Research On Process Mining Method For Hull Block Outfield Operation Logistics

In recent years,with the continuous improvement of the information and intelligence level of shipyards,a large amount of data will be generated throughout the entire life cycle of ship products.The full use of these data can effectively improve the quality and efficiency of shipbuilding.Process mining is a new discipline in the field of big data that provides a full set of tools to insight into facts and support process improvement.Using process mining technology to mine and analyze various types of data generated in the shipbuilding process has gradually become a new research direction.Under the modern shipbuilding mode,block is the most important intermediate product in the process of ship construction,and its outfield operation logistics process is also particularly important.In order to make full use of the data accumulated in the block outfield operation logistics process,to mine and diagnose problems in the process,so as to provide managers with decision support,this paper studies the process mining method for the hull block outfield operation logistics.It mainly includes the following four aspects:（1）A model discovery study of the block outfield operation logistics process was carried out.Firstly,the technical bases of the model discovery was introduced.Then,a hierarchical analysis model of “case shipyard-case ship-case data” was established.Afterwards,using α mining algorithm and heuristic mining algorithm,the process models were extracted respectively from overall event logs of block outfield operation logistics actual process of case shipyard.At the same time,through conformance check,it was verified that the model discovery effect of the heuristic mining algorithm was significantly better than α mining algorithm.（2）A trace clustering study for block outfield operation logistics process was carried out.A comparison of three main data clustering methods was carried out,and a basic clustering algorithm for trace clustering oriented to this study was determined.The feature vectors of the event logs of the block outfield operation logistics process were defined,and multiple similarity distance calculation methods between process instances,multiple similarity distance selection methods between clusters and the clustering result evaluation method based on silhouette coefficient were introduced,and a mathematical model of trace clustering for this study was built.Finally,a clustering and result analysis of case data was performed using trace clustering algorithm based on agglomerative hierarchical clustering,and the effectiveness of the algorithm was verified.（3）A comparative analysis study for the planning process and the actual process was carried out.A comparative analysis method based on multiple perspectives and multiple indicators from the perspective of process model fitness,the perspective of process model structure,the perspective of operation task,and the perspective of process instance was established.At the same time,a comparative analysis of the utilization rate of the block stockyard and assembling yard was performed.The problems of block outfield operation logistics process of the case shipyard were diagnosed,and the reasons of these problems were analyzed based on the opinions of experts.（4）A process mining system for block outfield operation logistics was built.The application test of the system through case data was performed,and the test results were consistent with the previous mining results,which verifies the effectiveness of the system in the process mining of block outfield operation logistics,and partly enhances the engineering application value studied in this paper.