Accurate and efficient multi-load forecasting is of great significance for the operation control and scheduling of integrated energy system（IES），in order to improve the load forecasting effect，a integrated energy system load prediction model based on least absolute shrinkage and selection operator（LASSO）and LSTM-GRU neural network was proposed. Firstly，in order to solve the problem of complex data caused by meteorological factors in the integrated energy system，a big data selection and analysis algorithm based on LASSO was studied to select and analyze the meteorological factors to obtain an effective data set. Secondly，the long short-term memory（LSTM）neural network was used to predict the system load，and the preliminary prediction value was obtained. Subsequently，the gated recurrent unit（GRU）was used to construct the error compensation model，and the compensation value of the prediction error was obtained through the training and learning of the prediction error. Finally，by reconstructing the output of the two，a more ideal prediction result was obtained. Through the simulation of the example，the proposed prediction model has higher prediction accuracy than the traditional LSTM neural network prediction model and the LSTM model optimized by particle swarm optimizer（PSO）.

The grid frequently exhibits weak grid characteristics because the impedance fluctuation range of the collector network is broad in high permeability distributed generation system. The coupling relationship between the phase-locked loop and the grid impedance leads to desynchronizing between the grid-following（GFL）converter and grid，which seriously threatens the stability of the system. However，the slow power response of grid-forming（GFM）converter is contradictory to the maximum power point tracking of source side，and the economy is poor. Consequently，in order to increase the grid-connected reliability of the system，some units are necessary to be configured flexibly，which will switch to the GFM mode. Concentrating on the grid-connected converters，a dual-mode adaptive flexible switching control strategy for distributed energy was proposed considering friendly interaction between grid and converters to maximize the utilization of new energy under the premise of system stability. The grid impedance identification algorithm based on non-characteristic harmonic injection was applied to sense the power grid strength，and the control strategy was adaptively switched according to the strength of the grid. Under the circumstance of robust grid，the constant power control method was adopted，which can quickly respond to the maximum power point instruction and improve the utilization rate of renewable energy. During the weak grid，converters flexibly switch to the virtual synchronous generator（VSG）control strategy to enhance inertia and damping support capabilities，realizing the friendly interaction between converters and grid. The proposed strategy enhances the robustness of grid-connected converters during the variation of grid strength，ensuring the stable operation of the system. The effectiveness of the proposed dual-mode control strategy was validated by PLECS simulation.

In order to solve the problem of parameter estimation in the high-performance control of squirrel cage asynchronous motor，a method for joint parameter identification of asynchronous motors with dual models based on improved whale algorithm was proposed. This method can effectively identify the stator resistance，the rotor resistance，mutual inductance and leakage inductance. In order to improve the identification accuracy of the algorithm，the nonlinear convergence factor was adopted，and the ideas of chaotic reverse learning，simulated annealing and adaptive mutation perturbation were integrated to overcome the shortcomings of the whale algorithm，which relied on the initial population，was easy to fall into local optimum，and had low convergence accuracy. Moreover，combining the advantages of the two traditional motor models，an improved dual-model joint identification was proposed，which further improves the accuracy of parameter identification. Based on this model，the improved whale algorithm was compared with the other two algorithms for motor parameter identification，and the experimental results show that the improved algorithm has high recognition accuracy，which proves the feasibility of applying the algorithm to identify the parameters of the squirrel cage asynchronous motor，and is of great significance for improving the control performance of the squirrel cage asynchronous motor.

A K-means clustering algorithm was proposed and a conditional Wasserstein generative adversarial network with gradient penalty（CWGAN-GP）to address the problem of imbalanced photovoltaic generation data caused by the low occurrence probability of extreme weather. A prediction approach combining bidirectional long short-term memory（BFLSTM）with convolutional neural network was introduced and incorporating channel attention mechanism to enhance the PV power prediction performance by integrating spatio-temporal features and dynamically adjusting the importance of feature channels. Firstly，correlation analysis and K-means algorithm were utilized to select and label various environmental factors. Then，extreme weather labels with fewer samples after clustering were selected，and CWGAN-GP was used for data augmentation.Finally，the augmented dataset was used to train the CNN-SE-BiLSTM prediction model for PV power prediction under extreme weather conditions.Simulation modeling was conducted using data from a certain PV power station，and the results demonstrate that augmenting the original extreme weather training set with CGAN-GP helps improve the prediction accuracy of the model. Moreover，CNN-SE-BiLSTM shows higher prediction accuracy among five weather categories compared to other traditional models，indicating that the proposed method is suitable for ultra-short-term photovoltaic power prediction.

A brake unit based on IGCT was designed for different applications of frequency converter in industrial scenes，especially the medium voltage frequency conversion system without feedback function. IGCT was used as the main power device，and the parameters and characteristics of IGCT was analyzed. According to the common topology of NPC medium voltage inverter in the market，the matching circuit of brake unit was designed，and the working principle of the system was described. In terms of device loss，the junction temperature of the power device was evaluated，and the water cooling circuit and the press assembly structure were designed. In order not to affect the midpoint balance of the DC bus voltage，an independent software control method was designed.Finally，the medium voltage brake unit designed has been applied to the occasion of rolling metal composite material in the indμstrial field，which confirmed the feasibility of the design scheme.

