Analyse van de integratiecapaciteit in passieve en actieve distributienetwerken

The large-scale electrification of energy end-use sectors is accelerating the deployment of distributed energy resources (DERs) in low-voltage distribution networks. Among these resources, electric vehicles (EVs) and photovoltaic (PV) systems are expected to dominate future residential energy consumption and generation. While these technologies are essential for decarbonization, their increasing penetration introduces significant operational and planning challenges for distribution system operators (DSOs). A central challenge is the determination of hosting capacity (HC), defined as the maximum amount of DER capacity that can be integrated into a distribution network without violating operational constraints. Hosting capacity assessment is inherently complex due to the stochastic and time-varying nature of DER behavior, particularly for EV charging, as well as the ongoing transition from traditional passive networks toward actively managed distribution systems. The aim of this dissertation is to develop scalable and risk-aware frameworks for assessing the HC of EVs and PV systems in passive distribution networks while also introducing new methodologies for HC assessment of DERs in active distribution networks (ADNs). To achieve these objectives, the thesis addresses three key research challenges: (i) the limited treatment of EV-driven uncertainties in existing risk-based HC studies due to their modeling complexity, (ii) the computational intractability of stochastic HC analysis under long-term temporal uncertainty, and (iii) the inadequacy of the conventional HC framework for ADNs that employ dynamic operating envelopes (DOEs). To address these challenges, the thesis proposes HC assessment frameworks grounded in realistic modeling of DER behavior and distribution network characteristics. For traditional passive distribution networks, a multi-period stochastic HC framework is developed using Monte Carlo scenario generation. Risk is quantified using Value at Risk (VaR), enabling DSOs to explicitly control acceptable levels of operational risk. The results demonstrate that EV-HC is constrained by different types of distribution network incidents, including voltage violations, line overloads, and transformer congestion, depending on feeder characteristics. A key finding is that EV charging simultaneity plays a decisive role in limiting HC, with critical violations occurring during both peak-load periods and high charging coincidence events. These results show that EV-related uncertainties are fundamentally different from PV-related uncertainties and must be explicitly modeled in HC assessment. To overcome the computational burden associated with long-term stochastic time-series analysis, the thesis introduces a grid-aware scenario-reduction approach based on representative days (RDs). Unlike conventional methods that rely solely on load or generation profiles, the proposed approach uses voltage distributions to identify representative operating conditions that are directly relevant to network constraints. Integrating RDs into the risk-based stochastic HC framework, where risk is modeled using conditional value at risk (CVaR), significantly reduces the number of required power flow and optimal power flow simulations while preserving accuracy in risk estimation. The results show that risk-based stochastic HC can be evaluated with low relative error using only a small number of RDs, leading to substantial computational savings. Beyond passive networks, the thesis extends HC analysis to ADNs, where DER behavior is constrained through DOEs. A network-aware HC framework is developed in which DOEs are used to ensure grid-safe flexibility provision. In this context, HC is redefined using aggregated quality-of-service (QoS) metrics to account for the impact of control actions on end users. The results demonstrate that DOEs can increase EV-HC while maintaining network security. However, the analysis also reveals spatial disparities in QoS impacts, with customers located farther from the transformer experiencing greater restrictions. To address the limitation related to HCA in ADN, a decentralized HC framework is proposed in which prosumers actively participate in network operations through active network management. By coordinating behind-the-meter EV charging and PV generation in response to DOE signals, the proposed framework simultaneously enhances EV and PV hosting capacity, preserves user privacy, and improves social welfare. The results confirm that the proposed HC framework enables higher overall DER integration compared to traditional approaches while aligning the objectives of the DSO and prosumers in ADNs. In conclusion, this dissertation advances the state of the art in HC assessment by introducing a risk-based stochastic framework that explicitly accounts for EV-driven uncertainties, developing grid-aware scenario-reduction techniques to address temporal scalability challenges, and redefining HC assessment for ADNs with controlled DER behavior. The proposed methods provide practical and computationally tractable tools that support DSOs in planning and operating future distribution networks under increasing electrification and uncertainty.