Increasing the share of Electric Vehicles (EVs) powered by renewable-based Distributed Energy Resources (DERs) is a key step towards climate neutrality. However, increasing the penetration of EVs and Photovoltaic (PV) generators may create large and hardly predictable fluctuations in power supply and demand, thus destabilising the grid. In this paper, an optimisation algorithm for smart EV charging is proposed to reduce the overall net-load variance through a more efficient exploitation of the available PV power, EV charging shifting, or vehicle-to-grid (V2G). Key distinctive features of the proposed approach are: (i) the formulation as a quadratic programming problem; (ii) the capability to enable a V2G charging policy, (iii) the inclusion of specific constraints regarding EVs’ availability, owners’ charging requirements and, partially, voltage stability; (iii) the study of the combined impact of EV and PV penetration on bus voltages, line currents, district self-sufficiency, and EV battery lifetime. The proposed approach is tested not only in ideal conditions, but also considering a basic persistence forecasting model of load and PV generation over subsequent days. The results of grid-level simulations in a case study show that the proposed approach could decrease the net-load variance by up to 60% if no forecasting errors occur and by about 50% when the persistence forecasting model is used. Additionally, the V2G policy notably decreases both the range of voltage fluctuations and the risk of line overloading, although at the expense of EVs’ battery lifetime, whose reduction actually depends on the battery capacity.

The transmission of sensor information from multiple locations through the communication channel significantly impacts the reliability of microgrid protection algorithms due to packet loss. The lossy information acquired at the control centre does not reflect the actual dynamics of the microgrid. Schemes reported in the literature for microgrid protection have not taken into account the inevitable phenomenon of data loss in the communication network. The present work proposes a scheme based on the combined framework of critical sensor identification (CSI), Markov chain-based packet loss modelling, spline interpolation and deep learning based stacked sparse autoencoder (SSAE) classifier for hybrid microgrid protection. The scheme initiates with the acquisition of sensor information from critical buses (identified using observability theory), followed by deriving the signal with packet loss of a specific rate. From the signal with packet loss, an approximate version of the original signal is obtained using spline interpolation and further fed to a set of SSAE classifier modules formulated to perform the tasks of fault detection/classification and section identification. The proposed scheme has been validated for wide range of fault scenarios (generated by varying the fault resistance, fault inception angle, fault location for all possible fault types) under varying degrees of packet loss (generated by varying the burst length and packet loss rate). The simulation result reflects that the proposed protection scheme is able to attain a high degree of accuracy and reliability along with properly recovering the post fault current–voltage dynamics from the lossy data under normal and N-1 contingency scenarios.

Nowadays, dynamic line rating (DLR) has attracted much attention. DLR leads to cost reduction and congestion management. Studies have shown that the maximum extractable capacity is obtained in the transient state. Load flow is the basis for many calculations in power systems. But here is the question: what is the impact of different load flow methods and standards on the temperature of lines considering transient conditions? This paper addresses such issues in a contingency analysis problem. The inclusion of new variables would extend the runtime. Therefore, parallel computing is employed. According to the results, there is a considerable difference in the line temperature obtained by the CIGRE and IEEE standards. This difference must be taken into account in power system calculations. Ignoring it may lead to damage to transmission lines. Also, the temperature obtained by the Gauss–Seidel method is a bit different from other load flow methods. Furthermore, the results show that temporal discretization (choosing an appropriate time step) has a significant impact on the convergence speed and runtime. Finally, parallel computing can significantly cover the increased volume of computations (5.8 times speedup in the IEEE 39-bus system using Newton–Raphson method).

Distribution grid companies and distribution system operators (DSOs) still mostly follow a traditional framework for grid planning. Such frameworks have so far served DSOs well in the economic assessment and cost–benefit analysis of passive measures, such as grid reinforcement. However, the development towards active distribution grids requires DSOs to also be able to assess an extended set of active measures. To this aim, this paper extends and implements a general planning framework for active distribution grids that builds upon the well-proven traditional framework. The methodology integrated in the framework includes: (1) decoupled models for (i) operation with active measures and (ii) optimal grid investment, and (2) methods for economic assessment considering active measures from both (i) a DSO cost–benefit analysis perspective and (ii) a willingness-to-pay perspective. In this paper, operational models are integrated for two examples of active measures, namely the use of fast-charging stations (FCS) and local energy communities (LEC). The methodology is demonstrated in a long-term grid planning case study for a realistic Norwegian medium voltage distribution system. For this case, grid planning with FCS as an active measure reduces the present value of grid investment costs by 70% compared with a passive grid planning strategy. The results also demonstrate how the methodology can be used in negotiating the price of active measures between the DSO and distribution system actors such as LEC and FCS operators.

