SMARTER artefact by nikeborome


									                         Stochastic Modelling Approach
                                              for the
             Reduction of Technical & Engineering-based Risk

                                       Presentation by
                                       Gordon Barr
                        Dept. of Engineering Mathematics, University of Bristol

                                     30th January 2003

University of Bristol
Research Background
 Initially developed through the EPSRC STARTright
  project, in collaboration with Cambridge University
 Brief: To provide software-based methods for engineers
  to minimise risk in the design process
 Companies were consulted from several sectors of the
  engineering industry and their project processes studied
 Prototype software and methods primarily developed with
  ALSTOM, and to lesser extent, Metso-Svedala

  University of Bristol
The SMARTER Software Development
 SMARTER endeavours to:
  – Allow design information to be inputted which has an
    impact on key project deliverables
  – Enable engineers to highlight the major causes of
    uncertainty and risk in engineering projects
  – Permit the entire project team to participate in the
    management of project risk and uncertainty
  – Reuse relevant historical data in current project
  – Make possible the exploration of project alternatives
    and what-if analysis

 University of Bristol
Key Performance Indicators (KPIs)
 KPIs are used to help make decisions
 SMARTER enables:
   – KPIs to be identified
          Using Sensitivity Analysis
   – KPIs to be tackled
          At the outset of engineering projects
          Given limited amounts of historical information
          Using the ability of experienced engineers
 The consequence of failing to meet KPIs
   – e.g. ‘How likely is it that the project will be over budget?’
   – e.g. ‘Which KPIs should we be most worried about?’

  University of Bristol
About the Case Study: Metso-Svedala
 Mechanical Design & Management Company
 Specialise in designing & coordinating the build of heavy
  mechanical plant.
 Product designs include:
   – Rail Car Dumper Systems
   – Stacker Systems for Loose Material
   – Reclaimer Systems for Loose Material
   – Ship Mounted Material Unloaders
 Selected Case Study:
  Tandem Random Car Dumper (Crescent)

  University of Bristol
Metso-Svedala - Single Unit Car Dumper

 University of Bristol
Metso-Svedala - Car Dumper In Operation

 Crescent Dumper (0º)    Crescent Dumper (45º)

 University of Bristol
Metso-Svedala – Car Dumper Pictures

 University of Bristol
Metso-Svedala Risk Issues
 Each Metso-Svedala project typically contains a large
  amount of original, adaptive or variant design
 Original designs cause problems!
   – Will the client change their requirements radically?
   – Will the components work together in a different
   – Suppliers and contractors are often new to the
          Performance issues
          Problems adhering to standards (local practices, etc)

  University of Bristol
SMARTER Method Overview
 Application split into two distinct sections:
   – Wizard
   – Advanced
 Reason: Generally engineers do not fully understand / like
  to use probability theory!
 The ‘Wizard’ section:
   – Removes the exposure of engineers to the underlying
     probability model.
   – Allows engineers to assert probability distributions
     based on linguistic statements.

  University of Bristol
              Software Demonstration
                  Please do feel free to ask questions!

University of Bristol
                        Additional Slides
                        These slides may be used to
                        elaborate on certain points.

University of Bristol
Project Structures

                  Assembly                                      PBS Tree
                  An artefact which must be designed,
                  manufactured, installed, etc.

                  e.g. Bought in components / materials used.

                  Work Package                                  WBS Tree
                  A collection of tasks which share a common
                  project goal.

                  e.g. Design a Widget, Manufacture a Widget.

 University of Bristol
PBS and WBS Interaction
 Product Breakdown Structure                   Work Breakdown Structure



 University of Bristol
Issues Important to Metso-Svedala
 The following Metso-Svedala issues were identified:
   – Conceptual design stage / tendering stage is critical
   – Uncertainty relating to ‘original’ design elements is not
     currently quantified
   – Monitoring project target values is critical (similar to
   – Metso-Svedala engineers need a fast method of
     evaluating data owing to time pressures
   – Not all Metso-Svedala engineers are familiar with
     probabilistic modelling methods

  University of Bristol

To top