Introduction

Water is one of the most basic needs for all living things, including humans (Geleta and Fufa 2021; Hunde and Itefa 2020). Water is a physical, social, cultural, economic, and political resource that is critical to human health and well-being. Drinking water is a human right that is critical to everyone's survival. As a result, the water distribution system forms part of key infrastructure that is vital to society (Abduro and Sreenivasu 2020; Adedoja et al. 2018). When water supply cannot meet demand, people suffer water shortage accelerated by undesirable pressure within distribution systems (Girsha et al. 2016; Jilo 2020).

In African countries, one-third of urban water supply systems are not continually functional. High population growth rates, scarcity of sources, treatment plant size, reservoirs, and storage tank capacity, power outages to run water pumps, high leakage problems, or some combination of these conditions are the primary causes of intermittent water supply in water distribution systems (Agunwamba et al. 2018; Alemu and Dioha 2020). The significant gap between the amounts of water produced into the distribution system and water lost in the systems is another major element that impacts water utilities. Water is lost in enormous quantities due to leaking distribution system pipes, couplings, valves, and fittings (Al-Washali et al. 2018; Gajbhiye et al. 2017). Access to adequate drinking water affects the majority of developing countries in Africa and Asia. This is aggravated by a lack of hydraulic performance and adequate demand analysis (Kwietniewski et al. 2019; Mekonnen 2018).

The availability of water resources and projecting future water demand are critical factors in most urban planning and sustainable development strategies. Forecasting water demand can be done by developing a mathematical water demand model based on the various elements that affect water usage (Enbeyle et al. 2022; Gökçekuş et al. 2019). Depending on the purpose and modeling approaches, water demand forecasting can be done as a short-term or long-term projection. Short-term forecasting consisting of a few days or weeks is frequently necessary for the operation and maintenance of existing water delivery schemes. Long-term forecasting takes into account a variety of elements that aid in the planning, growth, and design of new infrastructure, as well as the identification of effective strategies to fulfill water demand (Enbeyle et al. 2022; Mekonnen 2018; Mekuriaw 2020).

A water distribution system is a complex system of hydraulic control parameters connected to convey water from sources to consumers, as well as the physical state of all water pipes in the network (Hunde and Itefa 2020; Zhou 2018). However, water distribution systems, which consist of numerous hydraulic elements such as pipes, valves, reservoirs, pumps and pump stations, and various valves, are the most crucial in supplying communities with their water requirements (Pietrucha and Tchurzewska 2018). It is dynamic system designed to deliver water from the source to consumers' taps while instantaneously sustaining demand, pressure, and water flow (Abebe 2020; Kilinç et al. 2018; Wolde et al. 2020).

The quality of service that a utility can provide is measured by its effectiveness and efficiency of its water distribution network to deliver the required quantity of water under sufficient pressure and at an acceptable level of velocity during various normal and abnormal operational situations (Hamza et al. 2021; Milkecha and Itefa 2020). The ability of the water supply system to satisfy the necessary demand is influenced by the condition of the pumping units, network design, pipe material, and pipe age (pipe failure, leakage, and excess demand) (Geleta and Fufa 2021). Improving a water distribution system's hydraulic performance is the first and most obvious thing to do because it is necessary to supply a particular set of demand locations with the right flow and pressure while preventing substantial changes in the parameters (Beyene 2020; Jilo 2020). According to the national standard, the operating pressure in the distribution network should be 15–60 m under normal conditions and 10–70 m under exceptional conditions, and water velocity should be between 0.6 and 2 m/s, although one can find pipelines with zero velocity in the looped system (Milkecha and Itefa 2020).

Water shortages and frequent service interruptions are compounded by undetected leakage and complicated network systems, as well as a mismatch between demand and supply (Al-Washali et al. 2018). The efficient delivery of water supply distribution networks can help to reduce poverty and create an enabling environment for long-term development. However, water supply utilities in developing countries are faced with the challenges of low service coverage, high unaccounted losses of water, and low water quality (Mazumder et al. 2018; Hussein et al. 2021).

The performance of water distribution systems meeting consumers’ expectations is a major challenge for water authorities all over the world. Spatially, most developing countries' water distribution systems are frequently interrupted (Alemu and Dioha 2020; Awe et al. 2019). As a result, proper hydraulic performance measurements are needed to study the behavior of the water supply and distribution system, requiring flow, pressure, velocity, head loss in each pressure-raising main, efficiency, and pump operation point (Adedoja et al. 2018; Kabir et al. 2018).

