That mismatch is the target of a new open dataset published in Earth System Science Data Discussions by independent researcher Zach Dissington. The work, currently under peer review, presents a global registry of the world’s primary aluminium smelters operating in 2020 — 94 facilities spread across 36 countries, catalogued in 95 registry rows with coordinates, nameplate capacity, operating status, and per-row source notes. From this registry, the study derives a set of gridded relative-weight fields designed to serve as the aluminium-sector component of a spatial prior in atmospheric inversion models, the statistical frameworks used to work backwards from measured atmospheric concentrations to actual emission locations and magnitudes.

The logic behind the approach is disarmingly simple. Smelters are a finite, public, mappable set of industrial facilities. There is no mystery about where aluminium is produced: the plants are enormous, conspicuous, and documented by national statistics agencies and industry associations alike. Distributing a smelter-specific gas by population density, as many existing inventories effectively do through built-up-area and population proxies, is a category error that the new dataset aims to correct. The registry is accompanied by ready-to-use gridded maps in three variants — capacity-weighted, presence-only, and production-rescaled — at 1-degree resolution globally and 0.1 degrees over Europe, formatted for direct ingestion into inversion workflows and inventory disaggregation schemes.

Validation of the registry against independent data shows encouraging consistency. Registry capacity tracks United States Geological Survey national production estimates within roughly 25 percent in every listed producing country, with one glaring exception: China. The Chinese aluminium industry is vast and fragmented, and the registry handles it with 12 registered cluster-anchor plants standing in for a fleet of roughly 120 smelters. For China, the production-rescaled variant rescales the national total rather than attempting to resolve placement within the country — an honest acknowledgment of the limits of open-source facility mapping in a sector where plant-level data is harder to verify.

The real test, however, is whether the smelter-based maps actually improve the placement of CF4 emissions in the eyes of the atmosphere itself. To find out, the study compared the smelter field against the EDGAR v8.0 baseline — the standard European emission inventory — using an observation-driven posterior ensemble from the ICOS 2020 European CF4 analysis, generated with a flat prior under the RHIME inversion system. The results are nuanced but telling. In Iceland, home to power-hungry smelters fed by geothermal and hydroelectric energy and essentially devoid of other plausible CF4 sources, the smelter field correlated with the posterior better than the EDGAR baseline, scoring 0.25 versus 0.00, a difference that was statistically significant under spatial tests in both single-system ensemble members covering the island.

Across the pooled group of European smelter countries, the smelter field performed directionally better in five of the six ensemble members, suggesting a consistent if not always decisive advantage. But the comparison is not a clean sweep. France significantly favoured the EDGAR baseline in the same two members that favoured the smelter field over Iceland, and Norway, Spain, and the United Kingdom were all EDGAR-favoured in the evaluation. These mixed results reflect the inherent difficulty of the problem: CF4 emissions are small in absolute terms, atmospheric measurements are sparse, and inversion posteriors carry their own uncertainties. Where the measurements can clearly distinguish between the two spatial hypotheses — as in Iceland — the smelter map wins. Where they cannot, the outcome depends on the details of the inversion setup.

One finding from the evaluation is particularly striking. EDGAR assigns exactly zero CF4 emissions to 15 of the 22 European smelter cells in the analysis domain. In other words, for the majority of grid cells containing active aluminium smelters in Europe, the standard inventory places no aluminium-sector CF4 at all, presumably distributing the national totals elsewhere or omitting them entirely. For any inversion system trying to reconcile measured atmospheric CF4 with reported emissions, that kind of structural misplacement can bias the inferred emission estimates in ways that are difficult to diagnose. A prior that at least puts the weight where the facilities are gives the inversion a fighting chance of finding the right answer.

The author is careful to frame what the dataset is and, just as importantly, what it is not. The gridded fields are relative weights for a single sector, not emissions estimates in themselves. They tell an inversion model how to distribute a known or hypothesised aluminium-sector total across space; they do not say how much CF4 the world’s smelters actually emit in any given year. The paper’s sixth section lays out the workflow and the cautions required to use the fields correctly, an unusually transparent treatment of the assumptions baked into spatial priors — assumptions that are often hidden inside large inventory products and rarely examined by end users.

