What industries does newtsim serve?

newtsim serves oil and gas (flow assurance, drilling fluids, subsea pipeline), pharmaceutical and drug discovery, energy storage (battery materials, electrolytes), semiconductor process simulation, green hydrogen (electrolyser and catalyst design), nuclear decommissioning, aerospace and composites, corrosion and materials, carbon capture, agrochemicals, mining and geotechnical, and fermentation and bioprocessing.

How is newtsim different from standard CFD or simulation software?

Standard simulation tools use empirical models at a single scale. newtsim runs the full physics chain from quantum mechanics and molecular dynamics through to plant-scale CFD and structural mechanics, GPU-native, with direct data flow between every scale and no third-party solver dependencies. Constitutive behaviour is predicted from first principles rather than empirically fitted.

What does a fixed-price simulation study include?

A newtsim fixed-price study has a defined scope, clear deliverables, fixed cost agreed upfront, and is delivered in weeks rather than quarters. Studies typically include a scoping call, physics-grounded simulation using the appropriate newtsim solvers, results analysis, and a written report with actionable recommendations.

What is physics-grounded simulation?

Physics-grounded simulation derives material properties and constitutive behaviour from lower-scale physics (molecular dynamics and quantum mechanics) rather than from empirical fitting. This gives accurate predictions for novel materials and extreme conditions outside the range of experimental data, and enables root-cause analysis at the mechanistic level.

How does newtsim approach oil and gas flow assurance simulation?

newtsim runs a continuous physics chain from molecular wax crystallisation to full pipeline transient multiphase flow, with no empirical handoffs between scales. Wax appearance temperatures, gel strength, and deposition kinetics are predicted from molecular dynamics (newtsim Bond), passed directly to mesoscale morphology (newtsim Pack), and then propagated into multiphase CFD (newtsim Stream), giving flow assurance predictions that hold outside the range of laboratory data.

Can newtsim simulate battery electrolyte behaviour and degradation?

Yes. The Bond solver runs molecular dynamics for Li-ion transport and solid-electrolyte interphase (SEI) formation; Pack handles electrode morphology at the mesoscale; these results couple upward to cell-scale electrochemical models. This gives physics-grounded predictions of electrolyte degradation, capacity fade, and thermal behaviour without relying solely on empirical ageing fits.

How does physics-grounded drug discovery simulation differ from docking software?

Molecular docking software treats the protein and ligand as rigid bodies and scores binding poses by geometry. Physics-grounded simulation uses molecular dynamics to capture conformational flexibility and explicit solvation effects that rigid-body docking misses, predicting binding free energy, ADMET properties, and metabolic activation from first principles. This is particularly valuable for targets where induced-fit binding or solvent-mediated interactions drive selectivity.

What makes GPU-native simulation faster than CPU-based CFD?

GPU parallelism maps naturally to the computational kernels underlying molecular dynamics and CFD: large numbers of independent force or flux calculations that can execute simultaneously. newtsim runs all simulation scales GPU-native, avoiding the CPU-to-GPU data transfers that bottleneck traditional solvers which run physics on CPU and only offload selected operations to the GPU. This removes the transfer overhead and allows the full multiscale chain to run without the latency penalties of heterogeneous pipelines.

Does newtsim replace existing simulation tools like ANSYS or COMSOL?

newtsim fills the multiscale gap, connecting molecular-scale physics to the engineering-scale models that tools like ANSYS operate at. It is not a replacement for FEA-only or single-scale workflows where those tools are the right fit. Where newtsim adds value is in problems where constitutive behaviour, material properties, or interfacial physics need to be derived from lower-scale simulation rather than taken from a handbook or empirical fit.

How long does a fixed-price simulation study take?

Most studies complete in 4 to 8 weeks depending on scope. The standard timeline is: scoping call and problem definition in week 1; simulation execution across the relevant physics scales in weeks 2 to 4; results analysis, validation, and report delivery in weeks 5 to 6. More complex multiscale studies with broader operating envelopes may extend to 8 weeks. Scope and timeline are agreed and fixed before work begins.

What deliverables does a simulation study produce?

A newtsim simulation study delivers: a validated multiscale model calibrated to your system; an operating envelope with physics-based safety margins; a written technical report with findings and actionable recommendations; and optionally CFD visualisations or an ML surrogate model (newtsim Neural) for fast design-space exploration by your engineering team.

