Evaluating the value of fitting solar panels on a bus roof

Solar bus proposal

Evaluating the value of fitting solar panels on a bus roof

When

2023

Client

Nexport (internal)

Team

My role

Would adding solar panels to the roof of an electric bus really add range or save costs? I prepared a data-driven estimate for the energy gains to answer this question. In addition, to flesh out a what-if scenario, I designed a custom air conditioning packaging to maximise rooftop area for solar panels.

Sketch of a bus on the left, with an arrow coming from a grid symbol to signify energy flow. In the middle, a bus depot is drawn, with roof above a charging bus shown with solar panels.
Quick sketch to lay out the broader picture. The key challenge is no longer the electric bus itself, it is the infrastructure enabling charging that's holding back adoption. Solar panels on buses fit into a broader discussion on energy generation and utilisation, with grid power, bus depot energy generation and storage, and on-board generation, storage, and consumption all relevant topics.

Stage 1 – Estimating solar panel energy generation

Nexport’s commercial team was keen to explore solar panels as a value-add option for our buses. Some initial numbers shared for its potential appeared unrealistically high. I deemed it unrealistic a bus roof alone could generate “over 100 kWh (kilowatthour) per day.” If true, that would be nearly 40% of typical daily consumption for a city bus and indeed a massive saving. So I set out to establish a more thorough estimate, using a very clear process for how I’d get those numbers.

My self-formulated brief:

Establish a realistic estimate for daily energy generation of rooftop solar panels on a bus, including Return on Investment (ROI) to build a business case.

Iteration 1 - Basic calculation

Handdrawn calculations that show the calculations to get the roof area, nominal power output for a solar array of that size, and the expected total energy generated per day. In the right-bottom corner, it is shown that solar energy could only generate about 5% of total battery capacity per day.
Calculating daily energy generation by hand, using basic math and data available. The amount generated is equivalent to 5% of overall battery capacity, or 7.5% of typical daily consumption. Compared to pulling that energy from the grid, the cost savings would be around $6 per day.

For the first iteration, I made three assumptions to get started:

  1. We can use the entire roof area of a bus (that is, we ignore aircon units and emergency escape hatches);
  2. Solar panels behave as if placed at a perfect angle (this allows us to use common efficiency figures for fixed installations);
  3. All energy generated can be fully utilised.

A simple calculation will give me an initial ballpark figure:

  • Knowing our bus roof flat area is 2.1m wide and 11.8m long, we have a total area of 24.8 m2.
  • Typical efficiency for a solar panel is around 20% or 200 W/m2 (assuming a nominal 1000 W/m2 of sunlight energy), so we get a total power generation of 4.95 kW;
  • Across the whole day, we take a shortcut and multiply the power output by a factor of 3.8. Using this factor skips a more complex integration of sun position versus panel output at each time of the day.
  • Final outcome is 18.8 kWh generated per day.

This figure of 18.8 kWh is an absolute best case scenario but already less than one fifth of the earlier number!

Iteration 2 - Refining assumptions and layout variations

I began by revisiting the assumptions given earlier:

  1. I would include different configurations of solar panels, including configurations that no longer ignored existing rooftop elements.
    1. Also, I decided to only use existing solar panel components with known dimensions, using supplier data.
    2. The proponents of the 100 kWh/day figure raised that I forgot to add panels on the side of the bus, so I covered that option too.
  2. The angle relative to the sun should be considered when calculating a panel’s efficiency, so I get a better idea of the losses involved. I kept the kW-to-kWh/day factor to avoid complexity, in part because I lacked actual data on panel output versus angle versus sun intensity.
  3. System conversion losses (between panels and where energy is actually used on the vehicle) are still estimated, because proper modelling is too complex at this point. I still assume all energy can be utilised.

This time, I set up a spreadsheet to work through all calculations across various configurations. I also included cost estimates based on supplier info, with some assumptions and estimates for ongoing maintenance costs. Buses get lots of damage while on the road!

