Evidence
How was each claim checked?
Hosting capacity10
Generic 400 V model (sensitivity case; the headline is the Indian street model, E55): Phase-managed rural DTR (63 kVA, 700 m): P5 'Passes screening tests' solar is 34.4 kWp and its bootstrap 90% range is 30 to 36 kWp. The 50% cap is 31.5 kWp, so the range straddles the cap. 'Just past the cap' is NOT settled. P5 = 5th percentile over 30 random trials of the largest solar that passes the screening tests (see E40). P50 for the same case is 48.0 kWp.
How it was checked
Read hosting_p5_bootstrap_ci.csv row rural_long / phase_managed (boot_lo 30, boot_hi 36, cap_50pct 31.5).
Limits
Assumed network, 30 random adoption trials, 3 kWp per roof. A real transformer may differ.
results/hosting_p5_bootstrap_ci.csv
Generic 400 V model (sensitivity case; the headline is the Indian street model, E55): Ignoring phases overstates safe solar: at the normal tap (1.00 pu) the balanced model gives 1.3 to 3.5 times the 'Phase-unbalanced case (assumed split, not measured)' P5 across the three street types. P5 = 5th percentile over 30 random trials of the largest solar that passes the screening tests (see E40).
How it was checked
Ratios of balanced P5 to unbalanced P5 in hosting_lv_summary.csv at source_pu 1.0: 255/190.05 = 1.34, 147/73.35 = 2.0, 63/18 = 3.5.
Limits
Simulation on assumed typical streets; not measured on a real network. The case name was 'Unbalanced (field)' until 18 Nov 2026; it is a simulation with a random phase for each home, not a measurement (internal key stays 'unbalanced').
results/hosting_lv_summary.csv
Generic 400 V model (sensitivity case; the headline is the Indian street model, E55): The 50% transformer cap can be tight or loose depending on the street: for the long rural street in the phase-unbalanced case (assumed split, not measured) P5 (18 kWp) is below the 31.5 kWp cap, while for dense urban the working-reading cap (125 kW, unit assumed) is below the 190 kWp P5. P5 = 5th percentile over 30 random trials of the largest solar that passes the screening tests (see E40).
How it was checked
Compared p5 and cap_50pct_kwp row by row in hosting_lv_summary.csv and hosting_cap_unit_sensitivity.csv.
Limits
Depends on which reading of the cap unit is used (see E04) and on the assumed network.
results/hosting_lv_summary.csv
The unit of the 50% cap is not stated in TGERC Reg 1 of 2025 cl 5.1 or in TGSPDCL guideline IV(a): both say '50% of its rated capacity' of the DTR (identical words, no unit). Working reading: 0.5 x DTR kVA treated as kW (50 kWp per 100 kVA); unit assumed: kW on the DTR kVA rating; TGSPDCL to confirm. The 0.9 factor (the 2016 rule, 45 kW per 100 kVA) is an optional sensitivity, not law; the app shows all four readings and a toggle for the 0.9 case. Between R1 and R2 no verdict changes in 0 of 18 LV results (precomputed in results/hosting_cap_unit_sensitivity.csv).
How it was checked
Four readings (R1 to R4) are tabulated side by side in hosting_cap_unit_sensitivity.csv; the app shows all four.
because the unit is not confirmed by the regulator or the discom. (That neither text states a unit: checked - correct, SME label [checked-correct], official text read by the SME and not re-read here; see E25.)
Limits
Until confirmed, results are shown for every reading. To settle: a TGSPDCL feasibility reply, a TGERC clarification or an RTI to the Chief Engineer (IPC & RAC). Limits apply to AC export, not panel DC size: columns say kWp (panel) or kWac (inverter) where known, otherwise unit not stated.
results/hosting_cap_unit_sensitivity.csv
All network values (street lengths, transformer sizes, houses, 3 kWp roofs, load per home) are assumed typical values, not discom data.
How it was checked
Stated in hosting/README.md and in the data notes: no transformer, feeder or house data is held.
because no discom network or transformer data is held
Limits
The pilot replaces the assumed networks with the discom's GIS, transformer and phase data.
hosting/README.md
The standard hosting-capacity screen uses the minimum daytime load with the maximum PV at the sunniest hour.
How it was checked
Literature, as stated by the SME.
(SME label [checked-correct]; not re-checked here)
Limits
Our screen uses the sunniest hours and a flat assumed midday use of 0.25 kW per home (0.5 as a sensitivity), not a measured minimum load.
SME answer, item 4
P5 on the generic 400 V model (sensitivity case; the Indian headline uses 200 trials, E55): the 5th percentile (numpy, linear interpolation) over 30 random trials of the largest rooftop solar (3 kWp roofs) at which no tested limit is broken in the sunniest-hour screens; each trial redraws which homes adopt and (except the balanced model) the phase of each home. P5 is shown beside P50 and P95. Each trial value is the largest number of 3 kWp roofs that passes (roofs added one at a time, limit found by bisection), so every trial value is a multiple of 3 kWp; the percentile interpolates linearly between the sorted values, so the printed P5 can fall between 3 kWp steps (for example 34.35 kWp). Wording written 18 Nov 2026 from the code; the SME may adjust it.
How it was checked
Read from hosting/run_lv.py and hosting/lv.py. Recomputed here: the 5th percentile of the 30 trial values in results/hosting_lv_runs.csv and results/hosting_lv_load050_runs.csv reproduces the summary P5 in all 36 cases (largest difference 2e-15). All 1,080 trial values in the two run files are multiples of 3 kWp (checked 18 Nov 2026); some summary P5 values are not.
(checked here, local files). The SME could not confirm the definition, which is why it is printed under every P5 figure
Limits
91 of 540 roof-by-roof trials and a second engine (OpenDSS) show P5 can be optimistic; treat it as a screening range.
results/hosting_lv_runs.csv
Generic 400 V model (sensitivity case; the headline is the Indian street model, E55): The case formerly named 'Unbalanced (field)' is now 'Phase-unbalanced case (assumed split, not measured)': in hosting/run_lv.py and hosting/lv.py each home is put on a random phase (uniform over 3 phases), redrawn in each of the 30 trials; no real phase data is used.
How it was checked
Read from the code (lv.py build, run_lv.py _trial: rng.integers(0, 3, houses)).
