Assessment of Hydropower Potential in Wastewater Systems and Application in a Lowland Country, Lithuania

Abstract

This paper focuses on possible power generation by micro-hydro turbines integrated into lowland wastewater systems, which convert the potential energy of effluents in pipes into electric power. While other European countries have widely invested in this technology, Lithuania and other Baltic countries are still behind with their potential development rate. A search for potential micro-hydro sites was carried out, and a methodology for assessing water resources for an ungauged wastewater network is proposed herein. Particularities of wastewater flow patterns are briefly reviewed, and turbine operational constraints are analyzed. The hydro turbines available on the market to be installed in wastewater systems that meet lowland conditions are discussed. Available tools on the hydropower market to conduct a preliminary assessment of potential sites for urban water networks are considered. Multicriteria analysis is performed to select optimal projects by assessing the relevant economic, technical, and environmental criteria in water networks. The outcomes of this study can be used for unlocking the hydropower potential of wastewater systems in low-lying areas.

1. Introduction

Urban water management is known to be very energy-intensive. Therefore, engineers and researchers worldwide are looking for ways to recover the energy residing in water networks supplying drinking water or collecting sewage water and their treatment plants [1,2,3,4]. The untapped potential of small (1–10 MW), mini (100 kW–1 MW), and micro (5–100 kW) hydropower systems in engineered water conduits has largely remained unexplored, and this alternative energy source is receiving more attention from regulators in a number of countries.

The EU Horizon Europe Framework Program launched a call for the “development of hydropower equipment for hidden hydropower”, dealing with hydropower generation in water infrastructure and focusing on low-head schemes [5]. For example, this area of research certainly encompasses gravity-fed municipality wastewater networks or excessive pressure installations (e.g., pressure-reducing valves or break pressure tanks) in drinking water networks.

Unlike conventional hydropower systems operating in rivers, hidden or in-conduit hydropower systems have a minimal environmental impact as these systems are fully integrated into existing infrastructure [6]. These hydropower systems specifically harness the excess energy of water being used for a purpose other than energy generation.

The EU is fostering the exploration of the potential of micro to small hydropower energy recovery in municipal water systems. A project investigated the technical and economic feasibility of energy recovery from the municipal water networks through the installation of micro-hydro (MHP) turbines and established a GIS (geographic information system) database [7]. A recently completed project developed an institutional, social, and technological environment to increase resource efficiency in water networks by installing micro-hydropower technology [8]. An ongoing project has established a geodatabase of untapped potential energy recovery locations at existing hydro installations in selected European cities and countries [9,10]. Considering their outcomes, public water systems and publicly owned treatment works, including wastewater facilities, could benefit from installing an in-conduit hydropower system.

A comprehensive overview of innovations in in-conduit hydropower technologies and their applications was given in [11,12]. Real case studies with a brief technical description of small hydro turbines integrated into water infrastructure, including drinking water and wastewater networks, were presented in [6]. Assessments of hydropower potential, including the technical and economic viability of installing turbines in water and wastewater infrastructure, have been conducted in a number of countries [1,2,13,14,15,16,17].

The literature analysis showed that, in contrast to wastewater systems, there are significantly more available studies on energy recovery in raw or potable water distribution systems. Moreover, for the latter, hydropower potential can be relatively quickly assessed. Yet, there are significantly more studies on energy recovery from wastewater treatment plants (WWTPs) than from wastewater collection networks (upstream of WWTPs). Most studies have dealt with the installation of hydro turbines at the outfalls of WWTPs (downstream), where the water has already been treated. However, challenges appear upstream of WWTPs where raw wastewater quality and clogging of turbine runners can compromise normal operation. Experience from sewage pumps can be beneficial in addressing this hot topic.

Assessment of wastewater resources for hydropower generation is a crucial task. It is essential to know the amount of wastewater, as well as its flow rate and distribution over time. Flow rate frequency analysis must be performed to construct the flow duration curve (FDC), which is a key element for estimating hydropower energy [18,19]. The hydrological aspects of wastewater are not discussed here in detail. However, there are apparent differences between runoff formation in water networks (anthropogenic pattern) and natural streams or rivers. Sewage accounting is most accurately conducted at water treatment plants. However, not all WWTPs have flow meters installed. A big issue is accounting for wastewater flow in the collection network, i.e., upstream of the WWTP. A methodology is needed for the construction of the FDC for ungauged sites. In Switzerland, typical FDCs have been proposed for WWTPs [2].

The impact of variation in sewage flow due to storm events and changes in water demands on the optimal design of turbines was investigated [15]. Flow in WWTPs varies with daily water demand patterns, by season and year, and is subject to stormwater events depending on geographical location. Storm events are a random process; therefore, there is a need to examine wastewater networks and how their flow responds to these events at a particular site. Moreover, the effect of storm events is mostly considered at WWTPs, not in the wastewater network. Therefore, there is no firm response regarding the opportunities to account for hydropower generation from stormwater.

A wastewater treatment outlet is generally appropriate for installing turbines due to the sufficient and constant water flow. The parameters required for selecting hydro turbines, such as head and flow, are monitored continuously as part of the WWTP process. Therefore, the turbine’s performance can be relatively easy to monitor [12]. However, there might be a pitfall for low- or ultralow-head plants if the so-called tailwater effect is not considered. At most sites, depending on the receiving water body (e.g., river), during a flood period, the tailwater level at the outfall rises more than the level upstream of the intake and causes a reduction in the head. However, these cases are rarely reflected in the literature.

A comprehensive overview of turbines applicable to in-conduit and hidden hydro, especially highlighting recent advances in turbine technology, was given in [11,20,21]. Innovative technological solutions that improve conventional turbines with robust designs, better efficiency, and possibly lower cost were proposed [12]. At the same time, although newer or emerging technologies offer innovative approaches to in-conduit hydro generation, they are not always the most cost-effective choice [22]. The equipment cost comparison is complicated due to various sites and turbines. Yet, even if construction and installation costs are cheaper for modular systems, their hydromechanical and electric operating costs can be more expensive than those of conventional turbines [12,23].

