Design And Implementation Of Crude Oil Physical Property Modeling Software Based On Nuclear Magnetic Resonance Spectroscopy

Oil is an important source of energy for our daily traffic and national industrial production.Refining business refer to unprocessed oil as crude oil.But,the crude oil must go through a series of process of processing,in order to get useful all kinds of petroleum products.The physical properties of crude oil are the main indicators in the production and research to assess the quality of oil and control the processing,we need to determine its various physical properties,before processing crude oil,such as acid number,relative density,residual carbon,sulfur content.Crude oil comes from different fields in the world,and even in the same oil production plant,each oil production team's crude oil properties are not the same.And the refinery's crude oil processing process is complicated,Small changes in physical properties of crude oil will cause the fluctuation of the subsequent production and processing equipment.Therefore,it is very important for the stability of the production process and the benefit of the enterprise to master the crude oil data before the crude refining.Moreover,most domestic refineries adopt the crude oil blending method to mix and process each crude oil,if we use the traditional way of laboratory tests to obtain mixed crude oil's physical property data,there are a lot of problems,such as testing cycle is long,physical quantity is little,and it is do not have timeliness.The current rapid evaluation method of crude oil is that we can obtain the spectral data by using near-infrared instruments,and on the basis of this,the physical property data of crude oil are obtained through the property evaluation software,However,the quality of domestic crude oil is poor and the water content is high.Scanning using near-infrared instruments can not obtain accurate spectral data.In addition,the existing crude evaluation software is a direct encapsulated training model or a small number of algorithms,which can not meet the actual production needs of different enterprises.According at above problems,in this thesis,based on the oil refining production process,supported by the National Natural Science Foundation project "the operation theory and implementation technology of global synergistic optimization of oil refining production process(61590922)",we research crude oil physical property evaluation methods and modeling algorithms related processes and provides algorithm extension and data modeling capabilities to develop crude oil properties modeling software based on nuclear magnetic resonance spectroscopy.The main work is as follows:(1)In view of the request of crude oil physical property evaluation for oil refining process,this thesis analyzes the function and performance requirement of physical property modeling software for crude oil and designs a three-layer architecture model with front and rear ends separated,front-end adopts modular structure design,Service-oriented architecture design,to realize the algorithm expansion,data modeling,physical property evaluation analysis and other functions of the oil physical modeling software.(2)We use a variety of advanced programming techniques to complete the development of various functional modules.Based on the service-oriented architecture,the front-end humancomputer interface is separated from the back-end logic business processing.The Vue.js framework is used to realize the componentization development of the front-end interactive interface.The components are highly multiplexed to improve system maintainability and extensibility Using Vuex to complete the data communication between the components,unified management of the state of the data to decouple the components;the use of Node.js using Koa to complete the back-end database operations service development and use Flask to complete the algorithm service encapsulation;and to complete before and after End API data interface and MySQL database table design and implementation.(3)Finally,this thesis tests the performance of the software and the adaptability of each device,which proves that the software has higher running performance.At the same time,the actual production data of the refinery are used to validate the core functions of the physical property evaluation software of crude oil,including data modeling capabilities,the use of modeling algorithms to model the physical properties that need to be evaluated and the function of model analysis and preservation.The spectrum evaluation module provides users with crude oil One-click evaluation of physical properties,as well as evaluation results and standard results of the data contrast.Finally,the model is built by a variety of built-in algorithms to evaluate the physical properties.The accuracy and real-time performance of each algorithm model are compared and analyzed.The proven system has a high efficiency and accuracy in time property evaluation.