The scope of this section is to describe the most widely adopted technical solutions concerning equipment and support systems used in Mediterranean hatcheries, with emphasis on hydraulic aspects. Design parameters, size calculation and installation criteria are dealt with in this section.
Each subsection is divided in two parts, both treated in detail:
1. the main support systems and,
2. their technical components.
Civil works related to buildings or roads, including concrete calculations, electrical and thermomechanical systems, are not dealt with because:
they do not differ significantly from normal industrial standards in any country;
the professionals in charge of their design and construction bear full civil liability;
each country has its own rules, codes and standards for industrial plants, which may differ considerably.
The seawater supply system is a crucial element of any hatchery project as the entire production depends on a reliable supply. It includes the following components:
the seawater intake;
the pumping station;
the network connecting pumping station and hatchery;
the first water treatment system (coarse filtration up to 100 mm);
the internal distribution network;
the secondary water treatment system (fine filtration up to 1 mm);
the discharge system, that includes the drainage network, the wastewater treatment and the outlet.
Fig. 28  Seawater supply system diagram 
In Mediterranean hatcheries, seawater can be supplied from two sources:
1. by direct pumping from the sea;
2. marine/brackishwater wells.
In both cases a reliable water supply system is a key factor for the successful production of fry and for the whole economy of the farm as it accounts for a very important part of its energy cost. A well designed system, in terms of piping and proper choice of adequate pumps and ancillary technical equipment, assures the efficient functioning of the breeding centre. This system is crucial in large commercial hatcheries, in which the requirements for treated marine water can easily exceed 150l/s.
The direct pumping of seawater is the most widespread system to supply Mediterranean seabass and gilthead seabream farms and hatcheries. It is described in detail below.
There are different types of water intakes according to the type of coastline, distance from the hatchery to the sea and type of beach and sea bottom sediment. The three most common situations encountered are described in the following sections:
sandy coast with a low gradient,
rocky coast,
natural or artificial enclosure.
The peak water flow requirements of the hatchery must be carefully calculated to design properly the entire system. Future developments and system maintenance, which is an important aspect, should also be taken into consideration. If not properly kept in mind in the design phase, these two aspects may generate potential dangers such as abundance of fouling organisms in the pipelines or excessive silting in front of the water intake. These problems may easily become major drawbacks during the operation of the hatchery, requiring costly interventions to solve them.
Because of the risk of clogging by sediments, a water intake located on a sandy low coast would require the construction of civil works on the shore. Their design, related to their possible impact on littoral sediment transport, requires a detailed study of: wind regime, swell regime, sea level variations, tidal regime, bathymetry, and coastal currents.
Fig. 29  Water is pumped to a higher level at the end of the canal

LEGEND MPS = main pumping station St = strainer 
For this typology, the two main technical solutions usually adopted are: (i) a canal protected at sea by two breakwaters and, (ii) a submersed pipeline.
Water intake through a canal protected at sea by two breakwaters
The design of these dykes depends on the direction of the littoral currents and on the transport of littoral sediment. The water intake protection at sea has to be adapted to different situations as described in Figures 29 and 30.
Both solutions are not cheap because of the construction cost of the channel. However, they assure an abundant supply of water of very good quality, also protecting the pumping station, and are easy to maintain and to build. This type of water intake is seldom used because of its cost but it is suitable when large amounts of water are needed (the case of the nursery unit). Fig. 29 differs from Fig. 30 only in the presence of a boosting pumping station.
Water intake through a protected pipeline
This is one of the most commonly adopted solutions, even if it has to be carefully evaluated because of the risk of clogging, positioning and its difficult maintenance if the pipeline has a diameter smaller than of 1 000mm. The following figures describe the possible solutions depending on material adopted, skill of the contractor and the preference of the company. If and when possible, always seek a gravityfed water supply to bring water close to the hatchery (Fig. 31 to 34).
Fig. 30  Water is not pumped at a higher level

This type of seawater intake generally provides a good water quality due to the lower levels of sediments and suspended solids. These are intake systems generally with a marked gradient and the three types, described in the figures below, are installed depending on the type of pumping station adopted (Fig. 35 and 36).
In this typology the pipeline is fixed directly to the rocks, without protection, or to concrete blocks or is embedded in the rocks. The design of this water intake type requires the same data as indicated for the cases above to estimate the stress or damage which could be caused by storms to the pipeline and suction protection grating.
Fig. 31  Water is pumped to a higher level at the end of the pipeline

Fig. 32  Water is pumped to a higher level at the end of the pipeline

Fig. 33  Water is pumped to a higher level at the end of the pipeline

Fig. 34  Water is not pumped at a higher level

Fig. 35  types 3a and 3b: the pumping station is equipped with a classical centrifugal pump, with or without priming cap, according to the water level. These water intakes could be:

Fig. 36  type 3c: the pumping station is said to be "wet" and is equipped with a submersible pump. This water intake is a pipeline placed in the sea bed rocks below lowest sea level (lowest point of storm swell or lowest tidal level). This avoids problems during pumping due to a shortage of water at the grating. It also allows a sufficient load to obtain the flow required. 
A third type of water intake is the one placed inside coastal lagoons or manmade ponds connected to the sea through one or more openings in the sand bar (lagoon mouths) or which fill by seeping (percolation) through the sand. In the first case the pumping station has to be placed close to the mouth in order to pump seawater of the best quality. In the second case, the design of the water intake requires a good knowledge of the soil permeability to estimate the maximum amount of water that can be pumped (Fig. 37).
Fig. 37
LEGEND MPS = main pumping station 
The design and construction of water intakes in this type of coastline are closely related to the study of existing littoral transport, which in turn depends on the nature and size of local sediments. This study, together with the structural study, has to be carefully analysed in order to prevent disastrous effects such as the complete and rapid silting of the water intake facilities.
Any study of the local littoral transport is based on the wave/swell spectrum, which gives for each direction the amplitude (range) frequency of the different swells. From this spectrum it is possible to draw the diffraction curves (or swell forecast) of the swells from different directions on the nautical map.
These curves provide estimates of the decrease/increase coefficients (c) of the swells with a significant amplitude as these swells progress toward the shore, and particularly when they reach the shallower water area and break over (the breaking area), at the edge of the shore area. This is where the works have to be built and where almost all the littoral drift takes place.
These curves in (Fig. 38 to 40) also allow estimate of the obliquity angle (a) of the swell in relation to the shore line. The solid flow (see below) along the shore is closely correlated to this angle. Figure 40 shows the variation of (a), the wave efficiency coefficient.
