Methodology

8.3.1 Flow chart of the overall modified kinetic models process

Figure 8.1 shows the overall steps involved in the development of the modified kinetic model for batch, continuous substrate feed, and continuous feed with biological carrier material processes. There are four steps in the overall process as shown in Figure 8.1;

steps one and two identify and test the kinetic parameters using batch experiments, and these values are used in steps three and four. The CSTR kinetic, or “black box” model is used to identify the modified kinetic model under a continuous substrate feed as shown in steps three and four. The details of each step are discussed in Figure 8.1.

Step 1: The kinetic model for substrate removal and biomass growth rate prediction is described by the Monod equation (as detailed in Chapter 2). Batch testing is performed for initial nitrate concentrations of 40, 80, 120, and 160 mg-N/L under 1 L. Hydrogen gas is supplied continuously at 5 mL/min. The initial biomass concentrations were in the range of 450 - 550 mg-VS/L. The estimated kinetic values, such as maximum specific growth rate (𝜇_𝑚𝑎𝑥), half saturation concentration or Monod constant (Ks), yield coefficient (Y) using a Lineweaver-Burk plot and Monod model, are achieved.

Moreover, the model is validated with the results from the batch experiment. If the model fits the experimental data well, the Monod model will be used to describe substrate removal and biomass generation in the HD process. If the model does not fit the experimental data, a modified Monod model should be considered as shown in Step 2.

Step 2: The modified Monod model (i.e., double Monod model) considers two denitrification processes: nitrate to nitrite and nitrite to nitrogen gas. The nitrite has a significant effect on the degradation rate of nitrate. Therefore, the nitrite factor is important for the development of the modified model. Batch testing is performed under two substrates with controlled nitrate and nitrite concentrations. Initial nitrate concentrations are various at 40, 80, 120, and 160 mg-N/L, while nitrite concentrations are 20, 40, 80, and 120 mg-N/L. The initial biomass concentrations are found to be in the range of 450 - 550 mg-VS/L. Moreover, this step also considers the hydrogen gas limitation; as the hydrogen gas supply in the reactor is important for hydrogenotrophic denitrification process. As discussed in the previous chapter, hydrogen gas is expensive and has a relatively low solubility; more efficient utilization of hydrogen gas in the HD process is necessary in order to overcome operational challenges. Batch testing is performed under 5, 10, 15 and 20 mL/min of hydrogen gas supply. The estimated kinetic values, such as maximum specific growth rate (𝜇_𝑚𝑎𝑥 ), half saturation concentration or Monod constant (Ks), yield coefficient (Y) using a Lineweaver- Burk plot and Monod model are achieved. The double Monod model under two substrate limitations and a hydrogen gas limitation is developed and validated with data from the batch experiments. If the model fits the experimental data well, the double Monod model will be use to descript substrate removal and biomass generation in this HD process. If the model does not fit the experimental data, additional parameters should be used and validated again with experimental data until achieve a complete modified Monod model is generated that can explain all of the HD process mechanisms active in the batch experiments.

Step 3: The continuous substrate feed model is considered in this step. The CSTR, or

“black box” model is used. The parameters considered include flow rate (Q), inlet and outlet substrate (Sinlet, Soutet) and volume (V). The combination double Monod model and CSTR model is developed and validated with experimental data, which was generated by continuous nitrate feed into a 2 L of cylinder reactor, with an HRT of 12 hours and a H2 gas supply of 5 mL/min. If the modified model fits the experimental data well, the modified model will be used to describe substrate removal and biomass generation in this HD process. If the model does not fit the experimental data, additional

parameters should be used and validated again with experimental data until a complete modified Monod model is generated that can explain all of HD process mechanisms active in the continuous substrate feed experiment.

Step 4. The combination double Monod model and CSTR model with a carrier material is developed in this step. According to the results from Chapter 6, the Biofix biological carrier material is highly effective and enhances HD performance. Therefore, the kinetic model is developed with Biofix carrier material. The carrier factor coefficient and the maximum biological attachment capacity on Biofix are examined. The estimated values from the modified model are identified and validated with data from a continuous nitrate feed experiment in the Biofix reactor (Chapter 6). If the modified model fits the experimental data well, the modified model will be use to descript substrate removal and biomass generation in the HD process under continuous substrate feed with Biofix as carrier material. If the modified model does not fit the experimental data, the additional parameter should be used and validated again until achieve the completely modified Monod model under continuous substrate feed. The details of each parameter and the estimation of the kinetic values were summarized in Chapter 2.

