Solar PV systems and MPPT

In short

A solar photovoltaic (PV) system converts sunlight into DC electricity in a semiconductor junction and then conditions that power with DC-DC converters and inverters so it can feed a battery, a motor drive or the grid. Because a PV module's output power has a single peak on its P-V curve that moves with irradiance and temperature, the converter must continuously hunt for that peak - this is Maximum Power Point Tracking (MPPT). The topic sits at the junction of power electronics, control and renewable energy, and is asked in GATE EE/EC (PV characteristics, boost converter duty-ratio sums), university power-electronics and renewable-energy papers, and interviews at Tata Power Solar, Adani Solar, Waaree, Vikram Solar, Luminous, Delta, Sungrow, SMA, NTPC, SECI, NIWE and EPC firms. With India targeting 500 GW of non-fossil capacity by 2030 and rooftop/PM Surya Ghar schemes expanding, PV design, commissioning and O&M roles are among the largest electrical-engineering employers in the country.

Explained

PV cell physics and the equivalent circuit

• A PV cell is a large-area p-n junction. Photons with energy above the bandgap (silicon, Eg ≈ 1.12 eV, so λ < 1.1 µm) create electron-hole pairs; the junction's built-in field separates them, giving a photocurrent.
• Single-diode model: a current source I_L in parallel with a diode, a shunt resistance R_sh (leakage, edge paths) and a series resistance R_s (grid fingers, bus bars, contacts, bulk).
• Open-circuit voltage V_oc: about 0.60-0.75 V for a modern crystalline-silicon cell (≈0.68 V PERC, ≈0.72 V TOPCon, ≈0.74 V HJT; older Al-BSF cells ≈0.62 V); short-circuit current I_sc is nearly proportional to irradiance and scales with cell area (about 13-18 A for modern 182 mm / 210 mm wafers).
• Cell → module → string → array: cells in series add voltage (bypass diodes across sub-strings), modules in series form a string (voltage adds), strings in parallel add current.
• Technologies: mono-crystalline PERC / TOPCon / HJT (commercial module efficiency ~20-23%), multi/poly-crystalline (now largely phased out), thin film - CdTe, CIGS, a-Si (lower efficiency, better high-temperature and diffuse-light behaviour), and emerging bifacial modules that collect albedo on the rear.

I-V and P-V characteristics, and the effect of irradiance and temperature

• The I-V curve runs from (0, I_sc) to (V_oc, 0); the P-V curve rises, peaks at the maximum power point (MPP) (V_mp, I_mp) and falls to zero. Typically V_mp ≈ 0.75-0.85 V_oc and I_mp ≈ 0.9-0.95 I_sc.
• Fill factor FF = V_mp·I_mp/(V_oc·I_sc) - 0.7-0.82 for good c-Si cells; it falls when R_s rises or R_sh falls.
• Standard Test Conditions (STC): 1000 W/m², cell temperature 25 °C, air mass AM 1.5. Nameplate watts (Wp) are at STC. NOCT/NMOT ratings (800 W/m², 20 °C ambient, 1 m/s wind) are closer to field output.
• Irradiance: I_sc ∝ G almost exactly; V_oc rises only logarithmically with G. So current (and power) track irradiance nearly linearly.
• Temperature: V_oc falls at about −0.27 to −0.35 %/°C (roughly −2 mV/°C per cell), I_sc rises about +0.04 to +0.06 %/°C, and peak power falls at about −0.30 to −0.40 %/°C for c-Si. Hence output is lower on a hot Indian May afternoon than on a cool, bright winter day at the same irradiance.
• Left of MPP the module behaves like a current source; right of MPP like a voltage source - this is why the slope dP/dV changes sign and MPPT algorithms work.

Why MPPT is needed, and the impedance-matching view

• A PV source is non-linear: it delivers maximum power only into one particular load resistance, R_mpp = V_mp/I_mp. A fixed load (battery clamped at 12 V, a resistive heater) forces operation away from MPP and wastes energy.
• An MPPT is a DC-DC converter plus a control loop: by varying duty ratio D, it makes the PV side "see" the resistance it wants while the output side keeps its own voltage.
• For an ideal boost converter, R_in = R_L(1−D)²; for a buck, R_in = R_L/D²; for buck-boost, R_in = R_L(1−D)²/D². So the MPPT regulates D (not the load) to match impedance.
• Converter choice: boost when module/string voltage is below the DC-link or battery voltage (most grid-tied and small systems); buck when the array voltage is well above the battery; buck-boost / SEPIC / Cuk when the required ratio crosses unity. Boost is also preferred for continuous input current, which gives less PV-current ripple.
• Compared with a simple PWM charge controller (which just connects the array to the battery through a switch), an MPPT controller typically harvests 15-30% more energy, especially in cold weather or with high-V_oc arrays.

