During the flood discharge in large-scale hydraulic engineering projects, intense flow-induced vibrations may occur in hydraulic gates, gate piers, spillway guide walls, etc. Furthermore, the vibration mechanism is complicated. For the spillway guide wall, existing studies on the vibration mechanism usually focus on the vibrations caused by flow excitations, without considering the influence of dam vibration. According to prototype tests, the vibrations of the spillway guide wall and the dam show synchronization. Thus, this paper presents a new vibration mechanism of associated-forced coupled vibration (AFCV) for the spillway guide wall to investigate the dynamic responses and reveal coupled vibrational properties and vibrational correlations. Different from conventional flow-induced vibration theory, this paper considers the spillway guide wall as a lightweight accessory structure connected to a large-scale primary structure. A corresponding simplified theoretical model for the AFCV system is established, with theoretical derivations given. Then, several vibrational signals measured in different structures in prototype tests are handled by the cross-wavelet transform (XWS) to reveal the vibrational correlation between the spillway guide wall and the dam. Afterwards, mutual analyses of numeral simulation, theoretical derivation, and prototype data are employed to clarify the vibration mechanism of a spillway guide wall. The proposed mechanism can give more reasonable and accurate results regarding the dynamic response and amplitude coefficient of the guide wall. Moreover, by changing the parameters in the theoretical model through practical measures, the proposed vibration mechanism can provide benefits to vibration control and structural design.
Deep Reinforcement Learning (DRL) has been a promising solution to many complex decision-making problems. Nevertheless, the notorious weakness in generalization among environments prevent widespread application of DRL agents in real-world scenarios. Although advances have been made recently, most prior works assume sufficient online interaction on training environments, which can be costly in practical cases.
To this end, we focus on an offline-training-online-adaptation setting,
in which the agent first learns from offline experiences collected in environments with different dynamics and then performs online policy adaptation in environments with new dynamics. In this paper, we propose Policy Adaptation with Decoupled Representations (PAnDR) for fast policy adaptation.
In offline training phase, the environment representation and policy representation are learned through contrastive learning and policy recovery, respectively. The representations are further refined by mutual information optimization to make them more decoupled and complete. With learned representations, a Policy-Dynamics Value Function (PDVF) network is trained to approximate the values for different combinations of policies and environments from offline experiences. In online adaptation phase, with the environment context inferred from few experiences collected in new environments, the policy is optimized by gradient ascent with respect to the PDVF. Our experiments show that PAnDR outperforms existing algorithms in several representative policy adaptation problems.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.