Predictive Analytics have found fertile grounds in various
sectors and applications over the past 2 decades; mainly in the financial and telecom
services sectors and for such
applications in marketing, as for example, for customer
churn prevention, fraud detection, cross & upselling campaigns, credit scoring,
etc.
Most recently, the new buzzwords and capabilities in the analytics world are the ones like Machine Learning and Artificial Intelligence.
Most recently, the new buzzwords and capabilities in the analytics world are the ones like Machine Learning and Artificial Intelligence.
In the manufacturing operations and maintenance, Predictive Analytics
have been applied sporadically for preventive maintenance and even in more rare
occasions for failure root cause analyses and online fault modeling for
operator guidance and decision support.
Nowadays, the EPC industry is set to pay attention to Predictive
Analytics as part of their #digitaltransformation and #EPC4.0 initiatives. In late
2018, a leading EPC, Fluor, announced an initiative to apply Predictive Analytics for project
performance, and this is great news. Especially for those who pay the bill,
that is, owner/operators of projects and #megaprojects.
Yet, you may wonder how this can be achieved. On one hand, history
of projects data is needed. The more
projects’ history available, the better, and as long as the data is or can be
normalized.
Secondly, such history should also contain extensive asset –
specific work steps history for all phases; if possible, from FEL 1, 2, 3
stages, down to the detailed engineering, procurement construction and pre-commissioning
phases. When such data is collected,
SEMMA (Sample, Explore, Modify, Model, Assess) methodologies can be applied to analyze
and find related predictors and dependent variables, as for example cost,
schedule and quality variations.
With the emergence of the above – mentioned initiatives (#digitaltransformation
and #EPC4), in the EPC domain, it will be interesting to watch if and when owner/operators
and EPCs take advantage of such possibilities. Megaproject performances over
the years confirm that insight into past performances can be the mother of
improvement.