| In the northwest region of China, the population density is small, and the high energy consumption enterprises are widely distributed, and the user demand for power is diversified. However, because of the geographical conditions and the economic level, the electric power system construction level in this area is relatively low, and reliability of power supply is poor. Therefore, the development of all kinds of renewable energy according to the local conditions, and the improvement of microgrid energy supply mode using multi energy complementary will not only satisfy the users’needs of electricity, heat and other energy, but also have important significance to improve the efficiency level of the economic and the ecological environment. However, compared with the traditional power system, multi-objective scheduling of the micro grid shows their uniqueness and complexity, which makes the corresponding optimization problem with strong nonlinearity, various kinds of constraints, frequent changes on models and parameters. Therefore, it is very important to design the optimization strategy with strong flexibility, robustness and adaptability for the MGEES problems. In this paper, based on the evolutionary multi-objective optimization theory, the optimization strategies to solve the MGEES problems are studied as follows:(1) In order to realize the accurate management and scheduling of all kinds of micro sources and energy storage devices, the structure of microgrid and the operation mechanisms of various components are studied firstly. In this paper, a mathematical model of power supply, heating, charging/discharging of micro source and energy storage device is constructed according to the requirement characteristics of microgrid system and the diversity of various micro source and energy storage devices in the northwest region of China. On this basis, the MGEES objective functions are established including the minimum system comprehensive cost of the and the pollution emissions, considering the fuel consumption, the pollution emissions, the maintenance, the start-up of the unit, and the power from/to the main grid and other factors. Meanwhile, the typical constraints of MGEES are introduced, such as the capacity constraint, the power balance constraint, the controllable micro source climbing rate constraint, the minimum start/stop time constraint and so on.(2) The MOEA algorithm framework which is based on the multi-constraint handling strategy of microgrid (MG-MCMOEA) is proposed for the multi-constraint optimization problems of microgrid scheduling. The mechanism of the violation of constraints during the MOEA iterative process is analyzed, and the constraints violation is handled step by step. The step-by-step repairing strategy is designed to fix the violation of individual generation constraints, and the efficiency of the algorithm is improved by setting the infeasible solution repairing ratio. The equality constraints are transferred into corresponding inequalities, and the dimention of solution space is reduced. On the other hand, a self-adaptive ρg/μg-MOEA strategy is proposed to solve the multi-constraint optimization problems, which takes the corresponding methods to select individuals in the process of optimization. The framework makes the algorithm move forward to the feasible region gradually and obtains approximate feasible Pareto solution set eventually. On the basis of this, the MG-MCMOEA algorithm framework is designed to solve the multi-constraint optimization problem of microgrid.(3) In view of the problem that the distribution of MGEES solution is not balanced in the solution space and the optimization effect is not reliable with the scenario changes, the I-NSGAII-M2M algorithm is proposed to solve the multi-scenario problems of MGEES. The adaptive dynamic decomposition strategy is introduced. By changing the reference point, all of the computing resources are concentrated in the search space effectively, and the search in useless or infeasible region is avoided, and the feasible regions which have not been searched are protected. In this way, the ability of the algorithm is improved in solving engineering optimization problems with uncertain Pareto fronts with constraints. On the other hand, a heuristic based individual allocation method is proposed. The sub-population is made to select individuals from the nearest sub-space, thereby increasing the probability that the offsprings distribute in the original sub-space, and ensuring that the potential feasible solutions have greater priority in evolution by constraint handling criterions, which improves the diversity of feasible solution. At the same time, the MG-MCMOEA is combined with I-NSGAII-M2M to improve the ability of the algorithm to deal with the special constraints of MGEES.(4) Aiming at the problem that MOEAs cannot guarantee the reliability of optimizing MGEES, a MGEES comprehensive optimization strategy based on simplified MGEES knowledge base (SMGEES-KB) is proposed. In particular, by analyzing the characteristics of MOEA in solving the MGEES problems, the demand preference of model knowledge is summarized. According to "separation of the target-simplified model-combination of experience and optimization", the comprehensive optimization strategy of intelligent algorithm combined with mathematical programming is applied to extract SMGEES-K. And by using the second mutation and result self-correction, the extreme solutions are updated in solving MGEES problems. In addition, the decision maker’s preference based on I-NSGAII-M2M (PWV-INSGAII-M2M) is proposed, using SMGEES-K, decision information of compromise solution preference weight, and preference degree to obtain preference solutions close to the true Pareto front.The proposed optimization strategies are applied in a series of problems in MGEES, and the results show the conclusions as follows:By introducing MG-MCMOEA, I-NSGAII-M2M can effectively solve the multi-constraint multi-scene problems of MGEES, taking into account the minimizing cost of the system and the minimizing pollution emission simultaneously, and meet the decision-makers’diversified scheduling needs; By applying the optimization strategy based on SMGEES-K and PWV-INSGAII-M2M, the users can search the economic/ environmental protection scheduling solutions efficiently under different preferences, and need not to update optimization algorithm frequently because of MGEES problems changes. In summary, the comprehensive optimization strategies proposed in this paper is conducive to large-scale use of MOEA in practical MGEES problems, which can improve the reliability of the economic and ecological benefits in microgrid optimization, and has practical significance for the promotion of energy supply mode by microgrid in the northwest region of China. |
PDF Full Text Request