L.M. Schubotz
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>
2 records found
1
Vehicle-to-Grid (V2G) can provide flexibility to electricity systems, but its practical contribution to congestion mitigation depends not only on technical battery capacity, but also on user participation, vehicle availability, charging behaviour, and operational control. Existing research often treats participation as fixed or evaluates user acceptance separately from system performance. This study investigates which V2G system configurations are associated with lower local congestion pressure when participation emerges endogenously from heterogeneous contract preferences and behavioural feedback.
A full-year agent-based model of a stylised Dutch urban area was used to simulate the mobility, charging, contract acceptance, and V2G behaviour of 500 electric-vehicle drivers at five-minute resolution. Contract characteristics, charging policies, charging infrastructure, and renewable-energy conditions were varied across a large discrete design space. Congestion pressure was represented by a squared loss function penalising positive net-load exceedances above 600 kW. A learning-guided inverse-modelling procedure, based on a Bayesian Ridge surrogate and two independently seeded learning streams, directed simulations towards potentially favourable configurations. Each proposed configuration was evaluated through repeated stochastic simulations, and Patient Rule Induction Method was used to identify interpretable low-loss design regions.
The results show that participation is an enabling, but insufficient, condition for congestion relief. Higher participation was generally associated with lower loss, with the largest improvements occurring between low and medium-high participation levels and diminishing returns emerging at high participation. Guaranteed range was the strongest predictor of participation, followed by minimum plug-in time, while financial compensation and battery-degradation assumptions had smaller effects. Low congestion loss was most consistently associated with smart charging at residential locations, high renewable-energy coverage, and a medium-to-high guaranteed range. Smart charging and V2G performed complementary functions: controlled charging prevented avoidable EV demand from reinforcing existing peaks, while V2G reduced the remaining positive net load. Nevertheless, congestion remained concentrated in winter evenings, and a limited number of severe peaks persisted.
The inverse search identified favourable configurations relatively early and subsequently improved mainly by avoiding poor regions rather than converging towards a unique optimum. This study contributes by linking heterogeneous contract participation to realised, temporally available grid flexibility and by demonstrating how inverse modelling can identify multiple plausible socio-technical design regions. The findings indicate that effective V2G systems require the joint design of user contracts, charging control, renewable-energy conditions, and grid-responsive operation, rather than the maximisation of participation alone. ...
A full-year agent-based model of a stylised Dutch urban area was used to simulate the mobility, charging, contract acceptance, and V2G behaviour of 500 electric-vehicle drivers at five-minute resolution. Contract characteristics, charging policies, charging infrastructure, and renewable-energy conditions were varied across a large discrete design space. Congestion pressure was represented by a squared loss function penalising positive net-load exceedances above 600 kW. A learning-guided inverse-modelling procedure, based on a Bayesian Ridge surrogate and two independently seeded learning streams, directed simulations towards potentially favourable configurations. Each proposed configuration was evaluated through repeated stochastic simulations, and Patient Rule Induction Method was used to identify interpretable low-loss design regions.
The results show that participation is an enabling, but insufficient, condition for congestion relief. Higher participation was generally associated with lower loss, with the largest improvements occurring between low and medium-high participation levels and diminishing returns emerging at high participation. Guaranteed range was the strongest predictor of participation, followed by minimum plug-in time, while financial compensation and battery-degradation assumptions had smaller effects. Low congestion loss was most consistently associated with smart charging at residential locations, high renewable-energy coverage, and a medium-to-high guaranteed range. Smart charging and V2G performed complementary functions: controlled charging prevented avoidable EV demand from reinforcing existing peaks, while V2G reduced the remaining positive net load. Nevertheless, congestion remained concentrated in winter evenings, and a limited number of severe peaks persisted.
The inverse search identified favourable configurations relatively early and subsequently improved mainly by avoiding poor regions rather than converging towards a unique optimum. This study contributes by linking heterogeneous contract participation to realised, temporally available grid flexibility and by demonstrating how inverse modelling can identify multiple plausible socio-technical design regions. The findings indicate that effective V2G systems require the joint design of user contracts, charging control, renewable-energy conditions, and grid-responsive operation, rather than the maximisation of participation alone. ...
