New Reactive Power Optimization Model Enhances Grid Stability with EV and Renewable Integration
As the global energy landscape undergoes a profound transformation, the integration of electric vehicles (EVs) and renewable energy sources into power distribution networks has become a cornerstone of the clean energy transition. While the environmental and economic benefits of solar, wind, and EVs are widely recognized, their widespread adoption presents significant technical challenges for the stability and efficiency of existing power grids. The intermittent nature of solar and wind generation, coupled with the concentrated and variable power demands of EV charging stations, can lead to voltage fluctuations, increased power losses, and potential system instability. Addressing these challenges is critical to ensuring a reliable, resilient, and cost-effective power system for the future.
A groundbreaking study published in the Proceedings of the CSU-EPSA introduces a novel multi-objective reactive power optimization model specifically designed for active distribution networks (ADNs) with high penetration of new energy sources and EV charging stations. The research, conducted by Jiang Zhijun, Yuan Xuan, Qiu Wenhao, Huang Licai from the School of Information Engineering at Nanchang University, and He Wei from the Electric Power Research Institute of State Grid Jiangxi Electric Power Co., Ltd., offers a comprehensive solution to the complex interplay between modern energy resources and grid management.
The core of the study lies in its recognition that traditional grid management strategies are no longer sufficient. Previous research has often focused on optimizing either the power generation side (e.g., using inverters in wind and solar farms for reactive power control) or the load side in isolation. However, the authors argue that a holistic approach is necessary. The simultaneous influx of variable generation and highly dynamic, concentrated loads from EVs creates a dual challenge that can destabilize the system if not managed synergistically. The study explicitly highlights that many existing models have overlooked the significant impact of EV charging stations on the load side and have not given adequate attention to the resulting stability issues from both the source and the load.
To tackle this, the team developed a sophisticated multi-objective optimization model with three primary goals: minimizing the system’s operating cost, minimizing voltage deviation across the network, and maximizing system stability. The first objective, operating cost, is primarily driven by minimizing active power losses in the grid’s transmission lines, a significant expense for utilities. The second objective, voltage deviation, aims to keep the voltage at every node in the distribution network as close as possible to its nominal value (e.g., 12.66 kV), preventing equipment damage and ensuring power quality. The third and most innovative objective is the optimization of system stability, which the researchers approached by proposing a new static voltage stability index.
The proposed stability index is a key contribution of the paper. Traditional methods for assessing voltage stability can be complex and computationally intensive. The authors devised a new metric based on the analysis of the PV (Power-Voltage) curve, a fundamental concept in power system analysis that shows the relationship between the power drawn from a network and the resulting voltage. As a system approaches its stability limit, or “collapse point,” the PV curve exhibits a characteristic “nose” shape. The distance between the upper and lower branches of this curve at a given power level becomes very small. The researchers’ new index quantifies this closeness, providing a clear, single-value indicator of how close any given branch of the grid is to voltage collapse. By minimizing the maximum value of this index across all branches in the network, the model ensures that the entire system operates with a robust margin of stability, even under the most stressful conditions of high renewable output and peak EV charging demand.
The complexity of this multi-objective model is immense. It involves a vast number of decision variables, including continuous variables like the reactive power output from wind turbines, photovoltaic (PV) inverters, static var compensators (SVCs), and EV charging stations, as well as discrete variables like the number of capacitor banks (CBs) that are switched on or off at specific nodes. Solving such a mixed-integer, non-convex, and non-linear optimization problem is notoriously difficult for conventional algorithms, which can get trapped in suboptimal solutions or require prohibitively long computation times.
To overcome this computational hurdle, the research team introduced a hybrid optimization algorithm with a unique cross-feedback mechanism. This algorithm combines two powerful metaheuristic techniques: an improved Moth-Flame Optimization (MFO) algorithm for the continuous variables and a Genetic Algorithm (GA) for the discrete variables. The MFO algorithm, inspired by the navigation behavior of moths, is known for its ability to explore a solution space. However, it can suffer from a loss of diversity in its later stages, leading to premature convergence on a local optimum. The authors significantly enhanced the MFO algorithm by incorporating two key improvements.
First, they used a Tent chaotic mapping for the initial population generation. Chaos theory provides a way to generate sequences that are deterministic yet appear random and are highly sensitive to initial conditions. This ensures that the starting population of potential solutions is far more diverse and spread out across the entire solution space, giving the algorithm a better chance of finding the global optimum from the outset. Second, they introduced an adaptive mutation operator based on the population’s “aggregation degree.” This metric measures how closely clustered the current population of solutions is. If the population becomes too concentrated (indicating a risk of getting stuck), the mutation operator injects more randomness and diversity into the search process, effectively helping the algorithm escape local optima and continue its exploration.
