| With the continuous innovation and development of science and technology in the field of oil and gas production,just the section of exploration and production of oil and gas production industry,has formed more than 100 application systems in only ten years.These systems can avoid duplication of data entry,achieve switching control of remote device,realize automatic monitoring and diagnosis of the productive process,and so on.However,most of them have a phenomenon that the application services are programmed repeatedly.China Petroleum Oil & Gas Production Networking Project-A11 Project Team-proposes to build an integrated platform for oil and gas production management based on Paa S.This platform uses micro-service architecture to decompose a single application into a number of manageable micro-service methods,which can be used by many times only once been programmed.It greatly improves the quality and development efficiency of the software.Combining with specific domain features,this thesis mainly studies facet description method for oil and gas production and improved clustering algorithm from the perspective of software reusing.Specific design works are as follows:1.According to the large number of application services and deep domain of related terms in the China Petroleum management platform,a description method based on the combination of facet and domain ontology for oil and gas production is proposed.From the point of view of optimizing domain terms,it extracts micro services’ segmentations of named entity and converts them to synonyms,aimed to compresses the vector space of domain feature words.It uses the improved domain feature word selecting algorithm to emphasize the difference between domain words and common words,and further filter domain feature words.Experimental results show that after converting word segmentation,the vector space of candidate feature set is reduced by 47%.The feature word vector generated by the improved domain feature word selecting algorithm has strong domain.All of the micro-service applications in the platform can be described by the method independently and comprehensively.2.In order to solve the problem that the traditional K-means algorithm needs to determine the number of clusters in advance and get the initial clustering center randomly,which is easy to fall into the local optimal solution,an improved K-means algorithm is proposed.Based on the previous description method,it can obtain the faceted description vector set for all micro service applications in the platform.First,the appropriate clustering number k value is determined by establishing a change trend between the similarity parameter ε and the number of k.And then the initial clustering center points are confirmed according to the principle that the density of initial clustering centers is higher and the distance between each other is far.The experimental results show that the improved clustering algorithm has higher clustering accuracy and lower error compared with the traditional clustering algorithm and other improved algorithms.It can effectively realize the automatic clustering of micro services in the platform,and reduce the impact of human subjective factors on the classification results.In addition,this algorithm has good global convergence ability,which can effectively avoid clustering into local optimum.Although a series of studies on facet description methods for oil and gas production and clustering algorithm have been carried out in this thesis.Taking into account the implementation effect of micro services automatic classification,there are also many aspects still need further exp loration and research.For example,the complement of the ontology relation conversion table,the calculation of the similarity between the entity word and the domain feature word,the unique value of the similarity parameter ε,and so on. |
PDF Full Text Request