PV ROI Calculator Methodology SolarPlanner.id

1. Purpose & who this methodology is written for

Calculator purpose

Estimating annual energy production, electricity bill savings, and the payback period of rooftop PV systems for locations in Indonesia — based on auditable primary data sources.

Who it is for

• Homeowners and businesses considering a Rp 50–500 million capex who want to understand the basis of the calculation before deciding
• Academics and researchers who need a citation source for policy papers or energy-transition research
• Contractors and EPCs who want to verify the formula before issuing a quote to clients
• Regulators and policymakers seeking to understand baseline assumptions for the residential PV sector

Why it is open

A calculator whose output cannot be traced back to primary sources is a black box — not good enough for capex decisions worth millions of rupiah. This methodology is written so every figure can be audited back to its original source (NASA POWER, PLN, IEA PVPS, NREL), and so accuracy claims are validated against independent datasets — not asserted without evidence.

2. Primary data sources

What irradiance data source is used?

Primary: NASA POWER (Prediction Of Worldwide Energy Resources) Project — managed by NASA Langley Research Center (LaRC), parameter ALLSKY_SFC_SW_DWN (All Sky Surface Shortwave Downward Irradiance, kWh/m²/day).

Secondary validation: Global Solar Atlas 2.0 (Solargis/World Bank), ESMAP 2017, and PVGIS-ERA5 (EU JRC). Full cross-validation in Section 5 — Calibration & Validation.

What is the source of PLN electricity tariff data?

Primary: PLN Tariff Adjustment publication (PT PLN Persero, national service territory excl. Batam Island).

PLN tariffs have not risen for 9 consecutive quarters since Q4 2022 under the government tariff-freeze decision. That is why the numeric tariff values are identical across every quarterly file in our dataset — not an input error.

What is the source of PV component pricing data?

Refreshed quarterly (Jan / Apr / Jul / Oct) following the PLN tariff cadence. Current validity: Q2 2026 (refreshed 2026-05-20). Next refresh target: 2026-07-15.

Public Q2 2026 sources: Indonesian EPC public price lists (PowerSurya, Sobatbangun, Suryaqua, Solar Energy Indonesia, Solusi Power, Tuturasa, Sunergy) + FOB China TOPCon module benchmark (PV-Magazine International, Nastech Solar, JMBIPV Tech) + Indonesian retail distributor spot prices (Tokopedia, ROYALPV, Solusi Battery, BigGo, Deye Indonesia catalog). Manufacturer reference: Trina, JA Solar, Canadian Solar, Jinko, LONGi, Huawei, Sungrow, Growatt, Deye, Luxpower, Goodwe.

FOB China benchmark note, Q2 2026: TOPCon module FOB China $0.108–0.112/W (PV-Magazine International, May 2026). Up ~30% from December 2025 due to China's VAT export-rebate elimination on 1 April 2026. Landed Indonesia = FOB × USD/IDR × (1 + VAT 5%) × (1 + distributor margin 10–15%) ≈ Rp 1,700–2,400/Wp.

Refresh schedule: Quarterly runbook at docs/runbook/calc-pricing-quarterly-refresh.md — refresh target 15 Jan / 15 Apr / 15 Jul / 15 Oct. Drift rule: a 5–15% drift triggers a conservative figure update, >15% triggers an off-cycle refresh with an announcement banner to users.

3. Transparent formula: how the numbers are computed

How does SolarPlanner.id compute annual energy production?

kWh_tahun  = kapasitas_kWp × PSH × PR × 365
kWh_bulan  = kWh_tahun / 12
kWh_hari   = kapasitas_kWp × PSH × PR

Worked example (citable): A 5 kWp system in Jakarta with PSH 4.90 kWh/m²/day (NASA POWER 1984-2023, coordinates -6.2088°S, 106.8456°E) and a Performance Ratio of 0.80 yields 5 × 4.9 × 0.80 × 365 = 7,154 kWh/year or about 596 kWh/month.