The hybrid algorithm’s cross-feedback mechanism is its masterstroke. Instead of optimizing all variables simultaneously, it separates them. The GA works on the discrete capacitor bank switching decisions, while the improved MFO (dubbed TAMMFO) works on the continuous reactive power outputs. Crucially, the algorithms do not work in isolation. After a round of optimization by one algorithm, its best results are fed forward as fixed parameters for the other algorithm to use in its next round of optimization. For example, the GA might find a promising configuration for the capacitor banks; this configuration is then held constant while the TAMMFO algorithm searches for the optimal reactive power settings for the inverters and SVCs. The result of this search is then used to inform the GA’s next iteration, and so on. This iterative, feedback-driven process allows the two algorithms to guide each other toward a superior overall solution, leveraging the strengths of each method while mitigating their weaknesses. This approach is far more efficient than trying to solve the entire complex problem with a single algorithm.
To validate their model and algorithm, the researchers conducted a detailed case study using a modified IEEE 33-node distribution system, a standard benchmark in power systems research. The simulation scenario included wind turbines, a PV plant, and two EV charging stations connected at strategic points in the network. The researchers used realistic, forecasted data for wind and solar generation and a sophisticated Monte Carlo simulation to model the stochastic behavior of 400 EVs (a mix of private, official, and taxi vehicles) based on their travel and charging patterns.
The simulation results were compelling and demonstrated the superiority of the proposed approach. The researchers compared two cases: Case 1, which used their full hybrid algorithm (TAMMFO + GA), and Case 2, which used a standard MFO algorithm combined with GA. The results showed that the model successfully achieved its three objectives. After optimization, the system’s operating cost, primarily driven by power losses, was reduced by over 44% compared to the unoptimized scenario. Case 1 outperformed Case 2, achieving a 2.41% further reduction in cost, demonstrating the efficiency of the improved TAMMFO algorithm.
The impact on voltage quality was equally impressive. The total voltage deviation across the network was reduced by over 49%. Case 1 again showed a clear advantage, reducing the deviation by an additional 3.74% compared to Case 2. This translates to a more stable and higher-quality power supply for all customers connected to the grid. Perhaps most critically, the new voltage stability index (VSI) was significantly improved. The peak VSI value was reduced, indicating a much larger safety margin against voltage collapse. Case 1 achieved a more stable system than Case 2, with a further reduction in the peak VSI.
A major practical benefit of the proposed strategy was its ability to manage equipment wear and tear. The model explicitly includes a constraint on the maximum number of times capacitor banks can be switched on or off in a day, as frequent switching reduces equipment lifespan and incurs maintenance costs. In Case 2, the capacitor banks at two key nodes each switched 10 times, exceeding the allowable daily limit of 6. This is a significant operational flaw. In stark contrast, the optimized solution from Case 1 kept the switching within the limit, with one bank switching 6 times and the other only 4 times. This not only extends the life of the equipment but also reduces operational costs for the utility.
Finally, the computational efficiency of the hybrid algorithm was a standout feature. Case 1, despite its superior results, completed its optimization in 6,115 seconds, which is 38.06% faster than Case 2’s 9,873 seconds. This speed is crucial for real-world application, as grid operators need to run these optimizations on a daily or even hourly basis to respond to changing conditions. A faster algorithm makes this practical.
In conclusion, the research by Jiang Zhijun, Yuan Xuan, Qiu Wenhao, Huang Licai, and He Wei represents a significant advancement in the field of smart grid management. Their integrated model, which simultaneously considers the dynamic effects of renewables and EVs, provides a more realistic and effective framework for modern power systems. The novel stability index offers a powerful new tool for grid operators to proactively manage risk. Most importantly, their innovative hybrid optimization algorithm with cross-feedback demonstrates a practical and efficient way to solve one of the most complex problems in power engineering. As the world continues to electrify transportation and decarbonize its energy supply, this kind of sophisticated, data-driven, and holistic approach will be essential for building a grid that is not only green but also robust, reliable, and economical. This work provides a valuable blueprint for utilities and researchers worldwide as they navigate the challenges and opportunities of the energy transition.
Jiang Zhijun, Yuan Xuan, Qiu Wenhao, Huang Licai, He Wei, School of Information Engineering, Nanchang University; Electric Power Research Institute, State Grid Jiangxi Electric Power Co., Ltd. Reactive Power Optimization Model of Active Distribution Network with New Energy and Electric Vehicle Charging Stations. Proceedings of the CSU-EPSA. DOI: 10.19635/j.cnki.csu-epsa.001297