Control chart is a statistical process control tool to routinely monitor the quality of a process. In the past decade, using auxiliary information to enhance the ability of control chart to detect parameter shifts has attracted wide attention in literature. Most of the existing works in this topic use a linear model and assume performance and auxiliary variables follow a bivariate normal distribution to establish the auxiliary-information-based (AIB) charts. However, the normality assumption could be violated on some occasions. In this study, the skew-normal distribution is used to characterize the auxiliary variable to expand the joint distribution of the performance and auxiliary variables to a generalized joint family. Then, a new multivariate AIB chart is established to monitor the quality of process. The performance of the proposed AIB control chart method is verified using Monte Carlo simulations. An example with regard to the cement filling process is used for illustration.

控制图是一种统计过程控制工具，用于定期监控过程的质量。在过去的十年中，利用辅助信息来增强控制图检测参数变化的能力在文献中引起了广泛的关注。本主题中的大多数现有工作使用线性模型并假设性能和辅助变量遵循二元正态分布来建立基于辅助信息 (AIB) 的图表。但是，在某些情况下可能会违反正态性假设。在本研究中，使用偏态正态分布来表征辅助变量，将性能和辅助变量的联合分布扩展到广义联合族。然后，建立一个新的多变量AIB图来监控过程质量。使用蒙特卡罗模拟验证了所提出的 AIB 控制图方法的性能。以水泥填充过程为例进行说明。