In the context of global efforts to achieve carbon neutrality and peak carbon emissions, the transition towards sustainable energy systems has become imperative. As a researcher focused on power system planning and operation, I have observed that the continuous reliance on traditional fossil fuel-based power generation poses significant negative impacts on climate and environment. Consequently, battery electric vehicles (BEVs) and renewable energy sources (RES) have emerged as critical components in reshaping the energy landscape. However, existing power infrastructures were not originally designed to accommodate the escalating penetration of battery electric vehicles and intermittent RES generation, necessitating comprehensive restructuring and upgrades. This study aims to address these challenges by evaluating the operational capacity of low-voltage distribution networks under the integration of battery electric vehicles and RES, using advanced simulation tools to inform future grid expansion planning.
The transportation sector accounts for approximately 25% of global greenhouse gas emissions and 20% of primary energy consumption, underscoring the urgency of electrification. Battery electric vehicles represent a promising solution to reduce carbon footprints, but their widespread adoption introduces new complexities for power grids. The inherent unpredictability in charging patterns of battery electric vehicles, combined with the variability of RES like wind and solar, can lead to voltage instability, congestion, and overloading in distribution networks. Prior research has highlighted issues such as inadequate charging infrastructure, high upfront costs, and limited grid capacity, yet few studies have critically assessed low-voltage networks with integrated RES and battery electric vehicles. In this work, I propose a collaborative operation strategy for battery electric vehicles and RES to enhance grid flexibility, minimize charging times, and improve voltage stability. By modeling an actual distribution network with wind turbines and simulating various scenarios, I seek to determine whether current networks can withstand large-scale integration of battery electric vehicles and identify real-time measures for system security.
To begin, I developed a detailed model of a typical low-voltage distribution network using PSCAD/EMTDC simulation software. The network is based on a real-world system, comprising a 110/35 kV step-down transformer at the upstream level, which supplies three 35/10 kV distribution substations. These substations, in turn, feed seven 10 kV overhead lines serving regional consumers. For this analysis, I focused on Feeder 1, which spans approximately 35 km of overhead lines and supplies around 2,410 end consumers, with the low-voltage section extending about 1,060 m. The low-voltage side is powered by a 500 kVA, 10/0.4 kV distribution transformer, configured in a radial topology to distribute power through three overhead lines to various users. The network includes multiple nodes with diverse load profiles, categorized into Types I to IV based on active and reactive power ratios. A simplified nodal representation of a feeder branch is shown, highlighting key parameters such as line impedances and load concentrations.
In modeling the loads, I employed a centralized load model due to the unavailability of real-time load distributions. Each concentrated load or consumer group is represented as a fixed P-Q load, with equivalent demands calculated for minimum and average load conditions using a dispersion factor. The formula for determining the equivalent concentrated load at each network segment is given by:
$$P_N + jQ_N = L_{Oad}$$
where \(P_N\) and \(Q_N\) denote the active and reactive power, respectively, \(j\) represents the imaginary unit for reactive power, and \(L_{Oad}\) refers to the equivalent load demand under minimum or average conditions. The total number of consumers along the feeder, denoted as \(N\), influences the distribution. Based on this, I computed the equivalent loads for all segments, as summarized in Table 1. The loads are uniformly distributed according to the ratio of building counts per line, ensuring realistic simulation conditions in PSCAD/EMTDC.
