Waste heat recovery from internal combustion engines (ICEs) is one potential
option to improve overall vehicle efficiency. Rankine cycles based on engine coolant
and exhaust heat sources have demonstrated their effectiveness in enhancing brake
thermal efficiency. Critical to their success is the design of the heat exchanger
for the evaporator, with shell-and-tube heat exchangers (STHEs) a common hardware
choice. However, little experimental data exists evaluating STHEs with the pulsating
flow encountered in the exhaust of ICEs. In addition, correlations for periodically
varying flow do not appear to be used in the modeling of STHEs. To alleviate this
limitation, this study combined experiments using a pulsating exhaust heat source
from an ICE under low loads at a single engine speed with a one+one-dimensional
model to evaluate tube- and shell-side heat transfer correlations for a STHE without
baffles. Four working fluids, water, ethylene glycol, propylene glycol, and a 50/50
ethylene glycol–water mixture, were examined. The combined thermodynamic properties
of an ethylene glycol–water mixture were the most effective based on an evaluation
of heat exchanger effectiveness, overall heat transfer coefficient, exergetic
efficiency, and entropy generation. A Pearson correlation analysis identified the
inlet working fluid temperature as the parameter most strongly correlated with STHE
performance due to its higher enthalpy. From a modeling perspective, the pulsating
flow correlation of Al-Haddad and Al-Binally predicted greater heat transfer rates
in the STHE. In combination with all shell-side correlations tested, simulations
still underpredict performance relative to experimental results. An optimized
correlation developed specifically for the geometry of this unbaffled STHE matched
the experimental data more closely but likely overpredicted shell-side heat
transfer. Monte Carlo uncertainty propagation based on sensor uncertainties showed
that the differences in effectiveness and overall heat transfer coefficient exceeded
measurement uncertainty. Furthermore, sensitivity analysis demonstrated the
importance of accurate thermophysical property values and indicated that the
underprediction likely reflects coupled limitations in both tube- and shell-side
formulations, with correlations on each side exerting a comparable influence on
predicted heat transfer.
