Pipes: Conduits with a constant cross-section, with options to incorporate singular head losses and throttle valves. The values of Interior Diameter and Roughness can be loaded in the Pipes Database, according to manufacturer, Material and pressure rating, or imposed by the user. Optionally for each pipe an arbitrary number of accessories and a throttle valve (or retention valve) can be incorporated, with a variable degree of opening with single losses of the type    and generalised single losses formulated . The coefficients k of the singular losses in accessories is automatically communicated from the databases parameterised according to the type of singular element. In the case of valves the values of k are computed as: the discharge coefficient being included in the interval (0.1), and being a function of the degree of opennessa and valve type. is loaded from the valve databases. Alternatively, the values of K and N, reproducing the dissipation of total head in any devices, must be supplied by the user in the corresponding dialogue boxes, with utilities for adjusting to experimental data. Each pipe admits logical retention and independent closure devices and registers the design flow rate (equipped for the sizing phase) and speed in order to facilitate the migration of the results to transient analysis packages.

Pumps: Discharge elements defined in terms of increasing height through energy according to the flow rate, optionally considering total power consumption (or efficiency) and the NPSHR. The performance curves of the Element can be loaded using tabular values defined by the user, or automatically from the pump databases, which provide additional resources for choosing pumps adapted to a nominal point. Each pump incorporates a logical retention device. The performance curves are modelled using splines, permitting a precise fit throughout the range of flow rates for any type of geometry of the curve (maximums, minimums, inflection points). Thanks to this, GESTAR 2010 can exclusively offer a simple form of detailed modelling of the functioning of pumping stations as a whole, defining a pseudo-pump which takes as performance curves the conjoined curves of height-flow rate and power-flow rate, which GESTAR 2010 supplies directly in the “Pumping stations” module, according to the system curve, type and number of groups, and type of regulation.

Automatic regulating valves: Pressure reducing valves, pressure sustaining valves, flow limiting valves and acceptable combinations of these, configured by algorithms which enable a correct implementation of the operational states and limits of all of them in numerous contexts, including their placement in branches. The head loss coefficients K, for totally open positions, can be introduced manually or from the database of regulating valves.

In the case of networks configured as strictly branching with predetermined design flow rates and layout, GESTAR provides the sizing tool which includes economic optimisation criteria; this can find the combination of Pipes to meet the imposed pressure requirements for Design flow rates, with a minimum overall cost, or very close to the overall minimum. We present below the main features and resources of the optimisation module.

The most relevant numerical aspects related to the hydraulic solver are listed below: 

• Cuasi steady hydraulic solver uses a evolved Nodal Head type method, combined with Newton Raphson algorithm, for solving the system of non linear equations, with several extensions to handle with the specific irrigation network  components, (Aliod and Gonzalez 2007).

• Dynamic memory allow to analyze networks with unlimited number of components.

• Unknowns, either node heads or flow rate, are computed coupled and relaxed at each iteration step.
   
• Jacobian matrix computed for general non linear elements constitutive equations.

• Direct flow rate computation at known head nodes, coupled with the rest of unknowns.

• Pressure dependent flow emission modelled as specific element with constitutive equation given by the individual pressure-flow emitter response, with a final node where total head equals to altitude.

• Low resistance elements treated with the Campos (1993) technique allowing the direct and coupled computation of the low resistance element flow rate.

• Matrix direct inversion routines and matrix operation libraries specially adapted for a non symmetric matrix, by means of the linked lists technique (Duff et al. 1986). Routines are optimized for sparse matrices and compact storage.

• Solution convergence control by two independent criteria: maximum average flow residue at the nodes, respect the average flow rate in the elements, and maximum difference between the value of the head in the nodes, or the relative flow rate in elements, between two consecutive iterations.

• Regulation valves simulation by direct calculation of their operational parameters coupled with the whole equation system.
• Dual modeling of hydrants depending on the pressure level on the network: when the pressure rises above the hydrant pressure setting, the node is modeled as a pressure independent flow demand, but when the network pressure falls below that level, the node demand is modeled as an emitter.

• Pressure dependent continuous-like flow emissions (for modelling drip irrigation tubing) are  links modelled by means of a generalization (Aliod and González, 2007), of the integrodifferential approach of Warrick and Yitayew (1988).

• Inverse analysis allows the direct computation of unknown control parameters at node or elements, and computation of a common roughness or diameter for a group of elements, using a modified system of equations (Aliod and González. 2007).

• Volume level time evolution in extended period simulation is performed by means of the Rao and Bree (1997) formulation.

• Head loss generalized computation by monomic correlations or D-W expression with fully implicit formulae.

• Cubic spline fitting of characteristic curves (head, power, NPSHR) of  pumps vs flow rate, and of curves describing the reservoirs volume vs level.

Shown below are the main features and resources available in GESTAR 2010 for modelling, regulating and optimising pumping stations, by obtaining System Curves, Operation Curves and Probability Density Function of flow rates (PDF), which together establish energy costs for a certain period.

Determination of energy consumption: Combining the curves of the pumping station (head-flow rate and power-flow rate) and the Probability Density Function of the flow rates (with option to split into low, shoulder and peak) a calculation is obtained of energy costs reflecting the dependence of the results on the performance curves of the pumps, the number and size of the equipment, the regulation strategies used and the network system curve.