
Renewable Energy
Track the sun over Chennai, shade a real 540 W module, put a 2.1 MW turbine on a Weibull wind and size a family biogas plant
26 chapters, 26 interactives, few words. This book covers the renewable energy course of engineering degrees from the sunlight up: India's energy mix and capacity factors, solar radiation and sun-earth angles, tilted surfaces, flat-plate and concentrating collectors, solar water heaters and stills, photovoltaic physics, the solar cell model, shading and bypass diodes, off-grid and rooftop systems, the wind resource, rotor aerodynamics, power curves and annual energy, wind farms and wakes, biomass, biogas, gasification, small hydro, ocean and geothermal energy, green hydrogen, grid integration and the levelised cost of energy. The same sites (Chennai, Jodhpur, Pune and Bengaluru, with NASA's 25-year monthly sunshine data), the same 540 W module, the same 2.1 MW turbine and the same Chennai flat run through the book, so every chapter's numbers describe one world.
- AUCME365Renewable Energy Technologies
- AUEE3014Power Electronics for Renewable Energy Systems
- VTUBME654BRenewable Energy Power Plants
- VTUBEE654BTechnologies of Renewable Energy Sources
- SPPU402051BRenewable Energy Technologies
- SPPU203141Power Generation Technologies
- MUMEDLO7032Renewable Energy Sources
- MUEEDO5011Renewable Energy Sources
- KTUMET445Renewable Energy Engineering
- KTUPEEET412Renewable Energy Sources
By Skill-Lync · Updated 21 Nov 2026 · 26 chapters · 26 interactives · free, no sign-up
India's energy mix: installed capacity, generation and capacity factor
Which syllabus covers thisAU CME365 Unit 1; EE3014 Unit 1VTU BME654B Module 1; BEE654B Module 1SPPU 402051B Unit 1MU MEDLO7032 Module 1; EEDO5011 Module 1KTU MET445 Module 1; PEEET412 Module 1
What is the difference between a gigawatt and a billion units?
A gigawatt (GW) is power: how fast energy flows at one moment. A billion units (BU) is energy: one billion kWh, the 'units' on your bill. One GW running flat out for a year gives 1 × 8,760 h = 8.76 BU. India's plants made 1,845.9 BU in FY 2025-26. Capacity tells you what the grid could deliver at peak; energy tells you what it actually delivered. Mixing the two is the most common mistake in news about renewables.
Why does a gigawatt of solar make far less energy than a gigawatt of coal?
The capacity factor (CUF) is the energy a plant made divided by what it would make at full power all year. On FY 2025-26 data, India's solar fleet ran at 15.5 %, wind 22.8 %, large hydro 38.5 %, coal 65.7 % and nuclear 74.7 %. The sun sets, clouds pass and the noon peak lasts a few hours, so 40 GW of new solar gives 54.2 BU a year, the same as 9.4 GW of coal. Drag the slider: the mint fills (average output) always match.
What does India's electricity come from today?
On 31 Oct 2026 India had 563.1 GW: coal 225.1 GW (40 %), solar 174.5 GW (31 %), wind 59.5, large hydro 52.3 and nuclear 8.78 GW. Conventional sources (coal, gas, large hydro, nuclear) are mature and dispatchable; non-conventional ones (solar, wind, biomass, small hydro) are renewable but variable. Nuclear is the low-carbon alternative that runs day and night (75 % CUF) but is slow and costly to build. Coal still made 67.7 % of the energy.
How far is India from its 500 GW non-fossil target?
The goal (COP26, PIB) is 500 GW of non-fossil capacity by 2030. Non-fossil capacity was 283.47 GW on 31 Mar 2026 and 312.2 GW on 31 Oct 2026, leaving 187.8 GW. Spread over the 50 months to the end of 2030 that is about 45 GW a year; FY 2025-26 added 55.3 GW. Capacity is on track. Energy is the harder target: non-fossil sources made 29.2 % of the electricity in FY 2025-26.
What does a unit of electricity cost the environment?
CEA's CO2 database puts India's grid at 0.710 kg CO2 per kWh in FY 2024-25, and coal stations alone at 0.969 kg/kWh. The book's Chennai flat uses 16.45 kWh a day, about 4.3 tonnes of CO2 a year on the grid average. Coal also brings fly ash, sulphur and nitrogen oxides, particulates and heavy water use. Renewables are not free of impact: land for solar parks, bird strikes and noise near wind farms, submerged forests behind dams, and mining for silicon, copper and rare earths.
On 31 Oct 2026 India had 175 GW of solar and 225 GW of coal. Which made more electricity in FY 2025-26?
Coal made 1250 BU against solar's 174 BU, 7.2 times more. Coal ran at 65.7 % CUF and solar at 15.5 %, and much of today's solar was only built during the year.
Worked examples
Capacity factor of India's wind fleet in FY 2025-26 · Wind made 106.089 BU. Installed wind: 50,037.82 MW on 31 Mar 2025 and 56,094.84 MW on 31 Mar 2026 (CEA)
- Average capacity over the year ≈ (50,037.82 + 56,094.84)/2 = 53,066 MW = 53.07 GW
- Energy at full power all year = 53.07 GW × 8,760 h = 464.9 BU
- CUF = 106.1/464.9 = 22.8 %
- Using only the year-end capacity would give 21.6 %: too low, because new turbines did not run all year
Answer: CUF ≈ 22.8 %
How much coal capacity does 10 GW of new solar replace? · Use FY 2025-26 fleet CUFs: solar 15.48 %, coal 65.72 %. CO2: coal stations 0.969 t/MWh, grid average 0.710 t/MWh (CEA)
- Solar energy = 10 GW × 0.1548 × 8.76 = 13.56 BU a year
- Coal capacity for the same energy = 13.56/(0.6572 × 8.76) = 2.36 GW
- If it displaces coal: 13.56 BU × 0.969 = 13.1 Mt CO2 a year avoided
- Counted at the grid average instead: 13.56 × 0.710 = 9.6 Mt
Answer: 10 GW of solar matches the energy of about 2.4 GW of coal and avoids 9.6-13.1 Mt CO2 a year. It does not replace coal's evening output: that needs storage
Filling the 500 GW gap with solar alone · Non-fossil capacity on 31 Oct 2026: large hydro 52,314.66 MW, nuclear 8,780 MW, RES 251,148.81 MW (CEA). Target 500 GW
- Non-fossil = 52.31 + 8.78 + 251.15 = 312.24 GW; gap = 500 − 312.24 = 187.76 GW
- Rate needed over 50 months to Dec 2030 = 187.76 × 12/50 = 45.1 GW a year
- If all of it were solar at 15.5 %: 187.76 × 0.1548 × 8.76 = 254.6 BU a year
- That is 13.8 % of FY 2025-26's 1,845.9 BU, or the output of 44.2 GW of coal
Answer: 188 GW to go, about 45 GW a year; as solar it adds about 255 BU, 14 % of today's generation
Solar radiation: the solar constant, air mass, beam and diffuse light
Which syllabus covers thisAU CME365 Unit 2VTU BME654B Module 1; BEE654B Module 1SPPU 402051B Unit 1; 203141 Unit 5MU MEDLO7032 Module 1KTU MET445 Module 1; PEEET412 Module 2
How much sunlight arrives at the top of the atmosphere?
The solar constant is the sunlight on a surface facing the sun just outside the atmosphere at the mean Earth-sun distance: 1,361 W/m² from satellite measurements (NASA's TSIS-1 read 1,361.6 ± 0.3 W/m² in 2019). Older textbooks and FAO-56 use 1,366-1,367. The orbit is slightly elliptical, so the value on a given day is G0n = 1,361 (1 + 0.033 cos(360n/365)): 1,406 W/m² in early January and 1,316 in early July, a ±3.3 % swing. India's winter gets the stronger sun from space.
What is sunlight made of?
The sun radiates roughly like a black body at 5,778 K, so its spectrum peaks near 502 nm (Wien: 2,898 µm K / 5,778 K). Above the atmosphere about 8 % of the energy is ultraviolet. At the ground ozone has removed most UV and water vapour and CO2 have cut bands out of the infrared: roughly 3-5 % UV (below 400 nm), 42-43 % visible (400-700 nm) and 52-55 % infrared. The Heat and Mass Transfer book has a Planck-curve interactive; the PV chapters here use the standard AM1.5 spectrum.
What is air mass, and why does AM1.5 appear on every solar module?
Air mass is the length of the sun's path through the atmosphere relative to the path straight down: AM = 1/cos θz, where θz is the zenith angle. AM0 is space, AM1 the sun overhead, AM1.5 a zenith angle of 48.19°, the standard for testing PV modules. Near the horizon the flat-Earth formula fails: at θz = 85° it gives 11.47 while Kasten and Young's curved-atmosphere fit gives 10.31, and about 38 at the horizon. More air means more scattering and absorption.
How much of the beam survives a clear sky, and where does diffuse light come from?
Beam (direct) light arrives in a straight line from the sun's disc; diffuse light is sunlight scattered by air, dust and cloud and arrives from the whole sky; global = beam on the horizontal + diffuse. Hottel's 1976 clear-sky model gives the beam transmittance τb from the zenith angle and site altitude (tropical constants), and Liu-Jordan gives diffuse τd = 0.271 − 0.294 τb. Chennai, 21 June, 9:00 solar time: AM 1.38, beam 732 W/m² normal to the sun, diffuse 102 W/m².
How do the monsoons show up in solar data?
The clearness index KT = H/H0 compares the day's sunshine on the ground with what arrived above the atmosphere. NASA's 2001-2025 data give Chennai KT 0.63 in March but 0.46 in July, and Chennai's Oct-Dec average falls to 4.35 kWh/m²/day under the north-east monsoon. Pune in July has KT 0.31 and 67 % of its light diffuse. Jodhpur's dip is July (5.76 kWh/m²/day), but its winter KT of 0.65-0.69 makes it India's desert solar belt.
On which date does the top of the atmosphere over India receive the strongest sunlight (W/m² facing the sun)?
The Earth is closest to the sun in early January, so G0n peaks at 1,406 W/m² against 1,316 in early July. The June day is longer and the sun higher, which matters more on the ground.
Worked examples
Air mass and clear-sky sunlight at Chennai on 21 June, 9:00 solar time · Chennai φ = 13.08° N, altitude 0.05 km, n = 172, solar time 9:00. Hottel tropical constants, Liu-Jordan diffuse
- δ = 23.45 sin(360 × (284 + 172)/365) = 23.45°; ω = 15 × (9 − 12) = −45°
- cos θz = cos 13.08 cos 23.45 cos 45 + sin 13.08 sin 23.45 = 0.7219, θz = 43.8°; AM = 1/cos θz = 1.385
- a0* = 0.4237 − 0.00821 (6 − 0.05)² = 0.133, a1* = 0.753, k* = 0.3826; tropical: a0 = 0.1264, a1 = 0.738, k = 0.3903
- τb = 0.1264 + 0.738 e^(−0.3903/0.7219) = 0.5562; G0n = 1,361 (1 + 0.033 cos(360 × 172/365)) = 1,316.8 W/m²
- Gbn = 1,316.8 × 0.5562 = 732.4 W/m²; on the horizontal Gb = 528.7 W/m²
- τd = 0.271 − 0.294 × 0.5562 = 0.1075; Gd = 1,316.8 × 0.1075 × 0.7219 = 102.2 W/m²
Answer: AM = 1/cos θz ≈ 1.39 (Kasten-Young 1.38); clear-sky global on the horizontal ≈ 631 W/m² (529 beam + 102 diffuse)
Does altitude help? Clear-sky noon beam at Bengaluru · Bengaluru φ = 12.97° N at 0.842 km (NASA cell), 15 April (n = 105), solar noon
- δ = 23.45 sin(360 × 389/365) = 9.41°; at noon θz = φ − δ = 3.56°, cos θz = 0.9981
- a0* = 0.4237 − 0.00821 (6 − 0.842)² = 0.2053; a1* = 0.5055 + 0.00595 (6.5 − 0.842)² = 0.696; k* = 0.2711 + 0.01858 (2.5 − 0.842)² = 0.3222
- Tropical: a0 = 0.195, a1 = 0.6821, k = 0.3286; τb = 0.6857 (at sea level it would be 0.6211)
- G0n(105) = 1,350.5 W/m²; Gbn = 926 W/m², Gb = Gbn cos θz = 924.3 W/m²
- τd = 0.271 − 0.294 × 0.6857 = 0.0694; Gd = 1,350.5 × 0.0694 × 0.9981 = 93.5 W/m²
Answer: Clear noon: beam 926 W/m² normal, 1,018 W/m² global on the horizontal; the 842 m of altitude adds about 10 % to the beam transmittance
Clearness index and diffuse fraction: Jodhpur in May and July · Jodhpur φ = 26.29° N. NASA monthly means: May H = 7.34, Hd = 2.64; July H = 5.76, Hd = 2.9 kWh/m²/day. Average days n = 135 and 198
- H0 = (24 × 3,600 × 1,361/π)(1 + 0.033 cos(360n/365))(cos φ cos δ sin ωs + (π ωs/180) sin φ sin δ)
- May: δ = 18.79°, ωs = 99.68°, H0 = 11.03 kWh/m²/day; KT = 7.34/11.03 = 0.665
- July: δ = 21.18°, ωs = 101.04°, H0 = 11.1; KT = 5.76/11.1 = 0.519
- Diffuse fraction Hd/H: May 0.36, July 0.503
Answer: KT falls from 0.67 to 0.52 and the diffuse share rises from 36 % to 50 %: the south-west monsoon reaches even the desert
Sun-earth angles: declination, hour angle, altitude, azimuth and day length
Which syllabus covers thisAU not in syllabusVTU BEE654B Module 1SPPU 402051B Unit 1MU MEDLO7032 Module 1KTU MET445 Module 1
What is the declination and how do you find it for any day?
Declination δ is the latitude where the noon sun is exactly overhead. The Earth's axis is tilted 23.45°, so δ swings from +23.45° on 21 June to −23.45° on 21 December and passes 0° at the equinoxes. Cooper's equation gives it from the day number n (1 January = 1): on 15 May, n = 135 and δ = 18.79°. Spencer's Fourier series is more accurate (within 0.035°); Cooper's is the one exam papers expect.
Why is solar noon in Chennai not at 12:00 IST?
Indian Standard Time is the mean solar time of 82.5° E. The sun takes 4 minutes per degree, so Chennai (80.27° E) is 8.9 min behind, Jodhpur (73.02° E) 37.9 min. The equation of time E adds the effect of the elliptical orbit and the tilt: from −14.3 min (mid-February) to +16.4 min (early November). Solar time = IST + 4(L − 82.5) + E minutes; hour angle ω = 15° × (solar time − 12). On 21 June Chennai's solar noon is 12:10 IST.
Where is the sun in the sky at a given moment?
The zenith angle θz follows from cos θz = cos φ cos δ cos ω + sin φ sin δ; altitude = 90° − θz; azimuth is measured from south, west positive. At noon the altitude is 90° − |φ − δ|. In Chennai (φ = 13.08°) on 21 June δ = 23.45° is larger than φ, so the noon sun stands 79.6° high in the NORTH, and a south-facing panel is lit from behind at noon. On 21 December it is 53.5° high in the south.
How long is the day, and why do almanacs disagree with the formula?
The sun sets when θz = 90°: cos ωs = −tan φ tan δ, and day length N = 2ωs/15 hours. Chennai on 21 June: N = 12.77 h; Jodhpur 13.65 h in June but 10.35 h in December. Published times are longer: they count the sun's upper edge and the atmosphere bends light over the horizon, together 0.833°. With that, Chennai's 21 June sunrise and sunset come out at 05:43 and 18:37 IST, matching timeanddate.com to the minute.
How do the angles decide shading and the light on a panel?
For a south-facing surface tilted β, cos θ = cos(φ − β) cos δ cos ω + sin(φ − β) sin δ: the panel sees the sun as a flat surface would at latitude φ − β. Shadows set the spacing of panel rows. At Chennai's December noon a 1 m pole casts a 0.74 m shadow; at 9:00 solar time it grows to 1.55 m, 0.99 m of it pointing north, behind the row. Designers check the 9:00-15:00 window on 21 December.
At solar noon on 21 June in Chennai (13.1° N), where is the sun?
The declination (+23.45°) exceeds Chennai's latitude, so the sun is overhead at 23.45° N and Chennai sees it to the north: altitude 90 − |13.08 − 23.45| = 79.6°. It is overhead in Chennai twice a year, in late April and mid-August.
Worked examples
Declination, sunset hour angle and day length: Jodhpur on 15 May · Jodhpur φ = 26.29° N; 15 May is day n = 31 + 28 + 31 + 30 + 15 = 135
- δ = 23.45 sin(360 × (284 + 135)/365) = 23.45 sin(413.26°) = 18.79°
- cos ωs = −tan 26.29 × tan 18.79 = −0.1681, ωs = 99.68°
- N = 2 × 99.68/15 = 13.29 h
- Noon altitude = 90 − (26.29 − 18.79) = 82.5°, to the south
Answer: δ = 18.79°, ωs = 99.7°, day length 13.29 h (13 h 17 min)
Local apparent time: solar noon and sunrise at Chennai on 21 December · Chennai φ = 13.08° N, L = 80.27° E; n = 355; IST meridian 82.5° E
- B = (355 − 1) × 360/365 = 349.15°; Spencer: E = 2.17 min
- Solar time = IST + 4 × (80.27 − 82.5) + 2.17 = IST + (−6.75) min, so solar noon is at 12:00 + 6.75 min = 12:07 IST
- δ = −23.45°; cos ωs = −tan 13.08 tan(−23.45) = 0.1008, ωs = 84.22°
- Sunrise in solar time = 12 − 84.22/15 = 6.3856 h; in IST 06:30; sunset 17:44 IST
- With the 0.833° refraction and disc correction: sunrise 06:26, sunset 17:47 (timeanddate.com: 06:26 and 17:47)
Answer: Solar noon 12:07 IST; geometric sunrise 06:30 IST, seen at 06:26
Angle of incidence on a 20° roof in Pune at 10:00 IST on 15 March · Pune φ = 18.52° N, L = 73.86° E; surface facing south, β = 20°; n = 74
- E (Spencer, n = 74) = −9.64 min; solar time = 10:00 + 4 × (73.86 − 82.5) + (−9.64) min = 09:16
- ω = 15 × (9.2633 − 12) = −41.05°; δ = 23.45 sin(360 × 358/365) = −2.82°
- cos θz = cos 18.52 cos(−2.82) cos(−41.05) + sin 18.52 sin(−2.82) = 0.6986, θz = 45.69°
- cos θ = cos(−1.48) cos(−2.82) cos(−41.05) + sin(−1.48) sin(−2.82) = 0.7542, θ = 41.04°
- Beam tilt factor Rb = cos θ/cos θz = 1.08
Answer: θ = 41° (θz = 45.7°): the tilted roof gets 8 % more beam per m² than the flat ground at that moment
Radiation on a tilted surface, optimum tilt and how sunlight is measured
Which syllabus covers thisAU CME365 Unit 2 (measurement)VTU BME654B Module 2SPPU 402051B Unit 1; 203141 Unit 5MU MEDLO7032 Module 1KTU MET445 Module 1; PEEET412 Module 2
Why does tilting a collector towards the south catch more sunlight?
Beam light on a surface scales with cos θ, the angle to the surface's normal. The beam tilt factor Rb = cos θ/cos θz compares a tilted surface with the flat ground. For monthly design Klein's R̄b averages it over the month's average day. A low winter sun favours a steep tilt: Chennai's December R̄b at 29° is 1.36. Pune's best fixed tilt (21°) gives R̄b 1.36 in December but 0.85 in July, when the noon sun is almost overhead.
How do you estimate the sunlight on a tilted surface from horizontal data?
