Where does the energy come from? A three-way test for on-board harvesting

Every few months someone proposes putting a turbine on a vehicle. Or a generator on a wheel. Or piezoelectric film in the door panels. The argument that follows is always the same argument, and it is always about the wrong thing: how efficient the harvester is, what the conversion losses are, whether a better magnet or a smarter controller would close the gap.

None of that decides it. What decides it is a question nobody asks first: where is this energy coming from?

There are only three answers, and the answer fixes the sign of the energy balance before a single component is specified. I have started sorting every proposed on-board source into one of three classes before looking at anything else, and it has saved enough time that it is worth writing down.

Class A — external, or ambient

The source draws energy that is present in the environment and that the vehicle was not going to spend.

Photovoltaic conversion is the clean case. The photons arrive whether the vehicle moves or not; intercepting them costs the drivetrain nothing. Thermoelectric conversion of waste heat sits here too, as does piezoelectric conversion of road-induced vibration, and radio-frequency capture from the ambient field.

The energy balance is unambiguously positive. The only remaining questions are magnitude, mass and cost — real questions, but questions of degree.

Class B — recovered

The source recovers energy the vehicle is already committed to losing.

Regenerative braking is the definitive case. That kinetic energy is going to be dissipated as heat in the friction brakes. Recovering a fraction of it is strictly better than recovering none. Regenerative suspension dampers are the same argument applied to energy already being turned into heat in the damper.

The energy balance is positive, bounded by the fraction of the loss that is actually accessible.

Class C — parasitic

The source draws its input from the vehicle's own forward motion, while the vehicle is being propelled.

A turbine facing the airstream is in this class. So is a generator on a driven wheel.

The energy balance is negative during propulsion, and no improvement in harvester efficiency changes the sign.

That last sentence is the one that does the work, so it is worth deriving rather than asserting.

Why Class C cannot be rescued by better engineering

A vehicle moving at speed v through still air experiences a relative wind of exactly v. There is no external wind field. The apparent wind is the vehicle's motion.

A turbine placed in that stream extracts momentum from the air. By Newton's third law the air pushes back on the turbine with an equal and opposite force — which is to say, additional drag on the vehicle, transmitted through the mounting.

The power the drivetrain must supply to overcome that added drag is

P_drag = F_drag · v

and the power the turbine can deliver into the electrical system is

P_elec = η_mech · η_elec · C_p · ½ ρ A v³

where C_p cannot exceed the Betz coefficient of 16/27 ≈ 0.593 — a bound that follows from momentum conservation across an actuator disc and applies to any turbine of any design — and where η_mech and η_elec are each strictly below one.

The force that produces P_elec is the same force that produces P_drag, but it acts on the air at the reduced velocity through the rotor rather than at the vehicle speed. In actuator-disc terms, with axial induction factor a, the power extracted from the flow is T·v(1 − a), while the propulsion power needed to overcome the same thrust is T·v. The ratio is therefore (1 − a): two-thirds at the Betz optimum, before mechanical and electrical losses reduce it further. Add the parasitic drag of the turbine housing itself, which produces no output at all, and the deficit widens.

A perfect turbine, at the Betz limit, with lossless mechanical and electrical conversion, would return at most two-thirds of the energy it costs the drivetrain. This is not a manufacturing problem awaiting a better product. It is a conservation result.

The driven-wheel case is the same argument with the air removed. A generator on a wheel that the drivetrain is turning takes its input torque from the drivetrain: motor → wheel → generator → converter → battery, and every arrow is lossy.

The two exceptions — and they are precise

Class C hardware becomes legitimate under exactly two conditions, and stating them is the design guidance the whole exercise yields.

A non-driven, free-rolling wheel taps vehicle kinetic energy the same way regenerative braking does. If the vehicle is coasting or decelerating, that energy is being shed anyway. This is Class B behaviour and it is sound.

