Research On Unattended Heat Exchange Station’s Monitoring Technology Based On The Fault Tree Model

With the rapid development of urbanization, the urban heating demand is growing rapidly. Under the backgrounds of the government advocating ecological development and civilization construction, relevant technology about the energy saving and emission reduction is received extensive attention of the personage inside course of study. At present, some domestic urban heating system still remain extensive form, such as bad operation environment, excessive energy consumption, poor heating quality etc. All of the problems need to be solved. So design the automatic monitoring system of the urban central heating process to achieve energy conservation, emission reduction, and civilized production has important engineering application value and profound social significance. Heat exchange station is the key process of the urban heating system. This paper analyze the characteristic of the industrial process of the heat exchange station, design a kind of intelligent parameter analysis method based on the fault tree model combine with the BP neural network. Use this method to realize the intelligent parameter analysis so as to realize civilized, economical production and improve the heating quality.The heat exchange station parameters has many characteristics such as multiple input, disturbance, time varying, big lag and tight coupling etc. Firstly use the fault tree analysis, according to this method to establish the heat exchange station’s fault tree model, decide the top event, bottom event and the intermediate node event. Then construct the structure function, abstract the minimum cut sets, and then carry out the digital logic process to extract the heat exchange station’s fault tree model, get the expected vector matrix Y. This matrix contains all the parameter fault logical configuration message of the heat exchange station.In order to realize the match between the actual acquisition parameters and the vector matrix Y, take BP neural network as the training method. Take a large number of parameters as the input samples of the BP neural network. Take the vector matrix Y as the output of the BP neural network. After offline training get the template as the parameter analysis basis of the heat exchange station. The template contains all the relationship between the actual parameters and the expected vector matrix. And then online analysis the parameters of the heat exchange station follow this template. Through parameter analysis, charge the heat exchange station’s running status, carry out parameter control to realize the unattended monitor.This article finally proposed the MATLAB simulation test to the data analysis method based on the fault tree model combined with the BP neural network. Through offline training and test experiments to verify the feasibility of this scheme. The analysis scheme can instantly meet the training standards at limited time. The training effect is very obvious, the method has good repeatability. Use the well trained template to carry out testing experiment, it can clearly reflect the problems of the heat exchange station’s parameters. Using this scheme to the unattended heat exchange station monitoring system can get good results.This paper uses the ARM embedded controller to design the heat exchange station’s data collection module. Complete the whole module scheme of the data collection plan to lay the foundation of data analysis.