Solar Pump for Southeast Asia: The 10 Metres You Add Yourself

The short answer: in Southeast Asia the sun is not the problem and the water is usually shallow. Southeast Asia gets 4.1 to 6.7 kWh/m2/day depending on the month and the island, which is more than enough to pump with, and groundwater in the rice plains is often only 5 to 15 m down. What actually breaks systems here is the head you add yourself - where you put the pump, how much hose you run, how high you push the water - plus the one wet-season month when the sun drops, and coastal salt. Get the head down and you need a smaller, cheaper pump than you expected.

I ship into the Philippines and Indonesia regularly, and the failure pattern is the opposite of Africa’s. In Africa the water is deep and the pump has to work. Here the water is easy and the buyer still ends up short of water, because a low-head pump has almost no pressure to spare and every metre you add comes straight out of the flow.

Why Southeast Asia is a different problem from Africa

Buyers often assume the two markets are the same because both are hot and both are tropical. They are not.

Read the fourth row carefully. Archipelagic geography is why solar pumping makes more sense here than almost anywhere: running a distribution line to a small island or a remote barangay costs more than the pump will ever save, and diesel has to survive a boat trip. About 20 percent of Filipinos still live without reliable grid access, mostly on the smaller islands and in the uplands.

The month that catches people out is the wet season, not the dry one

This is the exact opposite of the African problem I wrote about in the Africa guide, and it is why you cannot reuse a template.

Here is real monthly data from the NASA POWER climatology, mean daily solar radiation on a horizontal surface in kWh/m2/day:

Mean daily solar radiation on a horizontal surface, kWh/m2/day, NASA POWER climatology.

Two things jump out, and neither is obvious from a marketing brochure.

First, the annual mean is not the danger. Manila averages 5.10 and never drops below 4.15. That is a small swing compared with Nairobi falling 41 percent, and it means the yearly average is a much safer starting point here than it is in East Africa.

Second, the worst month is not the month you would guess. Look at Manila: the lowest point is December at 4.15, with August close behind at 4.35. In Nueva Ecija it is August at 4.27, the heart of the southwest monsoon and the peak of the typhoon season. Cebu bottoms out in December at 4.06. Jakarta’s floor is February at 4.39.

Now put that next to the crop calendar. Irrigated rice in Asia needs 1,300 to 1,500 mm of water over a season, and per IRRI, land preparation alone - soaking and puddling - can consume up to a third of the total water requirement. In much of the Philippines the wet-season rice crop is transplanted in June and July and comes to heading in August and September. So the sun trough and the water peak do not sit comfortably apart the way they do in a temperate climate with a winter and a summer. They can land in the same month.

That is the design case for a Southeast Asian rice farm, and almost nobody sizes for it. Size the pump for a dry spell in the week the crop heads, at Nueva Ecija’s 4.27, not for the 5.08 annual average.

The derate is real here too, but for a different reason

Panels lose output as they heat up, roughly 0.35 percent of voltage per degree above 25 C, and a panel in full sun in the tropics runs at 60-65 C. Add humidity, dust from unpaved roads, and controller losses, and the practical performance ratio is 0.75 to 0.80. I use 0.78 here as well.

Effective full-flow pumping hours roughly equals GHI in kWh/m2/day x 0.78.

For Nueva Ecija in August that gives 4.27 x 0.78 = 3.3 hours. In April it gives 6.39 x 0.78 = 5.0 hours. Same array, same pump, a 50 percent difference in output between the two months - and the better month is the one you would probably design for if you sized on the annual average.

The 10 metres you add yourself

This is the single most useful thing on this page, so read it twice.

Shallow water and low head is the cheapest place in the world to pump. It is also where buyers make the most expensive mistake, because a low-head pump is designed for flow, not for pressure. It has almost no head to spare, so every metre you add comes straight out of your flow.

Most people assume the relationship is gentle. It is not. Here is the published measured curve for one of our real surface pumps, the DLP15-14-48-500, a 500 W 48 V unit with a rated 15 m3/h flow and a 14 m maximum head. This curve belongs to that single model, so every point below is attributable to it:

Look at the last two rows. At 13 m this pump delivers nothing at all, because its maximum head is 14 m. The rated “15 m3/h” is real, but it exists only at almost zero head.

Now look at the shape of the curve between 4 m and 10 m. It loses more than half its flow in 6 metres. Six metres is not a big distance. It is a small bank, a slightly raised pump platform, a longer run of undersized hose, or a nozzle you fitted to get more pressure.

Where those metres come from, and how to get them back:

• Mounting the pump above the water instead of at it. Every metre of suction lift costs you flow. Put the pump on the bank at water level, not on top of the bund.
• Undersized delivery hose. Long runs of small hose add friction head fast. Use 2 inch or 3 inch on a surface pump, and keep the run as short as the field allows. Our standard planning figure is about 10 m of pipe per 1 m of head, so 30 m of hose costs you 3 m.
• Pushing water up to a tank instead of into the field. Every metre of tank height is a metre of head. If you want to fill a tank, fill it from the field channel, or accept that the tank is a different job.
• A sprinkler or a nozzle. Sprinklers need pressure, and pressure is exactly what this class of pump cannot give. If you need spray pressure, move up to a higher-head model and accept lower flow.

