output of the movement patterns in the form of a data set that can be processed by the model. Fig. shows an example schematic structure of an AMB. Reference symbol 1 Adaptive module for movement monitoring of welding processes (AMB) 2 Position of the AMB on the welding / cutting torch 3 energy units 4 Computer unit 5 Mounting unit 6 Energy unit housing shell 7 Energy unit housing cover 8 Computer unit housing shell 9 Computer unit housing cover 10 Charger with battery management system 11 accumulators 12 spring contact pins (magnetic pogo pin connection) 13 switches 14 single-board computers 15 Antenna 16 Interface (Micro USB) 17 Energy level display 18 Sensor 19 Cover screw 20 threaded pin 21 Fixing screw 22 MIG / MAG welding torches 23 oxyacetylene welding torches 24 oxy-fuel cutting torches 25 TIG torches standard shape 26 TIG torches, elongated shape 27 TIG torches short form 28 Energy level display 29 Mounting surface fastening unit 30 mounting screw 31 Mounting clamp 32 fastening base element 33 Fastening clamping element 34 screw nut 35 Shielding QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] US 9 672 757 B2

[0017] EP 3 010 677 B1

[0018] EP 2 986 419 B1

[0018] Cited non-patent literature

[0000] Journal of Manufacturing Systems; Volume 57, October 2020, Pages 1-18: Application of sensing techniques and artificial intelligence-based methods to laser welding real-time monitoring: A critical review of recent literature

[0016] Wang Cai, JianZhuang Wang, Ping Jiang, LongChao Cao, GaoYang Mi, Qi Zhou; https: / / doi.org / 10.1016 / j.jmsy.2020.07.021

[0016]

Claims

[1] Adaptive module for monitoring welding processes during manual welding with motion sensors, wherein the adaptive module has at least one energy unit, a computer unit with at least one sensor and a fastening unit. [2] Adaptive module according to claim 1, characterized by that the respective energy unit comprises at least a housing, a charger with a battery management system and an accumulator. [3] Adaptive module according to one of the preceding claims, characterized by that the respective computer unit comprises at least a housing, a computer module and a sensor. [4] Adaptive module according to one of the preceding claims, characterized by that the respective fastening means comprises at least a base element and a tensioning element or a clamping fastening. [5] Adaptive module according to one of the preceding claims, characterized bythat the respective sensor can record the 6-axis acceleration with sufficiently accurate resolution as a combination of axial acceleration and rotatotic gyroscopy. [6] Use of the adaptive module according to one of the preceding claims with a welding / cutting torch or for torches of the metal inert gas welding, tungsten inert gas welding and oxyacetylene welding / cutting processes.

Images