Abstract:

In the petroleum system, process fault and its damage caused by the fault propagation chains have a serious effect on the oil and gas production and people's life. In this paper, from the fault causal chain perspective, a quantitative safety early warning method of fault propagation is proposed. First, the fault propagation chains is analyzed, and the GTST-DMLD (goal tree success tree, GTST; dynamic master logic diagram, DMLD) model is developed, by which the chain effect of the fault propagation behavior can be quantitatively modeled. It can be further used to evaluate the system safety state when the abnormal event happens. Based on the results of GTST-DMLD model, Markov process theory is introduced to establish a fault prognosis model, by which the fault consequence and its propagation direction can be predicted timely. The results given by the proposed method can help field operation personnel to carry out pre-active maintenance or emergency disposal. In the case study, the proposed method is applied to the atmospheric distillation unit and the vacuum furnace in a certain petrochemical company. It has been fully validated and provides a great help for the risk reduction and accident prevention in the field.

Key words: process systems, goal tree success tree-dynamic master logic diagram, fault propagation, safety, prediction

摘要：

炼化装置故障及其故障链式效应对油气生产和人民生命安全所造成的危害严重。从故障链角度进行事故风险研究，提出炼化装置故障链式效应定量安全预警方法。首先分析炼化装置故障链式关系结构，基于目标树成功树-动态主逻辑图（GTST-DMLD）建立其故障链式效应关系模型，揭示炼化装置故障链式效应行为规律和关联本质，从而评价装置异常工况下的安全状态。进一步以马尔可夫过程为理论基础，建立故障链式效应预测模型，预测故障传播的后果和方向，并计算各后果的发生概率，为现场操作人员进行主动维修或应急处置提供依据。案例分析中通过对某化工厂常压塔装置、减压炉装置为研究对象进行应用与验证，结果表明该方法可以准确地对系统故障发生后的状态进行评价和预测，方法有效、可行，便于操作人员在处置已有故障的同时，注意预防其他异常工况的发生，降低油气生产加工过程中的整体风险。

关键词: 过程系统, 目标树成功树-动态主逻辑图, 故障链式效应, 安全, 预测