Flow chart of overall process for modified kinetic models

Figure8.1 Overall flowchart for developed modified kinetic models

Monod model

Balance of Biomass Balance of Substrate

Validation (batch data)

Yes Batch experiment (Batch setup)

Kinetic parameters Step 1

Modified two step Monod model

Validation (batch data)

No

Balance of Biomass Balance of Substrate

No

Yes

Limitation of NO₃-N Limitation of NO₂-N Effect of H₂ flow modified parameters

Modified Monod model with continuous feed experiment

Modified model with carrier materials D, Q, V

Balance of Biomass Balance of Substrate

Validation

(Cont. feed data)

Yes

Balance of Biomass Balance of Substrate

Validation

(long term operation)

No

Yes Conclusion

modified parameters ^H2 limitation

Bio-fix

Non-carrier VS carrier material

modified parameters Maximum attraction Conclusion

No Step 2

Step 3

Step 4

8.3.2 Experimental setup

In order to identify the potential and kinetic values for hydrogenotrophic denitrification with nitrate as an electron acceptor and hydrogen gas as an electron donor. There experiment were separated into two types: batch tests (Steps 1 and 2) and continuous feed with substrate tests (steps 3 and 4).

8.3.2.1. Batch experiments

Batch tests were preformed using a 1 L cylindrical reactor with an L/D of 4. The synthetic medium was prepared as in Chapter 3 with various initial nitrate concentrations of 40, 80, 120, and 160 mg-N/L. Other batch tests were performed for nitrite factor using initial nitrite concentrations of 20, 40, 80, and 120 mg-N/L. The biomass taken from the HD reactor with Biofix carrier material (Chapter 7) after the nitrogen removal performance stabilized. A mixed-culture of Hydrogenophaga spp., Thauera spp., Rhodocyclaceae, Alcaligenaceae, and Rhodobacteraceae was detected in the HD reactor (Chapter 7), and the removal performance was found to be 95%; there was no nitrite accumulation in the reactor.

Figure 8.2 Batch experimental set up for identifies the kinetic parameter in step 1 and step 2

The initial biomass concentration was found to be in the range of 450 – 550 mg-VS/L range. Hydrogen gas was supplied continuously at 5, 10, 15, and 20 mL/min. A liquid sample was taken every 5 hours (approximately) and stored in the freezer (at -18℃)

X-Sampling

N-Sampling Reaction Preparing

until water quality analysis could be performed. The nitrate concentration and nitrite concentration were analyzed and calculated based on the analytical methods in Chapter 3. Biomass sample of 5 mL were taken concurrently with the liquid samples for biomass concentration analysis, which was accomplished via the dry weight method in terms of volatile solids (VS). Samples were taken continuously until the nitrate and nitrite concentrations reached 0 mg-N/L. Batch experiments were conducted for Steps 1 and 2, as illustrated in Figure 8.2.

8.3.2.2. Continuous substrate feed experiments

A CSTR, or “black box” model was used to identify the modified the kinetic model.

There are two types of modified models: one each for reactor with and without biological carrier materials. The cylindrical HD reactors measured 9 cm diameter and 31.45 cm in height; the working volume was approximately 2 L. They were fed continuously with nitrate concentrations of 40 mg/L. There was a 5 mL/min continuous hydrogen gas supply and the HRT was 12 hours. The temperature was controlled at 32

± 0.5 ℃ as used under optimized conditions (Chapter 5). The biomass inside the reactor was collected every day in order to investigate the biomass generation rate. These data were used for model validation under a continuous substrate feed (Step 3). Another HD reactor under the same conditions was operated with a biological carrier (Biofix) in the reactor. The operating conditions and results were shown in Chapter 6 for HRT changes from 12 to 4 hours. Lastly, the biomass concentration inside the reactor was examined via batch testing before the HRT was change; these data were used for model validation in step 4.

8.3.4 Measurement of microbial growth

There are 3 main techniques for measuring growth in the cell mass or cell number, including direct, indirect cell counting and visually (using instrumental). In this section, the direct counting technique is used to measure the dry weight of the biomass, or cell material, and give the volume of the culture. Biomass should be dried via the removal of water or medium. The biomass concentration is measured according to APHA (1998) in terms of volatile solids (VS). Volatile solids represent organic matter by the measured

weight lost after ignition at 550℃ for approximately 1 hour. The VS were calculated following the formula introduced below.

Biomass concentration

mg − VS/L = (A−B) x 1000

volume of sample (mL) (8.1) Where, A is weight of residue + dish before ignition (mg), B is weight of residue + dish after ignition (mg)