دانلود مقاله ISI انگلیسی شماره 25478
عنوان فارسی مقاله

بازپخت شبیه سازی شده ترکیبی و الگوریتم برنامه ریزی خطی عدد صحیح مختلط برنامه ریزی بهینه شبکه های توزیع شعاعی با تولید پراکنده

کد مقاله سال انتشار مقاله انگلیسی ترجمه فارسی تعداد کلمات
25478 2014 12 صفحه PDF سفارش دهید 8000 کلمه
خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.
عنوان انگلیسی
Hybrid simulated annealing and mixed integer linear programming algorithm for optimal planning of radial distribution networks with distributed generation
منبع

Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)

Journal : Electric Power Systems Research, Volume 108, March 2014, Pages 211–222

کلمات کلیدی
برنامه ریزی شبکه های توزیع - ژنراتور توزیع شده - روش تجزیه - برنامه ریزی خطی مخلوط عدد صحیح - بازپخت شبیه سازی شده -
پیش نمایش مقاله
پیش نمایش مقاله بازپخت شبیه سازی شده ترکیبی و الگوریتم برنامه ریزی خطی عدد صحیح مختلط برنامه ریزی بهینه شبکه های توزیع شعاعی با تولید پراکنده

چکیده انگلیسی

This paper presents a hybrid simulated annealing (SA) and mixed integer linear programming (MILP) approach for static expansion planning of radial distribution networks with distributed generators (DGs). The expansion planning problem is first modeled as MILP optimization problem with the goal of minimizing the investment cost, cost of losses, cost of customer interruptions due to failures at the branches and at DGs and the cost of lost DG production due to failures at branches. In order to reduce the complexity of planning problems the decomposition of the original problem is proposed into a number of sequences of sub-problems (local networks) that are solved using the MILP model. The decomposition and solution process is iteratively guided and controlled by the proposed SA algorithm that employs the proper intensification and diversification mechanism to obtain the minimum total cost solution.

مقدمه انگلیسی

Distribution expansion planning is a hard combinatorial optimization problem with a long history of contributions for improved solutions [1], [2] and [3]. One of the major characteristics of approaches proposed so far is whether they consider a single planning period or multi planning periods. The majority of models and approaches proposed for solving real-size multi-period planning problems produces a solution that is, among others, highly dependent on the effectiveness of static models integrated in the multi-period algorithm [4]. The models proposed for solving single-period (static) planning problems could be categorized as follows: mathematical programming based models, heuristic models and meta-heuristic models. Mathematical programming models, which can guarantee the optimality of the obtained solutions, are mostly based on mixed integer linear programming (MILP) [1], [5], [6], [7], [8], [9] and [10]. In [7] the MILP formulation of planning problems based on mixed integer conic programming and polyhedral relaxation is presented. The proposed approach enables accurate modeling of planning problems in which investment cost and cost of losses are considered (minimized). In [8] the MILP model is designed to minimize the investment and maintenance cost and cost of losses. The pool of solutions is obtained by varying relative optimality gap tolerance in the course of solving the MILP optimization problem and for each of them the cost of interruptions due to failures at branches is determined. The solution with minimum total cost becomes the best solution. The influence of distributed generators (DGs) at the planning process is discussed in [9] where the MILP model is presented with the goal of finding the solution with minimum investment and maintenance cost in the presence of DGs. In [10] the MILP model is designed to determine optimal type, size and allocation of DGs in radial distribution systems taking into account installation cost of different types of DGs and cost of energy supplied by the DGs and by the distribution system. Although continuous improvements are made, due to significant computational complexity the MILP models are not capable for solving large planning problems in reasonable time. Heuristic algorithms, [11], [12] and [13], although capable of finding “good” solutions (local optima) for real-size planning problems using relatively modest computational resources, do not guarantee the optimality of the obtained solutions. In order to improve the quality of heuristic methods, especially to overcome local optima, a numerous meta-heuristic algorithms are proposed [14], [15], [16], [17] and [18]. The model based on simulated annealing (SA) technique that finds a solution with minimum investment cost, cost of losses, and cost of interruptions due to failures at branches in passive radial distribution networks is presented in [16]. In [17] a combination of optimal power flow (OPF) and genetic algorithm is used to find the network development plan along with the sizes and sites of DGs that ensures minimal investment and operational cost and cost of interruptions due to failures at branches. A similar problem, where construction of tie-lines for improving reliability is considered, is handled in [18] by employing the modified particle swarm optimization technique. Although meta-heuristic algorithms may produce a better solution than heuristic ones, the quality of the obtained solution is uncertain, i.e. there is no guarantee of exactly how good the obtained solution is. A comprehensive survey of the above mentioned optimization approaches (heuristic, meta-heuristic, MILP), along with the analysis of their advantages and disadvantages, that have been applied in the other research areas in power systems could be found in [19], [20] and [21]. This paper proposes a new MILP model for static expansion planning of radial distribution networks with DGs that minimizes the investment cost, cost of losses, cost of customer interruptions due to failures at branches and at DGs and the cost of lost DG production due to failures at branches while taking into account a set of operational constraints (thermal constraints, voltage constraints, radiality constraints). In order to reduce the complexity of planning problems, which is significantly increased by taking into account failures at branches and especially at DGs, the decomposition algorithm based on SA technique is proposed. In the first step, the original problem is decomposed into a number of sub-problems (sub-networks/local networks) by employing the local network concept [22] and [23]. Each sub-problem is solved by applying the proposed MILP model and thus the initial solution of the original problem is obtained. This solution is further iteratively modified using the proposed simulated annealing algorithm that employs the proper intensification and diversification mechanism to search for the minimum cost solution. The obtained numerical results show that the proposed MILP model and the proposed decomposition approach (SA-MILP) can produce high quality solutions for static planning problems in radial distribution networks with DGs. The results also show that failures at DGs may have noticeable influence on the selection of the best expansion plan.

نتیجه گیری انگلیسی

The MILP model for expansion planning of radial distribution networks with DGs is presented that minimizes the investment cost, the cost of losses and the cost of interruptions due to failures at branches and at DGs and the cost of lost DG production due to failures at branches. In order to enable a solution of real-size planning problems to be obtained a new decomposition approach is proposed. This approach employs local network concept to decompose an original problem into a number of sub-problems (local networks) that are solved using the proposed MILP model. The solution process is iteratively guided by the SA algorithm. This algorithm employs intensification and diversification mechanism along with a non-monotonic temperature reduction scheme that ensures the most promising regions of the search space will be thoroughly exploited and the best solution will be found. The obtained numerical results show that the proposed approaches have the potential to be effective tool for obtaining minimum total cost solution for various size radial distribution networks with DGs. The results also highlight the necessity of considering failures at DGs in obtaining the best expansion plan in radial distribution networks.

خرید مقاله
پس از پرداخت، فوراً می توانید مقاله را دانلود فرمایید.