The rule I give every buyer in the Philippines: keep total dynamic head under 8 m for a surface pump, and buy the biggest hose you can afford. The pump is cheap; the flow is what you are actually buying. The full method for adding it all up is in how to calculate total dynamic head, and if you want to see what a curve is telling you, read how to read a pump curve.

What a rice paddy actually needs

Rice is a semi-aquatic plant and it is genuinely water-hungry, so the volumes here are much larger than an equivalent area of vegetables. The useful figures come from IRRI, the International Rice Research Institute:

• Total seasonal water input for irrigated rice in Asia is typically 1,300 to 1,500 mm, ranging from as little as 400 mm on heavy clay with a shallow water table to more than 2,000 mm on coarse sandy soils.
• Land preparation can consume up to a third of the total. Soaking and puddling is front-loaded and happens in a short window, which makes it the peak demand period, not the average one.
• After transplanting, ponded depth should start around 3 cm, rise to 5-10 cm as the plants grow, and stay there until the field is drained 7-10 days before harvest.
• From heading to the end of flowering, water must be kept at about 5 cm at all times. This stage is very sensitive to shortage, and shortage here costs yield through spikelet sterility.
• An unleveled field needs an extra 80-100 mm of water just to give even coverage, which is roughly 10 percent of the whole seasonal requirement. Levelling is a pumping decision, not just an agronomy one.

Two conclusions follow, and they are the reason rice and solar pumping suit each other so well.

First, you need volume, not pressure. A 5-10 cm pond is a very low head. That is exactly the duty where a solar surface pump is efficient and cheap.

Second, your critical day is a dry spell during heading, not the average day of the season. That is the day you must size for, and it lands in August in Luzon.

If you also run a drip system on a vegetable block beside the paddy, note that the two duties pull in opposite directions: drip wants pressure and filtration, paddy wants volume at almost no head. One pump rarely does both well.

Worked example: 0.5 ha of paddy, Nueva Ecija

A real brief, worked the way I would actually quote it.

The site. 0.5 ha (5,000 m2) of wet-season rice in Nueva Ecija. A shallow well with the water surface 1.5 m below the pump platform, and the field channel 1.5 m above it. 30 m of 2 inch delivery hose.

Head. 1.5 (lift) + 1.5 (field channel) + 3 (30 m of hose, at our standard 10 m per 1 m) = 6 m total dynamic head.

Worst case. A dry spell during heading in August: 4.27 x 0.78 = 3.3 effective hours.

Volumes. At 8 mm/day net for the paddy, 0.5 ha needs 40 m3/day. Take the sun trough month and the peak demand together and I size the pump for 40 m3/day.

Duty point. 40 / 3.3 = 12.1 m3/h at 6 m.

Pump. A 15 m3/h rated surface pump looks right, but its actual flow at 6 m is nearer 10 m3/h, which gives only 33 m3/day. So step up one model. The DCPM21-14-72-750 is rated 21 m3/h at 14 m maximum head on 750 W at 72 V, and at a 6 m duty point it has real margin left. That is the honest answer: the head figure, not the flow figure, decides which pump you can use.

Panels. 750 W x 1.3 = 975 W minimum, so two 550 W panels, 1,100 W. Match the string to the controller window - the method is in how many solar panels for a water pump.

Storage. 40 m3/day x 1.5 = a 60 m3 reservoir. On rice you can often use a field reservoir or a canal rather than a steel tank, which is far cheaper.

What happens if you get the head wrong. Suppose you mount the same pump on top of the bund, 4 m higher, and run 60 m of undersized hose instead of 30 m of 2 inch. Your head goes from 6 m to about 14 m. Flow does not fall by a third; on a curve this shape it collapses toward the shut-off point and can drop by more than half. The farmer concludes that solar is unreliable. The pump was fine. Four metres of elevation and 30 metres of hose were the entire problem.

If you want to run your own numbers rather than mine, put your water level, lift and hose length into the sizing tool and it will work out the head and the model for you.

Salt, rust and typhoons: the three things that end systems early

The electrical design in Southeast Asia is easy. The environment is not.

Salt. The Philippines has 36,000 km of coastline and the archipelago sits in an atmosphere that corrodes metal aggressively. Under ISO 9223 and ISO 12944-2, the corrosivity categories run from C1 up to C5-M for coastal and offshore areas with high salinity, and C5-M means a steel thickness loss of 80 to 200 microns in the first year alone. Then read the warning the standard itself carries: in coastal areas in hot, humid zones, the losses can exceed even C5-M. That is precisely where you are building.

The practical response is not to buy a “coastal” version of the same thing. It is to change the materials:

• Array frame: anodised aluminium rather than painted steel. Where you must use steel, use hot-dip galvanised and accept that the zinc is a consumable that will need replacing on a coast.
• Fasteners: 316 stainless, not 304 and never plain zinc-plated. A rusted bolt is how a panel leaves the frame in the first storm. If you handle water in a marine context you already know this trade-off; it is the same one covered in plastic vs stainless steel impeller.
• Cable and terminations: UV-rated cable, and glands or junction boxes that are actually IP65 rather than sold as such. The single most common field failure I see on coastal sites is water in the controller.
• Impeller: if the well is brackish - and many shallow coastal wells in Indonesia and the Philippine islands are - go stainless. Plastic is fine on clean fresh water and a poor choice on anything salty. Brackish water also changes your irrigation method, so test the water before you buy, not after.