(checked here, read from the code)
Limits
Internal key 'unbalanced' is kept in the database, CSV model column and code; the label mapping is in hosting/README.md.
hosting/run_lv.py
Headline street hosting capacity on the Indian street model (11 kV/433 V DTR, 415/240 V declared supply, named Indian conductors; 200 random trials per case), at 11 kV 1.00 pu: apartment block 315 kVA, 60 m, ABC 3x150, three-phase solar: P5 288 kWp (90% range 288 to 291), P50 291, deciding limit voltage, 50% cap 157.5 kW (unit assumed); city street 160 kVA, 300 m, ABC 3x95, phase unknown: P5 24 kWp (90% range 21 to 26.85), P50 48, deciding limit voltage, 50% cap 80 kW (unit assumed); city street 160 kVA, 300 m, ABC 3x95, phase-managed: P5 47.85 kWp (90% range 44.85 to 50.85), P50 63, deciding limit voltage, 50% cap 80 kW (unit assumed); town street 100 kVA, 400 m, ACSR Rabbit, phase unknown: P5 8.85 kWp (90% range 6 to 9), P50 18, deciding limit voltage, 50% cap 50 kW (unit assumed); town street 100 kVA, 400 m, ACSR Rabbit, phase-managed: P5 18 kWp (90% range 18 to 18), P50 30, deciding limit voltage, 50% cap 50 kW (unit assumed); village street 63 kVA, 700 m, ACSR Weasel, phase unknown: P5 0 kWp (90% range 0 to 0), P50 6, deciding limit voltage, 50% cap 31.5 kW (unit assumed); village street 63 kVA, 700 m, ACSR Weasel, phase-managed: P5 0 kWp (90% range 0 to 0), P50 9, deciding limit voltage, 50% cap 31.5 kW (unit assumed). P5 = 5th percentile of the 200 trials; the 90% range is the bootstrap interval stored with the results. The generic 400 V model (E01 to E03) is kept as a sensitivity case.
How it was checked
Read row by row from results/indian_summary.csv (written by hosting/run_indian.py; columns p5, p5_lo, p5_hi, p50, cap_kwp, binding) and loaded unchanged by scripts/load_data.py (scenario 'indian'); the API values are tested against the CSV in backend/tests/test_indian_headline.py.
(read from the local results file; a model result on assumed streets, not a measurement)
Limits
All network values are assumed (conductor, length, load points, 3 kWp roofs, 0.25 kW midday use); the neutral is modelled full size (optimistic). The OpenDSS and roof-by-roof checks were run on the generic model only. Screening values, not approved capacity.
results/indian_summary.csv
When a street is already above +6% voltage with no solar at all (for example a 433 V DTR with the 11 kV side at 1.03 pu), the hosting result is 0 and the binding is 'over limit before solar'. The app shows this as 'Over the voltage limit before any solar: fix the tap/voltage first', never as 0 kWp safe. In results/indian_summary.csv 7 of 21 cases are over the limit before solar: apartment block 315 kVA at 1.03 pu; city street 160 kVA at 1.03 pu; town street 100 kVA at 1.03 pu; village street 63 kVA at 1.03 pu.
How it was checked
Read from hosting/lv.py (hosting_one checks zero roofs first and returns OVER_BEFORE_SOLAR) and counted in results/indian_summary.csv column binding; the Street page and API state are tested (backend/tests/test_indian_headline.py, ui_tools/e2e/test_street.py).
(read from the code and the local results file)
Limits
The voltage before solar depends on the assumed 11 kV voltage, the DTR tap and the assumed midday load; a real transformer's tap position is not held.
results/indian_summary.csv
PM-KUSUM5
On the assumed 15 km feeder, where a PM-KUSUM plant connects changes the answer about four-fold (4.75 MWp near the substation vs 1.2 MWp at the tail, pumps off, sending-end 1.00 pu).
How it was checked
Read hosting_kusum.csv; 4.75 / 1.2 = 3.96.
Limits
Precomputed grid on an assumed feeder with 30 pump transformers; not recalculated live by the app.
results/hosting_kusum.csv
PM-KUSUM Component A: 500 kW to 2 MW within 5 km of the substation. Component C: feeder-level solarisation sized from the last 3 years of feeder farm energy at 19% CUF (example: 10 lakh units a year needs about 600 kW).
How it was checked
MNRE guideline, read by the SME.
(SME label [checked-correct]; not re-checked here)
Limits
Scheme text not held in this folder.
SME answer, item 3
Telangana: Component A was cut from 4,000 MW to 1,000 MW on 31 Jan 2025 and the state is pursuing restoration; Component B none; Component C 28,000 pump sets.
How it was checked
TGERC order 32/2025, read by the SME.
(SME label [checked-correct]; not re-checked here)
Limits
Order not held in this folder.
SME answer, item 3
Farm supply is 9 hours a day in the daytime in Andhra Pradesh; the Telangana pattern is not confirmed. Solar peaks when pump load peaks, so it helps reverse flow only if the pumps really run.
How it was checked
SME answer, item 3.
because the Telangana schedule is not confirmed
Limits
The app shows pumps off and pumps running as two scenarios.
SME answer, item 3
Pump-transformer ratings and midday pump load: nothing on file. The app's assumed 25 kVA pump transformer (15 kW per pump, pf 0.8) is a scenario, input needed from the discom.
How it was checked
No data held; the SME has none.
because no pump-transformer or midday-load data is held
Limits
Input needed from the discom.
hosting/kusum.py
Forecast9
Day-ahead forecast error on the test period (115,719 daylight hours, 13 points): 31.2 W per kWp for the model retrained monthly, 39.1 for raw physics, 38.2 for the earlier run A, 43.6 for same-time-yesterday.
How it was checked
Read test_monthly.csv and test_static.csv (row ALL).
Limits
Scored versus a physics model, not versus measurement (archive weather; E08).
results/test_monthly.csv
Forecast accuracy on real rooftops.
How it was checked
No metered rooftop output or ground station is held, so no comparison can be run.
because no metered rooftop output is held
Limits
The target is a PV-model output on Open-Meteo archive weather: versus a physics model, not versus measurement. Data requests in order: metered rooftop output (20-50 consumers, 15-minute or hourly, 12 months), ground GHI (IMD or a state station), MOSDAC.
results/test_monthly.csv
Whole-area P10 to P90 band covers 77.2% of hours (monthly retrain), below the 80% target: the area band is a little narrow.
How it was checked
Read area_test.csv, row new_monthly (coverage 0.772, 8,991 hours).
Limits
Band method is an approximation; error correlation 0.341 measured on 2024 only.
results/area_test.csv
The change of bias before and after October 2025 is caused by the weather-feed switch.
How it was checked
Review notes in the model README: the flip is mostly the 2025 monsoon, not only the feed switch; the cause is not established.
because the cause (weather models, archive source or a wet 2025 monsoon) is not separated
Limits
Do not describe the flip as proven to be the feed switch.
../README.md (review notes)
The newest archive days (from about 6 Nov 2026) are provisional: the archive is back-filled from forecasts.
How it was checked
Stated in the model README: archive within 1-18 W/m2 of the forecast from about 6 Nov. Dropping the last 4 test days changes nothing material (31.30 vs 38.33 vs 39.20). Only the model README states this: no independent check was made here (review fix M16, 18 Nov 2026).
because only the model README states it and no independent check was made here
Limits
Re-download the archive in a few weeks before re-scoring recent days.
../README.md
The 0.954 factor aligns the satellite-based model to PVGIS-ERA5 and was fitted on 2019 to 2021. After the factor the Hyderabad annual model/PVGIS ratio rises from 0.981 (2019) to 1.031 (2023). On the six-point average the model reads +2.0% (2022) and +2.63% (2023) against PVGIS. SME's ratios: 0.986 and 1.043 for Hyderabad; this project found 0.981 and 1.031 (difference stated, not hidden).