Does newtsim work for green hydrogen electrolyser and catalyst design?

Yes. The Stream solver handles proton exchange membrane and alkaline electrolyser CFD: multiphase flow, bubble dynamics, and membrane transport. Bond handles catalyst surface reaction molecular dynamics, giving adsorption energies and reaction barriers from first principles rather than from empirical correlations. Neural then produces fast-to-run surrogate models for design-space exploration and stack optimisation, reducing the iteration cost for new catalyst compositions and electrode architectures.

newtsim is a physics-grounded multiscale simulation consulting firm and software platform. We deliver fixed-price simulation studies and platform licences for industrial clients, covering the full physics chain from quantum mechanics and molecular dynamics through to plant-scale CFD and structural mechanics. All solvers run GPU-native.

Does newtsim use artificial intelligence or machine learning?

Yes. newtsim Neural produces ML-trained surrogate models derived from high-fidelity physics simulations. These surrogates run in milliseconds rather than hours, enabling rapid design-space exploration and real-time control without sacrificing physical accuracy. The ML models are trained on physics simulation data, not empirical measurements, so they inherit the accuracy and extrapolation properties of the underlying physics.

How do I get started with newtsim?

Book a free technical scoping call at newtsim.com/demo. Bring a description of your engineering challenge, the physics you think are important, and any data you have. The call is 45 minutes and produces a problem definition, proposed physics chain, and outline scope that forms the basis of a fixed-price proposal.

Can newtsim simulate non-Newtonian fluids?

Yes. newtsim specialises in non-Newtonian flow: yield-stress fluids, viscoelastic materials, shear-thinning and shear-thickening systems, and complex multiphase flows. The Stream solver handles macroscale CFD for these systems; Bond provides the molecular-scale rheology predictions that replace empirical viscosity fits; and the full chain gives accurate behaviour outside the calibrated range of laboratory data.

Does newtsim work for small companies or startups?

Yes. Fixed-price studies are available to clients of any size. The fixed-price, fixed-scope model means you know the cost before committing, making it accessible for smaller organisations without large simulation teams. Many newtsim clients are startups and scale-ups who need physics-grounded analysis early in product development before investing in laboratory or pilot-scale work.

What is the difference between molecular dynamics and CFD?

Molecular dynamics (MD) simulates the motion of individual atoms and molecules to compute material properties such as viscosity, diffusivity, reaction rates, and phase behaviour, from first principles. Computational fluid dynamics (CFD) models bulk fluid flow at the engineering scale using those material properties as inputs. newtsim connects the two: MD provides accurate constitutive inputs that CFD uses for engineering-scale predictions, rather than relying on empirical correlations.

What is a surrogate model in simulation?

A surrogate model (also called a reduced-order model or emulator) is a fast mathematical approximation of a complex simulation. newtsim Neural builds ML surrogates trained on high-fidelity physics simulation data. Once trained, a surrogate evaluates in milliseconds instead of hours, enabling real-time optimisation and design-space exploration. Because surrogates are trained on physics data, they are more accurate than empirical correlations, especially outside the training range.

What deliverables does a newtsim simulation study produce?

A newtsim simulation study delivers: (1) a validated multiscale model calibrated to your system; (2) an operating envelope with physics-based safety margins; (3) a written technical report with findings and actionable recommendations; and optionally (4) CFD visualisations and an ML surrogate model for fast ongoing design-space exploration by your engineering team.

Green Hydrogen

Green Hydrogen · Case Study

Post-Deployment Audit: Flow Field Design and Thermal Management in Large-Format PEM Cells

