ZK
Z. Kaseb
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>
10 records found
1
This PhD thesis addresses a fundamental challenge in power system applications: how to reliably model and control modern power systems under increasing uncertainty and operational constraints. As modern power systems evolve toward highly distributed, renewable-rich infrastructures, classical numerical methods for power flow (PF) and optimal power flow (OPF) face intrinsic limitations. These limitations are not merely computational but structural: nonlinear solvers, such as Newton–Raphson (NR), exhibit sensitivity to initialization, numerical instability in ill-conditioned scenarios, and scaling behavior that is increasingly incompatible with the dimensionality requirements of future power system operation. Despite decades of algorithmic refinement, these challenges remain only partially resolved within classical computation.
The thesis systematically investigates advanced computational paradigms, i.e., artificial intelligence (AI) and quantum computing (QC), as complementary tools for overcoming these fundamental challenges. Rather than proposing incremental solver modifications, the thesis develops new conceptual formulations for PF and OPF that rethink how physical structure, optimization, and computation are represented. First, a physics-informed neural network (PINN) framework is introduced that partially encodes power balance equations directly into the learning process. This contribution advances existing PINN approaches by reducing dependence on complete network topology and line parameters while preserving physical interpretability. Second, the thesis explores quantum neural networks (QNNs) as a novel learning paradigm for PF analysis under data scarcity and noise, offering new insights into how quantum representations in the fault-tolerant quantum computing era can provide compact, potentially more expressive alternatives to classical deep learning.
Beyond learning-based methods, this PhD thesis provides a foundational contribution by reformulating PF and OPF as quadratic unconstrained binary optimization (QUBO) and Ising models, enabling their solution using quantum and quantum-inspired hardware. This reformulation constitutes a conceptual shift from continuous, nonlinear equation solving to discrete energy minimization, revealing new pathways to address scalability. The thesis further contributes to the first systematic comparison of gate-based and adiabatic quantum computing paradigms for PF and OPF across system sizes and operating conditions.
To bridge the gap between theoretical quantum algorithms and real-world power system requirements, this PhD thesis further introduces a quantum hardware-in-the-loop (QHIL) framework. This framework enables the evaluation of quantum solvers within closed-loop, real-time power system simulators. Therefore, critical system-level challenges, such as communication latency, embedding overhead, and real-time feasibility, are identified. These challenges should be addressed before quantum computing can be meaningfully deployed in operational power systems.
Finally, recognizing the enduring relevance of classical solvers, this PhD thesis contributes a reinforcement learning (RL) framework that enhances NR convergence reliability through learned initialization strategies. By integrating a quantum-enhanced environment update mechanism, the thesis further explores how quantum optimization algorithms can accelerate learning and improve robustness without replacing well-established numerical methods.
In summary, this PhD thesis contributes new scientific knowledge by reframing PF and OPF through the lenses of scientific machine learning, quantum optimization algorithms, and hybrid computational architectures. It demonstrates that AI and QC are not merely faster solvers but fundamentally different tools that enable new representations, formulations, and system-level integrations. These contributions lay a conceptual and methodological foundation for next-generation power system modeling and control, with implications extending beyond power systems to other nonlinear engineering domains. ...
The thesis systematically investigates advanced computational paradigms, i.e., artificial intelligence (AI) and quantum computing (QC), as complementary tools for overcoming these fundamental challenges. Rather than proposing incremental solver modifications, the thesis develops new conceptual formulations for PF and OPF that rethink how physical structure, optimization, and computation are represented. First, a physics-informed neural network (PINN) framework is introduced that partially encodes power balance equations directly into the learning process. This contribution advances existing PINN approaches by reducing dependence on complete network topology and line parameters while preserving physical interpretability. Second, the thesis explores quantum neural networks (QNNs) as a novel learning paradigm for PF analysis under data scarcity and noise, offering new insights into how quantum representations in the fault-tolerant quantum computing era can provide compact, potentially more expressive alternatives to classical deep learning.
