Solar Pump for Africa: Size for the Worst Month, Not the Average

The short answer: a solar pump for Africa fails on a design mistake, not a technology problem. The continent has one of the best solar resources on earth, between 4.5 and 6.5 kWh/m2/day, so almost every site has enough sun. What breaks projects is sizing on the annual average and ignoring the cloudy months, or sizing on the borehole depth instead of the water level. Design for your worst month, derate the sun honestly for heat and dust, and the same pump will serve you all year.

I sell these systems into East and West Africa every week, and the pattern is always the same. The buyer has a good borehole, good sun, and a pump that was specified on a spreadsheet built for a European climate. Here is how to specify one that actually survives a Kenyan July.

Why Africa is both the easiest and the hardest place to sell solar pumping

The resource is genuinely world class. Studies of Kenya put global horizontal irradiation at roughly 2400 kWh/m2/year, and practitioners quote a national range of 4.5 to 6.5 kWh/m2/day. That is better than most of southern Europe.

The constraints are what make the job real:

Notice that the sun does not appear once on the problem side. If your project failed, the sun was not why.

Kenya’s sun is not the same in July as in January

This is the number almost nobody checks, and it is the one that decides whether your tank is empty in the dry season.

Mean daily solar radiation on a horizontal surface, kWh/m2/day, FAO data.

Read the Nairobi row again. July gives 3.72, against 6.34 in January. That is a 41 percent drop in the middle of the year, driven by the heavy overcast of the cool season, and it lands right when a rainfed farm has nothing and an irrigated farm is carrying the whole season.

If you designed your system on the 5.24 annual mean and July delivers 3.72, you are about 30 percent short on paper before we even talk about losses.

Now derate it, because panels in 35 C heat do not give nameplate power

Panels lose output as they heat up, typically losing around 0.35 percent of voltage per degree above 25 C, and a panel sitting in full sun at a hot site runs at 60-65 C. Add dust, controller and motor losses, and the practical performance ratio in this region is 0.75 to 0.80, not the 0.85 you would use in a temperate climate. I work with 0.78.

So the rule I use for every African estimate is:

Daily volume roughly equals flow at your total dynamic head x GHI in kWh/m2/day x 0.78.

For Nairobi in July that gives 3.72 x 0.78 = 2.9 hours of full-flow pumping. Not 8. Not 24. Nearly three.

A lot of rule-of-thumb tables quote 5 peak sun hours for the tropics. That is the resource, not what lands in your tank. The difference between 5 and 2.9 is the difference between a system that works and a customer who thinks solar is a toy.

The three numbers you must measure before you choose anything

Forget the borehole depth. It is the number buyers always lead with and it is almost never the number that sizes the pump.

1. Static water level - how far the water surface sits below ground, measured with a dip meter or read from the driller’s log. This is the number that matters. A 100 m hole with water at 15 m is an easy job; the same 100 m hole with water at 85 m is a completely different pump. This trips up more buyers than anything else, and it is why I wrote a separate guide on 100 m deep wells.
2. Drawdown - how far the water level falls once you start pumping. On a weak aquifer this can be 5 m, on a strong one 1 m. If nobody has measured it, allow 5 m and confirm after the first week of running.
3. Lift and pipe friction above ground - the height to the tank plus friction in the delivery pipe. Use about 10 m of pipe per 1 m of head for standard diameters, so 50 m of pipe costs you 5 m of head.

Add them: static level + drawdown + tank height + pipe friction = total dynamic head. The full method is in how to calculate total dynamic head.

Then convert your water need into a flow rate:

Required flow (m3/h) = daily water need (m3) / effective pumping hours in your worst month.

Worked example: 0.3 ha of vegetables near Machakos, Kenya

A real brief, with the numbers I would actually use.

The farm. 0.3 ha (3,000 m2) of vegetables on drip, at 4 mm/day, so 12 m3/day. Borehole 45 m deep, static water level 16 m, drawdown 4 m, tank 6 m above ground, 50 m of delivery pipe.

Head. 16 + 4 + 6 + 5 = 31 m total dynamic head.

Worst month. July, GHI 3.72, so 3.72 x 0.78 = 2.9 effective hours.

Duty point. 12 / 2.9 = 4.1 m3/h at 31 m.

Pump. A shut-off head of at least 1.3 x 31 = 40 m, and a maximum flow comfortably above 4.1 m3/h. That points to 4DSC6-45-48-500: 500 W, 48 V, 6 m3/h maximum flow, 45 m maximum head. Its shut-off head is 1.45 times the duty head, so the working point sits inside the curve rather than near its edge.

One honest warning about that 6 m3/h figure: it is the flow near zero head, not at 31 m. Always read your actual working point off the published H-Q curve instead of assuming the rated flow applies at your head, which is what how to read a pump curve covers.

Panels. 500 W x 1.3 = 650 W minimum, so two 550 W panels in series, 1,100 W. On a DF-48 controller that gives a string Vmp near 84 V, inside the 60-90 V MPPT window, and a Voc near 99 V against a 120 V limit. The wiring detail is in the controller wiring diagram.

Storage. 12 m3/day x 1.5 to 3, so a 20 m3 tank.