Fig. 38  Sediment littoral transport by currents and waves. Effects of breakwaters on sediment transport 
Fig. 39  Wave spectrum and diffraction plans 
Fig. 40  Curve of sediment transport in relation to wave obliquity angle (a) 
The solid flow is defined as the quantity of solid material maintained in suspension in the water (by wave energy) and passing through the beach per unit of time. It is a flow of material going in and out the beach. To help define the water intake design, the formula of the solid flow produced in the wave breaking area along the shore, is as follows:
where:
Q 
is the quantity of material transported 
K' 
is a coefficient function of the size of sand particles (d) 
L_{o} 
is equal to 1.56 T^{2} 
h_{o} 
is the swell amplitude in the open sea 
T 
is the period of the studied swell or wave period 
g 
is the gravity force acceleration, 9.81 m/s 
h 
is the amplitude at the first breaking swell or (h_{o} x c), with c being a coefficient taken from the "diffraction curves" of the swells when entering the breaking area 
a 
is the angle between the general direction of the waves and the the shore line at the entrance of the breaking area. 
If Q_{1} is the quantity of sand moved into one direction by all the oblique swells of quadrant (1) and Q_{2} the sand quantity moved into the other direction by all the oblique swells of quadrant (2), then, the ratio Q_{1}/Q_{2} (either greater or smaller than 1) will give the direction of the resultant littoral drift as follows:
This is, of course, an estimate of the predominant sediment circulation.
The calculation is based on the Hudson formula generally used to design the elements of the rock cover that should protect the structures from strong wave action. The formula is as follows:
where:
P 
is the weight of the rocks in tonnes 
KD 
is a coefficient (equal to 3.2 for rocks) 
d 
is the specific density of the rocks, in t/m^{3} 
d_{o} 
is the water specific density or about 1 t/m^{3} 
a 
is the angle with the horizontal of the external wall of the dyke 
H_{c} 
is the amplitude of the waves breaking on the structure 
Fig. 41  Crosssection of dyke built with rock in shallow water 
Figure 41 shows the theoretical crosssection of a dyke made of rocks in shallow water. It is at a level higher than is usually considered for Mediterranean situation (3m). It is designed using the above formula.
In this case the intake structures, which when wellpositioned do not interfere with the littoral transport, are designed taking into account only their resistance to the sea storms, and wave energy.
(a) Open canal
It is a structure with a trapezoidal section (see Fig. 42).
However, it can also be built as a structure in reinforced concrete with a rectangular section (as indicated in Fig. 43).
Fig. 42  Open canal crosssection 
The area of the crosssection to be used depends on the maximum water flow required, which is calculated by applying the Bazin formula:
Fig. 43  Square canal crosssection 
where:
Q 
is the water flow in m^{3}/s 
U 
is the water velocity in m/s 
S 
is the wet area of the canal crosssection, in m^{2} 
R 
is the hydraulic radius = S ÷ P, in m 
P 
is the wet perimeter, in m 
g 
is the roughness coefficient of the internal dyke walls 
i 
is the hydraulic slope of the canal in m/m 
Calculations are simplified by using the abacus that is usually found in the most important textbooks on hydraulics.
(b) Pipeline
In this case seawater is conveyed by pipeline to the pumping station either by gravity or by suction.
The size of the pipeline is calculated by applying the ManningStrickler formula, as follows:
Q = US and U = (1 ÷ n) R^{2/3} i^{1/2} 
where:
Q 
is the water flow, in m^{3}/s 
U 
is the water velocity, in m/s 
S 
is the pipeline crosssection area, in m^{2} 
n 
is the roughness coefficient of the pipe internal walls 
R 
is the hydraulic radius = S ÷ P, in m 
P 
is the wet perimeter, in m 
i 
is the hydraulic slope of the pipeline, in m/m 
The abacus for easier calculations and the data on head losses for pipes, elbows, ball valves and grating valves can be found in the most important hydraulic textbooks or are given directly by the PVC pipe and fitting producers.
When deciding on the construction of the seawater intake of a hatchery, two main groups of factors should be taken into account:
the technical ones, which depend on site conditions and the required water flow;
the economic ones, which are related to the cost of the structures to be built.
The final choice may be guided by the following considerations on water intakes and pumping stations:
· Water intake through an open canal
An open canal supplying seawater by gravity directly to the main and secondary pumping stations and with water moving at low speed has the following advantages:
it carries little suspended sediments since it works as an effective settling basin,
it can be easily maintained without any interruption of water flow.
· Water intake protected by converging dykes
From a technical standpoint, this is obviously the best solution. The littoral sediment transport is moderate and as the water intake is situated in a protected area (quiet water and low sediment load), the water can be fed to the main or secondary pumping station either through an open canal or a pipeline. However, because of the high construction costs, this solution can be considered only for very large hatcheries or, for those associated with large pondbased commercial farms.
· Water intake in open sea with pipelines
From an economic point of view, this solution is certainly the most attractive. But, as it is built without any protection and within the wave breaking area, it may transport plenty of sand during swell periods, which means strong wear on the pumps. Moreover, before using water in the hatchery, coarse sediments and suspended solids should be removed through a set of filters.
· Water intake with a screened pipe protected by a pile of rocks
This solution supplies water with less sediment at the pumping station. However, a set of filters to screen suspended solids is required.
· Water intake working by suction or by gravity
It is always better, whenever possible, to have a water intake working by gravity. If properly designed and installed, a pipeline working by gravity does not affect the functioning of the pumps. On the contrary, a pipeline requiring suction is often a source of problems such as air intake, priming failure and mechanical wearing of the pumps.
The pumping station is the structure where pumps are installed. When the hatchery is close to the sea or the difference between the hatchery level and the sea level does not exceed 2 meters, a single pumping unit is enough. In all other cases, it is better to bring seawater first to a reservoir close to the hatchery, from which it is pumped to the different units.
Wells are also used to supply water to the hatchery, but rather as a secondary pumping station, regardless of their type, number and dimensions.
Fig. 44  An aerial view of farm showing the main pumping station (TIMAR, Portugal) 
To bring seawater from the water intake to the hatchery, two types of pumping station exist:
a "dry" pumping station, built as a room outside of the water or watertight in which the pumps are installed functioning in open air;
a "flooded" or "wet" pumping station, built as a reservoir, in which submersible or vertical pumps are installed.
This type of pumping station is usually equipped with horizontal centrifugal pumps but sometimes vertical axial pumps are also employed. As these types of pumps work outside the water, the room where the pumps are to be installed should be located above the highest sea level. To eliminate the need for a priming cap on the pumps, which would be necessary when the suction pipeline is placed above water level, it would also be possible to install the pumps at a slightly lower level, close to sea level. This choice is, however, a risky solution and requires that the lower part of the premises of the pumping stations be watertight.
Fig. 45  “Dry” pumping station. Top and lateral views 
Seawater can be pumped in two ways:
when an open canal brings sea water to the pumping stations, from a sump located at the end of the intake canal, or
directly from the sea through a grating, which can be protected by a structure built of rocks, or else, installed in the open sea without protection.
Fig. 46  “Dry” pumping station below sea level 
This type of main pumping station can be equipped with either submersible pump sets or vertical axial pumps. It consists of a sump communicating directly with the sea through an open canal or a pipeline feeding the sump by gravity. It is important to place the pipeline or the canal below the lowest sea level so that grating and suction line would never be empty. Submersible pumps are installed directly into this sump, together with their lifting and backflow systems. Vertical axial pumps are installed above water level on a metal frame.
