DH
D.J. Huizer
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
1 records found
1
Hydrogen Pipeline Compressor Substation
A Physics-Based Optimisation Framework for Hydrogen Pipeline Compressor Substation Design: Configuration Selection, Intercooling, and Heterogeneous Train Dispatch
Hydrogen pipeline networks require compressor substations to overcome frictional pressure losses over long transport distances, yet no established design methodology exists for translating hydrogen’s distinct thermodynamic and safety properties into a substation configuration. This thesis addresses the research question: how can hydrogen compressor substation design be quantitatively optimised using a physics-based model under thermodynamic, operational, and safety constraints?
A python-based model combining the Leachman equation of state (via CoolProp), polytropic compression and intercooling relations, mechanical and utilisation-based feasibility filters, and a Casey & Robinson elliptic-curve representation of off-design centrifugal compressor behaviour was developed and applied to a representative case of 40 to 80 bar boundary conditions. A three-layer optimisation architecture reduces the combinatorial space of 4094 candidate configurations, first through independent mechanical and thermodynamic feasibility screening, then a joint SLSQP-based pressure-ratio and intercooler- placement optimisation, and finally a mass-flow-allotment optimisation across heterogeneous parallel trains, cutting the number of required thermodynamic property evaluations by 94%.
Three intercoolers (K = 3) emerged as the unambiguous optimum across all evaluated stage counts and both impeller tip speeds (450 and 600 m/s), driven by rapidly diminishing marginal cooling benefit against a fixed intercooler pressure drop. Stage count was found to affect mechanical and capital cost but not thermodynamic performance; optimising the pressure-ratio distribution across stage groups, rather than assuming a uniform split, yielded a modest 0.263% reduction in specific work, with most of the achievable efficiency gain instead coming from intercooler count and placement. Heterogeneous train dispatch was shown to substantially widen the usable operating window relative to an equal-split strategy, though the tightest surge/choke margin identified across the design space narrowed to approximately 1.19 times the design flow coefficient at the final stage of the (N=12, K=4) configuration.
The techno-economic evaluation, benchmarked against a Class 4 (AACE, ±30%) accuracy correlation, placed total CAPEX at 1421.95 and 1932.64 million euros for the titanium- and composite-impeller configurations respectively, though this comparison is confounded by the two configurations also differing in stage count and tip speed.
This work’s primary contribution is the layered optimisation architecture itself, together with the demonstration that heterogeneous train sizing offers a meaningful operational flexibility benefit over homogeneous dispatch. The framework’s predictions were compared against published elliptic-curve calibration data and existing compressor performance maps, though full off-design validation against independent experimental or vendor data was outside this thesis’s scope. As a decision-support tool for early-stage compressor substation design, the model is positioned to guide configuration selection ahead of detailed engineering, rather than to replace it. ...
A python-based model combining the Leachman equation of state (via CoolProp), polytropic compression and intercooling relations, mechanical and utilisation-based feasibility filters, and a Casey & Robinson elliptic-curve representation of off-design centrifugal compressor behaviour was developed and applied to a representative case of 40 to 80 bar boundary conditions. A three-layer optimisation architecture reduces the combinatorial space of 4094 candidate configurations, first through independent mechanical and thermodynamic feasibility screening, then a joint SLSQP-based pressure-ratio and intercooler- placement optimisation, and finally a mass-flow-allotment optimisation across heterogeneous parallel trains, cutting the number of required thermodynamic property evaluations by 94%.
Three intercoolers (K = 3) emerged as the unambiguous optimum across all evaluated stage counts and both impeller tip speeds (450 and 600 m/s), driven by rapidly diminishing marginal cooling benefit against a fixed intercooler pressure drop. Stage count was found to affect mechanical and capital cost but not thermodynamic performance; optimising the pressure-ratio distribution across stage groups, rather than assuming a uniform split, yielded a modest 0.263% reduction in specific work, with most of the achievable efficiency gain instead coming from intercooler count and placement. Heterogeneous train dispatch was shown to substantially widen the usable operating window relative to an equal-split strategy, though the tightest surge/choke margin identified across the design space narrowed to approximately 1.19 times the design flow coefficient at the final stage of the (N=12, K=4) configuration.
The techno-economic evaluation, benchmarked against a Class 4 (AACE, ±30%) accuracy correlation, placed total CAPEX at 1421.95 and 1932.64 million euros for the titanium- and composite-impeller configurations respectively, though this comparison is confounded by the two configurations also differing in stage count and tip speed.
This work’s primary contribution is the layered optimisation architecture itself, together with the demonstration that heterogeneous train sizing offers a meaningful operational flexibility benefit over homogeneous dispatch. The framework’s predictions were compared against published elliptic-curve calibration data and existing compressor performance maps, though full off-design validation against independent experimental or vendor data was outside this thesis’s scope. As a decision-support tool for early-stage compressor substation design, the model is positioned to guide configuration selection ahead of detailed engineering, rather than to replace it. ...
