Continuous Robotic Filament Winding for Tensile Applications

Abstract

With the rising urgency of building more sustainably and emitting less carbon dioxide into the atmosphere, optimising material use is becoming increasingly relevant. Globally, the civil engineering sector contributes to 8% of total CO2 emissions through cement production. In addition to reinforced concrete, it is estimated that 40% of these emissions originate from steel reinforcement. By optimising reinforced concrete elements, material waste can be reduced by up to 40%, while still meeting the required strength capacity. Due to the use of standard-sized steel bars and conservative detailing, this often results in significant oversizing of reinforcement. However, optimising reinforcement requires a change from using standardised steel reinforcement bars, mats, and cages towards long filament fibres. For example, when reviewing floors or walls with an open space, such as a door opening, around these locations the reinforcement net cannot continue its path, and the reinforcement needs to be strengthened locally. In these cases, the long filament fibres are capable of being placed in less fixed shapes and can be adjusted on a local scale to improve the strength. These filaments can be made from impregnated glass or carbon fibres, which harden over time.

Besides reducing the oversizing of steel, filament winding could also be made interaction-based, meaning that the winding is not limited to a fixed scaffolding point but can also intertwine with earlier wound filaments, reducing the material of these scaffolding elements. This led to the central question: How can material usage be reduced in both the filament and the scaffolding by using a construction process of continuous robotic winding? This research focuses on the topology and sequencing of tensile wound structures using automated robotic filament winding with a nylon rope. This study introduces a novel approach to filament winding by changing rope intersections into functional support points, replacing the need for traditional fixed scaffolding.

To create interactions, the separate winding layers must be connected to each other. This is achieved by transforming the concept of Reidemeister moves from knot theory into either back-and-forth or looped connections. To allow interactions to become connections, the 2D winding plane is combined with a 3D workspace. Additionally, scaling up the external scaffolding increases the number of possible intersection points that can be used to form interactions. In this research, scaffolding does not refer to the conventional way of supporting a structure, but the term is used to indicate fixed attachment points around which the fibre can be wound. These scaffolding supports are used to allow the wound structure to be created, but the shape of these supports can differ for each setup.

The research first identifies the design guidelines needed to create an interaction-based winding system. First of all, repeating patterns must be avoided, as they act like pulley systems and relocate nodes unpredictably. It is recommended to wind perpendicular to a single spanned rope, allowing a margin of 10 degrees. This prevents the nylon rope, which is used during this research, from sliding off and creating undesired topologies. Furthermore, connections help resist nodal movement and ensure all ropes are integrated into the interaction. For the winding sequence, it is preferable to begin winding between fixed supports to build resistance in the system before introducing interaction-based knots. Finally, straight-line winding is preferred over angled winding between fixed supports. Using these guidelines, a final design was developed based on a certain topology with spanned ropes. This reduced the number of scaffolding supports by 33%, from 12 to 8, while keeping the same topology. In contrast, the rope length increased by only 2%, from 6.42m to 6.56m. Another key observation is the nodal displacement in the interaction-based system. The average vector deviation amounted to 1.6cm, which is 6.3% of the distance between the supports. An additional winding resulted in a slight deviation, meaning that the system becomes stiffer and the nodes begin to behave like fixed supports.

In conclusion, the use of interaction-based methods enables a reduction in scaffolding material by replacing internal supports with rope intersections. Although this approach results in a slight increase in filament length, the difference is minimal. By minimising nodal displacement and improving the autonomy of the robotic arm to avoid collisions, continuous filament winding has significant potential to be applied effectively to non-fixed support setups based on interactions.