Design And Implementation Of Ultrasonic-based Detection System For Shell Internal Component

In recent years, a large number of unknown ammunition has been found in the process of construction operations and mining, and most of them are corrosive and rusty. It is difficult to judge specific types of ammunition according to the external characteristics, so it caused great difficulties for further destruction. Before destruct the waste shells, accurate identification and characterization is to ensure the process of packaging, transportation, handling and destruction, as it not only affects the efficiency of operations and brings serious security risk in the process of judging by experience and identification only. Therefore, army urgently needs a nondestructive testing system for shell internal component to adapt requirements for the development of the destruction of ammunition.This thesis researches the various waste shell body, investigates the detection principle of the internal components of the ultrasonic shell body, analyzes the relationship between ultrasonic characteristics and internal components, develops and designs internal ultrasonic shells element detection system according the above principle to realize signal transmission, reception, amplification, noise processing, data acquisition, echo graphics display, components judgment, test results storage, playback and output functions in the process of exploration.According to the requirement of internal components of ultrasonic shell detection on the system structure and function, this system mainly consists of ARM control board, acoustic collection plate, true color LCD screen, data processing and analysis program, acoustic transducer, shell detection platform of six parts component. The project completes overall design and debugging based on the hardware platform of ARM920T high-performance 32-bit embedded processor (EP9315). At the same time, this thesis uses data fusion technology to detect the internal components of the shell for the first time, gets a large number of ultrasonic data of fillers in connection with ultrasonic detector, the use of data fusion method for data processing and feature extraction, and the first wave, amplitude, frequency, gain and other characteristics of integration., about data fusion, taking into account the BP neural network data fusion algorithm optimization of accuracy, but easily leads the shortcomings of local minima and slow convergence. However, GA has a strong macro-search capabilities, can avoid local minima. This project will propose a GA-BP algorithm with the integration of their respective advantages, to provide comprehensive and detailed information for accurate and effective judgment of internal components, and will ultimately achieve the target of identifying shell body's internal fillers.