HYDRAULIC NETWORK MODELING AND OPTIMIZATION OF PUMPING SYSTEMS

Mitul Jani 1 , Shridhar Manure 2 and M G Dave 2 . 1. M. Tech. (Energy System), Dept. of Electrical Engg., Nirma University. 2. Bureau of Energy Efficiency (B.E.E.), Accredited Energy Auditor. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History


Hydraulic model:-
The hydraulic network simulation is done on EPANET platform, using an underground reservoir (R1), 4 pumps (P1-P4), 4 valves (V1-V4), 13 delivery nodes (D1-D13), 17 pipelines, viz.13 for delivery (L1-L13) and 4 for pump suction (F1-F4). The network developed as per actual configuration of the WDS is as shown in Figure 1 and detail as given in Table 1.  Existing condition analysis:-In order to calculate the efficiency of the pump, the performance parameters of the pump must be measured as per standard practices. These measurements should not be instantaneous and hence must be taken as an average of measurements over some span of running hours. In this case, the measurements are taken for pressure, flow and power for one hour for each pump and their efficiency is calculated using Eq. (1). Using the average values of one hour measurements, the efficiency is obtained for each pump as summarized in Table 2.
Eq. (1) Where, Q = Flow rate of fluid (in m 3 /hr or CMH) h = Total head developed by pump (in m) g = Gravitational acceleration (in m/s 2 ) ρ = Density of fluid (in kg/m 3 ) ƞ m = Motor efficiency P = Power input to motor The measurements taken at site are fed into the EPANET network module in order to evaluate overall performance of the hydraulic system. Per unit cost of electricity used for calculating the operating cost is taken to be as Rs. 6.25/kWh, the density of water was found out to be 992 kg/m 3 , the motor efficiency was calculated after conducting the no-load test to find out fixed losses and tested under full-load for copper losses, efficiency was calculated to be 90% as per de-rating factor and the acceleration due to gravity is taken as 9.81 m/s 2 . The pumps are operated for 8 hours a day which makes the per-day utilization / usage factor of 33%. The results of the simulation are given in Table 3. The pressure / head loss of the system components are as shown in Figure 2. The velocity of water in pipelines is an important parameter to evaluate the frictional losses, flow through each link and the unit head-loss as per the existing conditions modeled in EPANET are summarized in Table 4. These parameters are very useful in order to know about the operating conditions of the network. They determine the scope of improvement in network operation and the possibilities of future expansions or modifications as per the application demand. It is important to know how the network operates and in what condition so as to explore any avenue for optimization.
The velocity of water in each pipeline and the pressure available at different nodes / delivery points is as shown in Figure 3.  1) The condition of impeller of Pump No. 3 and 4 might be deteriorated; which is supported by the fact that the pump is producing only 28 m, due to which it is not able to develop its rated hydraulic power. 3) The velocity of water in all the pipelines was less than 2.0 m/s, which is satisfactory as per the design standards, as high velocity may incur more frictional loss. 4) The velocity are in limits, however, the pipelines L5, L9 and L12 have high unit head-loss due to the fact that the size of the pipelines are smaller (275 mm). 5) The pipelines L5, L9, L12 should be replaced with 350 mm pipes so that the head-loss is reduced as well as the loss-coefficient," C" used in Hazen-Williams equation, will increase and hence overall head-loss of pipeline will also reduce since the frictional losses would decrease. 6) The demand nodes D8, D13 and D11 are affected due to the poor performance of these pipelines, the pressure of water available at these nodes is very low as compared to other demand nodes. 7) Due to low pressure at demand nodes, the locality may not receive sufficient supply. Also if demand increases, the pump may not be able to deliver up to that node. 8) The pressure-head loss in pipelines is leading to waste of energy that is provided to the fluid by the pump and hence must be dealt with.