Because conventional hydro turbine technologies are not always competitive in the market, low-cost generators (e.g., pumps as turbines (PaTs)) are suggested. They are standard pumps utilized as turbines by reversing the flow direction across them. Research on PaTs and their application to small and micro-hydropower has been going on for about 100 years and is still relevant today [24,25,26,27]. PaTs are usually used at sites with higher heads; the experience of low-head application is seldom covered in the literature.

The availability of water pumps and their attractive cost for conventional hydro turbines make them a perfect technology for exploiting the untapped in-conduit hydro potential that can be technically feasible but not financially viable. However, PaTs have the shortcoming of a relatively low efficiency, which can be further reduced with high flow fluctuations, as in the case of the Francis turbine.

Table 1. Major differences between turbines and PaTs.

	Turbine	PaT
Advantages	Well-documented, accurate design	Cost-efficient
	Best efficiency	Widely available
	Wide range of control	Standardized, simple design product, short delivery, and low maintenance and repair costs
Disadvantages	Expensive	Not as well-documented as turbines, limited availability of turbine operation curves 1
	Limited local suppliers	Lower efficiency 1
	Complex design may be required	No variable guide vanes for varying flow

¹ Some large turbines or pump producers offer PaTs with high efficiencies (up to 87%), along with their operating ranges and guaranteed hydraulic characteristic data from prototype tests [28,29].

shows some major differences between PaT systems and conventional turbines.

The lack of in-depth studies dedicated to the impact of wastewater quality on hydro turbines, particularly the risk of clogging them in sewage networks upstream of WWTPs, can be a severe problem. In contrast, in the field of pumps, this critical issue is being addressed [30,31]. There is no significant difference between the operating modes of these hydraulic machines. Therefore, the experience in pumps can be straightforwardly applied to turbines. This is true, bearing in mind that the number of PaTs installed in water networks is increasing.

Tools to support water and wastewater utilities in assessing the technical and economic feasibility of installing hydropower facilities are of utmost importance. Because most in-conduit or hidden hydro systems are relatively small in capacity and, therefore, necessitate a relatively expensive feasibility analysis, assessment tools need to be as user-friendly and cost-effective as possible. Existing tools do not meet these requirements for small developers [32]. A few engineering design tools have been proposed to evaluate conduit projects, though they have not yet been applied to broader assessments. However, some exceptions can be attributed to the US-developed tools [12,33]. These tools operate in readily available spreadsheet software and can be accessed for free.

The literature review revealed no specific tool at the European level for even preliminary assessments for hydropower recovery in water networks. However, a slight exception can be made for a tool intended for water suppliers to perform the basic design of a site [34]. Some reviewed publications mentioned using standard Microsoft Excel calculations to complete a preliminary evaluation of water network sites; however, they are not openly available to hydropower designers.

There are several available assessment methods or tools for conventional small hydropower projects [35,36]. For example, a clean energy management software system, RETScreen Expert, is widely used [37,38]. It is the most comprehensive tool, allowing designers to model and analyze any renewable energy project to complete its pre- or feasibility analysis.

Water and energy management requires considering a variety of often conflicting factors. Multicriteria/multi-attribute analysis (MCA) is often a core part of this challenge. Many publications used MCA in the siting of RES facilities, including conventional hydropower generation sites [39,40,41]. However, there are not many publications on the application of MCA in hidden hydro networks, particularly municipal water networks. A turbine was selected at the outfall of WWTPs by assessing economic, technical, and environmental factors [23].

Researchers and practitioners largely use some five to seven MCA methods [42]. The paired comparison and outranking methods are popular in explicitly evaluating multiple conflicting decision-making criteria. The most known outranking method is ELECTRE [43,44]. To facilitate MCA practical application, software is needed. Several programs related to MCA exist on the market, categorized as free, semicommercial, and commercial [45].

Lowland or low-lying areas of a country represent challenges for installing hydro turbines in urban water networks compared with regions with a steep topography surrounded by mountains. The accuracy and completeness of the hydraulic data are crucial, especially when evaluating low- and ultralow-head sites. Ultralow head is defined for sites with a hydraulic head of less than 2–3 m [18,46]. Some authors even offered a feasibility threshold for such plants operating in urban areas: a minimum head of 4.5 m with a flow of 0.05 m³/s [16].

According to our survey, no operating hydropower turbines were detected in the water supply or sewage collection network in the Baltic states, including Lithuania. The main cause of this is a lack of awareness of such technology. Moreover, the renewable energy (RE) law of the Republic of Lithuania does not consider such technology; consequently, there is no kind of incentive. In contrast, some countries have implemented government-funded financial incentives for RE generation from gravity-fed systems, thus encouraging water companies to implement MHP schemes [13].

Despite the good practices reviewed above in a number of countries, a challenge remains during the planning phase of hydropower development, as the lack of access to available WSW data limits the quantification of water power potential, and thus the identification of potential sites. The lack of methodologies dedicated to quantifying hydropower parameters using limited data is prevalent. In an effort to address this problem, a methodology was proposed to quantify potential and identify in-conduit hydropower sites in the wastewater systems of lowland countries. This study is the first attempt to encourage water utilities and private owners to unlock the hydropower potential in their service areas. The overall aim was to evaluate the hydropower potential in urban networks in low-lying areas and propose methodologies for its practical use.

The specific objectives of this study were as follows:

• To review the available best practices of energy recovery in wastewater systems and identify methodology on the basis of local conditions;
• To search for the potential sites for the installation of hydro turbines and to evaluate wastewater resources for harnessing hydropower;
• To review and propose tools to facilitate preliminary and/or feasibility analysis of hydro schemes and to review turbines and their installation layouts in wastewater systems;
• To show best practice in performing multicriteria analysis for the selection of optimal sites.