Liu and Jordan's isotropic model splits the day's horizontal sunlight H into beam (H − Hd), scaled by R̄b, and diffuse Hd, of which a tilted surface sees the fraction (1 + cos β)/2 of the sky; the ground reflects H ρg into the rest (ρg ≈ 0.2 for dry ground, assumed). Chennai's December on a 13° roof: 4.6 kWh/m²/day against 4.18 flat. The model treats the sky as evenly bright and ignores the bright ring around the sun, so it slightly underestimates tilted surfaces on clear days.
Is the best tilt really equal to the latitude?
Roughly, but compute it. With NASA 2001-2025 data the best fixed south tilt is 12° in Chennai (latitude 13.1°), 26° in Jodhpur (26.3°), 21° in Pune (18.5°) and 14° in Bengaluru (13.0°). Pune's wet, cloudy summer pushes its optimum up. The gain over flat is modest in the south: Chennai +1.4 %, Jodhpur +8 %. Adjusting twice a year (Apr-Sep 0°, Oct-Mar 29° in Chennai) lifts Chennai to 1,995 kWh/m² a year.
Which instruments measure sunlight?
A pyranometer measures global radiation: a black thermopile under glass domes sees the whole sky (180°) from about 300 to 2,800 nm; ISO 9060:2018 classes them. Shade its sensor from the sun with a ring or band and it reads diffuse. A pyrheliometer points a narrow tube at the sun on a tracker and reads the direct normal beam. The Campbell-Stokes recorder (1853, improved by Stokes in 1879) focuses the sun through a glass sphere and burns a trace on a card; since 2003 sunshine duration means beam above 120 W/m².
How do you get daily radiation from hours of sunshine?
The Angström-Prescott relation H/H0 = a + b n/N links the measured sunshine hours n (from a sunshine recorder) to the day length N and the extraterrestrial H0. FAO-56 recommends a = 0.25 and b = 0.50 where no local calibration exists, so a cloudless day (n = N) gets 75 % of H0. Indian stations have their own fitted constants; use them when you have them. On Jodhpur's 15 April, H0 = 10.33 kWh/m²/day and a 10.2 h card gives 6.75 kWh/m²/day.
A fixed, south-facing collector in Chennai (13.1° N): which tilt collects the most sunlight over a year?
The annual optimum on NASA data is 11.9°, close to the latitude, and the curve is flat near the top: 1,933 kWh/m² against 1,906 flat (+1.4 %). From May to August the noon sun is north of overhead, which keeps the best tilt low.
Worked examples
Monthly H_T on a 13° south-facing roof in Chennai, December · Chennai φ = 13.08° N; NASA December means H = 4.1784, Hd = 1.9884 kWh/m²/day; average day n = 344; β = 13°; ρg = 0.2 (assumed)
- δ = 23.45 sin(360 × 628/365) = −23.05°; ωs = arccos(−tan φ tan δ) = 84.33°
- ω′s = min(ωs, arccos(−tan(φ − β) tan δ)) = min(84.33, 89.97) = 84.33°
- R̄b = [cos(φ−β) cos δ sin ω′s + (π/180) ω′s sin(φ−β) sin δ] / [cos φ cos δ sin ωs + (π/180) ωs sin φ sin δ] = 0.9149/0.7615 = 1.2014
- Beam (4.1784 − 1.9884) × 1.2014 = 2.631; diffuse 1.9884 × (1 + cos 13)/2 = 1.963; ground 4.1784 × 0.2 × (1 − cos 13)/2 = 0.0107
- H_T = 2.631 + 1.963 + 0.0107 = 4.605 kWh/m²/day
Answer: H_T ≈ 4.6 kWh/m²/day, 10 % more than flat
From a sunshine card to daily radiation: Jodhpur, 15 April · Jodhpur φ = 26.29° N; n = 105; a Campbell-Stokes card shows 10.2 h of bright sunshine (illustrative reading); FAO-56 a = 0.25, b = 0.50
- δ = 9.41°; ωs = 94.7°; N = 2ωs/15 = 12.63 h; n/N = 10.2/12.63 = 0.808
- H0 = (24 × 3,600 × 1,361/π)(1 + 0.033 cos(360 × 105/365))(cos φ cos δ sin ωs + (π ωs/180) sin φ sin δ) = 10.33 kWh/m²/day
- H = 10.33 × (0.25 + 0.50 × 0.808) = 6.755 kWh/m²/day
- A cloudless day would give 0.75 H0 = 7.75; NASA's April mean for Jodhpur is 7.13
Answer: H ≈ 6.75 kWh/m²/day (KT = 0.65)
Is a twice-a-year tilt change worth it in Pune? · Pune φ = 18.52° N, NASA monthly H and Hd, isotropic model, ρg = 0.2; tilts searched from 0° to 60° in 0.05° steps
- Flat: Σ H × days = 1,874 kWh/m² a year
- Best fixed tilt: 20.6° gives 1,960 kWh/m² (+4.6 %); tilt = latitude (18.5°) gives 1,959
- Apr-Sep best tilt: 0° (904 kWh/m²); Oct-Mar best tilt: 36.1° (1,136 kWh/m²)
- Twice a year: 2,040 kWh/m² (+8.9 % over flat, +4.1 % over the best fixed tilt); every month: 2,054
Answer: Two changes a year add about 4 % over the best fixed tilt; worth it on a ground-mounted array, rarely on a rooftop
Flat-plate collectors: energy balance, losses and the Hottel-Whillier-Bliss equation
Which syllabus covers thisAU CME365 Unit 2VTU BME654B Module 2; BEE654B Module 2SPPU 402051B Unit 2MU MEDLO7032 Module 2; EEDO5011 Module 2KTU MET445 Module 2; PEEET412 Module 2
Where does the sunlight on a flat-plate collector go?
Follow 800 W/m² onto the book's collector: a 2 m² copper absorber with 9 water tubes (risers), one low-iron glass cover and a selective coating. The glass passes 90.5 % and the coating absorbs 95 %, so S = 800 × 0.905 × 0.95 = 688 W/m² is absorbed. The hot plate then loses heat through the top, back and edges, in proportion to the loss coefficient UL = 4.33 W/m²K. With water entering at 40 °C in 30 °C air, 1,150 W reaches the water: efficiency 71.8 % and outlet 46.9 °C.
Why does the coating matter so much for the top loss?
Top loss is convection across the air gap plus infrared radiation from plate to glass. Klein's correlation gives both. Black paint absorbs well (α 0.97) but also emits like a black body (ε 0.91), so its top loss is Ut = 5.34 W/m²K at the default state. A selective coating keeps α 0.95 but emits only ε 0.04, and Ut falls to 2.79. A second glass cover cuts black paint's Ut to 3.11 but costs another 9.5 % of the sunlight. Back and edge losses (1.04 + 0.5 W/m²K) come from BIS IS 12933's minimum insulation.
What do fin efficiency, F′ and FR mean in plain words?
Heat absorbed between two tubes must flow sideways through thin copper to reach the water, so the plate between the tubes runs hotter than the tube. The fin efficiency F = 0.957 measures that, for 0.2 mm copper at 111 mm pitch. F′ = 0.913 adds the step from tube wall into the water. FR = 0.892 goes one step further: it is the share of the ideal you get compared with a plate held at the inlet water temperature everywhere, at the test flow of 0.02 kg/s per m². Slower flow lowers FR.
How is a collector tested and rated in India?
BIS IS 12933 runs the collector outdoors at about 0.02 kg/s of water per m², 2-5 m/s wind, at several inlet temperatures. Efficiency is plotted against (Ti − Ta)/G and a straight line fitted: the intercept is FR(τα), the slope is FR·UL. The collector must reach FR(τα) ≥ 0.65 and FR·UL ≤ 5.5 W/m²K (6.0 before 2007). The book's collector gives 0.772 and 4.35, a pass. With no water flowing it would stagnate at 161 °C under 800 W/m² (black paint: 114 °C). The interactive keeps wind at 1-3 m/s, where Klein's fit stays close to a full plate-to-glass heat balance; in still air or strong wind it drifts.
What does an evacuated tube collector change?
An evacuated tube collector puts the absorber inside a glass tube with vacuum around it. With no air there is no convection, so the top loss becomes almost pure radiation, which a selective coating keeps small. In the flat plate, even with a selective coating, radiation is only 0.64 W/m²K of the 3.23 W/m²K top loss at an 80 °C inlet: the air gap carries most of the rest. That is why tubes keep their efficiency at high water temperatures and in cold, windy weather, while flat plates suit warm climates and pressurised systems.
On the book's selective collector (water in at 40 °C, 800 W/m²), wind rises from 1 to 3 m/s. The efficiency:
η goes from 72.7 % to 71.4 %. The plate is only about 25 °C above the air and the selective coating radiates little, so a windier glass changes little. Black paint loses more: 67.5 % to 65.6 %.
Worked examples
Top loss of a black-painted collector by Klein's correlation · One glass cover (εg 0.84), black paint (εp 0.91), mean plate 60 °C, air 25 °C, wind 3 m/s, tilt 13°. Back + edge loss from the IS 12933 minimum insulation
- hw = 2.8 + 3.0 × 3 = 11.8 W/m²K; C = 520 (1 − 0.000051 × 13²) = 515.52; f = (1 + 0.089 × 11.8 − 0.1166 × 11.8 × 0.91)(1 + 0.07866) = 0.8609; e = 0.430 (1 − 100/333.15) = 0.3009
- Convection part = 1/[1/((C/Tp)((Tp − Ta)/(N + f))^e) + 1/hw] = 2.841 W/m²K
- Radiation part = σ(Tp + Ta)(Tp² + Ta²)/[1/(εp + 0.00591 N hw) + (2N + f − 1 + 0.133 εp)/εg − N] = 3.006 W/m²K
- Ut = 5.847; UL = Ut + Ub + Ue = 5.847 + 1.042 + 0.5 = 7.389 W/m²K
- Heat lost per m² at a 35 °C plate-to-air difference: 7.389 × 35 = 259 W/m²
Answer: Ut ≈ 5.85 W/m²K and UL ≈ 7.39 W/m²K, so about 259 W/m² is lost
Outlet temperature and efficiency of the book's collector with black paint · UL = 6.0 W/m²K (black paint, from the top-loss estimate), G_T = 750 W/m², water in at 35 °C, air 30 °C, 0.04 kg/s through 9 risers (12.7 mm outside, 11.5 mm inside, 111 mm pitch), 0.2 mm copper (k 390 W/mK), τ 0.905, α 0.97, Ac 2 m²
- m = √(UL/kδ) = √(6/(390 × 0.0002)) = 8.771 1/m; F = tanh(m(W − D)/2)/(m(W − D)/2) = 0.9422
- Flow per riser 4.44 g/s: Re = 684 (laminar), hfi = 4.36 k/Di = 235.7 W/m²K
- F′ = (1/UL)/(W[1/(UL(D + (W − D)F)) + 1/(πDi hfi)]) = 0.8832
- FR = (ṁcp/AcUL)(1 − e^(−AcULF′/ṁcp)) with cp 4,180 J/kgK: FR = 0.8558
- S = 750 × 0.905 × 0.97 = 658.4 W/m²; Qu = 2 × 0.8558 × (658.4 − 6 × 5) = 1,076 W
- To = 35 + 1,076/(0.04 × 4,180) = 41.43 °C; η = 1,076/(2 × 750) = 71.7 %
Answer: Outlet ≈ 41.4 °C, Qu ≈ 1,076 W, efficiency ≈ 72 %
Reading a collector test sheet and using it on a Chennai noon · An original test: η = 0.705 at (Ti − Ta)/G = 0.01 m²K/W and η = 0.545 at 0.05. Use: water in at 55 °C, air 32 °C, G_T = 850 W/m², 2 m² collector
- Slope FR·UL = (0.705 − 0.545)/(0.05 − 0.01) = 4 W/m²K; intercept FR(τα) = 0.705 + 4 × 0.01 = 0.745
- BIS IS 12933: FR(τα) ≥ 0.65 and FR·UL ≤ 5.5, so it passes
- x = (55 − 32)/850 = 0.0271; η = 0.745 − 4 × 0.0271 = 0.637
- Useful heat = 0.637 × 850 × 2 = 1,082 W
- Zero-output point x = FR(τα)/FR·UL = 0.1863: the line reaches η = 0 at Ti ≈ 32 + 0.1863 × 850 = 190 °C
Answer: FR(τα) = 0.745, FR·UL = 4 W/m²K (passes BIS); η ≈ 63.7 % and 1,082 W at the Chennai noon
Solar water heating: thermosiphon systems, storage and sizing
Which syllabus covers thisAU CME365 Unit 2 (storage)VTU BME654B Module 2 (storage); BEE654B Module 2 (SWH)SPPU 402051B Unit 2MU MEDLO7032 Module 2; EEDO5011 Module 2KTU MET445 Module 2
How does a thermosiphon heater work, and what does its day look like?
Water warmed in the collector becomes lighter and rises into a tank mounted above it; cooler tank water sinks back down to the collector. No pump, no electricity. The interactive runs a 100 LPD system (2 m² of the chapter-5 collector, 100 L tank) through an average January day in Chennai: the morning draw cools the tank, the sun lifts it to 83.5 °C by afternoon, and the electric backup adds only 0.56 kWh a day, a solar fraction of 86 %. A collector at the BIS minimum gives 70 %.
How big a system does a household need?
The Indian rule of thumb (Vikaspedia, the government portal): a 2 m² flat-plate collector heats about 100 litres a day, with about 50 litres of tank per m² of collector. Heating 100 litres from 24 °C (Chennai's January air) to 60 °C takes m·cp·ΔT = 98.3 × 4.18 × 36 = 4.1 kWh. Over a year at a 36 °C rise that is 1,500 kWh, which is where the portal's "up to 1,500 units saved per 100 litres a day" comes from. The book's model saves 1,183 kWh a year in Chennai after tank losses and monsoon days.
Sensible and latent heat storage: what is the difference?
Sensible storage stores heat by raising a material's temperature: Q = m cp ΔT. A 100 L tank heated from 30 to 60 °C holds 3.42 kWh. Latent storage melts a phase-change material (PCM) at a fixed temperature: paraffin waxes melt between 20 and 70 °C and store 150-250 kJ/kg, so the same 3.42 kWh needs 49-82 kg of wax with almost no temperature swing. The catch: waxes conduct heat poorly (about 0.2 W/mK) and need fins. Thermochemical storage (reversible reactions) packs more but is still largely in the lab.
Thermosiphon or pumped? Flat plate or evacuated tube?
A thermosiphon needs the tank above the collector and suits homes. Hotels, hostels and hospitals use pumped (forced-circulation) systems with many collectors and a controller that runs the pump when the collector is warmer than the tank. Flat plates are robust, take mains pressure and suit warm Indian sites. Evacuated tubes hold their efficiency at high temperatures and in cold, windy places such as the hills, but are more fragile. In hard-water areas an indirect system heats the tank through a heat exchanger, so scale stays out of the collector.
How much heat does a tank lose overnight?
A tank cools like a cup of tea: T − Ta decays as e^(−t/τ) with τ = M cp/UA. The book's 100 L tank with UA = 1.5 W/K (about 50 mm of foam, assumed) has τ = 76 hours, so a 70 °C tank in 25 °C air loses 6.6 °C in 12 hours. Morning hot water therefore depends on insulation, and a tank drawn in the morning starts the day at its coldest, which the interactive shows as a dip before the sun takes over.
The 100 LPD system with the chapter-5 collector is moved to Pune for an average July (monsoon) day. Its solar fraction is closest to:
Pune's July sun is only 3.31 kWh/m² a day on the horizontal (NASA POWER), so the tank peaks at 56.3 °C and the solar fraction is 51 %. The backup heater supplies 2.01 kWh a day.
Worked examples
Sizing a solar water heater for a Jodhpur family in January · 150 litres a day at 60 °C, mains water at the January mean air temperature 15.42 °C (NASA POWER), collector facing south at 26°, daily collector efficiency 45 % (assumed)
- Mass = 150/1.0171 = 147.5 kg; heat = 147.5 × 4.18 × (60 − 15.42) = 27,481 kJ = 7.63 kWh/day
- January sun on the 26° tilt (Liu-Jordan, NASA H = 4.2564 kWh/m²): H_T = 5.77 kWh/m²/day
- Area = 7.63/(0.45 × 5.77) = 2.94 m²
- Rule of thumb: 2 m² per 100 litres gives 3 m² for 150 litres
Answer: About 2.9 m² of collector, close to the 3 m² rule of thumb, with a 150 L tank
How much a 200 L tank cools overnight · 200 L at 68 °C at 6 pm, air 25 °C all night, tank UA = 2.4 W/K (assumed), no draws for 12 hours
- Mass 196.6 kg; τ = M cp/UA = 196.6 × 4180/2.4 = 342,467 s = 95.1 h
- T(12 h) = 25 + (68 − 25) e^(−12/95.1) = 62.9 °C
- Heat lost = 196.6 × 4180 × (68 − 62.9) = 1.16 kWh
Answer: The tank is at 62.9 °C by morning, about 1.16 kWh lost
Storing 5 kWh of heat: water tank or paraffin · 5 kWh between 45 and 65 °C. Water cp 4.18 kJ/kgK (NIST); a paraffin melting near 58 °C with 200 kJ/kg latent heat (middle of the 150-250 kJ/kg range); count only the latent part for the wax
- Heat = 5 × 3,600 = 18,000 kJ
- Water: m = 18,000/(4.18 × 20) = 215 kg, about 218 litres (NIST 985.66 kg/m³ at 55 °C)
- Paraffin: m = 18,000/200 = 90 kg
Answer: About 215 kg of water against 90 kg of wax, a 2.4× cut in mass, at the price of poor conductivity
Concentrating collectors: concentration ratio, troughs, dishes, towers and tracking
Which syllabus covers thisAU CME365 Unit 2VTU BEE654B Module 2SPPU 402051B Unit 2; 203141 Unit 5MU MEDLO7032 Module 2; EEDO5011 Module 2KTU MET445 Module 2; PEEET412 Module 2 (tower)
What is the concentration ratio, and what limits it?
Concentration ratio C = aperture area / receiver area. The Godawari trough in Rajasthan has a 5.67 m wide aperture (817 m² per 144 m assembly). On a 70 mm receiver tube that is W/(πd) = 25.8 on the tube's whole surface, or W/d = 81 on its shadow width, matching TERI's 70-80 suns for troughs. The sun is a disc of half-angle 0.2664°, not a point, so no mirror can beat 1/sin θs = 215 for a line focus or 46,266 for a point focus.
Why must concentrators track the sun, and how accurately?
A concentrator only accepts rays within its acceptance angle. With an 80° rim angle the trough's focal length is 1.69 m and the rim mirror is 2.88 m from the tube, so a 70 mm tube catches rays within ±0.7°. The sun's disc already uses ±0.27°, leaving ±0.43° for tracking and mirror errors. Troughs and linear Fresnel track on one axis; dishes and towers on two. A heliostat aimed wrong by ε throws its light off by 2ε, because the receiver stays still.
Trough, linear Fresnel, dish, tower, CPC: how do they compare?
From TERI's table (IEA-ETSAP/IRENA data): parabolic troughs run fluids at 390-550 °C at 70-80 suns; central receivers (towers) 250-565 °C at over 1,000 suns; compact linear Fresnel (flat mirror strips) about 270 °C at over 60 suns; dishes above 800 °C at over 1,300 suns, with an engine at the focus. A compound parabolic concentrator (CPC) is non-imaging: it accepts every ray within ±θa without tracking, with C = 1/sin θa, so ±30° gives only C = 2 but works on diffuse-rich days.
Where does concentrated solar power stand in India?
