SmartSignal is detecting in wider applications
Lisle, IL—SmartSignal's equipment monitoring and predictive capabilities are extending beyond airline and large utility turbines to locomotives and pipeline meters, a company representative told Control Engineering recently.
Lisle, IL151; SmartSignal 's equipment monitoring and predictive capabilities are extending beyond airline and large utility turbines to locomotives and pipeline meters, a company representative told Control Engineering recently.
The technology, advanced Equipment Condition Monitoring (eCM), is covered by 34 patents, and analyzes sensor input over time for deviations that anticipate equipment failure, says John Kerastas, SmartSignal's marketing communications director. 'We show people how the technology looks for patterns and deviations in sensor data sets across multiple variables,' either from their live data over time, or from a database or historian. 'The fun part is that when it works, you see the light go on in peoples' eyes.' The latest version of Smart Signal's software is eCM 2.5.
Still, convincing prospective customers sometimes takes time. 'At Delta, it took us nearly two years. When times get tough, people are often more willing to rethink how they do things.' Common thinking is: when it's broke, then we'll fix it. However, with assets that break in inconvenient locations or times, knowing when they're going to break is key to reducing costs, improving safety and reliability, and extending asset life.
SmartSignal's literature states that, 'In a week, one of your critical assets will suddenly shut down. Wouldn't you like to know that right now?' Mr. Kerastas explains by presenting some application examples:
Airplanes are least expensive to maintain when they're at key locations, and certainly no one wants an in-flight failure. A partial-fleet contract with Delta (Atlanta, GA) recently expanded to the full fleet after SmartSignal identified incipient faults in jet engines by analyzing real-time sensor data not detected by traditional systems.
Peaker gas turbines should be repaired during non-peak times. Also, in the future it may be necessary to prove to regulators the necessity of taking generating assets off-line to avoid any possible impression of market tampering. Many utility turbines serve in a peaking capacity now, cycling many times more than the original design anticipated, creating unplanned, unpredicted wear that can confound traditional maintenance schedules and off-the-shelf monitoring systems.
Railroads often run extra locomotives on trains and perform major overhauls "just in case." General Motors' Electro-Motive Division (LaGrange, IL) is using SmartSignal to detect pending equipment problems in locomotives using data from sensors that measure pressure, temperature, voltage, current, RPMs and locomotive speed. Analysis even detected an incorrect wheel size. This provides up to 20 days advanced warning, which has enabled more efficient planning of repairs and operations.
Failures related to environmental requirements pre- and post-combustion, such as pulverizers and electrostatic precipitators, require careful attention and documentation to ensure high asset availability.
Large compressors on a pipeline are more easily fixed on schedule, between material transfers, instead of during peak use.
On one of Chevron's (San Ramon, CA) pipelines, SmartSignal identifies faulty meters, so customers are not under- or over-charged. Pinpointing problems quickly reduces maintenance costs.
SmartSignal's focus up to now has been on capacity-constrained industries, which have assets that can't be readily stored.
Mr. Kerastas adds that models can be tuned more quickly and provide more advance notice, if asset experts can associate behavior or data deviations with known problems. This allows experienced staff to create an expert system, while they're still around. Then the software can learn what "Larry" knows before he leaves, while third-shift "Fred" can foresee what's happening before the process is dead.
Control Engineering Daily News Desk
Mark T. Hoske, editor-in-chief