Please use this identifier to cite or link to this item:
Title: Study on wind turbine wake characteristics and layout optimization with Hong Kong offshore wind power potential assessment
Authors: Gao, Xiaoxia
Advisors: Yang, Hongxing (BSE)
Lin, Lu (BSE)
Keywords: Wind power -- China -- Hong Kong.
Wind turbines -- Design and construction.
Issue Date: 2015
Publisher: The Hong Kong Polytechnic University
Abstract: It is reported that wind energy plays a significant role in the global renewable energy power markets with an increased installation capacity. However,most operating wind farm cannot generate as much power as predicted because of the wake effect. The inaccurate wake model and unsuitable turbine layout can enlarge this difference. Therefore, this research project aims to investigate the turbine wake model analytically and experimentally and then,to propose a program to conduct the turbine layout optimization in a large wind farm.Based on these, the offshore wind power potential of the Hong Kong offshore areas is assessed and the optimal layout pattern of wind turbines under Hong Kong offshore wind power exploitation conditions is recommended, which is very useful for Hong Kong offshore wind energy utilization. Firstly, the thesis develops a new model based on the Jenson's wake model, which regards the velocity deficit in a wake at a specific downwind distance as a linear one, with no consideration of the wind velocity profile on the radial direction. The new proposed wake model overcomes the shortcomings of Jenson's wake model and develops it to a 2D model using Gaussian function. The results from this model can well express the turbine wake characteristics such as the velocity deficit, wake width and centerline.Based on this model, the velocity deficit profiles at both downwind and radial directions can be calculated. The other improvement from this new model is that, when determining the turbine layout in a wind farm, the local velocity field of a turbine in the wake caused by upwind one is determined and calculated by the relative distance of the two turbines at two directions (downwind and radial) instead of downwind distance only.Moreover, a multiple wake model is developed to describe the wake profile generated by numbers of turbine in a row. The multiple wake model is the superposition of the Gaussian functions and in which, the sub-wake characteristics of each turbines can also be expressed with parameters in the formula.In order to validate the accuracy of the new model, the measured wind data from a 300 MW operating wind farm, i.e.the CWEX-13 located in Iowa of the United States, are analyzed.The undisturbed wind,wind after one row of turbine and wake after two rows were measured. The wind conditions at different downwind distance of four-turbines-in-row were measured too. The wake characteristics, such as the wake width,velocity deficit under different atmospherics stabilities and turbulence intensities are studied. For the first time,the wakes are distinguished as the inner wake and outer wake, depending on the relative locations of the turbines. The inner and outer wake characteristic are quantified and compared with each other.Results show that, the velocity deficit,immediately behind the turbine,is initially about 70%-75% for the inner wake and, 55%-60% for outer wake.Notably, significant velocity deficit is still apparent even as far as 10D behind the turbines, about 30%-35%. The same thing happens for the wake width. At the very beginning when the flow pass through the turbines, the wake expansion is 1.7D for outer turbines while 1.6D for the inner one. Noted that D represents the turbine rotor diameter. Another important conclusion is that, the outer wakes vanish faster than the inner wake (The inner deficit is 1.164 multiples of the outer deficit while the inner wake width is 0.87 multiples of that of the outer one). In this thesis,it is concluded that the ambient stability does not have significantly influence on the velocity deficit and wake width for these four turbines.
The second main contribution from this thesis is a proposed Multi-Population Genetic Algorithm (MPGA) simulation model for wind turbine layout optimization which is a long-history research hotspot started from 1983. The optimal turbine layout pattern can increase the power generation of the wind farm, reduce the wake interaction between wind turbines,which otherwise would increase dynamic mechanical load and cause higher fatigue load.Many researchers have been working on this topic with different methods, and the proposed MPGA program has been validated by solving the same wind farm optimization issues. More power generation with a lower cost of energy (COE) demonstrates the advancement of this model.For this optimization program, the focus in this thesis goes further. The new proposed wake model is used in this program with a comprehensive cost model, which considers the local wind farm development conditions including labour cost. By using this optimization simulation model, together with wind farm size, wind data and turbine characteristics as inputs,the simulation model can generate the optimal layout for the turbines with total power generations, COE and wind farm efficiency and so on. It is reported that, for the Case of 'constant wind speed of 12m/s with variable wind direction',using the newly-developed Jenson-Gaussian wake model in the MPGA optimization program make the power generation and wind farm efficiency decreased than that of the Jenson's wake model. The power loss caused by wake effect is about 20%, which is in accordance with previous literatures. Three layout patterns can be chosen before the program is started, i.e. aligned, staggered and scatter ones. Besides, the offshore wind developing conditions in Hong Kong are studied. The total water area suitable for offshore wind farm development is determined after considering the local water conditions, wind condition and water area usage purposes. The potential offshore wind farm area in and beyond Hong Kong's boundary (2 km) is about 357.78 km2 (21.68% of the HK's water area). Finally, four typical offshore wind farm sites located at different water areas in Hong Kong are selected. Using the MPGA optimization program, the top ten optimal layout patterns for each potential site are proposed. The Hong Kong offshore wind power potential is reported with the COE and wind farm efficiency. It is estimated that the optimal wind turbine layout separation is 14.5D in prevailing wind direction and 11.0D in cross wind direction (D represent the turbine rotor diameter). The levelized cost of energy (LCOE) is calculated in HK$ terms, i.e. 1.474/kWh (aligned), 1.467/kWh (staggered), nd 1.290/kWh (scattered). APG (annual energy generation) is determined to be 40.80×10x/y8 kWh (aligned), 40.42 ×10x/y8 kWh (staggered), and 33.98 ×10x/y8 kWh (scattered), representing 9.48% (aligned), 9.39% (staggered), and 7.89% (scattered) of the annual electricity consumption for HK in 2012.The results can provide guidance for the government or private developers to develop the offshore wind farms in Hong Kong. In summary,this research project developed a new analytical wake model and proved its practicability as the basic velocity deficits calculation models. The wake characteristics based on the new model are estimated. The newly-developed MPGA optimization program has a good performance on wind turbine layout optimization within wind farm and also, its availability in solving the real-world offshore wind farm turbine micro-siting has been validated. The optimization for Hong Kong offshore wind farm with the offshore wind energy assessment process can provide a new thought and filled the research gaps for the wind farm development and wind energy assessment in this research area.
Description: PolyU Library Call No.: [THS] LG51 .H577P BSE 2015 Gao
xxix, 193 pages :color illustrations
Rights: All rights reserved.
Appears in Collections:Thesis

Files in This Item:
File Description SizeFormat 
b28281548_link.htmFor PolyU Users208 BHTMLView/Open
b28281548_ira.pdfFor All Users (Non-printable)7.35 MBAdobe PDFView/Open
Show full item record

Page view(s)

Last Week
Last month
Checked on Jan 15, 2017


Checked on Jan 15, 2017

Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.