India built 329.5 MW of CSP under the National Solar Mission, but only 101 MW was running in 2023 (TERI). Godawari's 50 MW trough at Nokh was commissioned on 19 June 2013: 480 collector assemblies of about 817 m² make a 392,400 m² field heating oil from 293 to 390 °C. Its power sells at Rs 12.20/kWh, while solar PV auctions have reached Rs 2.55/kWh (NTPC, 2024). CSP's remaining case is cheap thermal storage for evening power.
How hot can a receiver get?
With no fluid flowing, a receiver heats until it loses what it absorbs: ε σ (T⁴ − Ta⁴) + h (T − Ta) = ρ γ α C Gb. For the trough at C = 25.8 with mirror reflectance 0.94, receiver α 0.95 and ε 0.10, h = 10 W/m²K and 800 W/m² of beam (all assumed), it settles at 879 °C. In service the oil carries the heat away and runs at 390 °C. Higher C means higher possible temperature, which is why dishes and solar furnaces reach far beyond troughs.
The trough's 70 mm tube is swapped for a 35 mm tube, doubling C. With a 0.4° pointing error, the share of reflected light hitting the tube:
Ray tracing gives an intercept of 100 % for the 70 mm tube but 70.9 % for 35 mm: rays from the rim, 2.88 m away, miss a small tube once the error plus the sun's ±0.27° exceeds its ±0.35° acceptance.
Worked examples
Geometry of a trough with a 90° rim angle · Aperture W = 5.674 m (Godawari collector assembly), rim angle ψ = 90°, receiver tube d = 80 mm, sun half-angle 0.2664°
- f = W/(4 tan(ψ/2)) = 5.674/(4 × 1) = 1.4184 m
- Rim distance r = W/(2 sin ψ) = 2.8368 m
- C = W/(πd) = 5.674/(π × 0.08) = 22.57
- Acceptance θa = asin((d/2)/r) = asin(0.04/2.8368) = 0.808°; check C = sin ψ/(π sin θa) = 22.57
- Left for tracking and mirror errors: 0.808 − 0.2664 = 0.542°
Answer: f = 1.418 m, C = 22.6, acceptance ±0.81°, so tracking must hold within about ±0.54°
How far is a dish from the concentration limit? · Sun radius 695,500 km at 1 au = 149,597,870.7 km; a dish 8 m across focusing on a 200 mm receiver
- θs = atan(695,500/149,597,870.7) = 0.2664°
- Line-focus limit 1/sin θs = 215.1; point-focus limit 1/sin² θs = 46,266
- Dish: C = (8/0.2)² = 1,600, which is 3.5 % of the point-focus limit
Answer: C = 1,600 against a limit of 46,266: real dishes stay well below it because of mirror errors and tracking
Stagnation temperature of a trough receiver · C = 30, intercept 0.90, mirror 0.94, receiver α 0.95, ε 0.10, h 10 W/m²K, beam 800 W/m², air 30 °C (assumed)
- Absorbed per m² of receiver: 0.94 × 0.90 × 0.95 × 30 × 800 = 19,289 W/m²
- Solve 0.10 σ (T⁴ − 303.15⁴) + 10 (T − 303.15) = absorbed by bisection
- T = 1,171 K = 898 °C
- Radiation alone (h = 0) would give 1,086 °C: convection matters
Answer: About 898 °C with nothing flowing
Solar thermal applications: stills, dryers, cookers and solar ponds
Which syllabus covers thisAU CME365 Unit 2VTU BEE654B Module 2SPPU 402051B Unit 2MU MEDLO7032 Module 2; EEDO5011 Module 2 (pond)KTU MET445 Module 2
How much drinking water does a solar still make?
A basin still is a shallow black tray of salty or dirty water under a sloping glass cover. Sunlight warms the water, vapour rises, condenses under the cooler glass and runs down into a channel; the salt stays behind. Dunkle's correlation gives the vapour transfer from water to glass. On an average April day in Chennai, 3 cm of water makes 4.14 kg of distilled water per m² of basin, a daily efficiency of 42 %, with the water peaking at 68 °C. Deeper water stores heat and produces more of its output after sunset.
How is a box solar cooker rated?
BIS IS 13429 rates a box cooker by two figures of merit. F1 = (stagnation tray temperature − air)/sunlight, from an empty cooker at steady state: at least 0.12 for grade A and 0.11 for grade B. F2 comes from heating 8 litres of water per m² of aperture from 60 to 90 °C and must exceed 0.40. In the model, a 0.25 m² grade-A cooker on an April day in Chennai would stagnate at 141 °C at noon and heats 2 kg of water to 90 °C in 133 minutes from 10:00.
What decides how fast a solar dryer works?
Drying is evaporation, so the heat needed is set by the water to remove. 20 kg of fresh chillies at 80 % moisture (wet basis, assumed) dried to 10 % must lose 20 × (0.8 − 0.1)/(1 − 0.1) = 15.6 kg of water. At hfg = 2,394 kJ/kg (NIST, 45 °C) that is 10.3 kWh of heat. A 2 m² air heater at 25 % efficiency (assumed) on an April day in Chennai removes 4.8 kg a day, so the batch takes about 3.3 days, faster and cleaner than open-sun drying on the ground.
How does a solar pond store heat for months?
A salt-gradient pond keeps saltier, denser water at the bottom. Warm brine stays down because it is still heavier than the fresher water above, so the gradient layer cannot convect and acts as insulation; the top stays near air temperature. With a 1.2 m gradient layer under 0.2 m of fresh water, 35 % of the April sunlight reaches the store, which settles at 94 °C while 20 W/m² is drawn off (December: 59 °C). Ponds reach about 90 °C but turn under 2 % of sunlight into electricity; Bhuj's pond supplied 80,000 litres of hot water a day.
What else does solar heat do? Air heaters to furnaces in one line each
Solar air heaters warm air for drying and space heating; air holds about a quarter of water's heat per kg (cp 1.005 against 4.18 kJ/kgK), so they need large ducts and fans. Passive space heating uses south glazing and thick walls; solar cooling drives absorption chillers (the Refrigeration and Air Conditioning book has a LiBr-water chiller interactive). Solar pumping today mostly means PV pumps. Heliostats are tracking flat mirrors (chapter 7). The solar furnace at Odeillo, France, concentrates sunlight to 3,500 °C for materials research.
The still's water depth is doubled from 3 cm to 6 cm (April, Chennai). The day's distilled water:
Output goes from 4.14 to 3.81 kg/m² and the share made after sunset rises from 10 % to 27 %. Deeper water heats more slowly, so it is cooler at midday, but it holds heat into the night. A still never needs boiling: evaporation runs at any temperature.
Worked examples
Distilled water from a basin still by the daily-efficiency method · Chennai in April: H = 6.4973 kWh/m² per day on the horizontal (NASA POWER); still daily efficiency 35 % (assumed); hfg at 50 °C from NIST = 2,382 kJ/kg; a household wants 20 litres of drinking water a day
- Evaporation energy = 0.35 × 6.4973 × 3.6 = 8.187 MJ/m²
- Distillate = 8.187 × 1,000/2,382 = 3.44 kg/m²
- Basin area for 20 litres = 20/3.44 = 5.8 m²
Answer: About 3.4 litres per m² a day, so roughly 6 m² of basin for 20 litres
Figures of merit of a box cooker from test readings · Stagnation test: tray at 125 °C in 33 °C air under 820 W/m². Load test (BIS method): 2 kg of water on a 0.25 m² aperture heats from 60 to 90 °C in 80 minutes at a mean 850 W/m² and 32 °C
- F1 = (125 − 33)/820 = 0.1122: grade B (BIS: ≥ 0.12 grade A)
- (MC)w = 2 × 4,180 = 8,360 J/K; t = 80 × 60 = 4,800 s
- F2 = (F1 (MC)w/(A t)) ln[(1 − (60 − 32)/(F1 × 850))/(1 − (90 − 32)/(F1 × 850))]
- F2 = (0.1122 × 8,360/(0.25 × 4,800)) × 0.5894 = 0.461
Answer: F1 = 0.112 (grade B) and F2 = 0.46, above the 0.40 minimum
Storage temperature of a salt-gradient pond · 24-hour mean sunlight 230 W/m²; mean transmission through the 1.2 m gradient layer 0.40; brine conductivity 0.6 W/mK; ground and side loss 1.0 W/m²K; 25 W/m² drawn off; top water at 28 °C (all illustrative)
- Heat reaching the store on average: 230 × 0.40 = 92 W/m²
- ΔT = (92 − 25)/(0.6/1.2 + 1.0) = 44.7 °C
- Store temperature = 28 + 44.7 = 72.7 °C; thermal efficiency 25/230 = 10.9 %
Answer: About 73 °C, delivering 11 % of the sunlight as heat
Photovoltaic physics: band gap, absorption and the efficiency limit
Which syllabus covers thisAU CME365 Unit 2; EE3014 Unit 1VTU BME654B Module 2; BEE654B Module 2SPPU 402051B Units 1, 3; 203141 Unit 5MU MEDLO7032 Module 3; EEDO5011 Module 3KTU MET445 Module 2; PEEET412 Module 2
Which part of sunlight can a solar cell actually use?
A photon carries energy E = hc/λ, which is 1,239.84/λ electronvolts with λ in nanometres. A semiconductor only frees an electron if the photon brings at least its band gap Eg. Silicon's gap is 1.12 eV, so its cut-off is 1,107 nm: longer, infrared photons pass straight through. Of the 1,000 W/m² in the AM1.5G reference spectrum, 80.8 % arrives in photons above silicon's gap. Each of them hands over only Eg; the rest becomes heat. That leaves 49 % usable, 32 % thermalisation heat and 19 % never absorbed. Drag the band gap and watch the three shares trade places.
How does a p-n junction make current, and why are silicon wafers so thick?
An absorbed photon lifts an electron across the gap and leaves a hole. In a p-n junction the built-in electric field of the depletion region sweeps electrons to the n side and holes to the p side, so they flow through the external circuit instead of recombining. Each cell gives about 0.69 V at open circuit in the book's module (49.61 V over 72 cells). Silicon has an indirect gap and absorbs weakly near it: at 600 nm light travels 2.4 µm before dropping to 1/e, at 1,000 nm 156 µm. So wafers are a few hundred micrometres thick and textured to trap light; direct-gap materials such as CdTe and perovskites absorb in far thinner films.
What is the highest efficiency a single-junction cell can reach?
Two limits. The ultimate efficiency counts only the band-gap loss: 49 % for silicon. The Shockley-Queisser detailed-balance limit adds the unavoidable radiative recombination: a cell at 300 K must glow back some light. On the AM1.5G spectrum it peaks at 33.7 % near 1.34 eV (the published value is 33.7 % at 1.34 eV); silicon at 1.12 eV gets 33.4 %, GaAs at 1.42 eV 33.1 %, a 1.55 eV perovskite 31.4 %. A hotter cell loses more. Real cells also lose light to reflection and contact shading and carriers to defects, so the best certified silicon cell in NLR's table is 27.1 % (LONGi, Nov 2023).
What kinds of solar cells exist, and how good are they?
First generation is crystalline silicon: monocrystalline PERC cells like the book's module, TOPCon and heterojunction; NLR's table lists 26.1 % for a single-crystal cell and 27.1 % for a heterojunction back-contact cell. Second generation is thin film: CdTe 22.6 %, CIGS 23.6 %, amorphous silicon 12.7 %. Third generation aims past the single-junction limit or at very low cost: perovskite 26.1 %, dye-sensitised 13 %, quantum dot 19.1 %, and tandems that stack two gaps, perovskite on silicon 33.7 %, above the 33.7 % single-junction limit. These are small certified research cells (dates in the table); modules run lower.
How is a silicon module made, from sand to panel?
Quartz is reduced with carbon in an arc furnace to metallurgical-grade silicon, then purified to polysilicon (the Siemens process deposits it from trichlorosilane). A single-crystal ingot is pulled from the melt by the Czochralski method and wire-sawn into wafers; the book's module uses 182 mm wafers cut into 91 × 182 mm half cells, 144 per module. Each wafer is textured, given a phosphorus-diffused junction, an anti-reflection coating and printed silver contacts. Cells are soldered into strings, laminated in EVA between 3.2 mm glass and a backsheet, framed, flash-tested and qualified to IEC 61215 (design) and IEC 61730 (safety).
Silicon's band gap is 1.12 eV. A material with a 2.0 eV gap would have a single-junction efficiency limit that is
Each absorbed photon does give more voltage, but at 2.0 eV the cut-off moves to 620 nm and most of the sunlight passes through. The SQ limit falls from 33.4 % to 22.8 %; the best single gap is near 1.34 eV.
Worked examples
Cut-off wavelength and photon energy · Silicon Eg = 1.12 eV, CdTe Eg = 1.5 eV, a green photon at 550 nm. hc = 1,239.84 eV·nm.
- Silicon cut-off: λg = 1,239.84/1.12 = 1,107 nm (near infrared).
- CdTe cut-off: λg = 1,239.84/1.5 = 827 nm, so CdTe ignores everything beyond deep red.
- Green photon: E = 1,239.84/550 = 2.25 eV.
- In silicon it delivers 1.12 eV; the extra 1.13 eV (50 % of it) is lost as heat.
Answer: λg = 1,107 nm (Si) and 827 nm (CdTe); a 550 nm photon wastes 50 % of its energy in silicon.
The most current light can give · 1,000 W/m² of monochromatic 600 nm light on a cell that turns every photon into one electron. h = 6.626 × 10⁻³⁴ J s, c = 2.998 × 10⁸ m/s, q = 1.602 × 10⁻¹⁹ C.
- Photon energy = hc/λ = 3.311 × 10⁻¹⁹ J.
- Photon flux = 1,000 / 3.311 × 10⁻¹⁹ = 3.02 × 10²¹ photons/m²/s.
- Current density = q × flux = 483.9 A/m² = 48.4 mA/cm².
- At 0.6 V the cell would deliver 290 W/m², an efficiency of 29 %.
Answer: At most 48.4 mA/cm²; even with perfect collection, 0.6 V gives only 29 %.
How much light a silicon wafer absorbs · A 180 µm wafer (typical, assumed). α(1,000 nm) = 64 /cm and α(1,100 nm) = 3.5 /cm (Green and Keevers). Beer-Lambert: absorbed = 1 − e^(−αd).
- 1,000 nm, one pass: 1 − e^(−64 × 0.018) = 68.4 %.
- With a rear mirror (two passes, 0.036 cm): 90 %.
- 1,100 nm, one pass: 1 − e^(−3.5 × 0.018) = 6.1 %; two passes 11.8 %.
- Near the band edge silicon is almost transparent, so cells texture the surface and add rear reflectors to make light bounce many times.
Answer: 68 % at 1,000 nm in one pass (90 % with a mirror), only 6 % at 1,100 nm.
The solar cell model: I-V curve, fill factor, STC, NOCT and temperature
Which syllabus covers thisAU EE3014 Unit 1 (characteristics)VTU BEE654B Module 2SPPU 402051B Unit 3; 203141 Unit 5MU MEDLO7032 Module 3 (efficiency factors); EEDO5011 Module 3KTU MET445 Module 2
What is the single-diode model of a solar module?
A current source IL (the light-generated current) in parallel with a diode, a shunt resistance Rsh for leakage, and a series resistance Rs for the fingers, busbars and solder. Five numbers describe a module. For the book's Waaree WSMD-540 at STC (CEC library): IL = 13.85 A, I0 = 2.75 × 10⁻¹¹ A, Rs = 0.161 Ω, Rsh = 143 Ω and a = 1.843 V, which is n × 72 cells × kT/q with ideality n = 1. The equation is implicit in I; the interactive solves it exactly. Slide irradiance, temperature, Rs and Rsh and watch the curve and its fill factor.
What are Isc, Voc, the maximum power point and the fill factor?
Short-circuit current Isc flows at zero volts, open-circuit voltage Voc appears at zero current, and between them power V × I peaks at the maximum power point. The datasheet gives 540 W at 41.75 V and 12.95 A, Voc 49.61 V and Isc 13.83 A. Fill factor FF = Pmp/(Voc × Isc) = 78.7 %: how square the curve is. Efficiency = 540 W/(1,000 W/m² × 2.574 m²) = 20.98 %. A high Rs bends the knee (four times Rs drops the model's FF to 67 %); a low Rsh tilts the flat top (Rsh ÷ 16 gives 55 %).
What are STC and NOCT, and why do datasheets quote both?
STC (standard test conditions) is 1,000 W/m², AM1.5 spectrum and a 25 °C cell: the rating. Real modules run hotter. NOCT is the cell temperature with the module open-circuited at 800 W/m², 20 °C air and 1 m/s wind: Waaree prints 43 ± 2 °C and 406.7 W at those conditions, 75.3 % of STC; the model gives 404.8 W at 43 °C. A typical Indian datasheet for a different module (Tata TP540HG10B) shows 396 of 550 W, 72 %. The CEC test of the WSMD-540 measured NOCT 44.94 °C. Some newer datasheets print NMOT, measured under the same 800 W/m², 20 °C, 1 m/s conditions.
How much power does a module lose on a hot Chennai afternoon?
Heat raises I0 sharply, so Voc falls; Isc rises slightly as the band gap shrinks. Waaree's datasheet gives Pmax −0.34 %/°C, Voc −0.25 %/°C, Isc +0.05 %/°C. The CEC model gives −0.353, −0.288 and +0.028 %/°C: the sources differ by a few hundredths, normal for fitted and measured values. In May Chennai's air averages 31.5 °C (NASA POWER), so at 1,000 W/m² the NOCT model puts the cell at 62.7 °C and the module at 468 W, 13 % below its rating.
How do I-V curves change with sunlight?
Isc is proportional to irradiance: 13.83 A at 1,000 W/m², about a fifth of that at 200. Voc changes only with the logarithm of irradiance, so it barely moves until light is very low. Efficiency therefore holds up in dim light: the model loses 3 % of its STC efficiency at 200 W/m², and Waaree states an average 4.5 % reduction (measured to IEC 60904-1, ± 3 % uncertainty). At high light the I²Rs loss in the series resistance grows, which is why fill factor falls slightly on the brightest days.
A module heats from 25 °C to 60 °C in full sun. Which falls the most, in per cent?
Isc rises a little (+0.05 %/°C), Voc falls 0.25 %/°C and Pmax falls fastest, 0.34 %/°C, because both the voltage and the fill factor drop: 11.9 % over 35 °C.
Worked examples
Fill factor and efficiency from the datasheet · Waaree WSMD-540 at STC: Pmax 540 W, Voc 49.61 V, Isc 13.83 A, Vmp 41.75 V, Imp 12.95 A; module 2,272 × 1,133 mm.
- Voc × Isc = 49.61 × 13.83 = 686.1 W (the outer rectangle).
- FF = 540/686.1 = 0.787; check Vmp × Imp = 41.75 × 12.95 = 540.7 W.
- Area = 2.272 × 1.133 = 2.574 m².
- Efficiency = 540/(1,000 × 2.574) = 20.98 %.
Answer: FF = 78.7 %, efficiency 20.98 % (the datasheet prints 20.98 %).
Module power on a May afternoon in Chennai · Air 31.51 °C (NASA POWER May mean), 1,000 W/m². Datasheet: NOCT 43 °C, Pmax −0.34 %/°C, Voc −0.25 %/°C.
- Cell temperature: Tc = Ta + (NOCT − 20)/800 × G = 31.51 + 23/800 × 1,000 = 60.26 °C.
- ΔT = 35.26 °C above STC.
- P = 540 × (1 − 0.0034 × 35.26) = 475.3 W.
- Voc = 49.61 × (1 − 0.0025 × 35.26) = 45.24 V.
- The full diode model with the CEC NOCT of 44.94 °C gives a hotter cell (62.7 °C) and 468 W.
Answer: About 475 W by the datasheet coefficients (12 % below rating); 468 W by the model.
