Skip to Main Content
Laser cladding is a multiple-parameter-dependent process, and a feedback control is critical for the process stabilization. This paper presents a generalized predictive control strategy with input constraints to stabilize the melt pool temperature during a high-power diode laser cladding process. A dual-color pyrometer was used to monitor the melt pool temperature. A state-space dynamic model relating the laser drive signal (laser power) to the melt pool temperature was identified experimentally using the subspace method. A generalized predictive controller with input constraints was implemented in real time using the state-space model. The closed-loop process was able to track the melt pool temperature to a reference temperature profile. Laser cladding of H13 tool steel on a substrate with uneven surface showed that the closed-loop process was able to compensate for an under-fill with 3-mm depth after 40-layer depositions.