An intelligent sensor network system coupled with statistical process model for predicting machinery health and failure

Abstract

Application of statistical process for the purpose of health monitoring of machinery and system is the main purpose of this paper. The system has two main parts. The first, sensing and transmitting of variable by a network of intelligent sensors and transducers. The second, analysis and prognosis of machinery health by a two-step statistical process model (SPM). Most contemporary intelligent sensors have signal-conditioning circuit integrated within the sensor as a single chip device. In many instances sensor and signal conditioning circuit are packaged together as one unit. However, intelligent sensors are becoming increasingly important for many critical applications. An intelligent sensor network system can perform sensing and transmitting of variables for the process controller and can also locally store and transmit sensed variable over wireless link to remote central processing computer. The central processing computer is used for continuous statistical analysis of multiple variables and can be used for predicting machinery failure. Operators at remote location can review processed information at any instant of time and determine imminent and prospective condition of the machinery. Currently we are researching to develop an intelligent sensor network system that will sense and temporarily store process variables and periodically transmit them to a remote location for processing, analyzing, and storing. The process variables are transmitted to the central computer as well as to the controller for continuous control of the process. The statistical process model (SPM) located in the central computer system will analyze the related data for predicting machinery failure and prognostic maintenance.