Industries / Industrial equipment and services / Industrial machinery / Generators
Alternators
This work builds the electrical half of a generator: the spinning rotor, the fixed stator with its copper windings, the cooling that carries heat away, and the exciter that feeds current to the rotor's magnet coils.
5
Milestones achieved, from 1866 to 1979.
1
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1866
Self-excited dynamo
Werner von Siemens found that a machine could build its own magnetic field from the current it was making, so it no longer needed weak permanent magnets. Siemens's history says this made large amounts of electricity cheap to generate; he patented it in Germany and England in 1867, and the dynamo was ready for series production in 1875.
LessonSiemens's own history says others were working on the same idea at the same time, so the credit rests on who patented and sold it.
Credit
Werner von Siemens, Siemens & Halske
1884
Turbine drives the generator directly
Charles Parsons coupled a steam turbine straight to a generator turning at about 18,000 revolutions per minute, where generators had been driven by slower piston engines. The Science Museum calls the machine "the forerunner of the turbo-generators that today provide most of the world's electricity."
Credit
Charles Parsons, Clarke, Chapman, Parsons and Co., Newcastle upon Tyne
1937
Hydrogen replaces air inside
General Electric put what it called the first hydrogen-cooled turbine-generator into commercial service: a 31,250 kVA (kilovolt-ampere, a measure of electrical output) machine at 3,600 rpm for the Dayton Power & Light Company. Hydrogen is light and carries heat well, so the same frame makes more power; oil seals on the shaft keep the gas in. GE reported more than a million kVA of such units in service or on order by the end of 1937.
Credit
General Electric; Dayton Power & Light
1957
Brushless excitation, rotating diodes
Engineers mounted silicon rectifiers, diodes that turn alternating current into direct current, on the spinning shaft, so a small exciter could feed the main rotor without carbon brushes rubbing on slip rings. It was first described for 40 and 60 kVA generators on the B-58 bomber, with rectifiers built to survive 15,000 g; a 1959 paper applied it to utility generators.
Credit
H. W. Henkels, H. R. Camp, J. R. Fortier, W. J. Shilling and R. E. Smith (1957); E. C. Whitney, D. B. Hoover and P. O. Bobo (1959)
1979
Cracking retaining rings
Retaining rings are steel bands that hold the ends of a rotor's winding against spin forces. One failed on 13 April 1979 at Sweden's Barsebäck-1 nuclear plant, and the US Nuclear Regulatory Commission counted 32 ring failures, 6 of which caused major generator damage. EPRI found in 1993 that rings of 18% manganese, 5% chromium steel crack under stress when moisture gets in, told utilities to keep them dry, and noted a newer alloy that resists this; a 2002 trade article names it as 18Mn-18Cr.
LessonA steel picked for strength failed from ordinary damp, and the fix was a new alloy plus keeping the rings dry.
Credit
US NRC; EPRI
Watch for
Generators for data-centre turbines
Baker Hughes says it will supply 25 BRUSH DAX 7 generators, two-pole and air-cooled, to be driven by Boom Supersonic's 42 MW Superpower turbines; with 6 ordered in 2025, Boom has firm orders for 31 generators of about 1.3 GW, for delivery from mid-2026 through 2028. The order was tied to Crusoe's data centres, and Crusoe withdrew from the turbine launch in September 2026.
Achieved whenthe generators run on a customer site, reported by someone other than Baker Hughes or Boom.
Claimed by
Baker Hughes, February 2026
Open
01
Small alternator fault heat maps
An open, labelled set of thermal images and electrical readings from a small brushless alternator of a few kVA on a test bench, covering faults put in on purpose: a blocked air inlet, an unbalanced load, one open rotating diode and an overloaded exciter.
The briefCredit · unclaimed02
Measured motor-start voltage dip
A public set of bench measurements of how far an alternator's voltage drops when a motor starts across the line, compared with what simple formulas predict from the alternator's published reactance, the per-unit figure that sets how deep the sag is.
The briefCredit · unclaimed
Open 01 · the brief
Small alternator fault heat maps
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
An open, labelled set of thermal images and electrical readings from a small brushless alternator of a few kVA on a test bench, covering faults put in on purpose: a blocked air inlet, an unbalanced load, one open rotating diode and an overloaded exciter.
Why it is still open
The open thermal fault data sets found are for induction motors, not alternators, such as the Najafi set on Mendeley Data with stator shorts, a stuck rotor and fan failures. Zenodo searches for thermal images of alternators and synchronous generators returned nothing, and a 2025 GitHub repository on generator partial discharge holds only an empty README.
Built with what exists
A small portable or bench generator set; a FLIR C5 or MLX90640 thermal camera; plug-in resistive loads; a Raspberry Pi or Arduino logger with a power meter; Python with OpenCV; Zenodo for release.
First step
Borrow a 2 to 5 kVA brushless generator and log thermal images and readings in the healthy state at no load, half load and full load.
Done looks like
At least 6 fault conditions plus a healthy baseline, each at 3 load levels, with images matched in time to voltage, current and temperature, published under an open licence with a baseline classifier and its accuracy on a held-out session.
Open 02 · the brief
Measured motor-start voltage dip
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public set of bench measurements of how far an alternator's voltage drops when a motor starts across the line, compared with what simple formulas predict from the alternator's published reactance, the per-unit figure that sets how deep the sag is.
Why it is still open
An open calculator, Standby Sizer, was published on GitHub in 2026, and its README says its figure "is an estimate (X″d model), not the alternator's published starting curve." Makers' sizing tools are closed. Searches of GitHub and Zenodo in November 2026 found no measured data set of motor-start dips on generators and no study comparing open predictions with measurements.
Built with what exists
A small generator set of 5 to 20 kVA; induction motors of 1 to 7.5 kW with flywheel loads; a power-quality logger such as a Fluke 435 or a PicoScope with isolated probes; Python; the Standby Sizer code.
First step
Record 10 starts of one small motor on one portable generator and compare each dip with Standby Sizer's estimate.
Done looks like
Dip and recovery-time readings from at least 20 starts covering 3 motor sizes, 2 alternators and 2 starting methods, released as open data, with each formula's average error in percentage points of voltage.