Storage Instead of Batteries: Sizing Tanks and Autonomy for Solar Pumping

Every solar pumping scheme faces the same awkward fact: the sun shines during the day and people want water at dawn and after dark. Something has to bridge the gap, and the choice of what does the bridging shapes the cost and reliability of the whole system for the next twenty years.

The instinctive answer is batteries. It is almost always the wrong one. A tank of water does the same job with no moving parts, no chemistry, no replacement cycle and no maintenance beyond an occasional clean, and in most schemes it does it for less money.

This article covers how to size that storage properly: days of autonomy, the difference between daily volume and peak demand, the elevation trade-off that quietly increases your array cost, and the water quality considerations that stop a well-designed tank from becoming a health problem. It follows on from our article on designing a solar water pumping system.

Why Storage Beats Batteries

The comparison is not close once you look at the whole life of the system.

A battery bank in a solar pumping application typically needs replacing several times over the life of the array and the pump. Each replacement is a significant capital event, often in a remote location, and the batteries themselves need protection from heat, which in the climates where most of this equipment operates is a real and continuing cost. Deep-cycle performance degrades with age and with every cycle, so the system's autonomy shrinks year by year even when everything is working.

A water tank has none of these properties. It stores the day's yield, it delivers it whenever anybody opens a tap, and its capacity in year twenty is the same as in year one provided it has not been physically damaged. It has no charge controller, no state of health, no thermal management and no disposal problem.

There is a second and less obvious advantage. Pumping directly to storage lets the pump run at whatever rate the available sunlight supports, which is exactly what a variable-speed solar controller is designed to do. Adding batteries to run the pump at constant speed throws away that flexibility and adds conversion losses on both sides of the battery.

Batteries earn their place in a narrow set of cases: where the pumped water genuinely cannot be stored, where the same array serves other loads that need overnight power, or where elevation for a gravity tank is impossible and pressurised delivery must be available at night. Outside those cases, the FAO guidance on solar-powered irrigation reflects general field practice: store water, not electricity.

Days of Autonomy

Autonomy is the number of days the system can meet demand with no pumping at all, and it is the primary sizing input.

One day of autonomy covers the overnight gap and nothing more. It is appropriate only where an interruption is genuinely tolerable and where consecutive poor-weather days are rare.

Two days is the common default for community water supply and general agricultural use. It covers a full cloudy day plus the following night, which handles the majority of weather events in most of the regions where solar pumping is deployed.

Three days or more applies where continuity is critical: health facilities, schools, livestock watering during dry periods, and any site where a failure would require emergency water trucking. It also applies where the weather genuinely produces multi-day overcast periods.

The temptation is to size for the worst case ever recorded, and it should be resisted, because storage volume is not free. Tank cost rises with capacity, the structure to support an elevated tank rises faster, and the footprint has to exist. Beyond about three days, the economics usually favour a modest increase in array size over a large increase in storage, because a bigger array recovers the deficit faster once the sun returns.

The honest way to choose is to look at the actual local weather record for consecutive low-irradiance days rather than at a rule of thumb, and to weigh the consequence of running dry. A clinic and an irrigated field have very different answers.

Days of autonomy by application. Beyond about three days, a modest increase in array size usually beats a large increase in storage. Sources: WHO Guidelines for drinking-water quality, 2017; MIMAH engineering practice.

Daily Volume Is Not the Whole Requirement

Sizing purely on days of demand misses a second constraint that catches out a surprising number of schemes: the rate at which people draw water.

Consider a village supply sized at two days of a 50 cubic metre daily demand, so 100 cubic metres of storage. That looks generous. But if the community draws most of its water in a two-hour window in the early morning, the tank has to deliver that peak flow through its outlet and distribution network, and the pump will not be replenishing it because the sun is barely up.

Two things follow. First, the outlet and distribution pipework must be sized for peak instantaneous demand rather than for average demand, or the tank will have water in it that people cannot get out fast enough. Queues at a standpipe are frequently a pipe sizing problem rather than a storage problem.