Checking the datasheet MPP with the five parameters · STC parameters: IL = 13.8456 A, I0 = 2.754 × 10⁻¹¹ A, Rs = 0.1612 Ω, Rsh = 142.91 Ω, a = 1.843 V. Find I at V = 41.75 V.
- Write f(I) = IL − I0(e^((V+IRs)/a) − 1) − (V+IRs)/Rsh − I and solve f = 0 by Newton's method from I = IL.
- Step 1: f(13.846) = −0.945 A, new I = 12.9518 A.
- Step 2: f(12.9518) = −0.0019 A, new I = 12.95 A.
- Step 3: new I = 12.95 A; P = 41.75 × 12.95 = 540.7 W.
Answer: I = 12.95 A at 41.75 V, matching the datasheet's 12.95 A and 540 W.
Modules and arrays: series and parallel strings, partial shading and bypass diodes
Which syllabus covers thisAU not in syllabusVTU BEE654B Module 2 (series/parallel)SPPU 402051B Units 3, 5; 203141 Unit 5MU EEDO5011 Module 3KTU not in syllabus
What happens when modules are wired in series and in parallel?
In series the same current flows through every module and the voltages add; in parallel the voltage is common and the currents add. The book's roof has 6 WSMD-540 modules: one string of 6 gives Voc 297.7 V and Vmp 250.5 V at 12.95 A; two strings of 3 give half the voltage at twice the current. A string can only carry the current its weakest module allows, so mismatched or dirty modules drag down the whole string, and parallel strings must have matching voltages. Cold mornings raise Voc, so string length is set by the coldest cell temperature against the inverter and the module's 1,500 V limit.
Why can shading one corner of a module cost a third of its power?
72 cells sit in series in three substrings of 24, each with a bypass diode. A shaded cell cannot carry the string current, so either the whole substring slows to its pace or the bypass diode turns on and the substring drops out. Shade four corner cells by 80 % and the model gives 354 W with the diode conducting, about two thirds of 541 W; remove the diodes and only 130 W is left. The P-V curve now has two peaks. Pick shadow patterns, depth, diodes and string length, and watch where a simple tracker gets stuck.
What is a hot spot, and do bypass diodes stop it?
A shaded cell forced to carry current is pushed into reverse bias and turns electrical power into heat in a few square centimetres. Without diodes the rest of the string can drive it to breakdown. A bypass diode limits the reverse voltage to roughly the other 23 cells' voltage plus 0.5 V, about 15 V. With the model's breakdown at −15 V (pvlib's example value; real cells vary), a fully shaded cell still dissipates about 152 W, which can scorch the backsheet. Diodes save energy and prevent the worst voltages; they do not remove the heating, which is why leaves, bird droppings and dust edges matter.
How does a maximum power point tracker find the peak?
Perturb and observe (P&O) nudges the operating voltage, keeps going if power rose and reverses if it fell, so it oscillates around the peak; the Power Electronics book has a perturb-and-observe interactive. Incremental conductance uses dP/dV = I + V dI/dV = 0, so at the peak dI/dV = −I/V: compare the two and move towards equality; it can stop at the peak and handles fast-changing light better. Both climb the hill they start on. With the bottom row of one module in a string of 3 shaded by 80 %, a tracker coming down from open circuit stops at 394 W while the global peak is 1062 W, so modern inverters sweep the curve now and then.
What do power optimisers and micro-inverters change?
A string inverter tracks one maximum power point for the whole string. Distributed MPPT gives every module its own: a DC-DC power optimiser behind each module, or a micro-inverter that turns each module's DC straight into grid AC. In the same 3-module string with one module's bottom row shaded, string tracking gets 1062 W; module-level tracking gets 2 × 541 + 129 = 1211 W, 149 W more. The price is more electronics on the roof; they pay off on shaded or multi-direction roofs and allow module-level monitoring and shutdown.
One cell of a 540 W module is completely covered by a leaf. The module has three bypass diodes. Roughly how much power is left?
The covered cell blocks its whole 24-cell substring, so that substring's bypass diode turns on and two of three substrings keep working: the model gives 354 W of 541 W (65 %), not 71/72.
Worked examples
How many modules in a string? · WSMD-540: Voc 49.61 V at 25 °C, β(Voc) = −0.25 %/°C, maximum system voltage 1,500 V. Coldest cell temperature (assumed): 5 °C for a Jodhpur winter dawn, 20 °C for Chennai.
- Jodhpur: Voc = 49.61 × (1 + 0.0025 × 20) = 52.09 V per module.
- Modules in series ≤ 1,500/52.09 = 28.8 → 28.
- Chennai: Voc = 49.61 × (1 + 0.0025 × 5) = 50.23 V → 1,500/50.23 = 29.9 → 29.
- The inverter's own maximum DC input (often lower than 1,500 V) is checked the same way.
Answer: At most 28 modules per string in Jodhpur and 29 in Chennai.
Heat in a fully shaded cell · One cell of the module fully shaded, bypass diodes fitted (0.5 V), cell breakdown −15 V (Bishop model, pvlib example). Model operating point at the global maximum.
- The module works at 354 W with one diode conducting.
- Around the bypassed loop the 23 lit cells plus the diode's 0.5 V drive the dark cell to its breakdown: it sits at −15.02 V.
- The lit cells then carry 10.09 A at 0.631 V each, 14.52 V in all, so the substring sits at 14.52 − 15.02 = −0.5 V, the diode's drop; the rest of the string current goes through the diode.
- Heat in the dark cell = 15.02 V × 10.09 A = 152 W.
Answer: About 152 W concentrated in one cell: a hot spot, even with the diode on.
Incremental conductance: which way to move? · The book's module at STC, two readings: 40 V gives 13.3154 A and 40.5 V gives 13.2407 A.
- dI/dV ≈ (13.2407 − 13.3154)/0.5 = −0.1494 A/V.
- −I/V at 40.5 V = −13.2407/40.5 = −0.3269 A/V.
- dI/dV is greater than −I/V, so dP/dV = I + V dI/dV = 7.19 W/V is positive.
- Power still rises with voltage: the tracker increases the voltage, towards 41.75 V.
Answer: Increase the voltage (dP/dV = +7.19 W/V).
Stand-alone PV systems: batteries, charge controllers, inverters and sizing
Which syllabus covers thisAU EE3014 Unit 3VTU BME654B Module 2 (off-grid)SPPU 402051B Unit 3; 203141 Unit 6MU MEDLO7032 Module 3 (1 kW case); EEDO5011 Module 3KTU not in syllabus
What does an off-grid solar system need, and how well does a standard design work?
An array, a charge controller, a battery and, for AC loads, an inverter. Size a village home in Chennai using 3 kWh/day (assumed) with 3 days of autonomy at 24 V: the method picks a 14° tilt, July as the design month, 2 modules (1.08 kWp) and 510 Ah. A year of seeded synthetic weather then gives 5 days of loss of load (1.4 %) and 16 % of the PV energy wasted because the battery was full. Change the system, site, autonomy, depth of discharge, voltage and controller.
Which battery, and how deep may it be discharged?
Tubular lead-acid batteries are the usual choice in Indian off-grid systems; lithium-ion is lighter and lasts more cycles. A battery standard draft for PV use (IEC 61427-based, NISE/CECRI report) designs for 3 to 15 days of autonomy, limits depth of discharge to 80 % at the end of autonomy and charge current to C/10. Its Ah efficiency for lead-acid is above 85 % at 90 % charge, above 90 % at 75 % and above 95 % below 50 %; lithium-ion above 95 %. Multiply by the voltage efficiency, 12.0/13.5 V, and lead-acid returns about 80 % of the energy put in.
PWM or MPPT charge controller, and what do the converters do?
A PWM controller is a switch: it connects the array straight to the battery, so a module sits at the battery's charging voltage. The book's module at 800 W/m² on a 24 V battery is held at 27 V and gives 297 W instead of 394 W. An MPPT controller is a DC-DC buck converter, Vout = D × Vin, run at the module's maximum power point (38 V, duty 0.71). A boost converter raises voltage, Vout = Vin/(1 − D); a PWM inverter switches an H-bridge so the filtered output is a 50 Hz sine, needing more than 325 V DC for 230 V AC.
How are the battery and the array sized?
Start from the daily energy at the DC bus: 3 kWh of AC load through a 92 % inverter is 3.261 kWh. Battery capacity C = E × N/(V × DoD) for N days of autonomy. The array must refill that energy in the worst month: kWp = E/(H × η), where H is the design month's irradiation on the tilted array in kWh/m²/day (equal to peak sun hours) and η collects the temperature derate, controller, wiring and soiling losses and the battery round trip for the night share. Round up to whole modules.
What is loss-of-load probability, and how do solar pumps fit in?
Loss-of-load probability (LOLP) is the share of days the system cannot meet the load. Monthly-average sizing hides runs of cloudy days, so a day-by-day simulation reveals it: doubling the home's autonomy to 6 days (1,019 Ah) cuts its loss-of-load days from 5 to 0, and a bigger array cuts it too. Under PM-KUSUM Component B the government targets 14 lakh stand-alone solar pumps with 30 % central assistance (50 % in north-eastern, hill and island states), up to 15 HP in some cases (PIB, 23 Dec 2023). The interactive's pump stores water in a tank instead of energy in a battery.
The Chennai home is redesigned with 6 days of autonomy instead of 3. What changes?
Battery Ah = E × N/(V × DoD) doubles to 1,019 Ah; the array is set by the design month and stays 1.08 kWp. Loss-of-load days in the synthetic year fall from 5 to 0.
Worked examples
Battery bank for the village home · Load 3 kWh/day AC (assumed), inverter 92 % (assumed), 3 days of autonomy, 24 V bank, depth of discharge 80 %.
- DC energy per day = 3/0.92 = 3.261 kWh = 3,261 Wh.
- Energy to store = 3,261 × 3 = 9,783 Wh.
- C = 9,783/(24 × 0.8) = 509.5 Ah.
- Choose two 12 V blocks in series of at least that rating, for example 2 × 12 V 550 Ah.
Answer: About 510 Ah at 24 V.
Array size for the design month · Chennai, tilt 14° (best for the worst month), July H_T = 4.601 kWh/m²/day (Liu-Jordan on NASA POWER). July derate × MPPT = 0.8668, other losses 0.90, battery Wh efficiency 0.8 for the 70 % of the load used at night.
- Usable kWh per kWp per day = 4.601 × 0.8668 × 0.90 = 3.589.
- Path efficiency = 1/(0.3 + 0.7/0.8) = 0.8511.
- kWp = 3.261/(3.589 × 0.8511) = 1.067 kWp.
- In 540 W modules: 1.067/0.54 = 1.98 → 2 modules = 1.08 kWp.
Answer: 2 modules, 1.08 kWp.
PWM versus MPPT on a 24 V battery · One WSMD-540 at 800 W/m², air 25 °C (cell 49.9 °C by the NOCT model), charging a 24 V lead-acid bank at 27 V.
- Model maximum power: 394 W at 37.98 V.
- PWM holds the module at 27 V: I = 10.99 A, P = 27 × 10.99 = 296.7 W.
- MPPT (buck, 97 % assumed): 394 × 0.97 = 382.2 W, duty D = 27/37.98 = 0.71.
- PWM delivers 78 % of what MPPT does; that is why the PWM design needs 3 modules instead of 2.
Answer: PWM 297 W versus MPPT 382 W from the same module.
Grid-connected rooftop PV: design, net metering, subsidy and payback
Which syllabus covers thisAU EE3014 Unit 3VTU BME654B Module 2SPPU 402051B Unit 5 (DPR); 402049 lab (net metering)MU EEDO5011 Module 3KTU not in syllabus
What is in a grid-connected rooftop system, and what does the inverter do?
Modules on a mounting structure, DC cables to a DC distribution box (fuses, surge arrester, isolator), a grid-interactive string inverter, an AC box and a bidirectional net meter. No battery. The inverter tracks the maximum power point, synchronises with the grid and must switch off within seconds if the grid fails, so it never feeds a line a lineman thinks is dead (islanding; tested under IEC 62116:2014). String check for the book's roof: 6 modules in series, each 51.2 V open-circuit on a cold 14 °C morning, give 307 V, well under a 600 V inverter limit (assumed).
How much energy does 3.24 kWp make on a Chennai terrace?
The model takes NASA's monthly sunshine, splits each average day into hours, tilts it to 12° and runs the Waaree module's diode model at the hour's temperature, then takes 4 % for the inverter and 12 % for dust, wiring and downtime (assumed). Result: 4,810 kWh a year, 1,485 kWh per kWp. Global Solar Atlas gives 1,565 kWh/kWp for Chennai, so the model is about 5 % cautious. March is best (15.9 kWh/day), November worst (11.6). Only 38 % is used as it is made: the flat's air conditioner runs at night.
How does net metering work in Tamil Nadu?
Under TNERC's 2021 rooftop regulations a domestic consumer can net-meter up to the sanctioned load. Exported units are subtracted from imported units and billed at the retail slabs, and spare units roll over to the next bill. The settlement year runs 1 April to 31 March, and any surplus left then lapses unpaid. The alternative, net feed-in, bills imports at the slabs and pays exports at a feed-in tariff: Rs 3.61/kWh in the 2021 order (valid to March 2023; a newer figure was not found). For the flat, yearly bills fall from Rs 46,454 to Rs 5,919 with net metering, Rs 16,167 with net feed-in, Rs 26,951 with no export.
What does the subsidy cover, and how fast does the system pay back?
PM Surya Ghar pays Rs 30,000 per kWp for the first 2 kWp and Rs 18,000 for the third, so Rs 78,000 for 3 kWp or more. MNRE's benchmark cost is Rs 50,000/kW for the first 2 kW and Rs 45,000 after: Rs 1,55,800 for 3.24 kWp. Modules must be made in India from Indian cells, and the vendor maintains the system free for 5 years. The flat sits in the top slabs (about 1,004 units every two months), so every solar unit saves Rs 11-12. At Rs 50,000/kW and a 3 % yearly tariff rise (assumed) the payback is 2 years. The government's partial subsidy of domestic tariffs is not modelled; it would lower the savings.
How much does dust cost, and what does maintenance involve?
Dust settles every day. A rooftop study in Lucknow (Scientific Reports, 2025) measured losses of 0.20 to 0.70 % per day and 10.2 % in a month without rain. At 0.2 % per day, 30 days without cleaning averages a 3 % loss: about 15 kWh in a Chennai March for the book's roof. Light rain can make it worse; a heavy shower cleans. Routine O&M: wash modules with soft water in the early morning, check torque on structure bolts, look for hot spots and cracked glass, read inverter alarms, test earthing and surge arresters once a year, and compare the monthly yield with the expected yield.
What is the balance of system, and what goes in a rooftop DPR?
Balance of system is everything except the modules: structure, cables, DC combiner or distribution box with string fuses and a surge arrester, isolators, inverter, AC box, earthing, lightning protection and meters. Structures are hot-dip galvanised steel or aluminium, bolted or ballasted, designed for the site's wind speed, and raised so water drains and modules are cleaned from a walkway. A detailed project report for a rooftop or MW plant covers: site survey and shadow study, layout and tilt, energy estimate (here 4,810 kWh/yr), single-line diagram, bill of quantities, cost, subsidy, cash flows and payback, O&M plan and approvals.
The book's 3.24 kWp roof exports surplus units under TN net metering. What happens to them?
TNERC's 2021 regulation nets exports against imports in units, carries the balance forward and lets it lapse at the end of the April-March settlement year. With 11 modules instead of 6 the flat would lose 2,813 kWh a year that way.
Worked examples
Annual energy and specific yield of the book's roof · 3.24 kWp facing south at 12° in Chennai; the kernel's tilted irradiation sums to 1,933 kWh/m² a year; take a performance ratio PR = 0.75
- E = P × H_T × PR = 3.24 kW × 1,933 kWh/m² ÷ 1 kW/m² × 0.75 = 4,698 kWh/yr
- Specific yield = 4,698/3.24 = 1,450 kWh/kWp
- The hour-by-hour model (diode model, temperature, inverter, 12 % losses) gives 4,810 kWh: its PR = 1,485/1,933 = 0.768
- Global Solar Atlas, Chennai at its optimum 14°: 1,565 kWh/kWp, so 3.24 kWp would make 5,069 kWh
- Capacity factor = 4,810/(3.24 × 8,760) = 16.9 %
Answer: About 4,698 kWh/yr at PR 0.75 (4,810 kWh by the full model), a CUF near 17 %
One bi-monthly bill with and without solar · April-May: the flat uses 1,004 units; with the roof it imports 685 units and exports 577 units (model, rounded); TN LT-IA slabs from 1 Jul 2025; no credit carried in
- No solar: 400 × 4.95 + 100 × 6.65 + 100 × 8.80 + 200 × 9.95 + 200 × 11.05 + 4 × 12.15 = Rs 7,773.60
- Net metering: billed units = 685 − 577 = 108, all in the first slab: 108 × 4.95 = Rs 534.60
- Net feed-in: imports 685 units cost 1,980 + 665 + 880 + 85 × 9.95 = Rs 4,370.75; exports earn 577 × 3.61 = Rs 2,082.97; bill Rs 2,287.78
- No export (export limiter): Rs 4,370.75, and 577 units are given away
Answer: Rs 7,774 falls to Rs 535 (net metering), Rs 2,288 (net feed-in) or Rs 4,371 (no export)
String voltage on a cold morning, then the subsidy and payback · Waaree WSMD-540: Voc 49.61 V at 25 °C, temperature coefficient −0.132 V/°C; design cold 14.1 °C; inverter limit 600 V (assumed). Cost at the MNRE benchmark; first-year saving Rs 40,535 (model, net metering)
- Voc(cold) = 49.61 − 0.132 × (14.09 − 25) = 51.05 V per module (the diode model gives 51.16 V)
- 6 in series: 306.3 V ≤ 600 V. Most in one string: 600/51.05 = 11.75, so 11 modules
- Cost = 2 × 50,000 + 1.24 × 45,000 = Rs 1,55,800; CFA = 2 × 30,000 + 18,000 = Rs 78,000; you pay Rs 77,800
- Simple payback = 77,800/40,535 = 1.92 years (no tariff rise, no O&M)
Answer: 306 V string, up to 11 modules per string; Rs 77,800 after subsidy, paid back in about 1.9 years
The wind resource: Weibull distribution, power density and wind shear
Which syllabus covers thisAU CME365 Unit 3VTU BME654B Module 3; BEE654B Module 3SPPU 402051B Unit 1; 203141 Unit 4MU MEDLO7032 Module 4; EEDO5011 Module 4KTU MET445 Module 3; PEEET412 Module 1
Where does wind come from, and how much does India have?
The sun heats land, sea and latitudes unevenly; air flows from high to low pressure and the earth's spin turns it. India's monsoons and the hill gaps of Tamil Nadu (Palghat, Aralvaimozhi near Muppandal) set the pattern. MNRE puts India's gross potential at 1,163.85 GW at 150 m: Rajasthan 284.25 GW, Gujarat 180.79, Tamil Nadu 95.1. Installed wind was 59.5 GW on 31 Oct 2026 (CEA), about 5.1 % of that. Chennai's 50 m wind (NASA) peaks at 6.05 m/s in Dec and dips to 4.36 m/s in Oct.
Why describe a site's wind with a Weibull distribution?
Hourly wind speed is never steady; the Weibull curve fits how often each speed occurs with two numbers: shape k (how peaked; 2 is typical inland) and scale c (close to the mean). With k = 2 it is the Rayleigh distribution, and c = 2v̄/√π. The interactive starts at Chennai, 5.3 m/s at 50 m, lifted to 5.93 m/s at a 110 m hub: c = 6.69 m/s. The most frequent speed is 4.73 m/s, but the speed that carries the most energy is 9.47 m/s, twice as high, because power rises with v³.
