Development of current condition assessment and target definition model for water balance practices in sustainable water loss management

Non-revenue water (NRW) includes components of apparent losses, real losses and authorized unbilled consumption. Apparent losses consist of losses due to water meter inaccuracies and illegal consumption. Real losses cover the leaks in transmission lines, mains and service connections. Various methods have been used to analyze and monitor these losses in water distribution systems (WDSs). A standard water balance was proposed by the International Water Association (IWA) in order to establish the water budget in a standardized structure (Lambert et al. 1999; Pearson 2019). The water balance method is the most basic approach applied to determine the main water loss (WL) components in WDSs. However, it is not always possible to obtain field measurements of the input and output components in the water balance and the data are estimated. As a result, an evaluation that does not represent the actual conditions is made (Silva et al. 2018). The water balance table proposed by the IWA allows the summarizing of the components and monitoring of the input/output components. However, the reliability of the water balance is directly dependent on the data quality of the components in the table. For this reason, Aboelnga et al. (2018) carried out component analysis by considering fault records for the most appropriate leakage management in a WDS. Water balance table, minimum night flow and component analysis (with failure records) methods have been applied for leak estimation in WDSs. The failure rates, network conditions and field data used in these methods are highly effective in estimating leaks (Amoatey et al. 2018).

Unreported leaks, which constitute a significant part of the leaks in WDSs, occur mostly at service connections. The active leakage control in district metered areas (DMAs) should be applied in order to improve these leaks in the water balance table and system performance should be monitored regularly (Boztaş et al. 2019). Calculating water losses and leaks in WDSs is essential for monitoring performance, active control, estimating its economic impacts, and establishing network management plans. In one study, minimum night flow rate and fault records were considered to estimate leaks (Negharchi & Shafaghat 2020). Apparent losses are another important component in water balance analysis. Meter tests and monitoring of customer consumption profiles are required to accurately determine these losses. Generally, however, apparent loss components are estimated without fieldwork as a percentage of input volume or NRW volume. Therefore, for sustainable water balance analysis, administrative loss components (illegal use, meter errors) should be determined according to field data and the data quality of their components should be questioned (Farley & Liemberger, 2005; Mousavi et al. 2017; Ncube & Taigbenu 2019).

Percentage of input volume and process indicators (losses per main length or losses per unit service connection) have also been used to monitor the performance of NRW components in water balance analysis (Al-Washali et al. 2019). Similarly, unavoidable real loss (UARL) volume, infrastructure leak index (ILI) and real loss process indicators have been calculated to analyze the performance of leaks (Lambert et al. 1999; Lenzi et al. 2014; Klosok-Bazan et al. 2021). The ILI indicator is calculated as the ratio of the annual volume of physical loss in the system to the UARL. This indicator considers the physical characteristics of a network (network length, number and length of service connections) and system pressure (Lambert et al. 1999; Cavazzini et al. 2020). The methods applied to reduce the real loss components in the water balance table are generally time consuming and costly. For this reason, economically recoverable leakage rates should be determined by considering the current rate of the components and the benefits and costs of the methods, and the most appropriate leakage rate target should be defined accordingly (Yilmaz et al. 2021a, 2021b). In this context, the economic leakage level model was proposed using optimization algorithm in WDSs. The basic data in the system should be obtained correctly in order to perform an accurate analysis with this model (Firat et al. 2021). Water utilities work hard to reduce water losses, resulting in high costs. To control water loss, first identify the relative contribution to loss of the various components. Next consider how much would have to be spent for loss reduction in that component. Based on this, determine which investment has the greatest impact for the least cost (Moslehi et al. 2021; Serafeim et al. 2022).

The water loss rate consisting of apparent and real losses is considered for the performance analysis of administrations in Turkey. The major disadvantage in considering the water loss rate is that, when entering data, authorized unbilled consumptions are often overestimated. Authorized unbilled uses are estimated at higher rates to calculate lower rates of apparent and real losses. This may be due to factors such as political and administrative pressures, lack of awareness and attention, failure to audit the accuracy of the data entered into the table, or the lack of a supervisory mechanism in order to show the institution's water loss rate performance well. To ensure that an administration appears to be doing a good job in reducing water losses, systems in place are needed to ensure that the system is not ‘played’.