Hydrogen pipeline networks require compressor substations to overcome frictional pressure losses over long transport distances, yet no established design methodology exists for translating hydrogen’s distinct thermodynamic and safety properties into a substation configuration. This thesis addresses the research question: how can hydrogen compressor substation design be quantitatively optimised using a physics-based model under thermodynamic, operational, and safety constraints?
A python-based model combining the Leachman equation of state (via CoolProp), polytropic compression and intercooling relations, mechanical and utilisation-based feasibility filters, and a Casey & Robinson elliptic-curve representation of off-design centrifugal compressor behaviour was developed and applied to a representative case of 40 to 80 bar boundary conditions. A three-layer optimisation architecture reduces the combinatorial space of 4094 candidate configurations, first through independent mechanical and thermodynamic feasibility screening, then a joint SLSQP-based pressure-ratio and intercooler- placement optimisation, and finally a mass-flow-allotment optimisation across heterogeneous parallel trains, cutting the number of required thermodynamic property evaluations by 94%.
Three intercoolers (K = 3) emerged as the unambiguous optimum across all evaluated stage counts and both impeller tip speeds (450 and 600 m/s), driven by rapidly diminishing marginal cooling benefit against a fixed intercooler pressure drop. Stage count was found to affect mechanical and capital cost but not thermodynamic performance; optimising the pressure-ratio distribution across stage groups, rather than assuming a uniform split, yielded a modest 0.263% reduction in specific work, with most of the achievable efficiency gain instead coming from intercooler count and placement. Heterogeneous train dispatch was shown to substantially widen the usable operating window relative to an equal-split strategy, though the tightest surge/choke margin identified across the design space narrowed to approximately 1.19 times the design flow coefficient at the final stage of the (N=12, K=4) configuration.
The techno-economic evaluation, benchmarked against a Class 4 (AACE, ±30%) accuracy correlation, placed total CAPEX at 1421.95 and 1932.64 million euros for the titanium- and composite-impeller configurations respectively, though this comparison is confounded by the two configurations also differing in stage count and tip speed.
This work’s primary contribution is the layered optimisation architecture itself, together with the demonstration that heterogeneous train sizing offers a meaningful operational flexibility benefit over homogeneous dispatch. The framework’s predictions were compared against published elliptic-curve calibration data and existing compressor performance maps, though full off-design validation against independent experimental or vendor data was outside this thesis’s scope. As a decision-support tool for early-stage compressor substation design, the model is positioned to guide configuration selection ahead of detailed engineering, rather than to replace it.
A python-based model combining the Leachman equation of state (via CoolProp), polytropic compression and intercooling relations, mechanical and utilisation-based feasibility filters, and a Casey & Robinson elliptic-curve representation of off-design centrifugal compressor behaviour was developed and applied to a representative case of 40 to 80 bar boundary conditions. A three-layer optimisation architecture reduces the combinatorial space of 4094 candidate configurations, first through independent mechanical and thermodynamic feasibility screening, then a joint SLSQP-based pressure-ratio and intercooler- placement optimisation, and finally a mass-flow-allotment optimisation across heterogeneous parallel trains, cutting the number of required thermodynamic property evaluations by 94%.
Three intercoolers (K = 3) emerged as the unambiguous optimum across all evaluated stage counts and both impeller tip speeds (450 and 600 m/s), driven by rapidly diminishing marginal cooling benefit against a fixed intercooler pressure drop. Stage count was found to affect mechanical and capital cost but not thermodynamic performance; optimising the pressure-ratio distribution across stage groups, rather than assuming a uniform split, yielded a modest 0.263% reduction in specific work, with most of the achievable efficiency gain instead coming from intercooler count and placement. Heterogeneous train dispatch was shown to substantially widen the usable operating window relative to an equal-split strategy, though the tightest surge/choke margin identified across the design space narrowed to approximately 1.19 times the design flow coefficient at the final stage of the (N=12, K=4) configuration.
The techno-economic evaluation, benchmarked against a Class 4 (AACE, ±30%) accuracy correlation, placed total CAPEX at 1421.95 and 1932.64 million euros for the titanium- and composite-impeller configurations respectively, though this comparison is confounded by the two configurations also differing in stage count and tip speed.
This work’s primary contribution is the layered optimisation architecture itself, together with the demonstration that heterogeneous train sizing offers a meaningful operational flexibility benefit over homogeneous dispatch. The framework’s predictions were compared against published elliptic-curve calibration data and existing compressor performance maps, though full off-design validation against independent experimental or vendor data was outside this thesis’s scope. As a decision-support tool for early-stage compressor substation design, the model is positioned to guide configuration selection ahead of detailed engineering, rather than to replace it.