DC-10: Eden Ngowi

Contact

E-mail: eden.ngowi@ntnu.no

Project

Learning-based health-aware operation and maintenance planning for improved safety

Host organization

NTNU

Supervisors

Prof. Prof. Johannes Jäschke (Main, NTUN); Prof. Alessandra Russo (co-supervisor, IMPERIAL)

Duration

36 months

Objectives

The main objective is to develop a framework for learning-based integrated optimization of long-term operation and maintenance decisions. Here knowledge-based models, e.g., mass and energy balances, and data-driven online learning of uncertain and unmodeled degradation phenomena will be combined to ensure safe and economical operation at all times.

The approach will use state estimation techniques to find parameter values of the first principle models, and advanced data-based techniques to determine the current equipment state and build a long-term prognostic model for planning operation and maintenance while guaranteeing safe plant operation.

The DC will work with DC-2 and -4 to build prognostic degradation models.

Expected results

  1. An algorithm for learning the current equipment condition as well as learning a model for predicting future degradation,
  2. A framework for integrated optimization of long-term operation and maintenance decisions,
  3. Validation of methodology by simulation of case studies (e.g., an LNG facility),
  4. Including models from DC-2 and -4,
  5. Prototype software implementation of algorithms for learning-based health-aware operation and maintenance planning

Planned secondments 

  1. KAIROS, M18, 3 months: Case study application of health-aware operation and maintenance
  2. Imperial, M15, 2 months: Training on deep learning and machine learning for process safety