There is an increasing need to decarbonise both heating and transport sectors in the UK, and the uptake of low carbon technologies (LCTs) will be central to this. The impact of LCTs on electricity network infrastructure varies both spatially and temporally, and is driven by the diversity in technology type, consumer behaviour, variable weather patterns, variation of the building stock and the incumbent network assets. In recognition of this diversity and household energy variability, LCT adoption and utilisation will be influenced by the distribution of socio-economic factors within a local area. This has the potential to impact network decision-making across different regions. As such, there is a requirement to consider socio-technical and socio-spatial dimensions when modelling LCT impact on network infrastructure. This research, presented within a UK context, demonstrates a novel high-resolution methodology that enables assessment of electrified heat and transport impact on transformer headroom using socio-economic indicators to inform the application of LCT consumption. This includes mapping of spatially linked datasets to identify relationships between consumption and social deprivation. These relationships are used as inputs to a heat pump modelling methodology that converts gas demand to equivalent electrical heat demand. This approach is compared with a generalised trial data approach to ascertain the impact of incorporating socio-economic elements. Electric vehicles are then introduced, where charging is based on socially disaggregated behaviour in the form of travel diaries showing the combined impact of different LCTs. Findings are considered from the perspective of the distribution network operator and other key stakeholders.

The integrated energy system (IES) is a multi-energy internet carrier with high efficiency, and environmentally friendly that has recently been the focus of much attention. Furthermore, graph theory has a superior solving ability to optimize quickly and effectively a topology programming problem for complex networks. The widely focused concern of graph theory is more proactive consideration of the potential complex topology of network optimization of energy supply network (ESN) at the planning stage of IES. To this end, this paper proposed a novel graph theory-based optimizing approach for solving the ESN configuration, which is called the layering and pruning method (L&PM) in this paper. First, a shortcoming of the minimum spanning tree (MST) to solve the planning problem of ESN is given in detail with an elaborate analysis. After that, the L&PM are designed with topological and economic analyses of local topologies, power flows, material consumption, etc., of pipelines of ESN, which can optimize the overall topology and pipeline capacities of ESN simultaneously, and avoid the problem that only optimizes the total length of ESN by MST. In addition, a two-stage planning model of IES considering ESN is formulated. Finally, via a real-world IES in South China, compared to the Prim algorithm (MST), the results of the case studies showed the developed L&PM’s advantages significantly in improving the overall planning scale and economy of IES and satisfy and reliability of the system, which verifies prominently the superiority, effectiveness, and feasibility of the proposed L&PM.

Hydro-thermal economic dispatch is a widely analyzed energy optimization problem, which seeks to make the best use of available energy resources to meet demand at minimum cost. This problem has great complexity in its solution due to the uncertainty of multiple parameters. In this paper, we view hydro-thermal economic dispatch as a multistage decision-making problem, and propose several Deep Reinforcement Learning approaches to solve it due to their abilities to handle uncertainty and sequential decisions. We test our approaches considering several hydrological scenarios, especially the cases of hydrological uncertainty due to the high dependence on hydroelectric plants, and the unpredictability of energy demand. The policy performance of our algorithms is compared with a classic deterministic method. The main advantage is that our methods can learn a robust policy to deal with different inflow and load demand scenarios, and particularly, the uncertainties of the environment such as hydrological and energy demand, something that the deterministic approach cannot do.

Solar photovoltaic generation and energy storage play an increasingly important role in supplying the electricity needs of remote areas. However, private energy storage systems are a significant encumbrance to consumers in remote areas. Moreover, communal energy storage has enormous economic constraints owing to the distance from remote areas. In this study, we propose a new model for shared energy storage using the Neighbor scenario, where each consumer can share an energy storage system with the nearest consumer. To validate the model, we propose three scenarios: Individual, Neighbor, and Communal. The system was designed as a mixed-integer linear programming (MILP) model to analyze optimal operational schemes. The results show that using the Neighbor scenario can save operating costs and increase the storage utilization rate by 6.15% and 2.65%, respectively, relative to the other scenarios. Finally, a sensitivity analysis is performed to demonstrate optimal operation in terms of economic and emission parameters.