Beyond learning-based methods, this PhD thesis provides a foundational contribution by reformulating PF and OPF as quadratic unconstrained binary optimization (QUBO) and Ising models, enabling their solution using quantum and quantum-inspired hardware. This reformulation constitutes a conceptual shift from continuous, nonlinear equation solving to discrete energy minimization, revealing new pathways to address scalability. The thesis further contributes to the first systematic comparison of gate-based and adiabatic quantum computing paradigms for PF and OPF across system sizes and operating conditions.
To bridge the gap between theoretical quantum algorithms and real-world power system requirements, this PhD thesis further introduces a quantum hardware-in-the-loop (QHIL) framework. This framework enables the evaluation of quantum solvers within closed-loop, real-time power system simulators. Therefore, critical system-level challenges, such as communication latency, embedding overhead, and real-time feasibility, are identified. These challenges should be addressed before quantum computing can be meaningfully deployed in operational power systems.
Finally, recognizing the enduring relevance of classical solvers, this PhD thesis contributes a reinforcement learning (RL) framework that enhances NR convergence reliability through learned initialization strategies. By integrating a quantum-enhanced environment update mechanism, the thesis further explores how quantum optimization algorithms can accelerate learning and improve robustness without replacing well-established numerical methods.
In summary, this PhD thesis contributes new scientific knowledge by reframing PF and OPF through the lenses of scientific machine learning, quantum optimization algorithms, and hybrid computational architectures. It demonstrates that AI and QC are not merely faster solvers but fundamentally different tools that enable new representations, formulations, and system-level integrations. These contributions lay a conceptual and methodological foundation for next-generation power system modeling and control, with implications extending beyond power systems to other nonlinear engineering domains. ...
This PhD thesis addresses a fundamental challenge in power system applications: how to reliably model and control modern power systems under increasing uncertainty and operational constraints. As modern power systems evolve toward highly distributed, renewable-rich infrastructures, classical numerical methods for power flow (PF) and optimal power flow (OPF) face intrinsic limitations. These limitations are not merely computational but structural: nonlinear solvers, such as Newton–Raphson (NR), exhibit sensitivity to initialization, numerical instability in ill-conditioned scenarios, and scaling behavior that is increasingly incompatible with the dimensionality requirements of future power system operation. Despite decades of algorithmic refinement, these challenges remain only partially resolved within classical computation.
The thesis systematically investigates advanced computational paradigms, i.e., artificial intelligence (AI) and quantum computing (QC), as complementary tools for overcoming these fundamental challenges. Rather than proposing incremental solver modifications, the thesis develops new conceptual formulations for PF and OPF that rethink how physical structure, optimization, and computation are represented. First, a physics-informed neural network (PINN) framework is introduced that partially encodes power balance equations directly into the learning process. This contribution advances existing PINN approaches by reducing dependence on complete network topology and line parameters while preserving physical interpretability. Second, the thesis explores quantum neural networks (QNNs) as a novel learning paradigm for PF analysis under data scarcity and noise, offering new insights into how quantum representations in the fault-tolerant quantum computing era can provide compact, potentially more expressive alternatives to classical deep learning.
Beyond learning-based methods, this PhD thesis provides a foundational contribution by reformulating PF and OPF as quadratic unconstrained binary optimization (QUBO) and Ising models, enabling their solution using quantum and quantum-inspired hardware. This reformulation constitutes a conceptual shift from continuous, nonlinear equation solving to discrete energy minimization, revealing new pathways to address scalability. The thesis further contributes to the first systematic comparison of gate-based and adiabatic quantum computing paradigms for PF and OPF across system sizes and operating conditions.
To bridge the gap between theoretical quantum algorithms and real-world power system requirements, this PhD thesis further introduces a quantum hardware-in-the-loop (QHIL) framework. This framework enables the evaluation of quantum solvers within closed-loop, real-time power system simulators. Therefore, critical system-level challenges, such as communication latency, embedding overhead, and real-time feasibility, are identified. These challenges should be addressed before quantum computing can be meaningfully deployed in operational power systems.
Finally, recognizing the enduring relevance of classical solvers, this PhD thesis contributes a reinforcement learning (RL) framework that enhances NR convergence reliability through learned initialization strategies. By integrating a quantum-enhanced environment update mechanism, the thesis further explores how quantum optimization algorithms can accelerate learning and improve robustness without replacing well-established numerical methods.
