Digital twins promise to revolutionize engineering by offering new avenues for optimization, control, and predictive maintenance. We propose a novel framework for simultaneously training the digital twin of an engineering system and an associated control agent. The twin's training combines adjoint-based data assimilation and system identification methods, while the control agent's training merges model-based optimal control with model-free reinforcement learning. The control agent evolves along two independent paths: one driven by model-based optimal control and the other by reinforcement learning. The digital twin serves as a virtual environment for confrontation and indirect interaction, functioning as an "expert demonstrator." The best policy is selected for real-world interaction and cloned to the other path if training stagnates. We call this framework Reinforcement Twinning (RT). The framework is tested on three diverse engineering systems and control tasks: (1) controlling a wind turbine under varying wind speeds, (2) trajectory control of flapping-wing micro air vehicles (FWMAVs) facing wind gusts, and (3) mitigating thermal loads in managing cryogenic storage tanks. These test cases use simplified models with known ground truth closure laws. Results show that the adjoint-based digital twin training is highly sample-efficient, completing within a few iterations. For the control agent training, both model-based and model-free approaches benefit from their complementary learning experiences. The promising results pave the way for implementing the RT framework on real systems.

The thermal management of cryogenic storage tanks requires advanced control strategies to minimize the boil-off losses produced by heat leakages and sloshing-enhanced heat and mass transfer. This work presents a data-assimilation approach to calibrate a 0D thermodynamic model for cryogenic fuel tanks from data collected in real time from multiple tanks. The model combines energy and mass balance between three control volumes (the ullage vapor, the liquid, and the solid tank) with an Artificial Neural Network (ANN) for predicting the heat transfer coefficients from the current tank state. The proposed approach combines ideas from traditional data assimilation and multi-environment reinforcement learning, where an agent's training (model assimilation) is carried out simultaneously on multiple environments (systems). The real-time assimilation uses a mini-batch version of the Limited-memory Broyden-Fletcher-Goldfarb-Shanno with bounds (L-BFGS-B) and adjoint-based gradient computation for solving the underlying optimization problem. The approach is tested on synthetic datasets simulating multiple tanks undergoing different operation phases (pressurization, hold, long-term storage, and sloshing). The results show that the assimilation is robust against measurement noise and uses it to explore the parameter space further. Moreover, we show that sampling from multiple environments simultaneously accelerates the assimilation.