Therefore, in this study, a new method is proposed to evaluate and highlight the reliability of data of water balance practices and hence the results obtained using the original scoring structure. The purpose of this study is not to propose a new water balance table. Rather, the quality of the data used for standard water balance analysis and the application levels of the applied methods are questioned. Thus, the data quality (quite good, good, doubtful, poor and quite poor) is classified according to the application level of the water balance practices. The class of the component is determined according to the examination and evaluation made in the administration being examined. This classification is not made by the personnel in the administration. For instance, it is not possible to talk about an accurate water balance analysis in systems where the system input volume and authorized billed consumptions, which are the most basic data in water balance analysis, are not measured regularly. These components that do not have a regular on-site measurement (only estimation) are not suitable for use in water balance analysis. Therefore, those components identified as poor data quality (with 0 or 1 points) based on scoring should be improved for an accurate water balance analysis. Similarly, the application quality of the methods applied to determine real and apparent losses in the field is also questioned. Thus, components with poor data quality and application level are identified within the scope of water loss management. Accordingly, it will be possible to provide more accurate data for sustainable water balance analysis in an administration and to make field applications more systematically.

In WDSs, the water supplied to the system consists of two basic components, authorized billed consumption (revenue water) and non-revenue water (Lambert et al. 1999). Non-revenue water consists of water losses (apparent and real losses) and authorized unbilled uses (Lambert et al. 1999; Pearson 2019). A standard water balance table was proposed by IWA to analyze and monitor losses in WDSs to a specific standard (Lambert et al. 1999) (Table 1). This table defines standard terminologies to express water balance. In this study, the standard water balance table recommended by IWA and widely used in many administrations is taken as reference. In addition, indicators or components used in establishing the water balance and monitoring the process/performance in an administration, which are not included in this table, have also been defined. Components that need to be determined directly in the field in this table are: system input volume, billed metered consumption, billed unmetered consumption, unbilled metered consumption, unbilled unmetered consumption, apparent loss components (illegal consumption, losses due to meter inaccuracies), and real loss components (leakages in mains, services connections and reservoirs). Other components in this table are calculated using these data.

In the literature, top-down and bottom-up methodologies have been used to analyze a water balance and define WL rates (Lambert et al. 1999; Mastaller & Klingel 2017a, 2017b; Mamade et al. 2018; Pearson 2019; Chawira et al. 2022). Moreover, component analysis based on failure records is also used to estimate the leaks in WDSs (Aboelnga et al. 2018; Amoatey et al. 2018; Negharchi & Shafaghat 2020). The bottom-up methodology includes detailed process steps such as planning and on-site implementation of district metered areas (DMAs) in a WDS, regular monitoring of input flows, minimum night flow analysis, and leak detection (Al-Washali et al. 2016; Amoatey et al. 2018; Mamade et al. 2018). The DMA could be expressed as a zone that is isolated from other networks with isolation valves and generally has 1 or 2 inputs. In a DMA, the input volume, customer consumption and network components are completely separated from other regions. Thus, leaks, customers and other components are managed more effectively (Pearson 2019). Therefore, the bottom-up methodology includes detailed field work and is generally more costly than the top-down method (Pearson 2019; Serafeim et al. 2022).

The ‘top-down approach’ is suggested as the simplest way to enter data into a water balance table (Liemberger & Farley 2004). This method involves less detailed work and is easier to implement. In this approach, the NRW volume is calculated by subtracting the authorized billed consumption volume from the system input volume measured by a flow meter. Then, water losses are calculated using the authorized billed consumptions and system input volumes. In order to analyze the strengths and weaknesses in the scope of WLM and to make a more accurate assessment of the water loss components, the apparent and real losses should be determined based on field data. However, data for some variables are estimated depending on factors such as lack of data, equipment and personnel in the administration, or technical and financial status. In this case, the sub-components of apparent and real losses are estimated and analysis is not made according to field data. Thus, an accurate assessment is not made about the components that need improvement in the administration. The current application levels of sub-components in the field should be analyzed first to define a more realistic and applicable roadmap for each component.

