基于Akima算法和失效概率分布函数的站场异常工况预警

打开文本图片集

Early warning technology of abnormal conditions in stations based on Akima algorithm and failure probability distribution function

HU Jianlin PetroChina Huabei Oilfield Company，Renqiu O62552，China

Abstract：Toimprovetheinherentsafetylevelofoilandgasstationsandrealizereal-timeearlywarningunderabnomalconditions， thispaperintegratesthemethodbasedonpriorknowledgewithdata-drivenmethods.TheAkimaalgoritmisfrstusedtosmoththe processparameters，andtentedeviationvelocityandaveragefaulttimearecalculated.Theprobabilitydistributionofpreviousfailure dataifittddvatioeloitistegatedtoueproabilitrtioc.allsiiit issetthroughdebging.earlyaingofbnoalconditiosisompleted，andthfldverificationisaidout.euls showthatheRungephenomenondosnotappearinthecurveobtainedbytheAkimainterpolatiomethod.Thecurvepasesthroughall theoriginaldatapointsandretainsthechangetrendoftheoriginalcurve.ThewaringresultsasedontheAkimaalgorithmandfailure probabilitydisributionfunctioncanbeadvancedfurtherthantheconventionalones.Thewaringtimeofheatertemperature，flashtower presure，andseparatrlevelcanbeadvancedyO7，234，and3sspectively.Theinstabilityofparameterchangeswilgeatly shortenthewaingtihileelativelyableparameterangeswillimittedvancingofthewaingi.Thresearchlsan provide a practical reference for improving the integrity management level of the stations.

Keywords：Akima algorithm;failure probabilitydistributionfunction;abnormalconditions;earlywarning; deviationvelocity

油气站场内涉及原油脱水、原油稳定、伴生气脱水、伴生气脱酸、轻烃回收和采出水处理等多个工艺过程，涉及的介质大多具有易燃、易爆、有毒等特性，且生产过程具有连续性，一旦出现异常工况，不仅影响正常油气生产，还可能引发火灾、爆炸等重大事故[-2]。（剩余5817字）