Processing of soft matter and composites: integration of material sensors with process models and intelligent control algorithms

This paper describes the development of an intelligent process-control syst em for the production of fiber-reinforced organic polymeric composites. The composite material consists of a polymer matrix (polyamide resin F174) and high-modulus quartz-fiber reinforcement. This composite material has good mechanical properties at high temperatures, and possesses a low dielectric constant, making it suitable for applications in missile radomes. The probl em is that the raw materials are chemically reactive, and the process-contr ol system must enable adaptation to variations in the temperature-time expo sure of the raw materials and/or variations which may occur in the material s received from different suppliers. The uniqueness of the control system l ies in that it is self-directing, and relies on information derived from se nsors (laser fiber-optic probes and dielectric sensors) placed within the m aterial. In addition, a materials-transformation model based on the chemica l kinetics of the polymerization process calculates a number of key polymer parameters, such as degree of imidization, degree of cure, molecular weigh t distribution, and polydispersity ratio in situ. The collective ability to collect high-quality sensor information, to run sophisticated but robust p rocess models in real time, to make complex decisions using artificial inte lligence (AI), and to implement these decisions for controlling the structu re of the actual material being processed represents a significant breakthr ough in materials and process capability in this field. The focus of this w ork, measuring and controlling the physical and chemical properties of the material, rather than the physical attributes of the processing machinery, is an important paradigm shift. (C) 1998 Elsevier Science Ltd. All rights r eserved.