Fig. 47  “Wet” pumping station. Section and top view 
In a separate building, close to the pumping station, the following equipment is usually installed:
the control panel of the pumps of the secondary station (main switch, controls, and protection circuits);
the hatchery emergency generator set together with its control panel.
Sometimes, the pumping station can be equipped with a water tower, which should allow water distribution by gravity. This choice is very interesting in order to avoid nitrogen supersaturation, but it is much more complex to run as it requires frequent cleaning to maintain a maximum control of hygienic conditions. In any case, one separate reservoir per circuit has to be contemplated in the hydraulic design.
Fig. 48  Secondary pumping station. Top view and section showing also alternative with submersible pump 
The design of a pumping station should take into account the following parameters:
type of pumps chosen (site conditions, design preference);
maximum water flow required to supply the hatchery at any time;
different water flows required during the annual production cycle;
preferred schedule of utilization of the pumps; this helps in defining the number of pumps to be installed;
hydraulic conditions under which the pumps operate, such as:
 for the main station: lowest pumping level, i.e. the lowest sea level as recorded in front of the hatchery;
 for the secondary station: partial flows to be distributed to the different hatchery sectors and necessary head for the equipment to function properly (which is a function of the hatchery design and type of equipment installed).
On the basis of these data, it is then possible to proceed with the design of the pumping stations by:
determining the number of pump sets required and their characteristics;
specifying the characteristics of the main and secondary pumping stations.
Generally the number of pump sets to be installed varies from two to four as a compromise between security (minimum two pumps) and economy (maximum of four pumps to cut running costs by fractioning the use) as follows:
two sets with a unit flow Qu = Qmax and the installed flow Qi = 2 Qmax. This solution does not allow the adaptation of the flow to the changing needs in water supply of the hatchery, but reduces investments;
three sets with a unit flow of Qu = 1/2Qmax and the installed total flow Qi = 1.5 Qmax. This solution allows a certain flexibility, but requires more investment;
four sets with two sizes of unit flow: Qu = 1/2Qmax for two sets and Qu = 1/4Qmax for the two other sets. Then the total flow Qi = 1.5 Qmax, as in the previous case, but with a better possibility of adapting the flow to the real needs. Of course, this solution requires even more investment than the previous one.
(In which Qu is the unitary flow of one pump. Qmax is the maximum flow needed by the hatchery operations. Qi is the maximum flow capacity when all the installed pumps are running).
For each unit in the hatchery, at least two pumps should be installed, one to be in operation while the other is kept on standby. This is, of course, the minimal configuration acceptable. If more pumps are desired the same consideration on the balance between investment and running costs discussed in the paragraph above should be taken into account. Spare parts should be readily available for each type of pump, and should be regularly replaced when used.
The final solution should be chosen taking into account factors like reliability of equipment and easy utilization, without neglecting the economic aspects related to investment and operational cost of the equipment. The following considerations can be of some assistance:
Because of the type of liquid to be transported (seawater) and the salty environment in which the pumps operate, the pump type that gives best results are the submersible pumps treated against marine corrosion. The reliability of such pumps comes from the fact that they are built very carefully to work continuously underwater. Due to their easy installation and maintenance, submersible pumps offer real advantages due to the practical mechanisms used for their assembly and dismantling. The very reduced chance to produce nitrogen oversaturation in the water pumped is another positive characteristic of this type of pump. Vertical and horizontal pumps are also frequently used as they are cheaper and are also easy to maintain. These pumps are also suitable for work with seawater but some additional precautions have to be taken:
the calculation of pump size and engines has to be very precise, in particular, for the total head. These pumps have a more limited range and working out of the optimal curve would affect their efficiency and can quickly destroy the impeller;
materials have to be carefully chosen between marine bronze, stainless steel AISI 316, titanium and plastic (in case of pumps used for closed and semiclosed circuits, bronze should not be employed due to the risk of contamination with metallic ions);
all joints have to be frequently checked in order to avoid nitrogen oversaturation in the water pumped.
Well water of good quality, due to the stability of its physicochemical characteristics, is an asset not to be missed by any hatchery. A well can be defined as a structure built in the ground, able to reach the water table and from where water can be pumped out. Its shape is usually cylindrical, developing along a vertical axis, and consists of two parts: an external wall, built from superimposed cylindrical elements, in concrete, plastic or other materials forming the wellcasing, and an internal filter.
The data required for a proper well design are divided in three main categories:
topographic data,
geotechnical data, and
hydraulic data.
Topographic data are essential to have a correct idea of the land in terms of its elevation, and to evaluate the variations of the water table levels.
Geotechnical data are fundamental for the construction of the well and have an impact also on construction costs. In fact, as the well casing will penetrate up to the water table, it is necessary to know in advance the characteristics of the soil. In particular, it is necessary to know:
the soil texture, to determine the size composition of the natural gravel filter;
the permeability of the soil, which is characterized by the permeability coefficient, a fundamental element to determine the flow of the well.
The soil texture is determined in the laboratory on the basis of samples taken by means of augers at various depths of the water table.
The permeability coefficient is determined by percolation tests carried out on the spot at different depths of the water bearing layers. This is the only way to provide a correct value for this coefficient, as laboratory tests on samples taken from the spot are usually unreliable. This is due to the difficulty encountered in bringing sand samples from waterbearing layers to the laboratory, without affecting their characteristics.
Hydraulic data essential to calculate wells are:
the well upper level, which is generally close to the average sea level in alluvial zones, with slight variations created by the sea level variations (alternated and out of phase);
the strength of the water table, i.e. the thickness of the waterbearing sands, and the eventual position, if it exists, of:
 the ceiling of the water table, formed by an impermeable layer;
 the bottom of the water table, also formed by an impermeable layer;
the level variation in the water table where pumping takes place.
To evaluate whether a well is feasible or not, two different types of analysis are required:
the estimation of the well flow on the basis of geotechnical and hydraulic data collected;
the calculation of the well structure in terms of civil works and filtering material.
For this estimation the following formula is used:
where:
Q 
flow in m^{3}/sec 
K 
permeability coefficients in m/sec 
H_{o} 
height of the exploited water table 
H_{p} 
height of water in the well (stabilized level for a given Q) 
R_{a} 
external limit of the draw down curve 
r_{p} 
well radius 
ln 
neperian logarithm = 2.3 decimal logarithm 
From this formula it can be deducted that the well flow is:
1. directly proportional to:
 the coefficient K;
 to a greater degree, to the variation of the water level in the well (H_{o}^{2}  H_{p}^{2});
 to a lesser degree, to the radius of the well: rp (ln rp).
2. inversely proportional to:
 the radius of the area where the well activity is felt (R_{a}) in the water table, this radius of activity being in itself inversely proportional to K, which increases still the importance of K in the flow calculation.
