Inspired by the principles of biomimetics, the start-up Acwa robotics is developing a robot capable of autonomously moving around drinking water networks that are still in operation, with the aim of collecting a large volume of information. Meeting with the technical director and the founder of this company.
In France, the drinking water distribution network is estimated at just under a million kilometers of pipes. About one in five liters does not reach the user because of leaks. Their origins are many and varied: age and type of pipes, corrosion, aging joints, soil movement, high water pressure inland… The communities and operators of these networks are implementing programs to replace or repair these pipes. To help them, start-up Acwa Robotics is developing a robot capable of autonomously moving through networks in operation to locate them and characterize their condition. Created in 2018, it currently has 14 employees, mainly engineers. Interview with Jean-François Rossi, technical director and founder of this company.
Engineering techniques: What needs do the technologies developed by Acwa robotics meet?
Jean-Francois Rossi: We are developing technologies for locating buried water pipes so that we can intervene more easily in these infrastructures. In France, regulations require each community to locate them within 40 cm, including depth. We then provide information about its maintenance status. Here again, this is a French obligation, since the 2012 Grenelle de l’environnement. Local authorities must not only establish the map of these networks, but also document them: diameter of the pipes, constituent materials, condition of the pipes, installation equipment, etc. All this data must then be transcribed into an open data file called Eaufrance. If this work is not carried out, the State reserves the right to double the water bill and, since last year, the fines have started to fall.
How do you manage to collect all this information?
We are developing robots capable of entering water networks in operation, then traversing them completely autonomously, collecting information there, emerging in another location or at their point of entry, and finally returning the collected data.
For trips, we opted for a mixed means of transport. The robot is either carried by the current, a preferred solution in straight lines, or else it moves in self-propelled mode. It does not have wheels, as the networks have very significant layers of rust, which can degrade them, generating red water unfit for consumption. Our machine is inspired by biomimetics and the movement of caterpillars. She moves as gently as possible, leaning against the pipe. To move forward, she releases the front supports, the robot stretches out, then takes back the front supports, then releases the back ones and shrinks. It progresses slowly but can make turns of up to 90 degrees.
Our robots can handle water flow of up to 2 meters per second. All of them are designed and tested to withstand pressures of 20 bar, which correspond to those found at a depth of 200 meters. With our first series of machines, we target tubes with a diameter from 250 mm. From a regulatory point of view, they must withstand pressures of up to 16 bar.
How to locate the tubes?
In a drinking water network there is no GPS signal, nor any means of communication with our robot. We use multiple data sources that we merge to accurately locate the pipeline layout. We use odometry, a technology that records the robot’s movements, and already present in cars, to position them when they pass through long tunnels. An inertial drive is also installed on our machines. This gives us additional information about gravity, for example, because in a duct the magnetometers do not work and we lose magnetic north. We also take pictures of the tube elements, which can be geolocated. This is the case with these small iron circles, called keyholes, which can be seen on the surface of the roads and which allow the opening and closing of the network valves. This equipment is another source of data and is also used to reconstruct the route.
What types of sensors are installed in your robots?
They are able to accommodate different types of sensors depending on our customers’ needs. We do not manufacture these sensors, but adapt them to operate in the restricted environment of pipelines. For example, we selected distance sensors that work in water and housed them behind specific Plexiglas glass so that they can withstand the pressure, taking care not to interfere with the measurement.
Many operators want to know the residual thickness of the tubes. For cast iron ones, we calculate using ultrasonic sensors, but there is no universal solution. For the other pipes, we have to work material by material and other technologies can be used, such as those based on eddy currents. We partnered with INSA Lyon to produce a complete overview of all measurable degradation in drinking water networks; this primarily concerns metal pipes.
Until now, the devices that were introduced in the water networks took videos in low definition, with images that were difficult to use, because the machines were shaken by the flow of water. Because ours rests on the barrel and can stop moving, it is capable of capturing high definition footage.
Our customers’ requests refer to sensors that we had not thought of at the beginning, such as measuring water turbidity, hardness, pH, etc. We are working to integrate them into our robots.
What stage is your project at?
We are testing a first pre-series of machines under real conditions with network operators. At the same time, we are studying the feasibility of setting up a first manufacturing and maintenance facility for our robots in Aix-en-Provence. One of the restrictions in their manufacture is that they must have neutral buoyancy, that is, have the same density as water, so as not to consume energy to float in the pipes. They weigh about 10 kg and this is within a few tens of grams. Another difficulty in the project is the sanitary level. We have to be very careful with the materials we choose, as our machines must not degrade the quality of the water in the pipes.
So far, we’ve done very little communication, but we’re already getting orders from all over the world. The problem of aging tubes and their poor condition does not only concern France, our market is global.