Twenty years of asset management research for Dutch drinking water utilities

This paper gives an overview of the asset management landscape on drinking water in the Netherlands and twenty years of research on this topic executed by KWR in close collaboration with water companies. A description is given of research questions and the international developments in the field of asset management. This is followed by the developments on asset management at the Dutch water company Evides. Twenty years of asset management research at KWR is presented in five phases, showing a transition from the question of how can the concepts of asset management help to better plan the replacement of distribution networks, towards integrated decision making on the asset system as a whole. A focal point for research could be how research can contribute to creating value for water companies. More formal information and improved modelling will continue to play a central role; however, attention is required for making use of expert knowledge, scenario building, data quality and the integration of information of technical, financial and societal origin. This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY 4.0), which permits copying, adaptation and redistribution, provided the original work is properly cited (http://creativecommons.org/licenses/by/4.0/). doi: 10.2166/ws.2020.179 ://iwaponline.com/ws/article-pdf/20/8/2941/812380/ws020082941.pdf R. Beuken (corresponding author) D. Savic A. Hummelen M. Blokker KWR Water Research Institute, P.O. Box 1072, 3430 BB Nieuwegein, The Netherlands E-mail: ralph.beuken@kwrwater.nl J. Eijkman Evides, Rotterdam, The Netherlands D. Savic Centre for Water Systems, University of Exeter, Exeter EX4 4QF, UK

can provide to the organization. Konstantakos et al. () see as the main advantage of the ISO 55000 standard that it provides a basis for discussion and improvement derived from experience and methods that have been widely acknowledged and approved.
The above definition of asset management has a more generic angle and challenges utilities to define the strategic purpose; that is, the value of assets for the organizations and their stakeholders. The value (which can be tangible or intangible, financial or non-financial) will be determined by the organisation and its stakeholders, in accordance with the organisational objectives. Woodhouse () clarifies the different perceptions of value and benefits associated with assets and their life cycle management. Woodhouse also provides a conceptual formula for calculating total value, as being the aggregation of: • the asset capital value, defined as the total costs of replacement; • a functional performance value multiplied by the remaining economic life, where: ○ functional performance value, comprising capitalised quantitative and qualitative usage benefits, minus total operating costs; • a risk value multiplied by the remaining economic life, where: ○ risk is defined as the capitalised sum of known opportunities and threats; This structured way of looking at value promotes an agenda for improvement and research, consisting of issues like: how to achieve reliable figures for capital value, how to express performance and risk, how to define the remaining life and how to quantify (monetise) intangible aspects for stakeholders.

DEVELOPMENT OF ASSET MANAGEMENT IN AN INTERNATIONAL CONTEXT
How asset management has been implemented depends on the typical geographical and regulatory circumstances. with the aim of creating a sound and systematic asset management approach to assist water utilities in the establishment of a rational framework for rehabilitation plans of water and wastewater networks. In the Netherlands, the water company's shareholders (mostly municipalities and in some cases provinces) focus on low tariffs, and water companies adopted asset management mainly to become more efficient and transparent and to get a grip on the expected large-scale replacements of distribution mains.

ASSET MANAGEMENT AT EVIDES
Evides is the second largest drinking water company in the Netherlands. It supplies drinking water to two and a half million inhabitants and industrial water to large industries in the South-West of the Netherlands, including the city and port of Rotterdam. Evides uses predominantly surface water from the river Meuse, treated at one centralized plant.

