Dynamic energy system modeling using hybrid physics-based and machine learning encoder–decoder models

Machalek, Derek; Tuttle, Jacob F.; Andersson, Klas; Powell, Kody M.

doi:10.1016/j.egyai.2022.100172

Cited by 19 publications

(4 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The desired metrics are more like composite functions of these directly measurable and well-defined quantities, and obtaining empirically a general form for the relevant equations is often not straightforward and can require significant expert knowledge and experience. Thankfully, a variety of other methods have built on or expanded beyond the original approach to incorporate information on chemical kinetics, [112,115,116] thermochemical/thermophysical information, [115,[117][118][119][120] mass/energy balances, [121][122][123] reactor engineering dynamics, [124,125] computed interatomic potentials, [126] and optics. [127] Many of these expanded approaches enable embedding of chemical information and may no longer be restricted to only finding solutions to and/or fitting PDEs with a previously known form.…”

Section: A Way Forwardmentioning

confidence: 99%

“…Thankfully, a variety of other methods have built on or expanded beyond the original approach to incorporate information on chemical kinetics, [ 112 , 115 , 116 ] thermochemical/thermophysical information, [ 115 , 117 , 118 , 119 , 120 ] mass/energy balances, [ 121 , 122 , 123 ] reactor engineering dynamics, [ 124 , 125 ] computed interatomic potentials, [ 126 ] and optics. [ 127 ] Many of these expanded approaches enable embedding of chemical information and may no longer be restricted to only finding solutions to and/or fitting PDEs with a previously known form.…”

Section: A Way Forwardmentioning

confidence: 99%

See 1 more Smart Citation

Leave No Photon Behind: Artificial Intelligence in Multiscale Physics of Photocatalyst and Photoreactor Design

Loh,

Wang,

Mohan

et al. 2024

Advanced Science

View full text Add to dashboard Cite

Although solar fuels photocatalysis offers the promise of converting carbon dioxide directly with sunlight as commercially scalable solutions have remained elusive over the past few decades, despite significant advancements in photocatalysis band‐gap engineering and atomic site activity. The primary challenge lies not in the discovery of new catalyst materials, which are abundant, but in overcoming the bottlenecks related to material‐photoreactor synergy. These factors include achieving photogeneration and charge‐carrier recombination at reactive sites, utilizing high mass transfer efficiency supports, maximizing solar collection, and achieving uniform light distribution within a reactor. Addressing this multi‐dimensional problem necessitates harnessing machine learning techniques to analyze real‐world data from photoreactors and material properties. In this perspective, the challenges are outlined associated with each bottleneck factor, review relevant data analysis studies, and assess the requirements for developing a comprehensive solution that can unlock the full potential of solar fuels photocatalysis technology. Physics‐informed machine learning (or Physics Neural Networks) may be the key to advancing this important area from disparate data towards optimal reactor solutions.

show abstract

Section: A Way Forwardmentioning

confidence: 99%

Section: A Way Forwardmentioning

confidence: 99%

Leave No Photon Behind: Artificial Intelligence in Multiscale Physics of Photocatalyst and Photoreactor Design

Loh,

Wang,

Mohan

et al. 2024

Advanced Science

View full text Add to dashboard Cite

show abstract

“…Similarly, a hybrid ML encoder–decoder architecture is used for dynamic energy prediction at a thermal energy facility in ref. [20]. In the study, different deep learning models anticipate the heat collected by steam and water in a boiler across multiple steps, with a focus on a new hybrid ML model by using an encoder–decoder and based on physics architecture that performs the most accurately overall.…”

Section: Introductionmentioning

confidence: 99%

Machine learning autoencoder‐based parameters prediction for solar power generation systems in smart grid

Zafar,

Che,

Faheem

et al. 2024

IET Smart Grid

View full text Add to dashboard Cite

During the fourth energy revolution, artificial intelligence implementation is necessary in all fields of technology to meet the increasing energy demands and address the diminishing fossil fuel reserves, necessitating the shift towards smart grids. The authors focus on predicting parameters accurately to minimise loss and improve power generation capacity in smart grids, given that accurate parameter prediction is essential for traditional power grid stations converting to smart grids. The authors employ an artificial intelligence‐based machine learning model, namely the long short‐term memory, to predict parameters of a solar power plant. After analysing the results obtained from the long short‐term memory model in graphical visualisation, the model is further improved using two different techniques namely, a convolutional neural network‐long short‐term memory and the authors proposed an autoencoder long short‐term memory. Comparing the results of these models, the study finds that autoencoder long short‐term memory outperforms the convolutional neural network‐long short‐term memory as well as simple long short‐term memory. Thus, the use of artificial intelligence in this study substantially enhances the precision of parameter prediction by augmenting the performance of rudimentary machine learning models, thereby facilitating the attainment of a resilient and resourceful power system that overcomes power losses and ameliorates production capacity in the context of Smart Grids.

show abstract

“…While promising, machine learning models for energy forecasting are not devoid of limitations, including complexity, future data dependency, fixed-length input sequences, challenges in handling longterm dependencies, and the specter of overfitting. Researchers have explored alternative machine learning methodologies, exemplified by hybrid encoder-decoder designs that exhibit exceptional accuracy in predicting dynamic energy patterns in thermal power plants [24][25]. In parallel, LSTM Autoencoder models have demonstrated efficacy in forecasting solar plant production a day in advance, adeptly addressing data uncertainties and distortions [26][27].…”

Section: Introductionmentioning

confidence: 99%

Enhancing Power Generation Forecasting in Smart Grids Using Hybrid Autoencoder Long Short-Term Memory Machine Learning Model

Zafar,

Che,

Ahmed

et al. 2023

IEEE Access

View full text Add to dashboard Cite

This study explores the implementation of advanced machine learning techniques to enhance the integration of renewable energy into smart grids, focusing specifically on predicting solar power generation for the upcoming year. Three distinct machine learning models are employed: the Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (Bi-LSTM) and a hybrid model that combines Autoencoder and Long Short-Term Memory (AE-LSTM). Using real-time solar power production data spanning a year, these models are trained and evaluated using mean absolute error (MAE) and mean squared error (MSE) as performance metrics. Results highlight the superior accuracy of the hybrid AE-LSTM model compared to the LSTM model as well as Bi-LSTM model, attributed to its capability to capture intricate temporal patterns and correlations within the data. This research underscores the significant potential of machine learning techniques, particularly the hybrid AE-LSTM approach, in facilitating the seamless integration of renewable energy resources into smart grids, contributing to more efficient and environmentally conscious power systems. Furthermore, a comprehensive analysis reveals the hybrid AE-LSTM model's capacity to produce superior predictions due to its additional training, solidifying its advantage over models solely reliant on the other model's architecture. In summary, this study demonstrates the effectiveness of advanced machine learning methodologies in revolutionizing renewable energy integration, with the hybrid AE-LSTM model standing out as a promising avenue for enhanced prediction accuracy.

show abstract

Dynamic energy system modeling using hybrid physics-based and machine learning encoder–decoder models

Cited by 19 publications

References 24 publications

Leave No Photon Behind: Artificial Intelligence in Multiscale Physics of Photocatalyst and Photoreactor Design

Leave No Photon Behind: Artificial Intelligence in Multiscale Physics of Photocatalyst and Photoreactor Design

Machine learning autoencoder‐based parameters prediction for solar power generation systems in smart grid

Enhancing Power Generation Forecasting in Smart Grids Using Hybrid Autoencoder Long Short-Term Memory Machine Learning Model

Contact Info

Product

Resources

About