Simulation tests for the operation of a water main with break pressure tanks I. H. T. ALLY1, M. MULHOLLAND2 AND C. A. BUCKLEY1 1 Pollution Research Group, Chemical Engineering, University of KwaZulu-Natal, Durban, 4041, South Africa. [email protected] 2 Chemical Engineering, University of KwaZulu-Natal, Durban, 4041, South Africa All correspondence to be addressed to Ismaeel Ally (address 1) Abstract Western Aqueduct system Results The Western Aqueduct project was conceived by eThekwini Water Services (EWS) to augment current • Skeletonization : infrastructure and to sustain adequate potable water supply capacity to the rapidly increasing Durban Umlaas Road Reservoir Normal Operation population. The study was initiated at the behest of the Water and Sanitation division of the eThekwini • Hydraulically simplify the Municipality, and is centered on the Ashley Drive (20 Ml) and Wyebank Road (10 Ml) break pressure tanks 20.01 km complex pipeline network Lumped (BPTs) that function as pressure-reduction devices for the new Western Aqueduct. Despite the advanced Demand Original Control Revised Control (2016) progress with the installation of the Ashley Drive BPT, a working, realistic, hydrodynamic model is still 20.01 km required to undertake simulation studies and evaluate the adequacy of the design. The Western Aqueduct • Account for: Oscillatory settling level is designed to accommodate anticipated 2036 water supply/demand conditions (peak flow – 400Ml/day). Ashley Road BPT (20Ml) Deadband zone These considerations include the proposed sequencing of the control valves, the speed of valve • reservoir draws from trunk 0.53 km 1400 mm Emberto movements and failure and maintenance modes. n Ashley Drive BPT - Flow & Level Ashley Drive BPT - Flow & Level mains 3.30 km 1400 mm Reservoir All valves fully open Tshelmnyam 10 initially 10 9 9 a Reservoir 7,0 m settling level Flow (into 8 BPT start level 70% 8 AD) • Consumer draws from Level 7 Level 7 BPT level BPT level Better adherence The hydrodynamic simulation model incorporates the trunk main in its entirety, while also accounting for 6.5 m settling level 1.06 km 1400 mm 6 6 initially in Haygarth to intended NOL the reservoirs that are supplied by the Western Aqueduct infrastructure, and incorporating vital aspects reservoirs 5 5 70% deadband Abelia /s)| Level (m) /s)| Level (m) 3 Reservoir 3 4 Flow (into AD) 4 (Kloof) pertaining to these reservoirs, such as the time-based demand profiles (consumer) and level control. 3 3 Reservoir Flow (out of AD) Flow (out of 2 Flow (m 2 Information utilized in the building of the model was sourced primarily from the design information and • Pipeline pressure profiles 1.14 km 1400 mm Flow (m AD) 1 1 Jerome current-operational results. MATLAB® was selected as the preferred program for the implementation of the 0 0 Drive 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 model. Inbuilt functions and simple mechanisms were used in conjunction with structured programming • BPT levels Reservoir 3.26 km 1400 mm Time (Minutes) Time (Minutes) principles in order to manage the complexity of the model code. Several solution methods were tested in Wyebank Road BPT order to attain the optimum trade-off between model accuracy, model complexity and the resources (10Ml) WR BPT globe valve activated Wyebank necessary to obtain meaningful results. 