基于数据驱动的石化装置静态报警抑制策略研究

由生产状态变化引起的误报警频发,为解决现有针对基于状态报警的抑制方法缺乏完整性与定量分析的问题,提出了新的报警抑制策略。通过分析报警记录等数据库,对引起报警的生产条件进行结构化整定并量化,利用分类筛选与数据过滤找出基于状态的报警,再结合关联性分析和概率判断,建立了基于数据驱动的静态报警抑制策略。使用现场数据的试验证明了该方法的有效性。

Strategic study for the data-driven based static alarming suppression for the petrifaction installations and facilities

An innovated alarming suppression strategy is proposed to reduce the false alarms.As is often the case,false alarms may come up due to the improper routine operation in the manufacturing process,which can be harmless in a direct sense,but is likely to make effects on the regular production process,particularly,in case that there is no quantitative alarm suppression strategy at hand.It is just because of the fact that we have seen the problem that has pushed us to settle the problem in mind and developed a new static alarm suppression strategy.For the above mentioned purpose,we have first of all proposed a model of the basic structure of the strategy on the basis of the qualified summary of the designing principles of the concerned strategy.By means of such databases as the alarming logs and DCS database,we have worked out the logical data and turned the processing parameters data in view of the processing conditions in the constructive data formulae,which can be divided into the conditions related to the alarm production and the suppression of the alarms so as to transform them into the quantified suppression triggering conditions.And,then,we have identified and classified the state-based alarming making and data-filtering means and devices for the alarm system to work out.And,in so doing,the paper refers to the probability judgment to the analysis of the relation between the state-based alarms and the suppression triggering conditions,so as to establish a formula to check whether the alarms and the corresponding suppressing conditions can be built into an alarm-production and the alarm-suppressing device system.Moreover,the paper has also developed an alarm-signal classification model so as to settle down the situation when the multi-conditions come up triggering the so-called synchronous alarms.And,finally,we have adopted the field data status-in-situ to testify the static suppression strategy under way to examine if it is feasible and available.The results we have gained demonstrate that the alarming system we are working at has turned out well-improved due to the application of the new static suppression system.As compared with the currently commercial-available similar false alarm-suppression systems,the new alarming suppression apparatus system,with a systematized and integrated process,helps to ensure no key information ignored,thus significantly raising its reliability and safety of application by collection and classification both of quantitative and qualitative data and information.