An eight-month study by the Region of Durham, Ontario, Canada, has confirmed the cost advantages of optimizing pump schedules using the Scheduler extension of InfoWater.
Energy costs represent a significant portion of any utility’s total operation and maintenance budget. The Region of Durham, Ontario, Canada is no exception, with 14% of the cost for operating transmission mains and 29% for treatment plants devoted to electricity bills. To help reduce these costs, the Region utilizes MWH Soft’s InfoWater software for hydraulic modeling and management. In conjunction with modeling, the Region also uses Scheduler to optimize pump scheduling and operational controls for more efficient operations.
System Background
The Region of Durham is home to over 500,000 people and its population is expected to nearly double by 2021. The region is on the shores of Lake Ontario, directly adjacent to the City of Toronto. The water system is composed of:
- 6 surface water supply plants
- 18 remote water storage facilities
- 17 booster pumping stations
- 28 groundwater wells
- 1900 km (1200 mi) of water mains
Optimization Approach
In an effort reduce energy costs, the Region of Durham decided to use InfoWater’s Scheduler extension to analyze and optimize a densely populated area that is home to over 75% of the customers in their system. They selected the southern area, composed of the communities of Pickering, Ajax, Whitby and Oshawa, which receives its water from Lake Ontario. This lake-based water system is composed of 1800 km of pipes (50mm - 1350mm), three major water supply plants, ten reservoirs and elevated tanks, twelve pumping stations, and 75 pumps.
They started by establishing an accurately calibrated 24-hour extended period simulation (EPS) of their system. To do this, they built an all pipe model of their system using their ArcGIS database for infrastructure data. The model is fully integrated to the GIS, so that updates to the GIS flow directly to the hydraulic model. The model was calibrated to a macro level (trunk mains), and actual operational scenarios were then performed and calibrated against Supervisory Control and Data Acquisition (SCADA) information.
To accurately start optimizing pump schedules, other key information is needed. The first is the energy price forecast for the following day. Energy prices in Ontario vary throughout the day. Each day the Independent Electricity System Operator (IESO) publishes a forecast of the next day’s energy rates to its website. This information is downloaded and loaded into InfoWater.
The second is forecasted demands for the next day. Region of Durham forecasts the next day’s demands based on historical demand information for that time of year and the weather forecast for the following day. From this raw demand forecast, a diurnal pattern for each region of the system is created. These patterns and demand information are also loaded into InfoWater.
Lastly boundary conditions for the system must be determined. Because the simulation is run daily, each day’s tank levels and pump status must be pulled from the SCADA system to ensure an accurate simulation.
InfoWater Scheduler
Using the power of genetic algorithms, enhanced with advanced Elitist and Global Search Control strategies, InfoWater Scheduler works quickly to define the optimal operational strategy for the target system based on hydraulic and water supply performance criteria and maximum cost savings.
The operating strategy (or schedule) for a pump, valve and pipe (e.g., interzone water transfer, valve control settings) represents a set of time-dependent control rules that indicate when it should be turned on or off over a specified period of time (usually 24 hours). The system operational criteria produced by InfoWater Scheduler prescribe lower and upper limits on system pressures, tank and water quality (water age) levels, maximum pipe velocities, maximum pump (and valve) flows or volumes, alternative supply sources, and other system constraints aimed at ensuring safe, reliable operations.
Scheduler outputs this information as a specific dataset of operational controls. This dataset is loaded in to a scenario for that day and a model simulation is run to double-check system operations. Once the engineer is satisfied that this operational control pattern is acceptable, it is turned over to the system operator for implementation. Once the operations team receives the results of the pump schedule, tank levels and pressure at the plants they are responsible for deciding whether to implement them. Currently, the optimized pump schedule produced by Scheduler is used about 80% of the time.
Conclusions
Results of the optimization study, conducted over eight months, have been encouraging. The savings range between 1% and 11% of energy costs from month to month, with an annual projected savings in the pilot area of over C$100,000 (Canadian Dollar).
The pilot study has also resulted in ancillary benefits:
• Identify system bottlenecks
• Identification of the need for an additional pump at one treatment plant
• Evaluations of higher efficiency pumps or variable speed pumps throughout the system
• Understanding of the energy cost of throttling valves
It has also produced a continuous operator training program where operations staffers learn new ways to operate the system more efficiently.