In Turkey, the ‘Regulation for Control of Water Losses in Water Supply and Distribution Systems’ published on 8 May, 2014, defined WL targets for administrations. Later, on 31 August 2019, the relevant regulation was updated and the targets were postponed for 5 years. In Turkey, generally the top-down method is used to analysis a water balance. Moreover, a regulation was published on 4 October 2021 in Turkey, which includes methods that water administrations should implement in order to reduce water losses in the next 5 years. These methods are basically the establishment of database systems (GIS, SCADA, and failure management systems), active leak control, DMA planning, night flow analysis and leak detection, meter management and fault management. The DMA method is applied in many administrations in Turkey in order to achieve the water loss rate targets specified in the regulation and to improve system efficiency. Minimum night flow analysis is performed in DMAs to accurately determine leaks (Özdemir 2018). In addition, a SCADA system is used to improve system performance and monitor DMA flow and pressure data (Güngör et al. 2019). Pressure control systems are applied in DMAs in order to reduce leaks in distribution systems (Özdemir et al. 2021).

The main problems experienced in water loss management and water balance analysis in distribution systems can be given as follows: (i) lack of roadmap for data collection, monitoring and analysis, (ii) lack of a methodology to assess the data quality of water balance components and to describe the current state, (iii) lack of a method to identify weaknesses/strengths of water balance components based on current status analysis, and (iv) lack of methods and tools to define targets and components that should be focused on according to the current situation analysis. A model that suggests the current conditions of an administration and queries the effectiveness of the methods applied for the components should be used for effective and sustainable water loss management and water balance analysis in water administrations. In this context, a sustainable strategic WL management model was proposed and outlined by Bozkurt et al. (2022), which contains matrices such as Module (Current Condition Matrix), Data Matrix, Module B (Current Performance Evaluation System), Module C (identifying weak and strong components within the scope of water loss management) and Module D (Methods and Tools Matrix) (Figure 1).

In the strategic WL management model, Module A aims to evaluate the data quality of components and current application level of methods in WLM. The data matrix contains the data required for analyzing the performance indicators within the scope of WL management. This matrix is associated with the components in Module A. According to the scoring of the components in Module A, the data quality of the components in the data matrix is defined. Module B includes key performance indicators used in WL management. Indicators are calculated based on data quality. In case of data defined as of poor data quality, a recommendation is offered as ‘data quality should be improved so that the indicator can be calculated’. Module C determines the strengths and weaknesses in WL management practices based on Modules A and B, and examines the risks and difficulties for WLM. Module D includes the methods and tools used to improve components in water loss management.

In this study, the current situation assessment structure for water balance components was developed and a scoring structure was determined for the components based on the strategic WL management model proposed by Bozkurt et al. (2022). This model aims to evaluate and highlight the reliability of data in a water balance analysis and hence the results obtained. Thus, it aims to identify the components with missing data, to perform gap analysis, to define the components that need improvement, and to identify the weaknesses and strengths of water balance components.

In this study, the standard water balance table given in Table 1 and recommended by IWA was considered in determining the model components. This table defines standard terminologies to express water balance. In addition, the indicators used in monitoring system performance according to the water balance in an administration are also defined as components in the model. Accordingly, a total of 27 components were defined. The components considered in this study are thought to cover all practices in water balance management. However, this number of components can be reduced or increased according to the requirements in future studies. These components can be given as:

• (i)

  Basic data measurement and monitoring: applicability and roadmap of water balance, system input volume, authorized billed metered and unmetered consumptions, authorized unbilled metered and unmetered consumptions)

• (ii)

  Practices for the determination of water balance components: systematic measurement and monitoring of real loss components for water balance calculations, systematic measurement and monitoring of apparent loss components for water balance calculations, losses due to customer water meter inaccuracies, leaks in the distribution system (mains and service connections), leaks in reservoirs, losses due to illegal use.

• (iii)

  Water balance analysis: estimation of leakage components with explosive and background estimates (component analysis), estimation of leakage and establishment of water balance based on minimum night flow, GIS-based water balance analysis, water balance calculation - comparison with different calculation methods, analysis of components that need to be reduced priority according to the water balance table, monitoring water consumption and resource efficiency, GIS-based Integrated Water Loss Management Model.

• (iv)

  Performance monitoring for water balance components: performance monitoring and integration of information systems, analysis and monitoring of NRW and components with different indicators, water balance monitoring, update system and analysis, monitoring of real loss performance indicators, monitoring of apparent loss performance indicators, monitoring authorized unbilled usage changes, target definition for NRW and basic components, definition of economic leakage level (ELL).

