Research On Abnormal Condition And Self-healing Of Fused Magnesium Furnace Based On Knowledge

Fused magnesia is a widely used refractory material,which is produced,in most cases,by a unique three-phase AC fused magnesium furnace(FMF).In the process,the frequent changes in raw materials impurities will lead to the set value of current unreasonable.The FMF will have abnormal conditions when it operates long-term at a unreasonable current setting.When abnormal conditions happen,it will lead to product quality decline or worse may result in accidents.The purpose of this paper is to discover anomalies in a timely manner,and constantly correct the current setting under the premise of ensuring safety in production,so that the FMF can gradually return to normal operation.FMF melting process is a typical complex industrial process,which can not establish an accurate mathematical model.However,on-site operators,by virtue of their own experience combined with the current simultaneously and observed the furnace,can give the appropriate current correction.This article aims at how to use the empirical knowledge of on-site operators to identify and self-healing the abnormal conditions.This article mainly completed the following aspects of work:first of all,the image features of the fused magnesia furnace are analyzed,and the image filtering processing and segmentation of the target area are carried out to extract the relevant features;On the basis of summarizing the previous researches,the short-time energy of current change rate,a new feature,is proposed to effectively distinguish the two abnormal states;And also done some work about data fusion,such as the information from different sensors were sequence synchronized and normalized.Secondly,the rules of abnormal identification are extracted,and the evaluation model of abnormal conditions of the FMF based on extension logic is established.In the determination of the weight of the relevant features,the method of game theory is adopted to get the best weight,which combines the subjective weight with the objective weight;And also identified the classic matter and the cut-off element.The accuracy of the model is verified by comparison at last.Finally,a self-healing control model based on case-based reasoning technology is established.The model realizes the current correction value based on the omni-directional description of the current running state of the FMF.The establishment process of 4R is introduced in detail,and the rationality and effectiveness of the model are verified by simulation experiments in this paper.It is worth mentioning that the case-based reasoning system established in this paper can be used to evaluate and correct the case results online,which is not found in the common case-based reasoning system.