How does SolarPlanner.id compute bill savings?

hemat_kWh_tahun     = produksi_kWh_tahun × tingkat_self_consumption
hemat_rupiah_tahun  = hemat_kWh_tahun × tarif_PLN_per_kWh
hemat_rupiah_bulan  = hemat_rupiah_tahun / 12

Default self-consumption is 100% for on-grid systems following Permen ESDM 2/2024 (Ministerial Regulation, effective 31 January 2024, net metering removed for new customers). Existing customers with a pre-2024-01-31 IUPTLU (electricity supply business license) keep the old export-import scheme (Permen 26/2021) for 10 years.

How does SolarPlanner.id compute the payback period?

capex_total_rupiah  = kapasitas_kWp × harga_per_kWp + ppn_11pct
payback_tahun       = capex_total_rupiah / hemat_rupiah_tahun_tahun_1

For multi-year ROI, the panel degradation factor is applied:

Tahun 1       : faktor = 1 − 0,02         (LID light-induced degradation)
Tahun 2 → 25  : faktor = 0,98 − 0,005 × (tahun − 1)   (linear 0,5% per tahun)

Tier-1 manufacturer performance warranty: ≥ 84.8% at year 25 (mono PERC) or ≥ 87.4% (TOPCon N-type).

How is GHI ↔ PSH converted?

PSH (kWh/m²/hari)        = GHI_tahunan (kWh/m²/tahun) / 365
GHI_tahunan (kWh/m²/thn) = PSH × 365

This conversion is lossless and commutative.

4. Explicit assumptions & constraints

Assumption #1 — default Performance Ratio 0.80

Source: IEA PVPS Task 13 (2018, 2021) for the tropical baseline; LSPMK/LIPI 2021 studies for the Indonesia correction (module derating 0.4%/°C × ΔT 25-30°C → 10-12% loss).

Assumption #2 — optimal tilt + azimuth for Indonesia

The calculator default assumes a tilt of 10-15° (tropical optimum) and an azimuth of 0° (north/south aligned, since Indonesia sits on the equator). Users with non-optimal orientation (east-west roofs, shading, slope >25°) should manually lower the PR to 0.70-0.75.

Assumption #3 — 100% self-consumption (post-Permen ESDM 2/2024)

Under Permen ESDM 2/2024 (Ministerial Regulation, effective 31 January 2024), rooftop PV export surplus is not compensated by PLN for new customers. Implications for the formula:

• 100% self-consumption assumption → all production is treated as offsetting the bill
• For oversized systems (production > consumption), effectiveness falls proportionally
• Existing customers pre-2024-01-31 with the old IUPTLU: the 65% export-import scheme (Permen 26/2021) applies for 10 years

Assumption #4 — quarterly PLN tariff snapshot

The calculator default uses the tariff R-1 1300 VA = Rp 1,444.70/kWh (Q2 2026, identical Q1 2024 – Q2 2026 due to a 10-quarter tariff freeze). The quarterly adjustment mechanism (Jan/Apr/Jul/Oct) applies only to the 13 non-subsidized tariff classes.

Assumption #5 — 40-year climatology, not a forecast

The PSH used is a climatological average for 1984-2023. An individual year can deviate ±20% from the 40-year average due to ENSO, volcanic eruptions, and tropical seasonal anomalies.

5. Calibration & validation: how do we know the numbers are accurate?

NASA POWER 1984-2023 PSH is valid for Indonesia, validated across three layers.

Layer 1 — Cross-validation NASA POWER vs Global Solar Atlas 2.0

Mean absolute deviation: 2.1% — below the declared calculator accuracy margin of ±15% (initial estimate).

Layer 2 — Cross-validation NASA POWER vs PVGIS-ERA5 (EU JRC)

A sample of 12 provincial capitals (latitude 5.55°N to 10.18°S), comparing NASA POWER 1984-2023 (MERRA-2) against PVGIS-ERA5 2005-2020 (ECMWF ERA5):

No calibration factor is required — two independent reanalysis datasets (MERRA-2 and ERA5) produce highly consistent values for Indonesia.

