HAZOP知识图谱构建方法

危险与可操作分析方法（hazard and operability analysis,HAZOP）主要采用头脑风暴的形式,将讨论结果记录在纸质文档中,该方法过于依赖专家经验且大量现有的HAZOP信息并没有得到共享与复用。针对此问题,本文提出了结合自顶向下和自底向上的HAZOP知识图谱半自动构建方法。在解析HAZOP信息基础上,结合HAZOP分析国际标准IEC 61882,以危险事件角度设计了HAZOP本体规则,完成了模式层的构建;以现有HAZOP数据为基础,在BiLSTM-CRF命名实体识别模型基础上引入了Attention机制,实现了关键HAZOP信息的自动提取并利用图数据库进行数据存储,完成了数据层的构建。以某油品装置HAZOP分析报告为例构建HAZOP知识图谱,验证了构建方法的有效性。结果表明HAZOP知识图谱能够清晰地展示HAZOP知识之间的联系,快速提供现有HAZOP分析信息,辅助人工HAZOP的开展,降低人工成本与时间成本,也为后期项目的生产与维护提供知识支撑,实现了HAZOP信息的共享与复用。     

基于图形化剧情对象模型的安全评价及其应用

HAZOP分析的信息包括内容信息与结构信息两方面。采用基于知识本体的剧情对象模型,能够全面地表达安全评价信息的内容和结构,解决了传统HAZOP方法存在的信息隐含、信息缺失等问题。在此基础上开发的图形化剧情对象模型,有效地实现了计算机辅助自动安全评价和安全信息的传递、复查和共享。通过对实际案例的分析表明,图形化的剧情对象模型为安全评价信息的标准化奠定了基础,同时也为提高安全评价的质量和效率提供了强有力的手段。     

基于本体的HAZOP因果知识描述模型

传统HAZOP分析通过专家头脑风暴的讨论形式进行,耗时耗力且无法保证分析结果完备性。针对此问题,本文在语义模型改进HAZOP分析方法的研究领域,提出一种基于本体的HAZOP因果知识描述模型。本文首先对传统HAZOP分析文本抽取共性知识,获得自然语言表述的HAZOP因果知识描述模型结构;再借助本体论的方法,通过对HAZOP因果知识中类、属性、实例的定义实现模型结构化修正;最后,使用本体开发软件Protégé将HAZOP因果知识描述模型通过OWL语言进行表示,构建了基于本体的HAZOP因果知识描述模型,可辅助人工HAZOP分析,提高人工HAZOP分析的效率,同时保障分析结果的完备性。本文以寻找储罐液位变化的致因、后果为例,对该HAZOP因果知识本体的推理过程进行验证,结果表明,该HAZOP因果知识本体可辅助人工推理分析过程,同时发掘HAZOP分析过程的浅层、深层知识,保障分析结果的深度和完备性。     

基于知识本体的过程安全分析信息标准化

过程危险分析的主要目标是识别危险剧情。危险剧情能够表达团队“头脑风暴”安全评价过程也能表达评价结论。危险剧情的知识本体是标准化过程安全分析信息的准确描述。知识本体是概念表达的明确规范。依据设计知识本体所遵循的规则提出了一种过程安全分析信息标准化方法,称为剧情对象模型（scenario object model,SOM）。SOM能够表达安全分析信息的内容和结构,能够实施计算机自动推理和半定量计算。应用知识本体SOM有效实现了计算机辅助自动安全评价和安全信息的传递、复查和共享。     

石油勘探开发领域本体的构建方法研究

石油勘探开发领域中涉及的专业有20多个,由于各专业之间的信息术语不统一,给应用集成和信息共享带来了很大的麻烦,而采用本体构建理论可以有效的解决这一问题。在了解和分析石油勘探开发业务特点的基础上,本体论提出了一套构建石油领域本体Petro-Onto的方法,建立了相应的顶层本体框架,提出了以数据模型和业务模型为参照体系的自动捕获本体的方法。目前,在油田信息集成中Petro—Onto得到了广泛应用。     