Meeting demand and system performance can have a severe impact on the socio-economic activities of a region. Currently, the town of Addis Kidam is facing the major challenges of inadequate quantity and performance of the water distribution system throughout the town, due to, high water loss, low system pressure, low flow velocity, and town expansion required to meet rapid population growth. In addition to these, the urbanization of towns and old infrastructure is also making it more difficult to meet required water quality requirements. Not meeting water quantity and quality adversely impacts socio-economic activities. In this study area, children and women were more affected than other groups of the population because that have to fetch water every day from several unimproved alternative water sources (hand-dug wells, rivers, and springs) across a long distance (Administration reports 2020). The main objective of this study was to evaluate the existing hydraulic performance of water distribution systems in Addis Kidam; using static analysis and WaterGEMS software together with considering future forecast demand. Statistical analysis was used to analyze the current water demand, water supply coverage, total water loss and future water estimation of the whole town. WaterGEMS V8i was used to evaluate the hydraulic performance parts of the water distribution systems of the town.

Methodology

Description of the study area

The study area is the town of Addis Kidam, which is the capital town of Fagita Lekoma Woreda in the Awi zone of Amhara regional state, Ethiopia; Fig. 1. It is located 100 km from the regional capital town of Bahir Dar, 17 km from the zone capital town of Injibara. The woreda has 29 kebeles, which have two urban and 27 rural kebeles with a total area of 67,733 hectares. There is a 24 h power supply in the study area and telecommunication access. As per the Ethiopian temperature zoning, 75% of the woreda is classified as Dega and 25% is classified under the woinadega climatic zone. The temperature record in the woreda ranges from 20 °C to 22 °C, and the annual precipitation ranges from 1,500 mm to 2,500 mm with an average record value of 2000 mm.

Fig. 1
figure 1

Map of the study area

Full size image

Water supply to the town of Addis Kidam comes from a spring called “Bilti” located west of the town. The spring is capped and was constructed in 1972 E.C. The spring has a yield of around 7.5 l/s (in 1 h pumping 35 m3), but the water level at the collection chamber has now dropped after consecutive pumping. The collection chamber is designed to store water for a larger operational time to ensure suitable operational pumping meet the demand of the service reservoir. Collection is from an underground masonry chamber. There are two existing service reservoirs supplying the town of Addis Kidam. The first service reservoir is a 75 m3 masonry basins that was constructed in 1972 E.C, and the second is a 200 m3 reinforced concrete basin which was constructed in 2004 E.C. Currently, both are functional but insufficient to satisfy the town water demand.

Data collection

The most important methods in any research are the collection of significant data for the study. The source of data was collected from both primary and secondary data sources. For this study, the primary data are elevation data (junctions, tanks, and pipes), pressure of junction, flow of pipes, water level in the tank, and water quality samples. Secondary data of this study are water production, water consumption, water line network, water reservoir, base population data, water consumption, and roughness coefficient of distribution pipelines. These data were collected using experimental tests, field observation, focus group discussion, and interviews. The pressure data at tested junctions, flow measurements, and water quality samples in the distribution networks were collected using experimental tests. To collect field measurements for this investigation, systematic random sampling techniques were used (Awe et al. 2019; Hossain et al. 2021). The water distribution network sampling size and techniques for model calibration and validation followed the standard methods of water distribution sampling size and techniques (Ahmed 2022; Kuma and Abate 2021). The size of the samples used for this study was 12% and 15% of all network pipes and junctions, respectively. Pressure readings were taken using a pressure gauge from both the high and low-pressure zones of selected places in the distribution networks at end-user taps (such as customers, institutions, and commercial tap points) throughout town. The data collection techniques were done by conducting a field observation or visit collecting data from Addis Kidam Town water supply and sanitation service office and distribution networks. The total numbers of focus group discussion and interviewers are 12. Water utility staff members of the town were also used to obtain additional relevant information. Data were collected to analyze water demand and forecasting future water demand and the hydraulic performance of water distribution systems using statistical and WaterGEMS software.

General frameworks of water demand and forecasting of future water demand and hydraulic performance of distribution systems, such as calibration and validation of water distribution system of the town, were as follows in (Fig. 2):

Fig. 2
figure 2

Flowcharts of data collection and methods of data analysis

Full size image

Materials and software used

The following materials and software were utilized in this study for data collection and analysis: (Tables 1 and 2). In addition to the pressure gauge, a water meter, stopwatch, and plastic jerrycan of 10 and 20 L were used to measure the required data for the calibration of the water distribution network.

Table 1 Materials utilized in this investigation and their intended uses
Full size table
Table 2 List of software used in this investigation and its purpose
Full size table

Method of data analysis

Statistical analysis of water supply coverage

A statistical analysis was used to analyze the amounts of water produced, water consumed, water losses, water coverage and level of connection per family. During the statistical analysis, a Microsoft Excel sheet was used to calculate water supply coverage, water losses, per-capital demand and level of connection per family.

The town’s water supply coverage was assessed based on average per capita use and family connection levels. The annual use was aggregated from individual domestic water meters to calculate the average per capita consumption. The amount of connection per family was examined in addition to the average per capita water consumption (Girsha et al. 2016).

