基于RBF网络的故障预测策略研究-控制理论与控制工程专业论文.docxVIP

II II Abstract Owing to the high temperature and high humidity in the section region of the continuous caster, the electrical wire skin of the electromagnetic stirring system is easy to be eroded. Consequently, earthing faults occur frequently in one point or more, causing the unevenness of the molten steel stirring, and resulting in direct or indirect economical loss. Bearing in mind of the above-mentioned reasons, the on-line electrical detection and fault prediction system of the electromagnetic stirring is in urgent requirement and thus developed in this dissertation. Hall voltage, current, leaky current sensors combined with their corresponding conditioning modules and information transmission modules are designed and applied in this system to online-detect the electrical signals and thereby supervise the operational status of the whole electromagnetic stirring equipment. In this way, the security status of the system is under effective monitoring. More significantly, the sufficient operational electrical information is sampled and processed for earthing fault prediction. According to the needs of application, an overall planning on the system is carried out. This dissertation describes the LCD module and the information transmission module. The LCD module, developed by using the SED1335, can provide immediate, accurate, high resolution display. The information transmission module which is developed by using the nRF905 and MSP430 has many merits such as a small size, low power consumption, reliable and steady data transmission and high resist interference performance. The module can serve well in short-range wireless data transmission areas. This dissertation also discusses the RS232/RS485 and Digital Storage Oscilloscope. At last it introduced the RBF neural network ,which is used to simulate and research the leakage current dynamic process. In this way, the predictive-maintenance manner of the Earthing Faults is achieved, and the continuity and security of th