Optimal Pump Scheduling for the Newly Built Sao Lourenco Water Supply System in the Metropolitan Region of Sao Paulo

摘要

The metropolitan region of Sao Paulo (MRSP) suffered a record-breaking drought period between 2013 and 2018. With 21.5 million inhabitants, the region has one of the largest water supply systems in the world, consisting of ten partially interlinked reservoir systems. A set of mitigation measures and expansion works were adopted during the last few years to recover from the recent drought and improve reliability and resilience of water supply. For example, the newly built Sao Lourenco water supply system is comprised of catchment structures, pumping stations, a water main, and water treatment plant (WTP). This system is able to supply water to approximately two million people and plays a complementary but important role in increasing water security in MRSP. While the system competes with energy production in a highly productive hydropower cascade, it has an installed capacity of 6 mVs, head near 400 m, and a 50 km long water main (its diameter is equal to 2,100 mm). Permitted average water withdrawal is 4.7 mVs. Other features of this system are three raw water tanks at the entrance of WTP, with a volume of 25,000 m~3 each, which helps to minimize pumping during peak hours. This paper studies linear and mixed integer programming applied to optimize the short-term pump scheduling and raw water tanks operation of the Sao Lourenco system, considering peak and off-peak prices of energy, variation of demand, cleaning and maintenance services, and stops/starts constraints.