Search

With the aggravation of energy crisis and environmental problems, microgrid has become a research and application hotspot in the field of electric energy, which puts forward higher requirements for the security and reliability of microgrid, so the research on fault diagnosis of microgrid has become the focus. In this paper, the domestic and foreign fault diagnosis research is analysed. Based on Bayesian, the theory of variational sparse Bayesian and fuzzy h-network is studied. The variational sparse Bayesian fuzzy h-network is used for fault diagnosis of microgrid, and the fault diagnosis model of variational sparse Bayesian fuzzy h-network is constructed. Based on the analysis of Bayesian signal processing, a diagnosis method combining fuzzy h-network and variational sparse Bayesian is proposed. The variational sparse algorithm is used to improve Bayesian, reduce the complexity and computation of signal processing and solve the uncertainty problem of microgrid fault diagnosis. The dual fault signal filter based on Bayesian is designed, and the fault diagnosis method of microgrid based on variational sparse Bayesian fuzzy h-network is further studied and designed. The effectiveness of the proposed signal extraction method and fault diagnosis method is verified by simulation in Matlab/Simulink platform. In this paper, a new fault diagnosis method of microgrid based on variational sparse Bayesian fuzzy h-network is proposed, which improves the speed and accuracy of microgrid fault diagnosis.

It is proposed to adopt two kinds of storage battery equipment in microgrid, energy storage zinc bromine flow battery and power storage lithium titanate battery. The hybrid energy storage system adopts micro-grid DC/DC converter and DC/AC converter two-stage control. The hybrid energy storage DC/DC converter control adopts a distribution method based on power fluctuations and constant current fast charge control, and the hybrid energy storage DC/AC converter adopts PQ control and an improved droop uniform frequency control method. The algorithm automatically switches in all kinds of operation modes of the microgrid to achieve the power quality indicators, such as voltage sag, frequency deviation, voltage deviation and voltage fluctuation, during the operation of the microgrid to meet the operating requirements. The simulation results prove that the control method is correct.

Wind farm operation is characterized by large scale, fast and disorderly, the construction of grid structure is not complete, the fast adjustment of power supply in the power grid does not match to others, the objective laws of electric power system and uncontrollable and intermittent of wind power, and adjustable power supply capacity constraints in power grid, all of them cause the consumptive ability of wind power grid, which leads to more and more abandoned wind. This paper studies the tower neural network method, according to different height wind speed and wind direction data of historical wind tower measuring, combining with the fan power of historical observation data of wind farm, and building the neural network model, then the sample data will be input to the neural network model which has built to get theoretical power fan and abandoned wind power. By comparing the wind tower method, neural network method, model machine method with area integral method calculate abandoned wind power data, evaluate the effect that based on abandoned wind power tower evaluation model of neural network method in low wind speed has a good reference value, relatively close to the measured wind speed.

The influence on distribution network of the distributed generation (DG) is closely related to the location and capacity connected to the DG, therefore it is very important to study the problems of site selection and capacity determination. The reactive power optimization is considered to improve the voltage level of power grid and reduce line losses. According to the capacity of the DG, The problem about selecting site and determining capacity is decomposed into two kinds of pattern with Parallel operation and island operation, which means comprehensive target of grid losses. voltage deviation and static voltage stability margin is optimal while DG is full load. Optimal allocation model of DG is established, the constraints condition includes node voltage, the transmission current of line, the single DG capacity and the total capacity of all DG. The genetic — ant colony hybrid intelligent algorithm is put forward to calculate the location and capacity of the DG. Simulation results show that the method is effective in problem of location selecting and capacity determing of DG.

Traditional power distributed network is transformed from single power supply radial network to more than one power supply loop network, direction of the power flow is changed, Therefore action of right protection cannot be completed in the micro-grid operating in the grid-connecting mode and two kinds of operation mode switching of interconnection and lonely island. The four standard of distribution system protection namely, selectivity, reliability, sensitivity and quick action cannot be effectively meet. In paper, influence on relay protection of the micro-grid is expounded in order to solve the problem of micro power grid relay protection, according to the load sensitivity of feeder circuit, current differential protection module and inverse time over-current protection module are established, the parallel operation protection strategy is designed, the simulation results with the typical micro-grid show that performance of the proposed protection module is more rationality and effectiveness.

For direct drive synchronous wind power generation system, the converter of direct drive motor is connected to the stator side of permanent magnet synchronous generator side, through controlling the converter, the maximum wind energy can be captured. Main circuit design of the direct-drive grid-converter is studied. Firstly, power grid side filter is designed and calculated, LCL filter is used instead of traditional L type filter, in this way, the value of the inductance、 the volume and cost are reduced effectively, and its inhibitory effect on high frequency signal is more obvious. Not only meet the vector triangle, but also grid connection harmonic current is restrained. Secondly, the filter of motor side is also designed, about the problem of voltage reflection for long cable drive motor PWM converter on the motor side, the solution is to reduce the output value of the du/dt and increase the pulse rise time, at the same time the reflected voltage of the motor side is also reduced. Finally ,Brake unit is researched, at the instant of the power grid voltage dips is detected, chopper control is started to IGBT of brake unit, making the machine side power in addition to the net from the side of the part, the rest are consumed in the brake unit, to ensure that the equipment can grid-connected operated. The last but one, the pre-charging circuit is included.