Back to the programme printer.gif Print

Delegates are invited to meet and discuss with the poster presenters in this topic directly after the session 'How does the wind blow behind wind turbines and in wind farms?' taking place on Tuesday, 11 March 2014 at 16:30-18:00. The meet-the-authors will take place in the poster area.

Diana Manjarrés Tecnalia Research & Innovation, Spain
Diana Manjarrés (1) F P Valentin Sanchéz (1) Javier Del Ser (1) Naïma Vande (2)
(1) Tecnalia Research & Innovation, Zamudio, Spain (2) 3E , Brussels, Belgium

Printer friendly version: printer.gif Print

Presenter's biography

Biographies are supplied directly by presenters at EWEA 2014 and are published here unedited

Diana Manjarrés, PhD. Researcher at the OPTIMA Unit of TECNALIA RESEARCH & INNOVATION working in the area of Telecommunications and Artificial Intelligence. Her research focuses on heuristic techniques for optimization problems related to different fields of knowledge. During her 4-year-long research career she has coauthored several scientific publications in various international journals (Applied Soft Computing, Engineering Applications of Artificial Intelligence, Expert Systems with applications…) and a number of contributions to renowned conferences such as MACOM, ISDA, etc.




During the last years wind energy has experimented a significant growth in comparison with other types of renewable energy sources. Accordingly, the number of wind farms has increased sharply to become one of the most developed worldwide infrastructures specially in the Southern Europe. Unfortunately, the high number of constraints and restrictions that must be considered nowadays when designing a wind farm deployment (e.g. protected environmental areas or geographical unfeasibility) calls for tools aimed at the cost-optimal placement of wind farms, along with an optimized micro-siting of their compounding wind turbines.


From an algorithmic point of view, the optimal micro-siting of wind turbines under multiple constraints entails a complex optimization problem, difficult to be solved by exact calculation methods. Thus, research efforts have so far delved into the derivation of cost-effective wind farm deployments by means of near-optimal evolutionary approaches applied to formulations where the maximization of the yield production is the unique optimization objective. Nevertheless, maximizing the yield production often implies an increase in the overall cost of the wind farm project due to the need for more involved road infrastructure, cabling, etc. In this context, an algorithm capable to balance differently the tradeoff between both conflicting metrics -- yield production and cost -- would provide the project designer with a set of possible wind farm deployments from which the final layout is selected according to specific requirements of the project at hand. Bearing this in mind, this research work proposes a novel multi-objective Harmony Search algorithm capable of 1) estimating such a set of wind farm deployments that differently trade yield production for cost; and 2) considering relevant factors that considerably affect the viability of a wind farm project, such as terrain constraints, environmental issues, restrictions related to radio communication infrastructure, archaeological heritage and visual impact, among others. To this end, information on these constraints is retrieved from an orchestrated database, based on which a set of forbidden areas are set and excluded from the overall search space, so as to place wind turbines on feasible zones. In addition, the derived algorithmic approach also incorporates a local search procedure by which wake effects between nearby turbines are mitigated and hence, an enhanced yield production is obtained. In order to validate the proposed multi-objective HS approach, an exhaustive performance assessment test is performed by means of several Monte Carlo realizations in a real scenario in the Basque Country (northern Spain). The good results obtained in these simulations highlight the practical applicability of the proposed algorithm for rendering the aforementioned set of Pareto optimal wind turbine placements, and fosters its utilization in cost-effective micro-sitting of wind turbines.

Main body of abstract

In the last year wind power underwent an annual market growth of almost 10% and a cumulative capacity of about 19%, and is expected to uphold this trend in forthcoming years. Thus, worldwide installed wind power will exceed 300 Gigawatts of capacity this year with a total of 225,000 turbines placed globally and an increase forecast to almost 500 Gigawatts by 2016 [1]. This significant growth finds its most recent roots in the growing prices of the traditional fossil fuel, the environmental concern (Kyoto protocol [2]), and the institutional support to renewable energy sources. In this context, wind energy has experimented the highest growth in comparison to other types of renewable energy, with most of the wind power consumed in the world being generated in large wind farms. This rationale positions wind energy as one of the most promising sources of renewable energy, and as an important stakeholder in the energetic mix of different countries, which are definitely betting for its development.

Nevertheless, the installation of wind farms and the micro-siting of wind turbines involve several important factors that must be carefully considered when starting a wind farm project, such as detailed terrain constraints, environmental issues and restrictions related to radio communication infrastructure and visual impact, among others. In addition, maximizing the yield production is one of the main goals of any wind farm deployment, even though it generally stands for an increase in the overall cost of the project. In this context, the SOPCAWIND (Software for the Optimal Place CAlculation for WIND farms) FP7 European project [3] aims at providing a specific tool for the optimum location of wind turbines within an area previously selected by the user. On this purpose, a meta-heuristic optimization algorithm is developed. Specifically, the optimization approach is based on the Harmony Search (HS) algorithm, a recently proposed population-based meta-heuristic algorithm which has obtained excellent results in the field of combinatorial optimization [4]. It mimics the behavior of a music orchestra when aiming at composing the most harmonious melody. As previously mentioned, HS is a population-based algorithm; it hence maintains a set of solutions in the so-called Harmony Memory (HM). An estimation of the optimal solution (i.e. the positions of the wind turbines compounding the wind farm at hand) is achieved at every iteration by applying a set of optimization parameters to the HM, which produces a new harmony vector every time. Due to its potential characteristics, HS has been widely applied for solving several optimization problems in different application fields, such as vehicle routing [5], multicast routing [6], multiuser detection [7,8], engineering design [9], radio resource allocation [10] and the access node location problem [11].