How it was checked
Annual sums of pvgis_1kwp and a re-run of the model for 2019 to 2023 (a_items_1_5_8.py).
because the claim carries the SME label [AGENT]: here only its local-file parts are reproduced (the 0.9538 factor fitted on 2019-2021 and the +2.0% / +2.63% 2022-2023 biases: checked - correct, local files only, see E43); the SME's ratios 0.986 and 1.043 are checked - wrong as I read them (0.981 and 1.031 here); 'lowest-yield years of 2005-2023' and 2024-2025 cannot be checked, because only PVGIS 2019-2023 is held
Limits
PVGIS-ERA5 is itself a model; the SME's ratios for 2019 and 2023 (0.986, 1.043) are not reproduced exactly; 2024 and 2025 cannot be checked, because PVGIS data held ends in 2023. SME note on 18 Nov 2026: the fit file is not on disk and the direction is supported by an independent test; in this folder pitch/data/pvmodel_1kwp/calibration_factor.txt (pooled 0.9538, fitted 2019-2021) and bias_check_2019_2023.csv are on disk.
pvgis_1kwp/hyderabad.csv
PVGIS-ERA5, Hyderabad, 17 degree tilt, 14% losses: long-term 1,501 kWh per kWp a year, yearly range 2005-2023 1,418 to 1,548. Roof-mounted (building) setting 1,441.6, about 4% lower: the default free-standing setting is optimistic for rooftops. Global Solar Atlas (satellite, optimal tilt 21 degrees): 1,589.
How it was checked
SME reproduced the PVGIS values and read the Global Solar Atlas value; not re-run here (the 2005-2023 series is not on disk; pvgis_1kwp holds 2019-2023 only, mean 1,456.4, range 1,418.0 to 1,521.1).
(SME label [checked-correct]; not re-checked here; the 2019-2023 part checked here against the local file)
Limits
PVGIS-ERA5 is itself a model. A roof-mounted case is offered on the Forecast page.
data/pvgis_1kwp/hyderabad.csv
Our model's annual yield, Nov 2025 to Oct 2026: Hyderabad 1,579.7 kWh per kWp; mean of the 13 TGSPDCL points 1,604.0 (range 1,558.9 to 1,638.7). That is above the PVGIS band (the 13-point mean about 3% above 1,550) and close to the Global Solar Atlas value (Hyderabad 0.6% below 1,589).
How it was checked
Recomputed here by hosting/yield_check.py: sum of column P_W_per_kWp_calibrated in pitch/data/pvmodel_1kwp/hyderabad.csv (8,760 hours) and of target_w per city in results/test_predictions_monthly.parquet, 1 Nov 2025 to 31 Oct 2026. The audit's 1,604 is the 13-point mean.
(checked here, local files; result in results/annual_yield_check.csv)
Limits
Modelled, not metered. High yield means more export (the cautious side for hosting capacity) but it overstates energy and savings.
results/annual_yield_check.csv
The model uses a 0.954 factor fitted on 2019-2021 (the lowest-yield years); it reads +2.0% high in 2022 and +2.6% high in 2023 against PVGIS on the six-point average.
How it was checked
SME: fit file not on disk, direction supported by an independent test. Here: data/pvmodel_1kwp/calibration_factor.txt (pooled 0.9538) and bias_check_2019_2023.csv are on disk; six-point mean of cal_bias_pct is +2.0 (2022) and +2.63 (2023).
(checked here, local files; the SME's label on direction: supported)
Limits
SME ratios 0.986 and 1.043 are not reproduced (0.981 and 1.031 here); the basis of the SME's ratios is unknown.
data/pvmodel_1kwp/bias_check_2019_2023.csv
Growth9
TGSPDCL net-meter services rose from 25,687 (Jan 2025) to 59,550 (Oct 2026); Oct 2026 is 68% above Oct 2025 (35,397).
How it was checked
Independent pandas sum of totservices by month on net_meter_clean_22_months.csv; 59,550 / 35,397 = 1.682.
Limits
Southern discom only; Aug 2025 is missing for the Gadwal circle and 8 Sangareddy sections (7, plus the one-service section Kangti; marked missing, not interpolated); Kodangal 6 to 789 services in May 2026 is treated as two series; July 2025 duplicates were removed in the clean file.
net_meter_clean_22_months.csv
Whether net-meter 'units' are gross generation, net or export-billed. The SME has not seen the units column, so its meaning is not checked by the SME either.
How it was checked
Units do not follow the seasonal shape of the forecast (correlation -0.16 over 22 months) and units per kW average 53 kWh a month against an expected gross 118-127, so gross generation is not supported; import, export and import minus export all remain open. No negative values (a zero floor would hide net export). Discom question: Is the net-meter units column import, export, or import minus export, in kWh, per month?
because the billing basis of the units is not stated in the open data
Limits
The app shows services (counts), not units, for growth. The units column is not used as installed kWp or generation until the discom confirms; the size test rests on reading the 'load' column as installed solar kW, which is not confirmed. Our own data-test result stands: gross generation is not supported; net import versus export-only is open (sme_review/A7_*).
data.md section 1d
Months-to-full is an estimate shown as a range (for example 28 of 50 kWp: 9.4 to 61.3 months across the slow, median and fast net-meter growth classes, computed in the loader on a billed kW basis; 20.5 to 34.5 months at the three MNRE reference rates). Display rule (project decision made 18 Nov 2026): the numbers stay in the API and the database; any figure above 120 months is shown as 'more than 10 years' and a transformer already at or over its cap or limit is shown as 'already full' (checked before the cap). The growth rate used is shown next to every months-to-full number.
How it was checked
Months = ln(cap / installed) / ln(1 + monthly growth), recomputed for the example transformers independently with pandas, without the loader (backend/tests/test_growth_classes.py: classes and ranges); hosting_saturation.csv is regenerated from the same computed classes by hosting/saturation.py and the loader recomputes the months from them at full precision. The older sme_review/A8_months_to_full_ranges.csv holds the earlier values from the stored classes 0.9 / 2.9 / 6.3 (superseded). One rounding function (one decimal, display only) serves every page. Since 19 Nov 2026 the stored months-to-full table on the Growth page uses the Indian street model: installed and cap from results/indian_months_to_full.csv, safe limit = the Indian P5 at 11 kV 1.00 pu (results/indian_summary.csv), months recomputed by the loader from the same growth classes.
because no per-transformer solar counts are held
Limits
Growth of services in a section is not growth of kWp on one transformer and is not an acceptable per-transformer proxy. Thresholds Act now under 2 months and Watch under 6 months are PROPOSED - NOT TESTED (set from the discom's real upgrade lead time). Show a range, not one number. MNRE reference rates, recomputed on 18 Nov 2026 with the true table date (31 Jan 2026): 1.69% a month (31 Jan to 30 Sep, 8 months), 1.82% (31 Jan to 31 Oct, 9 months), 2.87% (30 Sep to 31 Oct, 1 month), 1.96% (31 Dec 2025 to 31 Jan 2026, 1 month); derived and conditional on the SME-note values (source tables not held). Roofs up to 10 kWp skip the feasibility study (rule 11(7A), SME answer of 18 Nov 2026; official text not re-read here), so a transformer can fill without anyone checking (see E44).
results/indian_months_to_full.csv
Kodangal (Vikarabad circle) goes from 6 services in Apr 2026 to 789 in May 2026 and 1,498 in Oct 2026. The other 19 Vikarabad sections add 17 services in total that month and no section in the file loses more than 1 service, so it is not a move between sections of this file; the 783 added services average 0.34 kW of billed load each against about 6 kW elsewhere.
