Optimizing the Photovoltaic Solar Energy Capture on Sunny and Cloudy Days Using a Solar Tracking System Nelson A. Kelly and Thomas L. Gibson Chemical Sciences and Material Systems Laboratory General Motors Research & Development Center Warren, MI 48090 Annual Technical Meeting of the Council for Optical Radiation Measurements (CORM) Las Vegas, Nevada May 9-11, 2010 1
Statement of the Problem Future transportation systems need to: 1) become sustainable, 2) be removed from the air pollution and global climate change debates, and 3) reduce their dependence on petroleum. GM envisions future transportation with electric drive systems, such as fuel cell electric vehicles (FCEV) and battery powered extended-range electric vehicles (EREV). If solar energy is used to make the hydrogen, i.e. solar-powered water electrolysis, or to charge the batteries, then the vehicle uses no fossil fuels and emits no pollutants. Solar energy is also one answer to the terawatt challenge for future energy needs. We have addressed the optimization of the solar hydrogen and solar battery charging initiatives. In the present work we address the portion of that process involving optimizing the solar energy capture by a photovoltaic (PV) system (see Solar Energy, 83, 2009, pp. 2092-2102 for background information) Generally, solar tracking systems are optimized for sunny conditions when the solar disk is visible. A question of interest to those using solar energy on a day-to-day basis, is: can we improve the solar PV energy capture on cloudy days? 2
Solar Tracking Systems 2-axis tracking -- maximize the solar energy capture on sunny days by keeping the solar module pointed at the solar disk so the solar module is perpendicular p to the beam normal solar radiation, I bn I θ = I bn x cosine (Θ) = I bn if θ = 0 (Θ is the angle between the direct solar ray and a normal to the module surface) highest h solar energy to DC electrical l energy What happens on cloudy days? open-loop loop tracker still points at the obscured sun using very exact algorithms to know position of the sun in the sky Astronomically-determined location of the solar disk, i.e., http://www.usno.navy.mil/astronomy y Random nature of clouds no general model for solar positioning on cloudy days - empirical models are available 3
Four Solar Arrays and Electrolyzer GM Solar Hydrogen Fueling Station 4
Solar Measurements Four solar arrays containing solar modules wired in parallel and including a solar sensor mounted the plane of the array PV system (Sanyo 190-BA3, 190 watt solar modules) Solar sensors (LI-COR pyranometers, LI-200SL) Array tilt angle can be manually adjusted All four with same tilt to see variability in measurements Four different tilt angles (57, 42, 27, 0 ) can be used to simulate 2-axis tracking Measure solar irradiance over a one-year period Analyze measurements to improve the solar energy capture and smooth the day-to-day differences in solar energy capture Minimize system size and cost Minimize storage (batteries, hydrogen) Reduce day-to-day variability in solar energy due to clouds 5
Data Analysis Measure solar irradiance over a range of ambient conditions with four solar sensors (pyranometers) and four solar arrays (PV modules) with different four different tilt angles One array and sensor was always horizontal, H Near solar noon (sun directly south, β=180 ), one of the arrays was nearly perpendicular to any direct solar irradiation (it has a tilt angle, Θ, such that the cosine error was negligible); this array was thus pointed directly toward the sun, DTS We will analyze the data to determine the optimum solar collector orientation for sunny and cloudy conditions over the study period The variables we will utilize are the solar irradiances (W/m 2 ) for the LI-COR sensors and short-circuit current, I sc, for the Sanyo modules for the H and DTS configurations, and including the H/DTS ratio 6
Details on the LI-COR Pyranometer Sensor Placement (In-Plane Solar Irradiance Measurement) 7
Components of the Solar radiation on at a Horizontal Collector at the Earth s Surface G h = I bh + I dh I bh = I bn xcosine(θ) where I bn is the beam normal radiation where Θ is the solar zenith angle 8
Solar Energy Models for Beam and Diffuse Radiation Beam (direct) radiation On a sunny day, up to 90% of the total solar energy is beam radiation, so 2-axis tracking maximizes the collection of beam radiation by keeping the cosine term near 1, i.e., I θ = I bn x cosine (Θ) = I bn if θ = 0 Diffuse (sky) radiation On a cloudy day, nearly 100% of the solar energy is diffuse (sky) radiation (I bh = 0); Isotropic Diffuse Model (Liu-Jordan Model) I θ = I dh x (1 + cosine(θ))/2 where θ is the tilt from the horizontal I θ should be maximized for Θ=0 (0 tilt, horizontal surface) 9
