Intelligent Real-time Carbonization Control

"Deeper materials science knowledge from mathematical models and in situ material property sensors will be exploited in next generation factory process controllers to improve productivity and quality. These intelligent control systems will accelerate productivity growth and materials science application through the ease with which they can be modified to accept new materials knowledge and sensors. Carbonization of resin matrix carbon fiber composites, specifically, is a violent solid state chemical reaction with large stresses due to gas pressure, matrix shrinkage, and differences in thermal expansion. The matrix strength falls sharply, then rises again during the process. At least two competing reaction paths are important; the first cleaves methylene bonds, reduces toughness, density and stiffness; the second reforms those bonds, ultimately resulting in aromatic ring condensation that increases density, strength, and modulus. Since these reactions have different temperature dependencies, one can expect to shift the process towards the second path by changing the heating rate, but that change must be accomplished without producing macroscopic interlaminar crack growth. To accomplish this, we employ gas sensors to infer matrix chemical state and acoustic emission to identify microstructural changes and to recognize brittle fracture. The latter does not prevent delamination, but is important in testing models for the causes of delamination and in improving the processing cycle. This paper describes a control strategy to accelerate the process while still lowering the risk of failure.IntroductionDeeper materials science knowledge from mathematical models and in situ material property sensors will be exploited in next generation factory process controllers to improve productivity and quality. These intelligent control systems will be more easily modified to accept new materials knowledge and sensors, thus accelerating productivity growth. Such a control strategy for carbonization control is described below.Carbon-carbon composites are time consuming and costly to manufacture. Processing methods have been developed empirically and are usually very conservative because the process often fails when any change is made. Apparently minor changes in materials or geometry can produce failures that are difficult to diagnose and correct. We will describe below several essentially unavoidable sources of variability. These are best dealt with by applying materials science to adapt (i.e., closed loop control on materials properties) the process in real time to reduce the effect of unavoidable model idealizations, material variability and environmental noise. The most critical issue in control of the carbonization process control is avoiding catastrophic delamination. The strategy developed below makes use of a priori materials knowledge and real time sensors to accelerate the now very slow process without destroying the material, and without requiring unreasonable resources to determine the statistical properties of the material."