Evides introduced Asset Management officially in 2009
with the objective of optimising the total cost of ownership (both CAPEX and OPEX) while maintaining an agreed level of performance. During the first phase, the foundation for asset management was laid with the focus on defining roles and decision procedures on a strategic, tactical and operational level. Also a system of key-performance indicators and associated reporting tools was established.
As of 2011, more focus was given to predicting future investments. With regard to the distribution of drinking water, investment analysis shifted from being driven by external parties (reactive approach) to being driven by internal analysis (proactive approach), taking into consideration the expected life and the impact of failure on the company's values. With regard to the abstraction and treatment of water, the focus was on identifying vital installations and predicting the end of their technical life.
Particularly for distribution, better analytical tools and methods were required. This raised the awareness of the need to improve data quality. Currently, a number of projects are under way to improve the quality and availability of the data. Additionally to the Joined Research Program, Evides started an internal research programme for exploring, amongst others, future treatment concepts. These concepts should enable Evides to considering the system from source to tap and anticipate future water quality challenges like viruses, medicine residuals and diffuse pollution due to agriculture.
Better-founded plans, including the required capital costs, help Evides to establish realistic scenarios and to evaluate these considering performance, risk and costs (CAPEX and OPEX). Although regulation in the Netherlands is less strict than it is in countries like the UK, Evides is increasingly taking governmental and societal aspects into consideration. This results in more emphasis on efficiency and effectiveness, benchmarked on issues like cost, customer satisfaction, interruption of supply, water quality and sustainability performance indicators.
Currently, Evides is putting considerable effort into cooperation with partners on replacing distribution infrastructure. This cooperation with fellow network operators and municipalities requires an ever closer collaboration, not just on the scale of a single project, but on an entire programme considering a large number of projects over a number of years. This cooperation results in long-term relations with contractors, providing them with more certainty and continuity. In addition to these developments, the company is exploring the impact of the energy transition on asset management. This will potentially result in infrastructure interventions, such as demobilisation of gas pipes, construction of networks for heating and cooling and an increased capacity for electricity grids. The energy transition and associated changes could also have a huge impact on the water distribution network. These challenges are to be combined with adaptations for climate change and transformation towards a circular and low-CO 2 society. In order to be proactive, Evides invests in more knowledge on the current assets and their resilience towards a rapidly changing and more demanding future.

RESEARCH ACTIVITIES ON ASSET MANAGEMENT IN THE NETHERLANDS
The ten Dutch and one Flemish drinking water companies are shareholders of the water research institute KWR.
These water companies and KWR work closely together to co-create the so-called Joined Research Programme and its outcomes. This research programme is redefined every five In this period, field observations were collected from the networks that were constructed using the self-cleaning principles. Field measurements (particle counting, flow measurements and flushing experiments) showed that no accumulation of particles took place in self-cleaning networks (Blokker et al. ) [10]. Furthermore, cost savings were reported for construction due to shorter lengths and smaller diameters as well as for operation, as no flushing was required.

Start of making use of improved data: 2008-2012
The focus of the research period from 2008 to 2012 was on condition assessment, risk analysis and performance measurement of distribution networks. In this period, water companies applied predominantly asset management methods to water distribution networks. Information systems became more mature as improved software for enterprise resource planning, GIS and asset registers became operational. This development resulted in more and better quality data. The use of digital devices by pipe fitters enabled an easier and better controlled process for the registration of pipe failures. In 2009, five water companies started to register pipe failures in the USTORE database. Initially, data on failures and the network assets were stored in spreadsheets (Vloerbergh & Blokker ) [11]. In 2010, the USTORE web platform was launched enabling water companies to upload data into a database.
In the same year, two more water companies joined USTORE.
Information on the condition of large-diameter pipes can be obtained by mains inspection. In Beuken et al.
(), different options for inspections were described and inspection with the Acoustic Resonance Technology of a 300 mm cast-iron main was described in detail [12]. The report also provided business cases indicating that, in addition to technical limitations of the in-line inspection technology, it can only be applied cost-effectively for 300 mm diameter and larger pipes.
A simplified formula is being applied in the Netherlands to calculate the so-called erosion craters (pits resulting from a pipe burst). By using this formula, a buffer zone is calculated around pipes, which is used to identify what objects are potentially prone to damage in case of a pipe burst.
This formula is rather conservative, resulting in large buffer zones. This is unfavourable for water companies, since a large buffer zone means that high impact objects (dykes, highways, important buildings, etc.) could potentially be affected. A more realistic formula, resulting in smaller buffer zones, means that fewer mains are in high impact areas. Such a formula is available; however, it requires extensive computation, combining GIS and allmains hydraulic models. In van Daal et al. (), the adopted method and a case study are presented [13]. This research shows that more complex computation can help water companies make more realistic asset analyses. In this example the additional computational complexity resulted in a decrease in the number of high-risk mains.
In this period, water companies were also approached by various software distributors offering decision support software packages for mains replacement. As the water sector wanted an objective opinion on these products, KWR was engaged to assess potential results and the necessary requirements for their use. Several workshops were organized in 2010 to allow several software companies to present their solutions. Subsequently, pilots were organised with water companies whereby these packages were applied and assessed. Results of these pilots can be found in Beuken & Blokker () [14]. The main conclusion of these pilots was that the available operational input data and the strategic considerations (often applied as weights) were not yet sufficiently mature to obtain reliable results.
Intensifying and broadening asset management During this research period a more integrated approach for asset management was sought, applying asset management also to assets for (ground) water abstraction and treatment. Asset management for treatment works (including pumps and reservoirs) seemed to be more challenging than for distribution networks. This was due to the uniqueness of individual treatment processes and installations, where the design and operation are adjusted to the water quality of the source. Where a distribution network is described by a handful of different asset types, a treatment plant counts hundreds of them. Furthermore, data availability is a challenge as the Dutch water companies want to apply the concepts of asset management to treatment plants but are still setting up asset registers. Therefore, a database on asset failures similar to that for distribution networks is much more difficult to establish. Another challenge is the lack of a clear indicator on the quality of performance, such as customer minutes lost in water distribution networks. In some cases risk analyses have been applied, but companies are often reluctant to follow them because of the huge administrative burden for setting up such an analysis. Research done by KWR was focussed on making quantitative fault tree analysis and knowledge exchange on risk analysis and operation.
Asset management for groundwater abstraction wells appeared to be easier to implement, as most of the well systems have a rather uniform design. Research into clogging of wells done at the beginning of this century (see Van Beek ), showed that clogging due to particles could be prevented by switching pumps on and off on a regular basis. Clogging due to accumulation of chemical precipitates, however, may be best prevented by continuous abstraction [21]. The knowledge of those phenomena has been implemented into the operation of well fields. Within asset management research, methods were described for monitoring clogging behaviour and the regeneration of clogged wells. Switching procedures have been optimized, balancing between a stable water quality performance and minimized costs for regeneration.

