Local control schemes for real-time optimization of variable speed pumps

Abstract

In this work, we have developed local control schemes for the optimal operation of a pumping station that includes multiple variable speed pumps. The pumping station must maintain a target pressure for a closed supply network that usually does not contain storage facilities. The control loop includes repeatedly reading the control feedback (pressure) from the control system and then adjusting the operation of the pumps in the station as needed: changing the pumps' speed and on\off status. Firstly, we formulated a control algorithm that simulates the current practice. Next, an optimization algorithm was developed to achieve minimum energy operation considering the efficiency curves of the pumps. The algorithm utilizes the pumps' characteristics curves for ensuring hydraulic feasibility while sustaining the required pressure setpoint. The optimization problem incorporates bound constraints on the speeds, time gap constraints to prevent frequent pump changes, and physical constraints that quantify the operation point in the flow-head domain and the flow-efficiency domain. We considered two operation strategies. The first is the "free strategy" in which each pump can operate with a different speed inside a predetermined speed range, while in the "equal strategy" all active pumps must share the same speed. The methodology was demonstrated using a realistic case study. Our preliminary results indicate potential energy reduction compared to the current practice.