Layer 3 — Field measurement of 12 residential systems 2023-2025

What is the confidence level of the SolarPlanner.id calculator?

6. Methodology limitations

Every model is a simplification. The following are things we deliberately do not model, with the reasons and practical implications. It is better for users to know what the calculator does not handle than to trust the numbers without context.

What does the calculator not model?

1. Weather forecast for a specific year. The PSH used is a 40-year climatological average, not a forecast for any particular year. An extreme ENSO year can deviate ±20%. Cross-checking ERA5 reanalysis 2020-2025 shows a year-to-year PSH range averaging 9.65% (maximum 12.82% in Jakarta), with a significant delta against the 40-year average of only +0.91% — the climatological average remains representative for long-term residential and commercial planning. For industrial sizing >100 kWp and utility-scale, model the worst-case PSH × 0.90 (P90 bankability) before a financial decision.
2. Microclimate <50 km. The NASA POWER 0.5° × 0.625° grid does not capture micro-variation in high mountains (Wamena, Lanny Jaya, the Papua highlands >2,000 m above sea level).
3. Detailed shading. The calculator default does not model shading from trees, neighboring buildings, or roof structures. Local shading over 10% of the panel area = a potential 30-50% production loss (non-linear module-level shading).
4. Detailed panel degradation curve. The linear 0.5%/year model after a year-1 LID of 2% is a simplification — the actual derating factor is slightly non-linear; the difference is < 2% over 25 years.
5. Site-specific soiling. The tropical soiling factor in Indonesia varies 1-8% depending on industrial dust (high in Cikarang/Karawang), coastal aerosols, or a dry climate (low in NTT).
6. Inverter efficiency curve. The model assumes a constant PR; in reality an inverter has peak efficiency of ~98% at mid-load, dropping to ~93-95% at low-load (<20% rated).
7. Extreme weather & disasters. Hail, floods, earthquakes, volcanic eruptions — all outside the modeling. PV insurance is recommended for capex >Rp 100 million.
8. PLN Batam. The core data bundle focuses on PLN Persero. The Bright PLN Batam tariff is separate (a dedicated Permen ESDM / Ministerial Regulation). Available on request.

What is deliberately made conservative (under-promise, over-deliver)?

• Default PR 0.80 (not the 0.85 optimum) — realistic for new standard systems
• Default self-consumption 100% with a 0% margin — does not include the old export-import scheme (= silent upside for the user)
• Confidence interval ±10% (not the ±5% that may be statistically valid for a subset of large cities) — robust across all 121 coordinates

Is there any conflict of interest?

No. SolarPlanner.id (as of this document's writing) is an independent education platform + contractor directory. The calculator has no upsell bias toward any vendor — output is identical regardless of the component brand the user selects. Platform monetization via the contractor directory (lead-fee or subscription model) does not affect the calculator formula.

6.5 Update 2026-05-22 — Engine architecture refactor (Sprint 17)

On 2026-05-22, the calculator received an engine architecture refactor as part of Sprint 17 (Calculator MVP best-engine-possible). This change is transparent to current users but opens the path for Phase 2-6 features (persona profile, proper Monte Carlo, brand benchmark, beta tester). Full architecture documentation: ADR-015 Calculator Data Layer Architecture.

What changed in the engine?

• 5-dimensional data layer. The engine now supports data lookup that is cohort-aware (panel tier × age), region-aware (7 Indonesian climate zones), persona-aware (residential / small business / industrial), assumption distribution (Monte Carlo ready), and TKDN scenario branching (4 scenarios per Permen ESDM 5/2025 + Permenperin 34/2024 — Ministerial Regulations; TKDN = domestic content level).
• Singleton cache + TS const fallback. Data is looked up from the Postgres DB with an in-memory cache (5-minute TTL), falling back to a TypeScript const baseline if the DB is unresponsive. The pure-function engine is preserved for testability and a latency budget of <50ms p95.
• Strangler Fig migration. Phase 1 (cohort+region) shipped 2026-05-22, Phase 2-4 cascade over 2026-05-26 → 2026-06-30. Each phase is reversible in <1 sprint to minimize regression risk.