Typhoons. About 20 tropical cyclones enter the Philippine area of responsibility each year and eight to nine make landfall. This is a routine design input, not a rare event. The pump is usually safe, deep in the well and disconnected. The array is the vulnerable part, because panels are large, light and sail-like.

Design for removal, not for resistance:

• Choose a mounting tilt you can drop nearly flat in an hour.
• Use quick-release bolts so the panels can be stacked and stored before landfall. Panels that can be taken down will be taken down; panels bolted permanently to a high roof will be lost.
• Never mount the array on a high roof. Wind speed rises sharply with height, and a roof array also puts the most fragile part of the system in the least serviceable place.
• Prefer ground mount at 10-15 degrees. Near the equator the yield penalty for a shallow tilt is small, rain still self-cleans the glass, and the frame is easy to reach.

Dust and dirt. Not desert dust here, but unpaved road dust and volcanic ash in the Visayas and Luzon. A heavily soiled array loses 10 to 20 percent of output, and to a farmer a dirty array looks exactly like a broken pump. Plan a monthly rinse and put it in the maintenance routine.

Algae, sand and slow water: protecting a shallow-water pump

Shallow, warm, sunlit water is a different problem from deep borehole water, and it attacks a different part of the system.

• Algae and weed. Surface water grows things. A suction strainer or an intake filter is mandatory, and on a drip or sprinkler system the filter is not an option, it is a component. The design logic is in designing a solar drip irrigation system.
• Sand and silt. River and canal water carries abrasive particles, and a plastic impeller pumping grit will not last. The mechanics of this are in solar pumps in sandy wells. If your source is a river or a sand-lined well, stainless impellers are the specification, not an upgrade.
• Air locks. A surface pump that loses its prime will not recover on its own when the sun comes back. Fit a foot valve, keep the suction line airtight and short, and keep the pump as close to the water level as the design allows.
• Low, steady head beats high head every time. If you can arrange the site so the water has the shortest possible distance to travel, you can buy a smaller pump, a smaller array and a smaller controller. On this site that is the whole game.

One number worth keeping in mind: on the 48 V submersible line, the maximum flow anywhere in the range is 6 m3/h, across the 3DSC, 4DSC and 3DPC series. If your duty point needs more than 6 m3/h at low head, do not try to force a 48 V submersible to do it. Either move up to 72 V or 110 V, or use a dedicated surface pump built for flow, which is what the DLP, DCPM and HJSN families are for.

A real model ladder for shallow water and islands

These are actual models from our published specification tables, ordered by the job they suit. Every model and every figure below can be found in our catalogue.

The progression is consistent across the range, and the letters are the key: 2DPC, 3DPC and 4DPC for small plastic-impeller jobs on clean water, 3DSC and 4DSC when you need stainless or more head, DLP, DCPM and HJSN when the job is surface water and volume, and 6DSC and 8DSC when the customer is a scheme rather than a farm. The high-head end of the range, the 3DSS and the 100 m-plus models, is largely wasted in this region - the shallow water here does not need it, and you would be paying for head you never use. That world is covered in the deep well guide.

If your interest is specifically community supply on an island, the sizing logic shifts toward peak day demand and storage rather than a single field, which is covered in solar pump for village water supply. The electrical side, including the hot-climate MPPT behaviour where heat pulls the string voltage down toward the bottom of the controller window, is in the controller wiring diagram.

Five things that decide whether the system survives

In order of how often they actually kill a project here:

1. Head you added yourself. A bank, a raised platform, an extra hose run, a nozzle. This is the number one cause of a solar pump that “does not work” in Southeast Asia, and it is entirely within your control. Measure your head honestly before you buy.
2. Sizing on the good month. In the Philippines size for August or December, not for the 5.1 annual average, and remember that the water peak and the sun trough can coincide on a rice crop.
3. Corrosion and storm design. Anodised frame, 316 fasteners, an array you can take down, and never on a high roof. If your site is within a few hundred metres of the sea, assume C5-M and buy accordingly.
4. Filtration on surface water. Algae, sand and silt will destroy an unprotected pump faster than any electrical fault, and they are almost free to filter out. Size the strainer and the filter as part of the system, not as an afterthought.
5. Storage sized for the dry spell. 1.5 to 3 times the daily requirement. On a 7,641-island country where spare parts arrive by boat, a reservoir is also your buffer against a two-week repair wait.

Everything else - head, hose, panel count, voltage window, tank volume - is arithmetic, and the arithmetic is in this article.

Planning a system in the Philippines, Indonesia, Vietnam or Thailand? Put your water level, your lift and your hose length into the sizing tool to see your real head, or message me on WhatsApp with your water level, lift, hose length and crop, and I will come back with a model, a panel count and a tank size sized for your worst month.