Finally, and to be able to appreciate the importance of the various parameters intervening in the calculation of the flow capacity of a well, two examples are given below:
Parameter 
Example 1 
Example 2 
K (m/sec) 
0.0001 or 10^{4} 
0.0001 or 10^{3} 
Ho (m) Hp (m) 

Ra (m) 
50.00 
50.00 
rp (m) 
2.00 
1.00 
Q (m^{3}/sec) 
The analysis of these two examples highlights the importance of the coefficient K, depending on whether we are dealing with fine sand with limited permeability or with more permeable coarser sand. Even with a hydraulic load and a well diameter reduced to half, the flow is considerably increased.
Q_{2} 19 l/sec with 
K = 1 x 10^{3} 

Dh_{2} = 4m 

rp_{2} = 1m 
Q_{1} 7.8 l/sec with 
K = 1 x 10^{4} 

Dh_{1} = 8m 

rp_{1} = 2m 
It may be possible to use seawater wells when a hatchery is to be established on sandy or nonhomogeneous soils. They are easy to build and guarantee a constant water supply. Areas adequate for well construction are usually formed by sediments which go from fine sand to silt and for which the permeability coefficient K varies, in m/sec, from 1 x 10^{3} for coarse sand to 1 x 10^{9} for silt. In the case of rocky coastlines, apart from some very unusual locations, it is better to consider from the planning stages a standard pumping station, drawing water directly from the sea.
The strength of the littoral water tables is generally weak, usually only a few meters. This implies that the flow that could be obtained is modest, being in the order of a few litres per second rather than tens of litres per second.
As a final consideration, it is very important to bear in mind that in case of large wells we should avoid pumping more than 30 litres per second in order to reduce the movement of fine particles all around the well area and the resulting quick clogging. Moreover, it has to be borne in mind that wells are often temporary constructions with an expected life span of 510 years. Therefore locations for new wells should be found in due time.
Pipelines and canals (open canals, gutters) are used to carry seawater to and from the hatchery. Different materials are used for pipelines, depending on their use and on the hatchery sectors where they will be installed. Materials used for the piping outdoors could be protected steel, concrete and fibreglass, while polyethylene (PE) and polyvinyl chloride (PVC) are used for piping inside the hatchery, where nontoxic materials are required. Canals and gutters are mostly made of reinforced concrete, prefabricated concrete, bricks, PVC and metal.
In general, for pipelines that have to work under pressure, PVC or PE are the materials used more frequently, while for hatchery systems in which liquids flow by gravity, open channels or PVC and PE pipes are common. Inside the hatchery all the water circuits are normally built using PVC piping, which is the more flexible and easy to use material in terms of installation and repairs, and also because of the variety of existing PVC fittings such as valves, elbows, fast joints, etc.
Pipelines bringing water to the main pumping station, either using suction or gravity, are generally made of:
coated steel pipes (inside and outside), with a special protection coat to limit corrosion;
concrete pipes with a metal core;
PE pipes covered with concrete or ballasted with concrete blocks to counter buoyancy.
The working pressure limit of the pipeline should be at least 6 bars, but it is better to use piping that could work at 10 or 12 bars pressure, although it may be slightly more expensive. It is preferable to use oversize pipes so that they can stand severe working conditions in the sea. All the sections and fittings of the pipeline are gathered on site; then they are welded/assembled after having prepared the trenches in the sea bottom and on land. Immediately after, the pipeline is pushed into the sea on floats and is lowered into the trench. This work is often considered of secondary importance, but it is often a key element for the success of the hatchery and farm. Saving money on this installation, materials or studies could severely affect all farm sectors later.
This hydraulic connection is generally made using the following pipes:
PE pipes of the series N.P. 6 (nominal pressure 6 bars) for service under low pressure or of the series N.P. 10 (nominal pressure 10 bars) for service under medium pressure;
PVC pipes of series N.P. 6 or N.P.10. They are usually 6 m long and they are joined either by solvent welding, or by using flanges or adapter fittings; or
concrete pipes, generally 5 m long and assembled by adapter fittings.
Water is distributed inside the hatchery by pipeline systems that are either suspended under the roof or running on the floor. The first solution is preferable to limit the risk of possible damage to the pipes and to facilitate movements in all areas, whilst the second option is frequently an easier and cheaper solution.
Pipelines for internal distribution of water are made of PVC and are usually of small diameter (31 to 200mm), assembled by solvent welding or threaded sockets, or fast joints. In the case of long pipelines, fast joints are preferable since a piping system that can be easily dismantled can also be thoroughly cleaned and disinfected. Standard pipes usually available are 6 m long.
Water is drained from the hatchery through a network of secondary channels/gutters that convey the effluents from the various tanks to the main drainage canal. The internal gutter network is usually made of concrete, or of light PVC, and it can be covered with:
removable reinforced concrete slabs, which are essential for PVC gutters;
wooden boards, or metal slabs coated for protection against corrosion.
In summary, each pipe should be chosen and used according to specific needs. Stainless steel or concrete pipes should be used when a strong mechanical resistance is needed. PE pipes should be preferred when mechanical resistance is not the sole factor to consider, and the pipe is going to be exposed to atmospheric conditions. Finally, PVC pipes are generally used for internal pipelines as this material is not toxic and has a very small roughness coefficient, which allows the use of smaller pipes for accurate calibration of the water flow to the various outlets.
The final choice for materials should be based on an accurate survey to identify local manufacturers, to evaluate the quality and cost of available materials and to locate potential contractors with the knowhow and equipment necessary to put together the hydraulic systems.
Four types of information are essential for the correct sizing and design of a pipeline network:
1. the roughness coefficient of the material chosen;
2. the water flow required;
3. the internal water velocity;
4. the predefined head loss produced by the line lenght/fittings and by the equipment interposed.
The design of a closed pipeline is made using the ManningStrickler formula, (also applicable for water transfer by gravity), which is as follows:
Q = US = U (K R^{2/3} i^{1/2}) 
where:
Q = 
water flow in m^{3}/s 
U = 
water speed in m/s 
S = 
wet section area in m^{2} 
K = 
head loss coefficient = 1/n 
n = 
roughness coefficient 
R = 
average radius = S/P 
P = 
wet perimeter in m 
i = 
hydraulic slope in m per m 
Using this formula it is easy to calculate any of its elements, for example:
knowing section and flow, it is possible to determine the hydraulic slope;
in case of a fixed slope, it is possible to determine how much water can pass through the pipe;
viceversa, knowing the flow it is possible to determine the dimension of the pipe.
Sometimes, to simplify calculations, an abacus or graphic methods can also be used.
The head loss coefficient generally considered for smooth pipes under pressure is K = 95 which corresponds to a roughness coefficient n = 0.0105. For PVC pipes, K is about 120 and n = 0.0083.
To calculate the total head charge H in m (the height necessary to transfer a given flow), we can use the formula: H = i L, where i is the hydraulic slope (in m/m) and L the length (in m) of the pipeline. To this the sum of head losses due to the pipeline fittings (grating, elbows, valves, etc.) should be added.
