Published on September 26, 2014
Probabilistic analysis of marine fuels in emission controlled areas Kamal Soundararajan (Kamal) 6th ICAE 2014, 2 June
2 Overview •Motivations •Methodology •Results •Conclusion http://www.marineinsight.com/wp-content/uploads/2012/07/co2.jpg
Increasing global pressure to reduce GHG emissions •Mandatory regulations to curb the rise in SOx and NOx emissions •Energy efficiency would fuel the future of shipping •Ship owners starting to explore various fuel and technology options (LNG as a marine fuel, dual fuel options, solar and wind power, waste heat recovery systems) •Selection of fuel and technology options is complex, requires long- term planning and is often very costly to make changes 4 http://www.dieselnet.com/standards/inter/imo.php http://www.dieselnet.com/standards/inter/imo.php
How do future IMO regulations influence decision-makers’ selection of marine fuels? •No easy answers ▫Highly diversified sector (different ship types, varying operating profiles, varying pricing mechanisms) ▫Uncertainty over future fuel prices, infrastructure and regulations ▫Complex interactions between ship owners and operators (split-incentives) •Very common forms of analysis include cost benefit analysis •This paper attempts to provide some insights by partially resolving the uncertainty and the various risk- appetitites that affect the decision-making process 5
6 Methodology http://pilotproject.com/
Summary •Problem formulation •Deterministic Analysis •Probabilistic Analysis •Model Appraisal 7 The Decision Analysis Cycle/Process
Formulation: A purchase decision with three alternatives •Three types of ship engines configurations considered ▫Two stroke diesel engine fitted with SOx scrubbers : NPV1 ▫Gas fired engine with LNG as a fuel source : NPV2 ▫Dual fuel engine that allows a combination of LNG and fuel oil : NPV3 •Decision specifications estimated from publicly available sources (for example) ▫Cost of initial investment ▫Revenue estimates ▫Cost of operating SOx scrubbers •Defining the boundaries for a more impactful analysis ▫Mid-sized containership between Asia and Europe 8
Deterministic Analysis: Cash flow analysis for each alternative 9 7 3 7 1,2,3 3 k (1 )k k (1 )k i Fuel Costs i Earnings NPV Initial Costs Cash flow diagram of NPV1 Cash flow diagram of NPV2/3 Diagrams are for illustrative purposes only. Actual cash flow diagrams are varied
•Reliability of different engines ▫Downtime experienced : d1 , d2 and d3 •Uncertainty in future prices ▫Price estimates of LNG and Marine Gas Oil (MGO): P LNG and Pdies •Uncertainty over future LNG bunkering infrastructure ▫Additional distance travelled: D •Uncertainty over future NOx and energy efficiency regulations ▫Additional cost incurred at a later time frame: I 10 Probabilistic analysis: Uncertain system variables
Influence diagram of purchase decision 11 NPV 1 NPV 2 NPV 3 D d2 I d3 Plng Purchase decision d: downtime D: Additional distance traveled P: Price l: Additional cost incurred I D d3 d1 Pdies Plng NPV1 NPV2 NPV3 Purchase Decision
DPL software used to conduct various types of analysis •Tornado dominance •Probabilistic decision tree diagrams •Sensitivity analysis •Expected Value of Perfect Information (EVPI) •Risk analysis 12
13 Results https://brillianta3.files.wordpress.com/2012/
Deterministic policy tree diagram •LNG as a fuel source was chosen as the optimal solution 14 94.7 [94.7] Diesel 107.8 [107.8] LNG 103.3 [103.3] Dual Decision1 [107.8] Licensed by Syncopation Software for evaluation purposes only.