How it was checked
Section totals by month from net_meter_clean_22_months.csv (a_netmeter_tests.py).
Limits
Cause (reorganisation, re-coding, new population) cannot be checked yet, because the discom has not been asked. Kodangal is treated as two series; no growth across the break.
net_meter_clean_22_months.csv
The Aug 2025 file has no rows for the Gadwal circle (13 sections, about 186 services) and 8 Sangareddy sections: 7 that are present in July and Sep 2025 (73 services in July) plus the one-service section Kangti (no rows from Jul 2025 to Aug 2026). Aug 2025 total services (33,272) are about 0.55% low.
How it was checked
Circle and section presence by month; Gadwal 182 (Jul) and 189 (Sep) services. Sangareddy sections counted again on 18 Nov 2026 straight from data/TGSPDCL/*.csv: 33 sections in July and September 2025, 26 in August; the 7 missing are Zaheerabad Town and Rural, Koheer, Mannapur, Nyalkal, Jharasangam, Raikode; Kangti (1 service) has rows in Jan to Jun 2025 and from Sep 2026 only, so it is also absent in Aug 2025 (8 in all). Earlier notes said 7: that count left Kangti out.
Limits
Marked missing, excluded from growth rates, not interpolated. Ask the discom for the missing month.
net_meter_clean_22_months.csv
MNRE Telangana rooftop: 682.5 (31 Dec 2025), 695.9 (31 Jan 2026), 795.90 (30 Sep 2026), 818.76 (31 Oct 2026). The 695.9 table is headed 'as on 31.01.2026' (read in the original PDF). On those dates the compound rates are 1.96% (Dec to Jan, 1 month), 1.69% (Jan to Sep, 8 months), 1.82% (Jan to Oct, 9 months) and 2.87% (Sep to Oct, 1 month) a month.
How it was checked
Rates: arithmetic with exact month counts from the true date of 695.9 (sme_review/a_mnre_corrected.py). 682.5, 795.90 and 818.76 are from the SME note, source table not held. The date of 695.9 is item E24.
because the SME-note values 682.5, 795.90 and 818.76 are not on disk, so the rates that use them are derived and conditional
Limits
With the true date growth looks fairly steady at about 1.7 to 2.0% a month, so the SME's earlier reading of a recent speed-up (which assumed 31 Mar) is not supported; the median early-warning class (3.0% a month, computed) is above these multi-month rates and close to the latest single month only (2.87%). SME's earlier rates 2.35 / 2.26 / 0.65 assumed 31 Mar and are superseded (decision of 18 Nov 2026).
telangana_reports/mnre_2026_02.pdf
MNRE table headed as on 31.01.2026 (read in the original PDF): Telangana rooftop 695.9 MW, total 5,065.10 MW.
How it was checked
pdftotext -layout on telangana_reports/mnre_2026_02.pdf on 18 Nov 2026: heading 'State-wise (Location based) installed capacity of Renewable Power as on 31.01.2026', Telangana row rooftop 695.9, total 5,065.10 (4,360.49 + 695.9 + 8.71). Decision of the SME on 18 Nov 2026: 31.01.2026 is correct.
Limits
The SME's reading '31 Mar 2026' is not supported by the table we hold. A state total, not per transformer. The SME-note values for other dates are not held (E19).
telangana_reports/mnre_2026_02.pdf
Roofs up to 10 kWp skip the feasibility study (rule 11(7A)), so a transformer can fill without anyone checking. Telangana single-phase supply goes up to 15 kW (GTCS amended to 2016), while the 5 kW / 75 kW table is only the 2016 net-metering guideline.
How it was checked
SME answer, item 6. Label as the SME gives it: supported (earlier notes had it at AGENT level). Rule and GTCS texts not held here.
because the rule and GTCS texts are not held here (SME label: supported)
Limits
Months to full above 120 is shown as 'more than 10 years' (display rule, project decision 18 Nov 2026).
SME answer, item 6
Growth classes (P10 / P50 / P90 across sections of the compound monthly growth, first to last month, Jan 2025 to Oct 2026) are now COMPUTED in the loader from the net-meter table: billed kW of the kW categories 0.95 / 3.04 / 6.37% a month (440 sections; used for months to full); second basis, services 1.62 / 3.77 / 8.40% (441 sections); comparison, billed load of all categories as recorded (adds kW and hp) 0.91 / 2.98 / 6.43% (441 sections; this is the reviewer's recompute of 0.90 / 2.98 / 6.43 on 443 sections including Kodangal). Rules: Aug 2025 missing, not interpolated (only the two end months are used); Kodangal left out (two series); flagged-step sections included in the primary classes and reported without them as well (0.93 / 3.02 / 6.31). Classes now computed in the loader on billed kW basis (18 Nov 2026); previous stored values 0.93/2.91/6.26 came from a JSON not held.
How it was checked
Computed by hosting/growth_classes.py (called by scripts/load_data.py, table growth_class with the definition) and recomputed independently with plain pandas in backend/tests/test_growth_classes.py (18 cases agree to 1e-9).
(computed and independently recomputed from the local file); the earlier stored values 0.93 / 2.91 / 6.26: cannot be checked yet, because their source JSON (saturation_growth_rates.json) is not held, and they are no longer used
Limits
Section growth is not a per-transformer rate. Services growth (+68% in a year, 1.62 / 3.77 / 8.40% a month by section) and billed kW growth (0.95 / 3.04 / 6.37% a month) are different quantities: only billed kW feeds months to full. Choosing the all-category load instead of the kW categories changes the 28-of-50 range from 9.4 to 61.3 to 9.3 to 64.1 months; the kW-only basis was chosen because hp and kW cannot be added.
TGSPDCL_summary/net_meter_clean_22_months.csv
Feeder view1
Transformer status lights use the working rule accepted by the SME on 18 Nov 2026; not validated with the discom: green above 25% of cap left, amber 5 to 25%, red below, grey if there is no phase data. Installed solar is roofs today times 3 kWp (assumed roof size, an estimate). Months to cap above 120 are shown as 'more than 10 years'; the growth rate used is shown beside them.
How it was checked
Thresholds from the pitch draft; accepted by the SME as the working rule on 18 Nov 2026; roofs and caps from results/feeder_demo.json.
because the thresholds and the roofs-times-3-kWp estimate are a working rule accepted by the SME, not validated with the discom, and the 20 transformers are illustrative
Limits
Illustrative layout; the discom GIS replaces it in the pilot. Discom validation of the thresholds and of the installed-solar estimate is still needed. The early-warning limits (Act now / Watch) remain PROPOSED - NOT TESTED.
results/feeder_demo.json
Live forecast2
A live day-ahead forecast is available from this app.
