(628g) Feed-Forward Control of Dynamic High-Density Microalgae Cultures Using Model Based Predictive pH Control and a Novel Biomass Sensor

This talk will present the critical problem of accommodating proton imbalance during algae growth, measurement of stoichiometry and using these constraints in a process control model to achieve high-density algae growth.

We will present efforts that combine an algal culture growth model with dynamic environmental culturing conditions (e.g. pH, diurnal sunlight intensity, weather, available nutrients, dissolved inorganic carbon and temperature) into a nonlinear model based on a predictive process control system. We seek to implement our nutrient based metabolic pH control strategy along with real-time biomass concentration measurements for large-scale high-density algae cultivation for biofuel production.  Modern parameter estimation methods such as nonlinear Kalman Filter technique are employed to build an adaptive process control model by simultaneously updating model parameters from experimental measurements. Our group has quantitatively studied the stoichiometric proton imbalance in phototrophic algae growth for the development of a control model based on providing incremental additions of reduced and oxidized inorganic nitrogen sources to the culture. A novel optical density sensor was developed for feeding real-time growth measurements of the high density culture into our model. Our goal is to use OD sensor measured data, along with simultaneously acquired Internet based weather prediction information, to estimate the system’s state variables, feeding into our process model to deliver optimized feeding.  Following that, nutrient/water pump and CO₂ feed rates are optimized to match the growth rate (via stoichiometric mass balances) while keeping the pH/evaporation in balance. The main control system and model's objective is to maximize biomass and liquid hydrocarbon (objective function) growth rates for maximum economic viability while providing culture stability (inequality constraint) under the constraints of continuous high-density culturing outdoors.

pH instability is an inherent obstacle in the growth of continuous high-density microalgal cultures.  Our lab has undergone extensive quantitative measurements of the proton imbalance ratio, φ (phi), which is the result of the metabolism of different nitrogen sources.  The feeding of two common nitrogen sources, nitrate, a fully oxidized form of N, and ammonium, a fully reduced form of N, results alkalinization and acidification of the media respectively as they are metabolised. Common methods for controlling the pH in bioprocessing are based on brute-force (acids and bases) that cause accumulation of toxic counter-ions or use pH buffers which are cost prohibitive. We present an integrated strategy based on varying the stoichiometry of the two N-species within the feed-forward control (based on mass balances) implemented successfully over the past 3 years of bioreactor runs.  Due to the complicated change of pH in such processes, dedicated experiments were performed to quantitatively study the relationship between the amount of imbalanced protons and nitrogen assimilation. It was found that nitrate produces a net proton imbalance of -1 H⁺ and ammonium produces +1 H⁺. Such metabolic studies for various nitrogen containing molecules have been performed in microalgae Chlorella vulgaris and Chlamydomonas reinhardtii cultures, but are expected to be generally applicable to most eukaryotic microalgae because of their relations to fundamental biochemical process. In addition to the contribution from nitrogen assimilation on imbalanced pH, though not dominant, there are also contributions from other metabolic process, such as secretion of organic acid into the culture medium. As a result, by combining the contribution of both nitrogen assimilation and other metabolic products, a mixed nitrogen source of ammonium and nitrate was used to minimize the resulting proton imbalance; however, Chlorella vulgaris and Chlamydomonas reinhardtii were shown to preferentially utilize ammonium when both ammonium and nitrate were provided in the medium. Experimental results have shown that ammonium concentrations exceeding 9% of total nitrogen (0.3 gN / L) will result in a pH drop too low to sustain additional growth, while the same media could support healthy cultures if controlled by a pH stat. In order to achieve ammonium concentrations in the media exceeding 9%, our feed-forward model must utilize incremental additions of ammonium in the presence of nitrate.  Batch cultures have enabled growth to concentrations as high as 5 gDW/L on 0.3 gN/L, demonstrating that a fed-batch nitrogen strategy is necessary to balance the culture pH.  A prototype LabView program was generated to execute this feed-forward predictive control loop and has been successfully tested with preliminary validation. Due to differences in gene regulation in prokaryotic cyanobacteria, the behavior of nitrogen assimilation in Synechococcussp. PCC 7002  is also under separate investigation and suggest that additional control parameters would be needed for cyanobacteria.  

One major challenge to both lab-scale and industrial microalgae culturing systems is to monitor the growth of high-density cultures in a quick and efficient manner to enable immediate feedback control. Biomass concentration is measured using an online high-density OD sensor which allows for the specific growth rate to be measured over desired time intervals. We have developed an inexpensive sensor to monitor the growth of algae geared toward model based process control. One of the biggest limitations using OD as a proxy for biomass measurements is light saturation (or more precisely lack of sensitivity of measurement as nearly all of the light becomes attenuated). Referring to Beer-Lambert law the alternative to diluting a sample or using reduced absorbance region of the spectrum is to reduce the light path length. We have demonstrated this in a simple flow cell equipped with an LED light source with a dominant emission wavelength of 525 and peak emission wavelength of 528 to mitigate interference from chlorophyll molecules while remaining in the scattering regime.