TKDN scenario branching — 4 component-sourcing scenarios

Under Permen ESDM 5/2025 (TKDN 60% mandatory for PLTS IPP) and Permenperin 34/2024 (TKDN computation for solar modules), the calculator now offers 4 optional sourcing scenarios in the “Adjust assumptions” section:

• Default (baseline ×1.00): No explicit preference. The engine uses neutral assumptions.
• Scenario A — Strict TKDN (×1.08): Government project / mandatory IPP. Panels + inverters must pass the TKDN ≥40% threshold. ~5-10% premium vs full import due to the local manufacturing gap.
• Scenario B — Hybrid TKDN Preferred (×1.03): Voluntary TKDN, optimizing partial incentive access. A mix of local + imported panels/inverters.
• Scenario C — Full China Import (×0.90): Pure private project, absolute cost optimization. China supply-chain commodities (Longi/Trina/JA Solar + Huawei/Sungrow). ~10-15% discount vs baseline but exposure to trade policy + supply-chain risk.

The multiplier is applied only to the base PV CapEx — the hybrid upgrade (battery + premium inverter) is not affected, to preserve the existing battery economics model.

Uncertainty range — Monte Carlo N=1000 sampling (Phase 3 v2)

Since the 2026-05-23 update, the calculator includes a P10/P50/P90 range for 3 key metrics (payback period, total investment, annual savings) based on Monte Carlo sampling, N=1000 iterations over 5 distributions per calculation:

• Irradiance (PSH) — per BMKG region. Phase 3 v2 ship: CV per 7 Indonesian climate zones from 12 BMKG stations (Open-Meteo 2020-2025 reanalysis cross-validated against NASA POWER 1984-2023). CV range 1.85% (stable Sumatra Equatorial) to 4.24% (Java-Bali Inland Dry, variable season). Fallback 5% from literature for coordinates without a region tag.
• Tariff escalation. CV 35% (PLN historical Q1 2015–Q2 2026 catch-up variance per ESDM tariff adjustment).
• Soiling rate. CV 35% (NREL Soiling Research; IEC TR 63226; Maghami et al. 2016 — generic Indonesia tropical default).
• Capex variance. CV 15% (IRENA Indonesia 2024 + AESI Annual Report 2024-2025 contractor pricing spread).
• Panel degradation. CV 30% (NREL Real-World Module Degradation 2013-2023; Jordan & Kurtz 2013).

The sampler uses a Box-Muller transform with a seeded RNG (deterministic — the same input always produces the same output). The methodology label in the calculator results = monte-carlo (previously fallback-15pct, the ±15% conservative band from Wave 2c).

Phase 3 v2 status — irradiance ALREADY refined. The per-climate-zone CV for Indonesia is now real BMKG field data (12 stations covering 7 zones). The remaining 4 dimensions (tariff/soiling/capex/degradation) are still literature-derived — refined in batch Phase 3 v3: regional soiling per city + regional capex benchmark + degradation per panel brand × cohort (Sprint 19-20 HELIOS workstream).

BMKG-derived CV per IDN-CLIM region:

• IDN-CLIM-001 (Sumatra Equatorial): CV 1.85% — 1 station (MES Medan). The most stable equatorial climate.
• IDN-CLIM-002 (Java Coastal Urban): CV 4.14% — 2 stations (JKT Jakarta + SUB Surabaya). Urban aerosol + monsoon.
• IDN-CLIM-003 (Java-Bali Inland Dry): CV 4.24% — 1 station (BDO Bandung). Strong dry-season variability.
• IDN-CLIM-004 (NTT-NTB Savanna): CV 2.51% — 3 stations (KOE Kupang + AMI Mataram + WGP Waingapu). Stable arid.
• IDN-CLIM-005 (Kalimantan Equatorial Forest): CV 3.73% — 2 stations (PNK Pontianak + PKY Palangka Raya).
• IDN-CLIM-006 (Sulawesi Mixed): CV 3.71% — 2 stations (UPG Makassar + MDC Manado).
• IDN-CLIM-007 (Maluku-Papua Maritime): CV 2.29% — 1 station (DJJ Jayapura).

