Ammonia-powered car - impromptu description of fairly advanced concept

I recently calculated that current lithium production, if shared equally among all humans, would be just enough to produce around 0.3 kWh of LiFePO4 batteries per year (they even don't require cobalt). I would like a car that can drive for 3 hours at maximum speed, and then be quickly refueled. Small car requiring 22 kW of power during that scenario would require 66 kWh of batteries, which would weight approximately 500 kg. Pure hydrogen storage is bulky and extremely dangerous, that is why using ammonia as its carrier seems to be best option (it is the hydrogen carrier with highest volumetric density, that is easily stored as slightly pressurized liquid). If maximum payload of a car is 500 kg, and then I add 500 kg of batteries, then I need many components of a car to be twice as strong and powerful.

Initially I was thinking about a cabin motorcycle with a smaller engine, but that might have been a little dangerous, because of side winds, so I changed my priority to something that I could actually drive (I do not hold motorcycle licence). This car would be very peculiar, because I want rear windshield to be able to retract and extend. In retracted position aerodynamics would be better, but headroom for passengers in the rear seats would be extremely limited. Inflatable seal could be a part of main car body, it would prevent leaks between it and rear windshield. Other body variant could also be offered, such as a hatchback (with larger rear trunk and higher rear doors), a Cybertruck-style pickup (with buttresses sloping to the back), and a taller MPV.

I would like to make smallest possible car that can fit comfortably 4 people that are 2 meters tall. It is supposed to be only 1.4 m wide and tall (ride hight could be set by the driver). Wheelbase would be very long, 2.8 m, and overhangs short, giving total vehicle length of only 3.8 m. Flat engine would be located in the back, below rear trunk, so steering angle of front wheels could be large, as front-mounted engine restricts how much wheels can turn. Obviously front trunk would also be available.

By flat engine I understand transversely mounted horizontal inline-3. Engine heads would be close to the rear bumper, connecting rods would be quite long, so that balance shafts would fit around them. Four of them would be placed in total, so that eccentric masses on them would individually cancel vibrations of all the pistons. Counterrotating pairs of eccentric masses on two of them would balance first harmonic of forces produced by the the pistons changing their velocity. Another sets of eccentric masses on balance shafts, turning at twice the ratio of the crankshaft, would eliminate second harmonic. Acceleration, or force, of a piston can be viewed as as a sum of two sine waves, of which one of them has twice the frequency.

I published on this blog two earlier concepts of an ammonia engine. One was based on pre-pressurized two-stroke Lenoir cycle with 8 valves and 4 spark plugs per cylinder. Later I changed my mind and I wanted to do a four-stroke with 4 valves per cylinder and 0 spark plugs. This construction would be an HCCI engine with extremely high compression ratio in the order of 40:1. Ammonia flame propagates very slowly, so igniting whole charge all at once, would be very beneficial. Unfortunately too steep rise in pressure would result in ringing and possibly destruction of an engine, because of that lean mixture has to be used. 40:1 compression ratio is probably unachievable in practical, lightweight, construction, so this lean mixture would have to be highly preheated (it is easier to achieve autoignition temperature when initial temperature of a gas before compression is higher). Density of charge in HCCI engine would be very low, and displacement high.

My current concept combines characteristics of the previous ones. On the rim of cylinder head (each cylinder has its own) 8 valves are located. Three of them are open during exhaust, four of them would admit air during intake, and the last one would be admitting gaseous ammonia. I hope that mixing of gases inside cylinder will work well enough. In the center, module with 4 spark plugs located on its rim, would be placed.

This would be a four-stroke construction. My analysis lead me to conclusion that four-stroke engines won with the two-strokes, even uniflow ones with direct injection, because of better control of intake air (you can simulate Atkinson cycle for example), and scavenging. Intake ports at the bottom of cylinders probably increase wear on long and heavy pistons.

Valves would be of a special type. Each would be attached to a spring, making it a simple dampened harmonic oscillator. On top of this main valve, small hydraulic piston would be located, with two voice coil motor operated small valves per each side of the small main valve piston. Additive manufacturing would be great in creating complex system of channels for hydraulic fluid. Below moving parts of voice coil motors many circular channels that allow passage of fluid would be present, they could be opened or closed by voice coil motors. Two voice coil actuators could be used to increase pressure on either side of the small piston (adding energy to the harmonic oscillator), two other ones would allow fluid from the either side to exit, and if they would be both open, spring would be pumping hydraulic fluid through low pressure channels. Closing of all the voice coil valves would stop main valve in a given position. Voice coil valves should have geometry, that slows them down, when they are near limiting positions (narrowing channels around voice coil actuators for example). Hall sensors could be used to determine positions of main valves. Voice coil valves should be able to allow flow of fluid in one direction only.

Such precise control of main valves would allow to run engine in reverse. Normal four-stroke that runs backward has this problem, that after exhaust into intake manifold would take place, intake from exhaust manifold would occur (two-strokes do not have that problem). Precise stopping of main valves at the right times could replace the throttle. Cylinders could be easily disabled from cycle to cycle.