The significance of getting CF4 right extends well beyond the aluminium sector. Because the gas persists for 50,000 years, every tonne emitted is effectively a permanent addition to the atmospheric greenhouse burden on any timescale relevant to human civilisation. Global CF4 concentrations have been rising steadily since the mid-twentieth century, and while the anode effect can be reduced through better process control and emerging inert-anode technologies, the emissions already released will outlast every institution that recorded them. Accurate attribution of ongoing emissions is therefore a prerequisite for both scientific accountability and any future policy that treats perfluorocarbons with the seriousness their longevity demands.

The dataset itself, archived on Zenodo under a CC-BY-4.0 license alongside accompanying model code and software, represents a quiet but meaningful shift in how emission inventories for industrial gases might be built. Instead of relying on population proxies that smear a point-source gas across cities, the registry approach starts from the physical reality of the emitting infrastructure and builds outward. It is a method that could in principle extend to other sector-specific gases with well-defined facility footprints. For now, it gives atmospheric scientists something they have long lacked for CF4: an open, verifiable, smelter-resolved map of where the world’s aluminium — and its most permanent greenhouse gas — actually comes from. Whether the inversion community adopts it widely will depend on further validation, but the direction of travel is clear: when a gas has a factory address, the map should know it.

Subject of Research: An open global registry of primary aluminium smelters used to derive gridded spatial priors for CF4 (PFC-14) greenhouse gas emissions in atmospheric inversions.

Article Title: An open global registry of primary-aluminium smelters and a derived gridded spatial prior for CF4 (PFC-14) emissions

Article References: Dissington, Z. (2026). An open global registry of primary-aluminium smelters and a derived gridded spatial prior for CF 4 (PFC-14) emissions. https://doi.org/10.5194/essd-2026-480

Image Credits: AI Generated

DOI: 10.5194/essd-2026-480

Keywords: CF4, PFC-14, tetrafluoromethane, aluminium smelters, greenhouse gases, emission inventories, atmospheric inversion, EDGAR, spatial priors, anode effect, Earth System Science Data, open dataset

The dataset, described in the journal Results in Engineering by Qing Li and Hao Xu, documents experiments on a rotor-bearing-gearbox system in which five distinct blade conditions were tested: a healthy blade and four seeded faults, namely a groove, wear, a partially broken section and a bent blade. Each condition was recorded under a stable speed of 20 Hz and under four non-stationary scenarios: continuously increasing speed, continuously decreasing speed, increasing then decreasing speed, and decreasing then increasing speed. The result is a matrix of 25 datasets covering every combination of health state and speed trajectory, all synchronously sampled at 20 kHz together with the instantaneous rotational speed signal.

The motivation is straightforward but consequential. Incipient blade faults such as small cracks, grooves or wear patches change the local mass distribution and stiffness of a rotor, disrupting its circumferential symmetry. Under steady conditions these faults produce characteristic, well-understood signatures. But when the shaft speed is sweeping continuously, the fault frequencies smear across the spectrum, a phenomenon known as spectral smearing, and the diagnostic features that work beautifully at constant speed can dissolve into noise. Machines in the field almost always operate in these non-stationary regimes, which means algorithms validated only on constant-speed data may fail precisely when they are needed most, during start-up, shutdown and load transients, when the risk of resonance-driven failure is elevated.

The experimental platform itself is a carefully instrumented drivetrain. A three-phase asynchronous motor, governed by a variable frequency drive, turns a shaft through a coupling into double-support bearings of type PH205, carrying a 120 mm aluminium alloy rotor disc fitted with six blades. Thirty-six M5 threaded holes, spaced at 10-degree intervals around the disc, allow precise dynamic balance adjustment. Downstream, a reducer gearbox connects to a magnetic powder brake, whose adjustable input current lets the researchers impose a controlled resistive load anywhere from 0 to 6 Nm. For all the runs in this dataset the brake was held at a fixed 4 Nm, isolating speed variation as the changing experimental factor.

The four fault types are quantitatively defined rather than vaguely described, which matters enormously for reproducibility. The groove fault is a cut 3 mm wide and 35 mm long; the wear fault removes material to a depth of 1.5 mm; the broken fault removes a 35 mm region of the blade; and the bending fault deforms the blade by 10 degrees. The healthy reference blade is 2 mm thick, 60 mm long and set at a 30-degree blade angle. Because each fault is a precisely specified geometric modification, other laboratories can in principle reproduce identical damage on their own rigs and compare results directly, something rarely possible with naturally occurring faults.

