The Case for Fast-tracking Service Connection Renewals to Reduce Water Loss

Julian Fyfe (formerly of Wellington Water), Cedric Papion (Stantec)

Renewal of assets approaching the end of their serviceable life is one of the four pillars of leakage management. However the pace of renewals can often be too slow to prevent increases in the rate of rise in leaks and bursts, particularly on a deteriorating network. In the Wellington Region, the aging drinking water network is exhibiting high and growing water losses as leaks and bursts form faster than they can be repaired. More than 75% of the observed faults have been found to occur on service connection valves and pipes. Under the current renewals model, service connections (and ridermains) are typically scheduled for replacement when the supporting water main is due for renewal.

Recognising that multiple repairs of a single service connection can quickly amount to the cost of replacement, Wellington Water commissioned Stantec to find a cost-effective approach to renewing service connections (reactive, proactive or both) outside of the BAU mains renewal model. The aim is to test the theory that dedicated service connection renewals could reduce losses and slow the rate of rise in leak formation in the short term and avoid or defer significant investment over the longer term. This paper presents the analyses related to identifying target service connections for proactive renewals, quantifying the water loss amelioration benefits, and the comparative economics of different renewal approaches. 

The service connection leak history was first analysed against several factors known to contribute to leak formation such as age, pressure and material. Several machine learning algorithms were used to model the likelihood of leak for each connection. While the predictive power of the preferred model is insufficient to be meaningful at the level of individual connections, it provides a basis to determine geographical priorities and compare scenarios.

The model also allows the identification of areas where service connections renewals are the most likely to lead to a reduction in leak rate. This may be used to scope a programme of capital works dedicated to proactive service connection renewals. Unlike watermain renewals, service connection replacements are relatively straightforward and low-risk in that they cause minimal service disruption, involve shallow excavation, do not require heavy or expensive plant and equipment and have a limited footprint within the road carriageway. In theory, it should possible for dedicated crews to undertake the works efficiently at scale, particularly if unusually complex sites are omitted from scope.

The model was then used to quantify potential leakage reduction benefits and simulate various investment scenarios. The first considered reactively replacing faulty service connections instead of repairing them. Other scenarios consisted of combinations of reactive renewals and various levels of investment in proactive renewals.

The study’s findings will allow Wellington Water to benchmark service connection renewals against leakage management options and inform the organisation’s water demand reduction strategy.