Benefit:
Catastrophic failures can be prevented that may lead to serious/irrepairable equipment damage. Multi-million dollar business productivity loss can be avoided Environmental pollution that could result as of equipment malfunction/damage can be averted or significantly reduced. In certain cases human lives can be saved from accidents due to explosions or contact with dangerous components. Systematic analysis and early prediction of failure leads to effective management of the asset and hence the risk pertaining to insurance and finance of operation can also be accordingly well managed.
Solution:
Various machine learning techniques can be used along with physics models to predict failures. Especially, Deep Learning is a specialized area of Machine Learning that shows promise in detecting anomalies earlier ahead of a forced outage event thereby preventing several mechanical and financial damage to the business. Deep Learning techniques rely on complex networks of Neural networks that can uncover hidden relationship between parameters that significantly contribute to anomalies. These techniques can be developed generically such that they can be customized to target multiple failure modes with minimal effort along with domain expert input.
