Exploring Sustainable Development of New Power Systems under Dual Carbon Goals: Control, Optimization, and Forecasting

In the context of achieving carbon neutrality, the substantial integration of high proportions of renewable energy sources has significantly impacted the dynamic characteristics of power systems, including frequency stability, voltage security, and synchronous stability, thereby posing formidable challenges to the secure and stable operation of power systems [1]. Various measures can be undertaken to enhance power system stability. For instance, enhancements to floating platforms can bolster the stability of offshore floating wind turbine units [2]. Furthermore, the introduction of novel devices can fulfill requirements for power density or improve power generation efficiency [3,4,5]. Nevertheless, the key to enhancing system stability lies in aspects such as grid control and optimization, planning and scheduling, and the accurate forecasting of renewable energy generation.

With the introduction of new energy sources, the accurate prediction of photovoltaic (PV) and wind power in power systems becomes increasingly critical. A deep learning-based comprehensive multi-site wind speed prediction model graph embedding-graph isomorphism-based gated recurrent unit demonstrates superior forecasting accuracy. Experimental results indicate that this model achieves a mean square error of 0.8457 m/s and a root mean square error (RMSE) of 0.9196 m/s in predicting wind speeds at a height of 10 m, thereby enhancing the reliability of wind speed forecasts and contributing to improved power generation efficiency in electric systems [6]. On the other hand, Fu et al. [7] proposed a novel hierarchical-improved variational mode decomposition–temporal convolutional network-gated recurrent unit multi-head attention PV power prediction mechanism, integrating improved variational mode decomposition, time convolutional network-gated recurrent unit architecture, and enriched multi-head attention mechanisms. Experimental results demonstrate a substantial decrease in RMSE and mean absolute error by 55.1% and 54.5%, respectively, compared to traditional methods, particularly evident during fluctuations in PV power under adverse weather conditions. The adoption of this approach effectively enhances the accuracy of prediction, which is crucial for ensuring the secure scheduling and stable operation of power systems. In renewable energy generation systems, the importance of energy storage systems is increasingly emphasized, and accurate battery modeling is essential for optimizing system control and enhancing power generation efficiency. However, due to insufficient data and noise interference, employing a generalized regression neural network for denoising and predicting real-time V-I data proves to be an effective strategy [8].

Control and optimization of the power grid are pivotal for enhancing the overall operational stability of the electricity system. Precise control of various grid components can typically be achieved through rational methodologies or the development of efficient controllers. For example, an optimized voltage feedforward control strategy is proposed to reshape the phase characteristics of system output impedance, thereby significantly enhancing the system’s adaptability to grid impedance. This approach addresses issues introduced by voltage feedforward and equivalent parallel virtual impedance correction, ultimately improving system stability [9]. Moreover, Zhao et al. [10] introduced a receding Galerkin optimal controller combined with a high-order sliding mode disturbance observer scheme. This enables precise tracking of large-scale load setpoints for boiler–turbine units under varying unknown disturbances, enhancing system robustness against interference and improving tracking capabilities within operational constraints. Additionally, Li et al. [11] presented adaptive controllers such as a dynamic model-based adaptive controller and an enhanced quasi-proportional-resonant controller, effectively mitigating harmonic and thrust pulsations while maintaining stability in linear induction motor systems. Furthermore, Zou et al. [12] introduced an intelligent, nonlinear robust controller using a chaos particle swarm gravity search optimization algorithm for diesel generator speed and excitation control. This approach significantly enhances dynamic accuracy and disturbance suppression capabilities compared to traditional proportional–integral–derivative methods. Given the multifaceted influences on grid components, establishing a rational simulation platform and devising effective control strategies remain critical despite the considerable challenges involved [13,14].