Executive Summary

The GrInHy2.0 project at Salzgitter Flachstahl --- 720 kW of Sunfire SOEC electrolysis at 850 degrees C, FCH JU funded, the largest solid oxide electrolyser demonstration in Europe at the time --- documented something the project team described carefully in its published reports: during early testing phases, strong thermal gradients and suspected loss of electrical contact created performance issues that required fast current control strategies and insulated edge sealing to mitigate in later module generations. The mean degradation rate in well-managed modules was under 23 mOhm.cm²/kh. The lesson the GrInHy2.0 team drew was explicit: thermal gradient management is the primary scaling challenge for large-format electrochemical cells, and it must be explicitly designed for rather than assumed to follow prototype behaviour. Published PEM electrolyser studies confirmed the same finding: Z-type manifolds produce 3--4x worse flow distribution than U-type manifolds for channels longer than 400 mm, and the pressure drop non-uniformity scales approximately as the square of the channel length. Infrared thermography studies had already demonstrated that once peak-to-mean current density ratio exceeded 1.20, hotspot membrane lifetime was reduced by 3--5x relative to the cell mean; above 1.30, pinhole formation timescales shortened to 8,000--15,000 hours. The commissioning data to flag this problem had even been recorded: the anomalously high pressure drop across the bipolar plate at first flow was misinterpreted as a blocked filter, when it was actually manifold stagnation in the far-field channels --- a signature of Z-type maldistribution at the 600 mm channel length scale that the 240 mm prototype had never encountered.

Eighteen months after commissioning a 25 MW PEM facility with 3,000 cm² large-format cells, 12 cells in one stack had pinhole perforations. Infrared thermography showed a 28 degrees C temperature spread against a 12 degrees C design specification. The failed cells were clustered 100--150 mm from the inlet manifold and 50--80 mm from the far edge --- the superposition of manifold far-field starvation and corner edge heat accumulation, exactly where the Ito and Kang literature predicted the worst-case outcome. The affected stack was offline, and the remaining four stacks were running at de-rated current density, bleeding GBP 450,000/month in lost H2 production revenue.

Had a full 3D CFD simulation of the large-format bipolar plate and manifold been run at the stack design stage --- or as a scale-up validation step before committing to production tooling --- it would have flagged the Z-type manifold's 2.8x channel flow non-uniformity at 600 mm channel length, identified the corner/far-field superposition zone as the thermal failure location, and specified the U-type manifold re-plumbing and flow rate increase needed to bring temperature spread within the 12 degrees C specification. The CAD data and commissioning pressure drop measurements to validate that model were available before the first stack was loaded with membranes.

This post-deployment audit applies full 3D VOF CFD with conjugate heat transfer and electrochemical coupling to reconstruct the root cause, evaluate six retrofit scenarios, and deliver a re-plumbing specification. The recommended retrofit --- U-type manifold re-plumbing at GBP 80,000 per stack combined with a 50% water flow rate increase at GBP 15,000 per stack --- reduces temperature spread to 9--11 degrees C within the specification, at GBP 475,000 total for all five stacks versus GBP 2.0M for a full bipolar plate re-tool and GBP 4.5M in continued losses if the failure propagates. The thermal and flow distribution signatures identified by the simulation --- temperature maps, channel flow non-uniformity, local current density distribution, and pinhole risk by cell position --- define the sensor channels for newtsim livesim, enabling real-time thermal management health monitoring of large-format stacks in service with automated early-warning before non-uniformity crosses the accelerated degradation threshold.

Scenario Background

(illustrative reference case)

In this worked example, a hydrogen production facility at a UK industrial hub retained a simulation consultancy to diagnose thermal non-uniformity and pinhole failures in large-format PEM electrolyser stacks. The facility supplies hydrogen for direct-reduction steel trials at a major UK steel producer, using a bespoke large-format stack design developed in partnership with a UK electrolyser startup.

The installation totals 25 MW across 5 x 5 MW stacks, each containing approximately 350 cells of 3,000 cm² active area (500 mm x 600 mm). The stacks use CNC-machined titanium bipolar plates with a parallel-serpentine hybrid flow field (2 mm channel width, 1.5 mm rib width) and a Z-type manifold configuration where inlet and outlet are on the same cell face (left side). Membranes are Nafion XL reinforced (27 um), with cathode operating pressure at 20 bar and anode at 1.5 bar. The nominal current density is 1.5 A/cm² (450 A total cell current), with a design temperature target of 70 degrees C and a uniformity specification of +/-6 degrees C (<=12 degrees C total spread).

The measured reality diverges sharply from the design intent. At rated load, the temperature range across Stack 02 reached 72--100 degrees C (a 28 degrees C spread), and the peak-to-mean current density ratio measured 1.45 --- well above the established threshold of 1.20 for accelerated local degradation. Twelve pinhole perforations developed in the highest-temperature zone of Stack 02 at Month 16, with voltage rise exceeding 8% in the affected region.

