Planning a micro hydro plant19 March 2007
Giampiero Goretti and Raffaele Ruscitti illustrate the methodological peculiarities characterising the technical-economic feasibility study and planning of a low head micro hydro power plant in Italy’s Tiber valley which employs a Coanda-effect screen. Based on the site specific hydrogeology, the plant layout has been designed taking into account both the operational performances and the main economic parameters of investment for each possible configuration
In the context of the rational exploitation of alternative or integrative sources, small scale hydraulic energy takes considerable and renewed interest from more and different points of view: the high energy density, the land diffusion of a lot of unexploited resources, the availability in this sector of a well known technology, and the continuous and often very fast rising fossil fuels prices. In addition, these micro hydro plants can be easily inserted in irrigation or water distribution projects, taking advantage of lower costs.
In this paper particular attention is focused on peculiarities of the technical and economic feasibility study and the design of an ultra-low head micro hydroelectric power plant for a farm holiday centre also aimed at didactic ecological activities.
Characteristics of the farm
The Collevalle farm is located to the north of Rome, along the Vezza stream, close to its confluence in the Tiber, 75m asl. At the site a small dam is used to raise the water level for irrigation purposes. The farm is specifically aimed at an intense development, application, demonstration and popularisation of products of scientific research in the field of biological agriculture, with a wide interdisciplinary perspective in the field of agronomy, mechanisation, economy and management of the projects, and finally the transfer of the acquired knowledge in the form of case studies. Since the vegetable cultivation and animal breeding is operated by biological management, its basic strategy is directed towards sustainable management systems. The electric energy requirement of the farm and its share were quantified by data recorded in 2002.
For the global consumption of 32,779kWh, the total expenditure reached EUR 7692.00. This unexpected small annual consumption comes from the precise managerial behaviour: a wider employment of irrigation water would not be convenient, because the increase of the electric energy cost wouldn’t be overcome by the benefits of the production increase.
Flow and hydrologic data
A number of hydrologic data were acquired. During 1981 and 1982, direct flow measurements were recorded monthly by the Department of Geology of Roma Tre University (Table 1)
In addition, direct measures of the flow rate were achieved through the use of an electromagnetic induction instrument. The outflow value through one of the two gates of the existing dam was also set.
At the National Hydrografic Italian Service, as well as at the Authority of the Tiber Basin, instead of direct flow rate data, monthly cumulated precipitations series were found, for the periods 1921-1950 and 1980-1999.
The mean precipitation value recorded in these periods was 855.1mm/year, with a standard deviation of 180.42mm.
The river flow rate is the sum of both the linear source and the run-off contribution which reaches zero in dry periods. As the remarkable drainage values quickly exhaust, the flow of the Vezza is characterised by a great irregularity, with sudden floods rapidly decreasing to the less varying values; the water stored in the reservoir of volcanic rocks feeds the river throughout the year so that, despite the modest flow, the Vezza bed is never dry, and water resource is continuously available. Based on the 1981-1999 data, it was verified that from May to September the Vezza flow is typically characterised by the linear source regime, fed by the underground waters.
In order to better ascertain the hydroelectric potential of the stream, from July 2002 to November 2003 the Vezza flow was directly measured. Added to the hydrogeologic considerations and the statistic analysis of the rains, this direct study was aimed at a refined matching between such data and the Flow Duration Curve, namely the 1981-1982 one.
As a result of this synthetically reported study, in 50 years (1921-1950 and 1981-1999) it was observed that mean value of the precipitations in the zone – 855.1mm – were close to the ones registered in 1981 (843.2mm) and in the 1982 (842mm). In addition the flow measurements proved that the year 1981-1982 could be assumed as a representative year. Based on hydrological and ecological (Micro-habitat) considerations, a value of the minimum vital flow (QDMV) of 80l/sec  was estimated and the duration curve drawn in Figure 1 was obtained.
Estimation of the available power
By the above mentioned data and the corresponding duration curve, the computation of the design flow rate and power of the plant was based on 80l/sec as the minimum vital flow of the river, 5.1m hydraulic head, and a 0.8 mean efficiency value of the intake and diversion duct, having at first assumed a 15% of the gross jump value for the losses in the pipes. In fact, as is customary in micro hydro plants, the water covers a greater distance than in mini hydro schemes, in which this loss is estimated as 10% of the gross jump. The resulting figures were assumed as a first approximation of power and energy.
