Time series analysis and long short-term memory (LSTM) network prediction of BPV current density

Okedi, Tonny I.; Fisher, Adrian C.

doi:10.1039/d0ee02970j

Cited by 13 publications

(14 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the second, gradual photoresponse was not recovered, and overall photoresponse remained low. The dip in current at dawn was absent under the 3h:3h light-dark regime confirming previously reported results 12 . Iron stress expedites activation of exoelectrogenesis, amplifies mean current output, but reduces light gradients In line with previous results, Fe (-)|Air BPVs had higher current densities versus the control, but a markedly reduced photoresponse, Fig.…”

Section: Current Density Profilessupporting

confidence: 92%

“…Patterns previously reported when operating under a diel (24h) light-dark period were observed, namely a dip and recovery in current density at dawn (inset, Fig. 2d), and a period-to-period decay in the magnitude of the photocurrent 12 . Following the dip, photoresponse was two-sloped, with an initial steep spike in current, followed by a more gradual increase over the remaining duration of the illuminated interval.…”

Section: Current Density Profilessupporting

confidence: 76%

“…However, due to knowledge gaps, the non-linear and nonstationary nature of bioelectrochemical processes, and difficulties in the in-situ measurement of key variables, it has proved difficult to develop robust first-principle computational models. To overcome these challenges, the authors recently applied deep learning, in particular Long Short-Term Memory (LSTM) networks, to predict the one-step-ahead seasonal current density in BPVs using a three step process: (1) decomposing the observed current density into a trend, seasonal and irregular component via Seasonal and Trend Decomposition using Locally Estimated Scatterplot Smoothing or LOESS (STL); (2) training the LSTM network with the seasonal component; and (3) predicting the one-step-ahead seasonal current density using the trained network 12 .…”

Section: Introductionmentioning

confidence: 99%

“…In order to forecast the observed current density, a separate model for the erratic and difficult-to-predict irregular component is required. The LSTM forecasts were shown to be unsatisfactory without the removal of the irregular component 12 . Furthermore, the physical meanings of the trend, seasonal and, particularly, irregular components obtained via STL are unclear.…”

Section: Introductionmentioning

confidence: 99%

“…It is therefore imperative to find an alternative decomposition approach that results in more regular, physically meaningful sub-components in order to develop more robust models via LSTMs or other time series modelling techniques. To achieve this, the authors previously proposed Empirical Mode Decomposition (EMD), a time series decomposition method applied when analysing signals using the Hilbert-Huang transform (HHT) 12 .…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Green notes: The rhythms of cyanobacteria exoelectrogenesis as de-composed by the Hilbert-Huang transform

Okedi¹,

Yunus²,

Fisher

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Electrons from cyanobacteria photosynthetic and respiratory systems are implicated in current generated in biophotovoltaic (BPV) devices. However, the pathway that electrons follow to electrodes remains largely unknown, limiting progress of applied research. Here we use Hilbert-Huang transforms to decompose Synechococcus elongatus sp. PCC7942 BPV current density profiles into physically meaningful oscillatory components, and compute their instantaneous frequencies. We develop hypotheses for the genesis of the oscillations via repeat experiments with iron-depleted and 20% CO2 enriched biofilms. The oscillations exhibit rhythms that are consistent with the state of the art cyanobacteria circadian model, and putative exoelectrogenic pathways. In particular, we observe oscillations consistent with: rhythmic D1:1 (photosystem II core) expression; circadian-controlled glycogen accumulation; circadian phase shifts under modified intracellular %ATP; and circadian period shortening in the absence of the iron-sulphur protein LdpA. We suggest that the extracted oscillations may be used to reverse-identify proteins and/or metabolites responsible for cyanobacteria exoelectrogenesis.

show abstract

Section: Current Density Profilessupporting

confidence: 92%

Section: Current Density Profilessupporting

confidence: 76%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Green notes: The rhythms of cyanobacteria exoelectrogenesis as de-composed by the Hilbert-Huang transform

Okedi¹,

Yunus²,

Fisher

2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

Differentiation characteristics of karst vegetation resilience and its response to climate and ecological restoration projects

Chen

Wang

et al. 2023

Land Degrad Dev

View full text Add to dashboard Cite

In light of the recent pressure from global warming, extreme drought events, and deleterious human activity, the strength and long‐term change trends of vegetation in karst regions—in terms of their resistance to external disturbances—have not been studied systematically. Therefore, herein, we quantified the vegetation resilience and its nonlinear change trends in south China karst under different environmental gradients by measuring the lag‐1 autocorrelation to time‐series Normalized Difference Vegetation Index (1990–2018), clarifying the driving forces of vegetation resilience changes. It was shown that the vegetation resilience change in south China karst was not monotonous. In the first stage (pre‐2002), precipitation and warming promoted the increase of regional vegetation resilience (slope = −0.045, p < 0.0001). In the second stage (during 2002–2010), the increasing trend of vegetation resilience was not obvious and vegetation resilience was difficult to keep up with vegetation productivity, indicating the time‐lagged effect of ecological restoration projects to vegetation resilience. In the third stage (post‐2010), due to the continuous advancement of ecological restoration projects, vegetation resilience increased significantly and had the largest amplitude (slope = −0.128, p < 0.0001). Simultaneously, under different environmental gradients, vegetation resilience showed significant differentiation characteristics. In comparison to non‐karst regions, increases in the vegetation resilience were more obvious in karst regions especially in the post‐2010. With increases in the soil depth, the vegetation resilience exhibited an increasing trend, indicating its dependence on soil. At slopes >25°, the vegetation resilience increased most obviously and the resilience of meadows was the largest, which can be the preferred vegetation type for ecological restoration projects. This research provides another perspective to understand karst vegetation ecosystem and the results will facilitate the protection of karst ecosystems.

show abstract

A Review of Long Short-Term Memory Approach for Time Series Analysis and Forecasting

Kader

Yusof

Khalid

et al. 2022

Lecture Notes in Networks and Systems

View full text Add to dashboard Cite

Time series analysis and long short-term memory (LSTM) network prediction of BPV current density

Abstract: Biophotovoltaics (BPVs) have increasingly gained interest due to their potential to generate low-carbon electricity and chemicals from just sunlight and water using photosynthetic microorganisms. A key hurdle in developing commercial...

Cited by 13 publications

References 29 publications

Green notes: The rhythms of cyanobacteria exoelectrogenesis as de-composed by the Hilbert-Huang transform

Green notes: The rhythms of cyanobacteria exoelectrogenesis as de-composed by the Hilbert-Huang transform

Differentiation characteristics of karst vegetation resilience and its response to climate and ecological restoration projects

A Review of Long Short-Term Memory Approach for Time Series Analysis and Forecasting

Contact Info

Product

Resources

About