[WIP] Solar sailing with openinverter

[giplt]

I want to convert an inland cargo barge (Binnenschiff) to electric propulsion and purchased a second hand Kia Niro EV motor to that end. Eventually I would like to cover the ship's cargo hatches with solar panels and start sailing fossil free on energy harvested in situ.

Searching for an appropriate controller I came across the openinverter project. Truly delighted by the dedication to open source principles and the elegant balance between completeness and simplicity of the design. Thanks johu and all other contributors for all the time and skills invested!

Piecing together available documentation at kia-hotline.com and doing some additional measurements I believe these are the specs of the motor:
- Type: Interior Permanent Magnet Synchronous Motor
- Power: 150kW (204hp)
- Pole pairs: 4 (15 rpm/Hz measured with a low power variable-frequency drive)
- Voltage: 3 phase 356V, 245-421V input range for OEM Electric Power Control Unit
- Rotational speed: max 11200rpm, 3800 - 8000 rpm for max power
- Inverter frequency: 253 - 533Hz for max power, max 747Hz (derived from above specs for rotational speed and pole pairs)
- Torque: max 395Nm
- Sensors: resolver (+/-/S1/S2/S3/S4) and temperature (GROUND/SIGNAL), available through KET MG644803-5 connector

I just ordered the rev 3 main board which supports FOC but now I'm a bit in doubt as to what to do next. From the wiki, webshop, the EV Fundamentals video's and other docs I suspect that openinverter started as a full kit (with gate drivers and sensor boards included) but shifted towards repurposing inverters from donor EV's. So I think my options are either finding a fitting Kia/Hyundai power control unit or to order gate driver and sensor PCB's elsewhere and build the power stage myself around a set of appropriate IGBT half bridge modules. Is that correct?

Does anyone have experience with marine aplication of the openinverter design? Anything I should consider that is fundamentally different from a car?

And would there be a compelling argument to go look for a donor Kia/Hyundai Power Control Unit rather than build the power stage myself, or the other way around?

Also, does anyone have a clue as to whether the temperature sensor can be connected straight to the main board? I'm not quite sure what kind of sensor is inside the motor...

[crasbe]

The pictures you posted... that's the boat you want to power with that motor?

[Ev8]

The main issue with marine applications is the power requirements tend to be high and constant to push and keep pushing the boat through water, cars need low constant power as air is much easier to push though than water! with short peaks of high power for acceleration, an ev motor rated at 150kw is its peak power which it will not sustain for more than a minute or so, I’d personally be looking for a much bigger motor, something from an electric bus maybe… not to say the Kia motor won’t move the barge but it’ll be disappointing I would have thought

[johu]

Regarding the inverter you could try and use a Prius inverter with high voltage. Alternatively a Leaf Gen2 or BMW i3 inverter for which drop in logic boards exist. Building your own power stage has become somewhat deprecated.
I know that Tesla LDUs are being used in boat applications apparently to the satisfaction of their users. I reckon if you just cruise along slowly power demand might be low enough plus you have the ability to install a water-to-water heat exchanger for better cooling. You can always obtain a larger motor if the Kia one proves unfit.

[giplt]

Thanks for the replies!

For some background: the original motor that moved the ship was 175kW (235hp) and these things don't move fast. Transport by water is very efficient at low speeds. For comparison: this ship has the capacity of about 16 trucks. But indeed, the constant power that this EV motor can sustain it the key variable that I couldn't find anywhere.

Overall, ship power is roughly proportional to the cube of the speed, so doubling the speed requires 8 times as much power. The precise calculation is rather complicated but the best estimate I could make is that we should need about 10kW for 6km/h (which is the legally required minimum speed), 34kW for 9km/h and 80kW for 12 km/h (which is a common cruising speed). All well within the power range of the EV motor. Maximum power is used only for manoeuvring situations.

But only testing will tell what is really needed in terms of motor power, battery capacity and what could be possible in terms of solar panels. The EV motor is actually so small that I can simply couple it to the propellor shaft without having to do any reconstruction in the machine room. This this will make for an affordable first test drive to discover how ship and motor behave while having the existing diesel engine standby if needed. Also, I don't have to bother about regulations because I don't change anything to the certified existing propulsion.

As for the inverters that johu mentioned, would these tolerate a rectified 380V from a generator? That would be the fastest (and cheapest) route for testing in a no-battery setup.

[Jacobsmess]

Would a hybrid setup perhaps be better, utilizing the lexus gs450h gearbox as well?

[johu]

Don't talk him into fossil
The Prius inverter can take 380V rectified while the other two will overvolt

[janosch]

Exceptional!

I live on a narrowboat in England myself, much smaller scale.

I have 3 solar panels, so about ~800W peak, that is not enough for a fridge or freezer in winter, but good for most day to day appliances. It looks like you have 10x at least in surface area.

my displacement is 18t, Diesel is a Kubota with 28HP, you can take that against tide (a bit), or up a fast river. If you want to do river cruising I would seriously look at Tesla drivetrains or something else that closer matches your output. What is displacement of your boat? 200t?

Space: You have so much space! You could go LiPo battery bank, or have modules from 2 - 3 cars in parallel in your engine room.

Some thoughts to your points:

Maximum power is used only for manoeuvring situations.

