Layer-to-layer height control of Laser Metal Deposition processes

A laser metal deposition (LMD) height controller design methodology is presented in this paper. The height controller utilizes the particle swarm optimization (PSO) algorithm to estimate model parameters between layers using measured temperature and track height profiles. The process model parameters for the next layer are then predicted using exponentially weighted moving average (EWMA). Using the predicted model, the powder flow rate reference profile, which will produce the desired layer height reference, is then generated using Iterative Learning Control (ILC). The model parameter estimation capability is tested using a four-layer deposition. The results demonstrate the simulation based upon estimated process parameters matches the experimental results quite well. The experimental deposition using this methodology demonstrates good tracking of the height reference in terms of the finished track.