The underlying cause is a prototype-to-production scaling gap. The OEM conducted all development and qualification testing on 600 cm² cells in a temperature-controlled test fixture that maintained +/-2 degrees C external uniformity. This masked the thermal management inadequacy of the flow field design, because in the smaller prototype the shorter flow channel length (240 mm vs. 600 mm in production) ensures even distribution, and the external temperature control suppresses internal hotspots. At 3,000 cm² production scale, with no external thermal control and a 600 mm channel span, the flow field must provide all thermal management internally. The prototype qualification data provided no warning of the production-scale failure.

This scaling gap is documented as a recurring challenge in the published literature. Published studies demonstrate that PEM electrolyser cells above approximately 1,000 cm² require explicit manifold design optimisation; simple geometric scaling from smaller cells produces systematically poor flow distribution.

Challenge

The failure cascade from manifold maldistribution to pinhole formation involves five coupled physical mechanisms.

1. Flow maldistribution in the Z-type manifold

The bipolar plate manifold must distribute water across 300 parallel channels over a 600 mm span. For a Z-type manifold (inlet and outlet on the same cell face), the pressure distribution along the manifold header is inherently non-uniform: channels nearest the inlet port see the highest pressure and highest flow, while channels at the far end see significantly lower pressure and lower flow. Published studies demonstrate that Z-type manifolds produce 3--4x worse flow distribution than U-type manifolds for channels longer than 400 mm --- exactly the geometry relevant here. The consequence is a 2.8x ratio between maximum and minimum channel flow rates across the 600 mm cell width, measured by the commissioning pressure drop test (which showed an anomalously high pressure drop for the expected flow rate --- misinterpreted at commissioning as a blocked filter, but actually a sign of manifold stagnation in the far-field channels).

2. Thermal gradient formation and hotspot development

Two independent thermal effects combine at the dry/slow-flow regions. Reduced evaporative cooling in channels with low water flow means less heat is carried away from the electrochemical reaction. At 1.5 A/cm², the heat generation rate is approximately 0.7 W/cm² (from the voltage efficiency loss: (V_cell - V_thermoneutral) x J = (2.0 - 1.48) x 1.5 = 0.78 W/cm²). In channels with 2.8x lower flow than the inlet channels, the local temperature rise is proportionally higher. Simultaneously, cell corners receive water from two channel ends (corner convergence point) but lose heat to the titanium bipolar plate edge (which conducts heat to the external cell clamping structure). At the corner/manifold-far-end intersection, both effects coincide: low flow and high edge heat loss create the most severe thermal non-uniformity.

The infrared thermography (Month 14 archived data) confirms this analysis: the highest temperature region (+22 degrees C above mean) is located approximately 100--150 mm from the inlet manifold along the channel direction and 50--80 mm from the right (far) edge --- consistent with the combined manifold far-field starvation and corner edge effect.

3. Non-uniform current density distribution and positive feedback

At higher temperature, Nafion proton conductivity increases (sigma proportional to exp(-E_a/RT), E_a approximately 12 kJ/mol in the 60--90 degrees C range), and the Butler-Volmer exchange current density increases. Both effects make the hot, well-conducting membrane region carry more current than the surrounding cooler regions. The positive feedback loop --- hot region leading to higher sigma and i_0, leading to more current, leading to more Joule heating, leading to hotter still --- amplifies the initial maldistribution. Once the non-uniformity exceeds approximately 20% in local current density (the established threshold for accelerated local degradation), the degradation in the hotspot region is no longer representative of the stack mean.

4. Accelerated membrane degradation and pinhole formation in hotspots

At 90--100 degrees C (the measured peak temperature in the hotspot regions), membrane degradation rates are substantially elevated. Nafion chain scission rate scales with temperature via Arrhenius kinetics: at 95 degrees C versus 70 degrees C (mean), the degradation rate factor is exp(-E_a/R x (1/368 - 1/343)) approximately 2.3x (using E_a = 60 kJ/mol from published Nafion degradation kinetics). This yields a hotspot membrane thinning rate of approximately 3.5 um per 10,000 h versus the baseline 1.5 um per 10,000 h. Over 16 months (11,520 hours) of operation, hotspot thinning reaches approximately 4.0 um from an original 27 um membrane. Combined with mechanical stress from temperature-driven differential thermal expansion of the titanium bipolar plate at 90 degrees C versus the Nafion membrane (the titanium plate expands by 8.5 x 10-6 /degrees C x 20 degrees C x 1.5 mm rib width = 0.26 um additional compression per rib in the hotspot zone), the MEA is compressed by 15--20% more in the hot region than the cool region. The combination of accelerated chemical thinning and elevated mechanical compression triggers pinhole formation in the locally thinnest membrane regions.

