Nature-based solution optimised by responsible AI

Delivery of a 44,000m2 integrated constructed wetland to improve water quality in the River Dearne in Yorkshire, has benefitted from improvements in safety, cost and sustainability thanks to intelligent plant combined with AI.

Working through a wet and cold winter, the team from Mott MacDonald Bentley (MMB) used AI to efficiently schedule the cut and fill operations and soil distribution workflow before using intelligent plant to carry out the work. Guided by a driver, the plant directly followed the design requirements reducing the need for manual setting out techniques and banksman within the earthworks. Not only did this improve site safety by limiting people-plant interaction, it accelerated progress and improved accuracy. Despite moving over 63,000m3 of earth there was no overcutting and no need to transport any of the materials offsite.

"The use of the intelligent plant and AI allowed us to optimise the delivery of the Dearne Reach project in a multitude of ways, including reducing health and safety risk and maximising our efficiency,” says Ben Gouldsborough, capital delivery project manager for Yorkshire Water. “We were pleased to have the programme finished ahead of schedule - a direct result of the use of these new techniques."

This achievement was the result of ongoing investment in digital innovation by MMBC. "Our plant and digital teams are dedicated to upscaling and consolidating new digital solutions and innovative technologies, including AI, with the goal of enhancing safety and productivity,” explains Dean Holmes director at MMBC. “Achieving these objectives necessitates collaboration across various functions, project teams and the supply chain with our clients. This is a fantastic example of the brilliant outcomes that this is generating for our clients and communities.”

Improving water quality

Integrated constructed wetlands are sometimes described as “nature’s kidneys” because of their ability to filter and clean wastewater. “A constructed wetland is an engineered ecosystem that tries to mimic natural processes,” explains Olivia Rowe project leader for MMB. “At Dearne Reach the driver was removal of ammonia.”

Ammonia is a compound of nitrogen and hydrogen found in domestic sewage and agricultural runoff, which creates conditions that are toxic to aquatic life. Wetlands can remove this and other nutrients through biological adsorption by the plants, chemical conversion by bacteria living within the ecosystem, and physically slowing the flow of water which gives suspended solids more time to form sediments that drop out of the flow.

At Dearne Reach 52 species of plants have been designed to sit in 13 adjacent cells in which over 350,000 plants adsorb nutrients from storm overflow discharges and wastewater flowing out of the Clayton West wastewater treatment works.

With a regulatory deadline of July 2025 fast approaching, it was imperative that the team delivered the project on time. “We were under pressure because it is unusual to start an earthworks project in Autumn knowing it will run through the winter,” says Paul Pickering, site agent for MMB. But it was the only way to ensure that long-term client Yorkshire Water could meet the regulatory requirements for water quality set by the Environment Agency.

350,000 plants from 52 different species

4km length of river improved

Working with trusted suppliers

“From the outset we knew that early collaboration with the digital delivery team to aid programme efficiency, improve accuracy and elevate health and safety would be really important,” says Olivia.

Not only did it keep the project on track, but the use of digital tools also including AI, enabled MMB to use its existing equipment to deliver the work rather than hiring in larger machines or outsourcing work to subcontractors. “There can be a tendency to oversize equipment during earthworks especially when time is constrained,” explains Richard Goodhead, digital plant lead for MMB. “So, I wanted to use digital tools to analyse the gap between the design requirements and the machinery performance to confirm that we could use what we had to meet the deadline.”

For this analysis Richard worked with an existing supplier, machine control specialist MDiG which regularly supports MMB with control systems for its Komatsu intelligent machine control plant. Using the Smart Construction Dashboard and Smart Construction Simulator from Japanese company Earthbrain, Richard and the team took this to the next level, combining it with AI to analyse the site operations.