How it was checked
Wired on 18 Nov 2026: the backend calls Open-Meteo (read-only, no keys), runs the shipped model files in product/artifacts and falls back to the stored example if the call fails. Each result says source = live or stored.
(the live path exists and ran once on 18 Nov 2026; accuracy is a separate question, see E23)
Limits
Live forecast wired 18 Nov 2026. Accuracy not yet measured against real output (no clean live evidence until days after 9 Nov 2026 are final). Result is an estimate versus a physics model, not versus measurement. When the call fails the app says so and shows the stored example.
product/live.py
Live forecast wired 18 Nov 2026; accuracy not yet measured against real output (no clean live evidence until days after 9 Nov 2026 are final).
How it was checked
Backend integration with Open-Meteo plus the shipped model files; parity with the research code checked on a recorded response; one real call on 18 Nov 2026 checked against clear-sky and nameplate bounds (see BUILD_LOG.md).
because there is no clean live evidence to score it against until days after 9 Nov 2026 are final and no metered rooftop output is held
Limits
Estimate, versus a physics model, not versus measurement. The first day is the day-ahead the model was trained for; the second day is further ahead and less reliable.
product/live.py
Rules6
Voltage at the point where supply begins: LV +6%/-6%, HV +6%/-9%, EHV +10%/-12.5%; unbalance not above 3% (TSERC SoP Reg 5 of 2016, Part II cl 2.1(i) and 1.8).
How it was checked
Read by the SME from the 2016 gazette page images.
because the 2016 gazette is not held here and a later amendment cannot be ruled out (negative search)
Limits
Consumer-end limits, not transformer limits. +/-5% is a synchronisation-step limit (guideline text: fluctuation not greater than +/-5% at the point of connection), not steady state. Later, on 18 Nov 2026, the SME labelled the LV limit +6%/-6% (TSERC SoP Reg 5 of 2016) as [checked-correct] (see E39); the HV and EHV values and the 3% unbalance limit keep the label above.
SME review, 18 Nov 2026
Phase-managed connection and export caps are things the discom may require under the technical-feasibility step. No Telangana clause for either was found in the 2025 texts.
How it was checked
Text search of the TGSPDCL guideline held here (the word phase appears once, for the metering arrangement); the regulation text is not held.
because a negative search is not proof and the regulation text is not held here
Limits
Our proposed remedy, not a rule. Never write that the regulation requires it.
telangana_reports/TGSPDCL_Rooftop_Net_Metering_Guidelines_2025.pdf
TGERC Regulation 1 of 2025 cl 5.1 and TGSPDCL guideline IV(a) both say 'the sum of the rated capacities of rooftop solar on a DTR shall not exceed 50% of its rated capacity'; neither gives a unit.
How it was checked
Official texts read by the SME on 18 Nov 2026 (not re-read here: no gazette or live site access).
(SME label [checked-correct]: official text read by the SME; not re-checked here)
Limits
What the 50% is of (kVA of the DTR) and the unit of the result (kW, kWp, kVA) are not stated. Our reading is in E26.
SME answer, item 1
A DTR rating is in kVA; rooftop sizes are kWp elsewhere; no power-factor or DC/AC rule is stated. Working reading in the app: cap = 0.5 x DTR kVA treated as kW; shown on screen as 'unit assumed: kW on the DTR kVA rating; TGSPDCL to confirm'. Limits apply to AC export, not panel DC size.
How it was checked
SME reading, not a stated rule; the app keeps R1 as the labelled working reading and still shows R2 to R4.
because this is a reading, not a stated rule
Limits
TGSPDCL feasibility reply, TGERC clarification or RTI to the Chief Engineer (IPC & RAC) would settle it.
results/hosting_cap_unit_sensitivity.csv
The 2016 rule was 50% of the DTR kVA x 0.9, taken as kW. The app offers this as an optional 0.9 case (toggle on the street page, precomputed rows R2 and R4).
How it was checked
Library note given by the SME; the 2016 gazette is not held here.
because it is a library note and the 2016 text is not held here
Limits
Sensitivity, not law.
results/hosting_cap_unit_sensitivity.csv
Telangana LV voltage limit is +6%/-6% (TSERC SoP Reg 5 of 2016). The street 'Safe' screen tests the +6% side, loading and unbalance; a transformer tap at 1.03 pu already uses part of the margin; a transformer or cable can overload even if voltage passes.
How it was checked
SME answer, item 4 (gazette read by the SME). Our code: hosting/lv.py check() tests voltage at +6% (1.06 pu), transformer loading at most 100 per cent, cable loading at most 100 per cent, unbalance 3%.
(SME label [checked-correct] for the +/-6% limit; not re-checked here). That lv.py tests four limits: checked here, read from the code
Limits
The -6% side, reverse flow and the evening peak are not tested; the street page shows them as 'not tested'. The badge reads 'Passes screening tests', not 'voltage only', because the code also tests loading and unbalance.
hosting/lv.py
Network data6
Public data cannot show safe hosting capacity: it has no transformer kVA, no consumers per transformer and no 15-minute load.
How it was checked
SME answer, item 2.
(SME label [checked-correct]; not re-checked here)
Limits
Needed from the department: a DTR list with kVA and consumers, 15-minute DTR and feeder metered load, and one real service-number lookup result.
SME answer, item 2
TGSPDCL has a page 'NetMetering Capacity on Distribution Transformer' that takes ONE service number and returns that consumer's DTR. It is a lookup, not a public list of transformers.
How it was checked
Form seen and checked by the SME.
(SME label [checked-correct] for the form; not re-checked here)
Limits
Wording to use: 'lookup exists, no public DTR list found'. Never write 'not published'.
SME answer, item 2
What the TGSPDCL lookup returns for a real service number (the DTR kVA, load, room left).
How it was checked
The SME has never seen the output.
because nobody has seen the output of the lookup
Limits
One real service-number lookup result is needed from the department.
SME answer, item 2
TGNPDCL: no DTR-capacity page was found. Regulation cl 5.2 requires capacity to be published per DTR and feeder.
How it was checked
Negative search by the SME; the regulation clause as read by the SME.
because a negative search is not proof
Limits
Wording: 'lookup exists (TGSPDCL), no public DTR list found'; never 'not published'.
SME answer, item 2
Distribution transformer failure in Telangana is 11 to 13% a year (2015-17) and 15 to 17% in rural circles (Prayas, from discom filings). It is driven by agriculture, not by rooftop solar.
How it was checked
Secondary source as given by the SME.
because only a secondary source (Prayas, from discom filings); SME label [secondary]
Limits
Years 2015-17 only.
SME answer, item 2
Upgrade lead times: nothing on file. Rule 11(8) of the central consumer rules: strengthening for up to 5 kW is paid from the discom's revenue requirement.
How it was checked
SME answer, item 2; no label given for rule 11(8); official text not re-read here.
because no lead-time data is held and the rule text was not re-read here
Limits
Ask the discom for real lead times; they set the 'Act now' and 'Watch' thresholds (PROPOSED - NOT TESTED).
