Olivier Jouve is the Director of Predictive Analytics at IBM.
The big trend to help everyone “get fit this year” are devices that monitor workouts and the body’s response to environmental and physical conditions. Way more fun and (let’s hope) more effective than just an occasional finger on the pulse.
It seems that humans are borrowing a technique pioneered on mechanical equipment.
The use of analytics is keeping dailytabs on equipment, parts, usage and wear so organizations can make optimized, informed decisions and predict when to repair or replace a power line, a water pipe or a catalytic converter before they break down.
Organizations across the globe have been placing much more importance on the notion of “predictive maintenance,” which uses predictive analytics to capture and analyze data to proactively detect failure patterns or other issues to avoid downtime.
Predictive maintenance enables organizations to be proactive, rather than reactive, with their maintenance of assets. Though, what is interesting is that the definition of predictive maintenance has evolved in the last few years.
Traditionally, predictive maintenance meant just that, or predicting when maintenance would be needed on an asset, such as when a production line machine on the manufacturing floor would fail.
Today, it has moved from beyond the plant to include any physical asset, such as buildings, windmills, oil drills, cars, airplanes, and much more.
Having the insights to understanding when a part or machine is going to fail enables an organization to better plan their production schedule. This helps optimize inventory and logistics decisions, and ultimately an organization’s customers receive the right parts at the right time and increasing customer satisfaction and loyalty.
By improving maintenance, other areas of the business benefit as well.
Here are few examples:
Every year a heavy equipment manufacturer in the mining industry needs to strike a balance of keeping enough parts in the field so the equipment can be fixed immediately. The company gathers and analyzes data regarding machine usage and environmental conditions to determine which parts are likely to fail and inform their supply chain ahead of time of the parts that need be ordered throughout the year. Additionally, they also gather information about individual machine operators to understand usage so they could provide better training to the employees that had increased failures.
For gas stations, an “out-of-service” gas pump can cost hundreds of thousands a day and paralyze the business. A large company utilized predictive maintenance to predict when gas pumps would fail by taking into consideration flow velocity, nozzle pressure, number of remaining gallons, calibration measurements and other information. They are now able to ensure gas continuity to their customers, keeping their top line flowing.
Automotive manufacturers are using predictive maintenance to improve quality control on assembly lines. Initially, raw materials had to be molded or cast into specific shapes and certain casting processes can often take eight hours or more. One manufacturer was seeing parts that failed the quality checkafterthe process leading to loss of inventory and time. Using predictive maintenance, the car maker could predict which parts were likely to fail before this eight-hour process. Subsequently, those parts that were likely to fail were pulled from the process and either scrapped or fixed, saving millions of dollars.
Organizations across all industries are learning that analytics can predict a problem and help prevent it from becoming an even bigger one. After all, the simplest and most efficient way to fix a problem is to make sure the problem never happens in the first place.