Maintaining and validating a digital twin for industrial applications is key to derive reliable lifetime and performance predictions. Usually, engineers and analysts use test data and validated models from prior programs to calibrate and enhance their digital twins. However, the calibration and validation procedures are generally manual and can take several months due to the number of parameters to adjust, the number of sensors used, and the multiple operating conditions to consider. To keep the optimization complexity to a certain level, the actual workflow involves sub-modeling techniques. This is problematic as inter dependencies are ignored and often lead to divergence after reintegrating the sub models to the final digital twin. Optimization frameworks offer the possibility to automate and perform this validation procedure for the whole digital twin, but the computational cost associated is still extremely high and prevents the engineers to consider all the parameters and mission profiles. Maya HTT recently developed TMG Correlation, an adjoint-based solver that allows optimizing a digital twin using reference data from test or analysis in a faster and more efficient way than classic optimization methods. Indeed, this solver implemented in Siemens Simcenter 3D supports both steady-state and transient thermal correlation analyses with little overhead related to the number of design variables and sensors. This presentation focuses on the validation of a simplified and representative whole engine model containing hundreds of boundary conditions and corresponding number of design variables, using virtual test data and TMG Correlation. First, the presentation will focus on a description of the TMG Correlation tool and its associated adjoint-solver. Then, the whole engine model use case and the thermal correlation setup will be described. Finally, the performances of the tool and the results will be discussed, before giving some perspectives of work.