Model Based Predictive Control for the Tow-Placement Technique

Abstract

lNfRODUCTIONAutomated thermoplastic tow-placement is a non-autoclave, composite manufacturing technique developed for aerospace structures where processing of high temperature composites is required. It promises improved and reliable [mal part quality as well as cost reduction using an automated robot system. At the Center for Composite Materials a lab-scale wOlkcell consists of an ABB robot, a tow-placement head and an AGEMA thermal camera where process models and control systems can be developed and validated.The automated tow-placement head, includes a two roller, two hot nitrogen gas torch approach, that provides nip point heating as well as consolidation force (Figure 1). The incoming thermoplastic tape is heated up under the preheater and a first bonding of the tow on the substrate occurs under the laydown roller. Then, bonding strength is increased through thru-thickness heating of the material under the main torch. Finally, porosity is minimized under the consolidation roller.Quality of the material is quantified in terms of void content and degree of bonding between insitu consolidated layers. Process set-points consist of torch heights, velocity, consolidation forces, temperatures of the rollers and the tool. The amount of energy transferred to the material is determined by independently controlled nozzle heights [1] and deposition rate. The forces applied to each roller can also be independently controlled so that pressures are generated in the material to reduce void content. Process models for this tow-placement head configuration, including heat transfer, consolidation and bonding, have been developed and can now accurately predict optimal process conditions [2,3] Quality sensors for the tow-placement technique have not yet been developed for real-time monitoring of in-situ void content and polymer interdiffusion [4]. However, the current state of the material can be predicted using accurate process models. This software sensor provides the

Modeling of Heat Transfer and Void Dynamics for the Thermoplastic Composite Tow-Placement Process

“…This camera captures the full temperature field from the composite surface during processing. The camera is accurate to within 0.1 C, and the data can be extracted in real time for control purposes [24]. Figure 7 is a snapshot of the thermal signature of the ATP process.…”

“…These resulting gradients in material quality have prompted further research into suitable methods for on-line process control [23][24][25]. With adequate knowledge of the processing history, it is possible to tailor the bulk void content through a predictive artificial neural network based feedback control.…”