智能化风险及可操作性（HAZOP）分析系统研究进展

通过梳理智能化风险及可操作性（HAZOP）分析的各类方法,将其分为定性分析和定量分析两大领域,详细介绍了不同的研究方法及其优缺点,并结合实例对有代表性的方法作了详细阐述。智能化HAZOP定性分析系统的研究主要集中在建立专家知识库的方法上,其中基于深层知识建立专家知识库的研究相对较多,该方法虽然能够揭示偏差产生的机理、传递路径,但是分析结果准确度较差。基于经验知识、定性分析与定量分析相结合及基于动态模拟的HAZOP定量分析是实现智能化HAZOP定量分析的3种主要方法,是由模糊定量向精确定量、由静态分析向动态分析发展的过程。基于动态模拟的智能化HAZOP定量分析不仅研究偏差的数值,而且研究偏差的持续时间对过程系统的影响,分析结果更加精确,是较为完善的定量分析方法。该方法借助于商用模拟软件,使不具备丰富经验的分析人员也可以完成精确的HAZOP分析。"动态偏差的阶梯化求解法"使分析动态偏差方便、易行,基于该方法建立的偏差"层级"模型充分说明了处于不同层级的偏差对过程系统的不同影响。在分析智能化HAZOP分析系统发展历程基础上,结合最新研究进展,预测了其发展趋势。     

HAZOP分析结合LOPA分析方法在长输管道站场的应用研究

HAZOP分析源于石化行业,近几年在管道板块开始推广应用,但是危险与可操作性分析方法(HAZOP)在实际应用中存在先天缺陷,现提出将保护层分析(LOPA)与HAZOP分析相结合,能进一步提高HAZOP分析的安全评价能力。首先介绍HAZOP和LOPA的基本原理和分析步骤,然后从两者信息互相共享、分析流程互相紧扣的特点,阐述两者集成的可行性,并给出HAZOP/LOPA的分析步骤,最后将其用于天然气长输管道站场安全分析。结果表明,HAZOP分析结合LOPA分析方法,可以用具体数值来量化风险的大小、评估安全措施的预防效果、确定现有风险是否可以接受以及是否还需要增加安全措施,极大地丰富了HAZOP的分析内容。     

HAZOP分析技术在芳烃抽提联合装置上的应用

危险和可操作性分析对于石化装置的工艺设计和安全管理具有重要意义,本文介绍了HAZOP分析方法的作用、作业流程,并结合风险矩阵法,针对惠炼二期芳烃抽提联合装置进行了HAZOP分析,有效地识别了装置中潜在的风险,提出了降低风险的措施和建议,并明确了建议措施的落实责任方。基于在应用过程中发现的问题,提出了HAZOP技术分析的展望。     

基于过程模拟的HAZOP偏差量化研究与应用

以天然气凝液脱轻烃工艺为例,提出了基于动态模拟的HAZOP分析方法,利用HYSYS动态模拟对HAZOP偏差进行量化分析,通过过程模拟所获得的偏差偏离正常值所达到具体量,有效解决传统HAZOP分析偏差的模糊性问题,且对工艺潜在的风险以及风险发生与否有一个更直观认识。并在传统HAZOP分析表的基础上提出了基于动态模拟结果的HAZOP分析表,增加了模拟结果分析及安全措施模拟验证等内容;首次提出了"安全域"这一概念,该概念指的是动态模拟时某一参数由初始值到达危险峰值的时间,这段时间的长短可以反映出某偏差造成事故的危险程度及可能性,为处理危险事故场景提供了时间的参考临界值,从而进一步提高HAZOP分析的准确性与可靠性,使HAZOP分析更加完善。     

基于案例推理的HAZOP分析自动化框架

危险与可操作性（HAZOP）分析是一种广泛应用于化学流程工业的危险分析方法。为克服现有的HAZOP分析专家系统在“非常规”分析方面的局限性,提出了基于案例推理（CBR）的HAZOP分析自动化方法,描述了案例库及案例结构,给出了案例搜索策略。为便于案例库的知识管理,开发了案例构造器。工业实例应用结果表明,基于CBR的HAZOP专家系统突破了现有HAZOP专家系统没有学习能力和不能进行“非常规”分析的技术瓶颈。     