SME answer, item 2
Real data8
The 'Real data we hold' panel lists each real dataset with its period, row count, coverage, source file and last date, all computed by the loader from the files (TGSPDCL net-meter records, Open-Meteo weather actuals and day-before forecasts for 13 points, the MNRE table, quoted PM Surya Ghar figure, PVGIS reference, live forecast feed, guideline texts).
How it was checked
Row counts, first and last dates and point counts recomputed with pandas from the source files, independently of the loader (table in CODE_REVIEW_FIXES.md).
(checked here, local files)
Limits
Counts describe the files on this machine. Weather actuals are satellite/reanalysis-based, not ground-measured; the newest days are provisional (E11).
scripts/load_data.py (real_dataset)
The Forecast page compares the real Open-Meteo day-before forecast of sunlight (GHI) with the real Open-Meteo actuals for a chosen TGSPDCL point, last 14 days (daily totals and hourly), up to the last day held (9 Nov 2026). It is a weather-model forecast against satellite/reanalysis-based actuals, not metered solar output.
How it was checked
Daily totals (sum of 24 hourly W/m2 / 1000) and the 14-day mean difference recomputed with pandas from irradiance_openmeteo_tgspdcl/<point>.csv and prev_day1_tgspdcl/<point>.csv for the points listed in CODE_REVIEW_FIXES.md.
(checked here, local files)
Limits
Says how good the weather forecast was. It says nothing about the accuracy of the power forecast against real roofs (E08). Actuals are not a ground station.
irradiance_openmeteo_tgspdcl/*.csv, prev_day1_tgspdcl/*.csv
Every headline number on the Home page and in the Evidence summary table carries a tag: REAL (measured or published data we hold), MODEL (our model or a published model reference) or EXAMPLE (assumed values).
How it was checked
Read from the code and covered by backend tests (every overview tile has a tag, real tiles first; the summary table lists value, tag and source file).
(checked here, read from the code and tests)
Limits
The tag says what kind of number it is, not that it is right. EXAMPLE values are assumed typical values, never discom data.
backend/app/services/overview/overview_service.py
Not held: real transformer data (kVA, consumers, phase, 15-minute load), metered rooftop solar output, ground-measured sunlight, hourly net-meter data, TGNPDCL net-meter data.
How it was checked
Searched the project data and results folders for such files; none exists (see data/data.md section 7).
because a negative search of this folder is not proof that the discom holds no such data; we hold none
Limits
Each one is a request to the discom or the department (see the open questions).
data/data.md
The real billed load per domestic net-meter service is 5.91 kW in Oct 2026 (316,880.5 kW over 53,604 services), against the 3 kWp roof assumed in the street and feeder models.
How it was checked
Arithmetic on the Oct 2026 domestic rows of net_meter_clean_22_months.csv (recomputed with pandas); the unit of the domestic load is kW per TGSPDCL_summary/monthly_by_category.csv.
because the load column is not documented, so it is not confirmed that billed load equals installed solar kW (the arithmetic itself: checked - correct, local file)
Limits
Do not read 5.91 kW as the typical roof size. It is shown beside the assumed 3 kWp only so a reader sees the difference.
TGSPDCL_summary/net_meter_clean_22_months.csv
PM Surya Ghar in Telangana: 54,668 installations and 70,144 households at 27 Jul 2026 (quoted from Lok Sabha Q 1698).
How it was checked
Quoted from docs/Challenge16_DER_Final.md, where the SME reviewer marked it checked - correct; the Lok Sabha source file is not held here.
because the Lok Sabha source file is not on disk (SME reviewer label: checked - correct, not re-read here)
Limits
A state count of installations, no MW per state; not per circle or transformer.
docs/Challenge16_DER_Final.md
Nowcast backtest (next 1 to 3 hours), 1 Jan to 9 Nov 2026, days not used for fitting, 13 TGSPDCL points, hourly: mean error 22.5 W per kWp at 1 h ahead against 28.91 for the day-ahead forecast (22.2% lower); 27.08 at 2 h (6.3% lower); 28.46 at 3 h (1.6% lower). Persistence (same as the latest hour) is 98.07 at 1 h.
How it was checked
Read from nowcast/results/backtest.csv (nowcast/backtest.py: fade weights 0.65, 0.40, 0.25 fitted on 2025, scored once on 2026, 53,805 rows per lead) on 19 Nov 2026; percentages recomputed from the error columns. Scored versus the archive physics value, not measured output.
(backtest figures recomputed from the CSV; versus a physics model, not versus measurement)
Limits
MODEL, not measurement: the hourly archive is smoother than a real meter, so real gains may differ. The gain shrinks with lead time (22.2% at 1 h, 1.6% at 3 h). See E54 for live and 15-minute use.
nowcast/results/backtest.csv
Live nowcast and 15-minute steps: not built as a live feature. The method exists (nowcast/run_live.py takes the last hour's metered kW), but no live observation feed is held and no sub-hourly data is held.
How it was checked
Scope statement, 19 Nov 2026: the project holds hourly Open-Meteo data only; the satellite feed returns no data for Telangana; Open-Meteo's 15-minute values for India are interpolated from hourly data.
because no live observation feed (smart meters, SCADA or metered rooftops) and no sub-hourly data is held
Limits
Do not present live accuracy or 15-minute accuracy. In the pilot the last hour's reading would come from 20 to 50 metered consumers or SCADA; the live gain must be measured then.
nowcast/README.md
REAL measured or published data we hold MODEL our model or a published model reference EXAMPLE assumed typical values. Read from the same tables the pages read.
