Using the Internet of Things to detect asset failures before they occur

Senior Product Manager, Internet of Things, IBM

A few years ago, my air conditioner sprang a leak, and the first I knew of it was when my ceiling started to buckle and leak. Of course I discovered the leak on a weekend, so I had to pay for emergency service on top of a costly ceiling repair and repainting. After that, I figured I’d better sign up for a yearly service plan. For the next two years, everything was fine except for the money I shelled out for the inspections.

This year, during the yearly inspection and service, the service technician noticed a capacitor and coil that, based on their condition, needed fixing. He also recommended that I have a sensor installed to detect the presence of water in the drip pan.

All this got me thinking—what if the key elements of my air conditioner included sensors, allowing me to skip the cost of yearly service while at the same time preventing a catastrophic failure and attendant ceiling cave-in?

This is now possible in many industries thanks to reductions in the cost of sensors and advances in connectivity. Take a consumer example. The Internet of Things is enabling cost-efficient implementation of condition-based maintenance for complex assets such as electric transformers or wheel bearings on rolling stock, for which maintenance costs are high and whose failure can create significant safety, environmental or economic risks. David Berger of Western Management Consultants gives a good overview in The value of integrating health, safety and environment processes with enterprise asset management.

For those managing high-value assets and leveraging an enterprise asset management system such as IBM Maximo, taking advantage of real-time equipment condition is an important element in driving efficiency, maintaining safety and improving asset reliability. Many such assets may already be instrumented and controlled by proprietary supervisory control and data acquisition (SCADA) systems. Although this data is not always available to those who are doing preventive maintenance, open standards allow connection to an analytics platform, such as IBM Internet of Things Foundation, that can then predict failures or recommend maintenance based on asset condition.

Three steps can help in leveraging IoT and analytics technologies for improving asset performance:

  1. Collect sensor data to assess asset condition, defining thresholds or rules to initiate actions or notification to enable condition-based maintenance.
  2. Analyze historical sensor data, as well as asset failure and work order history, to uncover new patterns that can aid predictions of asset failure.
  3. Leverage asset condition using analysis tools to assess the economic, safety, environmental and public relations effects of failures while also analyzing alternative strategies for handling assets (repair, replace, load shifting, run to fail and so forth).

Getting to that third step may be the ultimate goal, but many organizations can derive significant benefit from taking even a first step toward leveraging IoT for condition-based maintenance. Next time, I’d like my air conditioner service company to call me a few weeks before my system needs repair. No leaky ceilings, and no cost for yearly inspections!