In this system, sub-components are classified as ‘Basic Level Applications’, ‘Intermediate Level Applications’ and ‘Advanced Applications’ (Tables 2–4). While grouping the sub-components, their applicability in the field, requirements (data, technological infrastructure, methods or processes to be applied beforehand), difficulty levels in implementation and economic requirements were considered. Thus, information that will form a reference in determining the priority sub-components for improvement will be produced.

The components considered in this study, the most basic components (data, method/field application, and indicator) for which data should be measured in water balance analysis were included in the ‘basic level applications’ group. Then, the components that need to be determined or calculated according to the experience, data and application results in basic level applications are listed in the ‘intermediate level applications’ group. Finally, advanced level practices were defined in an administration, considering their feasibility in the field, analysis difficulty and data requirements. For these advanced applications, it is important that the basic and moderate level components have a certain level of application and quality.

In the developed system, a scoring structure between 0 and 100 was defined for the water balance components. The scoring structure created is ‘quite poor: 0 points’, ‘poor: 20 points’, ‘insufficient: 40 points’, ‘moderate: 60 points’, ‘good: 80 points’ and ‘quite good: 100 points’. The application levels of the water balance components of an administration are determined in the field by considering these scores. Accordingly, the improvement targets of each component and the methods and processes to be applied are suggested considering the nature of the problem and its applicability in the field.

If a water balance component has a score of 0 or 20, that component is defined as a weakness in water balance practices. It is expressed as ‘a component with poor data quality, a component that needs improvement in the first degree’ for the data to which that component is linked. Moreover, it is recommended not to be used in any analysis. If a water balance component has a score of 40 or 60, that component is defined as a weakness in water balance practices. The data to which that component is linked is expressed as ‘data of dubious data quality, data that needs improvement’, but is used in analysis. If a water balance component scored 80, that component constitutes a strength in water balance practices. It is expressed as ‘data quality is good, economic and technical criteria should be considered for improvement’ for the variables in the data matrix to which that component is associated. Finally, if a water balance component scored 100, that component constitutes a strength in water balance practices. It is expressed as ‘data quality is very good, the current situation should be preserved’ for the variables in the data matrix to which that component is linked. Thus, ‘only good quality or regularly measured data’ is used in water balance analysis and management. According to these ratings, processes or components that need to be improved primarily in the administration are defined.

During the development of this model, efforts were made to ensure that the scoring structure was appropriate and applicable to the field conditions. Field experience, approaches suggested in the literature and on-site examination activities of administrations were considered as the basis for creating the scoring tables. Moreover, a lot of effort has been made within the scope of grouping the components as basic, intermediate and advanced level. In particular, the data and technical requirements of the components, application costs and applicability in the field were considered in grouping moderate and advanced practices. Also, during the model implementation phase, it was determined that decision makers and technical personnel have sufficient experience for systematic application and awareness in basic and moderate level practices. There was generally no problem in scoring the components in these two groups. However, in advanced applications, it was determined that there is not enough awareness about some components in certain parts of the administrations. The components in this group were scored after detailed explanations were given to the technical personnel in the administrations about the technical details of these components.

The current situation assessment matrix given in detail in the previous sections was applied to 4 pilot administrations in Turkey. Scoring was made for each variable in pilot administrations in accordance with the current situation. A unique scoring structure was defined for the components (Tables 2–4). The most important benefit of this model is this scoring structure, which is defined to express the current status of the components. These components are not scored by the administration's technical staff or decision makers. In order to define the current state of the administration, each administration was visited separately by the authors (expert team) of this paper. In this context, all practices in the administration were evaluated in 5 units, namely the Drinking Water Management Unit, the Customer Management Unit, the SCADA Unit, the GIS Unit and the Information Technology Unit.

For basic data measurement and monitoring: in the customer management system, the quality of basic customer data such as the measurement and monitoring quality of authorized billed consumption and monitoring quality of authorized unbilled users and consumption was questioned. In addition, the quality of the system input volume data was evaluated by the queries made in the SCADA unit.