| Headline number | Value | Tag | Source file | Evidence | Shown on |
|---|---|---|---|---|---|
| Net-meter services, 2026-10 | 59,550 | REAL | TGSPDCL_summary/net_meter_clean_22_months.csv | E12 | Home, Growth |
| Net-meter services, 2025-10 | 35,397 | REAL | TGSPDCL_summary/net_meter_clean_22_months.csv | E12 | Growth |
| Services growth over the year (with Kodangal) | +68.2% | REAL | TGSPDCL_summary/net_meter_clean_22_months.csv | E12 | Home, Growth |
| Services growth over the year (Kodangal left out) | +64.0% | REAL | TGSPDCL_summary/net_meter_clean_22_months.csv | E17 | Growth |
| Billed load, kW categories | 422.9 MW | REAL | TGSPDCL_summary/net_meter_clean_22_months.csv | E13 | Home, Growth, Street |
| Average billed load per domestic service | 5.91 kW | REAL | TGSPDCL_summary/net_meter_clean_22_months.csv | E50 | Street, Growth |
| Mean services per section office | 125.9 | REAL | TGSPDCL_summary/net_meter_clean_22_months.csv | E12 | Feeder, Growth |
| Telangana rooftop solar (MNRE, as on 31.01.2026) | 695.9 MW | REAL | telangana_reports/mnre_2026_02.pdf | E24 | Home, Street, Growth |
| Telangana solar total (MNRE, as on 31.01.2026) | 5,065.10 MW | REAL | telangana_reports/mnre_2026_02.pdf | E24 | Home |
| PM Surya Ghar installations, Telangana (quoted) | 54,668 | REAL | docs/Challenge16_DER_Final.md (quotes Lok Sabha Q 1698) | E51 | Home |
| Real weather actuals, hourly rows per point | 68,880 (2019-01-01 to 2026-11-09) | REAL | irradiance_openmeteo_tgspdcl/*.csv | E46 | Home, Forecast |
| Real day-before forecasts, non-empty hourly rows per point | 24,607 (from 2024-01-19) | REAL | prev_day1_tgspdcl/*.csv | E46 | Home, Forecast |
| Forecast error, our model vs raw physics | 31.2 vs 39.1 W/kWp | MODEL | results/test_monthly.csv | E07 | Home, Forecast |
| Annual yield, PVGIS long-term reference, Hyderabad | 1,501 kWh/kWp | MODEL | results/annual_yield_check.csv (SME value, not re-run) | E41 | Forecast |
| Annual yield, our model, Hyderabad (Nov 2025 to Oct 2026) | 1,579.7 kWh/kWp | MODEL | results/annual_yield_check.csv | E42 | Forecast |
| Town street (100 kVA, ACSR Rabbit), phase unknown, P5 | 8.9 kWp (90% range 6.0 to 9.0; 50% cap 50.0) | EXAMPLE | results/indian_summary.csv | E55 | Home, Street |
| Village street (63 kVA, ACSR Weasel), phase-managed, P5 | 0.0 kWp (90% range 0.0 to 0.0; 50% cap 31.5) | EXAMPLE | results/indian_summary.csv | E55 | Home, Street |
| Transformers on the illustrative feeder | 20 | EXAMPLE | results/feeder_demo.json | E15 | Home, Feeder |
| PM-KUSUM plant by position (assumed 15 km feeder) | 1.2 to 4.75 MWp | EXAMPLE | results/hosting_kusum.csv | E06 | Home, KUSUM |
| Months to full, 28 of 50 kWp installed (assumed), growth classes | 9.4 to 61.3 | EXAMPLE | TGSPDCL_summary/net_meter_clean_22_months.csv (growth classes computed by the loader) | E14 | Growth |
| Months to full at the MNRE state rates (derived, conditional) | 20.5 to 34.5 | MODEL | telangana_reports/mnre_2026_02.pdf (one value) + SME note (three values, tables not held) | E19 | Growth |
| Table | Source file | Rows | Loaded at (UTC) |
|---|---|---|---|
| annual_yield | results/annual_yield_check.csv | 10 | 2026-11-20T05:58:18+00:00 |
| area_band | results/area_test.csv | 4 | 2026-11-20T05:58:18+00:00 |
| cap_unit_sensitivity | results/hosting_cap_unit_sensitivity.csv | 24 | 2026-11-20T05:58:18+00:00 |
| circle_point | data/irradiance_openmeteo_tgspdcl/README.md | 20 | 2026-11-20T05:58:33+00:00 |
| evidence_item | scripts/evidence_items.json | 56 | 2026-11-20T05:58:31+00:00 |
| feeder_dtr | results/feeder_demo.json | 20 | 2026-11-20T05:58:18+00:00 |
| feeder_meta | results/feeder_demo.json | 3 | 2026-11-20T05:58:18+00:00 |
| forecast_accuracy[monthly] | results/test_monthly.csv | 12 | 2026-11-20T05:58:18+00:00 |
| forecast_accuracy[static] | results/test_static.csv | 12 | 2026-11-20T05:58:18+00:00 |
| forecast_by_point[monthly] | results/test_by_point_monthly.csv | 39 | 2026-11-20T05:58:18+00:00 |
| forecast_by_point[static] | results/test_by_point_static.csv | 39 | 2026-11-20T05:58:18+00:00 |
| forecast_example | results/live_area_example.txt | 24 | 2026-11-20T05:58:18+00:00 |
| forecast_hourly | results/test_predictions_monthly.parquet, results/test_predictions_static.parquet | 231438 | 2026-11-20T05:58:33+00:00 |
| growth_class | TGSPDCL_summary/net_meter_clean_22_months.csv (computed by hosting/growth_classes.py) | 18 | 2026-11-20T05:58:18+00:00 |
| hosting_bootstrap | results/hosting_p5_bootstrap_ci.csv | 9 | 2026-11-20T05:58:18+00:00 |
| hosting_runs[indian] | results/indian_runs.csv | 4200 | 2026-11-20T05:58:18+00:00 |
| hosting_runs[load025] | results/hosting_lv_runs.csv | 540 | 2026-11-20T05:58:18+00:00 |
| hosting_runs[load050] | results/hosting_lv_load050_runs.csv | 540 | 2026-11-20T05:58:18+00:00 |
| hosting_summary[indian] | results/indian_summary.csv | 21 | 2026-11-20T05:58:18+00:00 |
| hosting_summary[load025] | results/hosting_lv_summary.csv | 18 | 2026-11-20T05:58:18+00:00 |
| hosting_summary[load050] | results/hosting_lv_load050_summary.csv | 18 | 2026-11-20T05:58:18+00:00 |
| kusum_cap_sensitivity | results/hosting_kusum_cap_sensitivity.csv | 16 | 2026-11-20T05:58:18+00:00 |
| kusum_grid | results/hosting_kusum.csv | 16 | 2026-11-20T05:58:18+00:00 |
| net_meter_category_month | TGSPDCL_summary/net_meter_clean_22_months.csv | 163 | 2026-11-20T05:58:18+00:00 |
| net_meter_gap | TGSPDCL_summary/net_meter_clean_22_months.csv (derived) | 23 | 2026-11-20T05:58:19+00:00 |
| net_meter_section_month | TGSPDCL_summary/net_meter_clean_22_months.csv | 10099 | 2026-11-20T05:58:18+00:00 |
| net_meter_series_break | TGSPDCL_summary/net_meter_clean_22_months.csv (derived) | 7 | 2026-11-20T05:58:19+00:00 |
| nowcast_backtest | nowcast/results/backtest.csv | 6 | 2026-11-20T05:58:33+00:00 |
| real_dataset | computed from the files above | 8 | 2026-11-20T05:58:31+00:00 |
| real_fact | telangana_reports/mnre_2026_02.pdf; docs/Challenge16_DER_Final.md | 3 | 2026-11-20T05:58:31+00:00 |
| real_not_held | scripts/load_data.py (NOT_HELD) | 5 | 2026-11-20T05:58:31+00:00 |
| saturation | results/indian_months_to_full.csv (safe limit: indian_summary.csv P5 at 1.00 pu) | 21 | 2026-11-20T05:58:18+00:00 |
| weather_coverage | irradiance_openmeteo_tgspdcl/*.csv, prev_day1_tgspdcl/*.csv | 13 | 2026-11-20T05:58:31+00:00 |
| weather_daily | same files, every day, daily totals | 37310 | 2026-11-20T05:58:31+00:00 |
| weather_hourly | same files, last 14 days, hourly | 4368 | 2026-11-20T05:58:31+00:00 |
| weather_point | data/irradiance_openmeteo_tgspdcl/README.md | 13 | 2026-11-20T05:58:33+00:00 |
REAL = measured or published data held in this project. Weather actuals are satellite/reanalysis-based sunlight, not metered solar output. Our model, or a published model reference (PVGIS), run on real inputs. Not measured. Assumed typical values (street lengths, cables, houses, 3 kWp roofs, load per home). The pilot replaces them with the discom's data.
