Capacity Prediction Methods Research Articles

Low-rank coalbed methane production capacity prediction method based on time-series deep learning

Coalbed methane (CBM), a novel clean energy source, significantly contributes to reforming the energy supply structure, advancing carbon neutrality, and achieving peak carbon emissions. A precise production forecasting method is essential for understanding coalbed methane development and improving extraction efficiency. This paper proposes an improved multi-channel LSTM production forecasting model considering the influence of production dynamic factors, based on cutting-edge deep learning technology and widely applied production decline curve analysis methods. The model is experimentally tested on CBM wells in the Surat Basin of Australia. In the data processing segment of the model, we have improved the 3σ rule for handling outliers within a normal distribution. This enhancement maintains data integrity, achieves superior denoising, increases prediction accuracy, and significantly reduces model learning time. The test results demonstrate that the improved model exhibits excellent predictive performance for CBM wells, with 87.3 % of wells having a prediction accuracy exceeding 85 % overall. Additionally, the comprehensive multi-channel LSTM model identifies dynamic controlling factors in CBM production by analyzing variations in prediction accuracy with different dynamic factor combinations. This study provides new insights and scientifically rational references for production forecasting in CBM extraction. It's also crucial for effective production strategies and process optimization.

Dynamic data driven load-carrying capacity prediction method for composite laminates with delamination

The delamination of composites has invisibility and can reduce the load-carrying capacity (LCC) of structures. In this paper, a dynamic data driven LCC prediction method for composite laminates considering delamination is proposed. The method is divided into offline and online phases: In the offline phase, finite element modeling, prior information acquisition, and database construction are carried out. In the online phase, the Bayes factor is used to realize the rapid mapping between sensor data and the sample in the database to predict the LCC of the structure. In numerical cases with noise, the prediction errors of LCC are within 2.3% with sensor data, and the introduction of prior information can improve efficiency and accuracy of the prediction. The material parameters and delamination sizes of the identification results are the samples closest to the given solution in the database. In test cases, the information from strain arrays can effectively reflect the delamination growth of the laminate. After three times data assimilation, the prediction errors of LCC are within 10.5%.

An interpretable capacity prediction method for lithium-ion battery considering environmental interference

Predicting the capacity of lithium-ion battery (LIB) plays a crucial role in ensuring the safe operation of LIBs and prolonging their lifespan. However, LIBs are easily affected by environmental interference, which may impact the precision of predictions. Furthermore, interpretability in the process of predicting LIB capacity is also important for users to understand the model, identify issues, and make decisions. In this study, an interpretable method considering environmental interference (IM-EI) for predicting LIB capacity is introduced. Spearman correlation coefficients, interpretability principles, belief rule base (BRB), and interpretability constraints are used to improve the prediction precision and interpretability of IM-EI. Dynamic attribute reliability is introduced to minimize the effect of environmental interference. The experimental results show that IM-EI model has good interpretability and high precision compared to the other models. Under interference conditions, the model still has good precision and robustness.

Unsupervised feature extraction for lithium-ion battery electrochemical impedance spectroscopy and capacity estimation using deep learning method

In this paper, we propose an unsupervised feature extraction technique using a variational autoencoder and generative adversarial network (VAEGAN) that automatically extracts meaningful latent variables from the electrochemical impedance spectra (EIS) of lithium-ion (Li-ion) batteries. These variables accurately reflect the degree of Li-ion battery degradation and are closely related to its capacity. By analyzing the latent variables, it was found that the network can learn the effect of Li-ion battery capacity degradation on the EIS. The extracted latent variables are then used as input features for Gaussian Process Regression (GPR) to estimate the capacity of Li-ion batteries under uneven usage and with unknown historical data. The results show that the capacity prediction method proposed in this paper can significantly reduce prediction errors compared to the current state-of-the-art methods. The method not only provides a more reliable capacity estimation but is also robust to highly noisy EIS data, making it more suitable for practical application scenarios.

