The role of Water Operator Partnership in accelerating adoption of new working process for non-revenue water reduction: a case study – Waterworx WOP between Perumda Air Minum Tirta Moedal and VEI

Drinking water utilities play an instrumental role in society. Rural community and urban business activities rely on access to high quality and affordable drinking water for socio-economic development. For this important service to be carried out efficiently and effectively, it requires a high-performing and resilient water utility that can overcome the challenges it faces (Hukka & Katko 2021). Population growth, urbanisation, and an increasingly erratic supply of water (due to climate change) are some of the rising challenges for urban water utilities (Boretti & Rosa 2019). With water production not keeping pace with demand, the water availability per capita in many cities, towns and growth centres is dwindling.

While the global water community focuses on expanding universal service coverage (SDG6.1), in practice, this focus is often biased towards network coverage; much less so on the required operational budgets to improve the ‘water use efficiency’ (SDG6.4) of the existing and new (expanded) network. This is difficult to fathom when considering that the global NRW volume exceeds 126 billion cubic metres per year, amounting to 30% of the produced ‘System Input Volume’. Conservatively valued at only USD 0.31 per cubic metre, the cost/value of this water amounts to USD 39 billion per year (Liemberger & Wyatt 2019).

Not only is this an enormous financial concern, but elevated NRW also detracts water utilities, in a time of increasing scarcity and climate change, from reaching their goals of full-service coverage, at a reliable level of service and at an affordable price (Liemberger & Wyatt 2019). This underlines the importance of breaking the ‘vicious circle of performance decline’ in most NRW handbooks (Farley et al. 2008).

This Paper: (i) argues that infrastructure development programmes–beyond the expansion of water production and distribution capacity – should prioritize ‘small but smart (operational) investment’ in water use efficiency (i.e. Non Revenue Water reduction), and (ii) investigates if and how a Water Operator Partnership (WOP) can nurture a ‘virtuous circle of performance improvement’ whereby a modest investment sum can be recovered through achieved cost savings (reduced Real losses) and/or revenue increments (reduced Apparent losses). We do so on the basis of a case study of a WOP between Perumda Air Minum Tirta Moedal Kota Semarang (PDAM Semarang) and VEI Dutch Water Operators (VEI).

Water Operators' Partnerships (WOPs) are ‘peer-support partnerships between water and sanitation service providers. WOPs work by harnessing the skills, knowledge and goodwill within a strong utility to build the capacity and improve the performance of another utility that needs assistance or guidance’ (GWOPA 2023). In a WOP the ‘mentor’ utility shares knowledge and skills to strengthen their organisational capacity and operational performance in mutually agreed upon areas of need.

While innovation systems theory stresses the importance of knowledge (transfer) between WOP actors in the ‘innovation system’, methodological tools to systematically analyse the dynamics of these relationships are lacking. Efforts to develop and empirically validate a multidisciplinary approach that draws on social psychology (Wehn & Montalvo 2018) provide clear points of consensus as well as issues of conflict in the dynamics of knowledge transfer between WOP utility partners. Empirical evidence points to qualitative differences in goals of knowledge transfer as well as sources of differences and asymmetries in motivations, pressures and capabilities in the knowledge transfer process.

Usually, the developed ‘mentor’ utilities appreciate their capacity building role while partner ‘mentee’ utilities perceive a higher risk of failure of a unsuccessful project (Stephens et al. 2022). There are, however, also examples of WOPs that are perceived to be successful by both the ‘mentor’ utilities and partner ‘mentee’ utilities (Wright-Contreras et al. 2020). The varying context of the ‘mentor’ and ‘mentee’ utility, too, provides a foundation for peer-to-peer learning. Where a ‘mentor’ may introduce a best practice in NRW management to the ‘mentee’, for example, its applicability may be undermined e.g. under intermittent supply conditions which the mentor is not privy to. According to Doppenberg & de Blois (2020), the readiness of WOP to be introduced with a DMA approach, introduction of telemetry and commitment of senior management to establish a dedicated NRW Department resonate with contributing factors to success in other WOPs documented.