Pipe fittings are frequently expressed in equivalent length of pipe L_{1}. L_{2}, L_{3}, etc. The length of the pipeline L is lengthened by the sum of these equivalent lengths, so that finally the formula would be:
H = i (L + L_{1} + L_{2} + L_{3} +....) 
Example:
Thus, the total head charge necessary: H = 0.012 x 500 = 6 m or 0.6 bar 
In many cases, choosing the size of a small PVC/HDAD pipe is done by consulting a simple graphic abacus provided by the pipe manufacturer. This way it is easy to determine the pipe sections, and it also gives often the opportunity to choose the correct internal section and to evaluate head losses (in m/m).
It is very important to bear in mind that this way of calculating pipeline size is absolutely empirical and is easily applicable to pipes working under pressure. However, when the pressure applied consists of only gravity, a better evaluation is needed.
For overflow weirs with a free water fall, the flow calculations are based on the following formula:
where:
q 
is the water flow per meter of weir(in m^{3}/s) 
m 
is a coefficient (close to 0.45) 
Hd 
is the water head above the crest of the weir (in m) 
g 
is the gravity (9.81 m/s) 
For a weir that is L meters wide, total flow .
Knowing the weir width (L, in m) and the flow (Q, in m^{3}/s), it is possible to calculate the difference in level (Hd) between the weir crest and the water level upstream as:
The hatchery gutter network consists of rectangular canals of small size made usually of reinforced concrete, and either built on the spot or assembled using prefabricated sections.
Generally, the main drainage canal is a ditch of trapezoidal section, not covered, with a gentle slope and oversized for the flow expected.
The flow capacity of any type of canal is obviously related to its section and can be calculated with the Bazin formula:
where:
Q 
is the water flow in m^{3}/sec 
U 
is the water velocity in m/sec 
C 
is the Bazin coefficient = 
R 
is the hydraulic radius (in m) = S/P 
S 
is the wet section area in m^{2} 
P 
is the wet perimeter in m 
g 
is the Bazin roughness coefficient 
i 
is the hydraulic slope in m/m 
Note:
Bazin roughness coefficient varies as follows:
concrete smooth surface 
= 0.06 
surface in stones or bricks 
= 0.16 
surface in masonry 
= 0.45 
embankment 
= 0.85 
ordinary embankment 
= 1.30 
rock embankment 
= 1.75 
An abacus can be used to determine the hydraulic slope necessary for a given flow Q, with the crossection selected and the roughness coefficient known. It also allows to determine the difference in level between the channel upstream and downstream (Dh)
Referring to the above mentioned formulas and principles, this section gives some examples of practical calculations for pipe inlets and outlets in a marine finfish hatchery.
Let's assume that the internal network is made of three different circuits: A, B and C.
Fig. 49  Hatchery circuits MF = mechanical filter EX = plate exchanger = pumping station 
 Length: 210 m
 Equipped with a pump A: maximum flow Q_{A} = 10 l/s = 0.010 m^{3}/s
 Water velocity: U = 1.20 m/s
 Pipeline: rigid PVC 10 bars
Thus the theoretical section is given as S = Q_{A }÷ U = 0.010 ÷ 1.20 = 0.008m^{2}
 Selected diameter: N.D. Æ 100 mm, which means 98.8/110 mm
With a real section of 0.00785 m^{2} [Calculated as (98.8/2*1/1000)^{2} p] and the real velocity as U = 0.010 ÷ 0.00785 = 1.275 m/s
 Interposed circuit equipment: considering Q = 10 l/s
filter F1, which gives a head loss of 10m
filter F2, which gives a head loss of 20m
UV sterilizer, which gives a head loss of 5m
Secondary water distribution pipes to the tanks, with the following characteristics:
 Length: 50 m
 Equivalent flow (considering homogeneous water distribution to the tanks):
 Water velocity: 1.0 m/s
 Pipe type: rigid PVC 10 bars
 Selected diameter: N.D. ø 80 mm which means 80.6/90 mm
With a real section: S = 0.0395 m^{2} and the real velocity: U = Q_{e} ÷ S = 0.006 ÷ 0.0395 » 1.05 m/s
 Length: 210 m
 Circuit equipped with a pump B; maximum flow: Q_{B} = 50 l/s=0.050 m^{3}/s
 Water velocity: U = 1.20 m/s
 Pipeline: rigid PVC 10 bars
Thus the theoretical section is given as S = Q_{B} ÷ U = 0.050 ÷ 1.20 » 0.04 m^{2}
 Selected diameter: N.D. Æ 250 mm which means 224.2/250 mm
 Real section: S = 0.0395 m^{2}
 Real velocity: U = Q_{B} ÷ S = 0.050 ÷ 0.0395 Æ 1.265 m/s
 Interposed circuit equipment: considering Q = 50 l/s
filter F1, which gives a head loss of 15 m
secondary water distribution pipes to the tanks with the following characteristics:
 Length: 50 m
 Equivalent flow (considering an homogeneous water distribution to the tanks):
 Theoretical velocity: 1.0 m/s
 Pipe type: rigid PVC 10 bars
 Selected diameter: N.D. ø 200mm, which means 179/200 mm
With a real section: S = 0.025 m^{2} and the real velocity: U = Q_{e} S = 0.029 ÷ 0.025 » 1.16 m/s
Sea water reservoir feeding the tanks under the following conditions:
 Inlet flow coming from circuit B: 3 l/s = 0.003m^{3}/s
 Maximum flow to the tanks through circuit C: Q_{C}= 3 l/s = 0.003 m^{3}/s
 Capacity: half an hour of flow, which means: (3600 x 3)
 Length: 4 m
 Maximum flow: Q_{C}= 3 l/s
 Water velocity: U = 1.20 m/s
 Pipeline: rigid PVC 10 bars
Thus the theoretical section is given as S = Q_{C} ÷ U = 0.003 ÷ 1.20 » 0.04 m^{2}
 Selected diameter: N.D. ø 75 mm which means 63.2/75 mm
With a real section: 0.0031 m^{2} and a real velocity: U = Q_{C} ÷ S = 0.003 ÷ 0.0031 » 1 m/s
 Water distribution: considering Q = 3 l/s
 Length considered: 50 m
 Max. flow (considering an homogeneous water distribution to the tanks):
 Theoretical velocity: 1.0 m/s
 Pipe type: rigid PVC 10 bars
 Selected diameter: N.D. ø 75 mm which means 63.2/75 mm
With a real section: S = 0.0031 m^{2} and the real velocity: U = Q_{e} ÷ S = 0.00175 ÷ 0.0031 » 0.565 m/s
To finalize the design of the water inlet system, it is necessary to determine the sum of the head losses due to the equipment installed on the circuit, the friction of the water in the pipelines and the energy lost to move the water to obtain the flow required.