MPPT algorithms

1. Perturb and Observe / Hill climbing (P&O): change V_ref (or D) by ΔD, measure the new power. If P increased, keep perturbing in the same direction; if it decreased, reverse. Simple, sensor-light (V and I only), the industry workhorse. Drawbacks: steady-state oscillation about the MPP, slow or wrong-direction response under fast irradiance change, and ΔD is a speed-versus-ripple trade-off (variable-step P&O fixes much of this).
2. Incremental Conductance (IncCond): uses dP/dV = 0 ⇒ dI/dV = −I/V. If dI/dV > −I/V the operating point is left of MPP (increase V); if less, decrease V; if equal, hold. Better under changing irradiance than P&O, needs differentiation and so clean, fast sampling.
3. Constant Voltage (CV) / fractional V_oc: hold V_array ≈ k·V_oc with k ≈ 0.76-0.80. Cheap, robust at low light; needs periodic open-circuiting (or a pilot cell) and never sits exactly at MPP.
4. Fractional I_sc: I_mp ≈ 0.85-0.92 I_sc; needs a periodic short-circuit or a reference cell.
5. Ripple Correlation Control (RCC): correlates the natural switching ripple in v and p; dp/dt·dv/dt sign steers the duty ratio. Fast, no added perturbation.
6. Model/intelligent methods: fuzzy logic, ANN, and global search methods - Particle Swarm Optimisation (PSO), grey wolf, cuckoo search - used mainly for partial shading where the P-V curve has several local maxima and P&O can latch onto the wrong one.
7. Two-stage practice: a global scan or sweep of the P-V curve every few minutes to locate the global peak, then local P&O for fine tracking. Commercial inverters advertise this as "shadow scan"/"OptiTrac Global Peak".
8. MPPT efficiency = energy actually harvested / energy available at the true MPP; good converters achieve >99.5% static tracking efficiency, with dynamic (changing-irradiance) tracking a little lower. Do not confuse it with the inverter's conversion efficiency, whose European-weighted value η_EU is typically 96-98%.

Partial shading, mismatch and protection diodes

• One shaded cell in a series string limits the whole string's current and can be driven into reverse bias, dissipating power as a hot spot that cracks the cell or burns the encapsulant.
• Bypass diodes (typically 3 per 60/72-cell module, in the junction box) conduct around a shaded sub-string, capping its reverse voltage; they create the characteristic multi-step I-V curve and multiple P-V peaks.
• Blocking diodes in series with a string stop reverse current from stronger strings or from a battery at night (modern designs often use the controller's FET instead, to avoid the diode drop).
• Mitigation: string the modules so shading falls across one string, use multiple MPPT inputs per inverter, DC optimisers or module-level microinverters, or pick array topologies (SP, total-cross-tied TCT, bridge-link) that spread the mismatch.
• Mismatch also arises from manufacturing tolerance, soiling (a real problem in dusty north-Indian sites), differing tilt/azimuth and unequal ageing.

System architectures and power conditioning

• Stand-alone (off-grid): array → MPPT charge controller → battery (lead-acid / LiFePO₄) → inverter → AC loads. Sizing is driven by daily Wh demand, autonomy days and depth of discharge.
• Grid-connected: array → (DC-DC MPPT stage, optional) → DC link → voltage-source inverter with current control → LCL filter → transformer/grid. Control is usually dq-frame current control with a PLL, plus an outer DC-link voltage loop; the MPPT output is the DC-link voltage reference in single-stage designs.
• Topologies: central inverter (MW-scale plants), string inverter (commercial rooftops, several independent MPPT trackers), multi-string, microinverter (per module), and DC power optimisers.
• Inverter must provide anti-islanding (so it stops feeding a dead line), reactive-power/power-factor control, DC injection limits, LVRT/HVRT ride-through and surge protection; transformerless designs need leakage-current (RCMU) monitoring because of parasitic PV-to-ground capacitance.
• Solar pumping drives: array → boost → VSI → induction (V/f or vector control) or PMSM (vector-controlled) pump motor; MPPT is implemented by varying the motor's speed reference so the pump load naturally matches available power - a large PM-KUSUM market in India.
• Balance of system (BoS): mounting structure, trackers (horizontal single-axis; seasonal tilt adjustment on fixed structures), string combiner boxes with fuses and SPDs, DC and AC cables sized for <1-2% drop, earthing and lightning protection, monitoring/SCADA.

