Optimal Pump Scheduling By NLP For Large Scale Water Transmission System

In this paper an operational control for the Toronto’s Transmission Water System (TWS) is considered. The main objective of the ongoing Transmission Operations Optimizer (TOO) project consists in developing an advanced optimization and control tool for providing such pumping schedules for 153 pumps, that all quantitative requirements (such as pressure, flow, water level and water quality) with respect to the system operation are met, while the energy costs of delivering fresh water are minimized. We describe here, in general, the concept of TOO system and the large-scale non-linear, so-called Full Model (FM), based on system of hydraulic equations. The FM model is, in fact, a simplified version of EPANET hydraulic simulator with hourly hydraulic time-step, implemented as a complex NLP optimization model and usually solved over 24-hour horizon to deliver the aggregated optimal solution. To solve the resulting large-scale NLP we use the nonlinear interior-point method implemented in general-purpose large-scale IPOPT solver. Finally, we included the typical numerical example of application of the TOO Optimizer to solve the 24-hour and 7day FM problems, and compared obtained optimal FM results with results of hydraulic simulation performed under EPANET simulator. INTRODUCTION: TOO SYSTEM The City of Toronto water transmission system is a large complex hydraulic network of treatment plants, pumping stations, storage facilities (including underground reservoirs and elevated tanks) and transmission utilities (including hundreds of kilometers of water mains and significant number of control valves). The City of Toronto water supply system capacity is one of the largest in North America. The Water Supply function is responsible for providing services 24 hours per day, seven days per week. The system consists of treated water pumping at four filtration plants, 29 pumping stations, floating storage at 19 reservoirs and 9 elevated tanks, and approximately 520 km of water mains that transport treated water from the lake up through the system. Water is pumped through a hierarchy of pressure districts with elevated storage facilities. At present a large part of the system within the City of Toronto is essentially manually operated, where an operator decides for example when to turn a pump on or off. Even when there are no abnormal situations, manual decision making within the City of Toronto system is a complex process. With this background, the City of Toronto decided to develop the Optimizer that automatically determines control strategies for the Water Transmission System, based on certain criteria, including meeting service delivery levels (pressures, reservoir levels, water quality). The developed TOO Optimizer works as on-line tool alongside the SCADA control systems. The primary objective of the Optimizer (TOO) is to ensure that required water delivery standards are met, while minimizing electrical cost of water pumping. “The aim of pump scheduling is to minimize the marginal cost of supplying water while keeping within the physical and operational constraints, such as maintaining sufficient water within the system’s reservoirs, to meet the required time-varying consumer demands.” – (Methods, 2003). The complexity of the water system and dynamically changing energy rates present potential opportunities for optimizing operations by minimizing water pumping and treatment costs. The simplified flowchart diagram of the whole TOO system and its main software element, called the Control Strategy Component (CSC) is presented in figure 1. The CSC component consists of: • the Optimizer using the FM and FMBM optimization models, and also the library of optimization and matrix solvers • the EPANET hydraulic simulator using the hydraulic and water quality model of TWS system defined in EPANET INP format • the FM (NLP) and FMBM (LP) optimization models solved by COIN-OR IPOPT and Clp solvers, respectively • the library of COIN-OR optimization solvers (IPOPT and Clp) together with the library of HSL matrix solvers (MA27, MA57 and HSL_MA97) • the Simplifier, which produces a simplified hydraulic model • the Scheduler generating an optimized pump schedules on the base of FMBM and FM optimal aggregated solutions The FMBM and FM optimization models are automatically obtained from the hydraulic model of network in EPANET INP format and from additional data (provided by the TOO system) describing operational constraints, electricity tariffs and pumping station configurations. In order to reduce the size of the FM optimization problem the full hydraulic model is simplified by the Simplifier, while retaining the nonlinear characteristics of the model. In the simplified model all reservoirs, elevated tanks and all control elements, such as pumps and valves, remain unchanged, but the number of pipes and nodes is Proceedings 28th European Conference on Modelling and Simulation ©ECMS Flaminio Squazzoni, Fabio Baronio, Claudia Archetti, Marco Castellani (Editors) ISBN: 978-0-9564944-8-1 / ISBN: 978-0-9564944-9-8 (CD) Optimizer Scheduler Optimal