Full Flow — Testing times in water distribution monitoring

Ben Britton
Frontier Tech Hub
Published in
7 min readOct 27, 2020

Our story starts on a balmy autumn evening in the commercial outskirts of Amman, Jordan. Aziz Wadi, Field Ready’s technical lead for Jordan and Syria, is meeting a water truck driver to fit a prototype water monitoring device to his truck. The driver is late. When he does arrive he brings trouble. He’s in some kind of extended altercation with another driver. Before long the situation deteriorates into fisticuffs and bloody noses. Wadi quietly leaves, thankfully unharmed, and phones the next driver on his list.

Water tanker being fitted with IOT monitoring device. Photo Credit: Aziz Wadi/Field Ready

Field testing isn’t usually this high-adrenaline, but this is where the rubber meets the road. The clue is in the name for Field Ready. We work with designers, inventors and manufacturers to take designs from concept through to a ‘field ready’ humanitarian product. To get there we have to go through many processes — market research, ideation, design, prototyping, testing, user feedback, iteration, pre-production and batch or serial manufacturing. That’s a whole lot of work, so we have to be sure it’s going to be worth it. Fortunately, in Jordan, we have the Norwegian Refugee Council on side to guide the development of this potentially sector-changing product — an IOT enabled water monitoring device.

Humanitarian aid agencies around the world struggle to use real-world data for the basis of payments for water tanking companies, meaning that the water-supply sector is rife with misreporting, fraud and wastage. The need for a solution and its potential value is indicated by the recent 1,000,000 Euro prize won by UNHCR for a static water-monitoring system that tracks reservoir levels in water-distribution projects. Our system is mounted to the delivery trucks and logs water-delivery data through ingenious automated sensors; it then automatically uploads that information to a cloud-based database at the end of each day. A fully developed water-monitoring device like this would bring a level of accuracy and granularity to the monitoring data that water distribution partners can now only dream of.

This sprint, supported by UKAID’s Frontier Technologies Livestreaming programme, enabled us to test a prototype device for measuring volume of water delivered by tanker. We had some prototypes made in the U.S. and shipped to Jordan. We also made some in Jordan using parts sourced from the local marketplace.

Part of our water monitoring device being assembled in Jordan. Photo Credit: Aziz Wadi/Field Ready

What We Learned

Learning 1 — Localisation of production — Not ‘Whether?’, but ‘When?’

Making humanitarian aid items locally is Field Ready’s mission and model. But, during this project, we found that local supply and manufacturing was impractical or unavailable for many key items including printed circuit boards and flow switches. Replicating designs using only what you find in the local marketplace is often difficult and requires adapting the design to suit available parts.

Installing the water monitoring device prototype on a water tanker in Amman, Jordan. Photo Credit: Aziz Wadi/Field Ready

Components common in the U.S. were unheard of in Amman’s markets, while vital components, such as flow switches, were rare and the wrong size, type or quality. We had to ship components from the U.S. and China to complete the testing. We are pivoting to importing batches of key components and eventual regional production in a more developed market to solve this challenge. Once the device is fully developed, localisation of production might eventually look like import of key parts, local assembly and servicing in markets this product serves. We want to avoid a situation where the tail wags the dog and we adjust the design based on market availability of component parts - many of which are imported anyway.

Once we had all the parts and made two working devices, we had to wangle a ride on a water tanker. We were given the details of a water truck driver named Saqr, meaning “falcon” in Arabic. Saqr is a local water entrepreneur serving refugee camps and private neighbourhoods. He owns five water trucks, each driven by a family member. He agreed to allow us to place two devices in two of his trucks, driven by his uncle and father.

Learning 2 — Still waters run deep.

There is much to learn about water and water-truck deliveries; something as seemingly simple as routine deliveries can be affected by many factors. For instance, our data logger and the output spreadsheet depended on a single flow-rate multiplier. We assumed that the tanker output was a flat 700 litres per minute as it could empty its 21,000 litre capacity in half an hour. But there are so many variables that affect the flow rate — e.g. the tanker is parked on a hill, the tanker’s water levels are low, the tanker’s engine is running at a higher RPM, the height of the tanker relative to the tank being filled — all affect the flow rate. These could lead to anything up to an estimated+/- 30% margin of error when we are targeting 3% to 4%. To counter this in the next phase of development, we’ll experiment with using a flow meter to get accurate water-flow readings. We’ll also explore issues around water quality monitoring.

Some of the complex models of water trucking distribution. Credit: Thomas Wildman, Regional WASH Advisor, HECA

Learning 3 — Two can play that game

Water tanker drivers confided in us some interesting ways of gaming the system. Some unscrupulous drivers will fill their tankers with poor quality water and deliver that. Some will deliver less than the contracted amount of water and then deliver the remainder to private customers. An aim of this automated IOT water monitoring device is to remove the requirement to trust people with water deliveries. One driver said that if they had a data logger in their truck, they would game it by pumping water from one compartment of his tanker to another to activate the flow switch. Making this device as automated and ‘trust-free’ as possible has to be a priority.

Another gaming challenge for this prototype to overcome is the automated time-stamp method that we developed through the data-logging software. The data recorded was just the device ID, time since first activation and times the flow switch was activated in seconds. This system would be open to gaming because the data is imprecise. It gives a time that the pump was activated and a baseline time — however, there’s no way to know about location or if the driver is spoofing a delivery.

Eliminating trust from monitoring systems can be done. Next we will test GPS tracking and real-time mobile-data transmission to the cloud through a SIM-card-enabled device. We’ll be able to geo-fence the GPS coordinate data and match the data outputs from this device to distribution locations, making this a viable and accurate basis-for-payments for water trucking companies and contractors. Field Ready’s global technical team have a vision for a trust-free IOT-enabled, real-time, mobile data solution for water delivery monitoring. This is the holy grail for a global humanitarian industry estimated to be worth hundreds of millions of dollars a year. This device has implications for the viability and cost-effectiveness of emergency water trucking (EWT) and longer-term operations.

Installation of the IOT water monitoring device data logger and RF transmitter. Photo Credit: Aziz Wadi/Field Ready

Scale vision

Field Ready and our partners at NRC share a vision for the future and together we have developed a two-track approach to scaling.

Track one — Field Ready will continue to test the device in its current form and identify use-cases with our partners. One such use-case is during COVID and conflict situations when in-person monitoring of water-distribution is suspended, but water deliveries are ongoing. The MVP product we’ve tested in Jordan can provide a minimum standard of data right now. Once we’ve completed testing the ruggedness of the system, we will deploy it to NE Syria for further trials.

Track two — Monitoring water distribution at the individual water-tanker scale can give very accurate data, especially if the trucks are GPS-enabled and the water volumes recorded are accurate to the nearest few liters. So we plan to test GPS features for location and flow meters to measure output. We’ll also explore the addition of new sensors to collect water-quality data like dissolved oxygen, total dissolved solids and chlorine-level measurements. We plan, eventually, to make the system trust-free, the tech reliable and the data outputs accurate. Once we have all three then this is a groundbreaking device that could change the face of humanitarian water trucking forever.

Watch this space…

--

--

Ben Britton
Frontier Tech Hub

International Lead for Programmes. Working at Field Ready.