Automated restoration of windows from old buildings

The building sector accounts for around a third of energy consumption in Germany and therefore also contributes significantly to greenhouse gas emissions. A decisive lever for reducing this consumption is the energy-efficient refurbishment of existing buildings, especially windows. There are various options available for renovating windows, some of which require the non-destructive removal of the existing glazing from the window sash. Window renovation is currently carried out by hand, which leads to high costs and capacity bottlenecks in carpentry businesses. To make matters worse, there is also the risk of asbestos fibers being released from the window putty that was used until asbestos was banned in 1993. The applicable occupational health and safety guidelines also reduce productivity.

The 2ndLife project at the Chair of Production Systems (LPS) at Ruhr University Bochum, funded by the Central Innovation Program for SMEs (ZIM) in cooperation with the company Kilivations, is developing an automated system for the robot-assisted restoration of windows. 2ndLife aims to automate the removal of glazing in order to significantly reduce the exposure of employees to asbestos and increase productivity. Particular attention is paid to the Technical Rules for Hazardous Substances 519 (TRGS 519), which define binding guidelines for working safely with asbestos. Automation solutions in the construction industry, particularly in the field of renovation, have hardly been established to date. One of the main reasons for this is the often unpredictable and highly variable conditions in the construction site environment. The same applies to window renovation due to the high variability of window variants and geometries, for which no digital data is usually available. In addition, the condition of the windows varies considerably due to soiling or damage. The resulting need for repeated manual adjustment of the robot path for each individual window variant would significantly reduce the potential cost savings from automation. Against this background, a flexible and sufficiently autonomous automation system is required. This is to be achieved by developing a specialized image processing system for precise detection of the window contours and adaptive clamping.

Automatic adaptive path planning

The geometry and condition of the windows are recorded using a monochrome 2D camera attached to the robot tool. The captured images are then filtered and evaluated using edge detection algorithms to capture the contours of the window sash and glazing. Based on the contours, an offset cutting path is generated so that a narrow wooden web remains on the glazing, which enables the subsequent non-destructive removal of the glazing. The cutting paths are then transferred to an executable robot program. A downstream simulation of the robot movements ensures a collision-free and achievable robot path. For this purpose, program templates are created in advance, which also enable simplified adjustment of process-relevant parameters, such as the cutting and feed speed.