Energy yield, Indian standards and regulatory practice

• Yield metrics: specific yield (kWh/kWp/year - typically 1,350-1,700 in most of India, highest in Rajasthan and Gujarat), performance ratio PR (0.75-0.82 typical; it normalises for the irradiation received but not for temperature, so hot sites sit lower), capacity utilisation factor CUF (about 16-21% for fixed-tilt Indian plants), and availability.
• Losses chain: soiling, temperature, mismatch, DC cable, MPPT/inversion, AC cable, transformer, downtime, degradation (~0.5-0.7 %/year; warranties usually guarantee ~80-87% of rated power at 25 years).
• Standards (verify the current edition before use): IEC 61215 (module design qualification), IEC 61730 (module safety), IEC 61853 (energy rating), IEC 62109 (inverter safety), IEC 62116 (anti-islanding test), IEC 61724 (monitoring), IEC 62548 (array design requirements), IS 14286 (crystalline modules, aligned to IEC 61215), IS 16221 (module safety, aligned to IEC 61730), and CEA/CEI regulations for connectivity. IEC 61683 covers power-conditioner efficiency measurement.
• Indian framework: MNRE is the nodal ministry; SECI and NTPC run reverse auctions; CEA (Technical Standards for Connectivity of Distributed Generation) Regulations and state SERC net-metering/gross-metering regulations govern rooftop export (limits and compensation vary by state and change often - check the current state order). The Electricity Act 2003 with its amendments and the Energy Conservation (Amendment) Act 2022 provide the legal base; ALMM (Approved List of Models and Manufacturers) restricts which modules/cells may be used in government-supported projects; BIS certification (IS 14286 / IS 16221) and DCR content rules apply to subsidised segments. PM Surya Ghar: Muft Bijli Yojana (launched February 2024) subsidises residential rooftop systems - slab amounts are revised by notification.
• Tariff and cost scale: utility-scale solar tariffs discovered in recent Indian auctions have been roughly ₹2.4-2.7/kWh, and installed EPC cost for a residential rooftop system is roughly ₹45,000-65,000 per kW before subsidy; both move with module prices, so treat them as indicative.

Key points

• PV cell - a p-n junction converting photons above the bandgap into current; modern silicon cell V_oc ≈ 0.6-0.75 V.
• STC - 1000 W/m², 25 °C cell, AM 1.5; the nameplate Wp condition, and optimistic versus field behaviour.
• I_sc ∝ irradiance - current scales almost linearly with G, while V_oc only rises logarithmically.
• Temperature coefficient - power falls ~0.30-0.40 %/°C for c-Si; V_oc falls, I_sc rises slightly.
• MPP - under uniform irradiance, the single point where dP/dV = 0; V_mp ≈ 0.8 V_oc, I_mp ≈ 0.9 I_sc.
• Fill factor - V_mp·I_mp/(V_oc·I_sc); a direct measure of cell quality, degraded by series resistance.
• MPPT = impedance matching - the converter's duty ratio makes the array see R_mpp = V_mp/I_mp.
• Boost input resistance - R_in = R_L(1−D)²; the key equation for duty-ratio problems.
• Perturb and Observe - step and compare power; simple, oscillates about MPP, can misjudge fast irradiance changes.
• Incremental conductance - tracks dI/dV = −I/V; better under changing irradiance, needs differentiation.
• Constant voltage - hold V ≈ 0.76-0.8 V_oc; cheapest, never exact, good at low light.
• Partial shading - produces multiple P-V peaks; local hill-climbing can lock onto a local maximum.
• Bypass vs blocking diode - bypass diode routes current around a shaded sub-string; blocking diode prevents reverse current into a string or array at night.
• Hot spot - a reverse-biased shaded cell dissipating the string's power; the main reason bypass diodes exist.
• String sizing rule - maximum string V_oc at the coldest expected temperature must stay below the inverter/system limit (usually 1000 V or 1500 V DC).
• Performance ratio - actual kWh ÷ (irradiation × kWp); 0.75-0.82 typical in India, normalised for the resource but still lower at hot sites.
• CUF - annual kWh ÷ (kWp × 8760); roughly 16-21% for fixed-tilt Indian plants, not the same thing as PR.
• Anti-islanding - mandatory inverter protection that disconnects on grid loss (IEC 62116 test).

