Plant-floor machines have so much to say, if people would just listen. Trapped inside every machine is a wealth of information that can tell maintenance technicians whether small hiccups could eventually develop into larger problems. Tapping into such data is becoming easier with predictive maintenance tools.
By utilizing predictive technologies, machine maintenance is evolving from prescheduled routine tasks to the use of more precise indicators that increase maintenance efficiency. Scheduled maintenance is an effective way to reduce machine downtime and prevent costly repairs, but it doesn't take into account the ever-changing plant-floor environment, notes Jun Ni, a professor of mechanical engineering at the University of Michigan and co-director of the Center for Intelligent Maintenance Systems.
"Some forward-looking companies are starting to look at predictive tools and asking how they can get ahead because if you use CMMS (computerized maintenance management systems) to help you schedule maintenance, you do not consider the dynamics because the manufacturing shop floor changes from shift to shift, day to day, week to week," he says.
The Center for Intelligent Maintenance Systems, based at the University of Cincinnati, is focused on developing technologies that use embedded and remote monitoring and intelligent decision support tools to identify and prevent maintenance issues before they occur. One application the Center has developed is the Watchdog Agent, which analyzes equipment performance based on data gathered from sensors mounted on the machines.
The sensors extract data from such machine variables as temperature, vibration and motor current. With this information, manufacturers gain insight into the rate of machine degradation and historical comparison data that can be used to predict future issues. "We extract features from seemingly random information so we can give people information on the amount of degradation compared with a normal known condition, and also based on the rate of degradation, we can predict what remains in useful life," Ni says.
The wireless sensor system also allows technicians to continuously monitor vibration data as opposed to older-style wired configurations that only provided periodic reports, says Daniel Szoch, sales and marketing manager for the Timken Co.'s Reliability Solutions division.
Another predictive technology that's becoming more standard in many industries is thermography, notes Vlad Bacalu, a product manager with maintenance outsourcing firm Advanced Technology Services Inc. Utilizing an infrared camera, thermography, or thermal imaging, creates a visual picture of temperature that helps manufacturers identify hot spots, which can cause high electrical resistance or excess friction. Ultrasound measurements also can be used to identify sources of wasted energy such as air, gas, steam and vacuum leaks.
Each solution provides data that can head off potential maintenance disasters, but they're more effective when included as part of a comprehensive predictive maintenance program. One mistake many manufacturers make when deploying a predictive maintenance program is implementing technology without knowing how to properly operate the instruments or interpret the information, Timken's Szoch says.
Timken introduced a turnkey program for predictive maintenance in late May aimed at helping manufacturers establish their own programs. Timken, a provider of bearings and other steel products, helps manufacturers with the program design, startup, training and hardware selection. The consultancy service starts with an audit of each company's current program, including the types of machinery it's using, maintenance history and expertise level. The second phase is the program design, during which Timken helps the company decide what types of equipment will be needed, frequency of diagnostic readings and alarm thresholds that tell manufacturers when potential issues arise.
The actual predictive technologies implemented include vibration analysis, infrared thermography, oil analysis, ultrasound and motor circuit analysis. Vibration analysis is considered one of the most critical components because it provides information about rotating parts -- the equipment that requires the highest percentage of maintenance, Szoch says. "Rotating parts will give you the earliest indication on the equipment," he says.
The technology combined with trained, skilled maintenance technicians should reduce maintenance costs and unexpected failures. But Szoch says companies need corporate buy-in to make the programs fully effective. "Companies that are most successful have a high-level champion appreciating the value that can be derived from the program, and they seem to be an internal driver," he explains. "That's so important because it seems as though when you have one of those people, the program can move forward."
Szoch adds that program leaders need to document the success so they can validate the results to the rest of the organization and show the type of return on investment that's possible with predictive maintenance.