The traditional three-phase pulse width modulation（PWM）rectifier requires six power switches.Due to the shoot-through problem between two switches on the same bridge arm，the control difficulty is increased and overall system design becomes more complicated. A half-controlled three-phase PWM rectifier topology was used to achieve a three-phase rectifier. Only three power switches were used. The three switches were controlled by the same driving signal. The driving signal was generated by a dedicated controller，and the control circuit was simple. An experimental prototype was built using SiC MOSFETs as power switches. The experimental results show that the rectifier could operate at a higher switching frequency and the inductor and the capacitor are greatly reduced，thus the volume of the rectifierand the product cost are reduced，the conversion efficiency is improved and total harmonic distortion is low.

This research focused on the separation and calculation of electromagnetic losses in semi-direct drived permanent magnet wind generators（SDDPMWG）.Firstly，analytical method for calculating core loss of motor considering harmonic current was derived based on the Bertotti method. The eddy current loss of permanent magnet was obtained by analyzing axial and tangential components of the eddy current density of permanent magnet. The analytical methods for calculating DC copper loss，AC copper loss，and circulating current loss of the stator winding were also analyzed. Secondly，the changes of flux density over time in 6.5 MW SDDPMWG were analyzed according to finite element method. The spatial and temporal harmonics of flux density in the core were abundant. The impact of electromagnetic frequency on core loss of generator was calculated and analyzed. Then the relationship between eddy current loss of permanent magnet and electromagnetic frequency in 6.5 MW SDDPMWG was studied. Eddy current loss of permanent magnet increases with the increase of electrical frequency.Finally，finite element method and theoretical method were combined to calculate winding DC loss，AC loss and circulating current loss of 6.5 MW SDDPMWG quantitatively. The winding AC loss and circulating current loss are big.The research results can be effectively used for electromagnetic loss separation calculation of SDDPMWG. It can also provide support for the development of high-performance and low-cost SDDPMWG.

The structure parameters of dual-stator permanent magnet synchronous motors（DSPMSM）were optimized with finite element method and Taguchi method at the rated point，maximum torque point，and maximum speed point，respectively，aiming at the problem of large torque ripple of DSPMSM. The influence on electromagnetic torque performance of each optimization variable was also analyzed. Then，comprehensively considering the influence degree of different variables under each operation point and the proportion of each variable's influence degree to the total influence degree，the optimal combination of variables that can balance the electromagnetic torque performance under three operation points was finally obtained. The results indicate that the comprehensive optimization design method based on Taguchi method for multiple operation points can significantly reduce torque ripple and improve the electromagnetic performance of DSPMSM.

High-precision power converters are core components of ultra-precision motor motion control systems.Their output current harmonic distortion can cause unnecessary positioning fluctuations in ultra-precision motors，severely impacting the precision control system.Total harmonic distortion is an important index in high-performance precision control fields. A class of dual-buck topologies with no dead-time characteristics can eliminate the current distortion caused by dead-time in traditional bridge topologies and output high-quality，stable waveforms. To further improve the output current harmonic distortion of a class of dual-buck topologies，the sources of harmonic distortion in a class of dual-buck topologies were first clarified.For the harmonic distortion issues introduced by the power conversion stage and control system of a class of dual-buck topologies，the characteristics of low harmonic distortion a class of dual-buck topologies were summarized.The mechanisms of suppressing harmonic distortion by different control methods were revealed，and the relationship between various modulation strategies and harmonic distortion was studied. Finally，specific application scenarios for low-distortion a class of dual-buck converters were summarized，and reference directions for their future development were provided.

To address the supply-demand imbalance caused by renewable energy intermittency and load fluctuations in islanded microgrids，a two-layer optimization-based dynamic time-of-use（TOU）pricing scheduling model was proposed，aiming to enhance economic efficiency and operational stability.First，a comprehensive microgrid system model was established，integrating wind，photovoltaic，energy storage，marine energy，and diesel generators，while introducing a dynamic TOU pricing mechanism to guide user load behavior for peak shaving and valley filling.Subsequently，a two-layer optimization framework was constructed：the upper layer adjusts load distribution via dynamic pricing，and the lower layer optimizes generation dispatch and energy storage strategies to minimize operational costs，effectively resolving the limitations of traditional single-layer optimization in handling complex nonlinear constraints. Case studies demonstrate that the proposed model significantly reduces reliance on diesel generators，enhances adaptability to renewable energy fluctuations，and optimizes operational costs.The findings provide theoretical support for low-carbon scheduling of islanded microgrids，balancing economic and reliability objectives，and offer practical insights for advancing energy transition in island regions.