5. Prototype-to-production scaling gap

The 600 cm² to 3,000 cm² scaling (5x area increase) is not a trivial geometric scaling. Flow channel length doubles (240 mm to 600 mm), which increases the manifold pressure drop non-uniformity by approximately the square of the length ratio (x4). The prototype's 2 mm x 0.5 mm channel geometry, which produced adequate distribution at 240 mm length, produces 2.8x flow non-uniformity at 600 mm length under Z-type manifold arrangement.

Sunfire GrInHy2.0 SOEC --- Lessons from Large-Format Cell Scaling:

Although a different electrochemical technology (high-temperature solid oxide, 850 degrees C), the GrInHy2.0 project at Salzgitter Flachstahl (720 kWAC, FCH JU funded) documented directly analogous thermal management challenges during scale-up. The project team reported mean degradation rate <23 mOhm.cm²/kh in well-managed smaller modules, but strong thermal gradients and suspected loss of electrical contact in early testing phases --- mitigated by fast current control strategies and insulated edge sealing in later module generations. The cross-technology lesson: thermal gradient management is the primary scaling challenge for large-format electrochemical cells, regardless of technology, and must be explicitly designed for rather than assumed to follow prototype behaviour.

Real-World Basis

Published experimental and CFD studies of flow maldistribution in large-format PEM electrolysers (up to 2,500 cm²) establish the key quantitative finding: Z-type manifolds produce 3--4x worse flow distribution than U-type manifolds for channels longer than 400 mm, with pressure drop non-uniformity scaling approximately as the square of the channel length. This forms the theoretical basis for the U-type re-plumbing recommendation.

Infrared thermography of PEM electrolyser cells during operation at 0.5--2.0 A/cm² demonstrates that local current density non-uniformity above 20% (peak/mean ratio >1.20) correlates with accelerated local membrane degradation, and above 30% non-uniformity (ratio >1.30) membrane lifetime in the hotspot region is reduced by 3--5x compared to the cell mean. The 1.45x non-uniformity ratio measured in this scenario places it in the highest degradation risk category.

Temperature and Current Density Non-Uniformity Reference Data

Temperature Spread (degrees C)	Max/Mean Current Ratio	Hotspot Degradation Multiplier	Expected Time to Pinhole
<=6 (design target)	1.05--1.10	1.0x (no acceleration)	>80,000 h
6--12 (borderline)	1.10--1.20	1.5--2.0x	40,000--60,000 h
12--20 (exceeded spec)	1.20--1.35	2.5--4.0x	20,000--35,000 h
20--30 (severe)	1.35--1.50	4.0--7.0x	8,000--15,000 h
>30 (critical)	>1.50	>7.0x	<8,000 h

Published physics-based PEM electrolyser performance models provide the electrochemical sub-model framework adapted for this study, including Nernst equation formulation with temperature dependence of reversible voltage and exchange current density, and coupled flow-electrochemical simulation methodology for the electrolyser anode.

HyBalance operational data confirms thermal management as a critical challenge even at 1 MW scale: inlet water temperature control of +/-1 degrees C was required to maintain cell temperature uniformity within specification at 1.2 A/cm². At 3,000 cm² and 1.5 A/cm², the challenge is proportionally more severe.

Simulation Approach

The simulation strategy uses a domain decomposition approach to achieve full 3D resolution of the manifold while remaining computationally tractable at 3,000 cm² scale:

Full 3D CFD Model (newtsim Stream):

The domain decomposition separates two distinct regions, each chosen for its role in the failure mechanism. The manifold region is modelled in full 3D (500 mm width x 100 mm from inlet, full cell height) because this is where the flow maldistribution originates --- the Z-type header's pressure gradient across 300 parallel channels cannot be captured by a reduced-order model. Two-phase VOF resolves the anode water-O2 flow regime. Conjugate heat transfer captures conduction through the titanium plate (17 W/m.K), PTL (approximately 0.3 W/m.K), and MEA (approximately 0.2 W/m.K through thickness). The active area interior uses a 2D periodic model with the manifold model's outlet condition as its boundary, extending the thermal-electrochemical prediction across the full 600 mm cell length without requiring the full 3D mesh everywhere.

