Study On The Applications Of Information Fusion Technology

The concept of information fusion was brought forward in early 1970's. Information fusion techniques study how to process information from multisources and make the information with different form to supplement from each other, thus more impersonal and substantial cognition to a thing or an object can be obtained. Information fusion techniques was studied and applied in military field at first, and great success was obtained. For example, C3I (that is command, communication, control and information) system. Since 1980's, its application has gradually extended to civil fields such as robot and intelligent instrument system, image analysis and understanding, automatic target recognition, multi-source image integration. Presently, the information fusion techniques have become a very active research field.The thesis arms to study application of information fusion in the power system, that is to use information fusion to study criterion of contaminated degree of insulator, judgment of insulator's flashover is a new method by monitoring ultrasonic signal as polluted insulator discharges, there is a large amount of information included in the ultrasonic signal, it is obvious that exact criterion couldn't be obtained, if single information is analyzed, But mutual influence and action are not taken into account, therefore, correctness and reliability of criterion should be improved greatly by using information fusion to handle ultrasonic signal of polluted insulator discharge.First of all, The theoretical foundation, current research situation and developing direction of the fusion technology of information are introduced in this paper. At the same time, the application problem and relevant algorithm in the field of target recognition are discussed. Then the contaminant discharge mechanism of insulators is studied, the discharging course and the following acoustic emission phenomenon are analyzed. Afterwards, acoustic signal data are acquired by testing insulators of different contamination degree under different voltage grade. Statistical analysis of test data indicates that the criticality of contaminant insulator has close relationship with acoustic signal. According to the characteristic of the ultrasonic wave signal transmitted while the contaminant insulator is running, characteristic that figure the acoustical signal are confirmed here. Mathematic illation models that reflect the relationship between acoustic signal and criticality are also established by 4 kinds of methods, such as forward neural networks, fuzzy inference, ANFIS and D-S theory. Finally, the test data of acousticalsignal is calculated and simulated based on the mathematics models. The result proved that the mathematic models has high precision and the fusion technology of information can be used in measurement of polluted insulator by acoustic emission technology successfully.