Load-bearing capacity prediction of the orthogonal double corrugated SPSW under combined compression and shear load

The orthogonal double corrugated steel plate shear wall (ODC-SPSW) is composed of two corrugated steel plates placed orthogonally and connected by high-strength bolts at the corrugation valley regions. The load-bearing capacity prediction method of the ODC-SPSW subjected to both axial compression and shear should be investigated for engineering design. Firstly, with considering the embedded ODC-SPSW being simply supported on both sides and fixed on other both sides, its compressive elastic buckling load is derived using the equilibrium method, and the formula of the compressive elastic buckling load with a correction coefficient considering the differences in boundary simplification is corrected and verified through the finite element (FE) elastic buckling analysis to obtain the calculation formula of normalized high-thickness ratio. Additionally, the compression load-bearing capacity of the ODC-SPSW is approximately equal to the sum of the compression load-bearing capacity of the embedded ODC-SPSW and edge members. Finally, the load-bearing capacity of the ODC-SPSW under the combined load of the axial compression and shear is studied, and the design method of compression-shear interactive curves is proposed to guide engineering applications.

Digital Twin-Driven Multi-Factor Production Capacity Prediction for Discrete Manufacturing Workshop

Traditional capacity forecasting algorithms lack effective data interaction, leading to a disconnection between the actual plan and production. This paper discusses the multi-factor model based on a discrete manufacturing workshop and proposes a digital twin-driven discrete manufacturing workshop capacity prediction method. Firstly, this paper gives a system framework for production capacity prediction in discrete manufacturing workshops based on digital twins. Then, a mathematical model is described for discrete manufacturing workshop production capacity under multiple disturbance factors. Furthermore, an innovative production capacity prediction method, using the “digital twin + Long-Short-Term Memory Network (LSTM) algorithm”, is presented. Finally, a discrete manufacturing workshop twin platform is deployed using a commemorative disk custom production line as the prototype platform. The verification shows that the proposed method can achieve a prediction accuracy rate of 91.8% for production line capacity. By integrating the optimization feedback function of the digital twin system into the production process control, this paper enables an accurate perception of the current state and future changes in the production system, effectively evaluating the production capacity and delivery date of discrete manufacturing workshops.

Lithium-Ion Battery Capacity Prediction Method Based on Improved Extreme Learning Machine

Abstract Currently, research and applications in the field of capacity prediction mainly focus on the use and recycling of batteries, encompassing topics such as SOH estimation, RUL prediction, and echelon use. However, there is scant research and application based on capacity prediction in the battery manufacturing process. Measuring capacity in the grading process is an important step in battery production. The traditional capacity acquisition method consumes considerable time and energy. To address the above issues, this study establishes an improved extreme learning machine (ELM) model for predicting battery capacity in the manufacturing process, which can save approximately 45% of energy and time in the grading process. The study involves the extraction of features from the battery charge–discharge curve that can reflect battery capacity performance and subsequent calculation of the grey correlation between these features and capacity. The feature set comprises features with a high correlation with capacity, which are used as inputs for the ELM model. Kernel functions are used to adjust the ELM model, and Bayesian optimization methods are employed to automatically optimize the hyperparameters to improve the capacity prediction performance of the model. The study uses lithium-ion battery data from an actual manufacturing process to test the predictive effect of the model. The mean absolute percentage error of the capacity prediction results is less than 0.2%, and the root-mean-square error is less than 0.3 Ah.

Remaining capacity prediction of Li-ion batteries based on ultrasonic signals

ABSTRACT The estimation of Li-ion battery degradation performance is the key to the safe and effective operation. The current estimation methods for Li-ion battery degradation performance are mainly based on indirect electrical characteristic parameters such as voltage and current, which lack the direct characterization of internal materials, which will affect the accuracy and real-time performance of the prediction. Ultrasonic wave can directly characterize the changes in internal material properties of Li-ion battery during charging-discharging cycles. At present, the correlation between ultrasonic signal and Li-ion battery degradation process has been proven, and the relevant ultrasonic characteristics have been analyzed and established. On the above basis, this paper designs a battery degradation performance detection platform based on ultrasonic, which obtains the ultrasonic response signal and further selects four characteristics of battery aging. Aiming at the situation that the prediction starting point cannot be determined due to the possible loss of process data during the experiment, this paper uses random sampling to establish a test set to simulate the unknown prediction starting point, and establishes a battery remaining capacity prediction method based on the combination of the main model and residual correction model. Finally, the experiment proves that the proposed model has high accuracy.

