| With the application of Wireless Sensor Networks (WSN) in the Gas Pipelines Monitoring, Real-time monitoring and accurate localization can be achieved, and problems like low efficiency and location accuracy by artificial patrol can be solved. However, with the influence of the number of nodes, different monitoring place and types of sensor, the information of leakage on the net turns out variety and mass, and it is important to make sure that leakage information is processed correctly and reliable. Therefore, multiple information fusion technology is used and a new data fusion method based on Neural Networks for leakage information is proposed. The main contents and conclusions are as follows,1. The reason of leakages, the present research state and the main monitoring ways of pipelines were discussed; the theory and structural characteristics of WSN Monitoring System were emphasized and the main ways of information fusion were introduced.2. Main principium of Neural Networks and Ant Colony Optimization (ACO) were researched and a new data fusion system for leakage information was established based on NN. The results of simulation experiment shows that the data fusion system based on NN is effective for the monitoring data fusion and leakage diagnosis, but there were defects of slow rate of convergence, low precision of diagnosis and tending to converge to local optimum.3. In accordance with the problems from BP Networks, an improved fusion system based on ACO was present. The basic theory and algorithm process were particularly discussed and the program design and numerical simulation experiments were accomplished with MATLAB. At last, the performance test of the improved system was finished with the same sample data as mentioned before.The results of experiments show that the improved data fusion system can greatly increase the rate of convergence and accuracy of diagnosis without trapped in local optimum, and it was more useful for the leakage diagnosis of gas pipelines. |
PDF Full Text Request