基于传递熵与改进K2算法生成报警传播网络

工业报警变量数量增多所导致的“报警泛滥”问题, 严重影响了报警系统的应有功能. 针对此问题, 提出一种从过程报警数据集中学习报警变量传递关系的方法. 首先, 利用传递熵具有准确衡量一阶或多阶自相关性变量间因果关系的特点, 识别变量间的因果关系. 其次以变量间熵大小为准则, 保留传递信息量较大的节点, 最后考虑变量在不同状态所占时间比重对K2算法进行改进, 学习得到最终的报警传播网络. 通过在田纳西伊斯曼过程数据集上的验证发现, 该方法能够对报警发生的根本原因做出判断并较好地学习得到报警传播路径.     

智能报警管理若干研究问题

报警是及时发现生产过程中异常情况的直接手段,因此科学、高效、准确的智能报警管理对提高生产过程安全性和可操作性至关重要,也是进一步进行优化、控制和故障诊断的基础.本文就智能报警管理这一研究领域进行综述,首先从工业需求上论述报警管理的重要意义,说明报警对于过程安全的影响,指出数量过多和难以辨别是主要难题,并将现实情况与国际标准相对比;然后概述了报警管理的概念和内涵,介绍报警系统改造的全生命周期.其次总结该领域的主要研究问题,并针对单个报警和多个报警的设计展开讨论,揭示误报与漏报、检测速度与鲁棒性的矛盾,并给出常用的解决方法.针对单个报警,通过滤波、延迟和死区,对过程时间序列进行处理,实现报警的快速和准确.针对多个报警,通过多元统计分析全面表征多个变量之间的关系,利用相关性分析和传递熵分析来识别因果报警.最后通过一个实例来说明智能报警管理的过程.     

基于报警时序信息挖掘的报警关联分析方法

面对现代流程工业监控系统报警泛滥问题,为了寻找报警根源以减少无效报警,并针对传统方法在面对大量的报警数据时计算效率低下的问题,提出了一种基于报警数据时序信息挖掘的报警关联分析方法.通过将报警时序信息进行区块化处理,将报警时间序列转换为报警时间的节点序列,然后将区块之间的匹配度作为报警关联度的评价标准,减少了关联分析的运算量;采用滑动窗口比对计算相关报警时间序列的时间关联信息;采用TE过程（Tennessee Eastman process）数据,验证了所提方法的有效性.     

过程工业报警系统研究进展

系统、科学、有效的报警系统对提高生产过程安全、质量和可操作性能至关重要,引起了工业界和学术界的广泛关注,产生了大量研究成果。针对过程报警系统的最新研究进展,分别从报警系统优化设计、报警系统监控、报警系统性能评估这3个热点研究方向总结研究方法,并对不同方法进行了对比讨论。针对报警系统优化设计,利用ROC分析、相关性分析等方法对报警阈值进行优化,并对报警变量进行优先级划分;针对报警系统监控,利用相关性分析实现滋扰报警识别,基于过程知识和数据驱动实现因果报警识别,基于时间窗或运行长度分布实现抖振报警识别;针对报警系统性能评估,依据过程报警管理标准或动态风险分析等实现定性定量评估。最后,指出报警系统研究有价值的方向和趋势,并对报警系统研究前景做出展望。     

火灾报警系统报警处理与信息存储研究

火灾报警系统由多个区域报警控制器和集中报警控制器组成.文章介绍了一种用DS18B20作为测温报警元件,以RS485通信总线为主干网的火灾报警系统,阐述了系统的构成和报警功能基本实现.说明了对应房间号报警上限存储、报警处理与报警信息存储的方法.对修改报警上限时如何寻找存储器E2PROM对应的地址进行了分析.设计了报警处理的基本程序.实现了能够将报警温度与报警地点及报警时间进行存储、能够以声光等多种方式进行报警的功能.解决了传感器、存储器E2PROM、实时时钟电路和报警电路等多个方面的软硬件设计与协调.文章中介绍的方法较好地实现了火灾报警系统中当温度升高超出报警上限时,快速准确地预报.现场模拟实验表明,系统安全且稳定可靠.     

