Wind power generation accounts for an increasing proportion of the power grid, so efficient and accurate real-time wind power prediction is particularly important for wind power grid. In view of the strong randomness and fluctuation of wind and the difficulty of predicting wind power, a Salp Swarm Algorithms-Extremely Learning Machine (SSA-ELM) based ultra-short-term wind power prediction model is proposed. In this case, the multi-input sample set is composed of historical wind speed, temperature, wind direction, atmospheric pressure and other climatic factors that are highly correlated with wind power, and the network parameters are determined in the training process. In order to improve the adaptability and accuracy of the prediction model, the input weight matrix and hidden layer deviation of the Extreme Learning Machine (ELM) are optimized by exploring and developing the Salp Swarm Algorithm in the iterative process. Finally, the simulation experiment is conducted with the actual data of a wind farm in Henan Province, and the comparison with the traditional Extreme Learning Machine, Particle Swarm Optimization Extreme Learning Machine (PSO-ELM) and Back Propagation (BP) neural network model shows that the new method avoids falling into the local extreme value, and has faster convergence speed and higher prediction accuracy.

The virtual synchronous generator (VSG) strategy addresses the challenges of low inertia and weak damping in a grid, exacerbated by the high penetration of renewable energy sources, by simulating the external characteristics of synchronous generators (SGs). However, the power electronic nature of VSGs can lead to a significant overshoot in the grid frequency response and issues with low-frequency oscillations when the system experiences disturbances. To mitigate these effects, this study proposes an improved self-adaptive control strategy to enhance VSGs’ frequency response characteristics of VSGs during grid connection and under perturbations, analyzed from a parametric perspective. Initially, the VSG active control model is employed to analyze the mechanism by which various parameters affect power frequency regulation during dynamic processes. The optimal initial values for virtual inertia and damping were determined using an enhanced particle swarm optimization algorithm. Subsequently, the impact of the power quality is incorporated into the selection of the fitness function. In addition, an inertia link was integrated into the control strategy, and the optimal value of the inertia coefficient was established through simulation. Finally, the proposed control strategy is compared with the fixed-parameter control strategy and classical adaptive control strategy under various operating conditions. The results indicate that substituting the initial value obtained from the improved algorithm nearly eliminates oscillations during a one-time frequency modulation. When the active power command of the system is increased by 20%, the overshoot of the frequency waveform using the improved strategy is reduced by 25.2% compared with the classical strategy. This suggests that the proposed control strategy facilitates a smoother transition of the system to the next state when subjected to disturbances, thereby enhancing the frequency response of the system.

As the penetration rate of renewable energy continues to rise, the proportion of grid-following (GFL) inverters based on phase-locked loop synchronization is steadily increasing. Under weak grid conditions, increased equivalent impedance at the grid connection point and reduced short-circuit ratio make GFL inverters more susceptible to disturbances, limiting their voltage/frequency support capabilities. In contrast, grid-forming (GFM) inverters actively regulate voltage and frequency, enhancing dynamic performance in weak grids through equivalent inertia and damping support. Hybrid operation combining GFM units controlled by Virtual Synchronous Generators (VSG) with GFL units employing PQ control can achieve complementary advantages in voltage/frequency support and power tracking efficiency. This enhances system support capability while maintaining operational economy. However, the dynamic interaction between these two control systems may induce new oscillation issues, potentially reducing system stability margins. Therefore, this paper constructs a full-order small-signal model of the GFL/GFM hybrid grid-connected system. By integrating the characteristic root locus method with the participation factor method, it analyzes the influence patterns of system and control parameter variations on key eigenvalues. This approach identifies critical parameter thresholds and determines the dominant state variables influencing each oscillation mode. Finally, time-domain simulations and RT-LAB hardware-in-the-loop experiments cross-validate critical dynamic responses—including common coupling point voltage, inverter frequency, and output power—to validate the constructed model and analytical conclusions. These findings provide theoretical foundations for capacity configuration, parameter tuning, and controller optimization in GFL/GFM systems, thereby enhancing stability under small disturbances.

We present a theoretical investigation on simultaneous generation of two pairs of Stokes and terahertz (THz) waves from coupled optical parametric oscillations (OPOs) with a quasi-phase-matching (QPM) scheme. The two pairs of Stokes and THz waves are generated by stimulated polariton scattering (SPS) from periodically-inverted GaP. By analyzing the QPM conditions of coupled OPOs we find that the two THz waves with any frequency below the transverse optical (TO) mode frequency of GaP can be simultaneously generated with a suitable pump wavelength. We calculate the photon flux densities of the two THz waves by solving the coupled wave equations. The calculation results indicate that the two THz waves can be efficiently generated with high pump intensities, particularly in lower THz frequency band.

The integration of large amounts of uncertainties in renewable energies into integrated electricity and heating systems gives rise to new challenges and threatens security. This paper presents a distributionally robust risk dispatch (DRRD) model designed to mitigate the risks arising from uncertainty, with the objective of achieving controllable risk and cost-effectiveness. Uncertainties associated with the wind power forecast error are encapsulated within an ambiguity set that encompasses moment (e.g., mean and covariance) and unimodal information to produce solutions that are more reliable and less conservative. The proposed model constrained by Conditional Value-at-Risk (CVaR) not only limits the probability of violations but also controls the magnitude of potential violations, improving the out-of-sample performance. We employ tractable reformulations and some approximation techniques to recast the resulting distributionally robust CVaR constraints as second-order cone (SOC) programs. Finally, the numerical outcomes presented utilizing the modified Barry Island system showcase the risk mitigation and economic advantages of the proposed methodology.