For field practices to determine the water balance components: the activities and reports in the Drinking Water Management Unit were examined to evaluate the quality of the practices for the determination of leakage components in the field. Similarly, the quality of field applications for apparent loss components was evaluated in the examinations made at the Customer Management Unit.

For water balance analysis and performance monitoring: the quality of implementation of different calculation methodologies (component analysis, bottom-up, etc.) for water balance analysis was examined and scored. Similarly, the quality of indicators and processes used to monitor the performance of water balance components was scored.

The score results obtained for each administration are given and discussed in Table 5.

In basic level practices, the highest scores were achieved for authorized billed metered and unmetered components. These components constitute the strength of the administration and the data quality is good. However, the components of ID 1, ID 3, ID 4, ID 5, ID 6, ID 7, ID 8 and ID 9 (as indicated in Table 5) are generally of poor condition and constitute weaknesses in water balance practices. Especially since the system input volume has a direct effect on the water budget, all water sources should be measured and device calibrations should be checked regularly. In addition, it is understood that the components (ID 4 and ID 5) which cover the determination of the apparent and real loss components in the field, are of a poor level. Data should be obtained from the field in order to accurately analyze these losses and plan prevention methods. In moderate level practices, it can be said that the components are generally at a quite poor level and constitute weaknesses in water balance practices. These components constitute the sub-components of apparent and real losses, and the activities carried out to determine them in the field should be improved. Similarly, in advanced level practices, it can be said that the components are generally at a quite poor level and constitute weakness in water balance practices. These components constitute the sub-components of apparent and real losses, and the activities carried out to determine them in the field should be improved. Under these conditions, the water balance analysis made by considering these components in the administration is highly unlikely to represent the field. Therefore, these components should be improved as a priority for an accurate water balance analysis. In this developed system, targets are defined according to the dynamic structure and current score of each administration (Figure 2).

The components of the basic level practices (between ID 1- ID 9) are the components that are poor and insufficient level and need improvement (Figure 1, Table 5). The components of ID 4, ID 5, ID 6, ID 7, ID 8 and ID 9 should be brought to the average level first, then to the good level, and to a quite good level based on financial conditions (Figure 1). Hence, since ID 2 and ID 3 components are at a good level in the current situation, this situation should be protected first and then improved to a quite good level according to financial conditions. In moderate and advanced level practices, since all components except ID 14 are quite poor, these components should be improved first to average level (by doing pilot applications), then to good level and to quite good level according to the latest budget conditions.

Overall the scores are at good level in Administration II. Especially ID 1, ID 4 and ID 6 components are at a quite good level and constitute the strengths of the administration. Moreover, ID 3, ID 5 and ID 8 components are at a good level, again forming the strengths of the administration. These components are important for the systematic implementation of the water balance and for the determination of apparent and real losses according to field data. On the other hand, ID 2, ID 7 and ID 9 components are at an moderate level and need improvement. Especially since the system input volume has a direct effect on the water budget, all water sources should be measured and device calibrations should be checked regularly. In moderate level practices, it was determined that the ID 10 and ID 11 components are insufficient and constitute a weakness in water balance practices. Field activities should be improved as these two components directly cause loss of income. However, the components of ID 12, ID 14, ID 15, ID 16, ID 17 and ID 18 are generally at good level. In advanced level practices, components are generally at good level. The components of ID 19, ID 21 and ID 27 components are average. Practices in this level generally include advanced activities in WL management. Despite this, it is seen that the administration has a good performance. According to these results, it can be said that the data quality for the water balance filled by considering these components in these conditions is at a good level and will represent the field. However, the moderate and insufficient components should be improved to improve the data quality. The targets defined for each component according to the scoring results for Administration II are given in Figure 1. The components of ID 2, ID 7, ID 9, ID 10, ID 11, ID 19, ID 21 and ID 27 which are currently at moderate level should first be improved to good level and then to quite good level. Components with good scores in the administration should be improved by considering economic criteria.