Unfortunately, optimal performance according to one objective function often implies performance degradation in other objectives characterizing the problem at hand. Indeed this occurs when including cost as another objective conflicting with the yield production when optimally placing wind turbines, which gives rise to a multi-objective problem formulation. Consequently, the aforementioned single-objective solver for maximizing the yield production of a certain wind farm layout is evolved to a multi-objective Harmony Search (HS) algorithm capable of balancing the trade-offs between two objective functions: the maximization of the yield production and the minimization of the cost of the deployment. As a result, a set of Pareto-optimal solutions is obtained from which the desired solution is selected according to specific requirements of the wind farm project under consideration [12]. Moreover, the algorithm is designed so as to automatically exclude areas in which the placement of a wind turbine is not feasible due to its potential impact on the environment, radio-communication services or archaeological heritage, among others. The tool also incorporates a local search procedure, specially designed for wind farm deployment, that prevents the produced solutions from wake-effects between nearby turbines that might impact on the final yield production. The proposed multi-objective HS approach is exhaustively tested by means of several Monte Carlo realizations in a real-based scenario in the Basque Country (northern Spain). In light of the obtained results, the multi-objective HS algorithm is proven to render a wide, diverse set of wind turbine arrangements differently balancing yield and cost, of utmost relevance in any wind farm project potentially subject to location constraints.


Wind energy plays an essential role in the worldwide energy supply, and the number of wind farms is expected to increase accordingly in forthcoming years. From the designer's perspective, the main goal when harvesting wind energy is to sketch wind farm layouts by 1) meeting all existing geographical constraints; 2) maximizing the obtained yield production; and 3) minimizing the cost of the deployment. In this context, this paper presents a novel meta-heuristic approach based on the Harmony Search algorithm for efficiently solving the problem of optimally deploying turbines in wind farms under the aforementioned criteria and constraint set. The proposed optimization tool considers different conflicting objective functions, i.e. yield production and cost of the deployment, which gives rise to a multi-objective formulation of the problem. In order to effectively obtain different tradeoffs between both objective functions, a novel multi-objective adaptation of the Harmony Search algorithm is designed and developed. Additional mechanisms to counteract possible wake-effects between neighboring turbines are investigated and incorporated to the tool by means of local search procedures designed ad-hoc for the problem at hand. The performance of the obtained multi-objective HS algorithm is assessed by several Monte Carlo realizations in a real-based scenario in the Basque Country (northern Spain). The achieved results provide a diverse set of Pareto optimal solutions representing the encountered wind farm layouts within a predefined project area. Therefore, the proposed tool will help any given wind farm designer to optimally select the layout of the wind farm according to the specific requirements of the project and existing location constraints of the region.

Learning objectives
In this paper a novel multi-objective adaptation of the Harmony Search algorithm is developed and tested for efficiently solving the problem of optimally deploying wind turbines in wind farms by simultaneously addressing two conflicting objectives, i.e. the yield production and the cost of the deployment.

[1] Global Wind Report. Annual Market Update 2012. Retrieved on October 2013.
[2] Kyoto protocol. Retrieved on October 2013.
[3] SOPCAWIND FP7 Project. Retrieved on October 2013.
[4] Z. W Geem, J. H Kim, G. V Loganathan. A New Heuristic Optimization Algorithm: Harmony Search, Simulation, 2001.
[5] Z. W. Geem, K. S. Lee, Y. Park. Application of Harmony Search to Vehicle Routing, American Journal of Applied Sciences, vol. 2, pp. 1552-1557, 2005.
[6] R. Forsati, A. T. Haghighat, M. Mahdavi, Harmony Search Based Algorithms for Bandwidth-Delay-Constrained Least-Cost Multicast Routing, Computer Communications, vol. 31, no. 10, pp. 2505_2519, 2008.

[7] R. Zhang, L. Hanzo, Iterative Multiuser Detection and Channel Decoding for DS-CDMA Using Harmony Search, IEEE Signal Processing Letters, vol. 16, no. 10, pp. 917-920, 2009.

[8] S. Gil-Lopez, J. Del Ser, I. Olabarrieta, A Novel Heuristic Algorithm for Multiuser Detection in Synchronous CDMA Wireless Sensor Networks, IEEE International Conference on Ultra Modern Communications, pp. 1-6, 2009.

[9] T. W. Liao, Two Hybrid Differential Evolution Algorithms for Engineering Design Optimization, Applied Soft Computing, vol. 10, no. 4, pp. 1188-1199, 2010.

[10] J. Del Ser, M. Matinmikko, S. Gil-Lopez, M. Mustonen, Centralized and Distributed
Spectrum Channel Assignment in Cognitive Wireless Networks: A Harmony Search Approach, Applied Soft Computing, vol. 12, no. 2, pp. 921-930, 2012.

[11] I. Landa-Torres, S. Gil-Lopez, S. Salcedo-Sanz, J. Del Ser, J. A. Portilla-Figueras, A Novel Grouping Harmony Search Algorithm for the Multiple-Type Access Node Location Problem, Expert Systems with Applications, vol. 39, no. 5, pp. 5262-5270, 2012.

[12] E. Zitzler, L. Thiele, Multiobjective Evolutionary Algorithms: A Comparative Case Study and the Strength Pareto Approach, IEEE Transactions on Evolutionary Computation, vol. 3, no. 4, pp. 257-271, 1999.