The head losses i are calculated using the ManningStrickler formula:
Q = U S where
U = K R^{2/3} i^{1/2} and therefore
i = U2 ÷ K^{2} R^{4/3}
Adding this hydraulic load of the pipeline system to the final load to deliver at the end of the circuit, allows the calculation of the necessary total head of the pump.
Use an abacus to determine easily “i” knowing U, K and R (or diam. Æ).
(a) Residual load necessary at the end of the distribution lines in the secondary circuits: 5 m, which means a line of load = (+4.00) + 5.00 = (+9.00) m
(b) Head losses into secondary distribution circuit with (see (a) above) where:
L = 50 m
Q_{e} = 6 l/s
U = 1.05 m/s
K = 100 (n = 0.01)
Æ internal ~ 80 mm or 0.08 m
R = Æ ÷ 4 = 0.02 m
and Dh = 0.02 x 50 = 1 m
which means a load line in A3 = (+9.00) + 1.00 = (+10.00) m
(c) Head losses for primary internal circuit (see (a) above), where:
L = 60 m
Q = 10 l/s
U = 1.275 m/s
K = 100 (n = 0.01)
Æ internal ~ 100 mm or 0.10 m
R = 0.10 m ÷ 4 = 0.025 m
and Dh = 0.022 x 60 = 1.34 m
If local head losses are:
filter F1 = 10 m
filter F2 = 20 m
UV lamp = 5 m
for a total of 35 m, then total head losses = 1.34 m + 35 m =
36.34 m
which means a load line in A2 = (+10.00) + 36.34 = (+46.34)
m
(d) Head losses for primary external circuit (see (a) above), where:
L = 150 m
Q = 10 l/s
U = 1.275 m/s
K = 100 (n = 0.01)
Æ internal ~ 100 mm or 0.10 m
R = 0.10 ÷ 4 = 0.025 m
and Dh = 0.022 x 150 = 3.30 m, which means a head
loss at the outlet of the pump equal to (+46.34) + 3.30 = (+49.64) m
(a) Residual load necessary at the end of the distribution lines in the secondary circuit: 5m which means a line of load = (+4.00) + 5.00 = (+9.00) m
(b) Head losses into secondary distribution circuit (see (a) above), where:
L = 50 m
Q_{e} = 29 l/s
U = 1.16 m/s
K = 100 (n = 0.01)
Æ internal ~ 180 mm or 0.18 m
R = (0.18) ÷ 4 = 0.045 m
and Dh = 0.0084 x 50 = 0.42 m
which means a loadline in B3 = (+9.00) + 0.42 = (+9.42) m
(c) Head losses for internal primary circuit (see (a) above), where:
L = 60 m
Q = 50 l/s
U = 1.265 m/s
K = 100 (n = 0.01)
Æ internal ~ 224 mm or 0.224 m
R = 0.224 ÷ 4 = 0.056 m
and Dh = 0.0075 x 60 = 0.45 m
If local head losses for filter F1 = 15 m
then total head losses = 0.45 + 15 = 15.45 m
and the load line in B2 is (+9.42) + 15.45 = (+24.87) m
(d) Head losses in the external primary circuit (see (a) above), where:
L = 150 m
Q = 50 l/s
U = 1.265 m/s
K = 100 (n = 0.01)
Æ internal ~ 224 mm or 0.224 m
R = 0.224 ÷ 4 = 0.056 m
and Dh = 0.0075 x 150 = 1.125 m
This means a load line in B1 at the outlet of the pump equal to (+24.87) + 1.125
= (+25.995) m.
(a) Feeding pipe to the reservoir:
 load at the pipe feeding the reservoir: (+9.42) m
 load at the top of the reservoir where the water is distributed: (+5.50) m
 head losses into the distribution pipe to the reservoir (see (a) above), where:
L = 4 m
Q = 3 l/s
U = 1.0 m/s
K = 100 (n = 0.01)
Æ internal ~ 53 mm or 0.053 m
R = 0.053 ÷ 4 = 0.01325 m
and Dh = 0.032 x 4 = 0.13 m
which means a load necessary to distribute 3 l/s at the top of the reservoir
equal to (+5.50) + 0.13 = (+5.63) m.
(b) Distribution pipe
 Minimal load at the starting point of the pipe at the reservoir (with the
most unfavourable conditions):
(+4.25) m
 Head losses into distribution pipe (see (a) above), where:
L = 50 m
Q_{e} = 1.75 l/s
U = 0.565 m/s
K = 100 (n = 0.01)
Æ internal ~ 63.2 mm or 0.0632 m
R = 0.0632.. 4 = 0.0159 m
and Dh = 0.008 x 50 = 0.40 m
which means a load line at the end of the distribution pipe equal to (+4.25)
 0.40 = (+3.85) m.
Thus, the load available at the end of this pipe situated at
(+3.50 m) is:
(3.85)  (3.50) = 0.35 m, a very low load which will make use
of large valves compulsory.
There are several possible solutions to design the main outlet system. Three different ways are provided as example:
dug as a ditch in the ground, of triangular section and with a slope ratio of 2:1;
built in concrete or reinforced concrete, of rectangular section, open;
buried, using a large diameter pipe in reinforced concrete or centrifuged concrete.
In all cases, let us assume that the main gutter should allow the drainage (in good hydraulic conditions) of around 60 l/s with a total length of 150 m.
The calculation is based on various formulas according to the kind of system used (open or closed):
· For an open channel use the Bazin formula:
Q = US where:
Note:
Bazin roughness coefficieng varies as follows:
concrete smooth surface 
= 0.06 
surface in stones or bricks 
= 0.16 
surface in masonry 
= 0.45 
embankment 
= 0.85 
ordinary embankment 
= 1.30 
rocks embankment 
= 1.75 
The different values for C are thus function of g
· For a closed pipeline: use the ManningStrickler formula:
Q = U S where
U = K R^{2/3} i^{1/2}
with K = 90 for a free flow in concrete or reinforced concrete pipelines and K = 95 for a flow under pressure in concrete or reinforced concrete pipelines.
These two formulas are applied in the following examples where the relevant calculations are shown. However, an abacus for each formula can also be used for faster approximate calculations.