Average daily per capita consumption was calculated by combining the volume of water consumed for residential purposes at all sites within the town’s subsystem. This was used to examine the distribution of water supply coverage at various locations (Ayale 2018). The following expressions were used to calculate the town's average daily per capita consumption:

$${\text{Per capital consumption }}\left( {{\text{l}}/{\text{person}}/{\text{day}}} \right) = \frac{{{\text{Annual Consumption }}\left( {{\text{m}}^{3} } \right)*1000 {\text{l}}/{\text{m}}^{3} }}{{{\text{Numbers of population of town}}*365}}$$
(1)

One method for assessing the town’s water coverage was to consider the number of water connections per family. According to the CSA (2018) average family size of 4.5, the following equation was used to calculate the average number of connections per family.

$${\text{Connection per family}} = \frac{{{\text{Total number of connections in the subsystem}} }}{{{\text{Number of population/}}\left( {\text{Average family size}} \right) }}$$
(2)

Analysis of water loss

The total water loss analysis of the town was evaluated using the proportion of Non-Revenue Water (NRW) obtained from the total production and actual consumption of the study town (Rajasekhar et al. 2018). The difference between the water produced and the water billed or consumed is sometimes called water loss (Gajbhiye et al. 2017; Gökçekuş et al. 2019). According to Jilo (2020), Unbilled authorized consumption is the quantity of water used for operation purposes, to clean collection chambers, services reservoir, pipe flashing, and water discharge to preserve water quality, pressure test, chlorination of distribution, and the quantity of water used to a public area (like school, health center, etc.,), which is free from charge. Six to eight years of current water production and consumption in the system were used to analyze total water loss. The total NRW in the system was calculated for each recorded year, by using this equation.

$${\text{NRW }} = {\text{ SIV }}{-}{\text{ BAC}}$$
(3)

where, NRW = Non-Revenue Water, SIV = System Input Volume, and BAC = Billed Authorized Consumption.

A field visit was also used to collect the utility average leak flow, number of reported bursts, and average leak duration (Ahmed, 2022; Ayale, 2018). The physical loss in the mainline was calculated using the available data together with available data taking into consideration the minimum system annual physical loss, known as Unavoidable Annual Real Loss (UARL) (Kilinç et al., 2018; Rajasekhar et al., 2018). UARL was calculated using the following equations:

$${\text{UARL }}\left( {{\text{l}}/{\text{day}}} \right) \, = \, \left( {18{\text{Lm }} + \, 0.8*{\text{ Nc }} + \, 25 \, *{\text{Lp}}} \right) \, *P$$
(4)

where Lm is mains length (km), Nc is the number of service connections, Lp is the total length of private pipe, property boundary to customer meter (km), P is average pressure (m).

At the town level, total water loss was determined and expressed as a percentage of NRW, based on the number of connections and pipe lengths:

$${\text{Total Water Loss }}\left( \% \right) \, = \frac{{{\text{Total Production}} - {\text{Total Consumption}}}}{{\text{Total Production}}} *{1}00$$
(5)

Population forecasting

Population estimates are crucial in determining the existing and future water demand in the distribution system (Mekonnen 2018). To forecast the existing and future population of the study town, as well as to understand the key elements that determine population distribution, size, migration, and growth rate a number of factors were considered. Births, deaths, and population migration are the three major elements that influence population growth rates. Population forecasting is critical for a town to install infrastructures such as water supply, road access, and the establishment of a school and health center, among other things to fulfill economic development (Amorocho et al. 2019; Farmani et al. 2017). For this study, the geometric increasing method was used to forecast the current and future population of the town:

$$P_{n} = \, P_{o} *(1 + G)^{n}$$
(6)

where Pn is future population, Po is present population, n is per year or decade, and geometric increasing.

Analysis of water demand

Water demand analysis depends on three major factors. These are supply management, demand management, and supply–demand management (Abebe 2020; Enbeyle et al. 2022). The following techniques and steps were necessary to determine the amounts of present and future water demand at each junction in the town's distribution systems. These are required to forecast the total population in the study town, identification of the number of houses around each supply node, assign the number of people in each supply node, evaluate average daily water demand, and assign base water demand in each supply node. Forecasting the population up to the end of the design periods of the study area was the most significant components of all calculation. The next most important is delineating the study area which was achieved using topographical maps. From which a town water distribution network map using Arc-GIS and WaterGEMS was generated, respectively. This was then converted to an Arc-GIS shapefile and overlaid on the town's topography map. The number of houses near each node was then physically counted using the town’s overlapping map and used to determine the average number of people around each junction. A deterministic water demand estimating method was used to determine the town's average water consumption (Alemu and Dioha 2020; Awe et al. 2019). However, the town's per capita water consumption was calculated using annual water consumption data and the expected total population at the time of the study. Average daily per capita water consumption was calculated by dividing the annual consumption per total number of population towns and 365 days (Milkecha and Itefa 2020). The total consumption of water was calculated combinating the metered and estimated consumption for all public purposes. Therefore, metered and non-metered water consumption is important when compared with the total water production of the town. After calculating the average daily water demand of the system, the base water demand of the specific supply node of the distribution network was then calculated (Kilinç et al. 2018). Therefore, to calculate the current average number of people in each household, the average water demand (AWD) and base water demand (BWD) were as follows:

