Thermal hydrolysis pretreatment (THP) holds promise to intensify the solid handling capacity of mesophilic anaerobic digesters (AD) by improving the sludge biodegradability, pumpability, dewaterability, and biosafety. Comparing to conventional AD, the digester feedstock total solids (TS) concentration can be doubled from 5 to 10 %, and the solids retention time (SRT) can be halved from 30 to 15 days after THP is implemented, which directly enables existing AD to quadruple the loading rate capacity. Although the AD feedstock TS is hard to further increase due to its viscosity limit and the potential to cause free ammonia inhibition, the digester SRT has been believed to have room for further process intensification. To assess the potential impact of reducing SRTs on the sludge dewaterability and digestibility, this study compared the performance of three pilot-scale mesophilic AD reactors operated at SRTs of 10, 12.5, and 15 days, respectively. These digesters were fed with 9% TS sludge containing primary sludge (PS) and waste activated sludge (WAS) blended in a dry mass ratio of 3:2, which was processed in a pilot-scale CAMBI^TM THP unit operated at a steam pressure of 6 bar (equivalent to 165 ^oC) for 30 minutes. This study revealed that shortening SRTs from 15 to 10 days linearly decreased volatile solids reduction (VSR) from 60.7% to 53.6%, highlighting the reduction in hydrolysis rate as the major drawback. However, other key AD performance parameters such as methane yield and sludge properties in terms of rheology and dewaterability largely remained unchanged.

Anaerobic digestion (AD) has been implemented for resource recovery using agro-waste and waste-sludge to produce biogas and nutrient rich digestate, serving as a promising tool towards circular economy. Developing viable process control strategies for ADs has been held back due to its complicated anaerobic microbial dynamics. Lack of durable electrochemical sensors leads to futile attempts of real-time in situ monitoring the varying physical-chemical and biochemical status in ADs. To address this challenge, mm-sized microelectrode array (MEA) sensors have been recently developed with high durability and sensitivity. Meanwhile, input-output models using easily measurable variables (e.g., pH, redox potential (ORP)) to derive complex target variables (e.g., biogas, alkalinity), termed as soft sensors have gained recognition.

This study integrates high-fidelity in situ MEA sensing with soft sensing to achieve real-time system control and predict biogas yield in AD systems. Specifically, five MEA sensors targeting temperature (T), ORP, conductivity (EC), NH₄⁺, and pH were deployed in a lab-scale AD system (volume: 2L). To acquire the training data, system dynamics indicators alkalinity (A_lk), Volatile fatty acids (VFAs), phosphates (H_(3-x)PO₄^x-), biogas flow (Q_B) and biogas composition (BC) were manually measured. The data was then used to develop a soft sensor using support vector machine learning algorithms. This soft sensor consists of predictor variables (pH, ORP, EC, NH₄⁺, and T) and response variables (A_lk, VFA, H_(3-x)PO₄^x- , Q_B and BC). Obtaining these variables in a real-time mode enables system visualization, provides early warning of malfunction, and eventually advances the AD resilience as an efficient renewable energy source.

We developed a modeling platform derived from the anaerobic digestion model No. 1 (ADM1F, F stands for fast because of an efficient numerical solver) to predict the performance of a two-phase anaerobic dynamic membrane bioreactor system. In contrast to the original ADM1, modifications, such as separation of the solids retention time and hydraulic retention time, were included in ADM1F as the two-phase system utilizes dynamic membranes to achieve solid-liquid separation. Other modifications include adding the descriptions of carboxylic acid inhibition on methanogenesis, the effect of operating temperature on kinetics, and the Thiele Modulus determining the mass transport-reaction rate relationship. A lab-scale system including a 6.2-L, 39°C first phase and a 43-L, 20°C second phase was operated to co-digest food waste and sewage sludge. Over 350 days of operation, the first phase achieved an average volatile fatty acid yield of 0.47 g VFA/g volatile solids (VS)_fed. The two-phase system overall accomplished an average methane yield of 0.74 m³/kg VS_fed. After calibration, the model predicted the methane production rate with an average relative error of 2% compared to the experimental results and captured fluctuation caused by feedstock variations. Modeling results indicated that 1) co-digesting food waste with sewage sludge drastically increased the hydrolysis rates compared to mono-digestion of food waste; 2) the dynamic membrane in the second phase was not substrate mass transport-limited; 3) Increasing the operating temperature from 20°C to 30°C would only marginally enhance the performance of the second phase but decreasing it to 10°C would substantially deteriorate the performance.

Establishment of sustainable wastewater reuse practices requires the development of treatment technologies that are both low-impact and safe. Anaerobic membrane bioreactors (AnMBRs) are an attractive option in this regard, but many outstanding questions associated with emerging microbial contaminants remain. The current study assessed effluent microbial community and antibiotic resistance gene (ARG) profiles of AnMBRs treating two vastly different wastewater types: municipal wastewater and poultry slaughterhouse wastewater (PSW). This evaluation was based on the analysis of microbial communities for potential pathogens and the quantification of ARGs associated with sulfonamides, tetracyclines, and β-lactams. For the system to which municipal wastewater was introduced, a surge in microbial diversity was accompanied by a progressive increase in most targeted ARGs. Effluent sul1, sul2, and intI1 abundances increased by 1-2 orders of magnitude, reaching levels near those of the influent wastewater. tetC and tetQ also increased over time but with a sustained 3-log removal. The system exposed to PSW had an initial peak in sul1, sul2, ampC, and intl1 genes, followed by a gradual decrease (of at least one order of magnitude) after extended treatment. This observation coincided with a sudden drop in microbial diversity at initial exposure. These findings suggest that a finite horizontal gene transfer event likely occurred (for PSW). Both systems showed a progressive increase in potential pathogen abundance after extended exposure. The results of this work highlight the differences in effluent profiles that can be induced by various wastewater sources and, resultantly, the importance of the targeted assessment of their microbial safety.

Water Resource Recovery Facilities (WRRFs) with anaerobic digestion can co-digest food waste (FW) with sludge (S) to enhance energy recovery. Biomethane Potential (BMP) tests are commonly used to estimate the increase in methane production. However, these tests are time-consuming, so modeling can be used to predict methane production and the impact of implementing co-digestion. This paper presents a modeling-based stability assessment strategy for co-digestion blends using a ternary stability plot based on the substrates’ biochemical composition (Cook et al., 2017, http://doi.org/10.1016/j.watres.2017.01.027). To evaluate this strategy, the results of a modified BMP test for a novel two-phase anaerobic dynamic membrane bioreactor treating FW and sludge are also reported. Five blends of FW (pre- and post-consumer) and sludge (primary and secondary) were tested. The FW25 S75 blend had the highest methane potential (381 NmL CH4/gVS) and required a low NaOH dose (0.05 g/gVS) to maintain the pH around 7, indicating high stability. This result was consistent with the ternary plot prediction. When assessing stability, some blends fell in ambiguous areas on the plot where stable and unstable predictors coexisted. Therefore, we are re-evaluating the stability parameters developed by Cook et al. (2017) to determine their impact on the stability prediction model. Updated ternary plots are being developed using a modified Anaerobic Digestion Model No. 1. WRRFs will benefit from this stability assessment strategy when considering co-substrates by predicting if the selected co-digestion blends will result in stable digestion without running time-consuming BMP experiments.