As the penetration rate of household distributed PV in China's low-voltage distribution network continues to increase，the power flow of the distribution network has changed，leading to more serious problems such as node voltage overruns and three-phase imbalance，which poses a serious threat to the safe and stable operation of the distribution network. For situations where the distribution network topology and line parameters are unknown，the traditional voltage control strategy based on power flow calculation is no longer applicable， a data-driven voltage coordination control strategy was proposed for low-voltage distribution networks with high household photovoltaics penetration. First，a linear approximation model of the low-voltage distribution network was established，and the least squares method was applied to fit the relationship between node power and voltage based on the historical operation data of the low-voltage distribution network. Then，the photovoltaic inverter and energy storage system were used as regulation measures to minimize the degree of node voltage over-limit，minimize the three-phase unbalance and minimize the amount of equipment regulation as the objective function. The improved multi-objective particle swarm algorithm was utilized to achieve voltage optimization control. Finally，the effectiveness of the proposed voltage coordination control strategy was verified by simulation comparison and analysis with other control strategies and methods，taking the 21-node low-voltage station as an example.

With the wide application of solar photovoltaic panels，the detection of hidden dangers and faults of photovoltaic panels has become more and more important. In order to improve the inspection efficiency of photovoltaic panels and the accuracy of defect detection，a target detection algorithm based on attention mechanism was designed and a complete inspection scheme of photovoltaic panels combined with UAV technology was proposed. The method uses the high-resolution camera mounted on the UAV for image acquisition，and detects and identifies the defects of photovoltaic panels through the combination of the object detection of the attention mechanism and the UAV technology. The experimental results show that the proposed method has high accuracy and reliability in photovoltaic panel inspection，mean average accuracy （mAP） reaches 83.2%，F1 score is 84.5%，effectively improve the quality and effect of photovoltaic panel inspection，and has high practical value.

Modular multilevel matrix converter （M3C） is a multilevel topology that can directly realize AC-AC conversion，which has good application prospects in high-voltage transmission systems and DC distribution systems，but the existence of bridge-arm circulating channels inside the M3C leads to a very complex control structure. For this reason，a hierarchical system control strategy was proposed. First，the mathematical model of the M3C was derived based on the conventional double αβ0 coordinate transformation. Then，a dual-loop control structure consisting of an outer-loop and an inner-loop controller was employed in the M3C，which is usually employed in voltage source converters. In addition，in order to suppress the loop current component inside the system，a bridge arm energy equalization control strategy based on double αβ0 transform decoupling was proposed，through which the capacitive voltage control between the bridge arm and the bridge arm was realized by this control and carrier phase shift modulation. Ultimately，the simulation model and physical prototype were constructed，and the outcomes validated the viability of the proposed control strategy for the M3C system.

With the increasing penetration of new energy equipment，while alleviating the environmental problems caused by fossil energy，it also increases the frequency instability of the system. Photovoltaic （PV）and other new energy equipment to participate in the grid frequency regulation is an effective way to improve the frequency stability of the power system.First，the basic principle of frequency response of synchronous machine was analyzed，the correspondence was determined which is between active output of PV power generation and primary frequency regulation and inertia response，and the virtual synchronous control strategy of PV power generation was formed.Then，based on the power-frequency correspondence relationship of primary frequency regulation，it gave the parameter adjustment method that directly corresponds to the frequency and the load shedding rate，and realized the function of primary frequency regulation. Once again，based on the inertia response minus the frequency-power correspondence relationship，it gave the inertia response minus the frequency-power correspondence relationship，which is the most effective way to improve frequency stability of power system. Once again，based on the correspondence between inertia response frequency reduction and power，the parameter setting method of inertia response load shedding rate was given，and the negative influence of inertia in the process of frequency restoration was avoided through the determination of application conditions of inertia response. Finally，the PV grid-connected model was constructed，which verified that the PV load shedding operation has the ability to participate in the frequency regulation of the power grid.

The islanding detection method is a necessary method for grid-connected PV systems，but，the current islanding detection method is prone to misjudgement when encountering grid operation states such as high-resistance short-circuit ground faults and large-capacity load casting. For this reason，an islanding detection method based on reactive power perturbation and featured kurtosis was proposed based on analyzing the difference between the power at the grid point in the islanding operation state and other grid operation states. The method firstly used the three-phase voltage amplitude offset rate at the grid point to realize the starting criterion of active reactive power injection in PV system. Then，the apparent power waveform at the grid point was obtained，and after the normalization operation was performed to obtain the normalized apparent power （NAP） waveform，variational modal decomposition algorithm was used to decompose NAP at 8 layers，and the time-frequency component at layer 5 was obtained as the feature detection component. And then the feature detection component was characterized by a featured kurtosis to achieve islanding detection. Finally，a typical PV grid-connected system was constructed using Matlab simulation platform，and the effectiveness of the proposed method was tested through the working conditions of islanding operation with different quality factor loads，different high-resistance short-circuit ground faults，large-capacity load switching，and lightning strike faults，and so on.

