PERFORMANCE EVALUATION OF WIND ENERGY CONVERSION SYSTEMS

Abstract

Wind energy is making significant penetration into electric grids in various places around the world and the need for accurate analysis is increasing in importance. The energy generation by wind energy conversion systems is without fuel cost and nature dependent. Wind energy conversion systems (WECS) produce a variable power output unlike conventional power plants such as nuclear, thermal, hydro, etc. The magnitude of energy generated by wind turbines is related to different factors such as wind speed, wind direction, air density, seasons of a year, time of a day etc. Since these input variables keep on changing over time, the output of wind turbine also changes. Also, other technical constraints such as grid unavailability, turbine breakdown cause limitations to wind energy production. The power output from the wind turbine is also governed by the design features of wind turbine such as, cut-in speed - the minimum wind speed at which the wind turbine starts to deliver useful power, rated wind speed - the wind speed at which the rated power of the machine is reached, and cut-out speed - the maximum wind speed upto which the machine is allowed to deliver power and above this wind speed the wind turbine is shut downto prevent any untoward damage. Generally, the cut-in speed for the wind turbine is in the range of 2.5 to 3.5m/s and the cut-out speed ranges from 20 to 25 m/s. The rated power output is produced from wind turbines when the wind speed is above rated wind speed of the machine (approximately 14m/s). This phenomenon makes the total power output from WECS to vary hourly, daily and monthly. The wind energy technology has established a sound technical feasibility in spite of its intermittent and variable nature. The analysis of its performance is important in harnessing and utilizing the energy generation more effectively. This work focuses on the performance evaluation of wind energy conversion systems in relation to certain major performance aspects. To analyse performance actual field data are used. Field data are collected from seven wind farms with a total capacity of 37.225 MW in Muppandal, Tamilnadu (India) for a period of three years from April 2002 to March 2005. The Muppandal region in Tamilnadu has the largest concentration of wind turbines with the installed capacity of more than 1500 MW. The annual average wind speed in this region is 6 - 7 m/s. xiii The thesis studies the following aspects: 1. Assessment of energy yield from wind farms. 2. Assessment of reliability of wind power generation. 3. Assessment of economic of WECS. 4. Assessment of growth potential of WECS in India. As the output of a wind power plant fluctuates, it is necessary to assess the available energy yield of a wind farm for a given period in order to successfully operate wind energy with traditional power generation. Accurate estimations of wind energy output are also critical for economic viability, system reliability, scheduling and planning. Modeling techniques combine meteorological and historical generation data and in order to achieve the maximum possible accuracy, the methods should incorporate appropriate parameter and data. A model is developed using artificial neural networks (ANN) for the estimation of energy yield from WECS. This is useful in finding out the amount of energy that could be generated by wind farms. The input variables chosen for the ANNmodel are the monthly average wind speed, monthly average relative humidity and monthly generation hours. The output variable is the energy output from the wind farms. The most appropriate neural network configuration after successive trials is found to have a configuration 3-4-1 (3 neurons in input layer, 4 neurons in hidden layer and 1 neuron in output layer). The mean square enor for the training and testing values is obtained as 7.0x10" and 6.5 x 10"3. The model accuracy is evaluated by comparing the simulated results with the actual measured values of the wind farms and is found to be in good agreement. An estimation of the WECS reliability is necessary from the planning perspectives of the energy planners. The level of penetration of wind energy in the grid will have a definite impact on the overall reliability of the power system. Several studies have been reported on the reliability of the power system incorporating wind energy. These include methods such as probabilistic methods, chronological simulations etc. The contribution of a wind power plant to the reliability performance of a power system mainly depends on wind energy penetration level and wind conditions. Since the wind energy conversion systems are free of fuel cost, the energy generated by WECS must be fully consumed in the grid to reduce the overall generation xiv cost. On an average, a wind turbine installed at a potential site will generate electric powerfor 70-85% of the time, but not always at the rated power output. The poweroutput characteristics of wind energy conversion system are different from that of any conventional generation systems. A conventional power plant can be considered as a binary operated unit, either fully available or not at all but the wind power plant output vary from zero to its rated capacity. When the power generation through wind is inadequate or not available, the conventional sources must meet the deficit in load demand during those quiescent periods. In spite of the intermittent and variable nature of wind, the wind energy conversion systems contribute to a certain extent to meet the energy demand requirements of power system. The analysis on reliability aspects of wind power finds more significance in comparison with the conventional power generation systems. The reliability assessment of wind energy conversion systems is important in system planning and to determine appropriate generating resources to meet the expected total demand. The reliability of wind power generation is evaluated using Monte Carlo technique. Two indices are used, one on the basis of energy delivery factor and the other on the basis of load demand. Wind energy conversion systems mostly do not run at the rated capacity, therefore, the capacity value of WECS is generally taken as the energy delivery factor times the actual rated capacity. The reliability evaluation is done on the basis of an index, defined as the period during which expected wind energy is not supplied (EWNS). The evaluation is carried out with the daily and monthly energy generation data. The effect of increasing the hub height of wind turbines on the reliability index is additionally investigated. It is found that the increase of hub height improves the reliability. The reliability analysis is also done for WECS on the basis of another index LOLE, which is the loss of load expectation. LOLE is the expected period during which the load demand exceeds the wind power generation. It is expressed as number of hours per year. The analysis is carried out to find the extent of reliable contribution of wind energy conversion system in meeting the load demand. The chronological hourly average daily generations for each month of the wind farms are compared with a chronological hourly load demand model constructed for the purpose. The effect of increasing the hub height by 10 m is also investigated. xv The economic value of a wind energy conversion system mainly depends on the prevailing condition of the wind. There are other factors which influence the cost of wind energy, such as site specific factors, machine design parameters, government policies, etc. It is important to estimate the cost of generation and other economic benefits expected from a wind power project. A sensitivity analysis is canied out to determine the impact of six major variables on the cost of energy and the internal rate of return (IRR). The variables are: annual average wind speed, energy delivery factor, hub height, generation hours, capital cost and operation and maintenance cost. The sensitivity of each of the input variables is studied over a range of values around the base value. The analysis shows that the annual average wind speed is the most critical variable in influencing the cost of generated energy and the internal rate of return. In India, wind power generation has gained a high level of attention and acceptability and now ranks fourth in the world with an installed capacity of 6270 MWas on December 2006. Technological forecasting is used to estimate the growth and the direction of any technology. The technological forecasting is canied out using the Pearl (Logistic) curve to predict the overall future trends of Indian wind power industry and also in five potential Indian states, namely, Tamilnadu, Maharastra, Karnataka, Gujarat and Andhra Pradesh. These five states have high technical wind energy potential of more than 1000 MW. The installed wind power capacity data for a period of fifteen years from 1992 to 2006 is taken for forecasting the future trends of India and the five potential Indian states. It is found that 99% of India's technical wind energy potential may be achieved by the year 2030. The wind power technology in Tamilnadu has already gained importance and shows a good progress in the development. Maharastra and Karnataka show a relatively steeper rise compared with other states.