(664c) Overcoming a Lack of Data in Decision Analysis

To remain competitive in the faces of challenging market forces, process operations continually improve and adapt. This includes optimizing the scheduling of turnarounds and equipment overhauls [1]. Even with the rise of Big Data in the process industries [2], there are still frequent situations where the required data is limited or in a form that makes the decision analysis a challenge. However, by treating uncertainty as a state of knowledge we can reframe our problem and develop a credible comparison methodology. The reframing allows us to overcome the lack of data and access the power of established tools such as lifetime analysis, physics of failure and simulation. Using these tools we can then quantify the differences and economic impact of different overhaul schedules and policies. This talk will present a case study developed from work on turbomachinery.

[1] Matteo Biondi, Guido Sand, Iiro Harjunkoski, Optimization of multipurpose process plant operations: A multi-time-scale maintenance and production scheduling approach, Computers & Chemical Engineering, Volume 99, 6 April 2017, Pages 325-339, ISSN 0098-1354, http://doi.org/10.1016/j.compchemeng.2017.01.007.

(http://www.sciencedirect.com/science/article/pii/S009813541730008X)

[2] J.F. MacGregor, M.J. Bruwer, I. Miletic, M. Cardin, Z. Liu, Latent Variable Models and Big Data in the Process Industries, IFAC-PapersOnLine, Volume 48, Issue 8, 2015, Pages 520-524, ISSN 2405-8963, http://dx.doi.org/10.1016/j.ifacol.2015.09.020.

(http://www.sciencedirect.com/science/article/pii/S2405896315011015)