Electrochemical coupling:

Butler-Volmer kinetics at each computational cell capture the positive feedback loop that makes this problem self-amplifying: hotter regions conduct better, draw more current, generate more heat, and get hotter still. Local ohmic resistance combines the membrane contribution with the PTL contact resistance, both temperature-dependent. This coupling is why the 28 degrees C spread is so damaging --- it is not a passive temperature gradient but an actively amplifying non-uniformity.

Retrofit scenario matrix:

Scenario	Modification	Implementation Cost per Stack	Implementation Time
A (baseline)	As-built, current conditions	GBP 0	---
B	50% anode water flow rate increase	GBP 15,000 (pump uprating)	1 week
C	Inlet temperature reduced 60 to 50 degrees C	GBP 5,000 (control setpoint)	1 day
D	Current density de-rated to 1.2 A/cm²	GBP 0 (control setpoint)	1 day
E	U-type manifold re-plumbing	GBP 80,000 (pipe rerouting + flow testing)	3 weeks
F	Manifold taper re-machine	GBP 400,000 (bipolar plate EDM + stack rebuild)	3 months

Simulation Caveats

Classification: STANDARD (well-validated CFD regime). The 3D CFD simulation of BPP flow distribution and conjugate heat transfer sits in the well-established range of newtsim Stream validation for electrochemical applications, but carries domain-specific limitations:

BPP/PTL contact resistance. The electrochemical coupling in the CFD model uses a fixed contact resistance at the BPP/PTL interface of 8 mΩ·cm² (from the OEM's published stack specification at beginning of life). After 18 months of operation, TiO₂ passivation on the BPP surface has likely increased contact resistance by 2–5 mΩ·cm² in high-current-density zones, which would increase localised ohmic heating by 15–30% relative to the model. The predicted 28°C temperature spread is therefore a conservative lower bound; actual current conditions may be worse. Post-mitigation predictions (9–11°C) are based on BoL contact resistance and represent the target state after potential re-activation of the BPP surface (acid wash or electrochemical reduction protocol).
Electrochemical coupling simplification. The current density distribution is computed using a simplified Butler-Volmer kinetic expression with uniform catalyst layer properties. Local membrane resistance variations due to the hydration gradient (which is the downstream effect of the flow maldistribution) are not fed back into the electrochemical model iteratively. Full two-way coupling would increase the predicted non-uniformity by an estimated 10–15%.
Scale-up geometry assumptions. The simulation domain uses the nominal BPP geometry from the design drawing. Manufacturing tolerances (channel width ±0.05 mm, land width ±0.03 mm) in large-format plates can create significant channel-to-channel variation. A Monte Carlo study of manufacturing tolerance effects (not included in the base scope) would provide additional conservatism for the retrofit design specification.

Recommended framing: The scenario ranking (U-type manifold + 50% water flow increase as preferred retrofit) is robust to the model uncertainties. The predicted temperature uniformity improvement (28°C → 9–11°C) carries ±2°C uncertainty; the 12°C specification is met with margin for the P50 model prediction.

Key Predictions / Results

Table 1: Scenario Performance Predictions

Metric	Scenario A (Baseline)	Scenario B (High Flow)	Scenario C (Low T)	Scenario D (De-rated)	Scenario E (U-type)	Scenario F (Tapered)
Peak cell temperature	98 degrees C	85 degrees C	88 degrees C	80 degrees C	73 degrees C	71 degrees C
Temperature spread	28 degrees C	20 degrees C	22 degrees C	15 degrees C	9 degrees C	7 degrees C
Meets <=12 degrees C spec?	No	No	No	Marginal	Yes	Yes
Max/mean channel flow	2.8x	2.3x	2.8x	2.8x	1.2x	1.1x
Peak local J / mean J	1.45	1.30	1.33	1.18	1.08	1.06
Corner region Delta-T above mean	+22 degrees C	+15 degrees C	+17 degrees C	+10 degrees C	+4 degrees C	+3 degrees C
Predicted pinhole rate (hotspot)	0.8 per 10,000 h	0.25 per 10,000 h	0.30 per 10,000 h	0.05 per 10,000 h	0.018 per 10,000 h	0.010 per 10,000 h
DC efficiency at rated J	74.2%	76.1%	74.8%	75.5%	77.6%	77.9%
Scenario cost per stack	GBP 0	GBP 15,000	GBP 5,000	GBP 0	GBP 80,000	GBP 400,000