$${\text{No of people for a supply node }} = {\text{ No of houses assigned by that node }}*{\text{ No average person in each house}}$$
(7)
$${\text{AWD }} = {\text{ per capital water consumption }}*{\text{ Total numbers of the population of the town}}$$
(8)
$${\text{BWD for a supply node}} = \frac{{\text{Population served by that node}}}{{\text{Total population of the town}}}*{\text{Avg}}.{\text{daily water consumption}}$$
(9)

Analysis of water distribution networks

The hydraulic performances of water distribution networks were analyzed by using a computer software package called WaterGMES. The input data of software are size, type, length, and age of pipes, together with the global positioning system coordinates of reservoirs and junctions. The water model software’s output reflected critical hydraulic components such as pressure heads at junctions, head losses, velocities, and pipe flow rates. The pressure losses were calculated using the Hazen–Williams equation. This software was used to analyze and improve the existing water distribution system of the town for future times.

Calibration and validation of model

The estimated model parameters and field observations were not always the same. A variety of sample locations were used for the calibration and validation of pressures, flow, and tank levels in the study area's distribution systems (D’Ercole et al. 2018). The sample for calibration and validation was collected from a place of direct joining to the water mains, closer to the supply main junction, pipes at the homes outlet, and higher and lower elevations of distribution nodes. Validation was utilized to test the calibrated model with the self-determining dataset without the restriction of modifications (Milkecha and Itefa 2020). So, in this study, the model data quality analysis was done by calibrating and validating pressure, velocity, tank level, and flow link during real and simulated average daily demand. For model calibration and validation, the water distribution sampling size and techniques followed the standard methods of water distribution sample size and techniques recommended by (Mesalie et al. 2021).

Evaluation of model performance

Model performance evaluation for calibration and validation: there are numerous evaluation procedures that can be used to calibrate and validate pressure and flow performance. All compare the pressure and flow of the model simulation and calibration between the measured data to the simulated data (Hajibabaei et al. 2019). For this study, the performance of the developed models was evaluated using four statistical indices: root mean square error (RMSE), mean absolute error (MAE), correlation coefficient (R2), and nash–sutcliffe efficiency (NSE), which were selected to help us to select the best model with the highest accuracy and minimal error. The performance criteria were calculated as shown below (Nabavi-Pelesaraei et al. 2018; Roy et al. 2018). The RMSE and MAE are two standard metrics used in model evaluation, and the variety of RMSE and MAE that lie between 0 is a perfect fit and + ∞. Based on a rule of thumb, it can be said that RMSE values between 0.2 and 0.5 show that the model can relatively predict the data accurately. R2 magnitude is a linear relationship between the experimental and the simulated standards. Where R2 ranges from − 1 that indicates the poor model to 1 which indicates the model is good, with advanced values representing less error variance. R2 greater than 0.5 is considered acceptable.

Nash Sutcliffe coefficient of efficiency (NSE) computes the standardized comparative magnitude of residual variance in contrast with the measured data variance (Zaman et al. 2021). The variety of NSE that lies between 1 is a perfect fit and  − ∞. Therefore, RMSE, MAE, R2, and NSE were used to assess the model’s performance. This indicates how much of the variance in measured data is explained by the model. The size of the linear relationship between the experimental and simulated standards was the determining factor. Model simulations are generally worse for shorter time phases than for longer time phases. The value of RMSE, MAE, R2, and NSE is calculated by using (Eqs. 10, 11, 12, and 13).

$${\text{RMSE }} = \sqrt {\mathop \sum \limits_{i = 1}^{n} \frac{{\left( {X - Y} \right)^{2} }}{n}}$$
(10)
$${\text{MAE}} = \mathop \sum \limits_{i = 1}^{n} \frac{{\left( {X - Y} \right)}}{n}$$
(11)
$$R^{{2}} = \frac{{\sum ({\text{Xi}} - {\text{Xav}})^{2} { }({\text{Yi}} - {\text{Yav}})^{2} }}{{\sqrt {\sum ({\text{Xi}} - {\text{Xav}})^{2} } { }\sqrt {\sum ({\text{Yi}} - {\text{Yav}})^{2} } }}$$
(12)
$${\text{NSE }} = { 1 } - \frac{{\sum ({\text{Xi}} - {\text{Xi}})^{2} }}{{\sum {\text{Xi}} - {\text{Xav}})^{2} }}$$
(13)

where, Xi is the measured value (m3/s), Xav is the average measured value (m3/s), Yi is a simulated value (m3/s), and Yav is the average simulated value (m3/s).