The presented case study in this paper demonstrates how the VEI-PDAM Semarang WOP benefitted both utilities whereby each partner fulfilled mentor and mentee roles. On the one hand, VEI supported PDAM Semarang in successfully demonstrating a scalable approach to NRW reduction within one of its 11 Supply Zones.

Insights into contributing factors to success, on the other hand, have given VEI a sharper perspective on how to scale-up NRW reduction efforts from supply zone to utility level.

VEI was established in 2005 as an international joint venture of the two largest water companies in the Netherlands, Vitens and Evides. 5 other water utilities have since joined VEI. Driven by a Corporate Social Responsibility (SDG6) agenda and desire to invest in the personal/professional development of staff, WOPs aim to make partner water operators stronger, financially sustainable, and more (climate) resilient by facilitating ‘peer-to-peer’ learning between utility staff. VEI is currently implementing more than 30 projects (involving 60 + utilities) in more than 20 countries.

With a population of 1.65 million (BPS 2022), Kota Semarang is one of the largest cities in Indonesia. Perumda Air Minum Tirta Moedal Kota Semarang (PDAM Semarang) is the main supplier of water to the city. Its service coverage is 52% but with a high NRW of 46% . With high NRW value, PDAM is continuously looking for cooperation opportunities to be able to reduce its NRW. One of them is through WOP WaterWorX. With WaterWorX WOP, PDAM received numerous benefits such as an alternative source of funding, and a new way of working which will be explained in this article.

Inspired by the Utilities of the Future framework of the World Bank made by Lombana Cordoba et al. (2022), VEI developed an Excel-based ‘WOP Tool’ to evaluate the utility performance and maturity levels of individual work processes. It was shown that PDAM lack maturity in managing the water distribution topic, resulting in high NRW value.

Non-Revenue Water (NRW) is defined as the difference between the System Input Volume (SIV) and Billed Volume (consumption) of water. NRW reflects volumes of water lost through leaks (Real/Physical losses) and consumed water that is not invoiced to consumers (Apparent/Commercial losses), or a combination of both (Van Den Berg 2014). In areas with limited supply, high levels of Real losses can lead to supply interruption which is also what PDAM currently experiences.

Drawing on the ‘best practice’ from many countries, the first IWA Water Loss and the Performance Indicators Task Forces produced a standardised approach for water balance calculations with definitions for all terms involved and should be developed first to help in prioritising strategy (Farley et al. 2008). Various NRW assessment tools are available to quantify Apparent and Real losses using (historical) asset performance (leak reporting, meter registry, billing system) data (AL-Washali et al. 2020). In this case study, we applied the commonly used EasyCalc (Excel) tool developed by Liemberger & Partners and available for free on the website (www.liemberger.cc). Things that need to be filled in WB EasyCalc are as follows:

• 1. System Input Volume is water that enters the distribution system. The data are acquired from bulk water reading.

• 2. Billed Consumption is volume of water that is billed to customers because of customers, usage. The data are based on customers, meter reading data.

• 3. Unbilled Authorised Consumption consists of water used for flushing of the network and from fire hydrants. These volumes were estimated using technical work order data.

• 4. Apparent/Commercial losses consist of meter inaccuracies, unauthorised consumption, and data transfer errors. Water lost to meter inaccuracies is calculated using meter testing/calibration results or – in the absence of measurement data – is estimated based on meter age and through-put (registered no. of m³ on the dial). Water lost to theft is estimated using historical illegal water use data and customers survey. Data transfer errors have been reduced to a bare minimum through information system usage and mobile phone based meter reading.

• 5. Real/Physical Losses: The tool calculates the total volume of real losses by subtracting the above 2 captioned categories (volumes) from the total NRW volume. Using leak repair records, the volume and root causes of water lost through leaks on transmission/distribution mains, reservoir overflows, and leaks on service connections was further assessed using a Physical Loss – Component Analysis tool that was developed under the WOP since this functionality is not included in EasyCalc. The breakdown of the physical loss components is crucial in prioritising intervention strategies.