The first sizing of the plant components was carried out by means of an empiric curve of the costs, taking into account the corresponding producible energy. Maximum production economy was achieved at a flow rate of 0.56m3/sec; the corresponding first approximation values of the maximum power and of the producible energy were 19.5kW and 150MWh/year.
A nominal 0.56m3/sec flow rate was than assumed for the turbine selection and for the loss computation. Lastly, taking into account the actual power plant layout, the pipe losses and the productivity estimations were refined.
Different plant solutions were compared by an iterative design procedure with particular consideration of environmental impact.
The main points of study were the amount of the water resource and its potential, a detailed analysis of the topography and geomorphology of the site, the selection of the plants basic scheme, its location, and turbines and generators, and the eventual environmental impact and economic evaluation of the project.
The prominent characteristics of the proposed solution are: compactness, low environmental impact, installed power lower than 20kW, energy production entirely assigned to the farm self-consumption, with the so called grid connection regime of the ‘net-metering’, the careful observance of the law regarding minimum vital flow in the river bed (QDMV), parallel connection to the electric grid.
In order to reduce the heavy and prevailing cost of the civil works by means of a particular frontal intake, the spillway of existing weir was exploited, taking advantage of its gates to interrupt the plant water flow in case of maintenance or breakdown. In addition, due to its inexpensiveness, the traditional works of filtering and sedimentation of solids were replaced by a standardised Coanda-effect Screen element (Figure 2). A small forebay was located immediately below the frontal intake to fill the diversion duct. The height of the rear-edge of this small basin is lower so that the filtered overflow can escape along the front-side of the existing dam, causing minor erosive phenomena at the base of this structure.
Starting from the basin in a cross direction and resting on a short support, the headrace goes towards the left bank of the river and reaches the turbine in a rectilinear way. A calibrated hole below the pipe supplies the QDMV.
The optimal pipeline length and the corresponding head were calculated by a careful examination of orography of the surrounding area. In order to obtain an available gross head of 5.1m, the pipeline reached 350m; the diameter of a HDPE duct was 710mm. An asynchronous generator and a turbine, well suited to the available values of the net head and flow rate, connected by a poly-v multiplier, and devices of automatic control were located in a small box close to the river; the tailrace was made by an open channel.
As mentioned before, a Coanda-effect screen was chosen in order to limit both the civil works and the maintenance costs. This is a refinement of the trap-intake consisting of a stainless steel transverse tilted rack; its bars have an almost triangular section, with sharp edges in the upper part of the panel, and are rotated around their axis in order to modify the set up of the slot entrance between the bars.
This innovation, along with the Coanda-effect, is the base of the effectiveness of this configuration. This panel was located immediately underneath the spillway. Accelerated by a first element at the top of the weir, the overflowing water reaches the panel endowed of a parabolic or, for constructive simplicity, of a circular profile. The accelerated flow advances tangentially to the upper face of the panel bars and passes through the slots (about 0.5-1mm in width) falling down in the underlying basin; excess water, fish and silt are led towards the base of the panel, and released into the river. Filtering power of this panel is very high: in fact it excludes all particles greater than 1mm and 90% of those between 0.5-1mm in diameter. As a substantial advantage, these panels are self-cleaning (at Re>1000) because the sharp edges hold less sediment in comparison with the rounded ones – the Coanda-Screens are continually cleaned by the water flowing over them. If the panel is only partially wet, sediment accumulates on the boundary of the dry part until the next increase in flow, where typical speed values along the panel are about 2 or 3m/sec .
The intake design was made by the mathematical model reported by Tony L. Wahl, implemented as the software ‘Hydraulic Performance of Coanda Effect Screens, Ver. 0.43’ , in order to achieve: reasonably low head loss in comparison with the overall plant losses; high filtering power; discharge capacity allowing a future plant upgrade by replacing only the turbine-generator equipment; satisfactory performance of the acceleration element, even with the simplified circular profile, not exceeding longitudinal length of the screen; high efficiency; and maximum element standardisation.
Taking into account the above requirements several configurations were considered and run, with design input summarised in Table 2.
The resulting shape of the accelerating element is shown in Figure 3; the radius of the circular profiles of its left and intermediate parts is R2 = 0.0949m and R1 = 0.2173m; the profile of the right part follows the equation:
Y= - [0.5005*(X)1.868)*(0.1261-0.868)]
The overall element dimensions are: 0.1465m (width), 0.05m (height) and 4m (length).