Formulas and examples

Example 1 - Fill factor and efficiency. A module gives V_oc = 49.5 V, I_sc = 13.9 A, V_mp = 41.2 V, I_mp = 13.1 A at STC; area 2.58 m² (2.278 m × 1.134 m).

• P_mp = 41.2 × 13.1 = 539.7 W (a "540 Wp" module).
• FF = 539.7/(49.5 × 13.9) = 539.7/688.05 = 0.784.
• η = 539.7/(1000 × 2.58) = 20.9% - the usual figure for a 540 Wp TOPCon module. Always divide by the true module area, not the cell area: using only the cell area would give an impossibly high efficiency.

Example 2 - Temperature derating and cell temperature. Same 540 Wp module, NOCT = 45 °C, γ = −0.0035/°C. Ambient 40 °C, irradiance 900 W/m².

• T_cell = 40 + (45 − 20)/800 × 900 = 40 + 28.1 = 68.1 °C.
• Irradiance scaling: P at 900 W/m² and 25 °C = 540 × 0.9 = 486 W.
• Temperature factor = 1 + (−0.0035)(68.1 − 25) = 1 − 0.1509 = 0.8491.
• P_actual = 486 × 0.8491 ≈ 412.7 W - i.e. 76% of nameplate on a hot Indian afternoon.

Example 3 - Boost-converter duty ratio for MPPT. An array has V_mp = 120 V, I_mp = 8 A and feeds a 240 V DC link through a lossless boost converter.

• R_mpp = 120/8 = 15 Ω. Output power = input power = 960 W, so R_L = 240²/960 = 60 Ω.
• R_in = R_L(1−D)² ⇒ 15 = 60(1−D)² ⇒ (1−D)² = 0.25 ⇒ 1−D = 0.5 ⇒ D = 0.5.
• Check with the voltage ratio: V_o/V_in = 1/(1−0.5) = 2 ⇒ 240/120 = 2. ✔
• If irradiance halves so that I_mp = 4 A with V_mp ≈ 118 V, the MPPT must raise D to about 1 − 118/240 = 0.508 and the new R_in = 118/4 = 29.5 Ω matches R_L' = 240²/472 = 122 Ω × 0.242 ≈ 29.5 Ω. ✔

Example 4 - String sizing against the 1000 V limit. Module V_oc = 49.5 V at 25 °C, β = −0.30 %/°C, minimum expected cell temperature at a Leh site = −10 °C; inverter maximum DC input = 1000 V, MPPT window 200-950 V, module V_mp = 41.2 V.

• V_oc at −10 °C = 49.5 [1 + (−0.003)(−10 − 25)] = 49.5 (1 + 0.105) = 54.7 V.
• Maximum modules in series = 1000/54.7 = 18.3 ⇒ 18 modules (string V_oc = 984 V, inside the limit).
• At 60 °C, V_mp = 41.2[1 − 0.003 × 35] = 36.9 V ⇒ string V_mp = 18 × 36.9 = 664 V, comfortably inside the MPPT window. A 19-module string would give 1039 V at −10 °C and destroy the inverter.

Example 5 - Rooftop yield and payback. A 5 kWp rooftop system in Pune; annual irradiation on the tilted plane ≈ 1,850 kWh/m²; PR = 0.78; grid tariff ₹9/kWh; installed cost ₹2.6 lakh after subsidy.

• Annual energy = 5 × 1,850 × 0.78 = 7,215 kWh → specific yield = 7,215/5 = 1,443 kWh/kWp.
• CUF = 7,215/(5 × 8,760) = 16.5%.
• Annual saving (full self-consumption) = 7,215 × 9 = ₹64,935 → simple payback = 2,60,000/64,935 ≈ 4.0 years.

Where it is used

• Utility-scale solar parks: Bhadla (Rajasthan), Pavagada (Karnataka), Kurnool (Andhra Pradesh) - central inverters, SCADA-based MPPT and plant-level power control for NTPC, SECI, Adani Green, ReNew, Tata Power.
• Module and inverter manufacturing: Waaree, Vikram Solar, Adani Solar, Premier Energies, Tata Power Solar (modules); Delta, Sungrow, SMA, Luminous, Microtek (inverters/charge controllers) - roles in cell testing, flash-test I-V measurement, inverter firmware and MPPT algorithm tuning.
• Rooftop EPC and O&M: design engineers sizing strings against inverter windows, doing shadow analysis in PVsyst/Helioscope, commissioning under state net-metering rules; O&M engineers tracking PR, soiling and string-level under-performance.
• Solar water pumping (PM-KUSUM): VFD-based pump controllers with built-in MPPT for agricultural borewells - a major volume segment for Indian drive manufacturers.
• Telecom towers, railway signalling, street lighting and rural microgrids: stand-alone MPPT charge controllers with battery banks; Indian Railways' solar rooftops at stations and on coaches.
• Research and R&D roles: NISE (Gurugram), IITs and CSIR labs on module reliability, bifacial gain and advanced MPPT; hardware-in-the-loop testing of inverter control.