That is true if you are only on canals, there are other considerations if you want to go upstream on the Rhine in spring when current is fast, then you need more power and constant.

The EV motor is actually so small that I can simply couple it to the propellor shaft without having to do any reconstruction in the machine room.

Direct drive loses you a bunch of torque, 8:1 are typical reductions in automotive, I think marine gearboxes are 3:1 or something?
Just checked, mine is 2.5:1 reduction.

As for the inverters that johu mentioned, would these tolerate a rectified 380V from a generator?

Lets assume "yes" for now: what kind of generator have you got? How are you going to test if the motor with 80kW output is powerful enough if you only have a 3kW generator to feed him?


Fantastic project, keep us posted!

[giplt]

Here's some more background information.

A useful resource on ship propulsion and resistance can be found here: https://www.usna.edu/NAOE/_files/docume ... er%207.pdf

Especially this graph is instructive:


It's not a cubic power-relation but rather a complex interaction of various ship properties, most notable the hump caused by wave making resistance. The specific relation varies greatly depending on hull shape and surface lining, draft, current and wind.

Maybe it's worth mentioning that our ship will not be fully loaded. We converted it into a theatre and museum, see https://serendiep.nl. This make our case stand out from regular cargo shipping and gives us a numer of advantages:

- Although still considerable at about 150t, the water displacement is rather modest for a cargo barge. Also, the hull shape is such that the wave making really kicks in only above 14 km/h.
- Our sailing season coincides with sun availability, as this happens to be when audiences are out
- We don't sail on a daily basis, so we can solar-charge a battery for one or two weeks in row and then sail to the next city. Also, we can afford to make use of tide or preferential weather as we are less bound to strict shipping schedules to be met.
- As johu already pointed out, we have massive potential for watercooling both motor and inverter.

I can make a rough estimate for the needed power and also for the reduction, but testing in practice is vital to determining the best setup. Our gen set is 15kW, so the idea is to do a cheap first test with 1:3 reduction and 15kW power supply, and get a lot of data out of that.

Will look out for Prius Gen 2 inverter and see if we can get the motor spinning.

Next challenge is to find out the specifics of the Kia motor spline and cooking up a transmission to the propellor axis.

[celeron55]

If you know how much diesel it takes you to sail somewhere with the barge, you should be able to pretty much just calculate the electricity requirement as 1 liter = 3 kWh (of course realizing the accuracy isn't very good and actually depends on various factors). I would expect with one EV battery pack you won't be getting very far, so you truly need lots of solar panels for a continuous flow of energy while cruising.

You can also use the same rough estimation to estimate your required average power. Just divide energy by time. You can assume an EV motor won't be able to produce 1/2 of its maximum power continuously. Possibly more like 1/4, but that depends. Like, it's probably about 1 kg of motor = 1 kW of continuous power, as silly as that sounds.

[giplt]

It's precisely these calculations that have set off our EV-conversion plan. Here we go:

First, some key figures.

• The ship measures 42.4 x 6.6 m with a 1.3 m draft.
• The current engine is the Detroit Diesel 8V-92TA, a 369 hp two-stroke cycle, V-block diesel engine by GM. The engine is connected to the propellor shaft by a Twin Disc transmission with a 4.13 : 1 reduction.
• At full throttle the ship reaches a max speed of about 18 km/h with the engine running about 1750 rpm.

The relation between motor power and ship speed is rather complicated but overall can be approximated by a cubic relationship. Sailing twice as fast requires 8 times more power.

This yields the following equations:

P = Pmax * (V/Vmax)³, or
V = Vmax * ∛(P/Pmax), with
Vmax = 18 km/h and
Pmax = 369 hp = 271 kW.

Caveat: as noted above, much depends on the specific hull shape, especially with respect to its wave making properties. Our ship starts making waves only above 14 km/h so possibly the relationship is somewhat more forgiving below that speed.

We did one measurement sailing 198 km in 16 hours, which took 281 liters of diesel. This averages to 12.4 km/h at 89kW according to above equations. Caveat: This calculation makes use of the average speed. In practice parts of the trip were faster, others were slower. According to the equations the faster bits were consuming more energy proportionally than the slower parts must have spared.

We can check this against another approximation that can be derived from some key figures on the engine. Wikipedia states an energy content of about 36 MJ/l for diesel fuel and an engine efficiency of up to 55% for large two-stroke marine engines. This gives us 55% * 36e6 J/l * 281 l / (16 * 3600 s) = 97 kW. This is clearly in the same ball park. It even seems to account for the lesser efficiency at non stationary use of the engine and a probably overestimated efficiency of the 30 odd years old engine.

Anyway, the equations seem good enough to work with for now. This means that, utilising the full capacity of the 15 kVA generator we expect the electric propulsion to settle at around 7 km/h.

The big assumption here is that new losses in transmission and electronic conversions cancel out against profitable hull shape at lower speeds and unnaccounted disadvantages from above calculations.

As for solar capacity, the hatches that enclose the cargo hull measure about 132 m². Using standard 1.6 x 1 m panels the most efficient tiling could accommodate 72 of them. With an average 240 Wp for second hand panels we could maybe harvest around 17 kW in optimal weather conditions.

In the mean time, the PCB from johu and the Kia motor connector have arrived, Prius inverter still in transit.