Clean drinking water, made by the sun and smart engineering.
AquaPure Solar builds AI-enhanced solar stills: simple distillation units that turn salty or contaminated water into clean drinking water using only sunlight — with machine learning to get the most water out of every sunny day.
Billions of people still can't count on safe water.
For remote communities far from pipes and power lines, clean drinking water is often unreliable, expensive or unsafe.
people lacked safely managed drinking water in 2022 — about 27% of the world.
people use a contaminated drinking-water source.
diarrhoeal deaths each year are caused by contaminated drinking water.
people live in water-stressed areas.
Source: World Health Organization, Drinking-water fact sheet
Nature's water cycle, in a box.
A solar still copies the way rain is made: the sun evaporates water, leaving salt and impurities behind, and the pure vapour condenses into clean droplets.
- 1
Sunlight heats
Sunlight passes through a transparent cover and warms salty or dirty water in a dark basin.
- 2
Water evaporates
Only pure water vapour rises. Salts, germs and other impurities stay behind in the basin.
- 3
Vapour condenses
The vapour meets the cooler, sloped glass cover and turns back into clean water droplets.
- 4
Clean water collects
Droplets run down the glass into a channel and flow out as distilled drinking water.
Solar stills are simple, low-cost and environmentally friendly desalination devices — ideal for places without electricity or fuel supply.
A solar still's output changes with every cloud.
How much water a still produces depends on many interacting factors. That makes it hard to predict — and easy to run below its best.
What machine learning makes possible
Predictive models, sharpened by smart optimizers.
AquaPure Solar draws on the machine learning approaches reviewed for solar still performance: predictive models to forecast output, and hybrid models that pair them with metaheuristic optimization.
Predictive models
Learn from data how design, climate and operation translate into water production.
SVMSupport Vector MachinesMLPMulti-Layer PerceptronsANFISAdaptive Neuro-Fuzzy Inference SystemsDTDecision TreesMetaheuristic optimizers
Search for the best model settings — and the best operating conditions — inspired by nature.
GAGenetic AlgorithmsPSOParticle Swarm OptimizationSASimulated Annealing+Combined into hybrid modelsAn engineered still, built to work where infrastructure doesn't.
AquaPure Solar manufactures a desalination unit that combines a proven solar-still design with an AI layer for prediction and control.
100% solar-powered
Distillation runs on sunlight alone — no diesel, no grid connection.
Purifies by distillation
Evaporation leaves salt and impurities behind, producing distilled drinking water.
AI-enhanced performance
Machine learning forecasts output and guides operation for maximum yield.
Simple to maintain
Few moving parts and a simple design suit communities with limited technical support.
Clean water for places the network doesn't reach.
Remote villages
Communities far from water pipelines and treatment plants.
Off-grid schools & clinics
Safe drinking water where there is no reliable electricity.
Coastal & brackish areas
Places where the available water is too salty to drink.
Humanitarian settings
A fuel-free water source for camps and emergency response.
Engineers turning sunlight into safe water.
Faculty of Engineering
AquaPure Solar was founded in 2025 by a research team from the Faculty of Engineering, combining thermal and solar-energy design with machine learning to manufacture better desalination units.
Frequently asked questions
What is a solar still?
A simple device that uses sunlight to evaporate salty or contaminated water and condense the vapour into clean, distilled water — the same principle as the natural water cycle.
Does it need electricity or fuel?
No. Distillation is powered entirely by the sun, which makes solar stills well suited to remote and off-grid locations.
What does the AI do?
Machine learning predicts water production, identifies optimal operating conditions and supports real-time monitoring, control and energy management.
Which machine learning methods are used?
Predictive models such as Support Vector Machines, Multi-Layer Perceptrons, ANFIS and Decision Trees, and hybrid models that combine them with Genetic Algorithms, Particle Swarm Optimization or Simulated Annealing.
Why not just use a standard solar still?
A still's output depends heavily on climate, design and operating conditions. Data-driven prediction and optimization help each unit perform closer to its best, more reliably.
Built on a comprehensive review in Solar Energy.
Read the open-access review of predictive and hybrid machine learning for optimizing solar still performance.
Elsheikh, A., Faqeha, H., Hammoodi, K. A., Bawahab, M., Fujii, M., Shanmugan, S., Essa, F. A., Abd-Elaziem, W., Ramesh, B., Sathyamurthy, R., & Egiza, M. (2025). Integrating predictive and hybrid machine learning approaches for optimizing solar still performance: A comprehensive review. Solar Energy, 295, 113536.
https://doi.org/10.1016/j.solener.2025.113536