This paper proposes a novel approach based on adjustable robust optimization for the investment planning of a virtual power plant that participates in the energy electricity market. The virtual power plant behaves in this market as a price-taking agent that faces exogenous prices. The virtual power plant comprises conventional, renewable, and storage units, as well as flexible demands. Investment decisions on conventional, renewable, and storage units are made under the uncertainty related to future production costs of the conventional generating units, future consumption levels of the flexible demands, and future energy market prices. As a major modeling contribution, the nonconvex operation of both conventional generating units and storage devices is precisely accounted for, thus yielding a trilevel program with lower-level binary variables. The resulting model is solved using an exact nested column-and-constraint generation algorithm, which constitutes the methodological contribution. Results from several case studies are provided to show the effective performance of the proposed approach.

Variable renewable energy producers and electricity retailers encounter several uncertainties in their decision-making problems, such as intermittency of renewable energy sources, variability of consumption, and market price volatility. To cope with these uncertainties, this paper presents a new contract-based trading mechanism of power flexibility (FlexCon) between two parties, a variable renewable energy producer and an electricity retailer. The proposed mechanism is managed by a new entity, named FlexCon operator, to oversee the energy and financial trades through the contract and coordinate the transactions with the system operator. Through the FlexCon, the parties are able to exchange their energy imbalances as a source of power flexibility to alleviate the negative impacts of uncertainties in their decision-making problems. To this end, two two-stage stochastic linear problems are introduced from each party’s point of view. In the first stage, the variable renewable energy producer and the electricity retailer submit their bids to sell and purchase in the day-ahead market, respectively. Following the day-ahead market clearing, closer to the delivery time, the parties submit their decisions on the contract to the introduced FlexCon operator. The operator allocates possible power flexibility transactions based on the surpluses or shortages of the parties. Assuming that the imbalances are not completely resolved with the FlexCon, the remaining deviations are settled in the balancing market. The parties’ decisions related to the balancing market and the FlexCon are modeled in the second stage of the stochastic problem. The uncertainties associated with prices, renewable generation, electricity consumption, and the maximum exchangeable power flexibility through the FlexCon are considered via scenarios. Meanwhile, the profit risk is considered by the Conditional Value at Risk measure. The numerical results show that FlexCon effectively diminishes the impacts of uncertainties on the parties’ profit.

This paper introduces a probabilistic modelling approach, based on smart meter data and a novel agent-based electric vehicle (EV) simulator, to analyse the impact of various cost-based EV charging strategies on the power distribution network (PDN). We investigate the effects of a 40% EV adoption on three parts of the low voltage distribution network (LVDN) of Frederiksberg, a densely urbanised municipality in Denmark. Our findings indicate that especially cable and transformer overloading poses a challenge. However, the impact of EVs varies significantly between each LVDN area and charging scenario. Across scenarios and LVDNs, the share of cables facing congestion ranges between 5% and 60%. It is also revealed, that time-of-use (ToU)-based and single-day cost-minimised charging could be beneficial for LVDNs with moderate EV adoption rates. In contrast, multiple-day optimisation will likely lead to severe congestion, as such strategies concentrate demand on a single day that would otherwise be distributed over several days, thus raising concerns on how to prevent it.The wider implications of our study are that, while initial concerns have mainly been focused on congestion caused by uncontrolled charging during peak hours, a shift towards cost-based smart charging driven by a growing awareness of time-dependent electricity prices, could result in a sharp increase in charging synchronisation with undesirable implications for the PDN.

Lighting demand in built environments corresponds to more than 10% of the global energy use, with great potential for energy saving thanks to innovative smart materials and combination of passive–active strategies. This paper focuses on an innovative smart lighting system based on a blue-diode laser technology combined with photoluminescent pavements, to obtain a self-lighting surface with a substantial reduction in electricity consumption. A laser variable angle scanner prototype was developed and operated through a dedicated management software. The prototype was installed in lab to measure the emitted luminance/illuminance of the pavement, its time decay, and the electricity consumption. This application is yet a prototype, the scanning area and the investigated pavement are more compatible with pedestrian street and sidewalk because of the lower illuminances to be maintained. The results were compared to standard solutions installed in typical roads. The results presented promising benefits in terms of electricity saving (about 40% compared to standard LEDs) and luminance uniformity (higher than 0.9 compared to standard systems, i.e. 0.17÷0.33). Important future ideas derive from preliminary experimental campaigns to improve the smart laser system, and the coupled pavement properties to reduce the consumptions and improve the lighting conditions for pedestrians and vehicles, within potential self-lighting streets.