Combined Tornado Diagram 15 84.7 111.3 118.8 98.3 109.5 108.8 89.1 102.3 75.4 92.3 83.7 94.2 99.6 106.8 105.7 104.7 94.7 103.3 107.8 d3 d2 d1 Plng Pdies/lng Pdies I D
Combined Tornado Diagram •No tornado dominance among the three alternatives in the deterministic analysis •Selected variables are deemed sensitive enough for probability assessment 16
Probabilistic decision tree diagrams • Change of optimal decision to select diesel-fuelled engine ▫ Switch to cleaner fuel is strongly influenced by probabilistic events • Risk neutrality assumed 17 I Diesel [101.5] D LNG [100.5] d3 Dual [100.2] Decision1 [101.5]
Full probabilistic decision tree 18 Low NPV_1 Base NPV_1 High NPV_1 High b Pdies Base b Low b NOx and EE a d1 NOx a EE a None a Diesel I Low NPV_2 Base NPV_2 High NPV_2 High d Plng Base Low Yes c d2 No LNG D Dual Decision1 Relatively large base model consisting 84 various objective functions
Sensitivity Analysis •Most variables are found to be robust ▫Small changes in value do not change the optimal decision •Optimal decision changes to LNG for: ▫An increase of $2 for the base price of diesel ▫A reduction of $1 for the base price of LNG ▫An increase of 0.5 weeks of downtime for diesel engines 19
Expected value of perfect information (EVPI) analysis 20 Price of diesel Price of LNG Optimal decision Low Low Diesel Low Base Diesel Low High Diesel Base Low LNG Base Base Diesel Base High Diesel High Low LNG High Base LNG High High Dual Limitations in LNG bunkering infrastructure Optimal decision High Diesel Low LNG Future regulations in 2016 Optimal decision NOx and EE regulations LNG NOx regulations LNG EE regulations Diesel None Diesel
Expected value of perfect information (EVPI) analysis •Where LNG prices are cheaper than diesel, LNG as a fuel source is the optimal decision ▫Supports market-based measures for emission reduction efforts •Only when LNG and diesel prices are high, dual fuel engine becomes the optimal decision •Improving LNG bunkering infrastructure does help promote LNG being chosen as a fuel source •Future regulations to reduce NOx in ECAs does promote LNG as a optimal decision 21
Risk Analysis 22 Increasing risk aversion Assuming decision-maker has an exponential utility function and satisfies delta property
Risk Analysis •With decreasing risk tolerance the optimal decision changes from diesel to LNG and finally to dual fuel engines ▫Seems counter-intuitive since dual-fuel engines are often considered to be the ‘more’ risky option •While some factors make dual fuel engines risky such as reliability, other factors such as less uncertainty over future fuel prices and regulations tend to make the option less risky 23
Summary remarks •Price uncertainties have the largest impact on the optimal decision •Optimal decision for a risk neutral decision-maker is to invest in a diesel-fuelled engine •Future regulations and improvements to LNG bunkering infrastructure promotes LNG as the optimal decision •Price regulations on diesel also improve the adoption of LNG fueled engines •Investing in dual-fueled engines in most cases is a sub- optimal decision •The more risk averse decision makers are, the more likely they would consider LNG as an alternative fuel source 25
26 Thank you! Energy Studies Institute 29 Heng Mui Keng Terrace Block A, #10-01 Singapore 119620 For enquiries: Kamal Soundararajan Tel: (65) 6516 1456 Email: email@example.com
Probabilistic Analysis of Marine Fuels in Emission Controlled Areas ... Probabilistic analysis of marine fuels in ... Probabilistic assessment of ...
Probabilistic Analysis of Marine Fuels in Emission Controlled Areas on ResearchGate, the professional network for scientists.
Probabilistic Analysis of Marine Fuels in Emission Controlled Areas. International Conference on Applied Energy
Probabilistic analysis of marine fuels in emission controlled areas Author Kamal Soundararajan, Eulalia Han Source Name International Conference on Applied ...
Probabilistic Analysis of Marine Fuels in Emission Controlled Areas. Added by. ... Probabilistic Analysis of Marine Fuels in Emission Controlled Areas.
Emission Control Areas. ... the Baltic Sea became the first fully implemented SOx Emission Control Area in August ... Bunker Fuel Analysis: Marine Water ...
... comparisons, analysis, ... Marine Fuel Choice for Ocean Going Vessels within Emission Control Areas. Release date: ...
... that operate in specially designated Emission Control Areas. The quality of fuel that complies ... Analysis of Commercial Marine Vessels ...
Emission Control Area ... according to Directive 2012/33/EU as regards the sulphur content in marine fuel: ... an analysis of the trends in fuel oil ...