Why is the mean of v³ more than the cube of the mean speed?
Power density is ½ρ times the average of v³, not ½ρ times v̄³. Gusty hours count far more than calm ones: one hour at 10 m/s carries as much energy as 8 hours at 5. For a Weibull wind the ratio is the energy pattern factor Γ(1+3/k)/Γ(1+1/k)³: 1.91 for k = 2, 2.72 for k = 1.5, 1.4 for k = 3. At the Chennai hub with air at 1.159 kg/m³, ½ρv̄³ is only 121 W/m² while the true power density is 231 W/m². Never cube the mean speed.
How much faster is the wind higher up?
The ground slows the air, so speed grows with height (wind shear). Two models: the power law v2 = v1 (z2/z1)^α with α ≈ 1/7 for neutral air over open land, and the log law v ∝ ln(z/z0) with a roughness length z0: 0.03 m for open grass, 0.25 m for high crops with scattered obstacles. From 50 m to 110 m the 1/7 law gives 1.119 times the speed, the log law 1.106 (z0 0.03 m) or 1.149 (z0 0.25 m). Since power goes as v³, the 1/7 law's 12 % more speed is 40 % more power. That is why hubs keep rising.
How is wind measured, and how far can you trust NASA's numbers?
A bankable assessment puts cup anemometers and wind vanes on a met mast at two or three heights for at least a year, often with a sonic anemometer (no moving parts) or a LiDAR that measures speed at many heights by laser Doppler; MNRE ran two years of LiDAR off Gujarat and Tamil Nadu. NASA POWER's 50 m wind is a MERRA-2 reanalysis: a model average over a grid cell tens of kilometres across. It gives Muppandal 5.63 m/s and Chennai 5.29, nearly the same, although Muppandal sits in a wind-funnelling gap. Use it for teaching and first looks, not for siting.
Two sites both average 6 m/s. Site A has k = 3 (steady), site B k = 1.5 (gusty). Which has more wind power per square metre?
Power follows the mean of v³. The energy pattern factor is 2.72 for k = 1.5 and 1.4 for k = 3, so site B carries 1.94 times the power of site A at the same mean speed.
Worked examples
Weibull power density from a mean speed · Chennai, NASA mean wind at 50 m 5.29 m/s, take k = 2 (Rayleigh); air at 50 m: 1.166 kg/m³ (mean 27.9 °C)
- c = v̄/Γ(1 + 1/2) = 5.294/0.8862 = 5.973 m/s (= 2v̄/√π)
- Γ(1 + 3/2) = Γ(2.5) = 1.3293
- P/A = ½ × 1.166 × 5.973³ × 1.3293 = 165.2 W/m²
- ½ρv̄³ = ½ × 1.166 × 5.294³ = 86.5 W/m²: the cube of the mean misses a factor 1.91
Answer: About 165 W/m² at 50 m
From 50 m to a 110 m hub · Mean 5.29 m/s at 50 m; compare the 1/7 power law and the log law with z0 = 0.03 m
- Power law: (110/50)^(1/7) = 1.1192, v = 5.925 m/s
- Log law: ln(110/0.03)/ln(50/0.03) = 8.207/7.419 = 1.1063, v = 5.856 m/s
- Power density scales with v³: 1.1192³ = 1.402 (power law)
- A small turbine at 10 m on a crop field (z0 0.25 m) sees ln(10/0.25)/ln(50/0.25) = 0.696 of the 50 m speed, 34 % of its power
Answer: 5.92 m/s by the 1/7 law (5.86 by the log law): 40 % more power at the hub
How many hours does a turbine actually run? · Hub wind Rayleigh with c = 6.685 m/s; cut-in 3 m/s, cut-out 25 m/s, rated 8.91 m/s (the book's turbine at 30 °C)
- P(v ≥ x) = exp(−(x/c)²). Below cut-in: 1 − exp(−(3/6.69)²) = 0.1824, 1,598 h
- Running, 3 to 25 m/s: exp(−(3/c)²) − exp(−(25/c)²) = 0.8176, 7,162 h a year
- At or above rated: exp(−(8.91/6.69)²) = 0.1694, 1,484 h at full power
- Above cut-out: exp(−(25/c)²) ≈ 0.00000084, practically never
Answer: It turns about 7,162 h a year (82 %) but is at full power only 1,484 h
Wind turbine aerodynamics: Betz limit, tip-speed ratio and blade elements
Which syllabus covers thisAU CME365 Unit 3 (Betz)VTU not in syllabusSPPU 402051B Unit 4; 203141 Unit 4MU MEDLO7032 Module 4; EEDO5011 Module 4KTU MET445 Module 3; PEEET412 Module 1 (power numericals)
Why can no turbine capture more than 59.3 % of the wind's power?
To take energy the rotor must slow the air, but air that stopped completely would block the flow behind it. Treat the rotor as a disc that slows the wind by a fraction a: the air passes the disc at v(1 − a) and leaves at v(1 − 2a). Momentum and energy give Cp = 4a(1 − a)². Setting dCp/da = 4(1 − a)(1 − 3a) = 0 gives a = 1/3 and Cp = 16/27 = 0.593 (Betz). At 8 m/s the book's 120 m rotor sees 3.37 MW of wind; Betz allows 2.00 MW; its Cp of 0.48 and 94 % drivetrain deliver 1.52 MW.
What is the tip-speed ratio, and why does Cp depend on it?
λ = ωR/v, the blade tip's speed over the wind speed. Too slow and the blades meet the wind at a steep angle and stall; too fast and the rotor acts like a solid wall that air goes around. The book's turbine fit peaks at Cp 0.48 at λ 8.1. At 8 m/s that is ω = 1.08 rad/s, 10.3 rpm, with tips at 65 m/s, so a 1,500 rpm generator needs a gearbox of about 145:1. Variable-speed turbines speed up with the wind to stay near the best λ; pitching the blades (β) drops Cp on purpose above rated wind.
How does a blade element turn lift into torque, and why is the blade twisted?
A slice of blade at radius r sees the wind v(1 − a) along the axis and its own motion ωr(1 + a′) across it. The relative wind W arrives at the inflow angle φ; lift acts across W and drag along it. Lift's small component in the plane of rotation makes torque; most of it pushes the rotor downwind as thrust. At r/R = 0.75 and λ 8.1: φ = 6.23°, and the thrust part is 10.8 times the torque part. Because φ falls from 17.4° at r/R 0.25 to 4.7° at the tip, the blade is twisted about 13° to keep each slice at its best angle of attack.
Lift or drag, two blades or three: how are rotors chosen?
A drag machine like the Savonius is pushed by the wind, so its cups cannot move faster than the wind; theory caps it at Cp = (4/27) CD at a cup speed of v/3, about 0.18 with a CD of 1.2 (assumed). Lift machines run their blades several times faster than the wind and approach Betz. Solidity, blade area over swept area, sets the speed: many-bladed water-pumping mills have high solidity, run at λ near 1 and start under load; electricity rotors have low solidity and run at λ 6-9. Three blades balance smoothly on a rigid hub; two-bladed rotors need a teetering hub to ride out loads, and spin faster and noisier.
Upwind or downwind, tower, materials, yaw and noise: what else is designed?
Almost all large rotors face upwind, so blades do not pass through the tower's wake each turn. Yaw motors turn the nacelle into the wind; a 20° error cuts power to about cos³20° = 83 % (simple model). Thrust at 8 m/s on the ideal rotor is ½ρAv² × 8/9 = 375 kN, and the aerodynamic torque 1.5 MN·m: the tower and foundation are sized for these plus storms. Towers are tubular steel, concrete or lattice (Suzlon's S144 offers a hybrid lattice tower to 160 m); blades are glass fibre with carbon spar girders on the S144. Noise grows with tip speed (72 m/s at rated here) and falls about 6 dB per doubling of distance.
At the Betz optimum, how fast is the air moving as it passes through the rotor disc?
The optimum slows the air by a = 1/3 at the disc, so it passes at v(1 − a) = 2v/3, and ends up at v(1 − 2a) = v/3 far behind. That gives Cp = 4a(1 − a)² = 16/27.
Worked examples
Wind power, the Betz limit and electrical power at 8 m/s · The book's turbine: D = 120 m, Cp,max = 0.48 (fit), drivetrain η = 0.94; air at 30 °C near sea level, ρ = 1.1644 kg/m³
- A = π × 120²/4 = 11,310 m²
- Wind power = ½ρAv³ = ½ × 1.1644 × 11,310 × 8³ = 3.371 MW
- Betz limit = 16/27 × 3.371 = 1.998 MW
- Electrical = 0.94 × 0.48 × 3.371 = 1.521 MW; rated 2.1 MW is reached at v = (2P/(ηρACp))^(1/3) = 8.91 m/s
Answer: 3.37 MW in the wind, 2.00 MW at Betz, 1.52 MW delivered
Rotor speed, tip speed and gear ratio · Best tip-speed ratio λ = 8.1, R = 60 m, v = 8 m/s; a 4-pole 50 Hz generator turning near 1,500 rpm (assumed)
- ω = λv/R = 8.1 × 8/60 = 1.08 rad/s
- N = 60ω/2π = 10.31 rpm; tip speed = λv = 64.8 m/s (233 km/h)
- Gear ratio ≈ 1,500/10.31 = 145
- At rated wind (8.91 m/s) the tips reach 72.2 m/s
Answer: 10.3 rpm, tips at 65 m/s, gearbox about 145:1
A blade element at three-quarter radius · λ = 8.1, r/R = 0.75, v = 8 m/s, ideal a = 1/3; airfoil at α = 6° with Cl = 2πα and Cd = 0.011 (assumed)
- Local speed ratio λr = 0.75 × 8.1 = 6.075; a′ from a(1 − a) = λr² a′(1 + a′): a′ = 0.00599
- tan φ = (1 − a)/(λr(1 + a′)) = 0.6667/6.111: φ = 6.23°; blade set at θ = φ − α = 0.23°
- W = v √((1 − a)² + (λr(1 + a′))²) = 49.2 m/s, six times the wind
- Cl = 2π × 6π/180 = 0.658; torque coefficient Cl sin φ − Cd cos φ = 0.0604; thrust coefficient Cl cos φ + Cd sin φ = 0.655
Answer: φ = 6.2°, θ = 0.2°; thrust is 10.8 times the torque force at this slice
Power curves, pitch control, annual energy and capacity factor
Which syllabus covers thisAU CME365 Unit 3; EE3014 Unit 1VTU not in syllabusSPPU 402051B Unit 5; 203141 Unit 4MU EEDO5011 Module 4KTU MET445 Module 3
What are the regions of a wind turbine's power curve?
Below cut-in (3 m/s for the book's turbine) the rotor idles. From cut-in to rated, the controller holds the best tip-speed ratio and power rises as v³. At rated wind, 8.91 m/s in 30 °C air (8.76 m/s at 1.225 kg/m³), it reaches 2.1 MW and pitch holds it there. At cut-out (25 m/s, assumed) it feathers and stops to save the structure. Suzlon's real S144 (3.0 MW, 144 m rotor) lists cut-in 3.0, rated 9.2 and cut-out 18 m/s; with the book's Cp and drivetrain it would reach rated at 8.74 m/s, so either the real machine's Cp × η is below the fit's or its sheet assumes thinner air.
Pitch or stall: how does a turbine limit its power in strong wind?
A pitch-regulated turbine turns its blades to cut the angle of attack, shedding lift so power stays flat at rating. Older fixed-speed stall-regulated machines kept the blades fixed and let rising wind stall them: power peaks and then sags. The model's stall rotor, fixed at 8.4 rpm, peaks at 12.3 m/s and makes 15 % less energy than the pitch-regulated variable-speed curve on the same wind (illustrative); a fixed-speed pitch rotor loses 11 %. Stall machines also lose peak power in thin air: at 2,000 m and 30 °C the peak drops to 1.57 MW. Past about 20 m/s the stall curve's upturn comes from the Cp fit, not real aerodynamics.
How is annual energy worked out from a power curve and a wind distribution?
Split the wind into 1 m/s bins. For each bin, hours = 8,760 × (F(v + 0.5) − F(v − 0.5)) from the Weibull CDF; energy = power at the bin centre × hours; add the bins and take off losses. For the book's turbine at a 5.9 m/s Rayleigh hub wind and 30 °C: gross 7.05 GWh, and after 12 % for availability, wakes and electrical losses (assumed), 6.2 GWh a year. The bins carrying most energy are 7-10 m/s, not the most common speeds. A fine integral gives 6.99 GWh gross: 1 m/s bins are close enough.
What is a good capacity factor for wind in India?
CUF = annual energy ÷ (rated power × 8,760 h). The model's 6.2 GWh from 2.1 MW is a CUF of 33.7 %, or 2,953 full-load hours. CERC's 2020 tariff regulations set normative wind CUFs from 22 % (wind power density up to 220 W/m²) to 35 % (above 440 W/m²); the Chennai hub's 231 W/m² falls in the 221-275 band, 24 %. The model is higher partly because a low specific rating (186 W/m² of rotor, like the S144's 184) raises CUF, partly because its Cp fit is generous. The rooftop of chapter 13 runs at 16.9 %.
How is a power curve measured, and why correct it for air density?
IEC 61400-12-1 sets the method: measure the turbine's power against the free wind at hub height for months, average in short intervals, bin by wind speed, and report the curve with its uncertainty. Air density changes power directly, so readings are normalised to a reference density; for a pitch-regulated turbine the wind speed is scaled by (ρ/ρ0)^(1/3). Jodhpur in summer (35 °C, 253 m) has ρ = 1.112 kg/m³: the factor is 0.968, rated wind moves from 8.76 to 9.05 m/s, and power at 7 m/s drops 9.3 %.
The same turbine and the same wind move from a 15 °C site to a 35 °C site, both at sea level. How does annual energy change?
Air is 6.5 % thinner at 35 °C, which cuts power below rated, but above rated the pitch system holds 2.1 MW either way. The bin sum falls from 6.36 to 6.15 GWh, 3.3 %.
Worked examples
Energy from three wind-speed bins · The book's turbine at 30 °C (ρ 1.164 kg/m³); hub wind Rayleigh, mean 5.9 m/s (c = 6.657 m/s); bins of 1 m/s centred at 6, 8 and 10 m/s
- Hours in the 8 m/s bin = 8,760 × (exp(−(7.5/c)²) − exp(−(8.5/c)²)) = 746 h; P(8) = 1,521 kW; energy 1.135 GWh
- 6 m/s bin: 1,050 h × 642 kW = 0.674 GWh
- 10 m/s bin: 415 h × 2,100 kW = 0.872 GWh (above rated)
- All 25 bins: gross 7.047 GWh; less 12 % losses: 6.201 GWh
Answer: The 8 m/s bin gives the most of these three (1.13 GWh); the year totals 6.2 GWh net
Capacity factor and full-load hours · AEP 6.201 GWh from a 2.1 MW turbine; the chapter 13 rooftop makes 4,810 kWh from 3.24 kWp
- CUF = 6,201,053/(2,100 × 8,760) = 33.71 %
- Full-load hours = 6,201,053/2,100 = 2,953 h
- Rooftop CUF = 4,810/(3.24 × 8,760) = 16.95 %
- Same energy as the turbine from rooftops: 6,201,053/1,485 = 4,177 kWp of PV, 1.99 times the turbine's rating
Answer: Wind 33.7 % (2,953 h) against rooftop PV 16.9 %
A hot, high site: density correction · Jodhpur in summer: 35 °C at 253 m. The book's turbine (rated speed at 1.225 kg/m³ is 8.76 m/s)
- p = 101,325 × (1 − 2.25577 × 10⁻⁵ × 253)^5.25588 = 98,322 Pa; ρ = p/(287.05 × 308.15) = 1.1116 kg/m³
- Rated speed ∝ ρ^(−1/3): 8.758 × (1.225/1.1116)^(1/3) = 9.047 m/s
- Power at 7 m/s ∝ ρ: 1,072 kW × 1.1116/1.225 = 973 kW
- IEC-style normalisation of a measured speed: v_n = v × (1.1116/1.225)^(1/3) = 0.9681 v
Answer: ρ = 1.112 kg/m³; rated wind rises to 9.05 m/s and power below rated drops 9.3 %
Wind farms: turbine types, generators, wakes and layout
Which syllabus covers thisAU CME365 Unit 3; EE3014 Units 2, 4VTU BME654B Module 3; BEE654B Module 3SPPU 402051B Units 4, 5; 203141 Unit 4MU MEDLO7032 Module 4; EEDO5011 Module 4KTU MET445 Module 3; PEEET412 Module 1
Horizontal or vertical axis: which turbines are built, and why?
A horizontal-axis turbine (HAWT) points its rotor into the wind and its blades work by lift, like wings; three blades upwind of the tower is the standard machine. A vertical-axis Darrieus rotor also uses lift and accepts wind from any direction, but its blades see changing angles every turn and it sits low, where the wind is slow. A Savonius rotor is pushed by drag, like a cup anemometer, so its power coefficient stays far below the Betz limit. Size wins: Suzlon's S144 sweeps 16,286 m², 1.44 times the book's 120 m rotor. India had 59.20 GW of wind by October 2026.
Fixed-speed SCIG, DFIG or PMSG: what does the generator change?
A squirrel-cage induction generator wired straight to the grid turns at almost one speed (about 1 % slip), so the tip-speed ratio is right at only one wind speed. A doubly-fed induction generator (DFIG) lets the rotor speed swing about ±30 % around synchronous by feeding its rotor through a converter. A permanent-magnet synchronous generator (PMSG) puts all its power through a full converter, so any speed works. On the book's turbine at 6 m/s the model gives 589 kW fixed-speed (rotor set for 7 m/s, assumed), 642 kW DFIG and 642 kW PMSG; at 4.5 m/s the DFIG has hit its lower speed limit: 267 against 271 kW.
What does the AC-DC-AC converter chain do in a wind turbine?
The generator makes AC whose frequency follows the rotor speed; the grid wants 50 Hz. So a machine-side converter rectifies it to DC, a capacitor holds a steady DC link, and a grid-side PWM inverter switches IGBTs thousands of times a second to rebuild 50 Hz sine current, controlling active and reactive power. A transformer then lifts it to the collection voltage. In a PMSG the full 2.1 MW goes through this chain. In a DFIG only the rotor's slip power does, so ABB rates its DFIG converters at up to 30 %: about 0.63 MVA for the book's turbine. A matrix converter does AC to AC directly, without the DC link.
How much energy do turbines steal from each other in a farm?
Behind each rotor is a wake of slower air that spreads and recovers. The Jensen model treats it as a cone: the speed deficit starts at 1 − √(1 − CT) and shrinks as (D/(D + 2kx))², with k = 0.075 on land (WAsP). For the interactive's default farm, 5 rows × 3 of the book's turbines at 7 D along and 4 D across with 8 m/s along the rows, the second row gets 7.32 m/s and the last 7.24 m/s; the farm makes 18.3 MW, 80 % of what 15 free turbines would. At 10 D spacing it is 86.3 %; turning the wind 30° to the rows gives 88.9 %.
Why go offshore, and what is India doing about it?
Sea winds are steadier and stronger and there is room for very large rotors, but foundations, cables and ships cost much more. Over the smooth sea the air mixes less, so wakes recover slowly: WAsP suggests k = 0.04 offshore. Put the same default farm at sea and the model's efficiency falls from 80 % to 66.6 %. India set targets of 30 GW of offshore wind by 2030 (MNRE, June 2018) and in June 2024 approved Rs 7,453 crore of viability-gap funding for the first 1 GW off Gujarat and Tamil Nadu. On land, the 2023 repowering policy targets turbines below 2 MW; NIWE puts that potential at 25.4 GW.
Two of the book's turbines stand 7 rotor diameters apart, in line with an 8 m/s wind. What wind speed reaches the second one?