Vehicle-to-Grid (V2G) can provide flexibility to electricity systems, but its practical contribution to congestion mitigation depends not only on technical battery capacity, but also on user participation, vehicle availability, charging behaviour, and operational control. Existing research often treats participation as fixed or evaluates user acceptance separately from system performance. This study investigates which V2G system configurations are associated with lower local congestion pressure when participation emerges endogenously from heterogeneous contract preferences and behavioural feedback.
A full-year agent-based model of a stylised Dutch urban area was used to simulate the mobility, charging, contract acceptance, and V2G behaviour of 500 electric-vehicle drivers at five-minute resolution. Contract characteristics, charging policies, charging infrastructure, and renewable-energy conditions were varied across a large discrete design space. Congestion pressure was represented by a squared loss function penalising positive net-load exceedances above 600 kW. A learning-guided inverse-modelling procedure, based on a Bayesian Ridge surrogate and two independently seeded learning streams, directed simulations towards potentially favourable configurations. Each proposed configuration was evaluated through repeated stochastic simulations, and Patient Rule Induction Method was used to identify interpretable low-loss design regions.
The results show that participation is an enabling, but insufficient, condition for congestion relief. Higher participation was generally associated with lower loss, with the largest improvements occurring between low and medium-high participation levels and diminishing returns emerging at high participation. Guaranteed range was the strongest predictor of participation, followed by minimum plug-in time, while financial compensation and battery-degradation assumptions had smaller effects. Low congestion loss was most consistently associated with smart charging at residential locations, high renewable-energy coverage, and a medium-to-high guaranteed range. Smart charging and V2G performed complementary functions: controlled charging prevented avoidable EV demand from reinforcing existing peaks, while V2G reduced the remaining positive net load. Nevertheless, congestion remained concentrated in winter evenings, and a limited number of severe peaks persisted.
The inverse search identified favourable configurations relatively early and subsequently improved mainly by avoiding poor regions rather than converging towards a unique optimum. This study contributes by linking heterogeneous contract participation to realised, temporally available grid flexibility and by demonstrating how inverse modelling can identify multiple plausible socio-technical design regions. The findings indicate that effective V2G systems require the joint design of user contracts, charging control, renewable-energy conditions, and grid-responsive operation, rather than the maximisation of participation alone.
A full-year agent-based model of a stylised Dutch urban area was used to simulate the mobility, charging, contract acceptance, and V2G behaviour of 500 electric-vehicle drivers at five-minute resolution. Contract characteristics, charging policies, charging infrastructure, and renewable-energy conditions were varied across a large discrete design space. Congestion pressure was represented by a squared loss function penalising positive net-load exceedances above 600 kW. A learning-guided inverse-modelling procedure, based on a Bayesian Ridge surrogate and two independently seeded learning streams, directed simulations towards potentially favourable configurations. Each proposed configuration was evaluated through repeated stochastic simulations, and Patient Rule Induction Method was used to identify interpretable low-loss design regions.
The results show that participation is an enabling, but insufficient, condition for congestion relief. Higher participation was generally associated with lower loss, with the largest improvements occurring between low and medium-high participation levels and diminishing returns emerging at high participation. Guaranteed range was the strongest predictor of participation, followed by minimum plug-in time, while financial compensation and battery-degradation assumptions had smaller effects. Low congestion loss was most consistently associated with smart charging at residential locations, high renewable-energy coverage, and a medium-to-high guaranteed range. Smart charging and V2G performed complementary functions: controlled charging prevented avoidable EV demand from reinforcing existing peaks, while V2G reduced the remaining positive net load. Nevertheless, congestion remained concentrated in winter evenings, and a limited number of severe peaks persisted.
The inverse search identified favourable configurations relatively early and subsequently improved mainly by avoiding poor regions rather than converging towards a unique optimum. This study contributes by linking heterogeneous contract participation to realised, temporally available grid flexibility and by demonstrating how inverse modelling can identify multiple plausible socio-technical design regions. The findings indicate that effective V2G systems require the joint design of user contracts, charging control, renewable-energy conditions, and grid-responsive operation, rather than the maximisation of participation alone.
