The Forgotten Engineering of Windmills: From Persia to Modern Wind Farms

The schoolroom story of windmills compresses an extraordinary engineering history into a single image: Cervantes' Don Quixote tilting at them on the plains of La Mancha. The actual story is a thousand-year arc of structural and aerodynamic problem-solving that quietly worked out the answers to questions we now solve again with carbon fiber and computational fluid dynamics. The current wave of wind energy is not building on a clean slate; it is recovering and extending solutions that were already worked out in wood and brick.

The earliest documented windmills appear in eastern Persia around the ninth century CE, in the region around Sistan that still has some of the highest sustained winds in central Asia. These were vertical-axis machines — the rotor turned around a vertical pole, like a paddlewheel laid flat. The blades were fabric stretched over wooden frames, set into a cylindrical wall with one side opened to catch the prevailing wind from a known direction. The output drove millstones for grinding grain. Vertical-axis design has the advantage that the machinery does not need to track wind direction, but the disadvantage that only the downwind half of the rotor produces useful work — the upwind half resists. The Persian builders solved this with the half-walled enclosure, which let the wind through to one side only.

The horizontal-axis windmill — what most people picture when they hear the word — appears in northern Europe in the twelfth century. The earliest records are English and Flemish from the 1180s. The horizontal-axis design has fundamentally different problem set: the rotor must face into the wind, so the entire machinery has to rotate to track wind direction. The first solution was the post mill, in which the entire mill house pivoted around a single central oak post. The structure was massive — a multi-ton timber-framed building balanced on a vertical post and turned by hand into the wind via a long tail beam.

The post mill was operationally exhausting. The miller had to push the tail beam to reorient the entire building every time the wind shifted, and a wind shift during a storm could make the building catch wind from the wrong angle and overstress the structure. The fourteenth-century innovation was the tower mill — a fixed brick or stone tower with only the cap rotating to track wind direction. This decoupled the building from the wind-tracking problem and allowed taller, more massive structures than the post mill could support. The cap was rotated by hand-cranked wheels and gears that could be operated from inside the tower, a major safety improvement.

The Dutch and Flemish refined the tower mill into the technology that defines the regional landscape. Two innovations stand out. The first is automatic yaw control via the fantail — a small secondary windmill mounted at right angles to the main rotor, geared to the cap-rotation mechanism. When the main rotor was pointed into the wind, the fantail saw zero wind; if the wind shifted, the fantail caught it and spun, automatically rotating the cap until the main rotor was back in line. Edmund Lee patented this in England in 1745 and it spread quickly across northern Europe. It is a beautiful piece of mechanical control theory implemented in oak and iron, and the same principle (turning the rotor into the wind via mechanical feedback from a secondary sensor) is still how passive yaw control works on small modern wind turbines.

The second Dutch innovation is the polder mill — a windmill not for grinding grain but for pumping water out of below-sea-level land. The Beemster, the Schermer, the Purmer, and dozens of other Dutch polders were drained in the seventeenth century by chains of windmills lifting water in stages from drained ground up to canals that carried it to the sea. The reclaimed land totaled hundreds of square kilometers, and the engineering required individual mills capable of running continuously for decades while transferring multi-ton water flows in conditions where a few days of failure would re-flood the land they had drained. The economic logic was extraordinary: the windmills produced new land, which was then sold to fund more windmills, which produced more land. Modern offshore wind farms have a faintly similar economic structure but rarely the same direct closing-the-loop quality.

The blade aerodynamics of these classical mills were solved empirically over centuries before any of the underlying physics was formalized. The blades had a characteristic twist — steeper angle of attack near the hub, shallower toward the tip — because the relative wind speed seen by the blade increases linearly with distance from the rotor center. This is exactly the same blade-twist solution used in modern wind turbines and in aircraft propellers, and it was worked out by trial and error by Dutch millwrights who had no theory of lift, no concept of the angle of attack as we now formalize it, and no instrumentation beyond their own observation of mill output across blade redesigns. The American smock-mill builder Stephen Hooper introduced the roller reefing sail in the 1780s — adjustable canvas-and-roller blades that could be furled or extended like sails on a ship, allowing the miller to control rotor speed in changing wind. This is functionally the same as the variable-pitch blade control on modern turbines, implemented mechanically with rope pulleys and rollers two centuries before electrohydraulic actuators existed.

By 1850 there were perhaps 200,000 windmills operating across Europe. The American adaptation took a different form: the multibladed water-pumping windmill, designed for the Great Plains where farms needed to lift water from deep wells but did not need the heavy mechanical output of a grain-grinding mill. Daniel Halladay patented the self-regulating multibladed wind pump in 1854, and by 1900 something like six million of them dotted the American landscape. They were the dominant water-pumping technology of the western United States until rural electrification in the 1930s and 1940s replaced them with electric pumps. A surprising number are still operating; their slow rotation and low-stress design make them remarkably durable.

The collapse of windmill technology was nearly total. Steam power displaced grain mills in the late nineteenth century. Electric pumps displaced water-pumping windmills in the early twentieth. By 1950 the working-windmill population in Europe had declined by perhaps 95%, and the surviving structures were heritage objects rather than operational infrastructure. The institutional memory of the trade — the millwright's craft, the empirical knowledge of blade design and mechanism repair — survived only in a handful of preservation societies and a few working historic mills.

The modern wind-energy revival, beginning seriously in the 1970s in Denmark and California, was not a continuation of this tradition but a re-derivation of much of it from aerospace engineering. The blade-twist optimization, the yaw control via servos, the variable-pitch control via hydraulics — all of these were known to nineteenth-century millwrights, lost during the steam-and-electric interregnum, and reinvented from scratch using NACA airfoil profiles and finite-element analysis. There is something humbling in the discovery that the basic problems of wind power were already solved well enough to drain the polders and grind the grain of pre-industrial Europe, and that the engineering vocabulary for solving them developed twice in human history because the first round was forgotten.

The Vestas V164 turbine, currently among the largest production wind turbines in the world, has a rotor diameter of 164 meters and produces nine megawatts. The largest classical Dutch tower mills had rotor diameters of about 30 meters and produced perhaps 20 kilowatts of mechanical output. The ratio of output to size has grown by a factor of about 10,000 in two centuries. But the silhouette is recognizable, the underlying control theory is the same, and the question of how to point a giant wind-catching device into the wind without breaking it is the same question millwrights and Dutch polder builders worked out in oak and iron three centuries before anyone wrote down the Navier-Stokes equations. The classical windmill is one of those technologies that look quaint in retrospect because the visual style of the era is no longer current, and which prove on inspection to embody more genuine engineering depth than the surrounding mythology suggests.