Polestar HYDRA OS
Talk to our team
Rendered PEM electrolyser stack. Illustrative.

Scale-up

From a short stack to the stack you will ship.

Short-stack and early full-stack results become estimates for the full-size stack, including current and thermal distribution and, in alkaline designs, shunt currents.

Operating profile

From steady-state test to real operation.

Evidence from constant-load testing is translated to the profile the plant will actually run: fluctuating renewable power, start/stop cycles and standby.

Time

From months of data to the project's lifetime.

Stabilisation, linear or accelerating degradation, each with explicit scenario weights, so stack life becomes a P10/P50/P90 range rather than an assumption.

Continue to the page
Cells at full scale
Membrane electrode assembly
Skip

Two ways to buy it.

Illustrative render of a PEM electrolyser stack, assembled.

  • Evidence Pack

    Turns short-stack and full-stack test data into evidence structured for your customers' due diligence.

    For OEMs

  • TDD Support

    Estimates for a specific project and operating profile, ready for technical due diligence.

    For developers and owners

Commercial
Fixed fee, agreed before any data is shared. No integration and no live connection: you send existing data as files. Early engagements are priced lower in exchange for permission to use anonymised results in our validation work.
Typical duration
Six weeks from receipt of complete data.

The questions it answers

What HYDRA Assess tells you.

  • What will a full-size stack deliver, given short-stack results?
  • What happens under our operating profile rather than the steady-state test?
  • How confident can we be over the project's lifetime?
  • Which additional test or data would most reduce the uncertainty?

What you provide

  • Raw test data: short stack, full stack or both
  • Or 3 to 12 months of historian export from an operating plant
  • The intended operating profile
  • The datasheet and guarantee terms

What you receive

  • P10/P50/P90 for specific energy consumption, degradation rate and stack life
  • Uncertainty budget: measurement, parameter, model form, extrapolation
  • Representativeness assessment
  • Claim-versus-evidence comparison
  • Degradation-adjusted LCOH as a distribution
  • Recommended evidence to reduce the uncertainty
  • Data-quality statement

How an engagement runs.

  1. Week 1

    Data intake and quality checks

    Gaps, unit problems and sensor drift are listed back to you.

  2. Week 2

    Baseline

    Polarisation behaviour and correction to reference conditions.

  3. Week 3

    Degradation

    Separating reversible from irreversible losses.

  4. Week 4

    Projection

    Projection under your intended operating profile, with an uncertainty range on each figure.

  5. Week 5

    Draft report

    Draft report and review call.

  6. Week 6

    Final report

    The final report.

Methodology notes

Definitions, normalisation, uncertainty.

KPIs are complementary to DNV-RP-J302 and use the terminology of the JRC EU harmonised testing protocols. Measured values are reported separately from modelled values.

Data is corrected to reference conditions before any comparison. Every estimate carries its uncertainty, split into measurement, parameter, model-form and extrapolation components.

Physics gives the model its structure; statistics gives it its uncertainty. Language models may help draft report text. They never make the call.

Our standard of evidence

What it includes

  • Electrochemical engine calibrated against NREL benchmark data
  • Data-quality checks
  • Polarisation curve fitting and performance baseline (kinetic, ohmic, transport losses)
  • Normalisation of operating data to reference conditions
  • Degradation tracking, separating reversible from irreversible losses
  • An uncertainty range next to every figure
  • Projection under the intended operating profile
  • Short-stack to full-stack scale-up
  • Representativeness score (scale, stack generation, operating envelope)
  • KPIs complementary to DNV-RP-J302 performance tests
  • Remaining useful life estimate and stack replacement planning
  • Report text drafting (language model, from computed results only)

Questions.

Do we need to connect HYDRA OS to our plant?

No. An assessment runs on exports: raw test files or a historian export. Nothing is installed on your systems.

Our test data comes from a short stack. Is that enough?

It is where most evidence starts. The assessment states how representative it is of the full stack and your operating profile, what uncertainty that adds, and which additional test would reduce it most.

Who sees the results?

You do. The report states who commissioned it, and the methodology is the same whoever pays for it. Sharing it with your customers, lenders or their engineers is your decision.

What happens to our data afterwards?

It is held under NDA for the engagement and deleted afterwards with written confirmation.

Start with an assessment.

Send us an export. Our team returns verified KPIs, degradation rates and a clear view of where your voltage goes.

Request an assessment Talk to our team