Each speed scenario was executed with care. In the stable-speed runs, 60 seconds of data were collected at 20 Hz and split into two 30-second parts. The ramping runs varied in duration, from roughly 15 seconds for a simple deceleration to nearly 39 seconds for a full up-down sweep, with speeds ranging between about 9 Hz and 20.75 Hz depending on the trajectory. Multiple independent repeated trials were conducted for each health condition and speed profile, with the fault condition, sensor arrangement, load and prescribed speed curve held identical across repetitions. This repetition strategy reduces the influence of random experimental disturbances and measurement variability, and gives users a basis for assessing the consistency of vibration responses under nominally identical conditions.

The sensing arrangement captures three-axis acceleration, with channels in the X, Y and Z directions recorded from an accelerometer mounted on the right side of the rotor disc, alongside the input velocity signal. Each data file is organised into five columns: operating time, acceleration in X, acceleration in Y, acceleration in Z, and rotational speed. A systematic file-naming convention encodes the health condition, the speed scenario and the dataset number, so that a file labelled 2_3_Groove_Up_Down, for instance, unambiguously identifies a grooved blade run through an increasing-then-decreasing speed profile. This kind of disciplined organisation is unglamorous, but it is exactly what turns a pile of recordings into a usable benchmark.

The authors also demonstrate one analytical workflow that the dataset supports: envelope order spectrum analysis, a technique built to defeat spectral smearing under variable speed. The method band-pass filters the vibration signal, applies a Hilbert transform to extract the amplitude envelope of the modulated carrier, resamples the envelope uniformly in the angular domain using the integrated instantaneous rotational frequency, and then applies a discrete Fourier transform to produce a spectrum expressed in orders of the shaft rotation rather than in hertz. Because the angular domain is tied to shaft position rather than wall-clock time, features that would blur in a conventional frequency spectrum remain sharp regardless of how fast the shaft is turning.

Applied to the increasing-speed run of the grooved blade, the technique is revealing. The band-pass filtered signal shows periodic fluctuations confirming continuous pulse excitation from the damaged blade, and the overall envelope amplitude grows with time as the rising speed injects more impact energy. In the resulting envelope order spectrum, distinct peaks appear at 1.00X, 2.01X, 4.02X and 7.03X, matching the theoretical vibration features of a grooved blade, while noise components stay at a low level. The physics is intuitive: the local material removal disrupts the circumferential symmetry of the rotor, producing a periodic excitation once per revolution that yields the dominant 1X component, while the non-sinusoidal character of the excitation, filtered through the structural transfer path and nonlinear response of the system, generates the higher-order harmonics.

The research landscape that motivated this release is telling. The celebrated Case Western Reserve University bearing dataset, the Paderborn University bearing collections, the IEEE PHM 2012 run-to-failure bearing data, the PHM 2009 gearbox challenge data and the Tsinghua University gearbox datasets have collectively underpinned two decades of progress in data-driven diagnostics, yet nearly all were recorded at constant speed or at discrete speed steps, and almost none address rotor-blade damage specifically. By publishing labelled blade-fault data under genuinely continuous speed variation, the AAU team opens the door to benchmarking speed-conditioned feature extraction, cross-condition fault diagnosis, self-supervised learning and uncertainty-aware diagnostic methods on problems that actually resemble field conditions. The authors are candid about limitations: only five health states are covered, laboratory safety precludes the 3000 to 5000 rpm range of real turbomachinery, and the brake load is fixed. Future extensions, they note, will broaden the damage modes, widen the speed range, add adjustable loading, nonlinear ramps and random speed fluctuations. Even so, for a field long starved of non-stationary rotor data, this release is a welcome and overdue gift.

Subject of Research: A public rotor-blade vibration dataset for fault diagnosis under time-varying rotational speed conditions

Article Title: AAU Rotor Vibration Dataset Under Time-Varying Speed Scenarios

Article References: Li, Q., & Xu, H. (2026). AAU Rotor Vibration Dataset Under Time-Varying Speed Scenarios. Results in Engineering, 32, Article 113172. https://doi.org/10.1016/j.rineng.2026.113172

Image Credits: AI Generated

DOI: 10.1016/j.rineng.2026.113172

Keywords: rotor blades, vibration monitoring, fault diagnosis, time-varying speed, open dataset, envelope order spectrum, predictive maintenance, rotating machinery, machine learning benchmark, signal processing, Rotor, Vibration