Interview questions

• Q: Why does a PV module need MPPT at all? A: Its I-V curve is non-linear with a single power peak, so it delivers maximum power only into one specific load resistance; MPPT varies the converter duty ratio to present that resistance as operating conditions change.
• Q: How do irradiance and temperature affect a module differently? A: Irradiance mainly scales short-circuit current (and so power) nearly linearly, while temperature mainly reduces open-circuit voltage - power falls roughly 0.35 %/°C above 25 °C.
• Q: State the incremental-conductance condition at the MPP. A: dP/dV = 0 leads to dI/dV = −I/V; if dI/dV > −I/V the point is left of the MPP and the voltage should be increased.
• Q: What is the main drawback of perturb and observe? A: It oscillates around the MPP in steady state and can move in the wrong direction when irradiance changes rapidly, because it cannot tell a power change caused by its own perturbation from one caused by the weather.
• Q: Why does partial shading break conventional MPPT? A: Bypass-diode conduction creates several local maxima on the P-V curve, and a hill-climbing algorithm can settle on a local peak; a global scan or PSO-type search is needed.
• Q: Difference between a bypass diode and a blocking diode? A: A bypass diode is connected across a cell sub-string to carry current around a shaded section and limit reverse voltage; a blocking diode is in series with a string to prevent reverse current from other strings or from the battery at night.
• Q: Why is a boost converter the usual MPPT stage in grid-tied systems? A: The string voltage is normally lower than the required DC-link voltage, and the boost topology also draws continuous input current, which keeps PV voltage and current ripple low.
• Q: How do you choose the number of modules in a string? A: Maximum string V_oc at the lowest expected cell temperature must stay below the inverter/system DC limit, and string V_mp at the highest expected temperature must stay inside the inverter's MPPT window.
• Q: Difference between PR and CUF? A: PR compares actual output with what the array should have given for the irradiation it received (a quality measure, largely independent of how sunny the site is, though hot sites still score lower); CUF compares annual output with the plant running at rated power for 8,760 hours (a resource measure).
• Q: What is anti-islanding and why is it mandatory? A: It is inverter protection that stops export when the grid fails, so that line workers and equipment are not exposed to an energised "island"; it is tested to IEC 62116 and required by CEA connectivity regulations.
• Q: An MPPT charge controller versus a PWM one - what is the real gain? A: PWM clamps the array near battery voltage, losing the difference between V_mp and V_batt; MPPT converts that excess voltage into extra current, typically giving 15-30% more energy.
• Q: Why is the FF of a cell lower than 1, and what reduces it? A: Because the diode's exponential characteristic rounds the knee; high series resistance, low shunt resistance and a high ideality factor flatten it further.
• Q: A string's current is limited by one soiled module. Why does the whole string suffer? A: Series-connected modules carry the same current, so the weakest module sets the string current unless its bypass diode conducts and removes its voltage contribution.

Common mistakes

• Treating STC output as field output - forgetting the irradiance scaling and the temperature derating, which together can cost 20-25% at midday in India.
• Applying the temperature coefficient to ambient temperature instead of cell temperature (use NOCT or a thermal model).
• Sizing a string on V_mp instead of maximum V_oc at the lowest temperature - the cause of real inverter over-voltage failures.
• Using the voltage ratio 1/(1−D) alone and ignoring that MPPT actually works by impedance matching; or forgetting R_in = R_L(1−D)² in boost problems.
• Mixing up PR, CUF and efficiency, or quoting CUF when PR is asked for.
• Confusing bypass and blocking diodes, or assuming bypass diodes eliminate shading loss rather than just preventing hot spots.
• Assuming a single MPPT input handles multiple orientations or partially shaded sub-arrays equally well.
• Claiming incremental conductance has zero steady-state error - discretisation and sensor noise still cause small oscillation.

Related topics

Rectifiers and converters · Inverters · Wind energy conversion systems · Electric drives · UPS and SMPS design · Power semiconductor devices