Thermofluid optimisation of turboexpanders for mobile organic Rankine cycle systems
File(s)
Author(s)
Robertson, Miles C.
Type
Thesis
Abstract
The Organic Rankine Cycle (ORC) is a closed-loop thermodynamic cycle used for low-temperature heat recovery, from sources as diverse as solar, geothermal and industrial processes. Within the ORC, power is produced with the use of a dense-gas vapour expansion in a suitable device and the efficiency of this process is a key influence on cycle efficiency. This thesis therefore investigates the design, optimisation and validation of nozzled radial-inflow turboexpanders, applied to the recovery of vehicle exhaust waste heat.
A 1-dimensional meanline performance prediction methodology is developed, capable of modelling the real-gas effects prevalent in dense, organic working fluids. In addition, choking effects within flow passages are explicitly accounted for in both nozzle and rotor, allowing accurate prediction at the high Pressure Ratios typically observed in ORC expanders. A validation case is subsequently presented, tuning loss correlation coefficients to match 6 high Pressure Ratio experimental test cases, reducing average mass flow rate and efficiency errors to 1.33% and 2.04%, respectively.
This 1D model is used to create a methodology in which expander optimisation is performed across an entire vehicle duty cycle. A discretised heat exchanger model is developed, capable of modelling multiple fluid phases and real gas effects. A number of simplifications including surrogate models generated by neural network fits, lead to an O(10^3) reduction in computing effort. A case study is performed to quantify benefits to a 11.7 L diesel engine running a 1200-point Non-Road Transient Cycle, resulting in a 21.9% improvement in total cycle energy recovery against a single-point baseline case.
The methodologies contained within this thesis are used to design a turboexpander for an industrial research project, focusing on a 15 L diesel Tier 4 engine. Across two generations of design, a combination of computational and experimental methods are applied to produce an expander containing non-radial rotor inlet blading, producing a simulated power output of 12.4 kW at the design point.
Finally, in order to facilitate validation of real-gas Computational Fluid Dynamics (CFD) simulations and better understand the underlying flow physics, a blowdown facility is constructed for working fluid r1233zd(E), centred around a converging-diverging test section producing a Mach 2 expansion at the exit plane. Comparison of CFD and centreline pressure measurements for both nitrogen and refrigerant produced maximum errors in Pressure Ratio of 5.7% and 12.1% respectively, suggesting an overestimation by computational methods in predicting expander output power.
The methodologies, designs and experimental results contained within this thesis provide improvements to the modelling, optimisation and validation of dense-gas turboexpanders, with the aim of improving design processes and aiding the implementation of ORC for vehicle waste heat recovery.
A 1-dimensional meanline performance prediction methodology is developed, capable of modelling the real-gas effects prevalent in dense, organic working fluids. In addition, choking effects within flow passages are explicitly accounted for in both nozzle and rotor, allowing accurate prediction at the high Pressure Ratios typically observed in ORC expanders. A validation case is subsequently presented, tuning loss correlation coefficients to match 6 high Pressure Ratio experimental test cases, reducing average mass flow rate and efficiency errors to 1.33% and 2.04%, respectively.
This 1D model is used to create a methodology in which expander optimisation is performed across an entire vehicle duty cycle. A discretised heat exchanger model is developed, capable of modelling multiple fluid phases and real gas effects. A number of simplifications including surrogate models generated by neural network fits, lead to an O(10^3) reduction in computing effort. A case study is performed to quantify benefits to a 11.7 L diesel engine running a 1200-point Non-Road Transient Cycle, resulting in a 21.9% improvement in total cycle energy recovery against a single-point baseline case.
The methodologies contained within this thesis are used to design a turboexpander for an industrial research project, focusing on a 15 L diesel Tier 4 engine. Across two generations of design, a combination of computational and experimental methods are applied to produce an expander containing non-radial rotor inlet blading, producing a simulated power output of 12.4 kW at the design point.
Finally, in order to facilitate validation of real-gas Computational Fluid Dynamics (CFD) simulations and better understand the underlying flow physics, a blowdown facility is constructed for working fluid r1233zd(E), centred around a converging-diverging test section producing a Mach 2 expansion at the exit plane. Comparison of CFD and centreline pressure measurements for both nitrogen and refrigerant produced maximum errors in Pressure Ratio of 5.7% and 12.1% respectively, suggesting an overestimation by computational methods in predicting expander output power.
The methodologies, designs and experimental results contained within this thesis provide improvements to the modelling, optimisation and validation of dense-gas turboexpanders, with the aim of improving design processes and aiding the implementation of ORC for vehicle waste heat recovery.
Version
Open Access
Date Issued
2018-08
Date Awarded
2019-02
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Martinez-Botas, Ricardo
Costall, Aaron
Publisher Department
Mechanical Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)