The transition to renewable energy is vital for the Netherlands to meet the National Climate Agreement's goal of a 95% reduction in CO2 emissions by 2050. Residential energy, accounting for 20% of the nation's total, is a key area for local municipalities, which are crucial in implementing decarbonisation measures. Solar photovoltaics (PV) have rapidly emerged as a leading technology, achieving the government's target of 7 TWh from small onshore projects by 2022, eight years ahead of schedule.
However, the rapid adoption of solar PV presents several challenges for policymakers, particularly at the municipal level. These challenges include grid integration, financing, and ensuring equitable access to solar technology. Addressing these issues requires a deep understanding of the factors influencing solar PV adoption and the dynamics at play within different municipalities. Inverse modelling presents a methodology to explore these complexities. This technique involves working backwards from observed outcomes to identify and understand the underlying processes and parameters that generated those results. The application of inverse modelling is, however, very novel. Therefore, this thesis aims to lay the groundwork for future applications of inverse modelling by exploring the factors that influence solar PV adoption in Dutch municipalities. To achieve this goal, the following research question has been formulated:
How can inverse modelling contribute to uncovering plausible explanations for residential solar PV adoption dynamics in Dutch municipalities?
The study employs inverse modelling combined with agent-based modelling to analyse adoption patterns across different municipalities, utilising Random Search and Bayesian Search algorithms. Findings indicate that social factors are often more influential than economic incentives in driving solar PV adoption, emphasising the role of community engagement and peer effects.
These insights underscore the potential of inverse modelling to reveal the complex interplay of factors influencing solar PV adoption. This study demonstrates that inverse modelling can effectively identify patterns and key determinants within municipal data. The research suggests that future efforts should refine this approach, considering multiple agent-based models and alternative methodologies to enhance robustness and accuracy, paving the way for more comprehensive and accurate research in the future.
In conclusion, inverse modelling provides a powerful tool for uncovering the dynamics of solar PV adoption in Dutch municipalities by providing a systematic approach to understanding the underlying factors of the system and its intentions. Through IM, the key determinants that drive solar PV adoption can be inferred by analysing observed patterns and behaviours within the municipalities. This approach is additionally powerful because it does not rely solely on pre-existing theories or assumptions, as with forward modelling. Instead, it derives insights directly from the data.
...
However, the rapid adoption of solar PV presents several challenges for policymakers, particularly at the municipal level. These challenges include grid integration, financing, and ensuring equitable access to solar technology. Addressing these issues requires a deep understanding of the factors influencing solar PV adoption and the dynamics at play within different municipalities. Inverse modelling presents a methodology to explore these complexities. This technique involves working backwards from observed outcomes to identify and understand the underlying processes and parameters that generated those results. The application of inverse modelling is, however, very novel. Therefore, this thesis aims to lay the groundwork for future applications of inverse modelling by exploring the factors that influence solar PV adoption in Dutch municipalities. To achieve this goal, the following research question has been formulated:
How can inverse modelling contribute to uncovering plausible explanations for residential solar PV adoption dynamics in Dutch municipalities?
The study employs inverse modelling combined with agent-based modelling to analyse adoption patterns across different municipalities, utilising Random Search and Bayesian Search algorithms. Findings indicate that social factors are often more influential than economic incentives in driving solar PV adoption, emphasising the role of community engagement and peer effects.
These insights underscore the potential of inverse modelling to reveal the complex interplay of factors influencing solar PV adoption. This study demonstrates that inverse modelling can effectively identify patterns and key determinants within municipal data. The research suggests that future efforts should refine this approach, considering multiple agent-based models and alternative methodologies to enhance robustness and accuracy, paving the way for more comprehensive and accurate research in the future.
In conclusion, inverse modelling provides a powerful tool for uncovering the dynamics of solar PV adoption in Dutch municipalities by providing a systematic approach to understanding the underlying factors of the system and its intentions. Through IM, the key determinants that drive solar PV adoption can be inferred by analysing observed patterns and behaviours within the municipalities. This approach is additionally powerful because it does not rely solely on pre-existing theories or assumptions, as with forward modelling. Instead, it derives insights directly from the data.