Second, where several drawing points share one tank, the simultaneity of demand matters. Design guidance for community water systems typically applies a peak factor to average demand, and the factor depends on population size, with smaller populations showing sharper peaks because there is less statistical smoothing.

For irrigation the equivalent question is whether the scheme irrigates by gravity directly from the tank over a defined watering period, which sets a minimum flow rate the tank and its outlet must sustain.

The Elevation Trade-off

Elevating a tank gives gravity pressure at the point of use, which removes the need for a booster pump and gives water on demand at night without any electricity at all. It is a genuinely elegant solution and it has a cost that is easy to underestimate.

Every metre of elevation adds a metre to the total dynamic head the pump must overcome, and head drives array size directly. Raising a tank from 6 m to 15 m adds 9 m of head to every cubic metre pumped for the life of the system, which on a scheme with 60 m of existing head is a 15 per cent increase in energy and therefore roughly a 15 per cent larger array.

The structure is the other half of the cost. Water weighs approximately one tonne per cubic metre, so a 50 cubic metre elevated tank imposes a 50 tonne load plus the tank's own weight, at height, with wind loading, on a foundation that must be designed for it. Elevated steel towers at that scale are a significant civil engineering item and they are not cheap.

The alternatives are worth pricing rather than assuming. A ground-level tank with a small booster pump serving the distribution network costs far less in structure and in array size, at the price of needing power to deliver. A ground tank sited on naturally higher ground gives free elevation where the topography cooperates, and it is always worth walking the site to look for it. A hybrid arrangement, with a large ground-level tank and a small elevated header tank refilled by a booster, gives gravity delivery with far less elevated mass.

The circularity to watch, and it catches out first-time designers, is that raising the tank to improve distribution pressure increases the head, which increases the array, which increases the cost of the thing you raised the tank to avoid.

Tank Materials

Polyethylene tanks are inexpensive, light, easy to transport and install, and widely available. They should be UV-stabilised and, for potable use, of a food-grade specification. Their weakness is degradation under sustained high UV and temperature, which is precisely the environment they are usually installed in, so service life in hot climates is shorter than the catalogue suggests. Shading or painting a plastic tank extends its life and reduces algae growth.

Glass-reinforced plastic sectional tanks assemble on site from panels, which makes them practical for large volumes in locations where a one-piece tank cannot be transported. They cost more than polyethylene and last longer.

Steel, either welded or bolted sectional with an appropriate internal lining, suits large capacities and elevated installations. It requires corrosion protection and periodic maintenance of that protection, and the lining specification matters for potable use.

Concrete and ferrocement tanks are durable, resistant to heat, and can be built with local labour and materials, which matters in remote locations. They are slow to construct, need proper curing, and can crack if the foundation moves. Where they are well built they last a very long time.

The right choice depends on capacity, transport access, local skills, potable requirement and budget, and the pattern in the field is that the cheapest tank installed badly costs more than the appropriate tank installed properly.

The elevation trade-off, worked. Raising the tank to improve distribution pressure increases the head, which increases the array, which increases the cost of the thing you raised the tank to avoid. Source: MIMAH worked example.

Water Quality in Storage

A tank is a place where water sits still and warm, which is a set of conditions that favours things nobody wants in a drinking supply.

Residence time is the core issue. Water that sits for days loses any disinfectant residual and becomes hospitable to microbial growth. An oversized tank on a low-demand system is not conservative design; it is a water quality problem. This is a genuine argument against sizing autonomy far beyond what the risk justifies.

Light drives algae growth. A translucent tank in direct sun will grow algae, and the fix is opacity: an opaque tank, a painted one, or shade.

Contamination protection means a properly fitting lid, a screened vent to allow the tank to breathe without admitting insects, a screened overflow, and an inlet arrangement that does not allow backflow. Every opening is an entry point and each one needs deliberate design rather than whatever the installer had available.