g= 1.30
Q = 60 l/s or 0.06
m^{3}/s
Section: type I
 length: 150 m
 bottom width: nil
 side slope: 2:1
 water height (downstream): 0.50 m
 wet section (downstream): S = 0.50 m^{2}
U = Q ÷ S = 0.06 ÷ 0.50 = 0.12 m/s
Wet perimeter
P = 2.24 m
Therefore:
R = S ÷ P = 0.50 » 2.24 ² 0.22 m
C = 23.1
i = U^{2}
÷ C^{2} R = (0.12)^{2} ÷ (23.1)^{2} (0.22) =
0.000123 m/m
Dh = 0.000123 m/m x 150 m ~ 0.02
m
which means, following the scheme, the following water levels:
 downstream: (+0.50 m)
 upstream: (+0.50) + 0.02 = (+0.52 m)
Fig. 50  Aerial view of water circulation in hatchery and farm (TIMAR, Portugal) 
g = 0.06 m
Q = 60 l/s or 0.06
m^{3}/s
Section: type II
 length: 150 m
 width: 0.30 m
 water height (downstream): 0.30 m
 wet section (downstream): 0.09 m^{2}
U = Q ÷ S = 0.06 ÷ 0.09 = 0.67 m/s
Wet perimeter
P = 0.90 m
Therefore:
R = S ÷ P = 0.09 ÷ 0.90 » 0.10 m
C = 73.1
i = U^{2} ÷
C^{2} R = (0.67)^{2} ÷ (73.1)^{2} (0.10) » 0.00084 m/m
Dh =
0.00084 m/m x 150 m ~ 0.125 m
which means, following the scheme, the
following water levels:
 downstream: (+0.50 m)
 upstream: (+0.50) + 0.125 = (+0.625 m)
K = 90, "free flow"
Q = 60 l/s or 0.06 m^{3}
/s
Section: type III
Æ D = 0.50 m
S = tD^{2} ÷ 4 = 3.14 (0.50)^{2} ÷ 4 = 0.196 m^{2}
P = tD = 3.14 * 0.50 = 1.57 m
R = S/P = 0.196 ÷ 1.57
Therefore:
Dh = 0.000178 m x 150 m = 0.03 m
which means pipeline levels as follows:
 downstream: (+0.50 m)
 upstream: (+0.50) + 0.03 = (+0.53 m)
Fig. 51  Water outlets 
Fig. 52  Main pumping station using axial propeller pump (source: ETEC catalogue) 
For our purposes it is possible to define a pump as a device able to increase the mechanical energy of a liquid, or in more practical terms, a machine able to push a fluid from one point to another.
Pumps are frequently made of cast iron or 304 stainless steel. However, in marine hatcheries, because of corrosion problems linked to the use of seawater as pumped liquid, these two metals are not suitable and other materials like bronze, 316 stainless steel or plastic are preferable. When only cast iron or 304 stainless steel pumps are available, they have to be well protected outside and inside with an epoxy coating. For semiclosed circuits, in which pollution by metal ions should be avoided, plastic pumps are strongly recommended.
Pumps are generally driven by electric two/threephase engines or, in sites where electricity is not available, by diesel engines. The latter are mostly employed in the case of low pumping heads and large water flows and are only used if the hatchery is linked to a landbased growout farm.
Although a large number of pumping systems exists, it is possible to group them into three categories as follows:
Turbine (regenerative) pumps, where a rotating impeller equipped with paddles or blades transmits kinetic energy to the fluid. This is the type most commonly used in a hatchery.
Water lifts, such as the Archimedes screw and airlifts. These types of pumps are seldom used in a marine hatchery.
Displacement (volumetric) pumps, where the fluid transport is done through successive variations of capacity, the pumping being done through the alternate filling and emptying of an enclosed volume. These volumetric pumps are used only in very large hatcheries and ongrowing facilities for cleaning procedures or to transfer live food.
In view of the above, the following sections will only deal with electrical turbine pumps.
Turbine pumps are rotative and usually have a rigid connection to the engine. They are simple, relatively small, light and easy to maintain. According to the type of impeller used and the way it works, turbine pumps used in hatcheries and aquaculture farms can be of three types:
centrifugal pumps;
centrifugalpropeller pumps;
propeller pumps.
Centrifugal pumps are designed for medium water flows and great heights, while propeller pumps raise very large flows at low heights (only a few meters).
Fig. 53  Centrifugal pump 
Depending on construction criteria adopted, turbine pumps can also be classified in the following categories according to:
axis position: horizontal, vertical or leaning axes;
number of impellers present: mono or multistage;
pressure produced: low, medium or high pressure.
As far as position of the turbine pump in relation to the water level in the sump, pumps can be classified as follows:
a surface pump, when working completely outside the water;
immersed, when the pump is underwater but the engine is outside the water;
submersed, when both pump and engine are underwater.
Hatcheries are generally equipped with centrifugal turbine pumps, with a horizontal or a vertical axis. They are usually monostage and produce low or medium pressures. They can be surface, immersed or submersed pumps depending on the sites.
Concerning fluids:
type and origin of the fluid;
maximum flows needed;
working conditions;
hydraulic conditions at water intake and delivery points.
Information concerning pumps:
The total head (TH) of a pump is the pressure difference in meters of liquid column (MLC) between the suction and the discharge points. It is related to three elements:
 geometric head (GH), according to the hydraulic conditions defined above;
 pressure losses at suction point (J SUC), equal to the pressure (in MLC) necessary to overcome pressure losses in the suction pipe;
 pressure losses at discharge point (J DIS), equal to the pressure (in MLC) necessary to overcome pressure losses in the discharge pipe; these also depend on the fluid velocity and on the different fittings installed in the circuit.
If suction and discharge take place under atmospheric pressure, total head is calculated as TH (in MLC) = GH + J SUC + J DIS
If pressures (in kg/cm^{2}) at suction and discharge points are different, say P_{1} at suction and P_{2} at discharge, you can refer to a homogeneous system by using instead:
at suction (P_{1} ÷ g) x 10 (in MLC), and
at discharge (P_{2} ÷ g) x 10 (in MLC),
where g is the density (in kg/dm^{3}) of the pumped liquid, which is close to 1 for sea water. The above formula then becomes:
TH (in MLC) = GH + JSUC + J DIS + 10 [(P_{2}  P_{1}) ÷ g] 
Maximum suction height for centrifugal pumps: In theory, if a vacuum is created inside a vertical tube immersed in water by eliminating the atmospheric pressure at its upper end, the water will reach a height in the tube equal to the atmospheric pressure at that location, in MLC. At sea water level, this means a height of 10.33 m. In general, for an altitude A (in m), the height reached by the water inside the tube is reduced to 10.33 m  0.012 A.
In practice, however, the water height obtained by suction using a centrifugal pump is lower because part of the available pressure is needed to overcome the pressure losses in the suction pipe and to give the desired velocity to the fluid. To avoid pump cavitation (the formation inside the fluid of vapour bubbles), the absolute pressure at the pump inlet should never drop below the value of the vapour pressure corresponding to the temperature of the fluid to be pumped. To ensure that the pump will run safely, the pressure at the pump inlet should remain well above the vapour pressure of the fluid. The vapour pressure (in MLC) for sea water at 20°C is around 0.20 m. But it can reach as much as 1.3 m at 50°C at sea level.
Fig. 54  Section of submersible pump 
The suction performance of a pump, taking into consideration its technical characteristics and the way it is installed, is determined by the net positive suction head (NPSH). Two types of NPSH exist:
1. the available NPSH, which is the value of the absolute pressure measured at the pump intake considering the type, materials and equipment used for construction of the intake, such as pipe diameter, type and other fittings;
2. the required NPSH, which is a set of values given by the manufacturer for each type of pump and for a given speed of rotation of the engine, and which is shown as a curve relating NPSH to pump outflow.