Result and discussion

Analysis of potable water supply coverage

The water quantity, hydraulic performance of distribution systems and people’s ability to pay into the system was used to evaluate potable water supply coverage. The average daily per capita consumption of the town was calculated using population statistics the amount of water produced, consumed, and lost by the town. The number of domestic connections per household has also been used to assess the degree of bonding.

Average daily per capita consumption

The supply coverage of distribution systems depends on the capacity of water consumption for different domestic purposes. Per capita, water is one of the most important factors to characterize the water supply coverage of a town. According to Beyene (2020), a minimum quantity of 30 l/c/day domestic water supply was categorized as a basic level of service. On the other hand, according to Geleta and Fufa (2021), 28 l/c/day is taken as a basic need for average daily consumption. The average daily per capita consumption of this study calculated depended on the total projected water consumption and existing population, which was 157,297 m3/d and 22,100, respectively. From this, per capita consumption in the study town was calculated as follows:

$${\text{Percapitaconsumption}}(\left( {{\text{l}}/{\text{person}}} \right)/{\text{day}}) = \frac{{157,297{\text{m}}^{3} {*}1000{\text{L}}}}{{22100{*}365}} = {19}.{\text{5l}}/{\text{c}}/{\text{d}}$$

According to Ministry of Water, Irrigation and Energy (2016), the minimum per capita water demand for the fourth category of town with a population of 20,000 to 50,000 people was 50 L per capita per day. But the existing per capital demand of this town was only 19.5 l/c/d. This value showed that the current water delivery system was very poor compared to with standard Ethiopia Ministry of Water, Irrigation, and Energy and some literature on per capita demands.

Level of connection per family

Evaluates the level of potable water coverage of the town; which is also influenced by water loss of the distribution systems. The total number of billed data water meters was 2220, according to Addis Kidam Town water supply and sewerage offices; among those, 1835 were household customers and 385 non-domestic customers, which had an average number of connections per family.

$${\text{Connection per family }} = \frac{2,220}{{22100/4.5}} = \, 0.{452}$$

As a result, the water supply coverage per family in the study town was 45.2 percent. This study showed that on average, more than half of the town's families use water from the tap, indicating a significant imbalance between water supply and demand. According to Abduro and Sreenivasu (2020), the maximum level of domestic water supply connection per family required is 100%, which means one connection for one family. The connection per family in Addis Kidam town is significantly lower.

Analysis of water loss

One of the most serious problems facing water utilities is water loss in the distribution systems. There was a lot of water loss in the distribution system, which cause high challenges to meet the daily water demands of the town. The majority of the water loss is due to old networks and leaking pipes. The values of water production, water consumption, and water loss in this study are shown in (Fig. 3).

Fig. 3
figure 3

Water production, consumption, and water loss for years 2015–2020

Full size image

According to Fig. 3, a high amount of water production is lost every year. The average water loss in the study town was 37.9% and supply only 62.1% of consumers. According to Jilo (2020), the Ethiopian government anticipates reducing urban NRW to less than 20% of overall water production is required, but more study results showed that it is likely to remain above 20%. According to Milkecha and Itefa (2020), the average amount of water loss was 40% and only 60% of water actually reached the consumers. This suggested that the town's water supply system was in poor condition, and the water utility is losing a significant amount of money each year. This amount of water loss is due to excessive pressure from distribution and transmission pipes, poor working and use of non-standard fittings, pipes, valves, and a lack of immediate maintenance and operation during pipe breakdowns. The average water rate in the study town was 7 birr/m3 (birr is the basic monetary unit in Ethiopia). While the water supply offices lost 391,729.4 birrs per year.

Unavoidable average real losses (UARL): It is suggested to calculate the UARL in liters/day.

$$\begin{aligned} {\text{UARL }}\left( {{\text{l}}/{\text{d}}} \right) & = \, \left( {{\text{18Lm }} + 0.{8}*{\text{Nc }} + {25}*{\text{Lp}}} \right) \, * \, P \\ {\text{UARL}} & = \, \left( {{18}*{13 } + \, 0.{8 }*{2}0{35 } + { 25 }*{125}} \right) \, *{32}.{5 } = {162}0{77}.{\text{5 l}}/{\text{day }} = {59158}.{\text{3m}}^{{3}} /{\text{year}} \\ \end{aligned}$$

Total apparent water losses plus avoidable real loss in the system is estimated from the own water balance as; Apparent loss plus avoidable real loss = Total water loss − UARL = (85,663 − 59,158.3) m3/year = 26,504.7 m3/year.

