Research On Association Rules Of Material Between Surface Corrosion Grade And Environment Factors Based On Ontology

"Material corrode in natural environment" as a subject area which not onlyaccumulated the material corrosion data, but also focused on solving the the evaluation ofnature environmental corrosivity and predicting a long material corrosion behavior andworking life from results of short-term accelerated corrosion tests.Surface materialcorrosion is caused by interactions between environmental factors and material during itsservice time. The surface corrosion information of material, mainly determined thematerial corrosion grade through the size and area of corrosion feature.At present,in natural environmental corrosion test,researchers collect,deal with,storage and manage the needed information to achieve inquiries and applications bybuilding material corrosion databases and database management software.However,withthe deeper research was extensively tested and experimental parameters wasincreased,which made the information to be exponential growth.How to find potential andvaluable information with extensive experimental research data was the most concernedissue of researchers.Because surface corrosion characteristics determined the corrosiongrade, and it had a close relationship with environmental factors.Therefore,to research therelation between external corrosion grade and environmental factors was a practical issuethat needed introduce computer technology and others to be solved.With the information of corrosion features accumulated,the rapidly increasing ofinformation,the changing of test parameters, the traditional association rule miningmethods could not apply for the real needs, a lot of data could not be effectively extractedand analyzed to be under-utilization. Meanwhile, different corrosion test stations haddifferent ways of obtaining and storaging corrosion information,those producedheterogeneous datasets. How to resolve the relationship between external corrosion leveland environmental factors needed introduce association rule mining techniques based onontology, and used it into material corrosion field.According to material surface corrosion characteristics and composition, the paper didthe following studies:（1）Studied the composition of material surface corrosion characteristics informationand environmental factors,handling method and evaluating level of material externalcorrosive feature.（2）With the feature and purpose of corrosion data, established domain ontology and global ontology,according to data sources of each test station established localontology,defined the mapping and inference rules between loacal ontology and globalontology. Designed a material corrosion heterogeneous data integration system based onontology,which generated ETL rules through the mapping rules.And decomposed inquiriesinto sub-queries for each source database according to ETL rules,finally returned thesub-queries results to the queries interface of integration system,to achieve the integrationof heterogeneous data sources.（3）Designed a multiple association rule algorithm based on matrix.Because theapriori algorithm had a low efficiency when applied in multiple association rulemining,proposed a improved algorithm—QMMA.The experimental results showed thatQMMA algorithm has smaller time complexity and space complexity than Apriorialgorithm.（4）Divided the classes of environmental factors according to national standards ofenvironmental and harsh environment degree from each test stations.Using QMMAalgorithm to mine association rules between corrosion grade and environmental factors, gettest data of strong association rules.