Road to Data Driven Operations: How to get there

The overall proposal is based on the premise that a digital twin model of any equipment/system can be designed and developed using artificial intelligence and historical data, thereon evolved, using machine learning algorithm, with input from offline and online maintenance and operating data. The development of the digital twin model will need to be executed sequentially:

  1. Preparation data model with historical incidents and performance data with known outcomes, thereon
  2. The developed model is then used against Current State KPIs to predict the outcome bundled with other user actions/operational data.

The road to data driven operations is to make strategic decisions based on data analysis and interpretation rather than intuitions and making observations alone.

1. Machine Learned Data Model: The first step is to build a machine-learned data model by providing it an algorithm it can use to reason over and learn from those data (Historical data & Operating data). The model will be able to learn to recognize certain types of patterns.

2. Simulation Tool: Instead of representing a complete system as a statistical algorithm or generating a fixed data set, simulation captures the characteristics and relationships of system components to provide a dynamic model. The simulation tool helps to estimate the reliability indices by simulating the actual process and possible outcomes.

3. Next Action: Based on the analysis of the simulation tool, it helps to produce probabilities and predictions for systems in the future and thus make informed decisions regarding the next steps accordingly.

Execution Plan and Deliverables

The focus shall be on Phase 1, with a follow-on Phase 2 program as depicted in the figure below.

PHASE 1

  • Digitize historical data and prepare ongoing data capture process.
  • Analyse historical data to identify Key Performance Indicators and Data Requirements for performance optimizations.
  • Produce report and dashboard to visualize the data and decision-making process based on sampled data including associated data model(s).
  • Identification and recommendation of scope for Phase 2.

PHASE 2

  • Identification of Data Collection Requirement for the Key Performance Indicators along with their sources e.g. sensors, periodic inspection reports etc.
  • Data flow and processing architecture e.g. from field transmitters up to the onshore data processing and analysis system. In case of manual/physical inspection data, frequency of data collection and manual upload into the system via a standard data entry or management tool including from data from CMMS.
  • AI Driven Performance Dashboard with Simulation.