Research On Association Mining Based On Item Constraints In Safety Early Warning Of Drilling Operations

Enterprise should keep in mind production safety, especially drilling business. If the work of safety warning does well, and the hazard can be controlled effectively before the accident, then the drilling operation will be better. According to the different effects on accident, the hazards can be divided into three categories:inherent hazard, triggered hazard and tissue hazard. Once hazards indifferent types overlap in time and space along with the defense failure, an accident may occur. Association rules mining can find the links between those items. Thus it is natural for us to apply data mining to hazards. We expect the method mining hazard association rules which base on three types hazard theory. We concern the records whose hazards in different types overlap in time and space, so when the filter degree reaches the user-specified minimum degree, we can take advantage of the constraints on data filtering before mining.There aren't available algorithms since the specific constrain, "three types of hazard" principle, is different from the existing model constraint. The interesting hazard association rules also aren't filtered after mining by directly use Apriori algorithm. Not only because it was unable to take advantage of the constraints in the mining process, and the final rule pattern will be complicated and uncontrolled due to the increase classification. We propose DApriori (Dimensions Apriori) algorithm which match this constraint based on Apriori, then we verify the validity of the algorithm by applying it to the daily supervision and inspection data which comes from a drilling company. Experimental results show that association rules mining base on constraint can be applied to the safety warning. The hazard association rules mining by DApriori algorithm are all meaningful.Compared with the Apriori algorithm, the DApriori algorithm can filter the redundant association rules and has better efficiency.