2. Materials and Methods

2.1. Study Area

Lithuania is a low-lying country; therefore, the country’s SHPs operating in run-of-the-river mode mostly feature a low or medium head (up to 15 m). The same picture is valid for the remaining Baltic countries [47]. Due to the country’s topographic conditions, mostly sewage (wastewater) networks with free gravitational flow are attractive for harvesting water energy. Drinking water is extracted from a deep aquifer, and its distribution systems are artificially pressurized and can seldom be used for energy recovery.

The study area was the urban water network with potential micro-hydro sites deployed in Lithuania. More than 20 potential sites for installing hydropower turbines were identified (Figure 1 and Figure 2). Most of them were located in the sewage network, while only one site was spotted in the drinking water distribution network with a pressure-reducing valve [48]. Their preliminary power capacity was estimated as less than 100 kW. These potential sites with preliminary key characteristics, including those operating in other countries, are freely available [9].

Two large cities, Vilnius and Kaunas, including the town of Alytus, stretch around the largest rivers, the Nemunas and the Neris. The search for suitable sites for hydropower plants showed that the most significant potential exists in sewage networks, where their collectors (mains) descend into deep river valleys. It worth highlighting that 58 WWTPs are operating in the country. However, most of them exhibited no significant elevation drop (at outlets or inside water treatment installations).

This study examined the opportunities of MHP energy recovery at three WWTPs, four wastewater collectors (upstream of WWTPs), and one site in the drinking water network (

Table 2. Potential sites for installing hydropower turbines in urban water networks of Vilnius, Kaunas, and Alytus.

#	Name	ID Label	Population Equivalent (PE)	Service Area (km2)	Location 1	Head (m)	Flow (m3/s)	Outlet
1.	Kaunas (Jonavos street)	K1	104,300	25.3	U/S	35.0	0.3	Sewage network
2.	Kaunas (Raudondvario Street)	K2	36,800	16.4	U/S	27.4	0.18	Sewage network
3.	Kaunas (Pypliai)	K3	305,500	137.0	D/S	4.0	1.2	The Nemunas River
4.	Vilnius (WWTP-1)	V1	569,500	356.0	D/S	2	1.5	Outlet collector
5.	Vilnius (WWTP-2)	V2	569,500	356.0		2.9	1.5	The Neris River
6.	Vilnius (Prusu Street)	V3	35,000	18.2	-	61.1	0.11	Network
7.	Alytus (WWTP-1)	A1	49,900	39.4	D/S	15.0	0.11	Outlet collector
8.	Alytus (WWTP-2)	A2	49,900	39.4		10	0.11	The Nemunas River

¹ Site location relative to WWTP (upstream—U/S or downstream—D/S); see Figure 3.

). The latter represents a separate case and is not detailed in this study.

There are two layouts for installing a hydro turbine in sewage networks: upstream and downstream of the operating WWTP (Figure 3).

In the first case, untreated wastewater passes through the turbine, adversely affecting its operation. The hydro turbine should be more corrosion-resistant, and the intake must be provided with a trash rack. The second case, where the turbine is behind the wastewater treatment plant, is much more acceptable—the hydro unit encounters treated wastewater.

There are fundamental differences between installing turbines in WWTPs and installing them in their outlets or sewage collection networks. This is not only related to the quality of the wastewater (treated or untreated sewage). It is not always technically possible to create pressure in the collector of the sewage network due to existing side branches connected to the main (collector). Simple energy-dissipating devices are also installed in the high-slope gravity sewer mains to reduce the flow velocity. In this case, it is difficult to build up pressure in the collector; it is necessary to reconstruct the pipeline in order to make it work as a penstock. As a result, due to the additional civil engineering work, the MHP project can hardly be considered economically viable. Moreover, the quality of the pipe material, current technical state, and duration of service can also pose a problem. Consequently, converting gravity-fed pipelines into a high-pressure hydraulic regime (penstock) can be challenging.

The layouts of water network systems, their engineering drawings, and spatial information (GIS data) available from water companies were analyzed. In contrast, the assessment of the head did not present any difficulty.

2.2. Water Resources

A big challenge was to evaluate available water resources for power generation at individual sites. Over time, reliable trends in wastewater patterns are difficult to predict in Lithuania. For instance, in the town of Alytus (population of some 50,000), the amount of wastewater treated in the last 10 years decreased twofold. The same situation can be seen in other cities in the country. This was due to demographic conditions and the technological transformation of the water use in industry. Only in the last 5 years has the volume of wastewater treated by WWTPs stabilized.

Sewage volume data were obtained in various temporal formats (months, days, hours, and even minutes) from WWTPs and water utility companies. Additionally, measurements of wastewater-level fluctuations at the outlets of WWTPs and key structures of the collection network (upstream of the WWTP) were conducted to reveal the pattern of wastewater flow. Spot measurements (from 1 to 3 months) using data loggers at a 30 min interval were also performed. The recordings obtained were transformed into volumetric discharge values.

In hydrological analysis, as a common rule, the length of flow records to construct a reliable flow duration curve (FDC) must be at least 15 years [18,49]. The FDC is a graph that shows the percentage of time that flow in a waterway is likely to equal or exceed some specified value (Figure 4). Mean daily flows used for FDC construction represent the most accurate results for designing purposes. An FDC based on mean monthly flows can only be used for a preliminary study.

To compare the FDCs developed for each site, their normalization was performed according to the mean daily flow $\overline{Q}$ [50]. The parameter Φ describing the shape of the dimensionless FDC was used (Figure 4). The area under the curve represents the volume of the flow.