Quick combustion will be achieved by high compression ratio, of "only" 22:1, which is similar value to not turbocharged Diesels. Ammonia-air mixture compressed that much would start to ignite, but spark plugs, that are spaced apart, would made progression of combustion much faster.

I would like to use simulated Atkinson cycle, with valve opening timings different from regular constructions, so expansion ratio would be in order of 30:1. Control of timing of air and ammonia admission would allow change of compression ratio, which would work great during cold starting for example.

Ammonia and high compression ratio mean lots of NOx emissions. It would be best if more molecules of nitrogen oxides would be present in exhaust gas that the ones of ammonia, as control of gas mixture entering selective catalyst reduction system could be controlled by adding appropriate amount of ammonia before SCR (selective catalyst reduction). Onsite generation of additional NOx would be much more problematic, than "injecting" fuel, that is already stored onboard in large tank. Not mounting SCR system would be beneficial from the standpoint of combating global warming, as nitrogen oxides and ammonia cool the atmosphere, but this probably should not be done.

I would like to make piston and valves out of steel, while head and cylinder liner would be made out of cast iron. Steel piston would be quite heavy and could have problems with cooling, but it would have similar thermal expansion coefficient to the cast iron liner, so closer tolerances and smaller leakage could be possible. Cast iron can withstand higher temperatures and contains large amounts of slippery graphite, with would slightly resolve potential problems with liquid lubrication.

Powerful hydraulic system should make adaptive hydropneumatic suspension possible. In addition to regulating ground clearance, stiffness of suspension could also be regulated, by making passage of fluid in each of the dampening parts near the wheels narrower or wider.

Power for the hydraulic pump would come mostly from the engine probably, but it would be great if additional power from electric motor could be added. In simplest version additional pump would have dedicated motor. Possibly some CVT transmission or epicyclic gears like in Toyota Prius could achieve this goal.

Leveling of the load on each of the axis should be achieved by increasing our decreasing number of moles of gas in the suspension spheres (additional high pressure air pump would be required).

Double wishbone suspension should be used to connect front wheels to the rest of the car. MacPherson struts are compact, but do not work well with large ride height changes. At the back training arms would be used. They should be compact enough to allow connection of the chains (or belts), from the outputs of differential to the wheels. Sprockets should be concentrical with the trailing arm's pivots. Different number of teeth on them would be responsible for the final drive ratio.

Inside hollow output shaft of gearbox, which is concentric with the pivot of the trailing arm (it is located in front of the centers of rear wheels), a spur-gear differential would be located. With a simple open type differential, which splits momentum of input shaft in half on the output shafts, there is a possibility of one wheel being on a slippery surface, which means, that if speed of input shaft is an average of speeds on the output shafts, one of the output shafts can speed up rapidly, using the torque to increase rotational energy of the wheel, while the other wheel can not overcome stiction, making rotational speed of the slipping wheel twice that of the input shaft. Braking slipping wheel could improve traction considerably.

When it comes to the gearbox, I would make it mechanically a simplest possible 6-speed construction. Constantly meshed gears would be connected to the input shaft by the dog clutches, possibly without any mechanical synchronization, because computer should precisely predict when "teeth" should interlock, while controlling engine and motor-generator.

First gear would be similar to the second one in other cars, because of low great ratio and higher speeds that it can achieve. First gear would be simulated by powerful motors providing lots of torque right from the start. I would like 100 Nm of max torque from the motor connected in parallel to the engine, and 100 Nm from the slower speed and weaker motors in the front wheels each (in this case it is a series hybrid). Making a 4WD vehicle is important, as road conditions can can get difficult from time to time. During starting all the motors would be pulling the car (unless power to the front wheels is cut off or limited), and after some speed is reached engine "kicks in", while it is revolving slowly, because it is in equivalent off a second gear, not the first one. Maximum speed of the car probably will be reached at the 4th gear, I am hoping that engine running at 4 000 RPM will be able to produce 22 kW of power (generating 52.5 Nm).

There will be two overdrives. First one will not be able to reach maximum speed on the power of engine alone, but with additional burst off power from motors it will be able to drive a little faster for a while. I will call it "Super Overdrive", after a Billy Idol song. Lithium ion capacitors, that have between 1 to 2 kWh capacity will have to recharge after being depleted, so periods of underpower may come after using "nitro" (no nitrous oxide will be involved of course, but old video games liked to call short bursts of power that way.). There will also be an "Economy Overdrive", that will also allow to cruise at low RPM and high torque.

To make construction strong, yet lightweight and cheap I would like to use isogrids whenever possible. They are usually expensive, because large chunks of metal are substracted from initially massive parts. But extruded components, that are six lines radiating from the single point in cross section, could be welded to make array of triangles.

I will add additional illustrations and information in the future.