Impact of Change in Pipeline on the Network:-
From hydraulic network simulation, it is evident that pipelines L5, L9, L12 have significantly high head-loss per unit length, if the diameter of suction pipes are increased from 275 mm to 350 mm for all the pipelines, it would reduce the frictional head loss in pipelines and increase the pressure of water supplied at the farthest end node as evident from Figure 4. This simulation is showing significantly how this measure would improve the water distribution, also taking into consideration the future expansion of the network and increase in water demand to study / examine network behaviour in conditions of enhanced operating capabilities of the system. The change in pipeline diameter would yield results as obtained from the simulation given in Table 5. The results clearly show that the velocity of water in pipelines is significantly reduced and hence the frictional head loss is reduced as well. The effect of change in pipe dimension on pressure is also quite significant on the immediate demand node corresponding to the pipe and also increases the pressure on the last node of supply as the head loss for other pipeline is fixed. The results show that pressure at delivery node D13, D8 and D11 from pipeline L5, L9 and L12 respectively, is changed significantly due to change in pipe dimension and because of that, the end delivery nodes (D13, D8 and D11) have an increased pressure of supplied water. The difference in values of pressure at affected nodes is shown in Table 6. Corro-coating is a resin based chemical material applied on metals / alloys that are exposed to or are in contact with water / moisture. In this case, impeller and its casing might have been affected due to corrosion effect. The resin based chemical is applied as a coating to the internal parts of the pump.
Together these measures would increase the efficiency of the pump by 5% approximately in actual against claimed 10% by the manufacturer of resin coating material. It is most simple optimization measure. There is no requirement to make major changes in network component. Only by improving pump performance, the network performance is optimized.
The existing pumping network is estimated to work on the following improved parameters after implementation of suggested measures as shown in Table 7. It is also suggested for better performance and system operation, overhauling and maintenance of pumps should be done regularly and the system must be diagnosed properly in order to identify any problems, issues or factors that lead to degradation of efficiency of the pumping system. Timely maintenance of pipes is also an important factor and so is the cleaning of the strainer of the pump which is present on the suction side in order to filter out sludge / solid particles / waste, if present in the reservoir.
The improved flow and pressure of the network after implementation of energy saving measures is projected to be as shown in the network in Figure 5.
1042 The overall efficiency of the network has improved and is evident from the calculated results as given in Table 8. The summary of calculations for economic feasibility of implementing these measures is shown in Table 9. The total energy savings from these measures for present operational practice, is projected to be 1,86,180 kWh and in monetary terms, Rs. 11.65 lakh with an estimated investment of Rs. 34 lakh and average simple payback period of approximately 3 years.

[2] Construction of an Overhead Tank and pumping with an Existing Pump (1 x 200 HP):-
In current scenario, the WDS is not having any facility or infrastructure to supply water to the locality for 24 hours. The water is supplied in 4 hours duration for 2 times a day and hence only 8 hours a day. If an overhead tank is constructed in the premises of the WDS and a existing pump of 200 HP after overhauling, (in order to improve its efficiency) is used to supply water to the tank and then the tank could supply water to the network, it would not only reduce the no. of pumps operating and save energy, but also provide 24 hours of water supply to the locality. As per the data available from the network, in order to cater the demand of the locality, the capacity of an overhead tank should be approximately 45 Lakh litres in order to meet the demand for 24 hours. According to the capacity and pressure head requirement of network, the dimensions as well as the elevation required for the concrete tank are obtained to be as given in Table 10.  The costing of construction of an overhead tank and the civil, plumbing, electrical costs are obtained as a method of weighted average as given in Table 11. In this N.O.M., only one pump of 200 HP is supplying the water to the tank through a pipeline as per the level of water inside the overhead tank. This measure would save energy in terms of pumping requirement and the overhead tank would provide the network with 24 hours of un-interrupted water supply. In order to simulate this inside the network built in EPANET, the demand of each node over 24 hours has been distributed into average hourly demands with variable multiplying factors. This gives idea about the hourly performance of network as well as improves the distribution capacity of network by providing un-interrupted supply.

Pump
The modified network as simulated in EPANET, shows the pressure on individual nodes and flow in individual links as shown in Figure 6. The performance of the pump after the overhauling, working at improved efficiency of around 72% works on the estimated performance parameters as given in Table 12. The pump is required to run for 12 hours during the day in order to pump water to the overhead tank as per demand and water level variations. The techno-economical aspects of this N.O.M. as calculated in EPANET are as given in Table 13. A summary of feasibility study and economical advantages of this N.O.M. are projected to be as given in Table 14.   The new pumps are working at an efficiency of around 71% and estimated performance parameters as given in Table 15. The pumps are operating for 13 hours during the day to pump water to overhead tank as per demand and water level variations. The techno-economical aspects of this N.O.M. as calculated in EPANET are as given in Table 16. A summary of feasibility study and economical advantages of this N.O.M. are projected to be as given in Table 17.

Summary:-
The summary of all Network Optimization Measures (N.O.M.) are as given in Table 18. It gives a comparison of different optimization techniques for a water distribution network of WDS. The choice of most suitable option can be done on the basis of the requirements of the application. Also it depends on various factors like reliability of network supply, power consumption, scope of energy savings, amount of investment, suitability for the site and various other considerations are to be kept in purview of the subject while undertaking the hydraulic study of a large and complex network.