The CARL volume at utility/supply zone/DMA level can be calculated through a top-down estimation using EasyCalc. At the DMA level, bottom-up field measurement provides more accurate CARL figures. Consolidated at supply zone and utility level, the two can be compared. A combination of top-down and bottom-up assessments is considered best practice in prioritising NRW intervention strategies at utility, supply zone and DMA level. For comparison, the ILI value of utilities in Indonesia is shown in Table 1. This value will later be compared with conditions in PDAM Semarang.

When considering the wide range of possible Real (Physical) and/or Apparent (Commercial) loss reduction measures it is important to prioritize:

• Dominant NRW categories in the IWA water balance: with the ‘top-down’ assessment estimating the Real/Apparent Losses volume ratio. After the ratio has been determined, a strategy formulation can be made based on 4 pillars of intervention commercial/physical losses.

• Considering the dominance of physical losses in the case study area, therefore a discussion on physical losses intervention is explained in pillars below. ‘Low-cost high impact’ interventions at utility, zonal and/or DMA level need to be selected for quick win. Pressure management interventions in supply zones/DMAs with excessively high-pressure levels (pillar 3), for example, will generate a far higher return on investment than capital intensive replacement of a dilapidated network. Even so, dominant leaks (with relatively large leak flows) will still need to be found (pillar 2) and swiftly repaired (pillar 1) with good quality material based on Performance-Cost-Risk, i.e. asset management considerations (pillar 4). The point here is for cash strapped utilities with limited resources, the chronological order of interventions too is important.

The first pillar focuses on the efficiency and effectiveness of distribution network maintenance and repair in response to visible leaks identified by customers and utility staff. The leak awareness/location/repair time, quality of workmanship, and (choice of) materials determine the leak flow rate and duration.

Active Leakage Control (ALC) aims to pro-actively detect (in)visible leaks using leak detection equipment (e.g. ground microphones, leak noise correlators) and flow and pressure monitoring (Korlapati et al. 2022). The latter typically focuses on monitoring the Daily Volume (DV) and Minimum Night Flow (MNF) as two key parameters (Farah & Shahrour 2017):

• DV expresses the total volume of water consumed by water users in the zone/DMA in 24 hours.

• The MNF is a proxy for the leak flow. Typically occurring between 02.00 and 04.00 in the morning when everybody is asleep, the Net Night Flow (MNF minus the estimated night consumption) is a reliable indicator of the leak flow.

• Extrapolated to a monthly volume, this figure can be used to quantify the Real losses volume; by subtracting this volume from the total NRW volume, the Apparent losses volume can be calculated. This enables the NRW team to prioritize appropriate Real and/or Apparent NRW Loss reduction measures.

Pressure management involves regulating and optimising water pressure within the distribution system. This is achieved through the installation of Pressure Reducing Valves (PRVs) on feeder lines to DMAs with excessively high pressure. By maintaining optimal pressure levels, it reduces the stress on pipes and fittings, lowers the risk of new leaks and bursts, and reduces the leak flow rate of existing leaks. As such, pressure management helps conserve water, reduce energy consumption, and extend the infrastructure lifespan.

The initial PRV settings should reduce the pressure to an acceptable level but not yet to an optimal (minimised) level. If the pressure drop is too large, the visibility above ground (pillar 1), and/or ability to detect leaks at greater depth (pillar 2) is compromised. Similarly, when an increasing number of leaks is found and repaired, the system pressure will increase. This is when the PRV setting can be adjusted a second time or third time to achieve the optimal minimised pressure level that lowers the risk of leak recurrence.

Asset management involves the systematic management of water distribution assets, such as pipes, pumps, valves, and reservoirs. It includes strategies for prioritising maintenance, rehabilitation, and replacement based on asset condition and criticality. Effective asset management helps utilities make informed decisions about where to allocate resources, reducing the risk of infrastructure failures that can lead to water losses. It also extends the lifespan of assets, improving their long-term reliability.

NRW levels in Indonesia currently stand at an average of 33% of the System Input Volume (Ministry of Public Works 2022). In the Mid Term National Development Plan (RPJMN) 2020–2024 the national government has set a target of 25%. A previous study in Indonesia concluded that the Economic Level of Water Losses (ELWL) for utilities with groundwater as main source is 21% (Heryanto et al. 2021). Water utilities with surface water will have a lower ELWL since water production costs and thus the value of NRW losses are higher (Horváthová 2022).