As mentioned earlier, in view of a possible plant repowering by changing the electro-mechanical equipments, the screen was oversized – in flood conditions the panel can accept up to 1800l/sec, and filters 900l/sec. Compared with the incoming flow, the intake discharge capacity is shown in Figures 4 and 5. Due to the concavity of the screen along the profile, the acceleration of the water decreases while its speed increases; being almost in the entire screen over 1000, following the stream velocity the Reynolds’ number takes increasing values; that ensures the self-cleaning of the screen.
In the standard design of these panels the bars tilt angle is 5°, the panel inclination reaches 60°, and the accelerating part is considerably wide. In order to cut down the intake head losses the panel inclination was taken at 40° and the height of the accelerating plate was reduced up to 5cm, thus allowing it to reach the minimum height of the screen while preserving its functionality.
Having no driving mechanism and moving elements, the grid doesn’t need any regular maintenance, meaning the screen, as well as the plant, can work in a continuous way and the corresponding management and maintenance costs are avoided. In fact, due to the self-cleaning, in all the testing sessions no blockage was verified. In flood conditions the mechanical strength was proved satisfactory, while the freezing tests didn’t underline any problems even for air and water temperatures below 0°C (water temperature of the Vezza river never goes under 5°C). As the specific cost of this screen was about US$1625 m2/sec (2003), the total amount of the selected configuration (2.824m2) reached US$4589, which is below that of traditional solutions which include several mechanical and civil works along with their maintenance costs.
The structure between the screen and the pipeline is made by a prismatic basin, installed just below the filtering panel that conveys the water towards the pipeline entrance at one of its ends. This particular shape was chosen in order to gain good filling conditions of the pipe. As a reduced initial diameter of the pipeline (for example 400mm) would produce too high energetic losses, a diameter of 700mm was preferred for this section. In addition, in order to avoid outflow problems, a vent was planned to eliminate trapped air. The basin was designed for a maximum flow rate greater than the present one, giving excess water the chance to overflow close to the weir, thus limiting erosive phenomena at the foot of this structure. This basin is supported by two short pylons of concrete, from which the pipeline penetrates the left bank. In case of plant maintenance these new structures allow the opening of the existing gates without any interference or possible damage; the vital minimum flow is discharged below the transversally installed pipe.
For this specific application, ducts of four different materials were taken into consideration: polyethylene, steel, concrete and cast iron covered by enamel; the design discharge was fixed at 0.56m3/sec, and a pipeline length of 350m was assumed. To avoid deposits of thin sands of high specific weight, for example, the minimum water speed value was set at 1m/sec, while, in order to prevent the consumption of the pipeline walls, a maximum speed value of 2m/sec, typical of continually working plants, was assumed; based on these considerations the standardised diameter ranged from 560mm to 800mm. In view of the very small available gross head and construction and maintenance advantages offered by this material, a high-density polyethylene pipe (HDPE) was chosen: the comparison of the load losses for each diameter, and for the different materials is summarised in Figure 6. Taking into account the low available power and the technical life of the plant, the choice of the optimal diameter was made by the criterion of the lowest passivity, assuming a technical life of 25 years (at these low pressures, the pipes can even last 50 years) and a mean rate of interest of 5.1% taken from an IRS to 25 years, to which corresponded a rate of amortisation of 0.072. Both the regimes of the energy transfer to the Italian grid – net-billing and net-metering – were also taken into consideration. In net-billing, the produced energy is paid 0.0831EUR/kWh; in net-metering it is exchanged at 0.235EUR/kWh (quotation 2003). Some calculation results are shown in Figures 7 and 8. By the comparison of the two passivity curves (Pa), a marked difference between the values of the diameter of maximum ‘profit’ can be detected. In the net-billing hypothesis the minimum of passivity is found at a diameter close to 700mm; because of the greater weight of the missed production, in the other case this diameter falls between 800mm and 900mm; nevertheless, taking into consideration the transport of sand, the range was restricted between 700mm and 800mm. Due to uncertainties of the evolution of the energy market, a reduced cost of investment rather than a production increase was preferred, and eventually the standard diameter of 710mm was chosen. If the rate of interest had assumed values greater than the one considered in the calculations, it would have registered an increase of value of the optimal diameter. Corresponding to the chosen diameter the computed net head of the plant was 3.9m. For the net head values of approx 4m and the discharge capacity of 0.5m3/sec, well suited turbines are Banki-Michell and Kaplan turbines (Figure 9). Bearing in mind the above mentioned economic considerations, in order to limit the costs of investment, the Banki was preferred to a more expensive Kaplan; as a further advantage the Banki-Michell maintains a high efficiency to 15% of the nominal discharge capacity; in the designed plant the minimum flow rate is 42.5% of the maximum value, so that the turbine efficiency remains between 78.5% and 87%. Consequently, the maximum generated power reaches about 18.8kW and, assuming an overall efficiency of the poly-v multiplier electric generator group is equal to 0.89, a 126,600kWh annual energy production of this plant was estimated. Considering the 3.9m net head, the 0.56m3/sec capacity and a corresponding power of 18.8kW, the runner diameter of a customary Banki-Michell turbine is 400mm; by matching this turbine to a six polar-couples asynchronous generator (nominal tension 0.4kV, nominal frequency 50Hz, squirrel cage rotor, poly-v 1:2.5 running speed multiplier, 0.97 efficiency of the multiplier and 0.92 efficiency of the generator), the rated electric power of the plant was about 16.8kW.
