| Recently, global increasing energy shortage and environmental pollution have become serious. But reviews our country, the above problems are most prominent. In China, thermal power is the most important part in the electric power production. As a traditional way of power generation, the thermal power utilizes coal to produce electric energy. On the one side, coal is a kind of non-renewable fossil resource, and a large amount of load demand increasing in our country has further aggravated the problem of coal resource starvation. On the other side, for a large amount of noxious gas such as carbon dioxide and sulfur dioxide emitted by coal burning, damage of greenhouse and haze fog have revealed gradually. Consequently, our country needs a new resource to participate in electrical production to achieve the goal of national sustainable development. Since distributed photovoltaic generation has the advantages of environmental friendly, wide distribution, and abundant energy, it has developed rapidly, and has attracted widely attention of the world. However, output power generated from the distributed photovoltaic generation can be affected by the radiation intensity and environmental temperature, the distributed photovoltaic generation output power affected by the environmental influence can be far away from the maximum power point, which will decrease the generating efficiency. The distributed photovoltaic generation in the situation of islanding could still supply power to the utility, which will bring safety problems to the electric equipment and maintenance personnel. Moreover, if a large amount of distributed photovoltaic generation is connected in distribution network, and power flow of the distribution network can be change, which has a strong impact on network loss of distribution network. Network loss of distribution network can cause transmission line heating and accelerate ageing of transmission line, and it can affect safe operation of the distribution network seriously, the heating loss caused by the network loss can also waste huge electric power. Therefore, output power dispatching of the distributed photovoltaic generation should take into full account of the affection of the network loss, but uncertainty of output power of the distributed photovoltaic generation and power load increase the difficulty of optimal dispatching. Therefore, both control technique and output power optimal dispatching of the distributed photovoltaic generation are very significant to the wide connection of distributed photovoltaic generation in the distribution network. The main investigation contents include improving the adaptability of environment, increasing the efficiency of islanding detection and decreasing the network energy loss, and we study on the significant research of control and dispatching methods in this paper for the distributed photovoltaic generation.In allusion to the problem of environmental influence on the distributed photovoltaic generation, this paper proposes a two-stage maximum power point tracking algorithm based conductance ripple correlation. In order to adapt fast to the environmental variation of radiation and temperature, the proposed algorithm utilizes the math relationship among the conductance at maximum power point, radiation and environmental temperature to approach to the maximum power point in the first stage. In order to reduce oscillation effectively, the conductance defined as control object in the ripple correlation algorithm is utilized to track the maximum power point precisely in the second stage. To verify effectiveness of grid connected control method and feasibility of power regulation, photovoltaic grid connected inverter is designed and completed, it can be seen from the system test results that the platform conform to the grid connected standard.The islanding state of distributed photovoltaic generation will cause serious safety problem, so islanding detection is an indispensable technique when interconnecting distributed photovoltaic generation to the distribution network. The mainly islanding detection methods can be divided into passive detection methods and active detection methods. The passive detection methods are utilized to detect whether the system parameters such as voltage or frequency are within the normal range, if these system parameters are within the normal range when system in the islanding state, the passive methods can hardly detect the islanding state, so the passive methods have a large non-detection-zone. The active detection methods inject the perturbation to inverter current to detect islanding state and these methods can inject more current harmonic to the distribution network, so power quality of the grid can be polluted. In allusion to the problem of large non-detection zone and harmonic damage in the existing islanding detection methods, this paper proposes an islanding detection method based on code match of sequence perturbation. When the distributed photovoltaic generation keeps in the state of islanding, the inverter current and load voltage perform a strong correlation. With respect to the correlation, this paper uses sequence to perturb the amplitude of the inverter current periodically and calculate the islanding detection coefficient. This paper can detect the islanding state by periodic match degree of the islanding detection coefficient and the code in the sequence. In the meantime, this paper demonstrates the selecting principle of length and code pattern of the perturbing sequence, and also analyzes the effect of load variation on the islanding detection method. Based on the above content, islanding detection method based on code match of sequence perturbation can bring little harmonic to the grid, and can detect the islanding state effectively within a certainty range of load variation.Based on the demand of a large amount of the distributed photovoltaic generation connected in the distribution network and problem of output optimization, an output power optimization method based on K-means cluster for distributed photovoltaic generation is proposed. Based on the optimal target of network loss minimization, the output power of the distributed photovoltaic generation is optimized by the PSO method. To solve low optimal accuracy problem of the PSO method caused by initialized random solution, this paper employs cluster obtained by the K-means clustering method and formulas of the output power optimization to initialize solution of particle. Considering with the time-varying variation of power generated from the distributed photovoltaic generation and power load, the PSO based on K-means clustering initialization utilizes history data at each hour to optimize the distributed photovoltaic output power. Compared with the traditional PSO method of output power optimization, the PSO based K-means method for output power can reduce more annual network loss.Based on the problem of uncertainty of the power generated from distributed photovoltaic generation and power load, day-ahead dispatching method is proposed for optimizing the output power of distributed photovoltaic generation. In order to give full play to positive effect of the distributed photovoltaic generation on distribution network, the forecasting data is used to dispatch output power of the distributed photovoltaic generation ahead of time. The auto regressive moving average (ARMA) model with time-sharing coefficient is used to obtain the forecasting clearness index, environmental temperature and power load, the distributed photovoltaic generation output power is obtained by the formulation of clearness index and environmental temperature. Based on the target of hourly network loss minimization, the sequence quadratic programming or dynamic programming methods with the forecasting data is used to dispatch the output power of the distributed photovoltaic generation ahead of time, hourly network loss can be reduced effectively, the day-ahead dispatch of output power can be achieved. |
PDF Full Text Request