This PPP project is not without reason since customers in existing areas demand more water from PDAM (because some of the service area is intermittent) and this PPP project will help PDAM in expanding its service coverage. The benefit for existing customers was reflected in the customer satisfaction survey that was implemented by PDAM in 2019 as a baseline (Purwanto et al. 2019) and again in 2022 (Purwanto et al. 2022) to evaluate progress. It revealed two positive outcomes of the production level increase: (i) improved water availability (L/person/day), and (ii) increased customer satisfaction levels – reflecting the audited number of customers with pressure above national standards having increased from 56 to 60%. The continuous supply of water under higher system pressure, also increased the number and flow rate of (mostly invisible) leaks – as expected (physics!) and documented in literature (Al-Washali et al. 2019). It became evident that PDAM needed to step-up NRW intervention to exploit the sustained SIV increase under the PPP to improve network coverage and service levels. If PDAM is not able to control the NRW level under increased SIV/water pressure, it runs the risk of getting trapped in a vicious circle of performance decline with increasing production costs that are not recovered by increased sales.

System Input (m3): Volume of water entering the distribution network

Authorised Consumption (m3): Authorised consumption of water by customers (billed metered + billed unmetered consumption)

Non-Revenue Water (m3): System Input – Authorised Consumption; includes Real losses (leaks on mains, storage reservoirs, distribution lines, service connections), Commercial losses (meter inaccuracy, water theft and data transfer errors) and Unbilled authorised consumption (e.g. for flushing of lines after repair, firefighting, etc.)

NRW l/c/d: NRW expressed in L/connection/day

PDAM Semarang is pursuing various options to reduce NRW losses in its service area: (i) performance-based NRW reduction contracts with a private company (for designated supply zones), (ii) mobilisation of additional funds from the World Bank, and (iii) co-financing of ‘small but smart (operational) investment’ under the ongoing WOP with VEI. This paper zooms in on achievements under option 3.

Against the backdrop of (increasing) NRW levels, the WOP prioritised operational performance improvements – through strengthened working processes (maturity level growth) – in asset management, water production, water distribution, commercial operations, human resource management, and organisation & strategy – hereby:

• Establishing a dedicated NRW Department to quantify, localize and reduce NRW losses

• Quantifying NRW volumes ‘top-down’ (IWA Water Balance, and Component Analysis)

• Selecting a pilot area to demonstrate the ‘proof of concept’ and ‘business case’

• Localising and reducing NRW losses through Active Leak Detection

Evaluation of the business case to (convince senior management to) scale-up the approach.

While PDAM Semarang has a good leak reporting and repair (work order) system before the intervention, the data were not yet analysed to generate the desired ‘management’ information to guide decision-making. To upscale this, NRW department staff were trained and coached by VEI Short Term Experts (STEs) to estimate the Real losses in each of the above captioned categories using: (i) in-house data on the number of leaks, pipes sizes and length (to distinguish transmission mains/distribution mains/service lines), and (ii) empirical data (Liemberger & Farley 2004) on leak flow rates and leak awareness/location/repair time. This method of assessment is called Component Analysis (Aboelnga et al. 2018).

All of this transfer of knowledge was done via online coaching session because of the COVID-19 pandemic undermining travel. This coaching session is a substitute for the six weeks of (originally planned) on-site Short Term Expert (STE) visits. One aim was to act as a bridge for PDAM staff to share experiences, discuss challenges, and re-prioritised activities while VEI STEs monitored progress and adjusted the scope of their TA support to the emerging needs. Another aim was to facilitate the transfer of knowledge and swift translation to action, training and coaching targeted skilled, young, and motivated staff who could apply new knowledge in their day-to-day work.

PDAM Semarang had long (2009) established a dedicated NRW team to spearhead the utilities' NRW agenda. With coaching support by thematic experts and the project manager, this team was formalised as a department in 2021 (Figure 3) with a dedicated department for Real and Apparent Loss reduction under the WOP. Some benefit was gained by formalising the team to department. The first one is the capability to intervene by themselves, and the second is the ability to secure funding inside the department by ring fencing the saved money to improve the operational capability of NRW department.