Population forecasting

Forecasting the town’s current and future population is the most significant step in determining the town’s current and future water consumption based on current population numbers. The geometric increasing method was used in this study to forecast the current and future population of the town throughout a 20-year design period (2022–2042) (Table 3).

$$P_{n} = \, P_{o} *({1} + G)^{n}$$
Table 3 Projected population of study town
Full size table

At the start of the design period in 2022 and the end of the design period in 2042, the town's total projected population was 25,854 and 45,640, respectively. According to the findings of this study, the population is rapidly expanding every year without significant gaps. As a result, more daily water consumption is required to meet the town’s various household activities. Because the town’s existing water supply and distribution systems are insufficient to meet the current population’s needs. This is further compounded with additional people coming from both different suburban and rural areas to find better jobs, better living standards, education, health services, and different infrastructures, because this town is a capital town of Fegita Lekoma Worda in Awi zone.

Analysis of water demand

The term "water demand analysis" refers to the assessment of all types of water required in the one town (Gajbhiye et al. 2017). There are four types of mode services that make up the total demand for household water. These are house, yard, a public fountain, and none connections (Mesalie et al. 2021; Yang et al. 2020). Water use differs from town to town due to a variety of factors. The sanitation system, the size of the town, population numbers, types of water supply, amount and quality of water sources, climatic circumstances, socio-economic conditions, and water pressure in the town's distribution system should also be considered (Wolde et al. 2020).

The projected population of each mode of service: was calculated depending on the total projection of the town population and per-capital water demands of each mode of service of the town (Table 4).

$${\text{Projected Population }} = {\text{ Percentage of Population by Mode of Service}}*{\text{ Total No of Populations}}$$
Table 4 Projected population by mode of service
Full size table

Projected Daily Average Domestic Demand (PADDD): is the production of a projected population of each mode service and projected per-capital water demand of each mode service (Table 5).

$${\text{PADDD }}\left( {{\text{l}}/{\text{d}}} \right) \, = {\text{projected population}}*{\text{projected per capital water demand}}$$
Table 5 Projected daily average domestic demand
Full size table

Adjusted Domestic Water Demand—projected daily average domestic water demand is adjusted by climatic and socio-economic conditions of the study town (Table 6). The socio-economic factor of the study town is 1 as the town was under normal Ethiopian conditions and the climatic factors of the study town are 0.9 (Ahmed 2022; Kuma and Abate 2021).

$${\text{Adjusted daily average DD }} = {\text{ daily average DD *Socio}}\, - \,{\kern 1pt} {\text{economic factor*climatic factor}}$$
Table 6 Adjusted domestic water demands of study town
Full size table

This study was divided into two phases, which were the first and second phases. Then, total domestic water demand of the first phase and second phase was 1,593.94m3/d and 1,833.5m3/d, respectively. Divided the design period has different advantages during the construction and implementation of different infrastructures of water distribution systems. Because the construction and implementation of any infrastructure of water distribution systems of the first phase depending on the first phase water demand, while the construction and implementation of the second phase of any infrastructure were contracted and implemented at the end of the first phase years (2032) depending on the second phase water demand, because this demand was forecasted considering future expansion of the town into four directions (North, South, East, and West) of the town, but this expansion occurs at the end of the first phase. Therefore, the implementation of the water distribution infrastructure starting period (2022 year) up to the end of the design period (2042 year) once, it is not good, then construct and implementation of the first phase of the infrastructure of distribution systems at once and also implement the second phase of infrastructures starting from at the end of the first phase depends on expansion directions of the town or town populations. Then, the town water supply and sewerage office were used for the second phase money for the infrastructure of distribution systems for other projects to gate additional money up to the end of the first phase (for 10 years).

According to Table 7, the first phase of maximum daily demand and peak hourly demand was 7,122.4m3/d and 13,532.6m3/d, respectively, and the second phase of maximum daily demand and peak hourly demand was 9,683.8m3/d and 18,399.3m3/d, respectively. Regarded to above (Table 7), the per capital water demand during the first and second phases was 82.4 and 112 l/c/d, respectively. According to Gebrehiyot (2015), the quantity of water demand forecasted was 65 l/c/s. On the other hand, the amount of water demand predicted, according to Ahmed (2022), was 45–75 l/c/s. The result of this study indicated that the estimated water demand satisfied the town's current and future per capital demand. Therefore, the estimates of water demand are critical for the design, planning, implementation, and hydraulic performance of water distribution systems in order to ensure water availability and fulfill the demand for water design for a long-term and cost-effective water supply scheme.

Table 7 Summarized water demand assessments of the study area
Full size table

Analysis of water distribution network

The simulated model results were used to determine the high and low pressures and velocities at all junctions and pipes in the distribution network (Beyene 2020; Jilo 2020). Pressure and velocity in water distribution systems are the most important factor in providing appropriate and safe water for the whole populations in the study areas, with the same quantity at the same time. According to AL-Washali et al. (2018) and Hunde and Itef (2020), the maximum and minimum pressures and velocity in the water distribution network during average hourly flow ranged between 15 to 70 m and 0.6 to 2.5 m/s, respectively. Analysis of pressure and velocity in the distribution system of the study town is showed in (Table 8 and 9) during peak and minimum hourly demand.