The first rotor runs at CT = 0.58, so the deficit starts at 35 %, but by 7 D the wake has widened to 2.05 D and the deficit has fallen to 8.4 %. Only 8.4 % of speed, yet 23 % of power, because power goes as v³.
Worked examples
One wake at 7 D · The book's turbine (2.1 MW, D 120 m, Cp 0.48 below rated, η 0.94) in 8 m/s at 30 °C (ρ = 1.164 kg/m³). A second turbine stands 7 D downwind. Jensen model, k = 0.075
- Cp = 4a(1 − a)² = 0.48 gives a = 0.1773; CT = 4a(1 − a) = 0.583
- Initial deficit 1 − √(1 − 0.583) = 0.3546; wake width factor 1 + 2 × 0.075 × 7 = 2.05
- Deficit at 7 D = 0.3546/2.05² = 0.0844, so v = 8 × (1 − 0.0844) = 7.325 m/s
- Power: first turbine 1,521 kW; second 0.94 × ½ × 1.164 × 11,310 × 7.325³ × 0.48 = 1,168 kW
Answer: 7.32 m/s at the second turbine, 23.2 % less power
How far apart for 95 % of the wind? · Same turbine at 8 m/s (CT = 0.583). Find the distance at which a single wake leaves 95 % of the free wind speed, on land (k = 0.075) and at sea (k = 0.04)
- Need (1 − √(1 − CT))/(1 + 2kx/D)² = 0.05, so 1 + 2kx/D = √(0.3546/0.05) = 2.663
- Land: x/D = (2.663 − 1)/(2 × 0.075) = 11.09, that is 1,330 m
- Sea: x/D = (2.663 − 1)/(2 × 0.04) = 20.79, that is 2,495 m
- Even at 95 % of the speed a turbine loses about 14.3 % of power (0.95³), below rated
Answer: About 11.1 D (1.33 km) on land and 20.8 D at sea
Gearbox and converter of a DFIG · The book's turbine as a 4-pole, 50 Hz DFIG. Rated rotor speed: λ* = 8.1 at rated wind 8.91 m/s. Speed range ±30 % of synchronous, rated speed at +30 %
- Rotor speed at rated = λ* v/R = 8.1 × 8.91/60 = 1.203 rad/s = 11.48 rpm
- Synchronous speed = 120 f/p = 120 × 50/4 = 1,500 rpm; rated generator speed = 1.3 × 1,500 = 1,950 rpm
- Gear ratio = 1,950/11.48 = 169.8; lowest rotor speed = 11.48 × 0.7/1.3 = 6.18 rpm, reached at 4.8 m/s
- Converter ≈ slip × power = 0.3 × 2.1 = 0.63 MVA, against 2.1 MVA for a full converter
Answer: Gearbox about 1:170, converter about 0.63 MVA, variable speed above 4.8 m/s
Biomass: resources, calorific value and conversion routes
Which syllabus covers thisAU CME365 Unit 4VTU BME654B Module 5; BEE654B Module 4SPPU 402051B Unit 6; 203141 Unit 6MU MEDLO7032 Module 5; EEDO5011 Module 6KTU MET445 Module 5; PEEET412 Module 3
How much biomass energy does India have, and where is it?
Biomass is recent plant matter: crop residues (rice husk and straw, cotton stalks), sugar-mill bagasse, wood, animal dung and the organic part of city waste. MNRE puts India's biomass power potential at 42,312 MW: 28,446 MW from about 228 million tonnes a year of surplus agro-residue and 13,866 MW from bagasse cogeneration. Installed bio-power was 11.75 GW in October 2026, mostly in sugar mills: 9,821 MW bagasse cogeneration and 1,048 MW of other biomass. An energy plantation grows fast trees such as Casuarina or subabul for fuel on poor land; dry Casuarina wood holds 18.77 MJ/kg.
What do proximate and ultimate analyses tell you about a fuel?
A proximate analysis heats a sample in stages and weighs it: moisture, volatile matter (gas driven off at about 900 °C without air), fixed carbon (the char that burns later) and ash. An ultimate analysis gives the elements: C, H, O, N, S. Rice husk is 21.2 % ash and 38.5 % carbon on a dry basis, coconut shell 0.71 % ash and 50.2 % carbon, so the shell carries 20.5 MJ/kg against 14.7. The Channiwala-Parikh correlation predicts HHV from the elements; across the six fuels in the interactive its average miss is 3.4 %.
Why does wet biomass give so much less heat?
The higher heating value (HHV) counts the heat released when the water in the flue gas condenses; a boiler stack throws that water away as steam, so engineers use the lower heating value (LHV). Moisture hurts twice: a kilogram of wet fuel holds less fuel, and its water must be boiled off at 2.443 MJ/kg. Dry bagasse has an HHV of 18.73 MJ/kg. At 20 % moisture its LHV is 13.45 MJ/kg; straight from the mill at about 50 % it is only 7.49 MJ/kg, which is why mills dry and store it.
What are the routes from biomass to useful energy, and what does a TGA show?
Thermochemical routes use heat: direct combustion (boilers, cogeneration), gasification to fuel gas, pyrolysis to oil, gas and char, and carbonisation, slow heating without air to make charcoal. Biochemical routes use microbes: anaerobic digestion to biogas and fermentation of sugar to ethanol. Pressing oilseeds and transesterifying the oil gives biodiesel. Thermogravimetric analysis (TGA) weighs a sample as it is heated steadily. For ten Indian residues, Jha and Dass found drying up to 100 °C, torrefaction between 200 and 300 °C, the fastest loss between 300 and 500 °C, and an almost constant mass beyond 600 °C.
Cogeneration, co-firing and waste-to-energy: where is biomass used at scale?
A sugar mill burns its bagasse in high-pressure boilers, runs a back-pressure or extraction turbine for electricity and uses the exhaust steam to boil cane juice: that is cogeneration, heat and power from one fuel. Coal plants can co-fire biomass pellets with coal; the Ministry of Power's 2025 policy requires 5 % biomass pellets in every coal plant from 2025-26, plus 2 % torrefied charcoal made from municipal solid waste in the NCR. City waste can also be burnt or gasified, or its wet part digested: India had 324.24 MW of grid waste-to-energy and 554.16 MW off grid in September 2026.
Dry bagasse has an HHV of 18.73 MJ/kg. Fresh from the mill it is about half water. What is its LHV as fired?
Half the mass carries 18.73 × 0.5 = 9.37 MJ, then the steam costs 2.443 × (8.94 × 0.0596 × 0.5 + 0.5) = 1.87 MJ: the fuel's own moisture plus the water formed from its hydrogen.
Worked examples
HHV of coconut shell from its ultimate analysis · Coconut shell, dry basis: C 50.22 %, H 5.7 %, O 43.37 %, N 0, S 0, ash 0.71 %. Bomb calorimeter: 20.5 MJ/kg
- 0.3491 × 50.22 = 17.532; 1.1783 × 5.7 = 6.716
- −0.1034 × 43.37 = −4.484; −0.0211 × 0.71 = −0.015
- HHV = 19.75 MJ/kg against 20.5 measured: −3.7 %
- The correlation's stated average error is 1.45 % over its data set; single fuels can miss by more
Answer: 19.75 MJ/kg predicted, 3.7 % below the measured value
Rice husk at 12 % moisture: LHV and air · Rice husk, dry basis: C 38.5 %, H 5.2 %, O 34.61 %, N 0.45 %, ash 21.24 %, HHV 14.69 MJ/kg; fired at 12 % moisture
- HHV as fired = 14.69 × 0.88 = 12.927 MJ/kg
- Water out per kg = 8.936 × 0.052 × 0.88 + 0.12 = 0.5289 kg; × 2.4426 = 1.292 MJ
- LHV = 12.927 − 1.292 = 11.64 MJ/kg
- O2 per kg dry = 2.664 × 0.385 + 7.937 × 0.052 − 0.3461 = 1.0922 kg; air = 1.0922/0.233 = 4.688 kg, × 0.88 = 4.125 kg per kg as fired
Answer: LHV 11.64 MJ/kg; 4.13 kg of air per kg of husk; 18.7 % ash in the fuel as fired
Fuel for a 10 MW rice-husk power plant · The husk of the previous example (LHV 11.64 MJ/kg). Plant efficiency 25 % (assumed), 10 MW, running at an 80 % capacity factor (assumed)
- Electricity per tonne = 11.635 × 1,000/3.6 × 0.25 = 808 kWh
- Full load: 10 MW × 24 h = 240,000 kWh a day → 297 t of husk a day
- A year at 80 %: 297 × 365 × 0.8 = 86,732 t
- Ash to dispose: 86,732 × 0.1869 = 16,211 t a year
Answer: About 297 t of husk a day and 16,211 t of ash a year
Biogas: anaerobic digestion and sizing a family digester
Which syllabus covers thisAU CME365 Unit 4VTU BME654B Module 5; BEE654B Module 4SPPU 402051B Unit 6MU MEDLO7032 Module 5KTU MET445 Module 5; PEEET412 Module 3
What happens inside a digester?
Without air, four groups of microbes work in a chain. Hydrolysis breaks cellulose, proteins and fats into sugars, amino acids and fatty acids; acidogenesis turns them into volatile acids; acetogenesis makes acetic acid, hydrogen and CO2; methanogens finish the job (CH3COOH → CH4 + CO2, CO2 + 4H2 → CH4 + 2H2O). The overall sum for glucose is C6H12O6 → 3CH4 + 3CO2: half methane by volume, 0.373 Nm³ of methane per kg of glucose. The slow methanogens set the pace, which is why the slurry must stay in the digester for weeks.
What conditions does a digester need?
MNRE gives the working ranges: temperature 35 to 40 °C, pH 6.5 to 7.5, a carbon-to-nitrogen ratio near 30:1 and 8 to 10 % total solids, which is why cow dung is mixed 1:1 with water. Methanogens slow down sharply in the cold, so the slurry must stay longer: MNRE's hydraulic retention time is 30 days for tropical India and 40 days for sub-tropical regions. Too much feed lowers the pH (acids pile up faster than methanogens use them) and the plant 'turns sour'. Poultry droppings give 0.07 m³/kg against 0.04 for cow dung.
How do you size a family biogas plant?
Start from the dung. Five cattle at 10 kg a day give 50 kg of dung; mixed 1:1 with water that is 100 litres of slurry a day. At a 30-day retention time the digester holds 100 × 30 = 3 m³. At MNRE's 0.04 m³ per kg the plant makes 2 m³ of gas a day, MNRE's size for a family of five, enough for 4.4 hours on a 0.45 m³/h stove. The holder must store what is made between meals: 0.81 m³ for two cooking sessions a day.
KVIC floating drum or Deenbandhu fixed dome: what is the difference?
The KVIC plant has a brick digester with a steel drum floating on the slurry; as gas collects the drum rises, so pressure stays low and steady (up to about 20 mbar) and you can see how much gas is left. The steel rusts: the drum lasts up to 15 years, about 5 on a tropical coast. The Deenbandhu plant is an all-masonry hemispherical dome; stored gas pushes slurry up into the outlet chamber, so pressure varies (60 to 120 mbar) but there are no moving parts and the dome can last 20 years or more. Family models run from 1 to 10 m³ a day.
What is biogas worth as a fuel, and what is compressed biogas?
Biogas is 55 to 65 % methane, the rest mostly CO2 with a little H2S and moisture. Its calorific value is about 20 MJ/m³ (4,777 kcal/m³); MNRE equates 1 m³ with 0.42 kg of LPG. Remove the CO2 and H2S, compress it to the pressure of CNG and it becomes compressed biogas (CBG) for vehicles and gas grids. The SATAT scheme aimed at 5,000 CBG plants making 15 million tonnes by 2023-24, which was not reached; by November 2026 839 plants were registered and 210 commissioned, and GOBARdhan has Rs 23,731 crore for 2026-36.
A farming family keeps 5 cattle and feeds all their dung to a digester. How much biogas does it give a day?
5 × 10 kg = 50 kg of dung × 0.04 m³/kg = 2 m³: MNRE's plant for a family of five, about 4.4 hours of cooking.
Worked examples
A Deenbandhu plant for a colder district · 8 cattle, 10 kg of dung each per day, sub-tropical retention time 40 days (MNRE), 0.04 m³ of gas per kg, digester modelled as a hemisphere
- Dung = 80 kg/day; slurry = 80 + 80 = 160 kg ≈ 160 L = 0.16 m³/day
- Digester = 0.16 × 40 = 6.4 m³; hemisphere radius = (3 × 6.4/(2π))^(1/3) = 1.45 m
- Gas = 80 × 0.04 = 3.2 m³/day; on a 0.45 m³/h stove that is 7.1 h of cooking
- At 0.24 m³ per person per day (IIT Delhi) it cooks for about 13 people
Answer: A 6.4 m³ digester (1.45 m dome radius) giving 3.2 m³ of gas a day
How many LPG cylinders does a 2 m³ plant replace? · 2 m³ of biogas a day at 20 MJ/m³; LPG about 46 MJ/kg in 14.2 kg domestic cylinders; MNRE's equivalence 0.42 kg LPG per m³
- Energy = 2 × 20 = 40 MJ/day = 0.87 kg of LPG a day
- A year: 0.87 × 365 = 317.4 kg = 22.4 cylinders
- MNRE's equivalence: 2 × 0.42 × 365 = 306.6 kg = 21.6 cylinders
- The two agree within 4 %; stove efficiencies differ, so treat it as an energy-basis estimate
Answer: About 22 cylinders a year
Gas holder for two meals · The 2 m³/day plant; the day's gas is burnt in two equal sessions centred at 7:00 and 19:00 on a 0.45 m³/h stove. KVIC drum as wide as a digester of H = 2D (model)
- Cooking time = 2/0.45 = 4.44 h, so each session lasts τ = 2.22 h
- Between sessions the plant makes gas for 12 − τ = 9.78 h at 2/24 = 0.0833 m³/h: store = 0.815 m³
- Digester 3 m³ with H = 2D: D = (2 × 3/π)^(1/3) = 1.241 m
- Drum rise = 0.815/(π × 1.241²/4) = 0.67 m
Answer: Store 0.81 m³, 41 % of a day's gas; the drum rises about 0.67 m
Gasification and liquid biofuels: producer gas, ethanol and biodiesel
Which syllabus covers thisAU CME365 Unit 4VTU BME654B Module 5; BEE654B Module 4SPPU 402051B Unit 6MU not in syllabusKTU MET445 Module 5; PEEET412 Module 3
What happens in the four zones of a gasifier?
A gasifier burns biomass with too little air for complete combustion (the equivalence ratio is the fraction supplied), so the product is a fuel gas. In a downdraft unit fuel moves down through drying, then pyrolysis (volatiles and tar driven off, char left), then oxidation at the air nozzles, where some char and volatiles burn (C + O2 → CO2 releases 393.5 kJ/mol), then reduction, where hot char takes oxygen from CO2 and steam: C + CO2 → 2CO absorbs 172.5 kJ/mol and C + H2O → CO + H2 absorbs 131.3. Updraft units feed air at the bottom, cross-draft from the side; fluidised beds suit fine fuels such as husk.
How does an equilibrium model predict producer gas?
It assumes the reduction zone reaches chemical equilibrium: all carbon gasifies, and the water-gas shift and methanation reactions settle at constants set by temperature (from NIST-JANAF Gibbs energies), while an energy balance fixes that temperature. For the interactive's default, Casuarina wood at 15 % moisture and an equivalence ratio of 0.35, it gives 672 °C, 22.5 % H2, 18.5 % CO, 0.8 % CH4, 13.1 % CO2 and 45 % N2: 5.07 MJ/Nm³ and 83 % cold gas efficiency. At 0.25 the Boudouard check says char would be left, so the model fails there. Real gasifiers lose wall heat and leave char and tar: read the model as an upper bound.
How is producer gas used, and why is tar the problem?
Producer gas is lean, about 5 MJ/Nm³ against 36 for methane, because nearly half of it is nitrogen from the air. It can fire a boiler or kiln directly, or run a gas engine or a diesel engine in dual-fuel mode, after cooling and cleaning in cyclones, scrubbers and filters. Tar, the heavy vapour from pyrolysis, condenses when the gas cools and gums up valves; downdraft gasifiers crack much of it by passing the vapour through the hot oxidation zone, which is why they are chosen for engines. From the model's 1 kg of wood (2.37 Nm³ of gas), an engine at 25 % (assumed) makes about 0.83 kWh.
How is ethanol made, and what is E20?
Yeast ferments sugars to ethanol and CO2: sucrose plus water gives four ethanol and four CO2 molecules, so a tonne of sugar can give at most 538 kg of ethanol; at 90 % of that (assumed) it is 614 litres. Molasses, cane juice and grain are the Indian feedstocks; second-generation plants first break crop residue into sugars. Ethanol carries 26.8 MJ/kg (LHV). Under the Ethanol Blended Petrol programme India went from under 1.5 % ethanol in petrol in 2013-14 to 20 % (E20) in 2025-26, five years early, buying over 1,200 crore litres (projected).
What is biodiesel, and what are a biorefinery and IGCC?
Vegetable oils and fats are triglycerides, too viscous for diesel engines. Transesterification reacts them with methanol over a catalyst such as NaOH: one triglyceride and three methanol give three fatty-acid methyl esters (FAME, the biodiesel) and one glycerol. For triolein, a tonne of oil and 109 kg of methanol give 1,005 kg of FAME and 104 kg of glycerol; methyl oleate carries 37.2 MJ/kg. A biorefinery makes fuels, power and chemicals from one biomass stream; IGCC (integrated gasification combined cycle) burns cleaned gasifier gas in a gas turbine with a steam cycle on its exhaust.
A downdraft gasifier runs on wood at an equivalence ratio of 0.25. You open the air valve to 0.45. What happens to the gas?
In the model the reduction zone goes from 534 to 933 °C, the char that 0.25 leaves behind is all gasified, but the LHV falls from 6.5 to 4.1 MJ/Nm³ as more of the fuel burns and nitrogen dilutes the gas.
Worked examples
Heats of the gasification reactions · Standard enthalpies of formation at 298 K (NIST-JANAF): CO −110.527, CO2 −393.522, H2O (gas) −241.826, CH4 −74.873 kJ/mol
- Boudouard C + CO2 → 2CO: 2(−110.527) − (−393.522) = +172.468 kJ/mol (absorbs heat)
- Water-gas C + H2O → CO + H2: −110.527 − (−241.826) = +131.299 kJ/mol
- Shift CO + H2O → CO2 + H2: −393.522 + 110.527 + 241.826 = −41.169 kJ/mol (releases heat)
- Shift constant: K = 1 at about 1097 K; at 1,000 K, K = 1.438, so hotter gas keeps more CO and steam
Answer: +172.5 and +131.3 kJ/mol for the reduction reactions; the oxidation zone must supply that heat
LHV and cold gas efficiency of a producer gas · Dry gas: 18 % H2, 20 % CO, 2 % CH4, 10 % CO2, 50 % N2; 2.2 Nm³ of dry gas per kg of Casuarina wood at 15 % moisture (LHV 14.43 MJ/kg)
- Component LHVs from JANAF: H2 10.79, CO 12.63, CH4 35.79 MJ/Nm³
- Gas LHV = 0.18 × 10.79 + 0.20 × 12.63 + 0.02 × 35.79 = 5.183 MJ/Nm³
- Cold gas efficiency = 5.183 × 2.2/14.43 = 79 %
- Engine at 25 % (assumed): 11.4 × 0.25/3.6 = 0.79 kWh per kg of wood
Answer: 5.18 MJ/Nm³, 79 % cold gas efficiency, about 0.79 kWh per kg
Mass balance of a biodiesel batch · 1,000 kg of oil taken as pure triolein (C57H104O6), methanol at 6:1 molar (twice stoichiometric, assumed), full conversion
- M(triolein) = 57 × 12.011 + 104 × 1.008 + 6 × 15.999 = 885.45 g/mol: 1.1294 kmol of oil
- Methanol needed = 3 × 1.1294 × 32.042 = 108.6 kg; fed 217.1 kg, 108.6 kg recovered
- FAME = 3 × 1.1294 × 296.49 = 1,004.6 kg; glycerol = 1.1294 × 92.09 = 104 kg
- Check: 1,000 + 108.6 = 1,108.6 kg; energy in the FAME = 37.4 GJ
Answer: 1,005 kg of biodiesel and 104 kg of glycerol per tonne of oil
Small hydro and pumped storage: flow duration, head and energy
Which syllabus covers thisAU CME365 Unit 5VTU BME515D Module 4SPPU 203141 Unit 3MU EEDO5011 Module 6 (micro-hydro)KTU not in syllabus
What counts as small hydro in India, and how much of it is built?
