The longitudinal Varestraint test is the most widely used technique to quantify the weld solidification cracking susceptibility, which provides the total crack length, TCL, as an indicator. But experimental work requires a significant amount of time and resource. Data-driven approaches can serve as alternatives to estimate the TCL through influencing factors, including metallurgical factors, welding conditions, and test parameters. We develop the least-squares boosting model to predict the TCL, a solidification cracking susceptibility indicator, based on the alloy composition, welding parameters, and applied strain. The model manifests high accuracy and stability, and applies to a wide range of stainless steels. It serves as a fast, robust, and low-cost tool for the assessment of solidification cracking susceptibility.

纵向Varestraint试验是用于量化焊缝凝固裂纹敏感性的最广泛使用的技术，该技术提供了总裂纹长度TCL作为指标。但是实验工作需要大量的时间和资源。数据驱动方法可以通过影响因素（包括冶金因素，焊接条件和测试参数）来估算TCL。我们基于合金成分，焊接参数和应用应变，开发了最小二乘增强模型来预测TCL（凝固裂纹敏感性指标）。该模型具有很高的准确性和稳定性，并适用于各种不锈钢。它是评估凝固裂纹敏感性的快速，可靠且低成本的工具。