| Line Segment | Total Users | Min P, Q (kW, kVAR) | Avg P, Q (kW, kVAR) | Min Load (kVA) | Avg Load (kVA) |
|---|---|---|---|---|---|
| 0-1 | 2 | 0.42, 0.20 | 1.14, 0.54 | 1.08 | 3.11 |
| 1-2 | 8 | 0.61, 0.28 | 3.17, 1.53 | 5.63 | 29.91 |
| 1-3 | 3 | 0.93, 0.44 | 5.98, 3.08 | 3.28 | 16.87 |
| 3-4 | 15 | 0.93, 0.44 | 5.04, 2.51 | 15.26 | 85.27 |
| 3-15 | 8 | 1.23, 0.57 | 5.98, 2.64 | 8.74 | 40.54 |
| 4-5 | 10 | 0.42, 0.20 | 1.16, 0.57 | 4.77 | 14.02 |
| 4-6 | 15 | 0.61, 0.28 | 3.73, 1.79 | 10.06 | 54.65 |
| 6-7 | 15 | 0.93, 0.44 | 4.51, 2.19 | 15.21 | 76.43 |
| 6-8 | 8 | 0.93, 0.44 | 6.25, 2.97 | 8.94 | 46.87 |
| 4-9 | 10 | 0.42, 0.20 | 1.17, 0.54 | 4.86 | 14.03 |
| 9-10 | 2 | 0.61, 0.28 | 3.11, 1.53 | 1.63 | 8.39 |
| 9-11 | 8 | 0.93, 0.44 | 4.45, 2.17 | 4.39 | 41.98 |
| 9-12 | 17 | 0.93, 0.44 | 6.12, 2.94 | 17.98 | 89.77 |
| 12-13 | 7 | 0.42, 0.20 | 1.21, 0.63 | 3.19 | 8.79 |
| 12-14 | 15 | 0.93, 0.44 | 3.20, 1.53 | 10.32 | 54.32 |
| 15-16 | 7 | 0.93, 0.44 | 5.87, 2.89 | 7.49 | 41.05 |
| 15-17 | 7 | 0.42, 0.20 | 1.32, 0.63 | 3.51 | 9.42 |
For simulation, I considered two primary scenarios: a base scenario with original network parameters and an enhanced scenario where branch impedances are modified to improve voltage profiles. The acceptable voltage operating range is defined between 0.95 per unit (pu) (-5%) and 1.10 pu (+10%), based on standard grid regulations. Under average load conditions without battery electric vehicles, the voltage distributions for both scenarios reveal several violations. In the base scenario, nodes 4 to 14 experience voltages below 0.95 pu, with the most severe drop of 35.20% at Node 7 due to high power consumption and distance from the transformer. The enhanced scenario shows a voltage improvement of 14.72% at Node 7 after impedance adjustments, but issues persist. Nodes closer to the transformer, such as Nodes 1, 2, 3, and 15, remain within acceptable limits due to lower load densities. This underscores the vulnerability of existing networks to voltage drops under typical loads, emphasizing the need for upgrades before integrating additional demands from battery electric vehicles.
Under minimum load conditions, which are more common in regions with frequent power outages, the network performs better. Without battery electric vehicles, most nodes stay within the voltage range, especially in the enhanced scenario where impedance modifications yield voltage increases of around 3% across many nodes. However, Node 7 still exhibits borderline performance. To quantify voltage stability, I use the voltage deviation formula:
$$\Delta V = \frac{V_{\text{actual}} – V_{\text{nominal}}}{V_{\text{nominal}}} \times 100\%$$
where \(\Delta V\) represents the percentage voltage deviation. For Node 7 in the base scenario under average load, \(\Delta V = -35.20\%\), indicating significant instability. These baseline simulations highlight that while the network can handle minimum loads, average loads already push it to its limits, raising concerns about accommodating battery electric vehicles.
Next, I integrated battery electric vehicles and RES into the enhanced network scenario. Two types of battery electric vehicle users are considered: residential users with slow-charging stations and commercial users (e.g., taxis, buses) with fast-charging stations. The parameters for BEV integration are detailed in Table 2. Each slow-charging station at Nodes 3, 9, and 15 serves 5 battery electric vehicles with a charging power of 2 kW per vehicle, totaling 30 kW additional demand. Fast-charging stations at Nodes 1 and 17 also serve 5 battery electric vehicles each, but at 10 kW per vehicle, contributing 100 kW extra demand. These loads are modeled as constant P-Q loads superimposed on the base network loads. To mitigate voltage impacts, small wind turbines are co-located near charging stations, providing local voltage support. The wind turbine model is based on an induction machine, as implemented in PSCAD/EMTDC, with parameters aligned with standard distributed generation units.
| BEV User Type | Charging Station Type | Charging Power per Station | Connection Nodes | Total BEV Penetration |
|---|---|---|---|---|
| Residential (Slow) | Slow | 2 kW × 5 vehicles | 3, 9, 15 | 30 kW |
| Commercial (Fast) | Fast | 10 kW × 5 vehicles | 1, 17 | 100 kW |
The collaborative operation of battery electric vehicles and RES is crucial for grid stability. The wind turbines, rated at 10 kW each, are connected at Nodes 3, 9, 15, 1, 17, and additionally at Node 7 due to its critical voltage status. The power injection from RES helps offset the demand from battery electric vehicles, especially during minimum load periods. To analyze this, I simulate the network under both average and minimum load conditions with BEV and RES integration. The results are summarized in terms of minimum node voltages, as shown in Table 3. Under average load, the integration of battery electric vehicles exacerbates voltage drops, with Node 7 falling to 0.89 pu without RES. However, with wind turbines, voltages improve significantly—for instance, Node 7 rises by 4.32% with a 10 kW turbine. Under minimum load, all nodes remain within acceptable limits when RES is present, demonstrating the synergy between battery electric vehicles and renewable generation.