Specifying the Position of the Solar Disk Requires Two Angles, α and β Some people use the solar zenith angle, Θ Θ =90 - α 10
Four Solar Arrays Used for Testing Arrays with four different tilt angles (57, 42, 27, 0 ) MAJOR POINT:ATSOLAR NOON, ONE ARRAY WAS ALWAYS VERY CLOSE TO A DTS CONDITION AND ONE ARRAY WAS ALWAYS H Same tilt on all four arrays (and LI-COR sensors) 11
Solar Irradiance with all Four Arrays Having Identical Tilt Angles 1200 m 2 Solar Irr radiance, W/m 1000 800 600 400 200 0 6 8 10 12 14 16 18 20 Time of day Sunny day --- good agreement Cloudy day good agreement Sola ar irradiance, W/m 2 450 400 350 300 250 200 150 100 50 0 6 8 10 12 14 16 18 20 Time of day 12
Solar Irradiance with all Four Arrays Having Different Tilt Angles on a Sunny Day 1200 Solar irradianc ce, W/m 2 1000 800 600 400 200 Licor #1 Licor #2 Licor #3 Licor #4 0 6 8 10 12 14 16 18 20 Time of day LI-COR Tilt angle, Insolation (kwh/m 2 ) 1 57 6.84 2 42 6.42 3 27 5.53 4 0 3.41 13
Solar Irradiance with all Four Arrays Having Different Tilt Angles on a Cloudy Day 350 Solar irradia ance, W/m 2 300 250 200 150 100 50 Licor #1 Licor #2 Licor #3 Licor #4 0 6 8 10 12 14 16 18 20 Time of day LI-COR Tilt angle, Insolation (kwh/m 2 ) 1 57 1.00 2 42 1.08 3 27 1.20 4 0 1.30 14
Analysis of the H/DTS variable 181 days in the data base; 164 were used for the LI-COR and 161 for the solar arrays (Sanyo modules) One of the four LI-CORs and arrays was always H For a time period of ½ hour around solar noon (β = 180 ), for the south-facing arrays, one of the tilted arrays (57, 42, 27 ) was within 10 of DTS The cosine error was only 1.5% for the DTS measurement The solar irradiance was integrated for the ½ hour around solar noon for the H and DTS measurements to compute the solar insolation (kwh/m 2 ) during this period 15
H/DTS Ratio as a Function of the Solar Insolation 1.4 H/D DTS ratio 1.2 1.0 0.8 0.6 y = 0.771 * x -0.146 April May June July August September October November Power fit 0.4 0 0.1 0.2 0.3 0.4 0.5 0.6 Insolation, kwh/m 2, latitude tilt array 16
H/DTS Ratio vs. Tilt Angle on Sunny Days 1.0 0.8 H/ /DTS ratio 0.6 0.4 LI-COR Sanyo cosine(θ) 0.2 0.0 0 10 20 30 40 50 60 70 80 90 Angle θ, degrees 17
Isotropic Diffuse Model (IDM) Describing the Dependence of H/DTS on Array Tilt On Cloudy Days 20 2.0 H/DTS cloudy day ys 1.8 1.6 1.4 1.2 LI-COR Sanyo LI-COR, avg. Sanyo, avg. IDM 1.0 0 10 20 30 40 50 60 70 80 90 DTS array tilt angle, degrees 18
Summary and Conclusions Four identical solar arrays and sensors were oriented at four different angles for a period of approximately 8 months at the GM Milford, MI Proving Ground in 2008 to study the effect of the tilt angle on the amount of solar energy captured. On sunny days, with predominantly direct solar radiation, the H/DTS ratio approximately obeyed the well-known cosine response law. On sunny days, H/DTS approached 0.5 at low solar altitude angles, i.e., tracking the sun would yield twice as much solar energy as a fixed horizontal tilt. On cloudy days or during cloudy periods, our analysis shows that 2- axis tracking will reduce the solar energy capture versus a horizontally tilted sensor (or array). We observed that the H/DTS ratio reached values of up to 1.37 for the cloudiest days. 19
Summary and Conclusions A simple model for diffuse radiation, referred to as the Isotropic Diffuse Model, agreed well with the average angular dependence that we measured for the H/DTS ratio on cloudy days. An optimized solar energy system would utilize 2-axis solar tracking during sunny conditions to capture the direct irradiance, but would orient the modules toward the zenith for cloudy conditions. In the past, the emphasis has been on optimizing the capture of direct solar radiation on sunny days because it is by far the largest overall component of the solar irradiance. Maximizing the capture of diffuse solar radiation on cloudy days is important in order to minimize the system size and level out the dayto-day fluctuations in system output. The dramatic improvement on cloudy days is especially useful for a home solar-powered FCEV hydrogen fueling system or EREV battery charging g system that needs to the energy for daily commuting on a continuous basis throughout the year. 20