Abstract
There is an increasingly demanding for the development of renewable energy. The majority
of renewable energy sources are classified as low-grade heat, which needs an advanced power
cycle to recover energy. The transcritical carbon dioxide power cycle (T-CO₂) can provide
the advantages of the components’ feasibility and high Carnot efficiency with low-grade heat
sources. The performance of the expander plays a vital effect in the cycles, while the common
turbines are suffering extremely high rotational speed for micro-scale size. Therefore, it gives
the scroll expander (volumetric type) a chance for the micro-scale transcritical carbon dioxide
power cycle with the potential benefits of lower rotational speed, simpler geometry design,
and lower cost. However, the leakages of scroll type are reported in the previous organic
Rankine cycle (ORC) applications. The flank leakage is unavoidable because of the working
fundamentals between fixed and orbiting scrolls and it is caused by the pressure difference
between upstream high-pressure chambers and downstream low-pressure chambers. This
leakage can be more serious with high-pressure working conditions of the transcritical carbon
dioxide power cycle. Therefore, the development of scroll expanders applying to micro-scale
transcritical carbon dioxide power systems is still at the very early stage and limited literature
focuses on this field. The micro-scale scroll expander can potentially fill the gap if the
leakage issues get solved.
The main objective of this Ph.D. research is to construct a preliminary micro-scale scroll
expander prototype for transcritical carbon dioxide power cycle using computational fluid
dynamics (CFD) methodology. The transient three-dimensional CFD model is expected
to analyse the thermodynamic and aerodynamic performances of different scroll expander
designs. The model is also intended to assess the effectiveness of extra sealing designs for
flank clearance leakage as well as the optimal flank clearance size.
The current thesis uses the publication style and the major findings are related from
chapter 4 to chapter 7. Therefore, there are four major contributions of the current: 1) The
transcritical carbon dioxide scroll expander demonstrates better performance than the R123.
The average isentropic and exergy efficiency benefits are around 14 % and 7 %, respectively.
The specific power output of transcritical carbon dioxide scroll expander is about four times
higher than R123. Over-expansion and reversed flow are observed in the standard flank
clearance design (20 µm) and the flow characteristics are more sensitive to scroll geometry
rather than working fluids. 2) Increasing the size of flank clearance from 20 to 200 µm can
drop down the isentropic efficiency from 87.4 % to 44.8 % and the exergy efficiency shows a
similar trend. At the same time, increasing the size is beneficial to lower the manufacturing
cost and there is a trade-off relation between manufacturing cost and isentropic efficiency.
The surrogate-assisted multi-objective optimisation algorithm (SAMO) successfully solves
the trade-off relation and the final Pareto front indicates that the flank clearance of 127 to
200 µm has the most potential for the sealing design. In addition, the SAMO significantly
reduces the CFD computational time from 90 hours to 15 seconds. 3) Increasing the size of
flank clearance can effectively reduce the possibilities of over-expansion pheromones while
excessively enlarging the size results in under-expansion. The "turning point" is 40 µm for
the current design. The pressure imbalance between two symmetry expansion chambers is
gradually reduced when enlarging the flank clearance. The most efficient flank clearance to
balance the pressure was about 100 to 150 µm for this particular expander geometry under
investigation. 4) The function of the sealing cavity in the transcritical CO2 scroll expander is
achieved by the throttling process, where the jet divergence occurs in flank clearance and
then the dissipation process works in the seal cavity. The seal cavity is beneficial to reduce
leakage and increase the expansion ratio. The fundamentals of this design are similar to the
labyrinth seals for the turbine. The different geometrical sealing cavity designs of singlegroup
contribute a limited improvement to the isentropic efficiency, ranging from 0.91 %
to 0.95 %. The height and cavity number indicate the linear positive trend to the isentropic
efficiency, while the cavity spacing shows the opposite. There is no clear difference between
the three sealing cavity groups of rectangular-shaped teeth (RST), isosceles trapezoid-shaped
teeth (ITST), and right-angled trapezoid-shaped teeth (RTST). The transient leakage ratio
keeps a consistent trend with the isentropic efficiency. The transient leakage got massively
reduced by the sealing cavity from 55.3 % to 70.2 %. However, the transient leakage
reduction for one rotating angle has a very limited function in the overall performance. The
final decision of sealing cavity for multi-groups is based on the trade-off relation between
sealing performance and manufacturing. The locations of the sealing cavity are important
to solve the pressure imbalance between two symmetrical working chambers. The paper
suggests designing the upstream sealing cavity for a lower-pressure working process and
downstream for a higher-pressure working process, which can ideally achieve the maximum
pressure balance.