Study On Modeling And Simulation Of Steam Pipe Network And Fault Detection Algorithm

Steam has been used as the main energy transmission medium for heavy industry such as petrochemical industry for a long time due to its excellent characteristics of no pollution and low cost.As a steam transportation equipment,the steam pipe network transports steam to various steam consumption points.Due to the characteristics of large scale and wide distribution of the steam pipe network,and the fact that there are hardware equipment aging failures,it is difficult to identify and control the running state of the pipe network;the measurement instrument data usually shows incompleteness,low accuracy,and poor consistency.These problems bring huge challenges to enterprise management in terms of energy management,real-time scheduling and optimal control,which directly affect the economy and safety of enterprises.Faced with the above problems,this paper has done a lot of work in combination with the hydraulic and thermal coupling model of the steam pipe network and Flowmaster software simulation to improve the data quality of the field measurement data;the steam pipe network has been developed based on the IAPWS-IF97 water and steam thermodynamic property equations The automatic flow calculation software provides technical support for the design and renovation of the steam pipe network and the study of the management system;a method for applying the LSTM prediction algorithm to the fault detection of the steam pipe network is proposed.The main contents of this article include:(1)Aiming at the problem of poor data consistency in the steam pipe network,the pipe network simulation effects of the calculation model and simulation software are studied separately.The hydraulic and thermal calculation model of the steam pipe network is studied,and the simulation results of the steam pipe network are obtained by using the simulation software Flowmaster.The actual measurement data are verified and compared.The results prove that the calculation model and the simulation software have their own advantages.The network model is more reasonable.(2)Aiming at the problem of incomplete data of the steam pipeline network,an automatic calculation software for instantaneous flow of the pipeline network is designed.Improve the calculation accuracy of the steam pipe network to a maximum relative error of3%.(3)Aiming at the problem of low accuracy of online data monitoring of the steam pipe network system,the idea of applying the LSTM prediction algorithm to the fault detection of the steam pipe network is proposed.Various typical data-driven fault detection methods such as principal component analysis(PCA),support vector machine(SVM),and neural network are discussed.The method of applying the LSTM prediction algorithm to the fault detection of steam network is proposed independently.(4)According to the heating steam pipe network engineering example of Xiasha Industrial Park of Huadian Group,the design algorithm comparison experiment verifies the engineering application value of the LSTM prediction algorithm in the direction of steam pipe network fault detection.The accuracy rate of fault detection reaches 92%,compared with The accuracy of traditional fault detection algorithms is higher.In this paper,through the study of steam hydraulic thermal calculation model and steam pipe network simulation,and the development of software for the calculation of thermodynamic parameters of water and steam,the quality of on-site measurement data is improved.The mature time series prediction algorithm LSTM is applied to the precise and effective monitoring of the operation status of the steam pipe network in the industrial park,which effectively improves the operational safety factor and energy saving efficiency of the enterprise steam pipe network,and real-time monitoring and optimization of the steam pipe network Management is important.