| Dataset | Tag | Period | Rows | Coverage | Last date | Source file | Can show | Cannot show |
|---|---|---|---|---|---|---|---|---|
| TGSPDCL rooftop net-meter records, monthly | REAL | 2025-01 to 2026-10 | 146,665 | 22 monthly files; 20 circles and 473 section offices (circle + section name) in 2026-10; 59,550 services in 2026-10; July 2025 de-duplicated; Aug 2025 has no Gadwal circle (kept missing) | Oct 2026 | data/TGSPDCL/ (22 monthly files, Open Data Telangana) -> data/TGSPDCL_summary/net_meter_clean_22_months.csv | Real billed services, billed kW (load column) and units, by month, circle and section office. | Per-transformer counts, hourly values, what 'units' means (import, export or net). |
| Open-Meteo weather actuals, 13 TGSPDCL points (hourly) | REAL | 2019-01-01 to 2026-11-09 | 895,440 | 13 points, 68,880 hourly rows each; satellite/reanalysis-based archive | 9 Nov 2026 | data/irradiance_openmeteo_tgspdcl/*.csv (Open-Meteo, CC BY 4.0) | Real sunlight (GHI, direct, diffuse), temperature, cloud, wind by hour for the area. | Ground-station measurements, metered solar output, the newest days are provisional (back-filled). |
| Open-Meteo day-before forecasts, 13 TGSPDCL points (hourly) | REAL | 2024-01-19 to 2026-11-09 | 319,891 | 13 points, 24,607 non-empty hourly rows each of 25,056 (the empty hours are before the first real forecast) | 9 Nov 2026 | data/prev_day1_tgspdcl/*.csv (Open-Meteo previous-runs API, CC BY 4.0) | What the weather model forecast one day before, to compare with the actuals above. | Metered solar output; it is a weather forecast, not a solar-power forecast. |
| MNRE state-wise renewable capacity table | REAL | 31.01.2026 | 1 | one Telangana row: rooftop 695.9 MW, solar total 5,065.10 MW | 31.01.2026 | data/telangana_reports/mnre_2026_02.pdf | State-level rooftop and total solar capacity on one date. | A time series (the other dates in the SME note are not held), anything per district, circle or transformer. |
| PM Surya Ghar installations, Telangana (quoted) | REAL | 27 Jul 2026 | 1 | 54,668 installations, 70,144 households (quoted from Lok Sabha Q 1698) | 27 Jul 2026 | docs/Challenge16_DER_Final.md (source file not held) | A state-level count of subsidised rooftop installations. | MW per state, per circle or per transformer. The source file is not held, so the figure cannot be checked yet. |
| PVGIS 1 kWp hourly output (published model reference) | MODEL | not a dated file | n/a | source files not found on this machine | n/a | data/pvgis_1kwp/*.csv | A published model reference for 1 kWp output. | Measured output. |
| Open-Meteo live forecast (called when you ask, LIVE badge) | REAL | not a dated file | n/a | 13 points, next day and the day after; falls back to the stored example when the call fails and says so | n/a | Open-Meteo API (CC BY 4.0) via backend/app/integrations/open_meteo/ | A real weather forecast fed into our shipped model. | The power estimate is modelled: it is not scored against metered output. |
| TGSPDCL guideline and agreement texts, CSTEP feeder report, MNRE table (PDFs) | REAL | not a dated file | n/a | source folder not found on this machine | n/a | data/telangana_reports/*.pdf | The published rules: 50% DTR cap wording, 56 kWp self-certification threshold, the CSTEP feeder method. | The TGERC regulation text and the 2016 gazette are not held; the unit of the 50% cap is not stated in either text. |
- Real transformer data (kVA, consumers, phase, 15-minute DTR or feeder load). Not public; the DTR list has to come from the department. Every network value here is assumed.
- Metered rooftop solar output. Not public and not downloadable (PVGIS stops at 2023; no metered source was found). Forecast accuracy is versus a physics model, not versus measurement.
- Ground-measured sunlight (GHI). No ground station file is held. The Open-Meteo actuals are satellite/reanalysis-based; NIWE or IMD stations would be the check.
- Hourly or 15-minute net-meter / smart-meter data. Only monthly counts, billed kW and units are public. The meaning of the 'units' column (import, export or net) is not documented.
- TGNPDCL (northern discom) net-meter data. No TGNPDCL file was found; the net-meter data covers TGSPDCL only.
- REAL Telangana rooftop solar, MNRE state table: 695.9 MW (as on 31.01.2026; telangana_reports/mnre_2026_02.pdf)Table headed 'State-wise (Location based) installed capacity of Renewable Power as on 31.01.2026', Telangana row. State total, not per transformer.
- REAL Telangana solar power total, MNRE state table: 5,065.1 MW (as on 31.01.2026; telangana_reports/mnre_2026_02.pdf)Same table, 'Solar Power Total' column (ground mounted 4,360.49 + rooftop 695.9 + the third solar column 8.71).
- REAL PM Surya Ghar installations, Telangana: 54,668 installations (as on 27 Jul 2026; docs/Challenge16_DER_Final.md (quotes Lok Sabha Q 1698))Quoted from Lok Sabha Q 1698; the source file is not held, so it cannot be checked yet (the expert review found it correct). 70,144 households in the same answer; no MW per state found.
A traffic light per transformer: reason, room left, early warning. Network values are assumed examples until the discom supplies its own data. Screening only, never approved capacity.
Every claim in this app, how it was checked, and what it cannot tell you. Labels: checked - correct, checked - wrong, cannot be checked yet, because ...
Limit on the street results (E01 to E03 generic sensitivity; E55 Indian headline). P5 is a screening value and can be optimistic. Checks run on the generic 400 V model (not repeated on the Indian model): in 91 of 540 roof-by-roof trials "more roofs is never better" fails, and OpenDSS gives 69.0 vs 73.35 kWp (urban) and 16.35 vs 18.0 kWp (rural), both unbalanced. Treat it as a range, not a guarantee.
P5 as defined here (Passes screening tests). P5 = 5th percentile over 200 random trials (Indian street model, the headline; 30 random trials on the generic 400 V sensitivity) of the largest rooftop solar (3 kWp roofs, panel kWp) that passes the screening tests at the sunniest hour (voltage rise +6%, transformer 100 per cent of rating, cable 100 per cent, unbalance 3%); each trial redraws which homes adopt and (except the balanced and three-phase cases) each home's phase. Each trial value is the largest number of 3 kWp roofs that passes, so it is a multiple of 3 kWp; the 5th percentile interpolates linearly between the sorted values (numpy), so the printed P5 can fall between 3 kWp steps (for example 34.35). P50 is the median of the same trials. Screening value, not a guarantee.
Label notes: where an item says SME label, the label is the SME's as given: [checked-correct] = official text read by the SME, [secondary] = news or think-tank, [cannot be checked] = no source. The project could not re-read any gazette or live site, so none of these is upgraded to checked here.