Polestar HYDRA OS
HYDRA OS · Electrolyser performance intelligence
HYDRA OS turns test and operating data into P10/P50/P90 estimates of efficiency, degradation and stack life, from short stack to full stack and from steady-state test to real operation. These are numbers lenders, insurers and independent engineers can build a business case on.
The economics
Cell voltage decides how much of it each kilogram costs.
Uncertainty is the real risk
When evidence is thin, investors hesitate or assume the worst, sometimes planning stack replacements years earlier than the manufacturer expects.
Thin evidence means short stacks, steady-state tests, and months of data rather than years.
Representative by design
Physics you can inspect
Our electrochemical engine separates cell voltage into its contributions. Every output traces to a measurement, an equation and an uncertainty.
HYDRA OS at cell level
HYDRA OS separates each cell's voltage into kinetic, ohmic and transport losses and tracks each one over time. Catalyst loss, membrane ageing and transport limits stop hiding inside a single stack voltage.
Illustrative data
HYDRA OS at cell level
From short-stack and steady-state evidence, HYDRA OS estimates how your cells will age at full scale and on your operating profile, as P10, P50 and P90, and states whether the trend looks linear or accelerating.
Illustrative data
HYDRA OS at cell level
Cell-level losses roll up into specific energy consumption, stack life and degradation-adjusted LCOH, each as a range with its sources, and with the evidence that would narrow it most.
Illustrative data
From physics to evidence
Reports structured the way independent engineers review them: KPIs on recognised definitions, an uncertainty budget, and every figure traced to its source.
Continue to the pageIllustrative render · generic PEM stack
Make your test data count in your customers' due diligence.
Bring P50/P90 assumptions that survive due diligence.
Know whether your stacks deliver what was guaranteed, and watch the uncertainty narrow.
Review performance on recognised definitions, with the uncertainty stated.
See degradation and loss evidence before it becomes a claim.
IRENA identifies electricity as the largest single cost component of green hydrogen production.
CitedAdditional electricity per kilogram of hydrogen for every 10 mV of cell voltage.
DerivedThe annual cost of 20 mV for a plant producing 10,000 t/yr at €70/MWh.
Derived · illustrativeWhen evidence is thin, investors hesitate or assume the worst, sometimes planning stack replacements years earlier than the manufacturer expects. Thin evidence here means short stacks, steady-state tests, and months of data rather than years. HYDRA OS quantifies the uncertainty, shows where it comes from, and identifies the evidence that would reduce it.
From short stack to full stack, including shunt currents in alkaline designs and current and thermal distribution.
From steady-state testing to fluctuating renewable operation, start/stop cycles and standby.
From months of data to project lifetime: stabilisation, linear or accelerating degradation, with explicit scenario weights.
| Technology | Degradation |
|---|---|
| Alkaline (AEL) | 0.12 %/1,000 h |
| PEM (PEMEL) | 0.19 %/1,000 h |
| Solid oxide (SOEL), at thermoneutral voltage | 1.9 %/1,000 h |
P10/P50/P90 performance, degradation and stack-life estimates from your test or operating data, with scale-up, operating-profile translation and an uncertainty budget. Available as an Evidence Pack for OEMs, or as TDD Support for developers and owners.
Acceptance-test and guarantee verification, corrected to guarantee conditions, with a stated uncertainty and a clear verdict.
Continuous on-premise, read-only monitoring that narrows the uncertainty band as operating data accrues. Also covers data quality, early warning and stack planning.
The toolkit for independent engineers, lenders and insurers reviewing electrolyser performance.
Start with an export from your historian: no installation, no connection to your plant. For continuous monitoring, HYDRA OS reads data over OPC UA, Modbus or MQTT. It is read-only, on-premise, and never writes to control or safety systems.
Our electrochemical engine checks data quality, finds steady-state operating windows, corrects to reference conditions and separates cell voltage into its physical contributions. Every output traces to a measurement, an equation and an uncertainty.
You receive P10/P50/P90 estimates, KPIs on recognised definitions, an uncertainty budget, guarantee verdicts and the evidence that would most reduce the remaining uncertainty, in reports structured the way independent engineers review them. Language models may help draft the text. They never make the call.
Polestar makes no electrolysers, stacks or components, and takes no commission from those who do. HYDRA OS is OEM-agnostic and works alongside your OEM's monitoring and service. That independence is the product.
Our commitments
Our KPI definitions follow the conventions of DNV-RP-J302, Performance and testing of electrolyser systems. Our approach to model assurance references DNV's recommended practices on digital-twin and data-driven model assurance. Every tool on this site runs the same engine as HYDRA OS, with its method written out.
How the engine normalises, attributes and states uncertainty.
Calibration against published benchmark data, and validation by back-testing.
The engine's calculations, with the method written out.
Every figure on this page, with its status and source.
No writes to control or safety systems.
No cloud dependency.
NDA, deletion on completion, open export formats.
Documented data model; source-code escrow on request.
Keep it. HYDRA OS gives you an independent reading of the same data, which is what lenders, insurers and warranty discussions call for.
HYDRA OS is read-only and advisory. It never writes to control or safety systems. Assessments need no connection at all, only an export.
It stays on your site, or with us under NDA during an assessment, and it is deleted afterwards with written confirmation.
Every KPI follows a published definition, every figure carries its uncertainty, and every report records the data quality and the engine version used.
Open export formats, a documented data model and source-code escrow on request. Your evidence stays usable whatever happens.
Send us an export. Our team returns verified KPIs, degradation rates and a clear view of where your voltage goes, typically within weeks.