Predictive Modeling of Delivered Water Utilizing
Power Consumption Data

Lead Researcher: David Zoldoske
Industry Partners: REDtrac, Madera Pumps, and McCrometer

The need: Growers want to better understand the relationship between changes in kWh input and pump output. This project focused on integrating data that are usually unavailable or treated separately, including input energy changes, water flow rate, water pressure, water standing and pumping levels.  

The goal: The main goal of this project is to help growers understand the relationship between energy demand and pump performance (water and pressure). The project is designed to be accomplished in three phases: (1) establishment of proper data water and energy measurement infrastructure, (2) modeling of input energy and pump flow and pressure output, (3) impact assessment of changes in dynamic conditions of water pumping level and field irrigation conditions. 

 The impact: The team collected data from wells at the from the Fresno State farm and 10 additional wells to develop a computational model to assess input power vs. water output relationships. Major findings so far include producer resistance to allow data to be collected from their wells. Historically, California water users are not required to monitor and regulate water use on private property. Several irrigation districts planned to work with the team on implementing monitored wells within their service areas. This additional data will assist the team in generating algorithms to predict groundwater extraction from power consumption.