Research on Production Prediction Method of Multi-stage Fractured Shale Gas Horizontal Well

Abstract Shale gas is a crucial component of unconventional energy. Productivity evaluation of gas wells is essential to ensure efficient and stable production. However, predicting productivity has been challenging due to the complex characteristics of shale gas reservoirs and the use of multi-stage fractured horizontal wells (MFHW) technology in the development process. This study compares traditional production prediction methods with an optimized least squares support vector machine model (LSSVM). The traditional productivity prediction method involves establishing a mathematical model of the MFHW in a fully enclosed rectangular shale formation. The model considers the effects of shale gas adsorption, diffusion, and pressure sensitivity. The model’s analytical solution is obtained using Duhamel’s principle, Laplace transform, and inverse transform. An independent production data analysis software is developed based on the analytical solution of bottom-hole pressure to predict production. To implement the LSSVM model, the model’s input parameters must be determined first. The LSSVM model’s regularization parameters and kernel parameters are obtained through the particle swarm optimization (PSO) algorithm, and the prediction model is established. The model’s matching is evaluated by calculating the coefficient of determination (R2) and the normalized root mean square error (NRMSE). The results indicate that the traditional production capacity prediction method is suitable for stable production in terms of applicability. However, the LSSVM model does not have this limitation and generally provides more accurate predictions throughout the entire production process. For complex shale gas reservoirs that frequently switch wells and use multi-stage fracturing technology, the LSSVM model is more suitable for predicting shale gas well productivity.

Production Capacity Prediction Method of Shale Oil Based on Machine Learning Combination Model

Production Capacity Prediction Method of Shale Oil Based on Machine Learning Combination Model

Capacity Prediction Method of Lithium‐Ion Battery in Production Process Based on Improved Random Forest

Measuring capacity in the grading process is an important step in battery production. The traditional capacity acquisition method requires considerable time and energy consumption; therefore, an accurate capacity estimation is crucial in reducing production costs. Herein, a capacity prediction method for lithium‐ion batteries based on improved random forest (RF) is proposed. This method extracts features from the voltage data of the entire formation process and the first 25% of the grading process, saving 56.7% of the energy consumption and 74.6% of the time in the grading process. The importance of these features is ranked by RF, the best feature subset is selected, and the RF algorithm with parameters optimized by the genetic algorithm is applied to establish a capacity prediction model. The results show that the proposed model can accurately predict the capacity, with root mean square error (RMSE) and mean absolute percentage error (MAPE) values of only 191.89 mAh and 0.13%, respectively. Compared with other regression algorithms, the model shows a lower RMSE and MAPE and a higher capacity prediction accuracy for low‐capacity cells.

Battery state of health prediction based on voltage intervals, BP neural network and genetic algorithm

ABSTRACT Accurate prediction of lithium ion (li-ion) battery capacity is of great significance to battery health status management. In this paper, the different discharge time corresponding to the equal voltage interval is taken as the health factor. Three highly correlated health factors are extracted from the battery discharge curve, and the Back Propagation neural network (BPNN) optimized by a genetic algorithm (GA) is used to estimate the battery capacity accurately and quickly. Firstly, health factors related highly to battery capacity from the battery cycle life test are extracted, and the selected health factors are analyzed using the Spearman correlation coefficient and Pearson correlation coefficient. Secondly, this paper analyzes the prediction effect of different combinations of selected health factors using BPNN optimized by a genetic algorithm. Then, to verify the superiority of the proposed optimization algorithm, different optimization algorithms are used to adjust and optimize the parameters of BPNN automatically, and the experimental data are used for comparative analysis. Finally, the GA-BP is compared with other common battery capacity prediction methods. The results show that the GA-BP neural network prediction model can accurately and effectively predict the capacity of li-ion battery when using selected health factors.

Development of PV hosting-capacity prediction method based on Markov Chain for high PV penetration with utility-scale battery storage on low-voltage grid

ABSTRACT The previous stochastic hosting capacity prediction method using the Monte Carlo method for high photovoltaic (PV) penetration with a battery energy storage system (BESS) required a large number of computations to achieve the expected accuracy. The problem of high computational load must be addressed so that the electrical distribution planner can practically use the PV hosting capacity prediction in actual situations. Therefore, this study developed a Markov-chain-based PV hosting capacity prediction method for high PV penetration using BESS. The proposed method is described in detail, followed by case and validation studies. The obtained hosting capacity was 123.58 kW, which increased to 3676.4 kW after the utility-scale BESS implementation. The results demonstrate that the proposed Markov-chain-based PV hosting capacity prediction method outperforms the Monte Carlo method, which is the most popular stochastic hosting capacity method, in terms of accuracy and computational cost.

