Successive parabolic interpolation as extremum seeking control for microbial fuel & electrolysis cells

Abstract

Microbial Fuel Cell (MFC) power production and Microbial Electrolysis Cell (MEC) organic production depend strongly on their dynamic environment conditions, like inlet substrate concentration, temperature, etc. This work presents a discrete extremum seeking controller to quickly tune the MFC and MEC electrical settings in order to achieve maximum performance irrespective of these dynamic environment conditions using the successive parabolic interpolation iteration scheme. The controller converges in about 3.5 days within 5% of the cell's maximum performance and in about 5.4 days within 1% for an established MFC model. The proposed discrete parabola controller converges 3x faster than the state-of-theart controllers without requiring a time-consuming calibration procedure. Equally fast convergence speed is achieved on a MEC model.

Using human-in-the-loop optimization for guiding manual prosthesis adjustments: a proof-of-concept study

“…The algorithm uses gradient descent to guide the first parameter combinations and then uses successive parabolic optimization once a sufficient number of parameter combinations have been tested. These techniques are based on similar techniques adapted from previous studies ( Koller et al, 2016 ; Molderez et al, 2017 ), where the goal was to find the local minimum of an objective function, similar to a ball rolling toward the lowest point of a valley. After testing the first combination of shoe heel height and pylon height, two neighboring combinations within the grid of all possible shoe heel height and pylon height combinations were randomly chosen to test.…”

A Current-Driven Six-Channel Potentiostat for Rapid Performance Characterization of Microbial Electrolysis Cells

“…To characterize the MEC polarization curve quickly yet precise, an algorithm is required that starts from a broad measurement interval and efficiently narrows down the sampling range to the region of interest, i.e., the region where the performance metric is the largest. The used search algorithm is an improvement from [33]. Every iteration of this algorithm contains two steps.…”

“…more, take the one closest to mean(E l , E r )} 6: minimum measurement resolution. This algorithm contains two improvements in comparison to [33]. In the first step, the original algorithm determined the next characterization point based on the maximum of a fitted parabola through the previous three measurement values.…”

“…Measurement samples are only considered to have a different value, if |J i − J j | > J th (Algorithm 1, lines 7-11). Thus, in comparison with [33], the probability of a flipping error in worst-case scenario is strongly reduced because of this noise masking, equalling: P e = P(J k−1 − J k > J th ). Configuring J th σ k , however, reduces the obtained accurarcy in J * as more and more samples are treated equally.…”

“…where N = E r − E l is the original characterization interval length. As a result, in comparison with [33], worst-case convergence remains superlinear.…”