How to fight a pandemic (with the help of your feces)

Big Brother is watching you, but not from where you'd expect. Don't worry he (literally) wants you no harm!

If you're living in Ashkelon, a 130'000 inhabitants city located between Tel Aviv and Gaza, you've probably heard of the full-scale initiative KANDO is currently running with Ben Gurion University of the Negev.

The aim is to detect COVID-19 outbreaks where they are more likely to be identified. And that's neither at the entrance of the hospitals, nor in massive PCR testing campaigns.

The truth is out there, and has been for years: wastewater tells you everything about a population. Their activity, their habits, their health. And therefore, their possible contamination with the new coronavirus.

How? We'll come to that in a moment (read on!).

For now, let's meet Ari Goldfarb, CEO of KANDO and guest of today's episode of "(don't) Waste Water". As always, you can listen to the full story here. And by subscribing (obviously for free), you ensure you'll never miss any new episode!

The surest way to add value is to scratch your own itches

Ari's concern with wastewater and the way it is handled started long before even thinking of his career path. Born and living a few hundred meters from the Mediterranean, he was then "spending most of [his] time surfing in the sea". Thus, he saw "how polluted the seas become, year after year".

That's where he knew, he wanted to play a role, and to become an environmental engineer.

Operating a WWTP

Starting his career as process engineer on a wastewater treatment plant, he got frustrated to see that sometimes, he would "come in the morning to the WWTP to see that bacteria were disturbed in the reactor by an unexpected incoming pollution".

You can of course try to then mitigate things, but what he learnt the hard way there is that:

You don't have any control on the raw wastewater coming into the treatment plant

Rich of this field wisdom, Ari transitionned to...

Designing and Building WWTP

Now working for a consultancy engineering firm, Ari was striving to deliver "the best treatment plants". But despite designing "great treatments based on great technologies", he soon realized a powerful truth:

Then, reality comes.

Wastewater is not a stable stream, it has up and downs when it comes to pollution, and you have actually no control on it.

That fortified a strong belief for Ari: "the treatment is not starting and ending at the WWTP". There was instead much to do on the sewer network.

If you want to make it productive, to make it effective, you should look at the whole picture.

You should look at the source of the wastewater. You should look at the whole city. 

Actually, when dealing with wastewater there is much at stakes, especially in Israel. As Ari recalls: "70% of Israel is a desert" - you can imagine what this implies in terms of water scarcity. So, wastewater is before everything else a resource: "80% of wastewater is reused in agriculture".

"If you eat a watermelon in Israel, you are actually eating wastewater that's been reused"

Having identified and experienced those challenges first-hand, Ari took on to solve them. And that involved...

The two legs which, combined, create knowledge

When you look at cities in 2020, data is quite everywhere. Roads, drinking water networks, electrical grids... all of them have transitioned to the so-called smart grids, with the help of cameras and sensors.

When it comes to wastewater, the story is different.

Leg 1: The right to collect and analyze data

Nowadays, at best, you get flow and level data on the wastewater network, but nothing on quality.

There are two main reasons for that:

• Sewer networks are a challenging environment to collect data from (flow, corrosive atmosphere...)
• Sensors needed are very sophisticated and thus very expensive

That's where KANDO's approach kicks in:

in order to bridge this gap, we are measuring very indicative parameters and we can translate them to things that are interesting for our client using tools of AI and machine learning

The company relies on one hand on simple parameters (pH, conductivity, temperature), gathered on a proprietary data logger, cloud synchronized.

On the other hand, KANDO partnered to develop an advanced system, with a sampler automatically triggered by events, taking samples at the suited moment in time to best detect events.

With close to 10 years of data gathered that way, that makes for a solid database. But what for?

Leg 2: Cut through the fog and understand what you see

This is where the company's process expertise kicks in. Data gets interpolated, and feeds models to translate into insights, and finally knowledge.

"When a client opens our system, he actually doesn't see the raw data, what he sees, is if he has a pollution event somewhere in the city, what caused it, and what will be the impact on his WWTP" Ari explains.

This knowledge actually comes as the 3rd level of a 4-step ladder:

1. Collect Data
2. Analyze Data
3. Create Knowledge
4. Act and automate based on that Knowledge

But as much as the insights generated from the knowledge help operators to take decisions, that "action" level still remains in the hand of SCADAs and humans, first for safety reasons.

We are a knowledge system: we are not operating anything. We just give knowledge to the operator, so that he can operate the system in a better way!

Knowledge: a paradigm shift

If you recall Ari's experience as a process engineer working on a WWTP, operators used to be passive receivers of pollutions, trying to work the best out of it.

But now, equipped with knowledge on what's going on, where and when, they can take control, knowing that there are changes in quality.

Based on our experience we see between 2 to 4 events per week where wastewater quality dramatically changes in big cities.

Usually the operator would notice only once a month!

This new awareness kicks in on both ends of the value chain. Industrial managers that used to have only a vision on their product (and not on the wastewater) suddenly know live that their plant is generating a pollution event.

They would "sit with their family for dinner on a saturday when they receive a SMS telling them they are now discharging high pollution into the collection system". As they return to work the next morning, they will make sure that it does not happen again!

On the other end of the wastewater stream, WWTP operators get a notification, telling them a higher organic load is flowing down the sewer and will hit them in 3 hours. "They can make sure that their system is ready to get it" Ari explains.

All of this is enabled by KANDO's Clear Upstream, a full-package solution building on:

• Hardware (developed internally and in partnership, based on off-the-shelf sensors)
• Software (analyzing the data and translating it into knowledge)
• an user-friendly cloud interface (bringing the knowledge together, in an actionable way)

So, how to detect a pandemic outbreak?

As we've seen it, digitization acts as an enabler. But as Ari explains it's not about "liking technologies as technologies". It's about "building solutions".

When the indicative parameters measured by KANDO's IoT systems trigger it, the automatic samplers on the network collect some wastewater, that will be later analyzed in partner labs.

Within a few hours, results of the analysis can be fed back into a model, that calculates the amount of population infected, upstream of the sampling point.

Cutting down the "infected" area into sub-watersheds and repeating the procedure, you can narrow down the zone until you find a "hotspot". This was for instance demonstrated when KANDO's system detected a cluster, coming from an Hotel in Ashkelon where the government had quarantined sick people.

Over time, results will also allow to draw trends, and better understand the virus circulation.

Outlook

The application of network monitoring to pandemic monitoring is nothing new: it already was applied for instance in a polio event, back in 2013.

Yet, the bigger picture still contains a wide untapped potential. Combining the sewer data with other streams, such as "electricity production, water consumptions or worker's activities in manufacturing plants" would help to "get even more insights".

And as Ari explains:

the tools that we have today (AI, machine learning) for analyzing data are almost unlimited

So the limiting factor lies in the ability (not so much) and the willingness (that's the hardest part) of those various data sources to work together, building a common platform.

But for sure, that's only a matter of time. And if the COVID-19 outbreak ended-up being a trigger to this trend, we could happily call it a welcome side effect.

***

Download the full episode here

What do you think of this approach to network monitoring? Did you ever face WWTP performance challenges associated with incoming pollution? How do you think that COVID-19 monitoring in sewers can help to fight the pandemic? Share your feedback in the comment section!