Refresh path: the HELIOS data workstream (Sprint 19+) populates field data into the data-source.ts distribution layer — the engine + Monte Carlo sampler do NOT need code changes on data refresh. A pure swap of literature defaults for Indonesia field-measured values.

Why Monte Carlo, not just ±15%? Real distribution sampling produces a range that is asymmetric and realistic per metric — not merely a mirrored flat ±15%. The skewed tariff escalation distribution (truncated at the 0% floor, with a 2× tail in the catch-up scenario) makes the P90 benefit wider than the P10 — a fact captured in Monte Carlo but not in a flat ±15%. Backward compatibility: if Monte Carlo fails to initialize (extremely rare — a defensive guard), the engine falls back to the ±15% conservative band with the label fallback-15pct.

What has not changed in this Phase 1?

• Primary data (NASA POWER 1984-2023, BMKG 12 stations, quarterly PLN tariff) — still the same baseline, refreshed on a separate workstream.
• The Pasal 13 (Article 13) Permen ESDM 2/2024 formula — unchanged (commit 72062e0 remains authoritative).
• The 5-step on-grid UI wizard — unchanged. The TKDN picker was added to the optional “Adjust assumptions” section, without disrupting the default flow.
• Persona profile, brand benchmark, proper Monte Carlo, the interactive methodology-page comparison — deferred to Phase 2-5 (Sprint 18-23).

Backward compatibility

All previously shared calculator result URLs (share URL ?c=...) remain valid. The new TKDN field is optional + defaults to neutral, so old results do not change. The Prisma migration is non-destructive (all new columns nullable, ON DELETE SET NULL).

7. Citation format

Indonesian (academic)

SolarPlanner.id (2026). Metodologi Kalkulator ROI PLTS — Foundation. Basis data NASA POWER 1984-2023 (ekstraksi 2026-04-28) dan PLN Tariff Adjustment Q1 2024-Q2 2026, validasi PVGIS-ERA5 2005-2020 (mean abs delta 3,19%, no systematic bias). License CC-BY-4.0. URL: https://solarplanner.id/metodologi/kalkulator

English (academic)

SolarPlanner.id (2026). Methodology of the SolarPlanner.id PV ROI Calculator — Foundation. Indonesian rooftop solar irradiance dataset based on NASA POWER 1984-2023 (extraction date 2026-04-28) and PLN Tariff Adjustment Q1 2024-Q2 2026, cross-validated against PVGIS-ERA5 2005-2020 (mean absolute delta 3.19%, no systematic bias). Licensed CC-BY-4.0. URL: https://solarplanner.id/metodologi/kalkulator

Citation for a specific numeric fact

“A 5 kWp system in Jakarta produces ~7,154 kWh/year at PR 0.80 (PSH 4.90 kWh/m²/day NASA POWER 1984-2023, validated PVGIS-ERA5 +6.94% delta within the ±15% margin) — SolarPlanner.id, 2026, License CC-BY-4.0.”

8. Update cadence & contact

How often is the calculator data updated?

How to get in touch for questions / data corrections?

• Email: info@solarplanner.id
• Contact form: solarplanner.id/kontak

What is the license for this data + methodology?

CC-BY-4.0 (Creative Commons Attribution 4.0 International). May be used for any purpose (including commercial), with mandatory attribution SolarPlanner.id (2026), License CC-BY-4.0. Full license text: creativecommons.org/licenses/by/4.0/.

Primary source attribution (must be stated alongside the SolarPlanner.id attribution):

• NASA POWER:“These data were obtained from the NASA Langley Research Center (LaRC) POWER Project funded through the NASA Earth Science / Applied Science Program.”
• PLN Tariff: Official source PT PLN (Persero) — web.pln.co.id/pelanggan/tarif-tenaga-listrik/tariff-adjustment
• PVGIS:“PVGIS data © European Union, 2001-present.”