Table 2: Flow Distribution --- Z-type vs. U-type Manifold (Channel Flow Non-Uniformity)

Distance from Inlet (mm)	Z-type Manifold Flow (relative)	U-type Manifold Flow (relative)
0--50 (near inlet)	1.82x mean	1.12x mean
50--100	1.45x mean	1.09x mean
100--200	1.15x mean	1.05x mean
200--400	0.85x mean	1.00x mean
400--500	0.72x mean	0.97x mean
500--600 (far end)	0.65x mean	0.95x mean
Max/min ratio	2.80x	1.18x

Temperature distribution across 3,000 cm² PEM cell — Z-type vs U-type manifold

The U-type manifold reduces the flow non-uniformity ratio from 2.8x to 1.18x --- a 2.4x improvement --- by ensuring the pressure gradient along the manifold header is driven by the outlet port rather than the inlet port, creating a more uniform pressure distribution across all channels.

Table 3: Damage Assessment --- Remaining Cells at Risk

Region	Local J/Mean J	Local Peak T	Predicted Cumulative Pinhole Rate (18 months)	Current Status
Near-inlet central (0--50 mm x 200--400 mm width)	1.10--1.15	76--80 degrees C	<0.02 per 10,000 h	Low risk
Mid-cell (100--400 mm x 100--500 mm)	1.0 (mean)	70--75 degrees C	Baseline	Low risk
Far-end centre (500--600 mm x 200--400 mm)	0.92--0.95	68--72 degrees C	0.03 per 10,000 h	Negligible
Near-inlet corners (0--50 mm x 0--50 mm)	1.25--1.35	82--90 degrees C	0.25--0.45 per 10,000 h	Moderate risk
Far-end + inlet-face (failure zone)	1.40--1.50	90--100 degrees C	0.7--1.0 per 10,000 h	High risk (12 pinholes observed)

The far-end + inlet-face zone (100--150 mm from inlet, 50--80 mm from right edge) is the confirmed failure location. This zone experiences both manifold far-field starvation (low flow) and corner edge heat accumulation simultaneously --- the worst-case superposition of the two thermal effects.

Table 4: Economic Comparison of Retrofit Scenarios

Scenario	Meets Spec?	Retrofit Cost (5 stacks)	Production Recovery	5-Year NPV of Decision
A (do nothing)	No	GBP 0	None (further failures likely)	-GBP 4.5M (replacement MEAs + lost production)
B + E combined (recommended)	Yes	GBP 475,000	Full rated output at 1.5 A/cm²	+GBP 3.2M vs. do-nothing
D (de-rate only)	Marginal	GBP 0	Partial (20% production loss)	-GBP 0.8M vs. scenario E
F (full retool)	Yes	GBP 2.0M	Full rated output	-GBP 0.5M vs. Scenario E

Retrofit scenario comparison — temperature uniformity vs implementation cost per stack

Hotspot risk map — local J/mean J and pinhole rate by cell region, as-built vs retrofit

Expected finding and recommendation: Scenario E (U-type manifold re-plumbing, GBP 80,000 per stack) combined with Scenario B (50% flow rate increase, GBP 15,000 per stack) achieves 9 degrees C temperature spread --- within the <=12 degrees C specification --- at a total cost of GBP 475,000 for all five stacks. This is 4.2x cheaper than the full bipolar plate re-tool (Scenario F, GBP 2.0M for five stacks) and restores full rated output at 1.5 A/cm², recovering approximately GBP 450,000/month in lost H2 production revenue compared to the current de-rated operation.