For a pumping installation to work properly, it is necessary that the available NPSH is greater than the required NPSH by a few decimetres. The value of the available NPSH for water supply under depression in a free water basin such as the sea, is equal to: 10 m  (GH + J SUC).
Figure 55 gives an example of available and required NPSH curves. The operating point of the pump must be situated to the left of the vertical line passing by the intersection of the two curves, so that available NPSH be greater than the required NPSH.
Fig. 55  Example of available and required NPSH pump curves 
Rotation speed of centrifugal pumps:
The rotation speed of a centrifugal pump affects potential water flow (Q), total head (TH), and consumption of energy (P). If this rotation speed varies from V_{1} to V_{2}, the following three relationships exist:
1. (Q_{2} ÷ Q_{1}) = (V_{2}
÷ V_{1})
2. (TH_{2} ÷ TH_{1}) =
[(V_{2})^{2} ÷ (V_{1})^{2}]
3.
(P_{2} ÷ P_{1}) = [(V_{2})^{3} ÷
(V_{1})^{3}]
Operating point of a pump
The two main information elements needed for evaluating the operating point of a pump connected to a given system are flow (Q) and total head (TH) produced by the pump. To identify this operating point, you should superimpose on a graph the QH curve of the pump, and the characteristic curve of the pipeline, obtained by adding geometric head and total pressure losses.
The intersection point (S) of these two curves determines the operating point of the pump.
This clearly shows also that the operating point of the pump moves if the characteristic curve of the pipeline changes. For example, when the valve of the discharge pipe is partially closed, pressure losses increase. Similarly, if the pump is changed by a different one, there will be a new QH curve and thus a new position for the operating point S on the characteristic curve of the pipeline.
Pump curves:
There are three important and characteristic curves for a pump:
1. The outflow to height curve or QH curve, which shows the relationship between total head (TH) produced by a pump in relation to the flow (Q);
2. The efficiency curve, which relates efficiency of the pump to flow (Q). It always shows a peak value (optimum efficiency). It is best to use the pump around this peak value, which thus helps define the QH relationship to be respected. The efficiency of centrifugal pumps does not exceed 0.80, while for propeller pumps it can reach 0.90;
3. The brake power curve or PQ curve, which relates brake power (P) to flow (Q).
Grouping centrifugal pumps
The grouping of centrifugal pumps may be needed for two reasons: to increase the available flow or, to produce a greater head. In the first case, the pumps should be grouped in parallel. To obtain a greater head, group the pumps in series.
Fig. 56  Typical pump curves (Source: J. Fletcher, Zoeller Company) 
The calculation of the pumping system can be carried out after the data mentioned in the previous sections on pipelines and pumps characteristics have been collected. The steps required are as follows:
1. Define the characteristics of the hydraulic system(s) and calculate the following:
(a) geometric head (GH): the difference between the maximum level of discharge and the water level in the sump from where water is pumped.
(b) total pressure losses (J TOT): consider the maximum flow necessary for the circuit. Include pipe losses, losses due to special fittings (elbows, Tjunctions, valves, etc.), and losses due to all equipment installed in the hydraulic system. These partial pressure losses should be measured in metres of liquid column (MLC) and can be determined by using graphs or manufacturer's technical specifications. When water is pumped by suction it is very important to ascertain that the selected diameter of the suction pipeline fulfils the condition: (available NPSH) > (required NPSH), as described above.
2. Transfer the characteristic curve of the pipeline (defined by TH = GH + J TOT) onto a graph.
3. Transfer the three characteristic curves of the pump (QH, Rdt, PQ) explained above onto the same graph.
4. Finally, define the point S at the intersection of QH curve and characteristic curve of the pipeline. This gives the operating point of the pump, which should be located close to the maximum of the pump efficiency curve (Rdt).
The total flow necessary is obtained by increasing the number of pumps, as a multiple of the flow of a single pump. Any modification of the curves mentioned above will cause a change in the operating point at the expenses of pump efficiency.
The power (P in CV) absorbed by a pump can be estimated by the following formula:
where:
Q is the flow in m^{3}/s
TH is the total head in m
Rdt is the pump efficiency as given in the technical specifications by the manufacturer.
To make the correct choice when installing a pumping system, it is also very important to consider the three following points:
1. Use the equipment that best suits local conditions, and is the most reliable and easiest to maintain;
2. Choose the equipment necessary to guarantee a continuous supply of water to the hatchery, making sure that this supply is not undersized as water availability is essential for hatchery security;
3. Evaluate the investment cost in relation to the two previous points. Beware of proposals that appear to be very convenient at first but can be very expensive.
Such a choice should be made for each installation but it can be guided in general by the following further considerations derived from experience.
For a hatchery with a flow requirement varying from a few litres per second to up to 100 l/s, and with a maximum total head of 40 m, the best choice appears to be a monostage turbine centrifugal pump.
When water is pumped directly from the sea by suction and without the use of a sump, the only pumps adapted to this job are the horizontal monostage centrifugal pumps.
When water is pumped from a sump, the choice has to be made between a dry, a semisubmersed and a submersed centrifugal pump. The final choice should be made according to the above considerations on service and cost rather than strictly from a technical point of view, even though the submersed pumps appear to be more suitable for this kind of installation. They are much easier and quicker to assemble and dismantle.
Main pumping station: let us consider one main pumping station that should deliver a nonstop variable flow, reaching a maximum value Qmax:
1. the minimum equipment for this installation is two pump sets, each capable of delivering a flow equal to Qmax and working alternatively for 12 h each. Q installed is equal to 2 Qmax. This represents the cheapest solution in terms of investment, but it is the most uneconomical in terms of running costs, as it delivers too much water during a long period and the pumps are underutilized as they can operate for 16 instead of 12 h;
2. it seems thus more interesting to equip this main station with three pump sets of unit flow equal to Qu = 0.5 Qmax, which means that total Q installed is equal to 1.5 Qmax;
3. still better, one could install a group of four pump sets as follows: 2 sets with a unit flow equal to main flow/2 (Qu = 0.5 Qmax); and another 2 sets with a unitary flow equal to main flow/4 (Qu = 0.25 Qmax). In such a case, the total Q installed is equal to 1.5 Qmax as in the case above. Running costs would be much more variable and therefore adapting better to the various situations for flow requirements.
Fig. 57  Operating diagram for three options of pump sets 
These three possible solutions are further analysed in figure 57 where the operating diagrams of the pumping station equipped with two, three or four pump sets are shown.
Translating such calculations into operational costs, and particularly into annual electricity costs for pumping, is important before deciding which type of installation would be preferable.
Secondary pumping station: the equipment of the secondary pumping station consists generally of two to four types of pumps with different characteristics, which vary according to the service units to be supplied with water. The unit flow varies from a few litres per second to about 100 l/s, with a variable total head ranging from 10 to 40m.
A good choice in such cases is to install double sets of each type of pump. Obviously, it will be too costly to treble each pump set, all of them having a different QH. It would be more than reasonable to keep a good stock of spare parts available.