In summary, this PhD thesis contributes new scientific knowledge by reframing PF and OPF through the lenses of scientific machine learning, quantum optimization algorithms, and hybrid computational architectures. It demonstrates that AI and QC are not merely faster solvers but fundamentally different tools that enable new representations, formulations, and system-level integrations. These contributions lay a conceptual and methodological foundation for next-generation power system modeling and control, with implications extending beyond power systems to other nonlinear engineering domains.
The thesis systematically investigates advanced computational paradigms, i.e., artificial intelligence (AI) and quantum computing (QC), as complementary tools for overcoming these fundamental challenges. Rather than proposing incremental solver modifications, the thesis develops new conceptual formulations for PF and OPF that rethink how physical structure, optimization, and computation are represented. First, a physics-informed neural network (PINN) framework is introduced that partially encodes power balance equations directly into the learning process. This contribution advances existing PINN approaches by reducing dependence on complete network topology and line parameters while preserving physical interpretability. Second, the thesis explores quantum neural networks (QNNs) as a novel learning paradigm for PF analysis under data scarcity and noise, offering new insights into how quantum representations in the fault-tolerant quantum computing era can provide compact, potentially more expressive alternatives to classical deep learning.
Beyond learning-based methods, this PhD thesis provides a foundational contribution by reformulating PF and OPF as quadratic unconstrained binary optimization (QUBO) and Ising models, enabling their solution using quantum and quantum-inspired hardware. This reformulation constitutes a conceptual shift from continuous, nonlinear equation solving to discrete energy minimization, revealing new pathways to address scalability. The thesis further contributes to the first systematic comparison of gate-based and adiabatic quantum computing paradigms for PF and OPF across system sizes and operating conditions.
To bridge the gap between theoretical quantum algorithms and real-world power system requirements, this PhD thesis further introduces a quantum hardware-in-the-loop (QHIL) framework. This framework enables the evaluation of quantum solvers within closed-loop, real-time power system simulators. Therefore, critical system-level challenges, such as communication latency, embedding overhead, and real-time feasibility, are identified. These challenges should be addressed before quantum computing can be meaningfully deployed in operational power systems.
Finally, recognizing the enduring relevance of classical solvers, this PhD thesis contributes a reinforcement learning (RL) framework that enhances NR convergence reliability through learned initialization strategies. By integrating a quantum-enhanced environment update mechanism, the thesis further explores how quantum optimization algorithms can accelerate learning and improve robustness without replacing well-established numerical methods.
In summary, this PhD thesis contributes new scientific knowledge by reframing PF and OPF through the lenses of scientific machine learning, quantum optimization algorithms, and hybrid computational architectures. It demonstrates that AI and QC are not merely faster solvers but fundamentally different tools that enable new representations, formulations, and system-level integrations. These contributions lay a conceptual and methodological foundation for next-generation power system modeling and control, with implications extending beyond power systems to other nonlinear engineering domains.
We introduce a physics-informed neural network for power flow (PINN4PF) that effectively captures the nonlinear dynamics of large-scale modern power systems. The proposed neural network (NN) architecture consists of two important advancements in the training pipeline: (A) a double-head feed-forward NN that aligns with power flow (PF), including an activation function that adjusts to the net active and reactive power injections patterns, and (B) a physics-based loss function that partially incorporates power system topology information through a novel hidden function. The effectiveness of the proposed architecture is illustrated through 4-bus, 15-bus, 290-bus, and 2224-bus test systems and is evaluated against two baselines: a linear regression model (LR) and a black-box NN (MLP). The comparison is based on (i) generalization ability, (ii) robustness, (iii) impact of training dataset size on generalization ability, (iv) accuracy in approximating derived PF quantities (specifically line current, line active power, and line reactive power), and (v) scalability. Results demonstrate that PINN4PF outperforms both baselines across all test systems by up to two orders of magnitude, not only in terms of direct criteria, e.g., generalization ability, but also in terms of derived physical quantities.
...