MNRE looks after hydro plants up to 25 MW and splits them into micro (up to 100 kW), mini (100 kW to 2 MW) and small (2 to 25 MW). Bigger plants are large hydro under the Ministry of Power. MNRE puts the potential at 21,134 MW from 7,133 sites. By 31 December 2025, 5,159 MW was running from 1,196 projects: 24.4 % of the potential, at an average of 4.3 MW a project. CEA counted 5,189 MW on 31 October 2026. A scheme approved in March 2026 sets aside Rs 2,584.60 crore for about 1,500 MW of 1-25 MW projects from 2026-27 to 2030-31.
How much power does falling water give, and where does the head go?
Power = ρ g Q H η. One cubic metre per second falling 100 m carries 981 kW of hydraulic power. The gross head is the level difference between the intake and the tailrace. The turbine sees the net head: gross head minus friction in the penstock, h_f = f (L/D) v²/2g. In the interactive's default plant, 0.97 m³/s flows through a 0.7 m steel pipe 240 m long at 2.52 m/s and loses 1.36 m, 1.7 % of 80 m. Halve the diameter to 0.35 m and the loss jumps to 58 % of the head: loss rises roughly as 1/D⁵.
What is a flow duration curve, and how do you pick the design flow?
A hydrograph plots a stream's flow against the date. Sort a year of it from the largest flow to the smallest and you get the flow duration curve: the flow exceeded for each % of the time. The book's illustrative hill stream has a median of 0.6 m³/s and a mean of 1.1 m³/s, pulled up by monsoon floods. Keep the flow exceeded 95 % of the time (0.098 m³/s) in the river and size the turbine on the flow exceeded 30 % of the time (1.07 m³/s) less that: 0.97 m³/s. With 80 m of head that is a 625 kW Francis plant giving 2,845 MWh a year, a capacity factor of 52 %.
Pelton, Francis or crossflow: what happens when the stream runs low?
For most of the year a run-of-river turbine gets less than its design flow, so the efficiency curve at part flow decides the energy. RETScreen's curves for the default plant give, at 30 % of design flow: Pelton 87 %, crossflow 68 %, Francis only 58 %. At full flow they read 88, 79 and 88 %. Over the year the same site gives 2,993 MWh with a Pelton, 2,845 with a Francis and 2,585 with a crossflow. Two half-size units keep each one nearer full flow. Which type suits which head is the Fluid Mechanics book's specific-speed interactive.
How does pumped storage store energy, and what does it lose?
It uses two reservoirs at different heights and a reversible pump-turbine. When power is cheap (midday solar), it pumps water up; at the evening peak the water runs back down through the same machine. Stored energy = ρ g V H: 5 million m³ lifted 300 m holds 4,086 MWh. With 88 % pumping and 90 % generating efficiency, 4,643 MWh from the grid comes back as 3,677 MWh, a round trip of 79.2 %. Delivered over 6 hours that is 613 MW. Doubling the head doubles the energy from the same volume, so hilly sites with a big drop win.
Is small hydro free of environmental impact?
No. A run-of-river plant has no big reservoir, but its weir diverts water into the penstock, so the stretch of river between the intake and the powerhouse runs low. In the default plant that stretch is 240 m long, and for 70 % of the year, whenever the stream carries less than design flow plus residual, it gets only the residual 0.098 m³/s: about 9 % of the mean flow. Fish need passes past the weir. Roads, forest cleared for the penstock and silt flushed from the pond also matter in fragile Himalayan valleys.
In the interactive, you move the design flow from the flow exceeded 30 % of the time to the flow exceeded 10 % of the time. What happens?
The rating rises 125 % (to 1,406 kW) but the extra capacity only gets used in the few weeks of high flow, so energy rises just 24 % and the capacity factor drops from 52 % to 29 %.
Worked examples
Power from a micro-hydro site · A hill farm can divert 0.12 m³/s with a 45 m gross head. Welded steel penstock 300 mm, 135 m long (roughness 0.045 mm, water ν = 1.0035e-6 m²/s). Crossflow turbine 79 % at design flow (RETScreen), generator 95 %
- Area = π × 0.3²/4 = 0.0707 m²; v = 0.12/0.0707 = 1.698 m/s; Re = 1.698 × 0.3/1.0035e-6 = 507,519
- Colebrook: 1/√f = −2 log10(ε/3.7D + 2.51/(Re √f)) solved by iteration gives f = 0.0149
- h_f = f (L/D) v²/2g = 0.0149 × (135/0.3) × 1.698²/(2 × 9.807) = 0.99 m; net head = 45 − 0.99 = 44.01 m
- P = ρ g Q H η = 1,000 × 9.807 × 0.12 × 44.01 × 0.79 × 0.95 = 38,871 W
- Hydraulic power at the gross head = 53 kW, so the plant turns 73 % of it into electricity
Answer: About 38.9 kW: a micro hydro plant (up to 100 kW)
Annual energy from a flow duration table · A stream's year sorted into five equal slices: flow exceeded 0-20 % of the time 2.0 m³/s, 20-40 % 1.2, 40-60 % 0.7, 60-80 % 0.4, 80-100 % 0.15. Design flow 0.7 m³/s, net head 60 m, overall efficiency 80 % (constant), turbine stops below 30 % of design flow
- Turbine flow per slice: 0.7, 0.7, 0.7, 0.4 and 0 m³/s (0.15 is below the 0.21 m³/s minimum)
- Power at 0.7 m³/s = 1,000 × 9.807 × 0.7 × 60 × 0.8 = 329.5 kW; at 0.4 m³/s = 188.3 kW
- Each slice is 0.2 × 8,760 = 1,752 h: E = 1,752 × (3 × 329.5 + 188.3) = 2,061,750 kWh
- Capacity factor = 2,061,750/(329.5 × 8,760) = 71.4 %
Answer: 2,062 MWh a year from a 330 kW plant, capacity factor 71 %
Pumped storage for evening power · Upper reservoir 4 million m³, average head 320 m, pumping efficiency 88 %, generating efficiency 90 %. Pumping power bought at Rs 2.55/kWh (the L1 tariff of NTPC's 1,200 MW solar auction)
- Stored = ρ g V H = 1,000 × 9.807 × 4 × 10⁶ × 320 = 12.55 × 10¹² J = 3,487 MWh
- Energy to pump = 3,487/0.88 = 3,962 MWh; energy back = 3,487 × 0.90 = 3,138 MWh
- Round trip = 0.88 × 0.90 = 79.2 %; over a 5 h evening peak = 628 MW
- Energy cost per kWh delivered = 2.55/0.792 = Rs 3.22, before the plant's own capital cost
Answer: 3,138 MWh per cycle at 79.2 % round trip; the stored solar power costs Rs 3.22/kWh in energy alone
Ocean energy: tidal barrages, waves and OTEC
Which syllabus covers thisAU CME365 Unit 5VTU BME654B Module 4; BEE654B Modules 4-5SPPU 203141 Unit 6; 402049 Unit 6MU MEDLO7032 Module 6; EEDO5011 Module 6KTU MET445 Module 4; PEEET412 Module 3
How does a tidal barrage make power, and how much can one tide give?
A dam (barrage) closes off a bay. On the flood, sluices let the basin fill; at high water the gates shut. As the sea falls, the head between basin and sea grows, and water runs out through low-head turbines: ebb generation. A two-way scheme also generates on the flood, and a double-basin scheme smooths the output. The most one emptying can give is ½ρgAR²: a 30 km² basin with an 8 m range holds 2,680 MWh. The interactive's model, waiting for 3 m of head, gets about 1,420 MWh (53 %), which over 706 tides a year is about 1,000 GWh.
Where is India's ocean energy, and why is so little built?
MNRE (2019) puts the tidal potential at about 12,455 MW in the Gulfs of Khambhat and Kutch and large backwaters, wave at about 40,000 MW and OTEC at 180,000 MW, all theoretical. A barrage needs at least 5 m of tidal range. IIT Kharagpur's P. K. Bhaskaran gives about 11-12 m in the Gulf of Khambhat, about 8 m in the Gulf of Kutch and 6-7 m in the Sundarbans. Cost stopped two projects: 3.75 MW at Durgaduani for Rs 238 crore (Rs 63 crore per MW) and 50 MW in the Gulf of Kutch for Rs 750 crore (Rs 15 crore per MW). Ocean energy counts towards non-solar renewable purchase obligations.
How much power is in a wave, and how good is India's coast?
Waves carry energy along their crest. In deep water the power per metre of crest is ρg²Hs²Te/64π, about 0.49 Hs² Te kW/m: 2 m waves at a 9 s period carry 17.6 kW per metre, with a wavelength gT²/2π = 126 m. Power grows with the square of height, so storms dominate. CSIR-NIO's buoys off Goa measured a mean of 6 kW/m a year at 9 m depth, from 1.6 kW/m after the monsoon to 15.2 kW/m during it: a 9.5 times seasonal swing. Devices: oscillating water columns (air pushed through a turbine), point absorbers (bobbing buoys) and attenuators (hinged snakes).
How does OTEC turn a 20 °C temperature gap into power?
Ocean thermal energy conversion runs a heat engine between warm surface water and cold water pumped from about 1,000 m. With 29 °C and 7 °C (the Kavaratti figures) the Carnot limit is only 7.3 %. After heat-exchanger temperature drops, turbine losses and the big pumps, the interactive's model nets 2.3 %, and needs 3.5 m³/s of cold water per MW. Closed cycles (Anderson) boil ammonia; open cycles (Claude) flash warm seawater in a vacuum, and the condensed steam is fresh water. NIOT's Kavaratti plant, under construction, pairs 65 kW with 100,000 litres of drinking water a day.
What does ocean energy do to the sea around it?
A barrage changes the tide inside the basin. In the interactive's 8 m case the basin swings only 4.7 m instead of 8 m, a 42 % cut, so mudflats that used to dry out at low tide stay under water, and birds that feed on them lose ground. Silt that the tide carried out settles behind the barrage, and fish must pass the turbines or the sluices. OTEC brings up 3.5 m³/s of cold, nutrient-rich deep water per MW; where it is released changes the local plankton and temperature. Wave farms are gentler but take sea space from fishing.
A barrage site's tidal range is 4 m. Another, with the same basin area, has 8 m. How does the ideal energy per tide compare?
E = ½ρgAR²: doubling the range doubles both the volume of water and the average head it falls through. A 30 km² basin gives 670 MWh at 4 m and 2,680 MWh at 8 m.
Worked examples
Energy per tide in a Gulf of Kutch-sized range · A 20 km² basin behind a barrage, tidal range 8 m (about the Gulf of Kutch's), sea water 1,025 kg/m³, semidiurnal tide of 12.42 h
- Ideal energy per emptying = ½ ρ g A R² = 0.5 × 1,025 × 9.807 × 20 × 10⁶ × 8² = 6.433 × 10¹² J
- In MWh: 6.433 × 10¹² / 3.6 × 10⁹ = 1,787 MWh per tide
- Averaged over the 12.42 h tide = 1,787/12.42 = 144 MW; tides a year = 8,766/12.42 = 705.8
- Ideal per year = 1,787 × 705.8 = 1,261 GWh. The interactive's ebb model (start at 3 m head, 85 % turbines) gets 947 MWh per tide, 53 % of the ideal
Answer: 1,787 MWh per tide ideal, about 669 GWh a year from the model's ebb-only scheme
Wave power offshore and close to the coast · Significant wave height 2.5 m, energy period 9 s. Compare deep water with 9 m depth (linear theory, same wave height), and a device catching 30 % of the power along 100 m of crest (assumed)
- Deep water: P = ρ g² Hs² Te/64π = 1,025 × 9.807² × 2.5² × 9/(64π) = 27,578 W/m = 27.6 kW/m
- Deep-water wavelength = g T²/2π = 9.807 × 81/(2π) = 126.4 m
- At 9 m depth, solving ω² = g k tanh(kh) gives L = 78.2 m and group speed 7.48 m/s, so P = ρ g Hs² c_g/16 = 29.4 kW/m
- Device: 0.3 × 100 m × 27.6 kW/m = 827 kW while these waves last; Goa's yearly mean of 6 kW/m would give 180 kW
Answer: 27.6 kW/m offshore; a 100 m device averages far less than its storm rating
OTEC water flows for Kavaratti's 65 kW · Surface 29 °C, deep 7 °C. Model: each stream changes 3 °C, 2 °C pinch, cycle at 0.8 of Carnot between the ammonia temperatures, pumps take 30 % of gross, sea water cp 3.99 kJ/kg K
- Carnot limit = 1 − 280.15/302.15 = 7.28 %
- Ammonia evaporates at 29 − 3 − 2 = 24 °C, condenses at 7 + 3 + 2 = 12 °C: 1 − 285.15/297.15 = 4.04 %; gross = 0.8 × that = 3.23 %
- For 65 kW net, gross = 65/0.7 = 92.9 kW; heat taken from warm water = 92.9/0.0323 = 2,874 kW
- Warm water = 2,874/(3.99 × 3) = 240 kg/s; cold water = (2,874 − 92.9)/(3.99 × 3) = 232 kg/s
Answer: Net 2.3 % against a 7.3 % Carnot limit: about 227 litres a second of 1,000 m water for 65 kW
Geothermal energy: resources and flash and binary power plants
Which syllabus covers thisAU CME365 Unit 5VTU BME654B Module 3; BEE654B Module 3SPPU 203141 Unit 6; 402049 Unit 6MU MEDLO7032 Module 6; EEDO5011 Module 6KTU MET445 Module 4; PEEET412 Module 3
What kinds of geothermal resource are there?
Hydrothermal reservoirs hold hot water or steam in porous rock under a cap rock. Vapour-dominated ones give dry steam straight to a turbine; liquid-dominated ones, the common kind, give pressurised hot water that must be flashed or used in a binary plant. Geopressured zones hold hot brine at high pressure. Hot dry rock has heat but no water, so engineers fracture it and circulate water down one well and up another (an enhanced geothermal system). Magma is the hottest and hardest to tap. Heat can sit close to the surface: Puga's wells in Ladakh reached 135 °C only 400 m down. Low-temperature heat also serves directly: bathing, drying, and ground-source heat pumps, which MNRE is testing with IIT Delhi.
How much geothermal energy does India have, and what is happening with it?
The Geological Survey of India estimates about 10,600 MW of geothermal potential from exploring 381 hot springs. MNRE notified a National Policy on Geothermal Energy on 15 September 2025. At Puga, ONGC Energy Centre finished two 1,000 m wells in July 2026 and measured 135 °C at 400 m, for a 1 MW pilot plant; no date for power was given. MNRE has also funded IIT Madras to fit a 450 kW organic Rankine cycle plant to unused oil and gas wells, and a solar-geothermal hybrid at PDEU in Gujarat.
What fraction of hot reservoir water turns to steam in a flash plant?
Water at 230 °C stays liquid deep down because of the pressure. When it rises and enters a separator at lower pressure, part of it boils (flashes); the enthalpy balance gives the steam fraction. At 5 bar, x = 0.166: only 17 % becomes steam. That steam expands to a 0.1 bar condenser and gives 83 kJ per kg of brine, so a megawatt needs 12.1 kg/s from the wells. A lower separator pressure flashes more steam but each kilogram does less work; the best here is 3.1 bar, 86 kJ/kg. Try it in the interactive.
Dry steam, single flash, double flash or binary: which plant suits which reservoir?
Dry steam fields send steam straight to the turbine. Hot water fields use flash plants. A double flash re-flashes the separated brine at a lower pressure: at 230 °C the best pair of separators gives 108 kJ/kg against 86 for the best single flash, 26 % more. At lower temperatures a binary plant wins: the brine heats a low-boiling fluid such as isobutane in a closed organic Rankine cycle and all the brine goes back underground. With the interactive's assumed 40 % second-law efficiency and 70 °C reinjection, binary beats the best single flash below about 188 °C.
What are the environmental issues with geothermal plants?
The brine is mostly reinjected, which keeps reservoir pressure up and stops the ground sinking. A 5 MW single flash plant at the interactive's default draws 60 kg/s from its wells and sends 50 kg/s of separated brine back down. Dissolved gases such as hydrogen sulphide and carbon dioxide come out with the steam and must be treated or vented. Silica in the brine scales pipes as it cools, which sets how cold the reinjected brine can be (70 °C in the binary default). Land use is small and output steady day and night.
Puga's wells found 135 °C water. With 70 °C reinjection and a 40 % second-law efficiency (assumed), which plant gets more work per kg of brine?
The best single flash gets 20.6 kJ/kg at 0.67 bar; the binary plant gets 22.5 kJ/kg and keeps all the brine liquid. At such low temperatures little water flashes, so binary is the usual choice.
Worked examples
Flash fraction and brine flow for a 5 MW plant · Reservoir liquid at 230 °C, separator 5 bar, condenser 0.1 bar, turbine isentropic efficiency 85 %, no parasitic loads. Steam properties from IAPWS-IF97
- h_res = h_f(230 °C) = 990.2 kJ/kg. At 5 bar: T_sat = 151.8 °C, h_f = 640.2, h_g = 2748.1 kJ/kg
- x = (990.2 − 640.2)/(2748.1 − 640.2) = 0.1661
- Turbine: s_g(5 bar) = 6.8206; at 0.1 bar s_f = 0.6492, s_g = 8.1489, so x_2s = 0.8229 and h_2s = 2160.2 kJ/kg
- Work per kg steam = 0.85 × (2748.1 − 2160.2) = 499.7 kJ/kg; per kg brine = 0.1661 × 499.7 = 83 kJ/kg
- Brine for 5 MW = 5,000/83 = 60.3 kg/s, of which 10 kg/s is steam
Answer: x = 0.166; about 60 kg/s of brine for 5 MW
Choosing the separator pressure · Reservoir liquid at 200 °C, condenser 0.1 bar (45.8 °C), turbine 85 %. Rule of thumb: put the separator temperature halfway between reservoir and condenser
- Halfway temperature = (200 + 45.8)/2 = 122.9 °C, so p_sep = p_sat(122.9 °C) = 2.18 bar
- At that pressure x = 0.1533 and the work is 60.08 kJ per kg of brine
- A fine search over separator pressure finds the best at 1.96 bar with 60.19 kJ/kg
- The rule is 0.18 % short of the true optimum: the curve is flat near its top
Answer: About 2 bar and 60 kJ/kg; the equal-temperature rule lands within a fraction of a percent
Brine needed for a 1 MW binary pilot at 135 °C · Puga-like brine at 135 °C cooled to 70 °C at constant pressure in an ORC heater; dead state 25 °C; plant second-law efficiency 40 % (assumed)
- Heat from brine = h(135 °C) − h(70 °C) = 567.8 − 293.2 = 274.5 kJ/kg
- Exergy of that heat = q − T0 (s1 − s2) = 274.5 − 298.15 × (1.6872 − 0.9548) = 56.2 kJ/kg
- Work = 0.40 × 56.2 = 22.46 kJ/kg; thermal efficiency = 22.46/274.5 = 8.2 %
- Brine for 1 MW = 1,000/22.46 = 44.5 kg/s; the best single flash (20.59 kJ/kg) would need 48.6 kg/s
Answer: About 45 kg/s of 135 °C brine per MW with a binary plant
Green hydrogen and fuel cells: electrolysis, storage and efficiency
Which syllabus covers thisAU EE3014 Unit 1VTU BME654B Module 5; BEE654B Module 3SPPU 203141 Unit 6 (fuel cell); 402051B Unit 6 (bio-H2 only)MU MEDLO7032 Module 5; EEDO5011 Module 5KTU MET445 Module 5 (H2); PEEET412 Module 4
Why is India betting on green hydrogen?
