decisionLabsmart models by creative thinkers
dL
Page 2
>
dLOutline
1. Introduction to decisionLab
2. ATOM Overview
3. ATOM Use Case
4. ATOM Future Development
5. Artificial Intelligence Approach to Predictive Maintenance
6. Summary
Page 3
>
dLSome of the clients we help
Page 4
>
dLHow we do it?
SectorsAerospace
DefenceWaterPower
Transport
Services
Develop bespoke decision support
systems to optimisebusiness operations.
Agile Approach
Page 5
>
dLAgent-based
Turbine
Operations &
Maintenance
dL
Page 6
>
dL
Siemens Business Requirement• The introduction of the Siemens SGT-A65 aero-derivative gas turbine into Siemens Power Generation
product line has highlighted issues with the in-service performance and the supporting Repair & Overhaul.
• Existing excel-based methods for forecasting engine removals and the demand on R&O capacity and spare modules are inadequate given the complexity of the problem.
Page 7
>
dL
• How do you predict business performance and forecast KPIs to inform decision-making?
• How to evaluate investments options – investment in improving engine performance, Repair & Overhaul capacity, Increase in spares inventory?
Exam Question:
Page 8
>
dL
Digital Representation: “The digital representation that provides both the elements and the dynamics of how a system operates and lives throughout its life cycle”.
Digital Integration: “Integrates artificial intelligence, machine learning and data analytics to create a digital simulation models that update and change as their physical counterparts change. A digital twin is able to continuously learns and updates itself from multiple sources…..”
What is a Digital Twin?
Page 9
>
dLDigital Twin• Provides the capability to use sophisticated simulation and data-analytics methodologies to model
fleet operations of Siemens gas turbines.
Page 10
>
dLKey Benefits• The ability to model the entire fleet operations of Siemens aero-derivative gas turbines will enable:
• Forecasting of system KPIs• Visualization of fleet and maintenance facility operations• Identification of bottlenecks in the system and its impact of maintenance cost.• Run ‘what-if’ scenarios to aid investment decision-making.
Page 11
>
dL
DIGITAL TWIN ADOPTION:A DEEP DIVE INTO THEATOM MODEL
dL
Page 12
>
dL
Model Complexity
Page 13
>
dL
Page 14
>
dL
Page 15
>
dL
Page 16
>
dLMODEL ARCHITECTURE
• Decoupling model functionality
• Simultaneous development
• Version control (Git)• Cloud deployment
Page 17
>
dLATOM Future Development
Phase 1 – Standalone Application Phase 2 – Enterprise Solution Phase 3 – Introduction of AI to Simulation Modelling
Current capability Future Capability
Applying AI to optimization
Moving agents from rule-based decision making to agents that agents learn their own knowledge directly from historic data and by trial and error – Reinforcement Learning
Page 18
>
dL
• Not a Closed-Form Solution:• Agent-based modelling offers a simulation framework that is flexibility to solve a variety of business
process across varied levels of complexity.
• An enabler to the digital twin across various industry sectors
Extrapolation to Other Business Units
• Babcock International have taken on management of the tank power-pack repair system for the MOD
• £100s Million per annum
• Use Cases:
Page 19
>
dLAI Approach to Predictive Maintenance
dL
Page 20
>
dL
AI Capability
User Interface
Programme Nelson
Type 45 Destroyer: Use Case
Deployment
Page 21
>
dLSummary
• ATOM captures complexities of the lifecycle operations of industrial gas turbines – Digital representation of the assets operations.
• ATOM enables prediction of business performance and forecast KPIs to inform decision-making.
• ATOM is an enabler to the digital twin and is the first step towards the digital twin status.
• ATOM embodies agent-based modelling that provides a simulation framework that can be extrapolated to other business units.
decisionLabsmart models by creative thinkers
Thank you!
Top Related