To construct the dimensionless FDC, the ratio k was defined using the following formula:

$$k_i=\frac{Q_i}{\overline{Q}},$$

(1)

where $k_i$ is the normalized flow, $\overline{Q}$ is the mean daily flow (m³/s), and $Q_i$ is the daily flow at time $t_i$.

$$\overline{Q}=\frac{{{\displaystyle \sum }}_{i=1}^n{Q}_i}{n}$$

(2)

where n = 365 days. When $Q_i=\overline{Q},$ k = 1.

The Φ is the mean flow area (dashed area below the FDC) with respect to the whole FDC area (defined below the red curve).

$$\Phi ={{\displaystyle \int }}_0^{1}t{k}_I$$

(3)

It is evident that, for a steep curve, the Φ value is low, whereas for a flat curve, the Φ value is much higher. The latter is always desirable for any hydro scheme, as it indicates a relatively stable flow regime.

To account for a tailwater effect at the site of the treated wastewater outlet, daily water level records were used to construct the head-duration curve for the outlet of the WWTPs. For example, for site K-3, the recorded statistics of the Nemunas river at the Lampedziai gauging station (2014–2019) were considered.

2.3. Assessment Tools

As mentioned, methodologies or tools available to carry out hydro projects (preliminary designs or feasibility studies) are abundant. On the basis of previous experience, only the most suitable tools for water supply and WWTP infrastructure were reviewed. When reviewing tools designated for assessing the feasibility of installing hydro units in urban water networks, the differences in the interpretation of the terms software, computer package, and Excel^®-based workbook were ignored.

The following tools were tested: (1) RETScreen Expert [37,38], (2) In-Conduit Hydropower Project Screening Tool [33], and (3) Business Case Assessment Tool [12]. RETScreen Expert was identified as the most comprehensive tool for assessing the feasibility of eight potential sites.

2.4. Multicriteria Analysis (MCA) Methods

The multicriteria analysis of siting potential micro-hydro facilities in urban water networks was carried out using the software HYPSE [51]. The analysis considered a classical outranking technique, ELECTRE [43,44]. The data employed are shown in

Table 3. Data employed for multicriteria analysis for installation of hydropower turbines (the basic scenario, equal weights for all criteria).

#	Criterion	Unit of Measure	Direction	Weight (%)	Group and Weight		Alternatives (Projects)
							K1	K2	K3	V1	V2	V3	A1	A2
1.	Layout	Score: [1, 2]	Max	5.88	TEC	58.81	1	1	2	2	2	2	2	2
2.	Turbine type	Score: [1, 3]	Max	5.88	TEC		2	2	3	3	3	2	2	2
3.	Design flow	m3/s	Max	5.88	TEC		0.36	0.14	1.00	1.80	1.80	0.17	0.17	0.17
4.	Gross head	m	Max	5.88	TEC		35	27	4	2	2.9	53	15.5	10
5.	Substation	Score: [0, 1]	Max	5.88	TEC		0	0	0	1	1	0	1	1
6.	Transmission line	km	Min	5.88	TEC		0.05	0.05	0.1	0.05	0.05	0.05	0	0.4
7.	Power capacity	kW	Max	5.88	TEC		98	29	28	20	34	73	20	13
8.	Capacity factor	%	Max	5.88	TEC		43	40	76	65	64	73	39	38
9.	Tailwater effect	%	Min	5.89	TEC		0	0	25	0	20	0	0	25
10.	FDC type	Parameter	Max	5.89	TEC		0.47	0.42	0.56	0.6	0.6	0.57	0.62	0.62
11.	Total initial costs	k€	Min	5.88	Econ	29.42	101.1	61.7	76.3	50.8	80.9	47.3	18.7	14.0
12.	Electricity generated	MWh	Max	5.88	Econ		367	102	185	111	189	331	69	43
13.	Simple payback	yr	Min	5.89	Econ		4.3	15.5	5.5	7.3	5.8	1.6	4.4	6.7
14.	O&M costs	k€	Min	5.88	Econ		13.5	6.2	4.6	4.2	4.9	3.8	2.6	2.6
15.	Electricity revenue	k€	Max	5.89	Econ		36.7	10.2	18.5	11.1	18.9	33.1	6.9	4.3
16.	GHG reduction	tCO2/MWh	Max	5.88	ENV	11.77	99	28	50	30	51	89	19	12
17.	Use of electricity	Score [1, 2]	Max	5.89	ENV		1	1	1	2	2	1	2	2
Total (%)				100		100

.

It is beyond the scope of this paper to summarize all available MCA methodologies. The crux of MCA is the construction of a two-dimensional impact matrix, including alternatives or scenarios (e.g., projects or actions) and their criteria according to which the alternatives must be evaluated (Table 3). Collected field data and data generated by RETScreen Expert software were used as the input for this matrix. For the criteria, their properties were chosen to represent the prospective sites as objectively as possible, taking into account MCA experience [52]. A more detailed description of the criteria is given in

Table 4. Description of the criteria used for MCA (with reference to Table 3).

