MC-GRU:a Multi-Channel GRU network for generalized nonlinear structural response prediction across structures

TL;DR

This work tackles the generalization gap in AI-based seismic response surrogates by introducing MC-GRU, a multi-channel GRU that simultaneously ingests ground motion sequences and structural parameters. The architecture extends GRU by embedding structural information into the hidden-state update, yielding mappings of the form $Y = MC\text{-}GRU(X, S; \Phi)$ that generalize to unseen structures. Validation across a SDOF linear system, Bouc-Wen hysteresis, and an experimental RC column demonstrates high predictive accuracy (e.g., $R^2$ up to about 0.97) and robust cross-structure generalization, outperforming fixed-structure LSTM/GRU baselines. The approach offers a computationally efficient seismic-response surrogate with wide potential impact for PBEE and urban resilience, and code is planned for public release.

Abstract

Accurate prediction of seismic responses and quantification of structural damage are critical in civil engineering. Traditional approaches such as finite element analysis could lack computational efficiency, especially for complex structural systems under extreme hazards. Recently, artificial intelligence has provided an alternative to efficiently model highly nonlinear behaviors. However, existing models face challenges in generalizing across diverse structural systems. This paper proposes a novel multi-channel gated recurrent unit (MC-GRU) network aimed at achieving generalized nonlinear structural response prediction for varying structures. The key concept lies in the integration of a multi-channel input mechanism to GRU with an extra input of structural information to the candidate hidden state, which enables the network to learn the dynamic characteristics of diverse structures and thus empower the generalizability and adaptiveness to unseen structures. The performance of the proposed MC-GRU is validated through a series of case studies, including a single-degree-of-freedom linear system, a hysteretic Bouc-Wen system, and a nonlinear reinforced concrete column from experimental testing. Results indicate that the proposed MC-GRU overcomes the major generalizability issues of existing methods, with capability of accurately inferring seismic responses of varying structures. Additionally, it demonstrates enhanced capabilities in representing nonlinear structural dynamics compared to traditional models such as GRU and LSTM.

Figures (16)

• Figure 1: Comparison between the MC-GRU network and traditional response prediction networks.
• Figure 2: The flowchart of MC-GRU for generalized nonlinear structural response prediction.
• Figure 3: The architecture of MC-GRU cell.
• Figure 4: The architecture of the MC-GRU network.
• Figure 5: The MAE distribution across subsets with different structural information.