Multivariable process monitoring (MPM) algorithms are very popular for early detection of abnormalities and diagnosis of process conditions to ensure plant safety, reliability, and production quality during operation. Hazard and operability (HAZOP) study is a systematic brainstorming session to identify the credible causes and consequences of the process upsets. Then the process safeguards in place are reviewed to evaluate risk of HAZOP scenarios. According to IEC61882 HAZOP studies have proved to be very useful in a variety of different process industries. However, the method has some limitations to be considered in potential applications and particularly for complex processes. This paper aims at improving the HAZOP study by applying MPM algorithms to perform a holistic and more accurate consequence analysis and effective evaluation of the process safeguarding strategy. The proposed monitoring algorithm consists of two layers of decision boundaries. The 1^st layer is based on normal operating condition and the 2^nd layer (i.e. suggested in this paper) is based on safe operating condition of plant. The effectiveness of the proposed method is verified by investigations on an operating complex polymerization plant.

多变量过程监控 (MPM) 算法非常受欢迎，用于早期检测异常和诊断过程条件，以确保工厂运行期间的安全性、可靠性和生产质量。危害和可操作性 (HAZOP) 研究是一个系统的头脑风暴会议，用于确定过程干扰的可信原因和后果。然后审查现有的流程保护措施以评估 HAZOP 情景的风险。根据 IEC61882，HAZOP 研究已证明在各种不同的过程工业中非常有用。然而，该方法在潜在应用中存在一些需要考虑的局限性，特别是对于复杂过程。本文旨在通过应用 MPM 算法对过程保护策略进行全面和更准确的后果分析和有效评估来改进 HAZOP 研究。所提出的监控算法由两层决策边界组成。1^第一层是基于正常工作状态和所述2^次层（即在本文中所建议的）是基于植物的安全工作状态。通过对运行中的复杂聚合设备的调查验证了所提出方法的有效性。