Effective control of the power grid not only enhances its stability but also closely correlates with economic benefits. A multi-stage segmented control model based on sensitivity analysis is proposed to address the stochastic nature of wind power generation output [15]. The results demonstrate that this control method achieves a balance between economic efficiency and safety by minimizing the risk of cascading failures in large-scale wind power systems. Additionally, Wang et al. [16] developed a controller for LED lights that utilizes simulation devices to achieve multi-channel intelligent sensing, dimming, and control, which is significant for alleviating global energy shortages. Furthermore, Akdeniz et al. [17] introduced a novel preventive control approach using non-operational vulnerability indices. The suggested decision support system is expected to assist power system operators in making critical decisions to adjust power system configurations in response to potential risks of cascading failures and emergencies. In practical emergency situations, this system is anticipated to reduce total operating costs by more than 20%.

The planning and scheduling of power systems significantly influence their reliability and stability. Effective grid planning and scheduling ensure stable operation under diverse load conditions, reduce operational costs, optimize power resource allocation, and enhance power supply efficiency [18]. Dong et al. [19] proposed a two-level optimization configuration method to mitigate challenges like voltage standard violations, excessive currents, and power imbalances resulting from integrating distributed PV systems into distribution grids. This method effectively balances grid capacity, minimizes active power losses, and reduces operating costs. In addition to the two-level optimization configuration, game theory finds wide application in economic dispatch within power systems. For example, the master–slave game model is employed to analyze the supply–demand interaction between ‘grid-users’ and retailers for implementing demand response strategies [20,21]. These interactive games allow participants to continually share their interest information to achieve a Nash equilibrium solution that aligns with their respective interests. He et al. [22] utilized cooperative game (CG) methods to dynamically coordinate wind, PV, and thermal power generation scheduling, optimizing energy complementary transmission systems’ efficiency and facilitating the effective integration of new and traditional energy sources. Li et al. [23] utilized Stackelberg strategies to establish an equilibrium model for microgrid users and employed a multi-group genetic algorithm for iterative solutions. This model promotes user-centric energy utilization, enhances economic and system benefits, and effectively implements peak shaving and valley filling strategies. Current approaches to power system planning and scheduling are diverse, addressing uncertainties in renewable energy generation and decision-makers’ varying risk attitudes. Wang et al. [24] introduced a novel adaptive evaluation model for power system grid planning based on GRA-TOPSIS integrated CG and improved cumulative prospect theory.

The proliferation of electrical systems has led to an increasing demand for effective system state management. An interval state estimation method has been proposed that considers measurement correlations within these systems [25]. This approach involves comparing the state estimation interval with the system’s safe operating interval to ascertain whether the system’s safe operational range adequately covers the estimated state interval. Such methodology enhances situational awareness among system administrators, facilitating precise adjustments and control of energy systems as needed. Moreover, further advancements in system management efficiency can be achieved through the development of an intelligent question–answering system tailored for electrical system regulations. For instance, an enhanced BERTserini algorithm based on the BERT model is introduced for intelligent interpretation of electrical regulations [26]. This approach obviates the manual organization of professional question–answer pairs, thereby significantly reducing labor costs compared to conventional question–answering systems. Furthermore, it demonstrates superior accuracy and faster response times in providing solutions.

All in all, the urgency to enhance the stability and efficiency of power systems is growing, particularly with the large-scale integration of renewable energy sources into the grid. Renewable energy sources, such as wind and solar power, exhibit intermittent and fluctuating characteristics that challenge grid stability. As the focus on energy efficiency intensifies, there is a pressing need to leverage artificial intelligence technology and dynamic pricing mechanisms to balance supply and demand, thereby improving efficiency while maintaining economic viability. Looking ahead, the enhancement of power system stability will increasingly depend on advancements in artificial intelligence technology and big data analytics. These developments will enable more accurate forecasting and responsive measures, as well as optimized scheduling and planning. Concurrently, the advancement of novel control technologies and systems will facilitate precise control, ensuring high efficiency, economic feasibility, stability, and environmental sustainability. Such measures are critical for advancing the sustainable development of power systems.