Table of six variants: 1. blank roof without panels, 2. standards panels on roof while keeping AC and escape hatch, 3. as number 2 but with panels on top of the AC, 4. full roof covered with PV panels, 5. full roof and side strip covered, and 6. side skirting only, no roof panels.
Solar panel configurations used in my calculations, varying from no panels to entire roof plus side-mounted panels.

The max energy we can expect is now 14.9 kWh/day with roof and side panels fitted but at significantly worse ROI than a plain unmodded roof-only setup, which would yield 9.6 kWh/day. Based on this data, I concluded that (a) modifications to optimise solar panel quantity and placement were likely to yield worse ROI than a more restrained approach (resp. 153% vs 191% ROI over 20 years, generally poor figures with a break-even period of at least 8 years), and (b) final ROI figures are very dependant on ongoing maintenance costs. The slim margins could vanish if these costs proved higher in practice.

Reflection

Ultimately, the low ROI explains why rooftop solar on buses isn’t widespread. While it may be viable, the figures suggest that investing in a fixed installation at a bus depot is a safer investment.

I was happy with these outcomes relative to the time spent. It allowed us to assess likely outcomes and create a much clearer plan for how to proceed, if we were to investigate further. I effectively demonstrated a blueprint for how to approach similar innovation proposals, a significant process improvement for the Engineering team.

If I were to do another step, I would need to accurate model both panel performance and system performance to get a better idea of generated energy and how that energy may be utilised.

Stage 2 – Custom aircon concept

If we’re really keen to fit as many solar panels on a roof as we can, we’d ought to minimise the space taken up by the airconditioning unit. This would require a customised unit to split the parts that could go off the roof to other places, and leave just the condensator and noisy compressor parts on the roof. While a colleague looked into the internal layout with the lowered components, I primarily focused on the exposed roof components.

There are several ways to arrange the components, but the most viable overall packaging is some variant on a candybar shape. Using approximate geometry for the compressor and condensator tubing and fans, I mocked up eight variations.

3D render of 8 different shapes for the custom AC packaging, varying between a rectangular to a triangular cross-section.
Overview of eight shape variations for the split-system aircon's roof unit, shown with shadows. It is clear the triangular cross section of the toblerone shape minimises shadows.
Video overview of the shape variants described above.

The variations were then compared for their packaging efficiency, manufacturability for low volumes, and importantly, their minimisation of shadowing of surrounding solar panels. Shadows, even on a small part of a solar panel, reduce the power output drastically, so minimising shadows is desirable. A boxy shape would offer efficient packaging but the high walls would throw large shadows at low sun angles. A triangle shape (like a toblerone) would be the opposite. I settled on an ‘uneven toblerone’ shape as the best mix of packaging, fan placement, and shadowing reduction.

Shadows as a variation of time of the day, for mid Summer, mid Winter, and the equinox. In Summer, with the high overhead sunlight, differences aren't pronounced but the lower sun angles in Winter reveal greater contrast between designs.
3D render of a toblerone-shaped aluminium box on top of a semi-translucent bus roof, with additional notes pointing out components inside the AC unit. A technical drawing with dimensions is shown on the left.
View of the final concept design, with an early mockup for the in-ceiling evaporator box with blower fans and cooling rods. A dimensional drawing is shown in the upper left.

The final design was mocked up in 3D renders and used by executives for promotional purposes. A few draft technical drawings were prepared, but further technical development was slated for the medium-term future. In my view, it is unlikely the energy gains outweigh the potential development costs. Nonetheless, this was a useful exercise to explore design options and stimulate interesting discussions.

3D render of a bus with solar panels fitted, surrounding the smaller than standard custom AC unit.
Shown from above, the custom AC unit offers more space for solar panels and, not shown here, limits the AC unit's visibility on the roof when seen from the streetside. The latter gives a cleaner overall appearance for the vehicle.
3D render offering a top-down comparison of bus roofs with a regular AC (top) and smaller AC (bottom). It is clear the bottom bus can fit more PV panels.
Top-down view of a bus with a standard AC unit (top) and the slimmer concept (bottom). The bottom bus could fit several extra solar panels, for a 4 m^2^ (+23%) area improvement, which should yield up to 800 Watt extra power, for about 3 kWh extra per day.