Oxford, NASA PCoE, and NREL datasets. These reference studies show how the methods behave on public data before they are used in scoped client work.
ICA and DVA electrochemical fingerprinting used to detect lithium plating, SEI growth, and loss of active material from raw BMS CSV exports.
7,000+
Cycles Analyzed
Energy & Utilities
dQ/dV differential capacity analysis on Oxford University Li-ion pouch-cell data. The review identified the exact cycle where the battery moved from manageable degradation into accelerated failure using voltage logs alone.
Relevant stakeholders
Primary product
Proof format
What this study shows
Internal resistance profiling via voltage-transition analysis used to detect micro-short precursors hundreds of cycles before capacity drop.
Cycle 2
First Alert
Energy & Utilities
Internal resistance profiling on NASA Ames 18650 cells. The review detected high-frequency voltage transients, or micro-shorts, as early as Cycle 2, hundreds of cycles before a standard BMS would flag an issue.
Relevant stakeholders
Primary product
Proof format
What this study shows
Inverter clipping and string-level variance analysis used to identify and quantify ghost energy, power generated but discarded during peak irradiance.
4.2%
Yield Recovered
Renewable Energy
Inverter clipping analysis on NREL PVDAQ data. The review identified 4.2% annual yield recovery, energy produced by the panels but discarded by the inverters during peak irradiance, which standard site dashboards did not surface clearly.
Relevant stakeholders
Primary product
Proof format
What this study shows
If one of these patterns looks familiar, start with a fixed-scope Verify review. That baseline tells us whether Horizon belongs in the stack afterward.
Public data first • Live client reviews scoped separately • Managed or on-site delivery