Name	Brief Description
TEC Layout type: The facility can be installed downstream of or inside the WWTP, or upstream of the WWTP in the effluent network. Turbine type: PaT, Archimedean screw, and conventional submerged (in-conduit) turbines. Design flow: Typically taken as 30% of FDC. Gross head: Drop in elevation at the site. Substation: Cost depends mainly on the voltage and the installed capacity of the power plant. Transmission line: Cost depends on the line’s type, length, voltage, and location, as well as the installed capacity of the power plant being developed. Power capacity: Calculated hydro system power capacity or maximum power output of the site. Capacity factor: Ratio of the average power produced by SHP over one year to its rated power capacity. Tailwater effect: During high flows, a reduction in the gross head can be significant for low-head sites. FDC type: The shape of the FDC indicates the distribution of daily mean flow over a sufficiently long period; initially steeply sloped curve results from an uneven flow; FDCs that have a very flat slope indicate slight variation in the flow pattern.	1. Priority should be assigned to the hydro turbine downstream of the WWTP (clean water); turbine placement upstream will require extra O&M costs (trash rack cleaning). 2. It is recommended to avoid Archimedean screw turbines because of their excessive superstructure and visual pollution, especially in urban areas. 3,4,7,8. Design flow, gross head power capacity, and capacity factor must be as high as possible. 5. Presence of any substation nearby hydro installation. 6. Distance to the electric distribution grid or the point of use of power must be as short as possible. 7,8. Must be maximized. 9. Range of water-level fluctuations in receiving water body should be minimal, i.e., to avoid any reduction in the available gross head during times of high flows in the outlet. 10. A flat-sloped FDC resulting in a high j value is desirable for any hydro scheme.
ECO 11. Total initial costs: Total incremental investment that must be made to bring the proposed case facility online before it begins to generate savings and revenue. 12. Electricity generated. 13. Simple payback: The length of time that it takes for a proposed facility to regain its own initial cost. 14. O&M: Operation and maintenance costs. 15. Electricity revenue: Calculated by multiplying the electricity supplied to the grid (or used on-site) by the electricity export rate.	11,13,14. Total initial costs, payback period, and O&M costs must be as low as possible. 12,15. The amount of electricity generated and the revenue must be maximized; priority should be assigned to local use of electricity generation for wastewater treatment, not to the grid.
ENV 16. GHG reduction: The proposed facility’s greenhouse gas emission reduction (mitigation) potential. 17. Use of electricity: It can be on-site or delivered to the distribution grid.	16. GHG reduction is to be maximized. 17. The best way is to use the produced power directly on-site at the plants.

. To rank these hydropower sites, 17 criteria were identified, grouped into three identical groups: technical (TEC), economic (ECO), and environmental (ENV) factors.

The solution to any MCA using concordance analysis must be implemented through successive steps. Since most of the criteria had different units, it was necessary to perform their normalization. The min–max normalization was used, after which a pairwise comparison of the criterion scores (from 0 to 1) was carried out.

The approach of concordance analysis involves a pairwise comparison of all projects (alternatives or scenarios) in the matrix in each criterion. From each comparison between two projects i and k, we obtain concordance and discordance indices (C_ik and D_ik, respectively) [51].

The concordance matrix and resulting dominance index (c_i) consider positive aspects of alternatives. The discordance matrix reflects the negative aspects of projects, which evaluates the differences of two alternative projects. Four discordance dominance indices are commonly used: (1) simple discordance index (d_i,SD), (2) weighted discordance index (d_i,WD), (3) aggregate discordance index (d_i,AD), and (4) weighted aggregate discordance index (d_i,AWD). When they are low, i.e., have a negative value, there is a greater dominance of project i. Thus, the projects can be ranked in decreasing order. The global synthetic index (GS_i) integrates the information on each project’s positive and negative aspects. Its score must meet the condition ∑GS_i = 0, and the best option should favor those projects with GS_i > 0. The decision-maker most commonly chooses the project or alternative with the highest average ranking.

3. Results and Discussion

3.1. The Search for Potential Sites

Unlocking available SHP potential in the water network requires a framework that allows potential developers to locate this potential. There are obviously known sites in drinking water networks with excess head or pressure. For pipe network engineering, layouts and drawings are available; thus, their identification is not difficult. The same is true for WWTPs at inlets and outlets. However, the problem stems from the wastewater collection network placed in the areas upstream of WWTPs.

A number of authors have used spatial information (GIS data) to identify potential hydro sites in water distribution systems and compiled geodatabases [2,9,13,53]. Many countries offer spatial databases, i.e., high-resolution digital terrain or elevation models (DEMs). Global terrain data from Google Earth or other platforms, along with the Shuttle Radar Topography Mission (SRTM) DEMs from the United States Geological Survey [54], can also be used but with caution, i.e., only for the initial assessment of SHP locations and not for flat terrains or low-lying countries and regions with a low vertical resolution in topography. The SRTM DEM with a spatial resolution of 30 m has a reported accuracy of ±16 m [55], which is acceptable. However, the vertical accuracy crucial for determining elevation has also been reported to be <9 m for flat terrain and 4.3 m for mountainous regions. Such accuracy would exceed any project design standard.

As the analysis of prospective sites in wastewater networks in Lithuania has shown, gross head (a drop in elevation) or location coordinates can be determined from the GIS spatial data portal freely available in the country. However, the accuracy of this assessment will be low for ultralow- or even low-head schemes because the vertical resolution of the DEM would be insufficient.

Spatial regression analysis performed to assess the correlation between the hydropower recovery potential of the sites in water networks and the geographical data showed that there was no significant correlation between the population and terrain variability [56]. In other words, the potential of a hydro site is too site-specific for variables to explain the variation in potential.

Water utilities also operate their own georeferenced database from which wastewater network physical features can be identified. Unfortunately, these GIS databases rarely provide flow data. Flow rates of a particular wastewater collection service area can be roughly evaluated according to the population or number of residential areas connected according to the water use. However, it is hard to estimate the generated effluent volumes of industrial facilities.

Various GIS tools have been proposed to search for traditional hydropower sites or their potential [35,57]. Unfortunately, as our study has shown, no automatic site search using GIS tools is possible due to the technical complexity of sewage pipeline systems.

3.2. Wastewater Resources

As mentioned earlier, one of the main tasks is the assessment of wastewater resources streaming from the municipal network. In most cases, the wastewater volumes in the network are calculated from the data of the pumping stations or their electricity consumed. However, as the analysis of collected flow records and on-site measurements revealed, the reliability of such an evaluation is not high. When on-site wastewater flow data are unavailable, estimates can be derived from water use records or other relevant information. One of the key elements for energy estimation is the average annual wastewater flow rate, followed by the distribution of daily flow rates over time. The latter is usually expressed as the FDC, as discussed below.

The analysis of the mean annual wastewater flow at the outlets of the 56 collected WWTPs operating in the country showed that it correlates quite well with the population equivalent (PE) and wastewater collection (service) network area (A, km²) (Figure 5). Two separate cases were tested: (1) all 58 WWTPs and (2) WWTPs operating in smaller municipalities, excluding large cities (PE below 100,000).