...
The transition to renewable energy is vital for the Netherlands to meet the National Climate Agreement's goal of a 95% reduction in CO2 emissions by 2050. Residential energy, accounting for 20% of the nation's total, is a key area for local municipalities, which are crucial in implementing decarbonisation measures. Solar photovoltaics (PV) have rapidly emerged as a leading technology, achieving the government's target of 7 TWh from small onshore projects by 2022, eight years ahead of schedule.
However, the rapid adoption of solar PV presents several challenges for policymakers, particularly at the municipal level. These challenges include grid integration, financing, and ensuring equitable access to solar technology. Addressing these issues requires a deep understanding of the factors influencing solar PV adoption and the dynamics at play within different municipalities. Inverse modelling presents a methodology to explore these complexities. This technique involves working backwards from observed outcomes to identify and understand the underlying processes and parameters that generated those results. The application of inverse modelling is, however, very novel. Therefore, this thesis aims to lay the groundwork for future applications of inverse modelling by exploring the factors that influence solar PV adoption in Dutch municipalities. To achieve this goal, the following research question has been formulated:
How can inverse modelling contribute to uncovering plausible explanations for residential solar PV adoption dynamics in Dutch municipalities?
The study employs inverse modelling combined with agent-based modelling to analyse adoption patterns across different municipalities, utilising Random Search and Bayesian Search algorithms. Findings indicate that social factors are often more influential than economic incentives in driving solar PV adoption, emphasising the role of community engagement and peer effects.
These insights underscore the potential of inverse modelling to reveal the complex interplay of factors influencing solar PV adoption. This study demonstrates that inverse modelling can effectively identify patterns and key determinants within municipal data. The research suggests that future efforts should refine this approach, considering multiple agent-based models and alternative methodologies to enhance robustness and accuracy, paving the way for more comprehensive and accurate research in the future.
In conclusion, inverse modelling provides a powerful tool for uncovering the dynamics of solar PV adoption in Dutch municipalities by providing a systematic approach to understanding the underlying factors of the system and its intentions. Through IM, the key determinants that drive solar PV adoption can be inferred by analysing observed patterns and behaviours within the municipalities. This approach is additionally powerful because it does not rely solely on pre-existing theories or assumptions, as with forward modelling. Instead, it derives insights directly from the data.
However, the rapid adoption of solar PV presents several challenges for policymakers, particularly at the municipal level. These challenges include grid integration, financing, and ensuring equitable access to solar technology. Addressing these issues requires a deep understanding of the factors influencing solar PV adoption and the dynamics at play within different municipalities. Inverse modelling presents a methodology to explore these complexities. This technique involves working backwards from observed outcomes to identify and understand the underlying processes and parameters that generated those results. The application of inverse modelling is, however, very novel. Therefore, this thesis aims to lay the groundwork for future applications of inverse modelling by exploring the factors that influence solar PV adoption in Dutch municipalities. To achieve this goal, the following research question has been formulated:
How can inverse modelling contribute to uncovering plausible explanations for residential solar PV adoption dynamics in Dutch municipalities?
The study employs inverse modelling combined with agent-based modelling to analyse adoption patterns across different municipalities, utilising Random Search and Bayesian Search algorithms. Findings indicate that social factors are often more influential than economic incentives in driving solar PV adoption, emphasising the role of community engagement and peer effects.
These insights underscore the potential of inverse modelling to reveal the complex interplay of factors influencing solar PV adoption. This study demonstrates that inverse modelling can effectively identify patterns and key determinants within municipal data. The research suggests that future efforts should refine this approach, considering multiple agent-based models and alternative methodologies to enhance robustness and accuracy, paving the way for more comprehensive and accurate research in the future.
In conclusion, inverse modelling provides a powerful tool for uncovering the dynamics of solar PV adoption in Dutch municipalities by providing a systematic approach to understanding the underlying factors of the system and its intentions. Through IM, the key determinants that drive solar PV adoption can be inferred by analysing observed patterns and behaviours within the municipalities. This approach is additionally powerful because it does not rely solely on pre-existing theories or assumptions, as with forward modelling. Instead, it derives insights directly from the data.