Table 8 Distribution of pressure during peak daily and minimum hourly demand
Full size table
Table 9 Actual velocity of distribution system during peak hourly and minimum hourly flow
Full size table

According to WaterGEMS' hydraulic analysis, the optimum operating pressure head range was only obtained at 73.4% of the study area, with 26.6% of the identified nodes having a pressure head less than 15 m during peak hourly demand. WaterGEMS hydraulic analysis during minimum hourly demand identified that the optimum operating pressure head range was obtained at 81.5% of the study area. But 4.3% of the identified nodes had a pressure head less than 15 m, while 14.2% of a pressure head more than 70 m. The network's conditions were analyzed over a time that corresponded to the peak hourly and minimum hourly demand. According to Agunwamba et al. (2018) and Gökçekuş et al. (2019), low-pressure head systems have a direct impact on user satisfaction due to the reduction in water quantity delivered, as well as making the system more prone to negative pressure heads and pollutant incursion. According to Farmani et al. (2017), the pipe bursts and increased energy demand are caused by a high-pressure head in distribution networks. Therefore, the results of this study indicated that 26.6% and 14.2% of a pressure head, which have less and more than 70 m during peak and minimum hourly demand, respectively, cause a reduction in water quantity delivered and pipe bursts and increased energy demand in distribution networks of town. Maintaining the correct pressure head in the distribution network is accordingly critical for optimal operation.

The results of waterGEMS’ analysis of velocity in the distribution network during peak hourly demand found that 87.6% of identified pipes had the optimum operating velocity range within the study area, but 12.4% of the identified pipes had a velocity less than permissible limits. And also during minimum hourly demand, 65% of identified pipes had the optimum operating velocity range of the study area. But 29.9% of the identified pipes had a velocity less than 0.5 m/s, while 5.1% had a velocity more than 2.5 m/s. According to Adedoja et al. (2018) and Enbeyle et al. (2022), velocity in the distribution systems less than 0.5 m/s and greater than 2.5 m/s will lead to the deposition of sediment in the pipes and bursting the pipes in distribution networks, respectively. Therefore, adjusting the minimum and maximum velocity compared with the permissible velocity ranges (average 0.5–2.5 m/s) in the distribution networks was used to avoid the deposition of sediment and bursting of pipes.

Evaluation of pressures, flows and tank level in both the model calibration and validation

According to Milkecha and Itefa (2020), a model is required to be validated to ensure that its output accurately corresponds to values observed in the field. Therefore, a model needs to be calibrated in order to have confidence in its results. Similarly, Tufa and Abate (2022) also note that a phase in the calibration process called model validation makes ensuring that the calibrated model is properly assessed. The evaluation of the model performance of calibration and validation of pressures, flow, and tank level was accordingly completed with the results shown in (Figs. 4, 5, 6, 7, 8, and 9).

Fig. 4
figure 4

Calibration model of measured and simulated pressures at average daily demand

Full size image
Fig. 5
figure 5

Validation model of measured and simulated pressures at average daily demand

Full size image
Fig. 6
figure 6

Calibration of pipe flow during average daily demand at pipe No.86

Full size image
Fig. 7
figure 7

Validation of pipe flow during average daily demand at pipe No.86

Full size image
Fig. 8
figure 8

Calibration of Tank level during average daily demand

Full size image
Fig. 9
figure 9

Validation of Tank level during average daily demand

Full size image

The statistical correlation between the measured and simulated pressure during calibration and validation, from Figs. 4 and 5, recorded a R2 is 0.96 and 0.98, respectively, which is an acceptable limit. As a result, the calibration and validation of measured and simulated pressure values are done using the required coefficient of determination standards.

A high correlation of R2 of 0.96 and 0.97 was also obtained during calibration and validation at average daily flow, respectively, from Figs. 6 and 7, which provides a good agreement between measured and simulated values of pipe flow.

The values of R2 during calibration and validation at average daily flow were 0.85 and 0.91, respectively, Figs. 8 and 9, which demonstrated good agreement between measured and simulated values of tank level.

