(293g) Sensitivity Analysis and State Estimation for Microalgal Photobioreactor Systems

Microalgae are photosynthetic microorganisms, which have an ability to produce large amounts of oil that can be used directly as a high value bioactives or be used to synthesize biodiesel. The oil content in microalgae ranges from 15% to 77% depending on species and culture conditions. Although the oil production rate in microalgae is strain dependent, it has several advantages as a feedstock for biodiesel because of high growth rate and the ability of producing large amounts of lipid. However, biodiesel from microalgae is not economically competitive compared to biodiesel from conventional plant sources or petrodiesel. For the economic competitivenesss, optimal model-based control strategy is required to increase the rate of microalgae growth and the amount of stored oil.

However, there are some difficulties in applying model-based control to microalgal bioreactor systems. Microalgal systems are highly nonlinear and some of the parameters and states are difficult to measure directly or estimate. Furthermore, metabolism inside the cells makes process response very slow and its effect on process is difficult to understand.

In this study, a first principles ODE model for microalgae growth and neutral lipid synthesis related with photo effect is investigated for the purpose of maximizing the rate of microalgae growth and the amount of neutral lipid. The model follows the assumption of Droop model which explains the growth as a two-step phenomenon; the uptake of nutrients is first occurred in the cell, and then use of intracellular nutrient to support cell’s growth. In practice, microalgal bioreactor systems are complex and highly nonlinear. For this case, which has uncertainties in the model and has model-plant mismatches, robust model-based control using particle filter is studied. For this purpose, sensitivity analysis is performed to determine which parameters have a negligible effect on the model predictions. For state estimation, state variables are divided into frequent but inaccurate or accurate but infrequent measurements for same quality variable and Bayesian approach is exploited to account for multiple-source observations. To enhance the robustness, a practical Bayesian fusion formulation with time-varying variances is proposed and observation validity is taken into account. The Bayesian model calibration strategy is finally implemented by using a sequential Monte Carlo sampling approach based particle filter for simultaneously dealing with systematic and nonsystematic errors (i.e. bias and noise).