The decarbonization of the electricity sector calls for new operational schemes and businesses. In this context, traditional consumers have evolved towards prosumers, enabling the active participation of domestic installations in the system operation. A set of prosumers can be organized into energy communities to unlock different economic and energy benefits. This paper develops a bi-level scheduling strategy for robust optimal operation of active distribution networks in the presence of energy communities. The new proposal deals with energy management in communities at the lower level while managing distributed assets at the upper level. The uncertainties in demand, renewable generation, and energy prices are modelled using interval notation, which allows to adopt pessimistic or optimistic strategies. In addition, a novel Stackelberg-bisection sub-module is advocated to determine the distribution system operator profit, which considers collective welfare. The developed Mixed-Integer-Linear programming framework is tested in the IEEE 33-bus network incorporating energy communities, passive consumers, and distributed assets. The results obtained serve to validate the new methodology as well as analyse different results. For example, it is observed that the adopted strategy directly impacts the monetary balance, thus varying the incomes by up to 28 % depending on the robustness level. The proposed Stackelberg-based sub-module is also analysed, indicating that the expected profit may vary by up to 3 % depending on the strategy adopted and the level of robustness assumed.

With the increasing penetration of distributed energy resources in distribution networks, Volt-VAR control and optimization (VVC/VVO) have become very important to ensure an acceptable quality of service to all customers. System operators can rely on slow-responding utility devices, including capacitor banks and on-load tap changing transformers, along with fast-responding battery and photovoltaic (PV) inverters for the VVC/VVO implementation. Because of variations in response time of these two classes of devices, and different control actions (discrete versus continuous), coordinated and optimal scheduling and operation have become of utmost importance. This paper develops a look-ahead deep reinforcement learning (DRL)-based multi-objective VVO technique to improve the voltage profile of active distribution networks, decrease network and inverter power loss, and save the operational cost of the grid. It proposes a deep deterministic policy gradient (DDPG)-based approach to schedule the optimal reactive and/or active power set-points of fast-responding inverters, and a deep Q-network (DQN)-based DRL agent to schedule the discrete decisions variables of slow-responding assets. The reactive power output of PV and battery smart inverters are scheduled at 30-minute intervals and the capacitors’ commitment status is scheduled with several hour intervals. The proposed framework is validated on the modified IEEE 34-bus and 123-bus test cases with embedded PV and PV-plus-storage. To validate the efficacy of the proposed VVO, it is compared with several scenarios, including the base case without VVO, localized droop control of DERs, DDPG-only, and twin delayed DDPG (TD3) agent-based DRL techniques. The results justify the superior performance of the proposed method to improve the voltage profile, reduce network power loss, and minimize the look-ahead grid operational cost while minimizing the undesirable power losses in inverters as a result of power factor adjustments.

Integrating renewable energy sources in sectors such as electricity, heat, and transportation has to be planned economically, technologically, and emission-efficient to address global environmental issues. Microgrids appear to be the solution for large-scale renewable energy integration in these sectors. The microgrid components must be optimally planned and operated to prevent high costs, technical issues, and emissions. Existing approaches for optimal microgrid planning and operation in the literature do not include a solution for e-mobility infrastructure development. Consequently, the authors provide a compact new methodology that considers the placement and the stochastic evolution of e-mobility infrastructure. In this new methodology, a retropolation approach to forecast the rise in the number of electric vehicles, a monte-carlo simulation for electric vehicle (EV) charging behaviors, a method for the definition of electric vehicle charging station (EVCS) numbers based on occupancy time, and public EVCS placement based on monte-carlo simulation have been developed. A deterministic optimization strategy for the planning and operation of microgrids is created using the abovementioned methodologies, which additionally consider technical power system issues. As the future development of e-mobility infrastructure has high associated uncertainties, a new stochastic method referred to as the information gap decision method (IGDM) is proposed. This method provides a risk-averse strategy for microgrid planning and operation by including long-term uncertainty related to e-mobility. Finally, the deterministic and stochastic methodologies are combined in a novel holistic approach for microgrid design and operation in terms of cost, emission, and robustness.The proposed method has been tested in a new settlement area in Magdeburg, Germany, under three different EV development scenarios (negative, trend, and positive). EVs are expected to reach 31 percent of the total number of cars in the investigated settlement area. Due to this, three public electric vehicle charging stations will be required in the 2031 trend scenario. Thus, the investigated settlement area requires a total cost of 127,029 €. In association with possible uncertainties, the cost of the microgrid must be raised by 80 percent to gain complete robustness against long-term risks in the development of EVCS.