Tensile behaviors and configurations of double-headed bar overlap connections for precast concrete members

The superiorities in construction time, cost, and quality of precast RC structures are accompanied by the challenges in connection, particularly in steel bars connections between RC components. The extensively used steel bar connections for either grouted sleeve splice connection or grouted spiral-confined lap connection requires on-site grouting, which is a concealed construction process with increased the risk of grouting defects, thereby impairing connection reliability. The confined headed bar connection (CHBC) with visible on-site construction has been proposed to address the unseen grouting defects concerns. However, modern structures that use large-diameter or high-strength rebars present higher demands for steel bar connection capacity. For CHBC, it is suggested to use either a long connection length or a large-size head to satisfy the capacity demand. However, the long connection length can increase field wet operation, which is disadvantage in precast concrete construction. The large-size head can help reduce the connection length, but it can cause inconvenience including construction congestion and the increased eccentric bending moment under tension. This paper presents the CHBC with double-head configuration, which can reduce the single head size. The proposed CHBC consists of confinement stirrups and two overlapped headed bars and the stirrup confining the core concrete can also improve the CHBC capacity. Experimental investigation on 21 CHBC specimens has been carried out. The failure modes, bearing capacity, and stirrup strain development have been studied. Effects of the double-head configuration and confinement stirrup ratio have been analyzed through experimental and numerical investigation. Results show that the double-head configuration has a minor impact on the tensile capacity of CHBC while the increase of confined stirrup ratio can improve the tensile capacity of CHBC. Configuration details of size and spacing for stirrup reinforcement in CHBC are proposed. Based on the test results, CHBC tensile capacity prediction method is suggested, which can take the head configuration into account.

Testing and predicting the response of reinforced conrete shear walls retrofitted for recentering

Laboratory tests were conducted to investigate the feasibility of a new retrofit method for reinforced concrete shear walls to enable self-centering and to evaluate strength and drift capacity prediction methods for pre- and post-retrofit walls. The proposed retrofit method involved cutting the base of a wall, adding external post-tensioning, and providing additional confinement at wall toes. Three shear walls, one benchmark and two retrofitted specimens, were tested under quasi-static cyclic lateral loading. The two retrofitted specimens differed in the initial post-tensioning force, external confinement plate size, and base cut shape. The benchmark specimen exhibited both flexure and shear damage, whereas the governing displacement mechanism for the retrofitted walls was rocking at the wall-to-foundation interface. As intended, the strengths of the retrofitted walls were 80% to 100% of that of the benchmark specimen. The retrofitted walls had larger drift capacities, fewer cracks, smaller areas of concrete crushing, lower residual drifts, and sufficient energy dissipation. The existing methods predicted the strength and drift capacity of the pre-retrofit specimen reasonably well, yet overestimated the strength and underestimated the drift capacity of the post-retrofit specimens.

Ultimate bearing capacity prediction method and sensitivity analysis of PBL

The ultimate bearing capacity of Perfobond leiste (PBL) is one of the key parameters to evaluate the bearing capacity and reliability of steel–concrete structures, and it is very important to predict the ultimate bearing capacity of PBL more accurately. Based on the improved cuckoo search algorithm (CS), two prediction model optimized by back propagation neural network (BPNN) algorithm and extreme learning machine (ELM) were proposed. The local search ability of CS algorithm was improved by the triangular mutation operator and the distance-based distributed discovery probability, and the global search ability was improved by multi-step selection strategy. The weight, threshold, number of input parameters and number of nodes in the hidden layer of BPNN and ELM were optimized by the triangular multi-step cuckoo search (TMCS). The comprehensive sensitivity analysis (CSA) method and Morris sensitivity analysis (MSA) method were used to analyze the sensitivity of six key parameters of PBL, such as thickness of perforated steel plate, diameter of perforated holes, number of perforated holes, diameter of through reinforcement, yield strength of through reinforcement and compressive strength of concrete. The experimental data of push-out tests in published literatures were selected as samples, and the results show that the proposed TMCS-ELM algorithm and TMCS-BPNN algorithm can accurately predict the ultimate bearing capacity of PBL, and the average errors are 4.17% and 2.16% respectively. The sensitivity analysis results show that the compressive strength of concrete has the highest influence on the bearing capacity of PBL, followed by the yield strength of through reinforcement.

A novel method of prediction for capacity and remaining useful life of lithium-ion battery based on multi-time scale Weibull accelerated failure time regression

Lithium-ion batteries are essential energy storage components for electrical grid, and the health diagnosis determines the safety of the battery during usage and the rational classify of echelon utilization. In this article, a multi-timescale capacity and lifespan prediction method is proposed where capacity prediction and remaining useful life prediction are divided into the short-time scale and the long-time scale. For capacity prediction, the long short term memory neural network with five significant features is applied according to its accuracy performance in time series prediction. As for remaining useful life, the Weibull accelerated failure time regression is proposed to improve the prediction efficiency of a large amount of data. Finally, the predictive capability, robustness and effectiveness of proposed methods are verified using two datasets with different cycling test conditions within an error of 3.9 % in long-time scale and 2.7 % in short-time scale. The proposed method has great potential for targeted and accurate health state forecasting and long-term end of life prediction.

