Non-revenue water component specific drivers in Lilongwe city

Abstract

The Lilongwe Water Board (LWB), one of the water utility companies in Malawi, whose water losses as of 2022 were as high as 40%, employed network rehabilitation techniques in their 2015 to 2020 strategic plan to reduce the then water losses of 38.9% to 28%. However, as Non Revenue Water (NRW) is caused by myriad drivers, network rehabilitation activities have been proving futile. This study analysed the NRW component-specific drivers for LWB in Lilongwe City. A quantitative research design was adopted in which purposive sampling of the specific sites, Area 22 and part of Area 2 was used. R studio version 4.1.1 was used to perform descriptive statistics, pairwise Pearson Correlation tests, Minimum Night Flow Model, Fixed Effects Regression Modeling, and a feedforward backpropagation Artificial Neural Network based Improved Garson Algorithm's sensitivity analysis. The water balance framework for the two District Metered Areas (DMAs) (SZA1 and SZD3) was established, which confirmed that although post-rehabilitation NRW (38.95%) is below the 2022 national average (54.61%), it is still above the LWB’s target and global average NRW range of between 30 to 35%. The water loss components analysis showed that apparent losses (AL) contributed more to water losses than real losses (RL) and unbilled authorized consumptions (UAC). Different component-specific drivers were further identified for these NRW components. Maintenance works were the main driver for UAC, while accounting errors, illegal connections, and customer non-payment drove AL. Background leakages and bursts, connection density, type of pipe materials, and population density were the main RL drivers. In conclusion, this study has revealed the component specific NRW drivers for the LWB in Lilongwe City, Malawi. It recommends the need for targeted interventions to address the identified drivers and NRW to meet the LWB's goals and global standards. Keywords: Non-revenue water, potable water, district metered area, apparent losses, real losses, unbilled authorized consumptions