A computational thermomechanical model is proposed to predict the shape distortions of parts produced by additive manufacturing. This process induces material fusion and thus thermal and metallurgical phenomena. A simple experimental calibration is in to include their combined effect in the proposed model. As in additive manufacturing itself, the proposed model works in a layer-by-layer basis. The layers are linked using thermally activated contact elements. The model is validated experimentally, using a complex part, printed by Selective Laser Melting. The model was able to accurately predict the real shape deviations. This work represents a step toward a digital twin.