Coming: Better and Cheaper Condition Monitoring

Share

As president of an engineering firm that offers predictive maintenance services, Bill Marscher is a staunch proponent of condition monitoring.

But he is also quick to point out its shortcomings – its inability to precisely diagnose problems or to predict when a machine will fail. Plus, he says, thorough condition monitoring systems cost a lot.

Yet hope looms for plant and maintenance supervisors who want more from oil monitoring and vibration analysis. In a preview of an address that he will make next month at a conference sponsored by the Society of Tribologists and Lubrication Engineers, Marscher told Lube Report that technological advances will soon enable condition monitoring practitioners to pinpoint failing components and predict just how long they will last.

Moreover, he said, the costs to obtain this information will be practically negligible.

Condition monitoring encompasses a variety of methods used to detect mechanical problems early in order to avoid expensive breakdowns. Lubricating oil is analyzed for wear metal particles, contaminants and chemical breakdown. Vibration, thermal and electromagnetic patterns are analyzed for telling irregularities.

Advocates of these practices have documented millions upon millions of dollars saved by avoidance of unplanned downtime and by fixing problems before they become serious.
Marscher, president of the Parsippany, N.J., firm Mechanical Solutions Inc., is scheduled to deliver a keynote address at STLEs Condition Monitoring 2002 Conference and Exhibition, being held in Dallas, Tex., Feb. 18-20. A past-president of STLE, he is among these advocates of condition monitoring but hardly boosterish.

Theres no question that there is a cost benefit to applying condition monitoring, he said. But its got a long way to go. Plant operators who learn of a problem may not know how quickly they need to schedule a repair. And without knowing exactly which part is failing, they usually arm themselves with more replacement parts than they truly need.

Oil monitoring is very good at picking up issues that would signal overloaded bearings – things like increased oxidation or a high level of wear particles. But its not good at detecting problems inside a machine. For example, if there is cavitation occurring, that may not damage the bearings but it may be eating up a pump like nobodys business.

Other methods have their own weaknesses, he continued. As an example, he cited a motor used to drive a fan that ventilates a plant. Vibration analysis is often used to determine whether such machines are properly aligned, with elevated or abnormal patterns indicating they are not.

If these machines are misaligned, it can overload the bearings and cause damage, he said. But its possible for them to be misaligned in a way that limits vibration, and then vibration analysis wont work.

Cost is also a major obstacle for companies trying to establish effective condition monitoring programs. Plant operators can improve oil monitoring by installing probes to analyze lubricant continuously at a variety of points.

A lot of times people arent getting specific information because they need a lot more probes, Marscher said. But when you look at the cost for a probe, plus the wiring to connect it, its $15,000 to $20,000 per probe. For something like an oil turbine, that can add up to several hundred-thousand dollars.

The good news, Marscher said, is that these expenses will soon disappear almost entirely. A decade of research has led to development of small sensors that monitor oil online and transmit their information by radio, without wires. Their measurement of oil viscosity still needs work, Marscher said, but within two or three years, they will provide all of the analysis of a professional laboratory at a cost of $5 to $10 each.

Similar advances have already been made in software that processes the information gathered by condition monitoring. These improvements are partly due to enormous online data bases gathering information from machines operating all over the world.

The problem has been that we have systems that detect a problem and then they turn on a light to tell you youve got a problem, Marscher said. But the people in the plant dont know exactly what the problem is or they dont know how urgent it is that they fix it. The new software processes information in a much more sophisticated way. Its much more prognostic.

That kind of information, Marscher said, will spell good news for plant operators everywhere.

It will make our country and the whole world able to operate more efficiently and more safely.

Related Topics

Market Topics