Radial Turboexpander Optimization Over Discretized Heavy-Duty Test Cycles for Mobile Organic Rankine Cycle Applications

Abstract

Mobile organic Rankine cycle (MORC) systems represent a candidate technology for the reduction of fuel consumption and CO2 emissions from heavy-duty vehicles. Through the recovery of internal combustion engine waste heat, energy can be either compounded or used to power vehicle ancillary systems. Waste heat recovery systems have been shown to deliver fuel economy improvements of up to 13% in large diesel engines [1]. Whilst the majority of studies focus on individual component performance under specific thermodynamic conditions, there has been little investigation into the effects of expander specification across transient test cycles used for heavy-duty engine emission certification. It is this holistic approach which will allow prediction of the validity of MORC systems for different classes of heavy-duty vehicle, in addition to providing an indication of system performance. This paper first describes a meanline (one-dimensional simulation along a mean streamline within a flow passage) model for radial ORC turbines, divided into two main subroutines. An on-design code takes a thermodynamic input, before generating a candidate geometry for a chosen operating point. The efficacy of this design is then evaluated by an off-design code, which applies loss correlations to the proposed geometry to give a prediction of turbine performance. The meanline code is then executed inside a quasi-steady-state ORC cycle model, using reference emission test cycles to generate exhaust (heat source) boundary conditions, generated by a simulated 11.7L heavy-duty diesel engine. A detailed evaporator model, developed using the NTU-effectiveness method and single/two-phase flow correlations, provides accurate treatment of heat flow within the system. Together, these elements allow estimation of ORC system performance across entire reference emission test cycles. In order to investigate the limits of MORC performance, a Genetic Algorithm is applied to the ORC expander, aiming to optimize the geometry specification (radii, areas, blade heights, angles) to provide maximal time-averaged power output. This process is applied across the reference duty cycles, with the implications on power output and turbine geometry discussed for each. Due to the large possible variation in thermodynamic conditions within the turbine operating range a typical ideal-gas methodology (generating a single operating map for interpolation across all operating points) is no longer accurate — a complete off-design calculation must therefore be performed for all operating points. To reduce computational effort, discretization of the ORC thermodynamic inputs (temperature, mass flow rate) is investigated with several strategies proposed for reduced-order simulation. The paper concludes by predicting which heavy-duty emission test cycles stand to benefit the most from this optimization procedure, along with a comparison to existing transient results. Duty cycles containing narrow bands of operation were found to provide optimal performance, with a Constant-Speed, Variable-Load cycle achieving an average power output of 4.60 kW. Consideration is also given to the effectiveness of the methodology contained within the paper, the challenges of making ORC systems viable for mobile applications, along with suggestions for future research developments.

A methodology for the preliminary design and performance prediction of high-pressure ratio radial-inflow turbines

“…The reader is referred to Refs. [60,61] for examples of possible sets of design variables. The strategy shown in Figure 3(a) applies to a predefined set of input variables, and a dedicated optimisation procedure needs to be applied in order to determine the optimal set of inputs.…”

Analysis of impeller blade parameters and tip clearance of turboexpander in organic Rankine cycle system

“…To improve the performance of the turboexpander, it is necessary to optimize the essential geometric parameters of the turbine impeller blades, such as the blade inlet angle, the blade shape, the tip clearance height, and the trailing edge thickness. To study the limitations of ORC performance, Robertson et al applied the genetic algorithm to the turboexpander of the ORC system, aiming to optimize the geometry specification (radii, areas, blade heights, and angles) to provide maximal time‐averaged power output. Khairuddin et al used genetic algorithms to adjust the main features of the blade geometry—he hub, shroud, camber line, leading, and trailing edge profiles—to significantly improve the efficiency of the turboexpander.…”