Issue link: https://industrialengineer.epubxp.com/i/826827
52 ISE Magazine | www.iise.org/ISEmagazine Inside IISE Journals research Predicting the failure time of equipment using big sensor data Many capital-intensive assets, such as gas turbines and generators, are instru- mented with hundreds of sensors that monitor their performance and state of health. The key objective is to maxi- mize the asset availability by preventing unscheduled shutdowns and failures. To achieve this, inherent trends in the sensor data are extracted and modeled to predict the asset's remaining lifetime through a process known as prognos- tics. Advances in equipment prognostics always have been restricted by the avail- ability of the data that monitors its con- dition, which many companies consider highly proprietary. As a result, existing research has focused mostly on enhanc- ing the prediction accuracy under lim- ited data availability. Research on scal- ability and computational efficiency in the face of big data generally has been circumvented. Today, the volume and dimensional- ity of the data being generated by in- dustrial assets are very large, and many industries are becoming overwhelmed with the amount of data that needs to be processed in real time. For example, a gas turbine is equipped with more than 2,000 sensors that monitor vibra- tion, temperature, pressure, etc. Opti- cal sensors used for in-situ turbine blade crack detection generate 600 gigabytes per day – almost seven times Twitter's daily volume. In the paper "Scalable Prognos- tic Models for Large-Scale Condition Monitoring Applications," doctoral This month we highlight two articles from IISE Transactions. The first article addresses how to maximize asset availability and avoid unscheduled shutdowns and failures by proposing a prognostic modeling functional data analytic framework with algorithms in a big data environment. The solution presented has the potential to make a great impact in industries that rely on the continuous availability of capital-intensive assets like the energy and power generation sector. The second article proposes a novel statistical model for the unique data of nurse care coordination, along with further use of the model to examine and reveal how multiple aspects of nurses' care coordination activities are related to their demographic and workload variables at the individual/nurse level as well as the structure, policy and climate of the nurses' practice environment at the unit level. The authors demonstrated the developed method with real data sets to improve staff-nurse care coordination within hospital units to improve patient quality of life and save costs in the healthcare system. These articles will appear in the July 2017 issue (Volume 49, No. 7). Nagi Gebraeel (from left), Xiaolei Fang and Kamran Paynabar pose in the Analytics and Prognostic Systems Laboratory at Georgia Tech.