A summary of the regression analysis (ANOVA) is given in

Table 5. Summary of regression analysis.

#	R-Squared	Observations	Standard Error	Intercept	X-Variable	F	Significance F	t-Stat	p-Value
1.	0.88	56	0.023	0.0032	0.0032	378.2	ca. 0	19.45	ca. 0
2.	0.97	56	0.035	−0.0042	2 × 10−6	1673.8	ca. 0	40.91	ca. 0
3.	0.62	52	0.018	0.070	0.0013	79.5	ca. 0	8.91	ca. 0
4.	0.79	52	0.013	0.0054	1.51 × 10−6	192.0	ca. 0	13.85	ca. 0

; the significance level was set to p < 0.05. The previous figures and Table 5 show that satisfactory correlation results were obtained. Approximate values of mean annual flow for ungauged sites could be determined. However, the flow of each site was too site-specific for variables such as the population equivalent and collection network service area to accurately explain the variation in the wastewater flow rate.

Some studies examined the variation in wastewater flow due to heavy rain events in WWTPs to optimize turbine selection [15]. However, few such studies were conducted on the sewage network to install turbines. The wastewater flow hydrograph accounting for stormwater events at the intake of a possible location for installing a turbine is shown in Figure 6. A standard engineering and operation practice is that sewage flows must be separated from stormwater flows in the urban area. However, during flash floods, some rainwater unavoidably infiltrates the sewage system.

As can be seen, the relatively even wastewater flow was severely disrupted during short rains. As a result, short but very high spikes can be seen in the hydrograph. They exceeded 5–6 times the normal range of wastewater fluctuations. No turbine could accept these instantaneous sewage peaks without compromising performance. At first, installing a buffer tank for storing rainwater in the network would be hardly rational considering civil engineering costs. Of course, rainwater volumes would be higher in systems with a more significant wastewater system service area. However, the nature of the spikes would be similar but more delayed and smoothed (due to the longer surface water concentration time). A more detailed analysis of rainwater runoff frequencies could support this hypothesis.

Typically treated water fluctuations over 24 h at the outfall of the WWTP are shown for each month in Figure 7.

Flow in wastewater treatment plants varies with daily water demand patterns, varies by season and year, and is subject to heavy rain events. However, these hydrographs do not exhibit spikes in comparison to Figure 6; the rainwater concentration–time curve is flattened due to the enormous wastewater service area.

Normalized (dimensionless) FDCs for potential sites were constructed (Figure 8). Each 5% interval on the curve is equivalent to five percent of 8760 h (number of hours per year). Despite being in a dimensionless form (normalized to the mean annual flow), they varied widely. The parameter Φ was used to compare them, as shown in Table 3.

The proposed methodology allowed establishing a flow duration curve for an ungauged site using only three values: the highest flow, the mean flow, and the lowest flow, taking into account the parameter Φ.

A wastewater treatment outfall with a sufficient drop in elevation is generally suitable for an MHP installation due to the high and constant volume of flow. However, when installing turbine units, care must be taken to consider water-level fluctuations of the receiving water body in the outfall of WWTPs. This is especially relevant for low-head power plants, which were not sufficiently quantified in previous studies. The drop in head over time is significant, resulting in decreased power generation during a period with abundant water in a stream (Figure 9). The design head should be derived from the head duration curve at 30% of the exceedance level [58,59].

3.3. Hydromechanical Equipment

As mentioned above, the selection field for classical turbines is relatively narrow in a flat terrain where elevations are relatively low. This can be explained by the low flow rates and small size of turbine units, which increases the unit price of turbines (EUR/kW) compared to the larger capacity of hydro turbines. Only reaction-type turbines can be used at low-head schemes, e.g., propeller, Kaplan, seldom crossflow, and Francis.

It is also worth mentioning here the Archimedes screw, a gravity turbine that became popular several decades ago for low- or even ultralow-head schemes [46,60,61]. Despite the significant advantages of Archimedes screw turbines (low head, tolerance to water quality, and tolerance to debris or clogging), their applicability in urban areas is restricted. In municipal sewage networks, this kind of turbine is not likely to be accepted by the city’s residents due to bulky and heavy construction that causes visual pollution and incurs operational noise. These turbines can thus only be installed within the WWTP, within its sewage treatment outlets, or at a remote location away from residential areas.

The development of compact and modular turbines is a recent trend in turbine technology, e.g., a generator unit using a propeller turbine in an axial flow design [62,63,64,65,66]. A submersible turbine and a generator are combined in one unit; therefore, the need for the use of a powerhouse is eliminated and installation costs are reduced.

The commercially available modular units suitable for in-conduit hydropower are listed in

Table 6. Main features of modular turbines currently available on the market and suitable for in-conduit hydropower in a low-head segment.

	Turbine Type	Net Head (m)	Flow (m3/s)	Power (kW)	Comments	Reference
1.	Amjet ATS	1.5–12.8	0.2–26.0	3–2500	A range of series is available 1	[62]
2.	StreamDiver	2.0–8.0	2.0–12.0	50–1450	There are at least 7 modules	[63]
3.	Turbiwatt	1.2–8.0	0.1–3.6	3–120	Three (3) available modules/series	[64]
4.	Flygt	2.5–20.0	0.7–10.0	40–850	Six (6) available modules/series	[66]
5.	Hydromatrix	2.0–25.0	5.0–13.0	200–2200	Very large flow	[65]

¹ Depending on unit version/series.

. Unlike Kaplan turbines, their runner vanes are not adjustable. HYDROMATRIX units operate under significant flow rates and cannot be used for wastewater networks located in medium-size urban areas (population between 200 and 500 thousand) because of the insufficient generation of effluent volumes (maximum flow of some 1.5 m³/s). The same can be said for the StreamDiver turbine.

