The effects of short-term temperature variations on activated sludge settling

Abstract

Settling properties of activated sludge or mixed liquor suspended solids (MLSS) have been studied for more than 75 years at wastewater treatment plants. Temperature, together with MLSS concentration, has been acknowledged as important contributors to MLSS settling variations. Batch MLSS settling tests are performed on a regular basis at most of the plants. The majority of these MLSS settling test reports reflect the complete absence of any form of temperature compensation or even MLSS sample temperature (Ts) recordings. The objective of this study is to evaluate the effects of short-term temperature variations on MLSS settling parameters. This is done by means of simplified theoretical calculations, followed by operational reactor temperature (Tr) observations, and batch MLSS settling tests. The experimental work concludes with the implementation of an on-line MLSS settling test procedure at a full-scale plant reactor to develop settling models based on diurnal Tr fluctuations. These settling models illustrate that parameter correlations improve when Tr is included in on-line MLSS concentration-based settling models. The unhindered settling velocity of a single solid biofloc in water is considered in a simplified calculation to estimate the effect of temperature variations on MLSS settling. Over a Ts increase of 20°C, water density and viscosity reductions result in a calculated biofloc settling velocity increase of less than 0.5 m/hr. Similarly, biofloc density, shape, and size changes result in calculated biofloc settling velocity increases of about 11, 10, and 2 m/hr respectively over the 20°C Ts range. Plant temperature recordings show significant short- to long-term variations. Ambient temperature (Ta) and Tr fluctuate about 20°C and 1.8°C respectively per day, and Tr changes by about 4°C within a week, as measured on-line at local plants during the test period in winter. The aeration method can have a significant impact on Tr. Differences in Tr in adjacent surface and bubble aeration reactors in the same plant were about 5°C. Large enough Tr and Ta variations exist at these local plants to affect MLSS settling test results. The MLSS settling test cylinder environment and meteorological conditions have a direct influence on Ts during batch settling tests. Direct solar radiation increases the average Ts by 4.3°C, or by 0.15°C per minute, during a 30-minute MLSS settling test duration. This Ts change leads to a sludge volume index (SVI) change of 63 mℓ/g, at an average SVI decrease of 14.8 mℓ/g per 1°C Ts increase. Changes to other parameters include an initial settling velocity (ISV) increase of about 0.12 m/hr for every 1°C Ts increase, together with a clarified supernatant turbidity increase of about 1.4 formazine nephelometric unit (FNU) for every 1°C Ts increase. Ts adjusts towards Ta before and during a batch MLSS settling test, thereby influencing MLSS settling results. Compensation for Ts variations during routine MLSS settling tests is nevertheless not reported as a common practice. To some extent, this is due to a lack of temperature-controlled MLSS settling test equipment. An automated MLSS settling meter demonstrates a semi-continuous on-line method to determine settling parameters in situ at the operational Tr of a full-scale plant. A basic polynomial fits 11 MLSS settling parameters that indicate in most instances improved MLSS settling at increased Tr. The average SVI decreases by 14.8 mℓ/g for every 1°C Tr increase. Similarly, for every 1°C Tr increase, the maximum settling velocity (u_max) increase is 0.1 m/hr, and the time to reach maximum settling velocity (t_umax) decreases by 2.4 minutes. The incremental 5-minute duration average settling velocities increase over the first 15 minutes of a MLSS settling test, as the MLSS concentration decreases and the Tr increases. This direct incremental settling velocity trend with Tr is reversed between 15 and 30 minutes, as the average 5-minute MLSS settling velocity increases at a reduced Tr. The inclusion of Tr in MLSS concentration-based settling best-fit correlations with SVI, u_max, and t_umax improves the coefficient of multiple determinations (R2) by an average of 0.32. Best-fit SVI models with u_max and t_umax have R2-values of 0.90 and 0.95 respectively. The developed models are only valid for the individual reactor MLSS conditions within the experimental parameter ranges. The main contribution of this study is to present temperature-based MLSS settling models. These models illustrate that an automated on-line MLSS settling meter is suitable to identify and model temperature related MLSS settling data with minimal experimental effort. A suitable approach is provided to improve the reliability of MLSS settling data, as effects of short-term temperature variations can be practically eliminated from settling test.