Modeling Dual Fuel Internal Combustion Engine (ICE) Running on Methanol and Diesel

Abstract

Waterborn transport currently runs on diesel driven Internal Combustion Engines
(ICE). To reduce Green House Gas (GHG) emissions an option is to shift towards
alternative fuels, like alcohols. This thesis focusses on methanol in ICEs. The
combustion of methanol in ICEs can be done in numerous manners, where fuel
altering, external ignition and dual fuel concepts are the most commonly known.
Methanol is chosen as fuel due to its renewable character, which is a discussion
point. Methanol is mostly produced in a gasification process of coal and Natural
Gas (NG), whom are not renewable fuels, however methanol can be produced from
the gasification of biomass or hydrogenation of CO2. Producing methanol from
hydrogenation of CO2 is not economical beneficial at this stage, the knowledge and
experience for this method is limited, meaning this will be gained in the next
couple of years resulting in lower prices. A couple of vessels are already running
on methanol, where dual fuel and fuel blending concepts are predominance. The
conventional dual fuel is therefore modelled in thesis with a Vibe based model and
a two zone model.

Currently marine engines are Compression Ignition (CI) engines running on
diesel as fuel. Methanol has other properties compared to diesel. The biggest
difference is the lower cetane number, which is an indication on the fuels
self-ignition abilities. This defective self-ignition abilities of methanol results in the
necessity of adjusting a marine engine when using methanol as fuel. The most
redundant and known method is the conventional dual fuel engine. The diesel
injectors might need adjustments due to the smaller diesel input and there is need
for a methanol injector. Methanol can be injected in the input air duct. Two
different models are available for moddeling dual fuel combustion, namely the
Vibe based model and the two zone model. The two zone model is available for NG
and needs to be adjusted for methanol applications and the Vibe based model is
already available for a methanol applications. Experiments are done on a methanol
diesel conventional dual fuel engine and the results are described in a paper [1].

The Vibe model is a model that, in the basics, is build on reaction kinetics and
combustion duration resulting in ’shape parameters’, whom shape the Combustion
Reaction Rate (CRR). The model was constructed from the experimental pressure
trace converted into CRR by the ’Heat Release Model’, which was used in the
’in-cylinder model’ to establish the shape parameters. When the shape parameters
are known the in-cylinder model can run by itself (without the use of the heat
release model) producing a pressure and temperature versus crank angle traces.
The Start Of Combustion (SOC) and the End Of Combustion (EOC) are assumed.
The shaping parameters are valid for specific cases and therefore the model is not
predictive when changing the fuel composition or Start Of Injection (SOI).

The two zone model is a model based on two zones, the burnt zone and the
unburnt zone. The burnt zone is created by fuel injection and diffusion caused by flame speed. In the burnt zone combustion can take place where the CRR is
dependent on a version of the Arrhenius equation. The SOC is calculated with
another version of the Arrhenius equation. Predictive behavior was expected for
the two zone model, however the results contradict.

The two models generate a pressure trace. To compare the two models the so
called ’post processing’ model is introduced. This ’post processing model’ is based
on the ’Heat Release Model’ from the Vibe based model, meaning that for the
calibrated case the HHR of the ’Vibe based model’ is correct. The pressure traces
generated by the two models are inserted into the post processing model where the
output is the Heat Release Rate (HRR), Indicated Thermal Efficiency (ITE),
maximum mean in-cylinder temperature and Initial Mean Effective Pressure
(IMEP). The post processing model is first validated to the experiments, which
resulted in an acceptable match. The calculated mechanical efficiencies are between
80% and 90% which is conform reality.

The Vibe based model results in slightly higher values for IMEP and ITE
compared to the post processing model. When changing the blend ratio to less
methanol use the IMEP and ITE decreases and the mean max in-cylinder
temperature increases similar to the experiments. The HRR does not show the
premixed combustion, but it is in the same range as the experiments. The two zone
model results in a better fit for the HRR when looking at the shape for the
calibrated case. When changing the SOI or the blend ratio, the model generates can
even negative efficiencies or no results at all.

Both models are suitable for a calibrated case, however when changing the
input settings, both models do not work predictive as wanted.