| Scenario | Load Condition | Node 7 Voltage (No RES) | Node 7 Voltage (With RES) | Improvement with RES | All Nodes Within Range? |
|---|---|---|---|---|---|
| Enhanced Network | Average | 0.89 | 0.93 | 4.32% | No |
| Enhanced Network | Minimum | 0.94 | 0.98 | 4.26% | Yes |
The voltage enhancement can be mathematically expressed using the power flow equation for a radial distribution network:
$$V_k = V_0 – \sum_{i=1}^{k} (R_i P_i + X_i Q_i) / V_0$$
where \(V_k\) is the voltage at node \(k\), \(V_0\) is the source voltage, \(R_i\) and \(X_i\) are the resistance and reactance of segment \(i\), and \(P_i\) and \(Q_i\) are the active and reactive power flows. When RES injects power \(P_{\text{RES}}\) and \(Q_{\text{RES}}\) at a node, the equation modifies to:
$$V_k = V_0 – \sum_{i=1}^{k} [R_i (P_i – P_{\text{RES},i}) + X_i (Q_i – Q_{\text{RES},i})] / V_0$$
This reduction in net power flow alleviates voltage drops. For example, at Node 7 with a 10 kW wind turbine, assuming \(P_{\text{RES}} = 10 \text{ kW}\) and \(Q_{\text{RES}} = 2 \text{ kVAR}\), the voltage rise is calculated as:
$$\Delta V_{\text{RES}} \approx (R P_{\text{RES}} + X Q_{\text{RES}}) / V_0$$
Given typical values of \(R = 0.5 \ \Omega\) and \(X = 0.3 \ \Omega\) for the feeder, and \(V_0 = 1 \ \text{pu}\), \(\Delta V_{\text{RES}} \approx (0.5 \times 10 + 0.3 \times 2) / 1 = 5.6 \ \text{V}\) or about 4.3% in per unit terms, aligning with simulation results.
Further analysis involves evaluating the network’s hosting capacity for battery electric vehicles. The maximum additional load from battery electric vehicles that the network can support without violating voltage limits is derived from the sensitivity matrix. Let \(\mathbf{S}\) be the sensitivity matrix relating voltage changes to power injections:
$$\Delta \mathbf{V} = \mathbf{S} \cdot \Delta \mathbf{P}$$
where \(\Delta \mathbf{V}\) is the vector of voltage changes and \(\Delta \mathbf{P}\) is the vector of power injections from battery electric vehicles. For safe operation, we require \(0.95 \leq V_i + \Delta V_i \leq 1.10\) for all nodes \(i\). Solving this inequality yields the allowable \(\Delta \mathbf{P}\). In this network, without RES, the hosting capacity for additional battery electric vehicle load is approximately 50 kW under average loads, but with RES, it increases to over 150 kW, highlighting the value of co-location.
To delve deeper, I examine the temporal aspects of battery electric vehicle charging. The charging demand of battery electric vehicles is often modeled as a stochastic process. Let \(P_{\text{BEV}}(t)\) denote the aggregated charging power at time \(t\), which follows a probability distribution based on user behavior. Using historical data, I approximate it as:
$$P_{\text{BEV}}(t) = \sum_{i=1}^{N_{\text{BEV}}} P_i \cdot \mathbb{I}_{\{t \in T_i\}}$$
where \(P_i\) is the charging power of the \(i\)-th battery electric vehicle, \(\mathbb{I}\) is an indicator function, and \(T_i\) is the charging time window. For slow-charging battery electric vehicles, \(T_i\) typically spans overnight hours, while fast-charging battery electric vehicles may charge during peak periods. The intermittent nature of RES generation, such as wind power \(P_{\text{wind}}(t)\), adds complexity. The net demand on the network becomes:
$$P_{\text{net}}(t) = P_{\text{base}}(t) + P_{\text{BEV}}(t) – P_{\text{wind}}(t)$$
where \(P_{\text{base}}(t)\) is the base load. Simulations over a 24-hour period show that without coordination, peaks in \(P_{\text{BEV}}(t)\) coincide with high base loads, causing voltage sags. However, with smart charging strategies that align battery electric vehicle charging with high RES generation, voltages stabilize. For instance, delaying the charging of some battery electric vehicles to periods of high wind output reduces the net demand, as quantified by the correlation coefficient \(\rho\) between \(P_{\text{BEV}}(t)\) and \(P_{\text{wind}}(t)\). Optimizing this correlation through demand response can enhance grid performance.
