Our paper aims at assessing the way in which EPANET estimates the efficiency of variable speed pumps operating at speeds different from the nominal speed. An experimental setup, existing at the Technical University of Civil Engineering Bucharest, was modelled in EPANET 2. The facility allows the measurement of the characteristic curves of two identical pumps driven by electric motors with the rotational speed ratio of ½. The pumps curves were fed into EPANET and results for different rotational speeds were compared to measured values. Results show that EPANET does not alter pump's efficiency curve when operating at different rotational speeds

“Teiul Doamnei” Pumping Station in Bucharest supplies the WDN of a district with about 40,000 inhabitants. The pumping station was refurbished in 2004, and variable speed pumps were installed. Due to this refurbishment and other improvements of the water networks, performed by the Water Company of Bucharest, the total energy consumption for 2012 dropped to about 25% with respect to the value recorded for 2004. The numerical model of the pumping station and it's attached water network with 250 pipes and 212 nodes was created in EPANET. Variable speed driven pumps functioning algorithm was implemented in EPANET via ruled-based control statements. The overall variation of pumps’ speed allowed by the controls is between 70% and 100%. For each of the 12 months in 2012, from the water consumption recorded at the pumping station, a mean daily value of the flow rate was derived. The daily variation of the hourly flow rate has been considered according to Romanian standard for drinking water supply quantities in urban sites of about 40,000 inhabitants. This variation was introduced in EPANET as flow pattern over a 24h period. The mean daily value of the flow rate for each month, divided by the number of consumption nodes of the network, was added as base demand at each consumption node. The simulations were performed for a 24h period, for each month, with a time step of 5minutes (although flow rate variation is hourly, due to the limitations of EPANET, the hydraulic time step must be smaller than the flow rate pattern step). Monthly energy consumption computed values are in good agreement with the corresponding values recorded at the pumping station. The model we created can prove useful in performing other analysis in order to assess the functioning of variable speed driven pumps in a WDN

The case study points on the water distribution system (WDS) of Buzau City, a medium-sized city in South-Eastern Romania. We built a simplified numerical model of Buzau's main WDS in EPANET, which connects 4 pumping stations to 45 booster stations (end-users) and other users. The model was calibrated upon real-time recordings from 2014. The pumping scheduling is established upon control statements driven by the: discharge pressure of the corresponding pumping station; pressure levels at selected key points in the network (here, one pumping station can be shut down).Energy consumptions were calculated and compared with the current situation

Pump operation control under variable system demand is obtained either by throttling the discharge pipe, by using by-pass pipes, by modifying the rotor speed, or by separating the pump and the demand by means of a pressurized tank. We try to assess the efficiency of pump operation control by using a numerical model created in EPANET for a 24hours period. To avoid the demand driven algorithm used by EPANET, the network model consists of a throttle control valve followed by an emitter. Pump or valves parameters are modified using control statements. Results are consistent with the general theory

Pumping stations scheduling for water supply systems can be optimized by the Honey Bees Mating Optimization Algorithm (HBMOA). The study-case is the water distribution network of Oradea, Romania, consisting of 4 pumping stations, 3 tanks with variable water level and a constant level water source. The network was simplified, using a numerical network model and data recorded in August 2006, with two daily flow regimes (peak/off-peak). A modified form of HBMOA gave a suboptimal solution (the combination of pumps speed values), which is better, the daily pumping energy consumption being lower, than the one recorded in August 2006