“The dashboard first visualises the jobsite using point cloud data from the site survey, and target data from the design,” explains Daisuke Sakamoto, digital transformation manager at Earthbrain. This enabled the cut and fill volumes to be calculated automatically. The software can also monitor the progress of operations in real time as the intelligent machines report progress directly to the dashboard. The progress report function then enables the site team to generate visualisations of different views, for example cross sections comparing the current status with design requirements. It also provided real time measurements of the earthwork enabling the site engineers to accurately forecast completion dates.

Coupled with the work of the designers who had already optimised the site hydraulics to manage groundwater on the site, the team managed to achieve a perfect balance between cut and fill material meaning that there was no need to import or export soil.

Optimising operations

“Seeing how useful this was I wondered if we could take this even further and use digital tools to optimise the operations,” says Richard. For this the team turned to the Smart Construction Simulator which firstly used AI to calculate the optimal combination of cut and fill operations and therefore soil distribution based on proximity and the layout of the 13 wetland cells. It then activated its machine simulation function to allow the team to find out whether the equipment that it planned to use was optimal for the work.

“Before running the calculations, we have to set the parameters for the equipment,” says Daisuke noting that in each cell excavators, bulldozers, and trucks work together to meet the optimum soil distribution arrangement already defined. Richard wanted to know if the D61 dozers that MMB already owned could do the work efficiently. Inputting the data for the blade length, the capacity of soil moved, its speed and the height of spread along with similar details for the other equipment allowed the simulator to run algorithms that carried out the work virtually reporting data back to the team within seconds. The critical output for Richard was a figure called the operating factor. The higher the number and closer to 100 this was the more efficient the operations were. “If you are getting 50 or 60 per cent then you know that you need to change something,” says Daisuke. This could mean that the trucks are running half empty or the excavators are left waiting for trucks to move material.

“This really helped us to understand the balance between the number of trucks we had versus the number of excavators, and it proved that the D61 machines would be able to do the work,” says Richard.

Once finalised the earth shifting model was uploaded directly to the intelligent plant, which used on board GPS to navigate around the site. “The engineers imported the model from the site cabins and then the machines got to work,” says site agent Paul Pickering.

This was a change for the site engineers who usually spend their time physically setting out the site levels before guiding them around the site continually checking to ensure that the drivers were within the design levels. Instead, they uploaded the model, monitored progress on the Smart Construction Dashboard, and checked earthwork levels at the end of the day. “It made the site a lot safer as it eliminated the people-plant interface risk which was brilliant as these machines can weigh 50 or 60t presenting a significant risk if adequate controls are not in place,” says Paul.

Importantly it also accelerated progress as Paul no longer needed an engineer to physically monitor every machine, meaning that he could have more excavators operating simultaneously. “It probably cut the programme by a couple of months,” he adds. This proved to be crucial as the weather had other ideas. December was so wet that work had to stop and in January the team faced snow and freezing temperatures. But the time savings offered by the intelligent plant and the optimised schedule kept them on track, ultimately allowing them to complete work three weeks earlier than initial predictions saving £350,000.

“I have been doing this work for 40 years and there is no doubt in my mind that using intelligent plant is the future,” says Paul.

Future potential

“Using machine control reduced the amount of earth that had to be moved whilst still delivering the project requirements,” says Richard noting that the Smart Construction Simulator could also be used in the planning stages to support concept design and tendering, which wasn’t part of the scope of work for MMB in this project.

After successful completion of the work Richard and the team re-analysed the machine data to take a deeper dive into the construction methods. Interestingly this revealed some useful and not so useful insights identifying that deeper shorter drives were more efficient, but at the same time suggesting complex combinations of excavation routes and blade angles that would be difficult for human drivers to achieve.

All of these findings will feed in to future projects and it supports development of the Smart Construction Simulator too, which improves with every project it works on. “If we assembled the same team for a wetlands project tomorrow we’d do it even faster,” says Paul Pickering. This is good news given the need for sustainable solutions to help water companies like Yorkshire Water mitigate and adapt to the effects of climate change. By optimising the design and construction of this nature-based solution with state-of-the-art digital tools such as AI and intelligent plant, MMB has set a new benchmark for delivery.