Different studies that were conducted in different towns show the different values of the model in both calibration and validations of measured and simulated pressure, flow and tank levels. According to Ahmed’s (2022), study result of the model showed a good match between measured and simulated pressure, flow, and tank level of Jigjiga Town Ethiopia, both in calibration and validation periods with (RMSE, MAE, R2, and ENS = 0.68, 0.37,0.98 and 0.85) and (RMSE, MAE, R2 and ENS = 0.65, 0.42, 0.91 and 0.82), respectively. And also, according to Geleta and Fufa's (2021), study model of performance indicated that a good agreement between simulated and measured pressure, flow, and tank level of Alibo Town Ethiopia, both in calibration and validation periods with (RMSE, MAE, R2, and ENS = 0.58, 0.45,0.90 and 0.78) and (RMSE, MAE, R2 and ENS = 0.58, 0.34, 0.85 and 0.76), respectively. According to Table 10, the results of this study indicated that a good agreement between measured and simulated values of pressure, flow and tank level during both in calibration and validation period with (RMSE, MAE, R2 and ENS = 0.65, 0.40, 0.96 and 0.82), (RMSE, MAE, R2 and ENS = 0.62, 0.45, 0.96 and 0.68) and (RMSE, MAE, R2 and ENS = 0.52, 0.42, 0.85 and 0.75), respectively. In general, the model performance assessment showed a good correlation and agreement between the average daily demands measured and simulated of pressure, flows, and tank level of the distribution network. This indicates that waterGEMS can give a sufficiently reasonable result for Addis Kidam Town of performance model of the distribution system.

Table 10 Performance evaluation of pressure, flow, and tank level in both the model calibration and validation
Full size table

Management strategies of distribution networks

The water condition in the town of Addis Kidam, Ethiopia, as described and assessed in this study, was at an unsatisfactory level to meet basic human needs for reasonable hygiene situations. Because the average demands of water into peoples homes were less than 19.5 l/c/d to around 25,854 people in the town. According to Tufa and Abate (2022), the per capita domestic water consumption of the town is required to be 43 l/c/d to around 24,500 people. According to WHO (2016), the average per capital demand was 50 l/c/d. Therefore, compared with some literatures and standard of Ethiopia and WHO guidelines of per capital water demand of the town is lower than required.

The scope of this study does not include developing a practical water management plan. Nevertheless, based on the data gathered, the following basic principles are recommended to be combined into a thorough management plan that should be implemented immediately. First and foremost, a two-phase improvement strategy should be considered in light of the country’s socio-economic realities. The first phase should target the year 2032, while the second phase should target the year 2042. The plan should primarily focus on upgrading and expanding the water distribution network, and supplying the water above than 50 l/c/d.

Upgrading and extension of the water distribution network

The main goals of the improvement plan are to expand HHCs to at least 12% and 25% of the population by 2032 and 2042, respectively. In the first phase, public tap connections would be reduced to 10% of the population, with this kind of water supply becoming phased out during 2042. This would include expanding the water distribution system and upgrading the piping efficiency as appropriate in order to minimize water losses.

The main goal for the first phase would be to raise the unit water utilization rate over 30 l/c/d in order to ensure a minimal degree of hygiene for public tap users, as well as a progressively increase water supply up to 60 l/c/d. The required water supply in the second phase should fulfill a higher unit consumption requirement of 30–80 l/c/d, which corresponds to a daily water supply of 9,683.8m3/d.

Conclusion

The main objective of this study was to evaluate and analyze the water demand and forecast future water demand together with the hydraulic performance of the water distribution network of Addis Kidam a town in, Ethiopia. This was done using a statistical analysis and the WaterGEMS V8i software. It identified that the existing distribution network could not supply water demand for the current population of the town. Nor could it meet the hydraulic performance of pressure and flow velocity and recommended alternative methods and models for improving existing water demand for future times. The existing water distribution network faces numerous of challenges related to water demand, supply, and hydraulic performances of pressure and velocity. The results of this study indicated that the existing water distribution network of Addis Kidam Town was designed to meet a population of 8,906; however, the current population is 25,854. Hence, the supply cannot meet the required demand. The average domestic water demand and level of connections per family were 19.5 l/c/d and 45.2%, respectively. Water loss of the distribution network was 37.5%, which was very high, and the supply network only covering 62.1% of the town. Simulation of existing distribution network at junctions and pipes has 26.6% and 4.3% and 2.4% and 29.9% of lower pressures and velocities during peak and minimum hourly demand, respectively. Model performance of distribution networks was evaluated by calibration and validation of pressure, flow, and tank level using RMSE, MAE, R2, and NSE. The values of RMSE, MAE, R2, and NSE during calibration and validation of pressure, flow, and tank level were 0.65, 0.40, 0.96, and 0.82 and 0.56, 0.38, 0.98, and 0.78, respectively. The research recommends a two-phase strategic water distribution system response beginning by upgrading and expanding the water distribution network, to first achieve a supply of 30 l/c/d by 2032, and then lifting this to the 30–80 l/c/d range before 2042. The proposed water management upgrading approach is expected to establish a good transferable water supply design process that could be used for all residential communities in emerging countries facing comparable challenges. In general, this study concluded that the existing water demand and hydraulic performance of Addis Kidam Town in Ethiopia were inadequate for various types of water demand and supply for current and future times. Therefore, this study recommended that the existing water distribution systems needs to be improved incorporating current and future numbers of the population, water demands, and hydraulic performances.