The economic evaluations included taxes, along with the electro-mechanical equipment, the civil works and the engineering, as well as the building yard costs. In the cost breakdown summarised in Figure 10, the turbine-generation group takes the fundamental disbursement, as the pipeline is the main item of the civil works. The evaluation of the engineering expenses (Figure 11), was based on the Italian laws and regulations of the engineers and architects’ fees. It included: planning, direction of the works, coordination for the planning and execution of the works.
The economy of the designed plant configuration was finally verified through a comparative cost-benefits analysis of the different solutions, made both by the investment evaluation indexes in current coin (Return on Investment, Pay Back Time) and the updated indexes (Net Present Value, Profitability Index). To make the comparison easier, the further ratio NPV/I0 was added (I0 is the initial investment).
As an example, two reference solutions characterised by a rated power of 20.8kW and the net-billing transfer (solution A), and by a rated power of 18.8kW and the net-metering energy transfer (solution B) are compared in Table 3. As is clearly shown, particularly by the NPV, both the investments are highly profitable; the ROI and PBT economic indexes, immediate as they are in current coin, don’t keep any track of the economic devaluation. In addition they don’t underline the cash flow following the pay back period, and therefore are not able to represent the profitable investment throughout the plant life. As it is repaid in less time, the risk of Solution A is lower. Better comparative evaluations between the investments can finally be drawn by the values of profitability index (PI) and by the ratio between the net present value and the initial investment.
The values of all the considered economic indexes prove that both the solutions are convenient, but solution A is the most profitable in its own technical life.
Micro hydro power plants allow for water exploitation with a very low environmental impact, without modifying the prevailing use of the watercourse. They can be vital to the electrification of isolated communities, farms and remote areas not covered by the national electric grid. In addition, the plant hydraulic works often bring remarkable benefits to the water course itself. As a further environmental benefit, avoiding pollutant emissions, the hydroelectric generation involves some mitigation of the CO2 production. From this point of view, the production of one kWh of electric energy by fossil fuels needs nearly 2100 Kcal of heat, releasing in the air about 0.67 kg of CO2 (ENEL, 2003). The avoided emissions of the two previously considered plant solutions are respectively 85 and 62 tons of CO2 per year; a synthesis of these elements for the solution B is represented in the Table 4.
Flowing through the turbine passages, water is oxygenated thus improving the micro-habitat conditions as well. In these plants, the acoustic pollution can be practically negligible, reduced under 70dB inside the power house, and down to imperceptible levels outside.
Each element of the plant can considerably change the site visual impact, and therefore public opinion is generally reluctant to accept plants that notably modify the visual characteristics of the sites, even if they produce green energy. In order to mitigate this impact, in smaller plants some elements like vegetation and forms and colours integrated in the landscape can be easily employed, and plant fittings can be located in the subsoil.
Table 1 Table 2 Table 3 Table 4 Figure 1: Duration curve Figure 1 Figure 9: Well suited turbines are Banki-Michell and Kaplan Figure 9 Figure 3: Shape of the accelerating element Figure 3 Figure 4: Intake discharge capacity Figure 4 Figure 5: Intake discharge capacity Figure 5 Figure 6: Comparison of load losses for each diameter, and for different materials Figure 6 Figure 7: Calculation results of net-billing Figure 7 Figure 8: Calcuation results of net-metering Figure 8 Figure 10: Cost breakdown Figure 10 Figure 11: Engineering expenses Figure 11 Figure 2: Standardised Coanda-effect screen element Figure 2 Author Info:
Giampiero Goretti, Project Management Renewable Energy Power Plants, Tisol, Rome and Raffaele Ruscitti, University of RomeTablesTable 1 Table 2 Table 3 Table 4