Focusing on the central position of AC power frequency converter in the development of the national economy，emphasizingd their critical impact on the safety and efficiency of industrial production. In view of the rapid development of power electronics technology，the field of AC power frequency converter has made remarkable progress，but also faces many challenges，such as uneven product quality，and there is an urgent need to optimize the industry order to improve the overall quality level. By comparing and analyzing the development status quo of power supply systems at home and abroad，the high standard of product reliability required by the international community was revealed，and provided an in-depth analysis of the problems that exist in China's AC power frequency converter products in terms of reliability，standardization and regulation. In order to promote the high-quality development of the industry，a series of strategies were put forward：strengthening product reliability technology research and development，building a standardized evaluation system，ensuring the effective implementation of standards，promoting the power supply product certification system，and strengthening the operation and maintenance regulatory mechanism. These comprehensive measures are aimed at optimizing the industry ecosystem，improving product quality and safety，and thus promoting the high-quality，sustainable development of the AC power frequency converter industry.

Aiming at the problem that the circulating currents analysis of modular multilevel matrix converter（M3C） is not deep enough and the frequency component of circulating currents is not clear，the coupling process of submodule capacitor voltage fluctuation，bridge voltage fluctuation and circulating currents in M3C was deeply analyzed，and the primary coupling model of M3C circulating currents and its four main frequency components were obtained. The analysis results provided a theoretical foundation for using PR controller to suppress the circulating currents of corresponding frequency components. The theoretical analysis and suppression of circulating currents were verified by experiments. The experimental results prove that the theoretical analysis of the frequency components of circulating currents is correct，and it is also proved that the PR controller designed according to the analysis results has better suppression effect on circulating currents than the P controller.

An anti-interference domestic variable frequency multi-drive system solution for the five-stand continuous cold rolling line in the steel strip cold rolling field was presented. Firstly，the mechanical equipment and rolling process of the 900 mm five-stand continuous rolling line was introduced，and matched the models of domestic AC drive devices according to the motor table. The parallel connection mode of modular rectifier feedback power supply and inverter and the fiber optic communication were briefly described. Then，the problem of instantaneous drop of the incoming grid voltage encountered by the common DC bus system and how to improve the system stability by adding capacitor banks were discussed. At the same time，the method of using encoder isolation distribution board was introduced and providing separate power supply for the encoder to solve the encoder signal interference problem. Finally，the anti-interference performance of the system was verified by comparing the speed and torque accuracy before and after the improvement.

Aiming at the problem that the photovoltaic grid-connected inverter system operates under multiple disturbances，an linear active disturbance rejection control （LADRC） grid-connected strategy combined with photovoltaic system was proposed.Firstly，the dynamic process of PV grid-connected system was analyzed and modeled，and the LADRC-PM structure was designed according to the model characteristics of the photovoltaic system，further improving the speed on the basis of the classical structure. And then，the structure was introduced on the basis of the double closed-loop control framework，so that the system could compensate the disturbance through the feedforward way to enhance the system ability to suppress the uncertainty. Furthermore，the mathematical model of the proposed control strategy was dynamically analyzed using frequency domain analysis method，and its engineering stability，tracking convergence，and disturbance rejection superiority were theoretically verified，and the guidance of adjusting some parameters was given. Finally，the advantages of LADRC-PM control strategy was verified by the Matlab/Simulink simulation in various working conditions.

A fusion prediction method was proposed to predict and correct the calculation deviation of the top transformer oil temperature model in IEEE guideline，so as to realize the more precise prediction of the transformer top oil temperature（TOT）.Firstly，the characteristics of the transformer TOT model and the extreme learning machine（ELM） prediction model was introduced. Secondly，in order to avoid the problem of slow operation speed caused by double level intelligent prediction，the weighted multi-point extrapolation method combined with the load curve clustering algorithm was used to obtain the future load coefficient of the transformer which introduced as the load prediction level of the model. Finally，based on the calculation of thermal model，which the ELM was used to predict the deviation between the calculated value of thermal model and the measured value，and finally the accurate predicted value of the TOT of the transformer was obtained.The simulation platform was built and the simulation results show that the average prediction error rate of the proposed prediction method is only 0.59%，and the root mean square error is only 0.47 ℃. Compared with the other three methods，it has higher prediction accuracy and stability. The model training speed and prediction speed are only 1.21 ms and 0.39 ms，respectively，which proves that the fusion prediction model proposed and established has high prediction accuracy，stability and operation speed.