专家系统在火电厂汽包水位高报警诊断中的应用

为了及时、方便地根据火电厂的报警信息诊断出故障原因,使用Visual C 和专家系统工具CLIPS开发了火电厂报警诊断专家系统.给出了该系统的总体结构并介绍了其功能.同时详细说明了汽包水位高报警诊断的实现过程.实践证明该系统界面友好,功能完善,知识库易于维护和更新.     

基于相关系数的多变量异常数据段的检测

工业报警系统普遍存在报警过多的问题,为了获取过程变量的正常与异常历史数据段,从而设计工业报警系统的动态报警阈值,进行性能评估和优化,提出了一种基于过程变量之间的相关系数进行异常数据段检测的改进方法。该检测方法将相关系数和数据段长度作为分段依据,对过程变量的多元时间序列进行分段,采用Spearman秩相关系数获取变量之间的相关关系,从而检测出不符合正常趋势的异常数据段。仿真算例和工业案例表明该方法可真实地反映过程变量之间的相关关系,并准确检测出所存在的异常数据段。     

防盗报警类／Intrusion Alarm（新品精选）

AW-BM1600大型总线网络报警控制管理主机;NBX系列数码八变频红外栅栏;EP210防盗报警主机;KS-600C／D远程无线报警系统;YL-007M3A夜狼“慧信通”GSM短信拨号报警系统;SK-239GGSM／PSTN双网报警控制器主机。     

入侵检测报警聚合与关联系统设计与实现

入侵检测系统的大部分报警事件之间都存在某种联系。通过对这些报警的聚合与关联能够消除或减少重复报警，降低误报率及发现高层多步攻击策略。论文设计并实现了一种报警聚合与关联系统，系统主要包括报警聚合、报警校验、多步攻击报警关联和报告分析与规则控制等部分。实验证明：该系统能够减少报警数量，并能识别攻击意图，达到预警的目的。     

DCS报警信息的处理和报警的组态

自动控制技术人员应当熟练掌握分散控制系统（DistributedControlSystem，DCS）报警信息处理和报警组态的技能，并使整个工业生产过程始终处于有效的监视、控制和管理状态。本文以DCS市场占有量较大的Ovation系统为例，图文并茂地说明了DCS报警信息处理的基本内容和报警组态的主要步骤，也为不同品牌DCS报警信息处理及其组态，提供了参考。     

An accelerated alignment method for analyzing time sequences of industrial alarm floods

In industrial processes, analyzing and predicting process faults are quite important, which could help operators to take timely and effective responses to ensure process safety and prevent further losses, especially during alarm floods. Various fault analysis methods have been proposed so far, among which the alarm flood sequence alignment (AFSA) methods, unlike other traditional model-based or statistical methods, provide fault inference from the perspective of alarm sequence similarity assessment. A new AFSA method, the match-based accelerated alignment (MAA) is proposed to generate insightful and informative alarm sequence alignments. MAA focuses mainly on alarm match analysis and outperforms other methods in terms of robustness towards nuisance alarms and improved computational efficiency. More importantly, the alarm time information is well considered and explored in MAA, rendering its alignment results capable of revealing the real similarity of alarm floods. Numerical examples and a real chemical plant case are studied to demonstrate the effectiveness and efficiency of the proposed MAA method.     

煤矿瓦斯超限分矿分级防信息阻塞短信报警系统设计

针对目前煤矿瓦斯超限短信报警存在报警信息不分级别或只分级别不分矿名发送、短信发送设备容易死机造成信息阻塞等问题,设计了一种煤矿瓦斯超限分矿分级防信息阻塞短信报警系统。该系统通过对煤矿数字化瓦斯远程监控系统实时采集的数据进行分析处理,将超过各级报警阈值的数据根据不同的矿名和阈值级别,通过具有死机自动复位功能的多接口短信发送装置分别发送至相应矿相应级别的人员手机上。实际应用结果表明,该系统改善了煤矿安全分级监管、分级响应的能力,提高了安全管理的可靠性。     

A Review of Alarm System Design for Advanced Control Rooms of Nuclear Power Plants

Alarm systems are designed to provide cues to make operators aware of an operational problem, so that mitigation actions can be taken. New technologies are enabling innovative designs of alarm systems for safety-critical systems. To help better design alarm systems in advanced control rooms of nuclear power plants (NPPs), recent researches on alarm system design in the NPP domain are reviewed from the following aspects: (1) functions of an alarm system and opportunities and challenges of new technologies; (2) effects of alarm floods and different alarm management methods; (3) visual presentation of alarm information; (4) evaluation criteria for alarm systems. Through the comprehensive literature review, this article proposes research topics related to human factors and ergonomics that are worth studying in advanced alarm systems of NPPs.     

