Note
Go to the end to download the full example code.
Solar PV device reliability: evidence and access#
Review 14 sources and load 371 PV-specific records offline. Counts, rates, availability, survival and sensor drift remain separate. Existing Sandia PVROM fits are available through the original failure-database connector.
See the PV reliability research.

source_id access
baschel_2018 Public table; raw juwi O&M records confidential
lillo_2018 Public tables; 15 plants, 15 months
bucher_2022 Public survival summaries; individual histories not provided
martins_2026 Public paper; plant SCADA confidential
ruth_2018 Public aggregate benchmark
iea_trackers_2024 Public expert report and owner/operator survey covering 13.4 GW
brown_2026 Public observations; raw experimental data on request
deline_2026 Public slides and selected PVDAQ series; fleet data NDA-protected
lindig_2023 Published taxonomy and aggregate research; underlying ticket database not downloaded
gallardo_2019 Publisher abstract accessible; institutional full text restricted
jordan_2020 Public aggregate study
stanchev_2026 Public single-inverter case study
pvrom_2018 Existing packaged connector: energy_data_retrieval.failure_database
pvrom_2020 Public report; underlying 800+ system records not public
source_id device metric unit
baschel_2018 String monitoring unit failure_rate 1/operating_hour
lillo_2018 Junction / combiner box event_count events/15_months
lillo_2018 Monitoring system event_count events/15_months
lillo_2018 Monitoring system normalized_event_rate events/plant_year
ruth_2018 Single-axis tracker availability fraction
martins_2026 Single-axis tracker availability fraction
deline_2026 Single-axis tracker stuck_row_prevalence fraction
import matplotlib.pyplot as plt
from energy_data_retrieval.pv_reliability_database import (
load_pv_reliability_evidence,
pv_reliability_sources,
)
data = load_pv_reliability_evidence()
print(pv_reliability_sources()[["source_id", "access"]].to_string(index=False))
print(
data[data.device.str.contains("combiner|monitoring|tracker", case=False)][
["source_id", "device", "metric", "unit"]
]
.drop_duplicates()
.to_string(index=False)
)
counts = data.groupby("metric").size().sort_values()
ax = counts.plot.barh(figsize=(10, 5), color="#117b88")
ax.set(title="371 PV evidence records: metric type matters", xlabel="Records", ylabel="")
for index, count in enumerate(counts):
ax.text(count + 1, index, str(count), va="center")
plt.tight_layout()
plt.show()
Total running time of the script: (0 minutes 0.089 seconds)