Turbine costs comprise approximately half of the conventional hydropower project development costs [18]. Lower-end unit costs start from 1100–2800 EUR/kW [64]. Financial analysis has consistently shown that SHP projects in general and in-conduit projects in particular suffer from the rapid increase in development cost at lower head and capacity, resulting in a feasibility issue for low-head facilities. Significant opportunities to lower development costs through specific research and development are proposed [32]. To offset this drawback of conventional turbines, low-cost generators, e.g., pumps as turbines (PaTs), have been suggested.

The use of PaTs for energy recovery has been demonstrated to be cost-effective, as low as 12% of the cost of conventional small turbines [14,67,68]. As expected, a smaller-capacity PaT attracts a higher unit cost. For instance, at 3 kW, the unit cost varied from 500 to 1000 EUR/kW, whereas at 20 kW it was in the range of 200–500 EUR/kW.

Centrifugal pumps and reverse-mode PaTs are radial or mixed-flow types, similar to conventional Francis turbines. In contrast to an axial propeller (or Kaplan) turbine, both units are much more sensitive to clogging issues when operating in effluents charged with suspended particles. Moreover, another limitation of PaT and Francis turbines is their inability to operate effectively outside design flow.

The so-called bypass configuration is a classical layout for installing a turbine in a sewage pipeline system (Figure 10). The unit is usually operating in parallel with the existing pipeline.

3.4. Tools for the Assessment of Technical and Economic Feasibility of Installing In-Conduit Hydropower Systems

The main technical features of the following tools are presented in

Table 7. Review of small hydropower assessment software intended for the assessment of in-conduit hydropower at individual sites.

	Year	Developer	Operating System	Applicable Countries	Application	Accessibility	Features
							Units	Energy	Hydrology	Hydraulics	Turbines	Costing	Economic	GHG	Design Level
RETScreen [38]	1998	Canada	Software	International	Conventional HP	Open 1	SI	x	FDC	x	PaT not included	x	x	x	Feasibility and preliminary design
ICHPST [33]	2013	Alden, US	MS Excel	USA	WSW	Open	I 2	x	FDC	-	Mostly all types, including PaT	x	x	x	Screening
BCAT [12]	2019	Stantec, US	MS Excel	USA	WSW	Open	I 2	x	Design flow	-	Mostly all types, including PaT	x	x	x	Screening

¹ Only in viewer mode. ² Imperial. x—Feature included.

: (1) RETScreen Expert; (2) In-Conduit Hydropower Project Screening Tool for water supply and wastewater treatment facilities (ICHPST); and (3) California’s In-Conduit Hydropower Implementation Guidebook Business Case Assessment Tool (BCAT).

There were no significant differences between ICHPST and BCAT. However, their imperial units may not be convenient for those who use the metric system, despite their evident applicability to the municipal water.

ICHPST (Alden) was developed to complete a preliminary evaluation of a potential hydropower facility. It provides estimates of power, energy, cost, and financial viability based on user inputs. It can be applied to water going into a water supply system or being discharged from a wastewater treatment facility.

BCAT enables users to complete a preliminary evaluation of hydropower potential at urban water facilities. Considering the specific system and location (pressure, head, and flow), a user can estimate power generation potential and obtain estimates of power, energy, cost, and financial viability. BCAT considers pressure requirements downstream of the hydropower scheme. This is not the case with RETScreen (pressure is discharged to the atmosphere).

RETScreen software, a conventional hydropower project tool, requires a great deal of engineering preparation before it can be used to assess hydropower schemes in municipal water distribution systems. Its technical level is much higher than the above-reviewed screening tools; it can be easily adapted to complete feasibility studies and preliminary design of in-conduit hydro schemes.

3.5. Wastewater Quality and Possible Effects on Turbine Operation

Screening of solid-laden media is the first operation at wastewater treatment plants and wastewater collection systems. Screening removes objects such as rags, paper, plastics, and metals to prevent damage and clogging downstream equipment and piping [69]. It is evident that turbines should be designed explicitly for raw wastewater and operate without clogging or fouling caused by material in the fluid under any operating conditions within the range of service specified. For hydropower schemes using untreated wastewater, a trash rack chamber must be installed at the intake [2]. The trash rack chamber’s operational cost was identified to range from 0.03 to 0.08 USD/kWh.

Clogging or “ragging” taking place when pumping raw wastewater, as analyzed in a number of studies [30,31]. The causes of hydro turbine clogging are the same as those for pumps, i.e., the suspended solids transported by the sewage.

Wastewater flow quality issues are often mentioned for siting turbines, but no quantitative estimates are provided. Specially designed turbines can adequately handle wastewater sand particles or their chemical components. The actual cause of clogging is not solid, but fibrous material contained in the sewage. When long, stringy solids or fibers are present in the flow, problems can occur, particularly for axial (propellor) and radial flow machines, when these materials are caught on the rotating parts [70,71].

No studies have been performed in Lithuania on the impact of wastewater quality on the operation of hydraulic machines and their clogging. Water utilities have observed an ongoing reduction in water usage. However, the volume of waste in the sewage does not seem to be decreasing. As a result, the proportion of solids in the flow is gradually increasing, which is an issue for many water utilities. Standardized qualitative wastewater monitoring is carried out at WWTPs but not inside the sewage network. However, the main focus is on the chemical parameters of the effluent entering and exiting the WWTP. The total suspended sediments (TSS) are only occasionally recorded. Available data from water companies show that the average concentrations of TSS in the raw effluent can reach 500 mg/L. After treatment, they decrease at least 25-fold, down to 20 mg/L [72].

Large solids, rags, and other fibrous materials from wastewater can be a severe issue for operating turbines if not monitored. Spot measurements conducted upstream of Kaunas WWTP showed that manual cleaning of the K2 grating (rack gap = 5 cm, mean flow = 0.18 m³/s) is performed twice a week, and approximately 2–5 kg of fibrous dry matter is collected. Approximately 500 kg of dry material can be accumulated per year. This harsh environment can be considered when installing turbines in such locations.