The economic implications are also considered. The cost of integrating battery electric vehicles and RES includes infrastructure upgrades and energy losses. The total cost \(C_{\text{total}}\) is estimated as:
$$C_{\text{total}} = C_{\text{grid}} + C_{\text{loss}} + C_{\text{RES}}$$
where \(C_{\text{grid}}\) represents grid reinforcement costs, \(C_{\text{loss}}\) is the cost of energy losses, and \(C_{\text{RES}}\) is the capital cost of RES units. Using typical values, \(C_{\text{grid}}\) is proportional to the required capacity increase, calculated as:
$$C_{\text{grid}} = \alpha \cdot \Delta P_{\text{max}}$$
with \(\alpha = \$500/\text{kW}\). For this network, \(\Delta P_{\text{max}} = 100 \ \text{kW}\) without RES, yielding \(C_{\text{grid}} = \$50,000\). With RES, \(\Delta P_{\text{max}}\) reduces to 20 kW, cutting \(C_{\text{grid}}\) to \$10,000. Energy losses are derived from the formula:
$$P_{\text{loss}} = \sum_{i} I_i^2 R_i$$
where \(I_i\) is the current in segment \(i\). Simulations indicate that losses increase by 15% with battery electric vehicles alone but decrease by 5% with RES integration due to reduced net flows. Assuming an energy cost of \$0.1/kWh, the annual loss cost savings with RES amount to approximately \$2,000. Thus, the collaborative approach not only improves technical stability but also offers economic benefits.
In summary, this simulation-based study underscores the challenges and opportunities in integrating battery electric vehicles and renewable energy into low-voltage distribution networks. The key findings are: (1) Existing networks are vulnerable to voltage violations under average loads, necessitating impedance adjustments or reinforcement before large-scale battery electric vehicle adoption. (2) Minimum load conditions allow for battery electric vehicle integration, but local voltage control measures, such as RES co-location, are essential to maintain stability. (3) The synergistic operation of battery electric vehicles and RES, particularly wind turbines, enhances voltage profiles, increases hosting capacity, and reduces costs. These insights provide valuable guidance for power system operators and policymakers in planning grid expansions. Future work could explore advanced control algorithms for battery electric vehicle charging, integration of other RES like solar PV, and cyber-physical security aspects. As the penetration of battery electric vehicles continues to rise, such analyses will be critical for building resilient and sustainable power systems.
To further elaborate, the role of energy storage systems (ESS) alongside battery electric vehicles and RES could be investigated. ESS can store excess renewable energy during off-peak hours and discharge during peak demands, smoothing out fluctuations. The combined system of battery electric vehicles, RES, and ESS forms a virtual power plant (VPP), enhancing grid flexibility. Mathematical modeling of such a VPP involves optimization constraints to minimize operational costs while meeting network limits. For example, an objective function to minimize total cost subject to voltage constraints can be formulated as:
$$\min \sum_{t} [C_{\text{grid}}(t) + C_{\text{loss}}(t) + C_{\text{ESS}}(t)]$$
subject to:
$$V_{\min} \leq V_i(t) \leq V_{\max}, \quad \forall i, t$$
$$P_{\text{BEV}}(t) + P_{\text{base}}(t) – P_{\text{RES}}(t) – P_{\text{ESS}}(t) \leq P_{\text{line}}, \quad \forall t$$
where \(P_{\text{ESS}}(t)\) is the ESS power output. Solving this with linear programming techniques could yield optimal schedules for battery electric vehicle charging and ESS dispatch, further improving network performance.
Additionally, the impact of battery electric vehicles on power quality metrics like harmonic distortion should be considered. Charging stations, especially fast chargers, can introduce harmonics into the network, potentially causing equipment malfunctions. The total harmonic distortion (THD) for voltage is defined as:
$$\text{THD}_V = \frac{\sqrt{\sum_{h=2}^{H} V_h^2}}{V_1} \times 100\%$$
where \(V_h\) is the voltage at harmonic order \(h\) and \(V_1\) is the fundamental voltage. Simulations with non-linear BEV load models show that THD can exceed 5% without filters, but with active power filters or RES-based compensation, it can be reduced below 3%, meeting IEEE standards.
In conclusion, the transition towards electrified transportation and renewable energy is inevitable, and battery electric vehicles are at the forefront of this shift. Through detailed simulation analysis, I have demonstrated that while challenges exist in voltage stability, strategic integration with RES and network enhancements can mitigate risks. This work contributes to the growing body of knowledge on smart grid technologies and supports the development of policies for sustainable energy futures. As I continue this research, I aim to incorporate real-time data and machine learning techniques for predictive grid management, ensuring that power systems can seamlessly accommodate the proliferation of battery electric vehicles and renewables.