Unified problem modeling language for knowledge engineering of complex systems-part II

Real time applications to control industrial, medical, scientific, consumer, environmental and other processes is rapidly growing. Today such systems can be found in nuclear power stations, computer-controlled chemical plants, flight control, etc. This growth, however, has also brought to the forefront some of problems with the existing technologies. In domains like real-time alarm processing in a power system control centre existing technologies like expert systems cannot efficiently cope with. These problems have pushed for research into new techniques which could be used for solving these problems. The problems range from among other aspects, the enormous size of the power system and the fast response time constraints in emergency situations. In this paper we describe the application of the Intelligent Multi-Agent Hybrid Distributed Architecture for real-time alarm processing in a power system control centre. We show how the IMAHDA architecture is able to model the complexity and size of the power system as well as meet the desired response time constraints. Implementation of a large scale real time system like alarm processing involves realization of various objectives. These include methodology related objectives, domain related objectives, and management related objectives. This paper also describes the realization of these objectives.     

Performance assessment for generalized delay-timers in alarm configuration

False and nuisance alarms are major problems in the process industry. Generalized delay-timer is an effective technique to improve the performance of an alarm system. However, due to the increasing intermediate states, computation of the performance indices, false alarm rate (FAR), missed alarm rate (MAR) and expected detection delay (EDD), is hard and unsolved for generalized delay timer. In this paper, we first propose general formulae to compute FAR and MAR by adequately reconsidering the internal relationship between the no-alarm state and the alarm state. Then the computation of EDD is discussed via the probability of alarm or no-alarm at one sample by iteration. Simulation and industrial results show that alarm performance can be improved via generalized delay-timer with proper choice of the order.     

A distributed parallel alarm management strategy for alarm reduction in chemical plants

A distributed parallel alarm management strategy based on massive historical alarms and distributed clustering algorithm is proposed to reduce the number of alarms presented to operators in modern chemical plants. Due to the large and growing scale of historical alarms as the basis of analysis, it is difficult for traditional alarm management strategy to store and analyze all alarms efficiently. In this paper, by designing the row key and storage structure in a distributed extensible NoSQL database, the strategy spreads alarm data in a group of commercial machines, which ensures the capacity and scalability of the whole system. Meanwhile, Distributed Parallel Query Model (DPQM) proposed as a unified query model provides efficient query and better integration of distributed platform. Based on the characteristics of alarms and time-delay correlation of alarm occurrence, alarm similarity criteria are proposed to effectively identify repetitive and homologous alarms. In order to group massive alarm data, a new distributed clustering algorithm is designed to work concurrently in MapReduce frameworks. The test results using alarm data from real chemical plants show that the strategy is better than traditional method based on MySQL at system performance, and provides excellent redundant alarm suppression in both normal situation and alarm flooding situation.     

基于以太网可燃性气体报警器的研制

为了能够准确有效地对工业现场的可燃气体进行连续自动的多点检测,设计一种基于工业以太网技术的可燃性气体报警器.采用高性能的微处理器芯片LPC2214为主芯片,并在LPC2214上移植嵌入式操作系统μC/OS-Ⅱ和TCP/IP协议栈uIP,作为该报警器应用程序的开发平台.实现了现场气体远程检测,具有气体浓度显示和声光报警功能,并且提供了工业以太网接口,可以将检测到的数据通过工业以太网上传给上位机.为工业以太网现场设备的设计提供了可以借鉴的方法和应用范例.     

实验室智能安防报警系统研究

本文主要对实验室智能安防报警系统进行了研究,完成了一种远程安防报警装置控制系统的设计,该系统综合运用单片机、ZigBee无线通信技术,在介绍了系统基本架构及工作原理的基础上,完成了包括安防报警传感器选型在内的软硬件设计过程。该智能安防报警系统完成各传感器信息实时高效的采集和逻辑判断后,向单实验室信息集中节点反馈经处理后的信息,再由主节点汇总处理各集中节点信息并进行实时显示。