3.6. Multicriteria Analysis for Selecting Potential Sites

In total, 17 criteria were used for the multicriteria analysis (Table 3). Twelve criteria were to be maximized, while five were to be minimized. Their grouping was as follows: (1) technical-related (TEC; layout, turbine type, design flow, gross head, etc.), (2) economic-related (ECO; investment costs, electricity generated, simple payback, etc.), and (3) environmental-related (ENV; GHG reduction and use of electricity). These groups are only a suggestion, as the user can subjectively assign a criterion to the group of their choice, e.g., economic criteria, such as “electricity generated”, can also be considered as belonging to the group of technical-related factors. Moreover, the weighing system is a subjective element of the information [52].

A global summary of the results of this ranking based on the software HYPSE [51] is provided in

Table 8. Summary of rankings with the final ranking global synthetic index (GS_i): 1 = best position; 8 = worst position.

	Alternatives	Index
		ci Concordance Dominance Index		di.SD Simple Discordance Index		di.WD Weighted Discordance Index		di.AD Aggregate Discordance Index		di.AWD Weighted Aggregate Discordance Index		GSi Global Synthetic Index
A. Basic scenario: Criteria weights are equal, and group weights are different (ECO—29.42%, TEC—58.81%, ENV—11.77%)
a	K1	6	0.001	5	0.108	5	0.099	6	0.732	6	0.730	6	−0.728
b	K2	8	−1.530	7	0.301	7	0.299	8	1.931	8	1.929	8	−3.460
c	K3	5	0.059	4	0.0	4	−0.007	4	−0.444	4	−0.444	5	0.503
d	V1	2	0.764	1	−1.768	1	−1.758	1	−1.514	1	−1.511	1	2.275
e	V2	4	0.353	6	0.127	6	0.133	3	−0.855	3	−0.853	3	1.206
f	V3	1	1.060	2	−0.409	2	−0.417	2	−1.186	2	−1.186	2	2.246
g	A1	3	0.529	3	−0.086	3	−0.081	5	−0.148	5	−0.147	4	0.676
h	A2	7	−1.236	8	1.727	8	1.729	7	1.484	7	1.482	7	−2.718
B. High economic scenario: Criteria weights are not equal, and group weights are different (ECO—49.39%, TEC—42.18%, ENV—8.43%)
a	K1	2	1.076	5	0.108	1	−1.720	6	0.732	3	−0.345	2	1.421
b	K2	7	−1.606	7	0.301	8	2.371	8	1.931	8	1.276	8	−2.882
c	K3	5	0.098	4	0.0	6	0.633	4	−0.444	5	−0.120	5	0.219
d	V1	4	0.151	1	−1.768	5	0.198	1	−1.514	2	−0.439	4	0.591
e	V2	3	0.423	6	0.127	7	0.812	3	−0.855	4	−0.287	3	0.710
f	V3	1	1.722	2	−0.409	2	−1.560	2	−1.186	1	−1.039	1	2.761
g	A1	6	−0.073	3	−0.086	3	−0.773	5	−0.148	6	0.072	6	−0.145
h	A2	8	−1.792	8	1.727	4	0.039	7	1.484	7	0.882	7	−2.674

. The indices expressed by the numerical values are briefly explained in the methodology.

Two scenarios, as presented in Table 8, were considered:

A. “Basic scenario”: The criterion weights were equal (5.88%); however, the weights of the groups were different (ECO—29.42%, TEC—58.81%, ENV—11.77%).

B. “High economic (ECO) scenario”: a priority by a factor of 1.7 was assigned to the ECO-related group (49.39%). As a result, the individual criteria weights within each group were 9.88% and 4.22% for the ECO and TEC/ENV groups, respectively

In the case of similar criteria weights (basic scenario), the sites in Vilnius (V1, V3, and V2) were the top ranked (highest values of the index GS_i). When a greater weight was assigned to economy-related criteria (high economic scenario x), the top ranked site appeared in Kaunas (K1).

The concordance dominance index c_i considers the positive aspects of the alternatives. For all x scenarios, site V3 was at the top (c₆ = 1.060 and 1.722). The negative aspects of the alternatives are represented in four discordance dominance indices (d_i.SD, d_i.WD, d_i.AD, and d_i.AWD), which determine the relative discrepancy in the effects of the two alternatives. A low value of these indices (i.e., a negative value) implies a stronger dominance of alternative i. As a result, these indices slightly changed the site ranking, but the overall judgment remained similar.

4. Conclusions

• While the potential of energy recovery from wastewater systems using micro-hydro plants (MHPs) is an appropriate solution to improve the energy efficiency of the municipal water sector, it has seen no exploitation due to a number of technical and nontechnical issues in low-lying countries. Nontechnical problems include a lack of awareness about the scale of the existing resources available in water networks.
• The potential in lowland areas in terms of power capacity resulting from mostly low-head sites cannot be compared to that of elevated topography. In addition, for flat terrain, the selection field for turbines is relatively narrow; moreover, the low flow rates and small size of turbine units increase the unit price of turbines.
• A methodology was developed to quantify the potential and identify conduit hydropower sites in a lowland country’s wastewater systems, including resource assessment, suitable tools to make a preliminary assessment of potential sites, and choice of turbines and their operating parameters in a harsh environment. The lack of in-depth studies on wastewater quality’s impact on hydro turbines, particularly the risk of clogging them in sewage networks upstream of WWTPs, can be a severe problem.
• A conventional multicriteria analysis can help select the most appropriate site for constructing MHPs in urban water areas to achieve energy recovery. There are plenty of multicriteria tools available on the market for solving any real-world issue. However, at least preliminary site assessments and design procedures must be accessible beforehand for this analysis.