We introduce a physics-informed neural network for power flow (PINN4PF) that effectively captures the nonlinear dynamics of large-scale modern power systems. The proposed neural network (NN) architecture consists of two important advancements in the training pipeline: (A) a double-head feed-forward NN that aligns with power flow (PF), including an activation function that adjusts to the net active and reactive power injections patterns, and (B) a physics-based loss function that partially incorporates power system topology information through a novel hidden function. The effectiveness of the proposed architecture is illustrated through 4-bus, 15-bus, 290-bus, and 2224-bus test systems and is evaluated against two baselines: a linear regression model (LR) and a black-box NN (MLP). The comparison is based on (i) generalization ability, (ii) robustness, (iii) impact of training dataset size on generalization ability, (iv) accuracy in approximating derived PF quantities (specifically line current, line active power, and line reactive power), and (v) scalability. Results demonstrate that PINN4PF outperforms both baselines across all test systems by up to two orders of magnitude, not only in terms of direct criteria, e.g., generalization ability, but also in terms of derived physical quantities.
Conference paper
(2025)
-
Zeynab Kaseb, Matthias Möller, Markus Kirsch, Peter Palensky, Pedro P. Vergara
We investigate the performance of different annealers for power flow analysis. The annealers are D-Wave's simulated annealer Neal, D-Wave's quantum-classical hybrid annealer, D-Wave's Advantage™ system (QA), Fujitsu's simulated annealer, and Fujitsu's digital annealer V3 (DA). We implement Quadratic Unconstrained Binary Optimization (QUBO) and Ising model formulations, with the latter offering finer control over complex voltage adjustments. We also propose a new method for the iterative and adaptive updating of complex voltages for the QUBO and Ising model formulations. Six different test systems are experimented with. QA and DA show superior performance over classical annealers. DA effectively manages the 15-bus system, whereas QA encounters difficulties in embedding the problem graph onto the hardware graph because of the limited qubit connectivity. This constraint confines QA to the 14-bus system with the QUBO formulation and the 4-bus system with the Ising model formulation. The best performance is associated with different annealers across different test systems, which suggests that adjusting the threshold can improve precision if the compiler and annealer are capable of handling the variables involved.
...
We investigate the performance of different annealers for power flow analysis. The annealers are D-Wave's simulated annealer Neal, D-Wave's quantum-classical hybrid annealer, D-Wave's Advantage™ system (QA), Fujitsu's simulated annealer, and Fujitsu's digital annealer V3 (DA). We implement Quadratic Unconstrained Binary Optimization (QUBO) and Ising model formulations, with the latter offering finer control over complex voltage adjustments. We also propose a new method for the iterative and adaptive updating of complex voltages for the QUBO and Ising model formulations. Six different test systems are experimented with. QA and DA show superior performance over classical annealers. DA effectively manages the 15-bus system, whereas QA encounters difficulties in embedding the problem graph onto the hardware graph because of the limited qubit connectivity. This constraint confines QA to the 14-bus system with the QUBO formulation and the 4-bus system with the Ising model formulation. The best performance is associated with different annealers across different test systems, which suggests that adjusting the threshold can improve precision if the compiler and annealer are capable of handling the variables involved.
This paper investigates the impact of adaptive activation functions on deep learning-based power flow analysis. Specifically, it compares four adaptive activation functions with state-of-the-art activation functions, i.e., ReLU, LeakyReLU, Sigmoid, and Tanh, in terms of loss function error, convergence speed, and learning process stability, using a real-world distribution network dataset. Results indicate that the proposed adaptive activation functions improve learning capability over state-of-the-art activation functions. Notably, adaptive ReLU activation shows the most improved learning process, with convergence speed up to twice as fast as ReLU. Adaptive Sigmoid and Tanh activation functions exhibit superior performance in terms of loss function error, outperforming ReLU and LeakyReLU by up to two orders of magnitude. Furthermore, the proposed adaptive activation functions lead to smoother and more stable learning processes, especially during early training, improving convergence. The practical implications of this study include the potential application of these adaptive activation functions in distribution network modeling, forecasting, and control, leading to more accurate and reliable power system operation.
...
