Stringent demands in the sheet forming industry related to rapid and customized part realization coupled with rigorous requirements on geometric accuracy and product properties have outpaced the capabilities of traditional forming processes. As an emerging novel technology, Double-Sided Incremental Forming (DSIF) offers much higher flexibility with the complete elimination of the need for geometry-specific forming dies, and significantly reduces the forming time and cost in sheet metal production. However, challenges in achieving tight geometric tolerances at a reasonably short forming time and in maintaining structural integrity in incrementally formed parts limit the wide industrial adoption of the technology, even after decades of intensive laboratory developments worldwide. Due to the strong nonlinear nature of the DSIF process caused by material plasticity, geometric variations, and varying local contact conditions between the workpiece and the forming tool(s), only a few models succeeded in determining process inputs (such as toolpaths) to achieve the desired part geometry. However, each model was applicable only to a specific process parameter combination or geometry. Moreover, as a flexible process, IF is often performed under a vast variety of potential process parameter combinations, such as different sheet metal materials, sheet thickness, tool diameters, lubricating conditions and clamping forces, which further signify the difficulty in establishing a single generalized model that encompasses all these parameter variations. ', 'The central idea of the thesis, in response to the above-mentioned challenges and hurdles, is to explore and realize a control framework integrated with a mechanics-based model for the purpose of enhancing geometric accuracy and suppressing possible fracture. To bridge the gap between laboratory research and industrial applications for DSIF, the control framework is designed with high control accuracy, sufficient generalization capability and easy pragmatic instrumentation setup in mind. The envisioned framework integrates three control loops at different levels to ensure accurate tool positioning, consistent maintenance of the desired tool contact pressure, and effective springback compensations for DSIF. Specifically:', '\tA real-time tool co¬¬mpliance compensation and a real-time contact force control algorithm are introduced to compensate for tool compliance and motion error, providing a stable contact pressure and avoiding any early loss of contact between the tool and the part. The implementation of a contact stiffness model increases the algorithm’s generality and robustness for different machine setups and material choices. Both algorithms are directly superimposed over the conventional position servo control loop with minimal modifications and proven to be effective over a large range of process parameters with reasonable control accuracy (< 10 N deviation for a 300 N command force). The purpose of this control loop is to ensure the realization of a true DSIF process, which has demonstrated its effectiveness in achieving tight geometric tolerance, delaying fracture, and increasing fatigue life compared to single point incremental forming.', '\tA feedforward toolpath optimization algorithm is established during the off-line planning stage to reduce the geometric error for a newly developed Accumulative Double-Sided Incremental Forming (ADSIF) toolpath strategy. Two process parameters that determine the relevant tool positions, i.e., position angle and gap, are proposed, for the first time, to capture the bending and squeezing phenomena in ADSIF. These process parameters are then optimized with a response surface model as a function of the local geometric wall angle. A significant error reduction of 80% is achieved using the developed response surface model. The defined optimization framework can be implemented in ADSIF without the need for any feedback sensor as the starting process condition is well-defined and consistent throughout the process. Consequently, it should be noted that this strategy for geometric control is suitable for ADSIF only. A more general strategy for DSIF is to be shown next.', '\tAn on-line springback compensation algorithm is established, which dynamically modifies the toolpath to compensate for potential springback based on an accelerated FEM-based springback prediction model. The springback prediction model is a first-of-a-kind real-time model that links the target variable (final part geometry) to measurable process variables (forming forces). Moreover, since the FEM model can be easily adopted to different sheet materials, part geometries and boundary conditions, the predictive model can thus be easily generalized to all these parameter changes. A customized unsupervised classification algorithm is also proposed, for the first time, to extract the key control points for a given complex geometry, so that springback prediction only needs to be conducted at a limited number of selected locations. The experimental results demonstrate that springback error can be limited within the 1.5 mm range for the tested geometries, compared to more than a 5 mm geometric error for the uncontrolled parts. ', 'To summarize, this work directly combines the control and modeling efforts to develop a systematic control framework for DSIF that improves part accuracy/contact conditions with three control loops operating at different time scales. Specifically, the introduction of feedback control into the DSIF process permits a more error-/disturbance-tolerant model, while the establishment of the process mechanics model helps to determine the proper control strategy and increases control robustness. Moreover, the framework, built on a solid foundation of process mechanics, is designed with a low implementation cost, a non-machine-intrusive setup, and a high model generality, thus guarantying a great potential of the work to be scaled up from laboratory prototypes to industrial applications. In doing so, this work successfully improves various aspects of DSIF, and paves the way for continued commercial implementation of the DSIF technology in the near future.

Last modified

• 10/28/2019

Creator

• Huaqing Ren

DOI

• https://doi.org/10.21985/n2-9a24-qh83

Subject

• Mechanical Engineering

Keyword

• Incremental Forming
• Springback
• Control
• Optimization
• Modeling

Date created

• 2018-01-01

Resource type

• Dissertation

Rights statement

• In Copyright