A dual-method approach using autoencoders and transductive learning for remaining useful life estimation

Engineering Applications of Artificial Intelligence, Journal Year: 2025, Volume and Issue: 147, P. 110285 - 110285

Published: Feb. 24, 2025

Language: Английский

---

Artificial Intelligence in Lithium-Ion Battery Predictions

Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 203 - 236

Published: Feb. 28, 2025

In battery development, process from the discovery of a new material to testing performance is prolonged one. This time-consuming can be overcome by implementing Artificial Intelligence (AI). The need for AI in lithium-ion batteries (LIBs) predict specific capacity, efficiency, performance, and durability earlier. It could aid development most reliable advanced with reduced time consumption. Prospects LIBs facilitate predictions terms State-of-Art management systems (BMS) such as states charge (SOC), health (SOH), available power (SoAP), remaining useful life (RUL) various methodologies Machine Learning (ML). During pandemic situation experimentation laboratory may difficult however give an insight implement into research pave way do more innovation their respective fields. An overview applications LIB discussed this article based on review earlier works worldwide.

Language: Английский

---

Enhancing prediction accuracy of Remaining Useful Life in lithium-ion batteries: A deep learning approach with Bat optimizer

Deleted Journal, Journal Year: 2024, Volume and Issue: 2, P. 100003 - 100003

Published: June 1, 2024

Remaining Useful Life (RUL) prediction in lithium-ion batteries is crucial for assessing battery performance. Despite the popularity of deep learning methods RUL prediction, their complex architectures often pose challenges interpretation and resource consumption. We propose a novel approach that combines interpretability convolutional neural network (CNN) with efficiency bat-based optimizer. CNN extracts data features characterizes degradation kinetics, while optimizer refines parameters. Tested on NASA PCoE data, our method achieves exceptional results minimal computational burden fewer It outperforms traditional approaches, yielding an R2-score 0.9987120, MAE 0.004397067 Ah, low RMSE 0.00656 Ah. The proposed model models, as confirmed by comparative analysis.

Language: Английский

---

A Comprehensive Framework for Estimating the Remaining Useful Life of Li-ion Batteries under Limited Data Conditions with no Temporal Identifier

Reliability Engineering & System Safety, Journal Year: 2024, Volume and Issue: 253, P. 110517 - 110517

Published: Oct. 5, 2024

Language: Английский

---

A novel variable activation function-long short-term memory neural network for high-precision lithium-ion battery capacity estimation

Ionics, Journal Year: 2024, Volume and Issue: 30(5), P. 2609 - 2625

Published: March 23, 2024

Language: Английский

---

Green synthesis of polyimide by using an ethanol solvothermal method for aqueous zinc batteries

RSC Advances, Journal Year: 2024, Volume and Issue: 14(22), P. 15507 - 15514

Published: Jan. 1, 2024

Utilizing green solvents and raw materials, U-PIs with different morphologies improved performance were successfully prepared through a solvothermal method.

Language: Английский

---

Development of a DNN Predictive Model for the Optimal Operation of an Ambient Air Vaporizer of LNG

Processes, Journal Year: 2023, Volume and Issue: 11(11), P. 3143 - 3143

Published: Nov. 3, 2023

In this study, we conducted preliminary research with the objective of leveraging artificial intelligence to optimize efficiency and safety entire Ambient Air Vaporizer (AAV) system for LNG (Liquid Natural Gas). By analyzing a year-long dataset real operational data, identified key variables that significantly influence outlet temperature Gas (NG). Based on these insights, Deep Neural Network (DNN) prediction model was developed forecast NG temperature. The endeavor create an effective faced specific challenges, primarily due narrow range fan speeds safety-focused guidelines. To surmount obstacles, various learning algorithms were evaluated under multiple conditions. Ultimately, DNN exhibiting lower values both absolute mean error (MAE) square (MSE) successfully established.

Language: Английский

---

Prediction Model of the Remaining Useful Life of the Drill Bit during Micro-Drilling of the Packaging Substrate

Processes, Journal Year: 2023, Volume and Issue: 11(9), P. 2653 - 2653

Published: Sept. 5, 2023

The packaging substrate plays a significant role in electrical connection, heat dissipation, and protection for the chips. With characteristics of high hardness complex material composition substrates, drill bit failure is an austere challenge micro-drilling procedures. In order to monitor health state predict its remaining useful life (RUL) substrate, improved RUL prediction model established based on similarity principle, degradation rate, offset coefficient. And then, experiment carried out collect axial drilling force through precision measurement platform. Axial signals, which are processed via Wiener filtering method, used analyze effectiveness model. results indicate that, compared curves traditional model, present higher fitting degree with actual curves. average relative errors small stable all groups; values less than 15%, while fluctuation greatly large, maximum value even reaches 74.43%. Therefore, taking rate coefficient into account proper method enhance accuracy Furthermore, reliable theoretical support monitoring bits during also acts as potential improve micro hole processing efficiency substrates.

Language: Английский

---

Optimizing Remaining Useful Life Estimation of Lithium-Ion Batteries: A Particle Swarm Optimization-Based Grey Prediction Model

Journal of Energy - Energija, Journal Year: 2024, Volume and Issue: 72(3), P. 8 - 13

Published: Feb. 9, 2024

Accurate estimation of the age and condition lithium-ion batteries (LIBs) is paramount for their safe economically viable utilization. However, assessing degradation these power units proves challenging due to dependence on various environmental usage factors. In this study, we propose an efficient Particle Swarm Optimization (PSO)-based Grey Theory prediction model determine Remaining Useful Life (RUL) batteries. The proposed utilizes PSO optimize coefficients a grey model, enabling accurate forecasting remaining useful life LIBs. Our results demonstrate that presented outperforms conventional models in terms both accuracy stability. Furthermore, offers simpler predictions compared existing literature. By introducing promising technique, our study contributes precise RUL holds potential applications similar domains. This research serves as significant step towards ensuring effective management utilization LIBs, promoting reliability safety.

Language: Английский

---

Runoff Prediction for Hydrological Applications Using an INFO-Optimized Deep Learning Model

Processes, Journal Year: 2024, Volume and Issue: 12(8), P. 1776 - 1776

Published: Aug. 22, 2024

Runoff prediction is essential in water resource management, environmental protection, and agricultural development. Due to the large randomness, high non-stationarity, low accuracy of nonlinear effects traditional model, this study proposes a runoff model based on improved vector weighted average algorithm (INFO) optimize convolutional neural network (CNN)-bidirectional long short-term memory (Bi-LSTM)-Attention mechanism. First, historical data are analyzed normalized. Secondly, CNN combined with Attention used extract depth local features input weights Bi-LSTM. Then, Bi-LSTM time series feature analysis from both positive negative directions simultaneously. The INFO parameters optimized provide optimal parameter guarantee for CNN-Bi-LSTM-Attention model. Based hydrology station’s level flow data, influence three main models two optimization algorithms compared analyzed. results show that fitting coefficient, R2, proposed 0.948, which 7.91% 3.38% higher than CNN-Bi-LSTM, respectively. R2 vector-weighted 0.993, 0.61% Bayesian (BOA), indicating method adopted paper has more significant forecasting ability can be as reliable tool long-term prediction.

Language: Английский

---