Fault prediction system based on internet of things

What is IOT?

The internet of things incorporates sensor technology, communication network, internet technology, and smart computing technology and so on to accomplish sensor, reliable transmission and smart processing.

Common characteristics of IoT:

Comprehensive perception: That is, the collection of things in real time is accomplished dynamically by all types of means of perception available.

Reliable transmitting: That is, the perception information is reliably transmitted by all sorts of computer or network and the Internet

Intelligent processing: That is, smart computing such as cloud computing could be used to analyze huge data in order to achieve smart things control.

Fault prediction system based on internet of things:

Fault prediction scheme based on the internet of stuff develops on the three methods and can break through the initial framework of traditional surveillance and diagnosis of faults.

The combination of the internet of stuff and the prediction of faults can make complete use of more technology supports and generally share more data to achieve flexible distant prediction of faults, thus improving the effectiveness of work.

The system can create all linked, all communicating and exchanging information extensive condition surveillance, accurate transmission and smart fault forecast processing of the mechanical machinery groups ’ fault data:

Design of fault prediction with IoT:

The fault prediction scheme is intended to be an interconnecting network to predict the fault for important groups of mechanical machinery.

Its functional framework consists of four layers:

  • sensor monitoring layer,

  • middleware transmitting layer,

  • predictive application layer,

  • Decision feedback layer

Sensor-monitoring layer:

The sensor-monitoring layer is the basis of the system, and it is mainly for the monitoring of the equipment groups with the static and dynamic information of equipment collected.

The static data such as the name, type and material of the equipment is gathered through radio frequency identification (FRID) in this layer and stored directly in the labels.

The sensor monitoring layer can process up to 4 channels of main stage signals, 24 channels of vibration signals, 12 channels of static analogy signals, and up to 256 process variable information simultaneously.

Middleware-transmitting layer:

Middleware-transmitting layer is a significant connecting component, connecting the sensor-monitoring layer to the layer of prediction-application. The layer consists primarily of the Internet of Things middleware server and heterogeneous multi-protocol interfaces for transmitting multiple types of data.

In this layer the standardized information can be transmitted from the sensor-monitoring layer to the prediction application layer by means of wireless or GPRS.

Prediction-application layer

Prediction-application layer is the core application layer, which consists of remote expert teams, individual remote expert and data warehouse.

Different types of operation information about equipment organizations are intelligently processed in this layer. Intelligent computer interconnecting information fusion module is designed for smart processing of data and consists primarily of input interface, output interface and smart information processing part

Decision-feedback layer:

Decision-feedback layer is a prediction-application layer output module. Through this layer, reference strategies are given for EAM (Enterprise Device Management) inferred from the prediction-application layer.

The prediction information gained in the prediction-application layer can reveal the current mechanical equipment operating condition scientifically and effectively at any time and predict how long the future condition of the equipment groups will reach an unacceptable level and should be down for maintenance