0.24 km 1400 mm Reservoir The results of the study are presented as time-sequence plots depicting the results of various random and 0.43 km 1400 mm Mount Ashley Drive BPT - Valves Moriah Globe Globe valve 3 valve 2 stress-tests for conditions that range between 2015 and forecasted 2036 conditions. Support for the design Reservoir Globe valve 2 1 Globe concepts, additional recommendations and a heightened ability to anticipate system performance under 2.06 km 1400 mm Globe valve 3 Stepwise valve 1 0,8 sequential valve KwaDabeka Globe valve 1 varying conditions have been derived from the study and are reported in the paper. Concerns regarding the movements AD BPT 0,6 1 Reservoir Sleeve level in current control philosophy are also expressed. valve 2 deadband 8.00 km 1400 mm 0,4 Sleeve Sleeve Sleeve valve 1 Sleeve valve 2 valve 1 valve 3 Ntuzuma 0,2 Valves shut 5 (NR5) on high Sleeve valve 3 level Reservoir ValvePositionofstroke)(fraction 0 0 20 40 60 80 100 120 140 160 180 Background Time (Minutes) • Task: Analyze the system stability and vulnerability by simulating the BPT valve oscillations – wear risk 25% valve movement Normal Operation Stress tests tank levels and the system’s behaviour for each control system. • Profiled consumer draws • E.g. ‘triple step’ – disable reservoir controls - until 90 • Reservoir draws from trunk main as per control mins no reservoir draws from trunk mains, switch to system (deadband) maximum draws for 90 mins (to 180 mins), switch back 3-Step stress test (nil/full/nil) to zero draws after 180 mins. +200 m Original Control Revised Control (2016) Supply point 100% level 1400 mm overflow Original control SV: 25|0|0 Revised Control (2016) (undesirable) – no diameter GV: 0|0,5|1 Time-driven overflow 7,5 m 4,5 m sequence Valve rates restricted SV: 0,25|0|0 GV: 0,5|1|1 Deadband zone 7,0 m No valve movements Ashley Drive BPT - Flow & Level All valves fully All valves 10 SV: 0,25|0,25|0 open initially fully open Level (AD) Control philosophy 9 initially Level? 6,5 m GV: 1|1|1 3,5 m 8 Time-driven 7 0 m BPT level sequence BPT level 6 All reservoirs start drawing at Break Pressure Tank (BPT) 5 180 mins • Numerous interconnected variables includes: Level (m) | /s) Consumer 3 4 Flow (into AD) 2152 3 demands 1625 Mℓ/day • BPT tank levels Special Events: Longevity (m Flow 2 Flow (out of AD) All reservoirs All reservoirs 1 • Compartment maintenance 2106 start drawing cease drawing 869 Mℓ/day 0 • Consumer demands • Sleeve valve maintenance AD BPT at 180 mins at 360 mins 0 50 100 150 200 250 300 350 400 450 500 All reservoirs • Electricity outages Time (Minutes) cease drawing at 2036 WR BPT 360 mins • Reservoir intake flows 400 Mℓ/day 2036 400 Mℓ/day Maintains better (lower) Deadband & timer • System hydraulics (dynamic losses) settling level – all Dynamic settling level control reservoirs drawing • Static pressure and system topography (6,00 m) Break Pressure Tanks Control systems Findings To BPT compartments • Robust, capable of handling most special events From Umlaas Road reservoir Staggered valve Restricted valve Inlet launder Purpose: Reduce & • System viable until 2106 system pressure to movements movement rates DN300 Sleeve DN800 butterfly valve • Overflows likely with high demand variations– ample drainage (open) valve inside manageable levels Valve Position Water Level >4,5m Water Level <3,5m chamber DN600 globe • Water loss consideration – no override ‘smart’ controls – higher than DN100 gate 0% valve valve 25% intended maximum flow (~double) DN800 butterfly valve 30% DN800 butterfly valve (open) • Globe valve action – deviation from design intent of intervention only during (open) 35% 40% power outages. 45% position 50% valve • Valve oscillations – maintenance risk DN1400 common manifold Hydraulically 55% actuated* 60% • NOL (normal operating level – 50%) – not adhered to in original control – 65% Increase sleeve valve position valve sleeve Increase • Inlet control sleeve Decrease 70% improved with revision • Bottom-fed 75% 80% • Transient overpressures (7 bar) – significant consideration • Parallel offtakes – both globe valves and submerged sleeve valves 85% • Maintenance & wear • Common inlet chamber + overflow weir `.