研究目的
The objective of this study is to develop a LBAM modeling-monitoring framework that incorporates spatial-temporal effect in characterizing and monitoring melt pool behavior.
研究成果
The proposed hierarchical spatial-temporal modeling and monitoring framework effectively incorporates spatial-temporal effects in melt pool dynamics, offering superior detection power and false positive rates compared to existing methods. The two-level control chart system enables simultaneous monitoring of global and local patterns, facilitating the identification of out-of-control images and problematic regions. The framework is robust against data non-stationarity and computationally efficient for online monitoring.
研究不足
The study assumes layers are independently modeled and monitored, which may not fully capture the layer-wise dependence in physical processes. The STCAR-AR parameters are fixed and require manual updates for new data, indicating a need for automated updating mechanisms.
1：Experimental Design and Method Selection:
The study adopts Spatial-Temporal Conditional Autoregressive (STCAR) models for modeling melt pool dynamics, with STCAR-AR identified as the best fit. A two-level control chart system is constructed for monitoring, incorporating spatial and temporal effects.
2：Sample Selection and Data Sources:
Thermal images of melt pool collected during the LBAM process of a Ti-6Al-4V thin-walled structure are used. The data were captured via a dual-wavelength pyrometer.
3：List of Experimental Equipment and Materials:
Dual-wavelength pyrometer for thermal image collection, Ti-6Al-4V thin-walled structure as the manufacturing target.
4：Experimental Procedures and Operational Workflow:
The melt pool is divided into units forming a fixed spatial structure. STCAR-AR models are used to characterize temporal autocorrelation with an AR(1) process and accommodate the spatial structure with an adjacency matrix. A two-level control chart is developed for monitoring.
5：Data Analysis Methods:
Bayesian analysis with Markov Chain Monte Carlo (MCMC) simulation for model parameter estimation. The monitoring involves calculating control limits and summary statistics that incorporate spatial-temporal effects.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