Comparison Methodology

Infrared thermography validation:

Archived IR thermography images from the affected stack at 1.0, 1.25, and 1.5 A/cm² (three current density levels, from the Month 14 investigation) are compared against the CFD model's Scenario A (as-built) predictions at the same three conditions. Comparison is performed on a pixel-by-pixel basis, downsampled to 10 mm x 10 mm resolution to match IR camera spatial resolution. Required agreement: peak temperature within +/-3 degrees C, temperature spread within +/-4 degrees C, and hotspot location (centroid of the >85 degrees C region) within +/-20 mm of the measured centroid.

EIS spatial mapping:

Segmented EIS data (16-segment measurement grid using the segmented current collector, approximately 188 cm² per segment) provides spatial maps of HFR (10 kHz) and low-frequency (0.1 Hz) arc resistance. The model's predicted local water activity and local current density are converted to predicted EIS response using published Nafion impedance models. HFR spatial distribution must agree within +/-15% per segment; the low-frequency arc pattern (which correlates with water management quality) must reproduce the measured spatial gradient from inlet to far-field regions.

Pressure drop commissioning check:

Bipolar plate anode inlet-to-outlet pressure drop measured at 5 flow rates during commissioning is compared against CFD-predicted pressure drop (expected agreement +/-5%), validating the fluid dynamic model before thermal and electrochemical predictions are trusted.

Cross-validation with end-position stack:

The end-position stack in the 5-stack module has different thermal boundary conditions (one face exposed to the module housing air gap rather than an adjacent stack). The CFD model configured with these boundary conditions serves as an additional validation case. The end-position stack shows less severe maldistribution in the operational data (temperature spread approximately 18 degrees C vs. 28 degrees C in the affected stack), consistent with the end-position cooling providing partial relief. The model must reproduce this 10 degrees C difference between stacks to be considered valid for predicting the retrofit impact.

Deliverables

Root cause analysis report --- Definitive identification of the maldistribution mechanism (Z-type manifold at 3,000 cm² scale) and its thermal/electrochemical consequences (28 degrees C spread, 1.45x current non-uniformity, Fenton-accelerated hotspot membrane degradation). Includes 3D visualisations of temperature, current density, water saturation, and velocity magnitude maps for the as-built configuration. Published as a technical report formatted for submission to the OEM, the plant insurer, and the UK DESNZ green hydrogen project monitoring programme.
Retrofit recommendation --- Ranked assessment of all 6 scenarios by effectiveness (temperature spread achieved), cost, implementation time, and risk of unforeseen consequences. Primary recommendation: Scenario E (U-type re-plumbing) + Scenario B (flow rate increase), at GBP 475,000 total for five stacks, meeting the <=12 degrees C specification. Secondary recommendation: Scenario D (de-rating to 1.2 A/cm²) as an immediately implementable interim measure while the Scenario E+B retrofit is executed over 3 weeks per stack.
Revised operating envelope --- Safe operating current density and water flow rate as a function of achieved temperature uniformity specification, expressed as a 2D operating map. Allows the plant to continue operating at de-rated output (1.2 A/cm² delivering 80% of rated H2 output) during the retrofit period rather than shutting down.
Damage assessment for remaining cells --- Predicted cumulative thermal damage in each of the five stacks' cells over the 18-month operating history, using the maldistribution model driven by the historian data. Identifies the specific cell positions (by cell number within stack, by position within cell) at highest risk for near-term pinhole failure. Recommends priority EIS inspection pattern (full 16-segment EIS on the 50 highest-risk cells identified by position).
Future design guidelines for large-format PEM cells (1,500--5,000 cm²) --- A set of manifold and flow field design rules derived from the CFD parametric study, for the OEM's use in next-generation cell design to prevent recurrence in future installations. These include: U-type manifold mandatory for active areas >1,000 cm² (channels >300 mm long); manifold header cross-section must produce <10% pressure non-uniformity along the full span (verified by CFD before tooling commitment); corner cell edge insulation (5 mm ceramic foam strip) mandatory for cells >2,000 cm²; scale-up validation must include full-scale thermal testing with no external temperature control; and minimum water flow rate must be determined by thermal management requirement at maximum current density, not by stoichiometry alone.

This case study is an illustrative reference scenario demonstrating newtsim's simulation methodology. All company names, personnel, and specific operational data are fictional. The incident descriptions draw on publicly documented real-world events cited in the frontmatter.