Renewable energies have found a suitable place in the electricity grid due to their non-pollution and cheapness. From the point of view of the grid operator, the allocation of costs should be appropriately allocated to producers and customers in competitive electricity markets. A fair allocation strategy can lead to more efficient use of current transmission lines while providing economic signals to guide future generation planning and load shedding. In this paper, first, using a powerful structural decomposition method, the optimal location of electric vehicle charging stations (EVCSs) and wind farms is determined from the flow aspect of grid lines. Then, the transmission fixed cost is assigned to each item placed in the network. The proposed method is implemented on the IEEE 24-bus network, and its results are analyzed. The fair allocation of transmission fixed costs prevents collusion, and the proposed method can be an effective tool for the network operator.

This paper introduces a novel approach for the optimal placement of battery energy storage systems (BESS) in power networks with high penetration of photovoltaic (PV) plants. Initially, a fit-for-purpose steady-state, power flow BESS model with energy time shift strategy is formulated following fundamental operation principles. The optimal BESS placement methodology is subsequently developed in the realm of incremental modelling of power system losses, which permits to identify the best candidate node for the BESS connection that reduces the total power losses. And for practical results, daily varying patterns of nodal demand, generation dispatch and solar irradiance are considered, all this accompanied by the BESS operating strategy, i.e., electric energy time shift, which is one of the most sounding strategic applications in current BESS facilities. In addition to providing a suitable validation proof using the standard IEEE 5-bus test system, two practical test power network models with 24 and 118 nodes are used to showcase the usefulness of the incremental modelling approach for optimal BESS placement in power grids with high penetration of PV plants.

Increased electric vehicle (EV) penetration brings considerable challenges to the daily planning of the power grid operations. A careful coordination of the grid operations and charging schedules is needed to alleviate these challenges, and turn them into opportunities. For this purpose, we study the Multi-Period Optimal Power Flow problem (MOPF) with electric vehicles under emission considerations. We integrate three different real-world datasets: household electricity consumption, marginal emission factors, and EV driving profiles. We present a systematic solution approach based on second-order cone programming to find globally optimal solutions for the resulting nonconvex optimization problem. To the best of our knowledge, our paper is the first to propose such a comprehensive model integrating multiple real datasets and a promising solution method for the EV-aware MOPF Problem. Our computational experiments on various instances with up to 2000 buses demonstrate that our solution approach leads to high-quality feasible solutions with provably small optimality gaps. In addition, we show the importance of coordinated EV charging to achieve significant emission savings and reductions in cost. In turn, our findings can provide quantitative insights to decision-makers on how to incentivize EV drivers depending on the trade-off between cost and emission.

The transition to electric mobility could overload distribution transformers to unprecedented levels. Sustained overloads produce excessive heat within the transformer and hasten insulation deterioration. As EV deployment rises and the severity of overloads increases, the focus shifts from long-term effects that reduce transformer life to more immediate impacts that can disrupt normal functioning. The latter typically manifest as the blowing of protection fuse, the triggering of pressure relief device, or the breakdown of winding insulation, each of which could result in transformer outage. The transformer’s vulnerability can be monitored in terms of its apparent power flow, top-oil temperature and hottest-spot temperature. The IEEE C57.91 standard specifies the permissible range for these parameters, beyond which the transformer operation could get interrupted. A study to determine the range of EV penetration that can be accommodated by a distribution transformer without any discontinuity in its operation is presented in this paper. The non-linear nature of EV charging is incorporated in the analysis to account for the effect of higher-order harmonics on the transformer’s internal temperature. The effectiveness of a controlled charging strategy, which seeks to regulate transformer loading by minimizing its variance, in preventing service disruption is also validated in this simulation study.

This paper proposes a day-ahead stochastic operation planning for hybrid renewable/non-renewable multi-microgrid systems. The proposed model performs a multi-objective tri-stage decision-making framework to optimize the operating cost, generation flexibility, and demand-side flexibility simultaneously. The first stage of the proposed model presents a cooperative game to minimize the total operating cost of the multi-microgrid system. In this stage, microgrids consider the uncertainty of generation and consumption and share their local resources. This cooperation enhances the efficiency of energy scheduling and creates an overall gain. The Shapley value is used to consider the contribution of microgrids and define their operating costs. At the second and third stages, the generation and demand-side flexibilities are maximized to enhance the ability of the multi-microgrid system compared to the short-term changes in the system. Two new indexes Average Flexibility of Distributed Generation during Peak Period (AFDGPP) and Average Flexibility of Storage System during Peak Period (AFSSPP) are introduced to evaluate the flexibility of the system in different operating conditions. The proposed model is tested on a standard case study and the simulation results show that the AFDGPP and AFSSPP indexes have been improved by 161.8 kW and 92.32 kW, respectively.