Thermal spray in general, plasma spray in particular, is a highly complex process with numerous interacting variables associated with generation of the spray stream, deposit formation dynamics, and the resultant property linkages. Compounding this variability further are both the spatial (different booths and different locations) and temporal (process start-stops, hardware degradation, operator etc.) effects. As such, an understanding of process and coating consistency and variability offers significant challenges. Recent scientific advances as well as measurement tools have enabled elucidation of the intrinsic variabilities associated with each of the process sub-steps; however, integrated understanding of the system level reliability is still lacking. This article seeks an integrated assessment of process and coating reliability through systematic measurements of variabilities during each stage of the process subjected to different operating parameters. Through critical examination of first-order process maps, the influence of process parameters on particle state is reviewed for repeated spray runs with a single parameter effect as well as across a spectrum of process parameters. In addition, influence of these changes on designrelevant coating properties were obtained for plasma-sprayed zirconia through recourse to novel in situ and ex situ substrate curvature measurements. Finally, the implications of such integrated reliability studies have been explored through collaborative experiments conducted in the industrial sites.

Deliberate particle state variations were performed using atmospheric plasma spray (APS) and high-velocity oxy-fuel flame spraying (HVOF) to create a set of first-order process maps. Particle states were measured simultaneously using five in-flight particle sensors: DPV-2000, Accuraspray, SprayWatch, TDS, and SprayCam. While the sensors use similar methods for calculating particle characteristics, absolute values of temperature and velocity were considerably different. Process map trends among sensors are in agreement for the HVOF process, but differ when using plasma spray at high total gas flow conditions. After understanding the stochastic nature of particle detection, an open loop feedback control algorithm was implemented to achieve similar particle states with different hydrogen gas flow rates. The resulting particle state window measured by three different sensors under select fixed hydrogen flow rates was significantly narrowed.

No abstract

An empirical model was built for the prediction of HVOF sprayed NiCr in-flight particle properties based on the spraying parameters. The employment of factorial design in process parameter development, allowed determination of the contribution of the key process variables, such as flame energy (combustion pressure and O2/F ratio), spray distance and feed rate, on the resultant particle velocities and surface temperatures. The significance of each parameter was used to compose a simple model which enabled the description of the particles’ in-flight properties. Particles with velocities ranging as much as 300 m/s and temperatures ranging up to 350 °C were used to produce a range of coatings on an in situ curvature sensor enabling the determination of evolving and residual stress. These diverse particle states combined with the effect of flame impingement on the substrate, resulted in coatings of similar thickness, but significantly different stress states. Real time evolving stresses during deposition were interrelated to particle in-flight properties and, consequently, to spraying parameters.