This paper investigates the impact of adaptive activation functions on deep learning-based power flow analysis. Specifically, it compares four adaptive activation functions with state-of-the-art activation functions, i.e., ReLU, LeakyReLU, Sigmoid, and Tanh, in terms of loss function error, convergence speed, and learning process stability, using a real-world distribution network dataset. Results indicate that the proposed adaptive activation functions improve learning capability over state-of-the-art activation functions. Notably, adaptive ReLU activation shows the most improved learning process, with convergence speed up to twice as fast as ReLU. Adaptive Sigmoid and Tanh activation functions exhibit superior performance in terms of loss function error, outperforming ReLU and LeakyReLU by up to two orders of magnitude. Furthermore, the proposed adaptive activation functions lead to smoother and more stable learning processes, especially during early training, improving convergence. The practical implications of this study include the potential application of these adaptive activation functions in distribution network modeling, forecasting, and control, leading to more accurate and reliable power system operation.
Conference paper
(2024)
-
Zeynab Kaseb, Matthias Möller, Markus Kirsch, Peter Palensky, Pedro P. Vergara
We investigate the performance of different annealers for power flow analysis using adiabatic computing. The annealers include D-Wave's simulated annealer Neal, D-Wave's quantum-classical hybrid annealer, D-Wave's Advantages system (QA), Fujitsu's classical simulated annealer, and Fujitsu's digital annealer V3 (DA). We implement Quadratic Unconstrained Binary Optimization (QUBO) and Ising model formulations, with the latter offering finer control over complex voltage adjustments. Different test systems are experimented with to systematically evaluate the annealers. The evaluation is based on the accuracy, the annealer's capability to handle the decision variables, and the computational time needed. QA and DA show superior performance over classical annealers for our application. DA effectively manages larger test systems, whereas QA encounters difficulties embedding the problem graph onto the hardware graph because of the limited qubit connectivity. This constraint confines QA to the 14-bus system with the QUBO formulation and the 4-bus system with the Ising model formulation. The best performance is associated with different annealers across different test systems, which suggests that adjusting the threshold can improve precision if the compiler and annealer are capable of handling the number of variables involved.
...
We investigate the performance of different annealers for power flow analysis using adiabatic computing. The annealers include D-Wave's simulated annealer Neal, D-Wave's quantum-classical hybrid annealer, D-Wave's Advantages system (QA), Fujitsu's classical simulated annealer, and Fujitsu's digital annealer V3 (DA). We implement Quadratic Unconstrained Binary Optimization (QUBO) and Ising model formulations, with the latter offering finer control over complex voltage adjustments. Different test systems are experimented with to systematically evaluate the annealers. The evaluation is based on the accuracy, the annealer's capability to handle the decision variables, and the computational time needed. QA and DA show superior performance over classical annealers for our application. DA effectively manages larger test systems, whereas QA encounters difficulties embedding the problem graph onto the hardware graph because of the limited qubit connectivity. This constraint confines QA to the 14-bus system with the QUBO formulation and the 4-bus system with the Ising model formulation. The best performance is associated with different annealers across different test systems, which suggests that adjusting the threshold can improve precision if the compiler and annealer are capable of handling the number of variables involved.
Journal article
(2024)
-
Zeynab Kaseb, Matthias Möller, Giorgio Tosti Balducci, Peter Palensky, Pedro P. Vergara
This paper explores the potential application of quantum and hybrid quantum–classical neural networks in power flow analysis. Experiments are conducted using two datasets based on 4-bus and 33-bus test systems. A systematic performance comparison is also conducted among quantum, hybrid quantum–classical, and classical neural networks. The comparison is based on (i) generalization ability, (ii) robustness, (iii) training dataset size needed, (iv) training error, and (v) training process stability. The results show that the developed hybrid quantum–classical neural network outperforms both quantum and classical neural networks, and hence can improve deep learning-based power flow analysis in the noisy-intermediate-scale quantum (NISQ) and fault-tolerant quantum (FTQ) era.
...
This paper explores the potential application of quantum and hybrid quantum–classical neural networks in power flow analysis. Experiments are conducted using two datasets based on 4-bus and 33-bus test systems. A systematic performance comparison is also conducted among quantum, hybrid quantum–classical, and classical neural networks. The comparison is based on (i) generalization ability, (ii) robustness, (iii) training dataset size needed, (iv) training error, and (v) training process stability. The results show that the developed hybrid quantum–classical neural network outperforms both quantum and classical neural networks, and hence can improve deep learning-based power flow analysis in the noisy-intermediate-scale quantum (NISQ) and fault-tolerant quantum (FTQ) era.