Refineries and fertiliser plants already use hydrogen, made today from natural gas with CO2 released. Green hydrogen splits water with renewable electricity instead. The National Green Hydrogen Mission (₹19,744 crore, approved January 2023) targets 5 million tonnes a year by 2030, with about 125 GW of new renewables behind it. At 50 kWh per kg that is 250 billion kWh a year, so those 125 GW must run at a capacity factor of about 23 %, a typical solar-wind mix. About 8,000 tonnes a year of capacity was running by February 2026. IOCL's Panipat plant (10,000 tonnes a year) was bid at ₹397 per kg.
How much electricity does one kilogram of hydrogen take?
Faraday's law: each H2 molecule needs two electrons, so every volt of cell voltage costs 26.59 kWh per kg. The least possible voltage is 1.229 V (from NIST's Gibbs energy of water); at 1.481 V, the thermoneutral voltage, the electricity equals hydrogen's higher heating value of 39.39 kWh/kg. Real alkaline cells run near 1.8 V: in the interactive, 0.2 A/cm² at 80 °C gives 1.8 V and 49 kWh/kg, 80 % efficient on HHV. Push the current up and the losses (activation, then ohmic) grow.
Alkaline, PEM or solid oxide: what is the difference?
Alkaline cells use potassium hydroxide and a porous diaphragm, run at 70-90 °C and 0.2-0.8 A/cm², and are the cheapest: IRENA puts stacks at USD 270/kW and systems at USD 500-1,000/kW (2020). PEM cells use a solid polymer membrane, run at 1-2 A/cm² and follow a solar or wind supply quickly, but need platinum and iridium. Solid oxide cells run at 700-850 °C and use heat to cut the electricity to 40-50 kWh/kg for the whole system, against 50-78 for alkaline. The interactive models an alkaline cell.
Why is hydrogen hard to store?
Per kilogram it carries the most energy of any fuel (39.4 kWh on HHV), but it is the lightest gas: one kilogram fills 12.1 m³ at room pressure, only 3.2 kWh per cubic metre. So it is compressed to 350-700 bar in carbon-fibre tanks, liquefied (it boils at −252.8 °C, and liquefaction costs a large share of its energy), or carried as ammonia or methanol. Making it also takes water: 8.9 kg per kg of hydrogen by stoichiometry, more after purification losses.
What does a fuel cell give back?
A fuel cell runs electrolysis backwards: hydrogen and oxygen in, electricity, heat and water out. PEM fuel cells power cars and buses; solid oxide and phosphoric acid cells suit stationary power. The US DOE says fuel cells can exceed 60 % efficiency. Take 55 % of the lower heating value: 1 kWh into the electrolyser at 49 kWh/kg gives back 0.37 kWh, a round trip of 37 %, against about 85 % for a battery. Hydrogen wins on long storage and heavy transport, not on efficiency. The Electric Vehicles book has an interactive fuel-cell polarisation curve.
On a cloudy hour the electrolyser drops to a third of its rated current density (0.2 to 0.067 A/cm²). Per kilogram of hydrogen, does it now need more electricity, less, or the same?
Faraday's law fixes the charge, but energy is charge × voltage, and the voltage falls with current: 1.799 V at 0.2 A/cm² becomes 1.498 V at 0.067 A/cm² (80 °C). Faraday efficiency drops a little (97.7 % to 95.4 %), so the energy goes from 49 to 41.8 kWh/kg. The catch is cost: the same stack now makes a third of the hydrogen.
Worked examples
Hydrogen output and energy per kilogram of a stack · An alkaline stack of 200 cells, 2.5 m² each, carries 5,000 A at 360 V. Faraday efficiency follows the interactive's model (f1 = 120 mA²/cm⁴, f2 = 0.98).
- Current density j = 5,000/2.5 = 2,000 A/m² = 200 mA/cm²; ηF = 200²/(120 + 200²) × 0.98 = 0.9771
- H2 = N I ηF/(2F) = 200 × 5,000 × 0.9771/(2 × 96,485) = 5.063 mol/s = 36.75 kg/h
- Power = 360 × 5,000 = 1,800 kW, so energy = 1,800/36.75 = 48.99 kWh/kg
- Check: 26.59 × 1.8 V per cell/0.9771 = 48.99 kWh/kg; HHV efficiency = 39.39/48.99 = 80.4 %
Answer: 36.7 kg/h at 49 kWh/kg, about 80 % efficient on HHV
Reversible and thermoneutral voltage from NIST data · NIST: ΔfH°(H2O, liquid) = −285.830 kJ/mol; S° of H2O(l) 69.95, H2 130.680, O2 205.152 J/mol K. F = 96,485 C/mol. A cell runs at 1.9 V.
- ΔS = 69.95 − 130.680 − ½ × 205.152 = −163.306 J/mol K
- ΔG = ΔH − TΔS = −285,830 − 298.15 × (−163.306) = −237140 J/mol
- Reversible voltage = |ΔG|/(2F) = 237140/(2 × 96,485) = 1.2289 V; thermoneutral = 285,830/(2 × 96,485) = 1.4812 V
- Between them, T|ΔS| = 48.69 kJ/mol must come as heat; at 1.9 V the cell makes its own heat. HHV efficiency = 1.481/1.9 = 78 %
Answer: 1.229 V and 1.481 V; a 1.9 V cell is 78 % efficient on HHV
Tonnes a year and LCOH of a solar-powered electrolyser · A 1 MW electrolyser at Jodhpur follows a 1 MW solar plant (CUF 20 %, the book's PV model) at 49 kWh/kg. Electricity ₹2.39/kWh (the plant's LCOE). Capex USD 770/kW (IRENA 2020) at ₹93.765/USD, 9 %, 20 years, O&M 2 % a year.
- Energy = 1,000 kW × 0.20 × 8,760 h = 1,752 MWh; hydrogen = 1,752,000/49 = 35.8 t/yr
- Capex = 770 × 93.765 = ₹72,199/kW → ₹7.22 crore; CRF(9 %, 20) = 0.1095
- Annual capital + O&M = 7.22 crore × (0.1095 + 0.02) = ₹93.5 lakh → ₹262/kg
- Electricity = 2.39 × 49 = ₹117/kg; LCOH = 117 + 262 = ₹379/kg
- At a 56 % load factor (solar + wind) the same stack makes 100 t/yr and capital falls to ₹93/kg
Answer: 35.8 t/yr at about ₹379/kg: capital, not electricity, dominates at a 20 % load factor
Grid integration: variability, the duck curve, storage and curtailment
Which syllabus covers thisAU EE3014 Units 3-5VTU BME654B Module 3 (wind grid integration)SPPU 402051B Unit 1 (forecast), Unit 5; 203141 Unit 6MU EEDO5011 Module 3 (MPPT, grid)KTU not in syllabus
What changes when a grid runs on sun and wind?
Two things. Variability: solar follows the sun and wind follows the weather, so output swings hour to hour. Uncertainty: tomorrow's clouds are only forecast. India met a record 256.1 GW in April 2026 with 57 GW of solar at the peak hour (22 %) and about 81 GW at midday. On rainy, low-demand days in August 2024, solar and inflexible plants pushed the frequency above 50.05 Hz for up to 38 % of the day (CERC order, March 2025). Spinning generators also give inertia that inverters do not, unless designed to.
What is the duck curve?
Subtract solar and wind from demand and the net load sags at midday and climbs steeply after sunset: a duck's belly and neck. In the interactive's state grid (original numbers, 20 GW peak), 14 GW of solar in April gives an evening ramp of 3.4 GW per hour and 18 GWh of curtailed solar with coal at a 55 % minimum. An 8 GWh battery cuts curtailment to 2.1 GWh and the evening peak from 16.3 to 14.3 GW. The shapes are illustrative; India's April 2026 record peak still came in daylight, with 57 GW of solar helping.
Why does a coal minimum force curtailment?
Coal units needed for the evening peak cannot be switched off at noon; they can only turn down. The Grid Code sets a 55 % minimum turn-down for inter-state units, and the CEA's 2023 regulations ask for 40 % in phases, with ramp rates of at least 3 % per minute above 70 % load. CEA's 2029-30 study found 14.68 % of renewable energy curtailed on the windiest-sunniest day at a 55 % minimum and 10.56 % at 40 %. Tamil Nadu ran Tuticorin and Mettur units on two shifts 308 times in 2022, through a coal shortage and the high-wind season.
How does storage help, and what does it cost?
A battery charges on the midday surplus and discharges into the evening peak, losing energy each cycle: SECI tenders demand at least 85 % round trip, and CEA plans with 80 % for pumped hydro. Prices fell fast. A two-hour battery bid at ₹2.8 lakh per MW per month in April 2025 works out to ₹4.6 per kWh at one cycle a day and ₹2.3 at two; the Ministry of Power put storage at about ₹2.1 per kWh in December 2025. Pumped storage, flexible hydro, gas and demand response are the other flexibility tools.
Forecasting, scheduling and hybrid systems: how do they fit?
Wind and solar plants forecast their output a day ahead in 96 fifteen-minute blocks and are paid or charged for deviations under CERC's deviation settlement rules, so good forecasts save money. Off the grid, hybrids pair a variable source with a steady one: diesel-PV and biomass-diesel for villages and telecom towers, wind-PV because wind often blows when the sun is weak (India's monsoon), micro-hydro-PV in hills. The worked example sizes a diesel-PV hybrid and counts the diesel it saves.
In the interactive's state grid (April, no battery, coal minimum 55 %), solar doubles from 10 GW to 20 GW. What happens to the solar curtailed that day?
Curtailment goes from 2.3 GWh to 47 GWh, about 20 times. Only the solar above the coal-minimum line is lost, so every extra gigawatt adds more spill than the one before. Storage, a lower minimum or midday demand (pumping, electrolysers) must grow with the solar.
Worked examples
Can the coal fleet follow the evening ramp? · A state's coal fleet of 6,000 MW sits at its 55 % minimum at 5 pm. Net load then rises by 3,200 MW in one hour. Ramp rates per the CEA 2023 regulation: 2 % per minute between 55 and 70 % load, 3 % per minute between 70 and 100 %.
- Head-room = 6,000 × (1 − 0.55) = 2,700 MW, less than the 3,200 MW needed
- 55 → 70 %: 900 MW at 120 MW/min = 7.5 min; 70 → 100 %: 1,800 MW at 180 MW/min = 10 min
- The whole range takes 17.5 min, so speed is fine; the shortfall is head-room (megawatts): 3,200 − 2,700 = 500 MW
- That 500 MW must come from hydro, gas, a battery, imports or demand response
Answer: Fast enough (17.5 min end to end) but 500 MW short; flexibility must come from elsewhere
What an 8 GWh battery gives back · A 4-hour battery of 8 GWh charges fully on midday solar and discharges into the evening peak at 85 % round trip. Storage costs ₹2.3 per kWh delivered (a two-cycle reading of the April 2025 bid).
- Power = 8 GWh/4 h = 2 GW
- Energy back = 8 × 0.85 = 6.8 GWh, lost as heat = 1.2 GWh
- At full power it covers 6.8/2 = 3.4 hours of the evening peak
- Daily cost = 6.8 million kWh × ₹2.3 = ₹1.56 crore
Answer: 6.8 GWh over 3.4 h at 2 GW, about ₹1.6 crore a day
Diesel saved by a diesel-PV hybrid · A village micro-grid near Chennai uses 40 kWh a day from a diesel genset (3.3 kWh per litre, diesel ₹90 per litre, both assumed). Add 8 kWp of PV (the book's Chennai yield 4.07 kWh/kWp/day) with a battery; 85 % of the PV energy reaches the load (assumed).
- PV energy = 8 × 4.07 = 32.5 kWh/day; useful = 0.85 × 32.5 = 27.7 kWh/day
- Diesel now supplies 40 − 27.7 = 12.3 kWh/day, a solar fraction of 69 %
- Fuel: 12.1 → 3.7 litres/day; saved 3,059 litres a year
- Saving = 3,059 × ₹90 = ₹275,332 a year
Answer: About 3,059 litres and ₹2.8 lakh of diesel a year
Economics and policy: LCOE, CUF, tariffs and incentives
Which syllabus covers thisAU CME365 Unit 1 (plans)VTU not in syllabusSPPU 402051B Units 1, 3, Unit 4 (LCC, payback, LEC); 203141 Unit 4MU MEDLO7032 Module 6 (economics)KTU MET445 Module 5
Why is a rupee next year worth less than a rupee today?
Money can earn interest, so future rupees are discounted: at 9 % a year, ₹100 ten years from now is worth ₹42.2 today. The capital recovery factor turns a lump of capital into equal yearly payments over a plant's life: at 9 % for 25 years it is 10.18 % a year, so every ₹1 crore of capex costs ₹10.18 lakh a year. Solar and wind are almost all capital, so the discount rate moves their cost more than anything except the capacity factor.
What is the levelised cost of energy?
LCOE is the constant price per kWh that repays every cost over the plant's life: discounted costs divided by discounted energy. In the interactive, at the book's 9 %, utility solar costs ₹2.17 against NTPC's ₹2.55 auction, wind ₹3.22 against SECI's ₹3.78, solar with a 4-hour battery ₹3.54 against ₹3.52, and a new coal plant ₹4.02 with CEA's 2021-22 capex (new coal tariffs are now ₹5.85-6). Tariffs sit above these because bidders also price transmission, risk and equity returns.
Why does the capacity factor matter so much?
The capacity utilisation factor (CUF) is the energy a plant makes divided by what it would make at full power all year. Capex is paid per kW, but revenue comes per kWh, so LCOE scales almost as 1/CUF for solar and wind. CEA expects about 22 % for solar and 25 % for wind by 2029-30; the book's Chennai rooftop reaches 16.9 %, which is why it costs ₹3.94 per kWh against utility solar's ₹2.17. In the tornado, CUF and capex are always the two longest bars for renewables.
Which policies shape what renewables cost and earn?
Reverse auctions (SECI, NTPC) let developers bid tariffs down: ₹2.55 for solar in December 2024 and ₹3.12 for solar with storage in January 2026. Round-the-clock and peak-power auctions (FDRE) clear at ₹5.25-6. Since 6 June 2022 the Green Energy Open Access Rules let consumers of 100 kW or more (down from 1 MW) buy renewable power directly. Rooftop owners get net metering and PM Surya Ghar subsidy (₹78,000 for 3 kW). The ALMM lists require approved Indian modules, and from June 2026 Indian cells, in net-metering and open-access projects.
Energy audit, ECBC and zero-energy buildings: where do they fit?
The cheapest kWh is the one not used. An energy audit measures where energy goes and ranks fixes by payback; the Energy Conservation Act 2001 lets the government direct large designated consumers to get audits by accredited auditors. ECBC 2017 covers new commercial buildings above 100 kW or 120 kVA; ECBC+ and SuperECBC buildings use 35 % and 50 % less energy than typical. A zero-energy building makes on site what it uses in a year: a 2,000 m² Chennai office at 180 kWh/m² needs 360 MWh, too much for its roof PV (212 MWh), until SuperECBC halves the need.
A solar developer's discount rate rises from 9 % to 11 %. Its LCOE rises by about how much?
Solar is almost all capital, and the capital recovery factor for 25 years goes from 10.18 % to 11.87 %, up 17 %. With O&M unchanged the LCOE rises about 15 %. Cheap finance is why falling interest rates made solar tariffs fall.
Worked examples
LCOE of a 50 MW solar plant by hand · 50 MW at ₹3.5 crore per MW (BNEF figure quoted by Berkeley IECC, May 2025), O&M 1 % of capex a year, CUF 22 %, 25 years, discount rate 9 %. Ignore degradation first.
- Capex = 50 × 3.5 = ₹175 crore; work per MW
- CRF(9 %, 25) = 0.09 × 1.09²⁵/(1.09²⁵ − 1) = 0.10181
- Yearly cost per MW = 3.5 crore × 0.10181 + ₹3.5 lakh = ₹39.13 lakh
- Energy per MW = 1,000 kW × 0.22 × 8,760 h = 1,927,200 kWh
- LCOE = 3,913,219/1,927,200 = ₹2.031/kWh; with 0.6 %/yr module degradation (discounted sums) ₹2.13
Answer: About ₹2.03/kWh (₹2.13 with degradation), ₹0.42 under NTPC's ₹2.55 auction
Life-cycle value of the book's rooftop, and why slabs matter · Chapter 13's roof: 3.24 kWp in Chennai, 4,810 kWh in year 1; net cost after PM Surya Ghar ₹77,800; first-year bill saving under net metering ₹40,535 (TN slabs, chapter 13). Savings fall 0.6 % a year with the modules, no tariff rise; inverter ₹25,000 in year 12 (assumed); 9 %, 25 years.
- Present-value factor of a saving falling 0.6 %/yr = Σ 0.994^(t−1)/1.09^t over 25 years = 9.3774
- Savings = 40,535 × 9.3774 = ₹380,114; inverter = 25,000/1.09¹² = ₹8,888
- NPV = 380,114 − 77,800 − 8,888 = ₹293,425
- Why slabs matter: each solar kWh here saves ₹8.43, because it removes the household's top slabs (₹9.95-12.15). Valued at the first slab (₹4.95) the saving is only ₹23,811 a year and the NPV ₹136,593
Answer: NPV about ₹2.93 lakh; the first-slab shortcut would understate it by more than half (₹1.37 lakh)
A lighting audit in a college block · An audit finds 300 fluorescent tubes drawing 52 W each with their ballasts, on 10 hours a day, 300 days a year. LED tubes draw 18 W and cost ₹500 each; electricity ₹8/kWh; grid factor 0.71 kg CO2/kWh (all assumed).
- Saving per tube = 52 − 18 = 34 W; total = 300 × 34 = 10.2 kW
- Energy saved = 10.2 kW × 10 h × 300 days = 30,600 kWh/yr
- Money saved = 30,600 × ₹8 = ₹244,800/yr; investment = 300 × 500 = ₹150,000
- Payback = 150,000/244,800 = 0.61 years (7.4 months); CO2 avoided = 21.7 t/yr
Answer: 30,600 kWh and ₹2.45 lakh a year; pays back in about 7 months
Keep going
- Heat and Mass Transfer interactive book
- Power Electronics interactive book
- Power Systems interactive book
- Battery Management Systems interactive book
- Fluid Mechanics interactive book
- All Skill-Lync interactive books
Recommended reading
- Solar Engineering of Thermal Processes, J. A. Duffie and W. A. Beckman
- Solar Energy: Principles of Thermal Collection and Storage, S. P. Sukhatme and J. K. Nayak
- Non-Conventional Energy Sources, G. D. Rai
- Wind Energy Explained: Theory, Design and Application, J. F. Manwell, J. G. McGowan and A. L. Rogers
- Renewable Energy Resources, J. Twidell and T. Weir
- Solar Photovoltaics: Fundamentals, Technologies and Applications, C. S. Solanki
How this book was made
Chapters follow the official syllabi listed above. Every interactive's formulas and every worked-example number were checked by script. Explanations, animations and examples are original to Skill-Lync.