BY KIP EDGLEY
The use of Variable Frequency Drives (VFD) within the water and wastewater industries has evolved over the past three decades as motor speed control technology has developed and grown, resulting in dramatic changes to process control. The result of these gains has been the ability to look into a drive system and extract multiple parameters representing selected data, and to integrate this information into the ultimate control of the device.
Testing period on left. Note that due to differential and PSI component, bias held AFD solid. Note instability after loop bias is taken out, but PSI PID loop recovers quickly.
In most field installations, Adjustable Frequency Drives (AFD) are implemented as a finished product and are often considered inviolable once operating. This perception is usually incorrect. The ability to reconfigure or understand AFD data can be achieved with moderate programming skills. Users with these skills can, with some investment of time, acquire new or different data from the AFD to tune and optimize specific processes and performance.
As an example, Veolia Water decided to look at the use of four parameters on a 200 HP AFD which is controlled by a PLC and only provides discrete signals for Run, Fault, and Alarm. All other control and command is from the Ethernet peer-to-peer messaging with the PLC. The AFD controls water applications where well water is distributed either to a reservoir through Flow Mode, or a pressure zone directly in a Pressure Mode.
Using AFD data specific to this application, adjustments to the PLC programming provided the ability to bias the PSI PID loop and allow the differential between torque and flux current to be used as a trigger to introduce additional gain to the PID loop. By adding bias to the PID, the speed command would be increased. The resulting change in the process was evaluated against the limits of the process set point and the increased differential to determine if additional bias could be added.
Torque current is the portion of the total current to the motor that is in phase with the output voltage. Flux current, simplistically described, is the portion of the total current which magnetizes the motor. The efficiency of any motor will improve if more of the total current is in phase with the output voltage. The Power Factor will also improve as it is an expression of the phase angle between the voltage and the current.
By using the two components as individual data values and calculating the differential, the goal was to hold the pump within pressure parameters with an allowed PSI variance of ±10 percent of the setpoint (5 psi in this case), and to increase the differential current between Torque (higher desired) and Flux (lower desired).
During normal duty this 200 HP pump typically keeps the pressure at an average of 50 psi, but short term swings in pressure can be tolerated as long as the system remains lower than 60 psi. Within this system, the PLC stored the average differential over a period of time to establish a baseline differential.
When the bias loop was activated, the PLC would increase pump speed by adding to the PID loop and determining if the differential was increasing beyond the baseline. If this condition was true, the PLC would look at the pressure and determine if additional bias could be imposed on the loop. If the psi was approaching the upper level, it would do nothing. If the tolerance was exceeded, the speed bias would be reduced to bring the loop into pressure tolerance.
Results
The overall loop performance was improved based on the imposed bias, but the power gains were mostly in improved power factor as the KW expended was used more efficiently. Staff also observed that the PID loop itself was at times more stable as the bias would increase the torque current component. Evaluating the PSI and differential provided periods where the loop was held at a steady state while the reduced flux current state existed.
This project was intended to show test data from a difficult application (real time water delivery) as it is much more challenging than simple pump functions such as multiple hour runs for reservoir fill cycles or wet well pump down applications where the pump discharge is not regulated by a tight process boundary.
A plant’s efficiency is tied not only to environmental benefit, but to cost savings. By increasing motor efficiency, operators contribute to a reduction in the plant’s overall energy footprint and maximize the use of ratepayer money. Veolia Water plans to test multiple applications and locations, and anticipates that the power factor correction from the reduction of the flux component will be very beneficial to those with similar applications and interest in maximizing pump performance per KW consumed.
About the Author: Kip Edgley is a Manager of Automation/Integration at Veolia Water North America (www.veoliawaterna.com). He has been involved in water and wastewater operations since 1988 and has worked with motor controls and electrical systems, as well as PLC programming, instrumentation, communications and SCADA design and deployment for more than 30 years.
New Desalination Plant to serve Chilean Mine
Severn Trent Services has been awarded a contract to provide a seawater desalination plant that will serve the potable and process water needs of the Minera Esperanza copper and gold mine in Antofagasta, Chile.
The plant will draw seawater from the Pacific Ocean, which will be treated using ultrafiltration (UF) and reverse osmosis (RO) systems. The seawater will be pumped 145 km to the mine located at 2,200 meters above sea level.
Construction of the 634,000 gpd plant, which is owned by Antofagasta PLC of London and Marubeni Corp. of Tokyo, is expected to be completed in the fourth quarter of 2010. The equipment will be installed and commissioned by Proequipos Ltda. of Santiago, Chile.
The Minera Esperanza project is one of the world’s first such projects in which 16-in membranes have been used rather than standard 8-in membranes. By using larger-diameter RO membranes, fewer membranes and pressure vessels are required to produce the same flow of water. Using 16-in membranes also will require fewer pipe connections, shorter pipe runs and a smaller building to house the seawater desalination system.
Two Severn Trent Services UAT™ 705,000 gpd UF trains using Dow™ UF membranes and two UAT 634,000 gpd RO trains using Dow Filmtec™ 16-in membranes will be installed at the plant. The UF pretreatment system will not require a coagulant and will operate at 90 percent recovery, while the single-pass RO system will operate at 45 percent recovery, producing water quality below 400 ppm total dissolved solids.