Experimental study on demountable CFST K-joints designed with blind bolts

Developing demountable structures that enable the replacement of damaged components and the reuse of intact parts in their extended service-life is an effective approach to reducing carbon emissions and prompt sustainable construction. This paper presents an experimental investigation on the behavior of novel demountable K-joints formed by concrete-filled steel tubular (CFST) chords and circular hollow sectional (CHS) braces connecting through blind bolts and curved endplates. A series of tests are conducted on the demountable K-joint specimens, where pined supports are applied at the ends of two braces whilst one end of the chord is under tension and the other is left free. Based on the tests, the failure mode, the ultimate capacity, the joint stiffness and the deformability of the demountable joints are investigated. After the initial loading, the damaged components are disassembled and replaced by new ones to form new joints which are then loaded again for the 2nd and 3rd service periods. It is found that the differences in joint ultimate capacity during multiple service periods are generally within 5%, testifying the effective reuse of the main structural components such as CFST chords and blind bolts. The effects of important parameters related to the demountability design, the joint configuration and the material strengths are evaluated through the tests. Finally, a simplified ultimate capacity prediction method for such demountable CFST K-joints is proposed and validated.

Experimental and analytical investigation on shear mechanism of steel-UHPC composite T-Perfobond shear connectors

This paper proposed a T-Perfobond shear connectors to increase the end bearing area of the intermittently arranged PBL shear connectors. The failure modes, load-slip curves, and strains of perforating rebar of T-Perfobond shear connectors were discussed. Following that, a finite element analysis (FEA) model was established to investigate the influences of design parameters on the behavior of T-Perfobond shear connectors. A theoretical shear bearing capacity prediction method on T-Perfobond shear connectors was proposed, which can consider the effect of end concrete bearing pressure and the reduction coefficient of welding effect. The results showed that the specimens failed in three stages with bond damage, concrete dowel fracture and end concrete crushing. The shear bearing capacity of T-Perfobond shear connectors increased with the increase of web plate thickness and decreased with the increase of flange plate thickness. The hole diameter weakened the web plate strength and reduced the stability of T-Perfobond shear connectors. The theoretical predictions agreed well with the experimental and numerical calculation results. This study can be used to guide the design and application of shear connectors of steel–concrete composite beams.

A Data-Driven LiFePO4 Battery Capacity Estimation Method Based on Cloud Charging Data from Electric Vehicles

The accuracy of capacity estimation is of great importance to the safe, efficient, and reliable operation of battery systems. In recent years, data-driven methods have emerged as promising alternatives to capacity estimation due to higher estimation accuracy. Despite significant progress, data-driven methods are mainly developed by experimental data under well-controlled charge–discharge processes, which are seldom available for practical battery health monitoring under realistic conditions due to uncertainties in environmental and operational conditions. In this paper, a novel method to estimate the capacity of large-format LiFePO4 batteries based on real data from electric vehicles is proposed. A comprehensive dataset consisting of 85 vehicles that has been running for around one year under diverse nominal conditions derived from a cloud platform is generated. A classification and aggregation capacity prediction method is developed, combining a battery aging experiment with big data analysis on cloud data. Based on degradation mechanisms, IC curve features are extracted, and a linear regression model is established to realize high-precision estimation for slow-charging data with constant-current charging. The selected features are highly correlated with capacity (Pearson correlation coefficient < 0.85 for all vehicles), and the MSE of the capacity estimation results is less than 1 Ah. On the basis of protocol analysis and mechanism studies, a feature set including internal resistance, temperature, and statistical characteristics of the voltage curve is constructed, and a neural network (NN) model is established for multi-stage variable-current fast-charging data. Finally, the above two models are integrated to achieve capacity prediction under complex and changeable realistic working conditions, and the relative error of the capacity estimation method is less than 0.8%. An aging experiment using the battery, which is the same as those equipped in the vehicles in the dataset, is carried out to verify the methods. To the best of the authors’ knowledge, our study is the first to verify a capacity estimation model derived from field data using an aging experiment of the same type of battery.