New model offers ‘clear path’ to keeping clean water flowing in rural Africa

More than 184 million people in rural sub-Saharan Africa rely on shared handpumps for clean water. However, more than 50,000 of the pumps are broken, leaving millions in jeopardy of losing their safe water supply. 

Research from the University of Notre Dame studies how local nongovernmental organizations in Ethiopia, Malawi and the Central African Republic decide when to maintain and repair 3,584 community handpumps when information about pump functionality is incomplete or uncertain. The ongoing research has helped NGOs provide more consistent and safe access to water for more than a million people across the three countries. 

“Our results show that well-timed preventive maintenance can substantially reduce downtime and, in many cases, lower logistics costs — an important factor for NGOs with limited budgets,” said lead author Chengcheng Zhai, assistant professor of information technology, analytics and operations at Notre Dame’s Mendoza College of Business. Her findings in the paper, titled “Keep the Water Flowing: The Hidden Crisis of Rural Water Management,” are forthcoming in the journal Manufacturing & Service Operations Management and won the 2024 Service Science Best Cluster Paper Award from the Institute for Operations Research and the Management Sciences, a prestigious international research recognition.

Along with co-authors Alfonso Pedraza-Martinez, Notre Dame’s Greg and Patty Fox Collegiate Professor of IT, Analytics and Operations, and Rodney Parker, Kurt Bretthauer and Jorge Mejia from Indiana University, Zhai examined detailed data on mechanic visits and water-point functionality from three countries that use very different approaches to water handpump maintenance. 

Based on field research, Zhai and colleagues learned that the NGOs operating in the Central African Republic visit each water handpump once per year as part of a pre-scheduled rotation. The NGO in Ethiopia has an incoming call center that relies on community reporting of handpump breakdowns. The NGO in Malawi runs an outgoing call center, which contacts each community to inquire about the handpump condition so broken ones can be repaired. 

The team developed a dynamic optimization model, called the Markov Decision Process, that adapts to each NGO’s current maintenance model and identifies optimal schedules for mechanics to visit and conduct the necessary maintenance and repair services, with the goal of reducing water pump downtime while balancing the logistics cost. 

“Under this approach, NGOs perform both preventive maintenance and any needed repairs during each scheduled visit,” Zhai said. 

Using this model, the NGO observes how many pumps are reported broken in each cluster and how long it has been since the last visit, and then decides which cluster its mechanics should visit next. 

There are a number of complicating factors in determining maintenance practices. NGOs responsible for maintaining the water sources operate with limited staff and budgets. They must carefully plan their mechanics’ travel routes to keep pumps working while minimizing downtime and the costs of transportation and spare parts. At the same time, they must decide how much to invest in collecting information from communities about pump failures and whether their maintenance strategy should be more proactive (scheduled visits) or reactive (responding to reported breakdowns). 

It can be tough to figure out if fixing things as they break or sticking to a strict schedule actually works best. 

“If there is plenty of data and the water pumps are reliable, reacting as things break is best,” Zhai said. “But if repair demand is high, sticking to a scheduled, proactive plan works better.”

 Applying their analytics model across thousands of water handpumps and multiple time periods revealed substantial reductions in downtime. 

Enhanced logistics including well-scheduled preventive maintenance and repair service reduced maintenance downtime by 47 percent to 62 percent in Ethiopia, up to 53 percent in Malawi, and by 42 percent to 55 percent in the Central African Republic), with varying cost impacts ranging from savings in Malawi and Ethiopia and a 15 percent to 19 percent increase in the Central African Republic. 

“This vital research offers a clear path to keeping water flowing, delivering sustainable, safe water access for underserved rural communities across Africa and strengthening health and long-term resilience,” Zhai said. “Our findings are critical for NGOs working with tight budgets and incomplete information.” While NGOs may assume preventive maintenance increases costs, the study shows that when preventive maintenance is scheduled optimally, significant logistics savings are achievable. For organizations with limited budgets, these savings can free up resources for other critical activities. Even in cases where costs rise, the major improvements in pump uptime can justify the added expense. 

“This paper is part of our comprehensive research agenda on water management in sub-Saharan Africa. We are using analytics to improve the location, allocation, maintenance and funding of water projects in countries like Malawi, Ethiopia and the Central African Republic,” said Pedraza-Martinez, who also serves as faculty director of the Meyer Business on the Frontlines program. “Moreover, thanks to a collaboration with the Frontlines program at the Mendoza College of Business, MBA students are building on our research to advise NGOs and social enterprises focused on water management, helping these organizations improve their strategies and operations. This is research with a positive and tangible social impact.” 

Contacts: Chengcheng Zhai, 574-631-5445, czhai2@nd.edu; Alfonso Pedraza-Martinez, apedraz2@nd.edu