In basic level practices, it can be said that the scores are generally at moderate level. It was determined that the ID 5 and ID 8 components are at insufficient level in the administration. The data on these components are generally considered suspicious for the administration and need to be improved. In particular, determining the apparent and real losses in the field, regularly measuring all sources and calibrating the devices is quite important for an accurate water balance analysis. Moreover, it was determined that the ID 10, ID 11, ID 12 and ID 16 components were at insufficient level and needed to be improved. It can be seen that the ID 13 and ID 14 components are at a good level, while the ID 15, ID 17 and ID 18 components are at moderate level. In addition, the components of ID 26 and ID 27 are at poor level and constitute a weakness. The components ID 20–24 were determined as moderate level, while the ID 19 component was determined as insufficient level. According to these results, more field work should be carried out for a correct and representative analysis of the water balance in the administration. The possibility that a water balance created by using data of dubious accuracy to represent the field can also be evaluated as doubtful. In basic level applications, the scores available for the components are generally at moderate level, and all components in this group should be improved to good level first and then to quite good level (Figure 1). Components that are insufficient and poor level in the administration should be improved as a priority. Components with good current scores should be improved by considering economic criteria.

In general, it can be said that the scores are at a good level. The components of ID 2 and ID 3 are at a quite good level and constitute strengths of the administration. Moreover, the components of ID 1, ID 4, ID 6, ID 7, ID 8 and ID 9 are at a good level, again forming strengths of the administration. On the other hand, the ID 10 component is insufficient, the ID 11 and ID 16 components are at a poor level, and the ID 15 component is at moderate level and these components should be improved. The components of ID 12, ID 14, ID 17 and ID 18 are at good level or at quite good level and constitute strengths of the administration. The ID 27 component is at quite poor level, ID 22, ID 23 and ID 24 components are at moderate level. These components should be also improved. According to these results, the activities of systematically applying a water balance, determining the real losses based on field data and monitoring the performance changes are at a quite good level. It can be said that the data quality for a water balance filled by considering these components is at good level and will represent the field. The current scores for the basic level components are at good level, and the ID 5 component should first be improved to good level and then to quite good level. Moreover, the ID 11 and ID 16 components should be improved to moderate level first, then good, and then quite good level. Also, the target defined for components of ID 10, ID 13, ID 15 ID 19, ID 22, ID 23 and ID 24 are at good level then quite good level.

As a result, the developed model considers the dynamic nature of a system in the evaluation of water balance components and defines targets for each component accordingly. Gradual improvement is recommended by considering the current scores and defined targets. Moreover, components with very poor and poor data quality should not be used in analysis and should be improved first. This ensures that only components with good data quality are used. The components that need improvement are defined by considering the current situation of an administration. Thus, an applicable and accessible roadmap can be determined for the improvement of the components.

In this study, a methodology is proposed to test the current application level of water balance components, which are widely used in WL management. For this purpose, a total of 27 components were defined to cover water balance management practices in administrations. A scoring structure was defined that allows scoring between 0 and 100 for these components. Thus, the application levels of water balance components in administrations were scored gradually and improvement targets were defined accordingly. The developed model was applied to 4 pilot administrations. The improvement targets were defined in accordance with the dynamic structure according to the score obtained by the administrations for each component. In addition, the components that need improvement were determined according to the current scores of the administrations. The water balance analysis made by considering these components in Administration I is highly unlikely to represent the field. Therefore, these components should be improved as a priority for an accurate water balance analysis. Moreover, it can be said that the data quality for the water balance filled by considering these components in these conditions in Administration II is at a good level and will represent the field. However, the moderate and insufficient components should be improved to improve the data quality. On the other hand, more field work should be carried out for the correct and representative analysis of the water balance in Administration III. The possibility that the water balance created by using data of dubious accuracy to represent the field can also be evaluated as doubtful. In Administration IV, the activities of systematically applying the water balance, determining the real losses based on the field data and monitoring the performance changes are at a quite good level. It can be said that the data quality for the water balance filled by considering these components is at good level and will represent the field. The model will especially contribute to the evaluation of the quality of the data of the components used in the water balance analysis. According to this assessment, measurement quality needs to be improved in the field for components with low scores. This will also enable the monitoring of the impact of the actions taken to improve the data quality of the components. Since the performance analysis according to the water balance table is generally performed annually, it is also recommended to evaluate the quality of the data used in this analysis annually, using this model.

This study was produced from the PhD thesis of Cansu Bozkurt. The authors thank Inonu University, Scientific Research Project Funding for their financial support (FBA-2021/2457). This research was supported by TUBITAK (Turkish National Science Foundation) under the Project Number 220M091.

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