Current research
In the current research period, which started in 2018 and will finish in 2023, asset management research is executed following two pathways. Research topics of a mono-disciplinary nature are considered in a disciplinary theme (such as abstraction, treatment and distribution). The complementary theme of Integrated Asset Management covers topics that transcend disciplinary themes and focusses on integrated decision making. Also, attention is given to the semi-public role of water companies and the involvement of stakeholders into decision making.
The mono-disciplinary asset management research in the distribution theme focusses on deterioration processes of PVC and asbestos cement mains, a statistically representative method for extrapolation of monitored pipe failure frequencies towards future performance and network blueprints.
The projects make use of more data becoming available, for example failure data from the USTORE database, condition data from pipe inspections and data from various GIS sources. In the theme focussing on groundwater abstraction, guidelines for well operations are further being developed.
The overarching goal of the Integrated Asset Management theme is to develop a framework for integrated decision support, taking into account: • internal factors, with a focus on reliable information for decision making, the link between activities and the company's strategic objectives, risks and integrated performance measurement and risk analysis;

LINKING RESEARCH TO VALUE FOR WATER COMPANIES
Asset management is about creating value, and research on asset management should help achieve that goal.

LESSONS LEARNED
Looking back at twenty years of asset management research, the most important lessons learned are as follows: 1. The drinking water companies participating in the Joint Research Programme feel an increased urgency for an integrated model-driven approach. This will enable them to transition from organisations based on expert knowledge into those based on formal information systems and being able to make use of new technologies and complex decision support models.
2. Where the drinking water companies considered future developments mostly as extrapolations from the past, it is now felt that the future is unpredictable and that the paradigm of supplying drinking water will change.
Asset managers should therefore not try to rebuild existing assets but focus on redesigning them according to a set of future states.
3. Developments in the field of information management and methods for analysis will result in more good-quality data and better based decisions. Promising developments in the KWR research are, amongst others: • optimisation techniques (the Gondwana platform) for asset management planning; • machine learning for a better understanding of pipe breaks, and • building digital twins for analysing the asset management system as a whole, subject to various future scenarios.
4. Obtaining objective, representative and complete data sets is crucial for a better management of assets. This is, however, a time-consuming task and the need for data of appropriate quality is often not sufficiently understood.
Moreover, water companies prefer to make use of data that is opportunistically available rather than based on systematic and targeted data collection.
5. The problem of aging assets continues to play an important role in asset decisions. Condition assessment and definition of a risk-based residual life remains a crucial aspect. This is especially true for complex and tailormade installations.
6. Financial figures are used for business case and scenario evaluations. Effective asset management will only be possible when water companies possess detailed figures on capital and operational expenditure, as well as on the cost of failure or interruptions to supply. 7. Water companies apply registration on asset performance.
These registration activities serve merely for organisational issues; that is, they provide instructions for maintenance activities. A secondary use of these activities is for asset performance analyses. These analyses serve the more long-term purpose of accessing the system behaviour for defining, amongst others, risk-based inspection or scenario analysis. This is done less frequently. It also appears that these analyses require much higher levels of data quality and that considerable cleansing activities are required to perform these more detailed analyses.