The water balance revealed that Real (Physical) losses make up 80% of the total NRW volume. While Apparent (Commercial) losses are considered ‘low hanging fruit’ (with a better Return on Investment e.g. in replacement of aged/under-registering meters of large consumers) in NRW reduction, the water balance revealed that the Real (Physical) losses volume utility is much higher and thus requires urgent attention. This is also reflected in high ILI ratio of 48.

For benchmarking purposes, the ILI value of PDAM Semarang is compared with other utilities in Indonesia (Table 1). This research reveals that the ILI for PDAM Semarang is relatively above average in Indonesia. A high ILI means that water distribution network assets (infrastructure) are in poor shape, maintenance is poor, and/or supply conditions (e.g. excessively high and/or fluctuating pressure under intermittent supply) are far from perfect.

Considering the relatively small volume of Reported leaks and slightly larger but still SMALL volume of Found leaks, the need to step-up Active leak detection to identify and repair the Unreported/Unknown leaks with a relatively LARGE estimated volume (80% of the total Physical losses volume) became evident. Therefore, intervention to actively search for leaks is recommended to PDAM.

As a result, the NRW Department has since been updating this assessment annually utility-wide and for all individual supply zones to prioritise and implement NRW reduction measures. Later on, after the COVID-19 travel restriction is lifted, the on-site visit is continued.

Prior to the formation of the ALD team, zonal network maintenance teams spearheaded leak detection and repair activities (with the focus only on the repair activities). To upscale this the NRW team came in to step-up this effort by establishing the new Active Leakage Detection (ALD) team, with purpose of increasing the number of Found leaks – estimated at 0.1% of the total Real losses volume (Figure 5). Many unreported (mostly invisible) leaks, resulting in a far greater volume loss (97%), were still going unnoticed. At the time, the repair of ±25 reported/searched for leaks per day was not reducing the NRW volume as expected.

With this understanding, PDAM Semarang management agreed to form the ALD team for dedicated (full-time) leak detection. As a start, 2 ALD teams were formed. Team members were recruited from the existing plumbers in the zonal network maintenance teams. The selection process included interviews, and hearing test using acoustic leak detection equipment. An interesting fact is that all the ALD team members shared a talent or interest in music.

The training was conducted in one DMA with repair teams on standby; this so participants can experience the impact of their (team) effort for themselves first-hand. The training is also a source of inspiration to utility staff who are learning new things and using state-of-the-art equipment. Further Train the Trainer (ToT) activities focused on coaching prospective candidates to train and mentor existing and future ALD team members. Hence PDAM can sustain the (transfer of) knowledge and apply it in practice, in reducing NRW levels, in the future. The training was positive evaluated by participants with 4.6 score out of 5.

In the planning phase, the baseline NRW levels were established and DMAs with relatively high NRW levels were prioritised. ALD teams were subsequently deployed to these DMAs. Thanks to the newly trained and well-equipped ALD team, a lot of (previously Invisible) leaks were found (on average 2–4 leaks per day for both teams). To ensure that the additional leaks could be repaired within the shortest possible time, the NRW Department assigned 2 leak repair teams to repair all leaks found by the ALD team first – in doing so reducing the Awareness, Localising, Repair (ALR) time. The identified leaks (by the ALD team) that were repaired (by the repair team) are shown on a ‘leak heatmap’. The red colour indicates that leaks are concentrated in specific ‘hotspots’; these maps can be used to identify the root causes (old age of the network, high pressure, sub-standard material or pipe depth, quality of leak repair) and address these. Where applicable pipeline rehabilitation plans can be drawn up.

Besides forming new ALD teams the process of ALC requires good monitoring of flow and pressure to make an intervention decision based on data (Kowalski & Suchorab 2023). Prior to the WOP intervention PDAM Semarang already had some flow and pressure loggers installed in the distribution network, but these were not installed at critical locations to conduct MNF measurements and/or were not functioning reliably. Based on assessment and interview with PDAM, good telemetry (flow/pressure sensor, data logging and transmission) equipment should:

• 1. Have a reliable power supply: the battery should last long enough and be replaceable for sustainable operation

• 2. Be capable of transmitting data using a stables GPRS connection for remote flow/pressure monitoring on the PDAM – SCADA Dashboard.

