Frequency Domain Quantification Of Manufacturing Process Resolution

Manufacturing processes are input/output systems subject to disturbances and uncertainties. Hence, process control is a necessary compensation. However, control tools cannot be chosen arbitrarily, but are limited by several process characteristics. Primary among these, process resolution determines the degree to which changes in process inputs produce changes in process states and in process outputs. In turn, process resolution dictates the applicability of specific control tools. Two other limitations are feedback time and process state measurability. This thesis provides a means of quantifying process resolution and thereby gives a metric of process controllability. Using Fourier Transforms, parameterized process-inputs and process-outputs are expressed as input frequencies and output frequencies. The process itself is represented as a transfer function, and process resolution is defined as the transfer function bandwidth. The use of Fourier Transforms is a generalization of digital signal processing techniques. These have been used previously in manufacturing to describe the shape of sheet-metal-stamping parts and tools. Reviews of control problems in injection molding, sheet metal forming, and composite processing are included as background and examples.