A turbine or generator engaged only during coast-down or braking is likewise analogous to regen, and avoids the steady-state drag penalty entirely. A clutched or electrically de-excited turbine that free-wheels under power and loads only during deceleration is a Class B device wearing Class C hardware.

Both exceptions share one feature: the harvester is active only when the vehicle is shedding energy, never when it is spending it. That is the whole of the rule.

What this changes about the claims worth making

Sorting the sources this way immediately narrows what can honestly be said about each.

Solar. Peak output density follows from irradiance and module efficiency: at the AM1.5 reference of 1,000 W/m² and commercial silicon module efficiencies of roughly 15–22 %, that is on the order of 150–220 W/m². Yield therefore scales with available horizontal area and very little else — which is why a car roof, a bus roof and a trailer roof differ by an order of magnitude, and why any yield figure quoted without the area it assumes is unusable. Against a heavy vehicle consuming on the order of 100–150 kWh per 100 km, the defensible claim is that vehicle-integrated photovoltaics meaningfully offset auxiliary and thermal-management load, especially on duty cycles with substantial stationary time. The claim that does not survive is that it powers the vehicle.

Regenerative braking. The recoverable share depends strongly on duty cycle: it is highest on stop-start urban routes, where braking events are frequent, and lowest on steady highway operation. Nothing about that is novel, and it does not need to be — it is simply true.

The small ambient sources. Piezoelectric and radio-frequency capture are Class A and therefore positive in sign, but small by orders of magnitude relative to traction demand. They are genuinely useful for sensor power — a tyre-pressure monitor that never needs a battery change is a real product — and they are not a propulsion contribution. Multiplying an unmeasured per-device output by a device count is how a milliwatt becomes a marketing watt.

A note on my own patent, since it is the obvious objection

The specification of my vehicle power-electronics family describes eight generator sub-systems, and a wind-turbine sub-system is among them. Someone reading the above is entitled to ask about that.

The independent claim does not depend on any one of them: it recites “at least one first generator”, so the family covers implementations that use different subsets of sources. Describing a sub-system in the specification is not an assertion that every one is commercially attractive, and it is not a specification of a product containing all of them.

What is actually claimed and granted there is the system-level integration: one power processing unit that reads the state of charge of the energy storage device, compares it against a predetermined value, and provides sustained power to the storage device and to the motor simultaneously. The arbitration is the invention.

So the version of that architecture I would build first integrates photovoltaic input, the regenerative return, and one suspension or free-rolling-wheel path — Class A and Class B only. Instrumenting a Class C source during propulsion would measure a deficit, and you do not need the measurement to know that.

Why bother with a taxonomy

Because it costs nothing to apply and it settles the question before any money is spent.

An engineering argument about a proposed harvester can run for weeks on efficiency figures and never reach a conclusion, because efficiency is the wrong variable. One question — does this draw from the environment, from energy already being discarded, or from the vehicle's own propulsion? — sorts it in a sentence. The first two are worth engineering. The third is worth declining, however well built.

Use it if it is useful. It is not mine in any meaningful sense; it is just momentum conservation, applied early enough to be cheap.

Cite this note
A referenced edition of this note, with a permanent DOI, is deposited on Zenodo: Rohera, H. K. (2026). Where does the energy come from? A three-way test for on-board energy harvesting [Technical note]. Zenodo. https://doi.org/10.5281/zenodo.22700047

Hemant K. Rohera is an independent inventor based in Pune, India, and the sole named inventor on three granted patent families in three unrelated engineering fields — bioelectronic medical devices, hybrid energy storage and vehicle power electronics — with grants in force in India, the United States, Japan, South Korea, Canada, Mexico, Australia, Vietnam, Saudi Arabia, South Africa, Russia, Kazakhstan, Germany, the United Kingdom and Switzerland. The full register, with every grant number and date, is at hemant-rohera.vercel.app/patent-register.html.

Search This Blog

Hemant Rohera — Bioelectronics, Energy Storage and Power Electronics