What this system actually delivers. In January, with 6.34 x 0.78 = 4.9 hours, it moves about 22 m3/day, enough to push the plot to 0.5 ha. In July, with 2.9 hours, it moves about 13 m3/day, just covering the crop. Same pump, same panels, same pipe. The season does the rest. That gap is what you are designing for, and it is why a tank is not optional.

If you want to run your own numbers rather than mine, put them into the sizing tool and it will work the head and the model for you.

Buy panels for voltage, not for watts: the hot-climate MPPT trap

Here is the part that surprises people who have read European solar guides. In a cold climate the danger is a cold morning pushing panel Voc above the controller limit. In the tropics there is no cold morning, so the Voc risk basically disappears, and the real trap is the opposite: heat pulling Vmp down below the MPPT window.

The practical consequence is that on higher-voltage pumps you often have to add a panel for voltage reasons even when you already have enough watts. Work through it with the real controller windows:

The DF-110 rows are the ones that cost people money. Three 550 W panels look correct on paper, and in a 35 C site at midday they drift to roughly 107 V, just under the 110 V floor of the window. The pump keeps running, so nobody notices; it simply gives less water at exactly the hottest hour of the day. Four panels in series fixes it, and because tropical sites see no cold Voc spike, the 200 V string Voc sits safely under the 220 V limit.

If your duty point needs more than 6 m3/h, this is worth knowing early, because the whole 48 V line-up tops out at about 6 m3/h. Above that you move to the 110 V class, and you should budget four 550 W panels even if the pump power alone would only justify two or three.

Sand, silt and salinity: why impeller material matters more here

A European buyer pumps clean groundwater. An African buyer often pumps from a seasonal river, a sand-lined well, or a borehole in the arid and semi-arid lands where the water is hard and sometimes slightly saline.

That changes the pump you should buy, not the size, the material.

• Plastic impellers (2DPC, 3DPC series) are the budget choice and are perfectly good on clean water with modest running hours.
• Stainless steel impellers (3DSC, 4DSC, 6DSC, 8DSC) are what you want when the water carries sand or silt, when the pump runs 2,000 hours a year or more, or when the water is aggressive. A grit-laden flow chews a plastic impeller in a season or two.

The full trade-off is in plastic vs stainless steel impeller, and the mechanics of sand damage are in solar pump sand handling. If you take one thing from this section: on a scheme in the semi-arid lands, stainless is not an upgrade, it is the specification.

Also test the water before you buy, not after. High salinity or iron changes both the pump choice and the irrigation method.

Mounting, tilt, dust and theft

Four installation details that are different here, and that installation crews trained in Europe routinely get wrong.

Tilt. The usual advice is to tilt panels at your latitude. Kenya sits between roughly 5 degrees north and 5 degrees south, so that rule tells you to mount them almost flat. Do not mount them dead flat. Use 10 to 15 degrees, facing north in the southern hemisphere and south in the northern hemisphere. The small tilt costs you almost nothing in yield and lets rain wash the dust off instead of letting it bake into a crust.

Dust. In the arid north, from Marsabit to Turkana and across the Sahel, soiling is not a rounding error. A heavily dusted array can lose 10 to 20 percent of its output. Plan for a monthly clean and put it in the maintenance contract, because a dusty array looks identical to a broken pump to a farmer.

Theft. Panels are portable, valuable and often the only expensive object on a rural site. Raise the mounting frame well above reach, use anti-tamper bolts, and where possible mount the array on the tank tower or inside a fenced compound with the pump house. On community schemes, community ownership is the only security measure that reliably works.

Cable. Long drops to deep water mean long cable runs, and voltage drop eats the difference between a pump that performs and one that limps. At 100 m of drop cable this is a measurable, real loss, and the correct sizes are in solar pump cable sizing.

A real model ladder for African boreholes

These are actual models from our range, ordered by the job they suit. Every model and every figure below is from our published specification tables.

The progression is consistent across the whole range: 2DPC and 3DPC for small plastic-impeller jobs on clean water, 3DSC and 4DSC when the water carries sand or the pump runs long hours, 6DSC and 8DSC when the job is a scheme rather than a farm. If your project is a community supply rather than one farm, the sizing logic shifts toward peak day demand, which is covered in solar pump for village water supply.

The five things that decide whether the project survives

In order of how often they actually kill a project:

1. A tank sized for the cloudy season. Size storage at 2 to 3 times the daily requirement, not 1. A tank that only covers a good day will be empty in July, and the farmer will conclude that solar does not work.
2. Sizing on the sunny month. Anyone quoting you a single daily volume without asking about your worst month is guessing. Ask them what figure they used and check it against the table above.
3. Parts and service distance. A pump you cannot get an impeller for is a pump with a two-year life. Buy a frame size that is stocked in your region, and keep one spare impeller on site.
4. Security. Fencing, raised mounting and community ownership. I have seen functioning systems stripped for scrap, which is a harder lesson than any technical failure.
5. Water quality tested before purchase. Sand, salinity and iron all change the specification. Test first; it costs almost nothing and it protects the whole investment.

Everything else, head, cable, voltage window, tank size, is arithmetic, and the arithmetic is in this article.

Planning a system in Kenya, Nigeria, Senegal or Tanzania? Put your static water level, daily water need and tank height into the sizing tool, or message me on WhatsApp with your static water level, borehole depth and daily water need, and I will come back with a model, a panel count and a tank size sized for your worst month.