LELABAPP
Case study

A spring's water for a village
on the highlands of Madagascar

In Mahalavolona, on the Malagasy highlands, a captured spring flows below the village. The distribution reservoir sits 18 metres higher and 500 metres further away. The study covers a fully solar supply, with no grid and no generator.

Highlands, Madagascar Captured spring: 1.2 to 1.9 m³/h Need: 32 m³/day, year-round Elevation gain 18 m · pipe 500 m No grid, no generator
The project

The need, as the municipality expresses it

The village has grown and the existing gravity supply no longer keeps up. The good news is already in place: the captured spring fills a 32 m³ buffer tank by gravity at the foot of the village. What remains is the lift: carrying 32 m³ a day up to the distribution reservoir that overlooks the houses, so the water can then flow back down on its own to the standpipes.

32 m³/daythe village's water need, year-round: everyday water, not seasonal irrigation
1.2 to 1.9 m³/hthe spring's yield depending on the season, that is 29 to 46 m³ per day, collected in the 32 m³ buffer tank
18 mthe elevation gain between the buffer tank and the village reservoir, 500 m of pipe away
20 m³the elevated distribution reservoir, which supplies the village by gravity

Site specifics: no borehole. The pump is immersed directly in the buffer tank, 2 metres below the surface. And no reliable electricity nearby: fully solar pumping, without batteries, is the natural choice.

Step 1

The input data in LE LAB

A few minutes of input. The solar reference city is Antananarivo, around a hundred kilometres away, on the same highlands and at the same altitude as the site. The depth entered is only 2 m, since the pump is immersed in the tank; the 21 m delivery head covers the elevation gain and the reservoir height. The total head is calculated automatically: 38 m, of which 41% comes from the pipe's head losses. Over 500 metres, the pipe weighs almost as much as the lift: that is exactly what a sizing must account for.

Study summary in LE LAB: Antananarivo, 32 m³ per day year-round, depth 2 m, delivery head 21 m, 500 m of pipe, calculated TDH 38 m, no electrical backup, water reserve 20 m³

The summary before calculation: nine answers are enough, the total dynamic head is calculated for you.

Step 2

What LE LAB proposes

Four panels. To lift the water for a whole village, the chosen solution is a LORENTZ PS2-1800 with the HRE-32 end, a helical rotor, powered by 4 panels of 515 Wp, for 2,060 Wp in a single string. Average production: 44.0 m³ per day, from 38.1 m³ in June to 48.5 m³ in October. LE LAB recommends three panels at the tightest; the study keeps four, through the pro solar-array adjustment, to lift June and July above the need. The array stays this small because the head stays moderate and the highlands' sunshine is generous and steady. This is the frugality typical of a well-sized fully solar pumping system: no batteries, very few panels, no fuel.

Solution proposed by LE LAB: LORENTZ PS2-1800 HRE-32 pump, 2060 Wp solar array with 4 panels of 515 Wp, average production 44.0 m³ per day, with the note that the solar array was adjusted manually, recommendation 3 panels

The solution at a glance: pump, solar array and average production.

Why a LORENTZ? LE LAB also offers an alternative from another brand, but it requires four panels where the recommended LORENTZ needs three, for a slightly lower average output. And beyond the figures, there is the field reality: LORENTZ is well established in Madagascar, with installers who know these pumps and parts that can be found. For a village system that must last for years far from everything, the local maintenance network matters as much as the datasheet.
Step 3

Panels laid almost flat, without losing anything

At this latitude, the calculated optimal angle is 21°, with panels facing due north since we are in the southern hemisphere. But a low, gently tilted frame is simpler to build locally and offers less grip to the wind, a real concern in a country exposed to cyclones. LE LAB's angle setting quantifies the trade-off.

Panel angle set to 11 degrees in LE LAB: annual production stays at 100% of the 21-degree optimum

At 11°, annual production stays at 100% of the optimum: the simple frame costs nothing, and this pitch is enough to let rain rinse the panels.

