Experimental and modeling analysis of plasma spray nonlinearities for advanced process control design.

摘要

Plasma spray is a high-throughput, economical and low environmental impact coating process that can be used to meet the demanding performance requirements for turbines, pumps, and emerging biomedical applications and solid oxide fuel cells. However, current plasma spray process capabilities are limited by its significant process variations which pose challenges for engineering coating structure for advanced applications as well as optimizing process yield and economics. This dissertation investigates the critical issues needed to develop an advanced plasma spray process control system that can compensate for these process variations, including: characterization of different fluctuations observed in plasma spray process, development an understanding of the sources of the dominant variations, the dominant torch input-output dynamics and nonlinear behaviors of the process, and the control sensing requirements for detecting these variations.; A taxonomy of the different torch unforced arc fluctuations as well as the characteristics of forced arc fluctuations (e.g. due to torch input changes) is experimentally determined in order to characterize the dominant disturbances and control design challenges. Torch voltage and particle intensity centroid are identified as the most important measurements for detecting and understanding different process variations.; To determine the torch input/output dynamic characteristics for different operating torch conditions, a series of open-loop experiments were conducted. The nonlinear behavior of the process open-loop gains and dynamics are identified as important problems for development of a viable process control strategy design.; A low-order dynamic model of the plasma spray process is presented that captures the dominant interactions between torch inputs, plasma states and particle trajectory. The model is used to obtain a better understanding of the relation of torch inputs to centroid position and gain insight into the corresponding arc fluctuation effects. Experimental verification is also presented.; Based on the experimental and modeling analysis, implications for developing an advanced closed-loop control system and the need for supervisory process control strategy are presented to reduce the impact of process variations on coating quality by keeping coating related in-flight particle states consistent.