Power flow analysis using quantum and digital annealers
A discrete combinatorial optimization approach
Power flow (PF) analysis is a foundational computational method to study the flow of power in an electrical network. This analysis involves solving a set of non-linear and non-convex differential-algebraic equations. State-of-the-art solvers for PF analysis, therefore, face challenges with scalability and convergence, specifically for large-scale and/or ill-conditioned cases characterized by high penetration of renewable energy sources, among others. The adiabatic quantum computing paradigm has been proven to efficiently find solutions for combinatorial problems in the noisy intermediate-scale quantum (NISQ) era, and it can potentially address the limitations posed by state-of-the-art PF solvers. For the first time, we propose a novel adiabatic quantum computing approach for efficient PF analysis. Our key contributions are (i) a combinatorial PF algorithm and a modified version that aligns with the principles of PF analysis, termed the adiabatic quantum PF algorithm (AQPF), both of which use Quadratic Unconstrained Binary Optimization (QUBO) and Ising model formulations; (ii) a scalability study of the AQPF algorithm; and (iii) an extension of the AQPF algorithm to handle larger problem sizes using a partitioned approach. Numerical experiments are conducted using different test system sizes on D-Wave’s Advantage™ quantum annealer, Fujitsu’s digital annealer V3, D-Wave’s quantum-classical hybrid annealer, and two simulated annealers running on classical computer hardware. The reported results demonstrate the effectiveness and high accuracy of the proposed AQPF algorithm and its potential to speed up the PF analysis process while handling ill-conditioned cases using quantum and quantum-inspired algorithms.
...
Power flow (PF) analysis is a foundational computational method to study the flow of power in an electrical network. This analysis involves solving a set of non-linear and non-convex differential-algebraic equations. State-of-the-art solvers for PF analysis, therefore, face challenges with scalability and convergence, specifically for large-scale and/or ill-conditioned cases characterized by high penetration of renewable energy sources, among others. The adiabatic quantum computing paradigm has been proven to efficiently find solutions for combinatorial problems in the noisy intermediate-scale quantum (NISQ) era, and it can potentially address the limitations posed by state-of-the-art PF solvers. For the first time, we propose a novel adiabatic quantum computing approach for efficient PF analysis. Our key contributions are (i) a combinatorial PF algorithm and a modified version that aligns with the principles of PF analysis, termed the adiabatic quantum PF algorithm (AQPF), both of which use Quadratic Unconstrained Binary Optimization (QUBO) and Ising model formulations; (ii) a scalability study of the AQPF algorithm; and (iii) an extension of the AQPF algorithm to handle larger problem sizes using a partitioned approach. Numerical experiments are conducted using different test system sizes on D-Wave’s Advantage™ quantum annealer, Fujitsu’s digital annealer V3, D-Wave’s quantum-classical hybrid annealer, and two simulated annealers running on classical computer hardware. The reported results demonstrate the effectiveness and high accuracy of the proposed AQPF algorithm and its potential to speed up the PF analysis process while handling ill-conditioned cases using quantum and quantum-inspired algorithms.
Towards CFD-based optimization of urban wind conditions
Comparison of Genetic algorithm, Particle Swarm Optimization, and a hybrid algorithm
Journal article
(2022)
-
Z. Kaseb, Morteza Rahbar
Urban morphology can significantly impact urban wind conditions. Finding an optimum morphology to improve the wind conditions for a given urban area can be very challenging as it depends on a wide range of parameters. In this perspective, meta-heuristic algorithms can be useful to reach/approximate optimum solutions. While the satisfactory performance of meta-heuristic algorithms has been shown for different complex engineering problems, a detailed evaluation of these algorithms has not yet been performed for urban wind conditions. Therefore, this study aims to systematically evaluate the performance of meta-heuristic algorithms for CFD-based optimization of urban wind conditions at street scale. Three algorithms are considered: (i) Genetic algorithm (GA), (ii) Particle Swarm Optimization (PSO), and (iii) a hybrid algorithm of PSO and GA. The focus is on a compact generic urban area, while the height of the involved buildings is considered as the optimization variable. In total, 714 high-resolution 3D steady Reynolds-averaged Navier-Stokes (RANS) CFD simulations are performed in combination with the standard k-ε turbulence model. The results show that the hybrid algorithm is superior as it can improve the wind conditions by about 425% and 100%, compared with GA and PSO, respectively.
