Research On Optimization Model And Algorithmfor AC/DC Interconnected Power Grid’s Operation

With the construction of HVDC transmission system and interconnection of regional power networks, the scale and the complexity of topology in AC/DC interconnected power grid increase. As it relates to the transmission of a large capacity power, the optimization of each component in AC-DC interconnected power grid operation can create enormous social benefits, which makes the optimal operation of AC/DC interconnected power grid becomes an important issue. On one hand, with the promotion of conserving energy and reducing emissions, multiple optimization targets need to be considered while wind power integration makes the AC/DC interconnected power grid scheduling problem become more complicate. On the other hand, the optimization results considering regulating times of DC transmission power in a whole day is more compatible with the real operation, and the reliability of the day-ahead data is closely related to the accuracy of AC/DC parallel transmission corridor’s active power allocation. Therefore, the optimal operation of an AC/DC interconnected power grid has been investigated deeply in this paper from four aspects such as multi-objective optimization, stochastic optimization, dynamic reactive power optimization and the active power allocation of AC/DC parallel transmission corridor. Details are givenas follows:1) A piecewise normalized normal constraint method to solving multi-objective reactive power optimization problem of AC/DC interconnection grid is proposed. In order to minimize both active power loss and square sum of voltage deviation in all key nodes, a multi-objective reactive power optimization model of AC/DC interconnection grid is built and a piecewise normalized normal constraint method is proposed to obtain Pareto optimal solutions of this model. The proposed method, considering the difference of the angle between Utopia line and tangent line of each Pareto frontier point, is an improvement of the normalized normal constraint method. In this piecewise normalized normal constraint method, the Pareto frontier is divided into four segments according to certain rule, each segment corresponds to a sub-Utopia line, and hence the original multi-objective optimization problem is translated to four sets of multi-objective optimization problems. The four sub-Utopia lines can be divided by equidistant respectively, and each set of multi-objective optimization problem can be converted to solving a series of single objective optimization problems. The numerical results demonstrate that the proposed method has the capability to generate a more evenly distributed set of Pareto solutions when compared with the normalized normalconstraint and weighted sum methods. Thus, it can provide more comprehensive information for operators.2) Piecewise normalized normal constraint method and extreme scenario method are applied to solving multi-objective optimal power flow problems of AC/DC interconnected grid with wind power integration. In order to reduce the solving difficulty, stochastic scenarios are replaced by extreme scenarios to deal with fluctuations of wind power outputs, and a multi-objective optimal power flow model of AC/DC interconnection grid with wind power integration is built. This multi-objective optimization problem can be converted to a series of single objective optimization problems by applying piecewise normalized normal constraint method. When applying nonlinear primal-dual interior-point algorithm solving single objective optimization problem, the coefficient matrix of the reduced correction equation can be arranged in a bordered diagonal structure according to the order of the scenes. By means of the sparse storage and elimination technique, the demand of storage and the calculation among non-zero element can be reduced. The numerical results show the proposed method is feasible in solving the stochastic optimization problems of real AC/DC interconnection grid.3) A method for solution of dynamic reactive power optimization problemof AC/DC interconnection grid considering regulating times of DC transmission linesis proposed. To avoid the regulating power of DC transmission lines operating frequently, the dynamic reactive power optimization model of AC/DC interconnection grid is built. In the generalized Benders decomposition process, the power variables of DC transmission lines is divided in the master problem together with integer variables, therefore the sub-problem can be transformed into a set of independent optimization problems. Also, a tighten Benders cut is used to represent the sub-problem information into the master problem. To obtain the optimal power of DC transmission lines in a successive linear way, a dynamic step adjustment strategy is proposed. Results demonstrate the correctness and effectiveness of the proposed method in restricting regulating times of DC power. Also, the effect of different maximum regulating times constraints on dynamic reactive-power optimization results is analyzed.4) A method for optimal active power allocation of AC/DC parallel transmission corridors is proposed. In order to improve the finesse and accuracy of power allocation schedule among DC lines and AC lines in day-ahead dispatch plan, a framework to generate power grid operation data in the next day is presented based on dispatcher power flow data and dispatch management information system data. Therefore, the optimize power allocation among DC lines and AC lines can be performed. The application results show that the proposed method can effectively reduce the power loss in day-ahead operation and can provide decision support to realize the optimum operation of AC/DC interconnection grid.