Access for cleaning should be designed in. Tanks accumulate sediment and biofilm and need periodic cleaning, and a tank with no practical access will not be cleaned.

Draw-off arrangement should take water from above the sediment layer, with a separate low-level drain for cleaning.

Where the supply is for drinking, the WHO Guidelines for drinking-water quality set out the health-based framework, and the relevant national standard applies alongside. The engineering point is that storage is part of the treatment train whether or not anybody designed it that way: it can preserve water quality or degrade it, depending on choices that cost very little at the design stage.

Controls and Overflow

The pump must stop when the tank is full, and how that is arranged matters more than it appears.

A float switch in the tank signalling the controller is the common approach and it works well provided the switch is accessible for maintenance and the cable run is protected. It should be wired to fail safe, so that a broken cable stops the pump rather than running it continuously.

Hysteresis between the stop and restart levels prevents the pump cycling rapidly around the full point, which is hard on the motor and on the controller. A generous dead band costs nothing.

An overflow is mandatory regardless of the controls, sized to pass the full pump output, discharging somewhere that will not erode the foundation or create standing water. It is the backstop for a failed level control and it will eventually be needed.

A level indicator visible from the ground is a small addition that transforms operation, because it lets the operator see the system working and notice when it stops.

Storage preserves water quality or degrades it, depending on choices that cost very little at design stage. An oversized tank on a low-demand system is a water quality problem. Source: WHO Guidelines for drinking-water quality, 2017.

Common Sizing Errors

Sizing for autonomy and ignoring peak draw rate. Produces a full tank that cannot deliver fast enough at the morning peak.

Oversizing storage on a low-demand system. Long residence time degrades water quality and wastes capital that would have bought array capacity.

Elevating without pricing the head. Every metre of tower is a permanent addition to pumping energy for the life of the scheme.

Ignoring the structural load. Water is a tonne per cubic metre and elevated tanks impose serious loads. This needs designing, not estimating.

Translucent tanks in full sun. Algae, and shortened tank life.

No practical cleaning access. The tank will not be cleaned, and eventually that shows up in the water.

Overflow sized smaller than pump output. The one time it is needed, it will not cope.

Frequently Asked Questions

How many days of storage should a solar pumping system have? Two days is the common default for community and agricultural supply. Health facilities and other critical uses justify three or more. Beyond three days, increasing array size is usually better value than increasing storage, and long residence time starts to affect water quality.

Is a tank really better than batteries? For almost all pumping applications, yes. A tank has no replacement cycle, no degradation, no thermal management requirement and no disposal problem. Batteries are justified mainly where water cannot be stored or where the array serves other overnight loads.

How high does an elevated tank need to be? High enough to deliver the required pressure at the furthest and highest draw-off point after friction losses in the distribution network. Work backwards from that requirement rather than picking a height, because every unnecessary metre is permanent additional pumping energy.

What size should the tank outlet be? Sized for peak instantaneous demand, not average daily demand. Community supplies draw most of their water in short morning and evening peaks, and an outlet sized on daily average will not pass it.

Storage Is Where Solar Pumping Gets Cheap

The reason solar pumping works so well as a technology is that the storage problem, which is the expensive part of almost every other off-grid application, has a nearly free solution here. Water stores itself. There is no other renewable energy application where the buffer is this cheap, this durable and this maintenance-free.

Getting the sizing right is mostly a matter of asking two questions instead of one: how much water over how many days, and how fast does it have to come out. Schemes that ask only the first end up with full tanks and queues. Across water and solar infrastructure work in Sudan and Nigeria, the storage decision has repeatedly proved to be the one that determines whether a scheme feels reliable to the people using it, regardless of how well the pumping side performs.

Planning storage for a pumping scheme? Talk to our engineering team. We will size it against your actual demand pattern and weather record, and price the elevation options properly so you can see what the tower is really costing you.