Improving Aeration Controls While Being Limited by Blower Flexibility

Background
The Agua Nueva WRF (Figure 1) was designed as a scalping plant with a capacity of 32 MGD currently receiving an average daily flow of 25-28 MGD, following a distinct daily pattern. Secondary treatment includes 4 parallel basins in a 3-pass step-feed, 5-stage Bardenpho bioreactor system, followed by secondary clarifiers, tertiary disk filters and is disinfected using chloramines (Johnson et al. 2014). The objectives of the aeration controls were to manage the extreme dynamics of the influent flow and loads, and control ammonia to a band of 1.5 ±0.5 mgN/L to minimize external ammonia dosage for the chloramination system and the dosage of sodium hypochlorite.

The A zones (Figure 2) were designed to carry out most of the nitrification and handle the load dynamics via a high-low feedforward controller based on influent flow or ammonia load. The downstream B and C Zones of each step-feed pass were designed to adjust aeration intensity based on an ammonia feedback controller such that the Disc Filter effluent maintains the targeted ammonia range. Additional coarse bubble diffusers were installed in the A zones to supplement the oxygen supply of the fine bubble aeration in the 30 ft deep tanks.

Problem Statement
The steep influent load changes (Figure 3) require a faster reaction in aeration intensity than a feedback controller could deliver.

Triggered by insufficient nitrification performance, a study identified the likelihood of Quaternary ammonium compounds in the influent inhibiting nitrification. Even at a higher SRT additional aeration capacity was required to fully compensate the slower nitrification rates. The result was that the B and C Zones had to be aerated to a higher DO setpoint than originally designed. Pushing more air through the steep diffuser tapering caused significant pressure drops and required a higher system pressure. This issue was exacerbated by turbo blowers with limited turn-down capabilities. At high pressure, the surge margin restricted the available speed range, resulting in frequent blower cycling. At low pressure, gaps in airflow delivery also caused blower cycling. The need for higher pressure was further aggravated by aging diffusers. The inability of the blowers to maintain the required higher pressure led to drops in DO concentrations and significant ammonia breakthroughs.

The original controls for feedforward were implemented as high-low DO controllers and the coarse bubble aerators were switched on or off according to the fine bubble airflows. This caused additional disturbances to the very sensitive overall aeration system and often triggered blower cycling. Over the years, several control workarounds were implemented to deal with the blower issues but were difficult to follow. One key aspect of the redesign was to inform operators of the active controls and their performance.

Control System Redesign
The conceptional design uses a flow or ammonia load feedforward controller for the calculation of DO setpoints for the A Zones, whereas an ammonia-based feedback controller feeds a DO setpoint to the B and C Zones. A key improvement was the addition of secondary integral windup protection (SIWP) (Carlsson et al. 2017) to prevent that an underlying control loop causes integral windup in the outer loop. This is especially important in situations such as at Agua Nueva with limited abilities to increase system pressure or situations where the controller setpoints cannot be reached.

To coordinate fine and coarse bubble aeration in the A Zones, a control concept is applied that uses the filtered fine bubble airflow as input into a proportional coarse air control logic with bounds. Another goal was to redesign the control system for seamless integration with Jacobs' digital twin-based decision support system (Johnson et al. 2024).

Results Figure 4 shows the ammonia concentration at the end of aeration basin 1 before and after the control improvements. The ammonia peaks have been eliminated by the control improvements and the ammonia setpoint is maintained in both directions resulting in better water quality and lower chemical and energy consumption.

Figure 5 shows the improvements from the new coarse air bubble controller and Figure 6 of the feedforward flow controller. Both controls smoothen out the response and stabilized the overall aeration system performance. The SIWP prevented integral windup accumulating in the controller cascade when the control loop setpoints could not be reached (Figure 7).

Figure 8 shows an example of a faceplate providing insights to operators on the functioning, status, and performance of the aeration control system.

Broader Impact
The project improved an existing aeration control system by thoroughly analyzing the existing system constraints and redesign accordingly. The constraints made it necessary to develop and implement control solutions with smooth responses to load changes so that the control outputs would not cause disturbances themselves. A feedforward controller targets the reduction of dynamics, so that the ammonia feedback controller can deliver the targeted tight ammonia range. Major performance improvements are due to introducing secondary integral windup protection.

Special attention has been put on providing insights and access to operators via graphical user interfaces. The findings are easily transferrable and are currently applied to other aeration projects.