Research On Semantics And Modeling Application Of Metallurgical Coking Process Based On Ontology And Internet Of Things

Coking is an important process of iron and steel production. The quality of coke produced directly influences the efficiency and quality of the subsequent production. The stable operation and optimal management of coking production process are the key factors to determine the quality of coking products and the energy consumption of coke technology. Studying the semantization of coking and its related modeling application is significant to informatization and intelligence of the production of metallurgical industry.Metallurgical coking process is a typical complex system, whose processing mechanism is complicated and has the characteristics of strong non-linearity, large inertia, strong coupling, slow time-variation and so on. Therefore, it is very difficult to establish a precise mathematical model of industrial objects, which makes it harder to manage and control effectively according to the mathematical model. At the same time, it is difficult to achieve the anticipated results through traditional or simple management control methods. Thus, this thesis focuses on three scientific issues, namely data obtaining, feature representing, and knowledge applying aiming at the whole process of coking. Based on the technology of semantic Ontology and description logic, adapting techniques like IOT (Internet of Things), this thesis studies the scene data acquisition and semantic processing, formal description for knowledge, the construction of domain Ontology library and semantic reasoning method in coking production process. Finally, based on reality in actual process of coking, this thesis researches some semantic applications on the basic data management in coking, multi-source heterogeneous resource processing, massive knowledge processing and optimization of heat consumption influence factors etc. Experimental verification has been conducted on the actual data of three large iron and steel enterprises in coking production. The main contents and conclusions of this thesis are as follows:(1) Semantic coking process based on extended description logicCoking involves many ingredients which have different attribute states and some transformation process. In order to achieve better formal description of coking process, class matter-element, extension transformation and anti-relationship builder are introduced based on traditional description logic ALC (Attribute Language Complement). Then according to the characteristics and requirements of the metallurgical coking process, the description logic DL-MCP (Description Logics extended for Metallurgy Coking Process) system of metallurgical coking process is proposed; consequently, the grammar, semanteme and axiomatic system of DL-MCP is constructed; semantic elementes are extracted; and the related knowledge of metallurgical coking process is formally descripted. Example verification shows that the DL-MCP system is stable, practicable, completable and effective. The proposed DL-MCP and the formal description of coking process provide logical basis for semantic management, reasoning and application of coking process.(2) The IOT-based data acquisition and semantic processing in coking processAiming at each subprocesses like heating, gas gathering, gas pipe management, regenerative chamber management and etc., the monitoring index, monitoring attribute and the co-movement characteristics of parameter management are extracted. Then association model of main factors in coking is constructed, and model structure of data acquisition and management is proposed, together with the monitoring network strategy and model for it's corresponding process. On the issues of the construction and communication of network, clustering network structure based on ZigBee, data routing and transmission algorithm based on improved AODV (Ad hoc On-demand Distance Vector Routing) are presented in this thesis. At the same time, The semantic description and expression method of sensor network and sensing data in coking are proposed, and the Sensor Network Ontology and the Sensing Data Ontology are also constructed.Through constructing an IOT-based data acquisition network of coking process, comprehensive accurate and real-time semantic data can be obtained, which could provide an automated and intelligent data acquisition model and method for semantic management and application of coking process.(3) Semantic management of coking processing informationBased on the field investigation of three major iron and steel enterprises, CPD (Coking Process Database) is designed after extracting and concluding basic data and their correlations of the coking process. Meanwhile, the related concepts, their relations and the real examples are extracted, and CPO (Coking Process Ontology) is stored by OWL (Web Ontology Language) based on Ontology technology. In order to facilitate the unified semantic management of data, information and knowledge in coking, the conversion process and conversion rules from CPD to CPO database are proposed. In the purpose of solving the problem of heterogeneity of CPD, a multi-method Ontology mapping model is introduced. After calculating the similarity between entities of different ontologies, the precision ratio and recall ratio of the Ontology mapping results were evaluated. The experimental results indicate that the Ontology mapping model can effectively find a link between Ontology and lay the foundation for Information Inference across ontologies. Therefore, to establish the Ontology repository of coking process, to store CPO by OWL, to solve the heterogeneous issues through the Ontology mapping, and so as to realize the semantic information management of the coking process, which lay the foundation for the reasoning and retrieval service based on semantic Ontology in coking process.(4) Reasoning and retrieval service based on Ontology in coking processBased on coking Ontology database and coking process database, a knowledge reasoning model based on Ontology is constructed. With the help of Jena Inference Engine, knowledge reasoning is conducted based on certain rules. The RDF (Resource Description Framework) representation and transformation method for coking process are presented and SPARQL (Simple Protocol and RDF Query Language) is used to check the consistency of the coking process triple sets. Finally, coking process knowledge retrieval service model and algorithm are given based on the Ontology database. The results of experiments suggest that the proposed semantic reasoning model and data retrieval method are reasonable and could provide effective mechanisms and methods for the subsequent applications of semantic coking process.(5) Semantic applications of coking process based on OntologyOn the basis of semantic reasoning in coking process, applied researches are conducted and some examples were verified aiming at the several key issues existing in the coking process. Four aspects are mainly included, which are: ?the application of basic data management in coking process based on database and Ontology technology, ?the application of semantic fusion technology in the process of multi-source heterogeneous information resources, ?the application of semantic inference method based on RDF graph in the process of mass knowledge processing in coking process, and ?the optimization of heat consumption influence factors of coking process based on semantic reasoning. The experimental results indicate the effectiveness and feasibility of the Ontology based semantic application on coking process. The above mentioned four aspects of the semantic applications solved the following four problems respectively, ? basic data management of coking process is not standardized and unified, ?data and information for the coking process comes from more scattered sources and their forms are not unified, ? it is difficult to conduct analyzing and reasoning efficiently due to large amount of data formed in the coking process, ? there are multi-factors affecting the heat consumption of coking process, and it is hard to realize intelligent managment.Based on the Ontology and the IOT technologies, this thesis studies the data collection, information semantic management, knowledge semantic reasoning and retrieval service, and semantic application of coking process. First of all, based on Ontology technology, an extended description logic and formal description of coking process are established, and the Ontology knowledge base is constructed, which made certain theoretical innovation. Secondly, methodological innovation is made. Based on the IOT designing technology, field data acquisition intelligent network in coking process is constructed, which break through the limitations of wired network technology, combine with the expansion of the existing network, and solve the problems of data monitoring accuracy and limitation of the poor real-time due to bad environment, personnel, hardware equipment difficult to reach. Finally, making the semantic information management and semantic application of whole process of coking as the starting point, the mechanism and algorithm of semantic knowledge representation and semantic reasoning in the process of coking are studied and stated, which consequently provides the basis for semantic and intelligent metallurgical coking process. The study has strong universality, which can be extended to the production process of metallurgical industry, so as to provide a reference for semantic and intelligent application in metallurgical industry production process, and to promote the metallurgical process management automation and intelligent development.