Distributed On-Line System for Process Plant Monitoring

Abstract

This paper describes a methodology for monitoring industrial processes and plant that can be implemented cost-effectively within small-to-medium enterprises. The methodology is based on a network of 8-bit microcontrollers that communicate with each other on a controller area network bus. Ethernet connectivity is provided so that remote users can access the system on the internet. The software models developed for data acquisition nodes and the design of remote user interfaces and supervisory nodes are also explained. The system is aimed at providing specific maintenance guidance and fault identification, rather than gathering data for off-line analysis. Overly complicated processing is avoided to make real-time implementation possible, using 8-bit microcontrollers. The methodology emphasizes the use of process controller signals for fault detection and sensor signals for fault isolation. The suitability of the methodology is explored by acquiring signals from a laboratory-based process rig. Suitable monitoring techniques for the system in time and frequency domains are also discussed.

A Discussion of the Prognostics and Health Management Aspects of Embedded Condition Monitoring System

“…Figure 3 shows the hardware developed, in this case for a (Ahsan et al, 2006). The 4 vertical circuit boards, each with a PIC microcontroller, are connected to analog signals measuring flowrate, temperature, liquid level and pump power, and are the horizontal circuit board includes the connectivity module and CAN bus communications (brought physically close together in this implementation).…”

Use of parameter estimation for the detection and diagnosis of faults on electric train door systems

“…Given that any solution needs to address the practicalities of implementation and also economic feasibility, it would seem that fault diagnosis should be carried out in a distributed decentralized architecture that incorporates fault-tolerant designs, onboard train communication methods, powerful fault diagnosis tools as well as maintenance interfaces [4][5][6][7]. This paper demonstrates the necessary extraction of the physical parameters of a particular door using a continuous-time parameter estimation (PE) method with simple design criteria and low computational demands.…”