Stevens Institute for Artificial Intelligence Forges Partnership with Eastech Flow Controls

The Stevens Institute for Artificial Intelligence (SIAI) and Eastech Flow Controls recently formed an industrial alliance to enhance the development of the company’s AI-driven solutions.

Eastech Flow Controls develops and manufactures smart-sensing technologies coupled with advanced analytical software for wastewater control throughout North America. The partnership, which spanned the 2021-2022 academic year, involved Eastech sharing proprietary technologies with three teams of students and their project faculty mentors in order to develop advanced AI solutions for predicting flows and risks of failure in pipelines after unanticipated rainfall.

Projects incorporate interdisciplinary study, real-world data

One of the teams was led by Kaijian Liu, an assistant professor in Stevens’ Department of Civil, Environmental and Ocean engineering, who purposely chose his four research students — two undergraduate, two graduate — from diverse academic backgrounds.

“I didn’t want all of them to be from just one area of study. I wanted them to inspire each other and I think they did,” he said.

Liu and his team developed a deep-graph learning-based model that can predict the flow of infiltration and inflow of water from rainfall or the ground for wastewater systems in order to prevent severe pipe damage and flooding.

The group chose to use a deep-graph model because “when we are talking about our civil infrastructure systems, they naturally form a graph,” Liu said. This would make it possible to “leverage the natural structure of the wastewater infrastructure system” and visualize the areas most affected by the infiltration and inflow anomalies.

Jolene Ciccarone ’24, one of Liu’s student team members, is studying software engineering and applied for the partnership after her friend forwarded her an email that advertised the project.

“One of my main interests as a software engineer is AI machine learning, and this project was particularly interesting to me because we got to work with a real-life company and leader in wastewater management systems,” she said. “We were able to use Eastech’s resources and access all the different types of data, which I thought was awesome.”

Using historical data, “we were able to simulate the inflow and feed the predictive model and then see how those results related to the actual data we were given. Through those comparisons, we could see if this predictive model is a good indicator,” Ciccarone explained. “Our predictions were pretty similar to the observable infiltration and inflow that actually happened.”

Liu, who will continue research in the same direction, sees the project’s success from several perspectives.

“The first is that we were able to provide the innovative AI capability to solve the longstanding problem with the wastewater infrastructure sector and we were able to provide AI algorithms in a method that can actually work,” he said.

“The second benefit was that I had the opportunity to work with Stevens students from different departments and disciplines, which was a really enjoyable experience. I think we are all happy with our results and deliverable. That also matters a lot to me.”

Collaboration is key to partnership

Brendan Englot, director of SIAI, applauds not only the success of the individual research projects but also the success of the new collaboration with Eastech.

“It’s exciting to see faculty from across the institute come together to solve a grand challenge like this one. The goal is to bring faculty and students together to tackle tough real-world problems in an interdisciplinary way.”

Englot added that having one of the first industrial alliances at SIAI to have great success provides a blueprint for future projects.

"Our collaboration with faculty and students at SIAI was a great opportunity to enhance our algorithms and for the students to work with our engineers on advancing solutions to a global hazardous waste situation,” added Eastech Vice President of Marketing Bob Donnelly. “We hope this project is the beginning of a host of future projects for SIAI so that it represents the value of technological collaborations that can be between industry and Stevens.”