With the existing equipment not meeting these requirements, VEI and PDAM Semarang co-invested in 17 units for monitoring the water distribution/pressure (and NRW level) real-time. These sensors/loggers are very powerful – enabling the PDAM Control Centre team to monitor the impact e.g. of (multiple) leak detection and repair interventions without having to wait for the results of monthly water balance calculations.

Complementary to the flow and pressure sensors/data loggers, WaterWorX also invested in a portable ‘clamp-on’ ultrasonic flowmeter for measuring the flow on offtakes from the incoming lines (fitted with flow sensors/data loggers) to locate where the MNF is resulting in (in)visible leaks. This Step Testing process guides the PDAM Semarang team in prioritising lines for Active Leak Detection and repair activities. It also enabled the team to measure flow rates and/or calibrate existing (mechanical) meters on large diameter (200–800 mm) pipes. Combined with a portable pressure logger, this equipment allows PDAM Semarang team to also conduct MNF/Step Testing measurements on DMAs that are not fitted with permanent flow/pressure sensors and data loggers.

The installation of flow and pressure sensors/data loggers enabled the PDAM Semarang team to close the Plan-Do-Check-Act learning loop with flow (including DV and MNF) and pressure data to evaluate the impact of last month's ALC interventions. When NRW levels are reduced to an Economic Level of Loss (ELL), the PDAM Semarang team will continue monitoring the DV/MNF levels and monthly water balance results to sustain a well-functioning (effective and efficient) water distribution network.

To evaluate the business case, hardware and staff costs were assessed and compared with the generated benefit, i.e. cost savings and/or revenue increase. Hardware cost total is Rp. 560 million ($37,500) Staff costs Rp. 360 million ($24.500) consist of ALD team who were fully dedicated to the Gunungpati supply zone during the pilot project and the zonal network maintenance team (time share spent on the project was assessed using work order data) salaries. To calculate the monetary value of the reduced water losses per year, the reduced physical losses per volume (80% of the total NRW volume of 711.000 m³ pear year) volume was multiplied with the marginal cost of (produced) water $0.036/m³). Given the 24/7 supply of water (no suppressed demand for water), the monetary value of Apparent losses (20% of the reduced NRW volume) was considered to be negligible. By dividing the total investment amount with the monetary value of the reduced water losses per year, the calculated ‘pay back’ (Return on Investment) period was 1.2 years.

In conclusion, the WOP between VEI and PDAM accelerated the reduction of NRW by strengthening the NRW organisation (establishing a dedicated Active Leak Detection team), improving existing working processes, and introducing new approaches (e.g. NRW Component Analysis, Active Leak Control) and tools (telemetry for MNF measurements, acoustic leak detection equipment). The demonstrated ‘proof of concept’ is backed by an evaluated business case: the prioritisation of ‘small but smart (operational) investment’ to improve the water use efficiency (SDG 6.4) in infrastructure development programmes that are predominantly focused on expanding network coverage (SDG 6.1) generates good return on investment as demonstrated by the short payback period of 1.2 years. This resonates with related research (Doppenberg & de Blois 2020) though ‘business case material’ (evaluations) remains scarce.

The successful pilot is also contributing to Semarang-wide scale-up of the approach. Triggered by the return on investment in 34 flow/pressure monitoring sensors (incl. data loggers) in 2021, PDAM Semarang has since procured and installed 95 additional units at strategic locations across the water distribution network. The demonstrated success has also brought a positive change within the organisation, departments and individuals. At the onset of the Gunungpati pilot project, some staff were sceptical about the new approach e.g. the ‘high’ cost of telemetry, intensified (active) leak detection etc. The achieved results are contributing to a change in mindsets – inspiring and motivating utility staff and management to adopt a ‘new way of working’. It is evident that the WOP is contributing to working process maturity level growth and utility performance improvement.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.