Equipment Health Management System Based On Internet Of Things

During the using process of mechanical equipments and systems, fault diagnosis is inevitable. The traditional fault diagnosis technology is mainly limited to fault detect and diagnosis. In addition, the fault detecting methods only focuses on certain equipment, which makes it not so systematic. The internet technology applied in traditional methods is confined to data transmission, unable to achieve real-time data collecting on site. The detecting systems of equipments are usually established software. This leads to a problem that a great amount of duplicated codes will occur when the system is to be updated.According to the drawbacks and limitation existed in the current fault diagnosis system, equipment health management system has been studied in four aspects in this paper. The system scheme has been established based on the model of the internet of things, and the networking of sensor networks has been analyzed and studied from the perspective of theory. With respect to data transmission, the transmission schemes based on networking protocol and on database access have been designed. After testing and comparing the two schemes, the latter scheme has been adopted and applied in the system software. In terms of system software, the software structure based on "plug-in/platform" mode has been adopted; and the plug-in framework has been designed through programming. In this way, it’s convenient for the system to test kinds of equipment and to reduce the workload on secondary development of software, making the system more flexible in software in a bid to facilitate its updating and maintenance. As to the fault diagnosis and prediction, the algorithm of artificial neural network has been adopted on the basis of careful study on relevant algorithm for fault detecting and predicting, and comparison of principles, realization conditions and feasibility analysis between expert system and artificial neural network. Combined with the system and equipment to be tested in the system, the programming has achieved the fault testing algorithm based on SOM neural network and fault predicting algorithm based on Elman neural network. Also function checking and testing have been conducted for the system by setting up simulated testing environment in the laboratory.The prototype of equipment health management system based on Internet of things technology has been completely designed in this paper. It collects real-time data via sensor network, transmits the data detected to the master computer of the management send via internet in real-time, and supervises and controls the running status of equipment at the meantime. Besides, the system is able to predict the fault that might occur in the equipment based on the current running status and detecting data of the equipment, so as to maintain the equipment before any fault occurs. The "plug-in/platform" software structure is adopted in the system to load equipment management scheme to the system in the form of plug-ins. In this way, the system can achieve supporting multiple equipment management as well as upgrading and updating the system conveniently, so as to avoid heavy duplicated work.