Abstract:Active-learning surrogate model–based reliability analysis is widely employed in engineering structural reliability analysis to alleviate the computational burden of the Monte Carlo method. To date, most of these methods are built based on the single-fidelity surrogate model, such as the Kriging model. However, the computational burden of constructing a fine Kriging model may be still expensive if the high-fidelity (HF) simulation is extremely time-consuming. To solve this problem, an active-learning method based on the multi-fidelity (MF) Kriging model for structural reliability analysis (abbreviated as AMK-MCS+AEFF), which is an online data-driven method fusing information from different fidelities, is proposed in this paper. First, an augmented expected feasibility function (AEFF) is defined by considering the cross-correlation, the sampling density, and the cost query between HF and low-fidelity (LF) models. During the active-learning process of AMK-MCS+AEFF, both the location and fidelity level of the updated sample can be determined objectively and adaptively by maximizing the AEFF. Second, a new stopping criterion that associates with the estimated relative error is proposed to ensure that the iterative process terminates in a proper iteration. The proposed method is compared with several state-of-the-art methods through three numerical examples and an engineering case. Results show that the proposed method can provide an accurate failure probability estimation with a less computational cost.

What problem does this paper attempt to address?

### Problems Addressed by the Paper The paper aims to address the computational burden in structural reliability analysis. Specifically, traditional reliability analysis based on Monte Carlo Simulation (MCS) incurs high computational costs when dealing with high-dimensional, nonlinear, and implicit limit state functions. To alleviate this computational burden, the paper proposes an active learning method based on the Multi-Fidelity Kriging Model (MF Kriging Model), referred to as AMK-MCS+AEFF. ### Main Contributions 1. **Multi-Fidelity Kriging Model**: - Utilizes data from both Low-Fidelity (LF) and High-Fidelity (HF) models to improve prediction accuracy while reducing computational costs through the Multi-Fidelity Kriging Model. - By introducing the Augmented Expected Feasibility Function (AEFF), it comprehensively considers cross-correlation, sampling density, and cost queries to determine the location and fidelity level of the update samples. 2. **New Stopping Criterion**: - Proposes a new stopping criterion based on relative error estimation to ensure the iterative process terminates at an appropriate number of iterations, thereby guaranteeing the accuracy and reliability of the estimated failure probability. - Compared to traditional stopping criteria based on learning functions, the new criterion is simpler and more effective, avoiding premature or delayed termination. ### Method Overview 1. **Generate Initial Samples**: - Use Optimized Latin Hypercube Sampling (OLHS) to generate low-fidelity samples. - Use the exchange algorithm to generate high-fidelity samples from low-fidelity samples. 2. **Construct Monte Carlo Population**: - Generate the Monte Carlo population based on the statistical characteristics of the design variables. 3. **Evaluate Sample Responses**: - Evaluate the responses of the samples using the true model. 4. **Construct (Update) Co-Kriging Model**: - Construct the Co-Kriging model based on the samples and their responses. - In the adaptive update process, update the Co-Kriging model by adding new samples and re-estimating the hyperparameters. 5. **Check Stopping Criterion**: - Check if the stopping criterion is met; if so, terminate the iteration, otherwise continue updating the samples. 6. **Determine Update Points**: - Construct AEFF and evaluate its value, finding the sample with the maximum AEFF value as the update point. 7. **Evaluate Coefficient of Variation**: - Check if the required accuracy level is achieved by calculating the Coefficient of Variation (Cov). ### Experimental Validation The proposed method's accuracy and effectiveness are validated through three numerical functions and an engineering case. The experimental results show that the AMK-MCS+AEFF method can provide accurate failure probability estimates with lower computational costs. ### Conclusion The AMK-MCS+AEFF method proposed in this paper has significant advantages in structural reliability analysis, maintaining high prediction accuracy while reducing computational costs. This method provides an effective tool for reliability analysis in practical engineering applications.

An active-learning method based on multi-fidelity Kriging model for structural reliability analysis

An improved active Kriging method for reliability analysis combining expected improvement and U learning functions

An active learning Kriging model with approximating parallel strategy for structural reliability analysis

A New Active Learning Kriging Method for Structural Reliability Based on the Error Analysis of the Predicted Failure Probability

A novel fidelity selection strategy-guided multifidelity kriging algorithm for structural reliability analysis

An active learning Kriging model with adaptive parameters for reliability analysis

A high-dimension structural reliability method based on active learning Kriging and dimension reduction technique

A new adaptive analysis method based on the Kriging model for structural reliability analysis

A Semi-Parallel Active Learning Method Based on Kriging for Structural Reliability Analysis

An active learning Kriging‐based multipoint sampling strategy for structural reliability analysis

An active learning method combining adaptive kriging and weighted penalty for structural reliability analysis

CSP-free adaptive Kriging surrogate model method for reliability analysis with small failure probability

AGP-MCS+D: An active learning reliability analysis method combining dependent Gaussian process and Monte Carlo simulation

REIF: A novel active-learning function toward adaptive Kriging surrogate models for structural reliability analysis

Active learning for structural reliability: Survey, general framework and benchmark

Structural reliability analysis under evidence theory using the active learning kriging model

An improved active learning method combing with the weight information entropy and Monte Carlo simulation of efficient structural reliability analysis

The proposed rank-based active learning Kriging for efficient structural reliability analysis using weighted average simulation method.

A data-driven Kriging model based on adversarial learning for reliability assessment

Active learning for structural reliability analysis with multiple limit state functions through variance-enhanced PC-Kriging surrogate models

Active Kriging-based conjugate first-order reliability method for highly efficient structural reliability analysis using resample strategy