基于未确知的化工系统HAZOP分析

基于传统的HAZOP分析存在量化困难的缺陷,提出未确知数-HAZOP分析模型,该模型不仅解决传统HAZOP分析的存在的问题,还可以根据专家经验程度的不同整合专家给出的原因事件的概率区间值。文章利用该模型对运行过程中的化工装置进行了定性定量的分析。未确知数-HAZOP的分析思想对于进行化工事故预防具有重要的意义。     

An integrated qualitative and quantitative modeling framework for computer‐assisted HAZOP studies

The article proposes a novel practical framework for computer‐assisted hazard and operability (HAZOP) that integrates qualitative reasoning about system function with quantitative dynamic simulation in order to facilitate detailed specific HAZOP analysis. The practical framework is demonstrated and validated on a case study concerning a three‐phase separation process. The multilevel flow modeling (MFM) methodology is used to represent the plant goals and functions. First, means‐end analysis is used to identify and formulate the intention of the process design in terms of components, functions, objectives, and goals on different abstraction levels. Based on this abstraction, qualitative functional models are constructed for the process. Next MFM‐specified causal rules are extended with systems specific features to enable proper reasoning. Finally, systematic HAZOP analysis is performed to identify safety critical operations, its causes and consequences. The outcome is a qualitative hazard analysis of selected process deviations from normal operations and their consequences as input to a traditional HAZOP table. The list of unacceptable high risk deviations identified by the qualitative HAZOP analysis is used as input for rigorous analysis and evaluation by the quantitative analysis part of the framework. To this end, dynamic first‐principles modeling is used to simulate the system behavior and thereby complement the results of the qualitative analysis part. The practical framework for computer‐assisted HAZOP studies introduced in this article allows the HAZOP team to devote more attention to high consequence hazards. © 2014 American Institute of Chemical Engineers AIChE J 60: 4150–4173, 2014     

Automatic HAZOP analysis method for unsteady operation in chemical based on qualitative simulation and inference

Comparing with continuous production process, unsteady operation process, such as startup and shutdown, tends to abnormal situations due to a large number of operations of operators and dynamic state changes involved. To guarantee a safe operation, process hazard analysis (PHA) is very important to proactively identify the potential safety problems. In the chemical process industry, hazard and operability (HAZOP) analysis is the mostwidely used method. In this paper, based on proposed qualitative simulation and inferencemethod, an automatic HAZOP analysis method for unsteady operation processes is proposed. Mass transfer and relationships among process variables are expressed by Petri net-directed graphmodel based fuzzy logic. Operating procedure is expressed according to a formal expression. Possible operation deviations from normal operating procedure are identified by using a group of guidewords. Hazards are identified automatically by qualitative simulation and inference when wrong operation process is performed. The method is validated by a rectification column system.     

Active training sample selection and updating strategy for near-infrared model with an industrial application

Training sample selection is widely accepted as an important step in developing a near-infrared (NIR) spectroscopic model. For industrial applications, the initial training dataset is usually selected empirically. This process is time-consuming, and updating the structure of the modeling dataset online is difficult. Considering the static structure of the modeling dataset, the performance of the established NIR model could be degraded in the online process. To cope with this issue, an active training sample selection and updating strategy is proposed in this work. The advantage of the proposed approach is that it can select suitable modeling samples automatically according to the process information. Moreover, it can adjust model coefficients in a timely manner and avoid arbitrary updating effectively. The effectiveness of the proposed method is validated by applying the method to an industrial gasoline blending process.     

Systematic generation of cyclic operating procedures based on timed automata

Manual synthesis of cyclic operating procedure in a realistic system is widely regarded as a difficult task since it is both time-consuming and error-prone. It is thus desirable to develop a systematic approach to automatically generate the optimal schedule of operation steps so as to achieve one or more specific production goal. The timed automata are utilized in the present work for such a purpose. In particular, all components in a given system and the corresponding control specifications are characterized with automata according to the proposed modeling rules. By using parallel composition, a system automaton can be produced with these models and the most appropriate operation path can then be identified accordingly. For any practical application, a sequential function chart and the corresponding Gantt chart can also be easily extracted from this path. Three examples are presented in this paper to demonstrate the feasibility of the proposed approach.