Step 4

Production, month by month and hour by hour

Monthly production is read against the need, over sixteen years of the region's PVGIS solar history. The low point is in June, the austral winter: 38.1 m³ per day, above the 32 m³ required; July follows at 39.3 m³. The other months range between 40.4 and 48.5 m³. The rainfall curve shown tells the other half of the climate: from 475 mm in January to 46 mm in September, the long dry season is exactly when the spring's water becomes precious.

Estimated monthly production against the need of 32 m³ per day, critical month June at 38.1 m³ per day, rainfall curve from 475 mm in January to 46 mm in September

With the fourth panel, every month stays above the need line, June and July included. The critical month is shown, not hidden.

Daily output for a typical day in August: hour-by-hour flow rate, peak at 5.5 m³/h, 43.0 m³ produced, need of 32 m³ covered with 34% margin

The typical August day, hour by hour: the pump climbs to 5.5 m³/h during the hours of full sun and produces 43.0 m³, a 34% margin on the need.

Two rhythms, one tank. The spring flows around the clock at under 2 m³/h; the solar pump draws up to 5.5 m³/h during the hours of sun. It is the 32 m³ buffer tank that reconciles the two: it fills slowly day and night, and the pump drains it as the sun allows. Each works at its own pace, and the night recharges what the day has lifted. The validation is therefore twofold: the pumping energy on one side, the spring's low-water flow on the other — the latter is measured on site.
Step 5

The 20 m³ reservoir put to the test

LE LAB simulates the water reserve day by day over sixteen years of PVGIS weather history. With the 20 m³ reservoir alone, 355 of 365 days are fully covered. LE LAB still flags insufficient coverage, because incomplete days remain: 10 in an average year, which still deliver 73% of the need. The volume missing over the year comes to 83 m³ out of 11,688, or 0.7%. The worst episode of the sixteen simulated years: from 2 to 11 January 2009, 10 days of low sun.

Autonomy and water reserve block: 355 of 365 days covered with the 20 m³ reservoir, insufficient coverage warning, 10 incomplete days delivering 73% of the need, missing volume 83 m³ out of 11,688 per year, benchmarks 61 m³ and 91 m³

The benchmarks quantify the next step if the municipality aims for zero shortfall: 61 m³ would bring incomplete days down to no more than one per year, 91 m³ would cover every single day of the sixteen years.

The margin the calculation does not count. LE LAB simulates the 20 m³ reservoir alone here. On the ground, the 32 m³ buffer tank keeps filling all night and acts as an extra cushion: real robustness is better than what is shown. The 10 incomplete days weigh 83 m³ over the year, about 8 m³ on average per affected day: a volume that tank absorbs in practice. And for a village's water, an "incomplete" day at 73% is managed by rationing distribution for a few hours, not by cutting the water off.
Step 6

Manufacturer curves, at the real operating point

The pump's Q/H curves, plotted from official LORENTZ data, with the project's head highlighted: 5.5 m³/h on the 38 m curve, at the pump's maximum. The HRE-32 accepts up to 5.8 m³/h and 50 m: the project's head stays well below that ceiling, while the flow sits at the top of the range.

Manufacturer curves for the PS2-1800 HRE-32 pump: flow according to available power by head, operating point 5.5 m³/h at 38 m

Flow according to available power, by head. The yellow curve is the project's, at 38 m.

What this case shows

The tool frames, the trade decides

4 panelsthe frugality of a well-sized fully solar pumping system: no batteries, no fuel, very little equipment
2 m³/hthe real limit of the project is the spring, not the sun: in a severe low-water period it barely gives 29 m³ a day for 32 required, and measuring the flow on site remains a prerequisite for any offer
16 yearsproduction and the reservoir are tested day by day over sixteen years of the highlands' PVGIS solar history
Honestythe 10 incomplete days are shown, the uncounted margin from the buffer tank is stated, and the pump choice openly assumes the local-network argument

Your project deserves the same study

Replay this study with your own figures: location, need, spring or borehole. Free, no registration, results in a few minutes.

Size my project See the other case studies