...
Urban morphology can significantly impact urban wind conditions. Finding an optimum morphology to improve the wind conditions for a given urban area can be very challenging as it depends on a wide range of parameters. In this perspective, meta-heuristic algorithms can be useful to reach/approximate optimum solutions. While the satisfactory performance of meta-heuristic algorithms has been shown for different complex engineering problems, a detailed evaluation of these algorithms has not yet been performed for urban wind conditions. Therefore, this study aims to systematically evaluate the performance of meta-heuristic algorithms for CFD-based optimization of urban wind conditions at street scale. Three algorithms are considered: (i) Genetic algorithm (GA), (ii) Particle Swarm Optimization (PSO), and (iii) a hybrid algorithm of PSO and GA. The focus is on a compact generic urban area, while the height of the involved buildings is considered as the optimization variable. In total, 714 high-resolution 3D steady Reynolds-averaged Navier-Stokes (RANS) CFD simulations are performed in combination with the standard k-ε turbulence model. The results show that the hybrid algorithm is superior as it can improve the wind conditions by about 425% and 100%, compared with GA and PSO, respectively.
Data-driven optimization of building-integrated ducted openings for wind energy harvesting
Sensitivity analysis of metamodels
Metamodels are developed and used for aerodynamic optimization of a ducted opening integrated into a high-rise building to maximize the amplification factor within the duct. The duct consists of a nozzle, a throat, and a diffuser. 211 high-resolution 3D RANS CFD simulations are performed to generate training and testing datasets. The space-filling design and Genetic algorithm are used for data sampling and optimization, respectively. The performance of five commonly-used metamodels is systematically investigated: Response Surface Methodology (RSM), Kriging (KG), Neural Network (NN), Support Vector Regression (SVR), and Genetic Aggregation Response Surface (GARS). The investigation is based on (i) detailed in-sample and out-of-sample evaluations of the metamodels, (ii) annual available power in the wind (Pavailable), and (iii) annual energy production (AEP) for a 3-bladed horizontal-axis wind turbine (HAWT) installed in the mid-throat for the optimum designs obtained by the metamodels. The results show that converging-diverging ducted openings can magnify the experienced wind speed by the turbine and enhance the available wind power. In addition, the use of different metamodels can lead to a variation of up to 153% in the estimated Pavailable. For a small dataset, crude yet still acceptable accuracy can be achieved for Genetic Aggregation Response Surface and Kriging at a very low computational time.
...
Metamodels are developed and used for aerodynamic optimization of a ducted opening integrated into a high-rise building to maximize the amplification factor within the duct. The duct consists of a nozzle, a throat, and a diffuser. 211 high-resolution 3D RANS CFD simulations are performed to generate training and testing datasets. The space-filling design and Genetic algorithm are used for data sampling and optimization, respectively. The performance of five commonly-used metamodels is systematically investigated: Response Surface Methodology (RSM), Kriging (KG), Neural Network (NN), Support Vector Regression (SVR), and Genetic Aggregation Response Surface (GARS). The investigation is based on (i) detailed in-sample and out-of-sample evaluations of the metamodels, (ii) annual available power in the wind (Pavailable), and (iii) annual energy production (AEP) for a 3-bladed horizontal-axis wind turbine (HAWT) installed in the mid-throat for the optimum designs obtained by the metamodels. The results show that converging-diverging ducted openings can magnify the experienced wind speed by the turbine and enhance the available wind power. In addition, the use of different metamodels can lead to a variation of up to 153% in the estimated Pavailable. For a small dataset, crude yet still acceptable accuracy can be achieved for Genetic Aggregation Response Surface and Kriging at a very low computational time.
Journal article
(2020)
-
Z. Kaseb, Mohammadreza Hafezi, Mansoureh Tahbaz, Shahram Delfani