Software agents monitor machinery health

Edited by the Engineeringtalk editorial team Mar 3, 2005

Dexter Agents Version 3.3 is a software tool that allows companies to build and deploy software agents for machinery health monitoring and predictive maintenance.

Dexter Agents Version 3.3 is a software tool that allows companies to build and deploy software agents for machinery health monitoring and predictive maintenance.

Software agents automate the bulk of the time-consuming and complex data processing tasks typically associated with condition-based maintenance.

The agents acquire real-time machinery sensor data, monitor alarms, detect trends, diagnose equipment faults, and predict impending problems.

Prognostic agents are designed to predict equipment problems at their earliest stage of development.

These agents automatically scan historical machinery data to discover degrading performance trends and then use probabilistic neural network reasoning to predict machinery faults associated with those trends.

The agents can generally isolate a fault much faster than a human operator, who may be inundated with large amounts of process data to analyse.

Prognostic agents estimate remaining useful life, or equivalently time to failure, for each predicted equipment fault.

These estimates are based on the time available to run the equipment until the initial predicted alarm condition is expected to occur.

By reporting the earliest predicted alarm, the agent provides the operator with the maximum amount of warning time, such that corrective action can be taken prior to actual equipment failure.

Dexter Agents allows companies to rapidly deploy software agents without costly software programming.

Software agents keep a continuous watch over equipment and process health, detect abnormal conditions, pinpoint developing problems, and alert operators before process interruption can occur.

Dexter Agents can monitor sensor data in virtually any machinery application.

Improved reliability, machinery uptime, and production profitability are the goals of every manufacturer, no matter how large or small the operation.

The prediction of future equipment faults and remaining useful life can help maintenance staff determine when maintenance should be carried out.

By predicting machinery problems before they occur, unexpected breakdowns can be avoided, along with their attendant costs.

In the absence of significant degrading trends, decisions to extend equipment overhaul periods can be confidently made, backed by empirical data generated by the prognostic agents.

This type of predictive maintenance information leads to improved maintenance planning, cost reductions, and overall plant reliability.

Maintenance and repair decisions can now be tied to actual equipment operating conditions, versus the traditional time-based manufacturer recommendations.

Dexter Agents easily integrates with most process automation and data historian software on the market.

This means that existing investments in networks, process automation, and asset management systems are further leveraged with the addition of Dexter Agents.

Attaching intelligent software agents to monitor and assess process data will add significant value to the plant IT infrastructure that is already in place.