Levenberg-Marquardt Backpropagation Research Articles

Artificial neural network-based prediction model of elastic floor response spectra incorporating dynamic primary-secondary structure interaction

The evaluation of the Floor Response Spectrum (FRS) holds paramount significance in assessing the seismic behavior of secondary structures. Precise FRS prediction empowers engineers to make informed decisions concerning structural design, retrofitting, and safety precautions. This study aims to scrutinize the impact of dynamic interaction between primary and secondary structures on FRS. Both the elastic primary structure (PS) and elastic secondary structure (SS) employ a single-degree-of-freedom (SDOF) system. Governing motion equations for both coupled (with dynamic interaction) and uncoupled (without dynamic interaction) systems are formulated and solved numerically. The study investigates how variations in the vibration period of PS (Tp), tuning ratio (Tr), mass ratio (μ), and damping ratio (ξs) of SS influence FRS. The FRS impact remains minimal at μ = 0.001 (0.1%); however, with increasing mass ratio, PS-SS dynamic interaction significantly affects SS's spectral acceleration response. Coupled analysis is crucial only for secondary structures tuned to the primary structure's vibration period (0.8≤Tr≤1.2). This study utilizes two-layer feed-forward Artificial Neural Networks (ANNs) for FRS prediction. The Levenberg-Marquardt (LM) backpropagation (BP) algorithm trains the network using a comprehensive dataset. In summary, it is evident that the ANNs, once trained, enable accurate prediction of the FRS, exhibiting a R2 of 99%. Additionally, a design expression is formulated utilizing the ANN model and subsequently compared with the existing formulation.

A neural network computational structure for the fractional order breast cancer model

The current study provides the numerical performances of the fractional kind of breast cancer (FKBC) model, which are based on five different classes including cancer stem cells, healthy cells, tumor cells, excess estrogen, and immune cells. The motive to introduce the fractional order derivatives is to present more precise solutions as compared to integer order. A stochastic computing reliable scheme based on the Levenberg Marquardt backpropagation neural networks (LMBNNS) is proposed to solve three different cases of the fractional order values of the FKBC model. A designed dataset is constructed by using the Adam solver in order to reduce the mean square error by taking the data performances as 9% for both testing and validation, while 82% is used for training. The correctness of the solver is approved through the negligible absolute error and matching of the solutions for each model’s case. To validates the accuracy, and consistency of the solver, the performances based on the error histogram, transition state, and regression for solving the FKBC model.

Soft computing through supervised neural network based on Levenberg–Marquardt back propagation for numerical treatment of steady flow of Oldroyd 4-constant fluid between two rolls

This study presents a new intelligent computing method based on the Levenberg–Marquardt back-propagation-based supervised neural network (LMB-SNN) for the flow of Oldroyd 4-constant fluid in the forward roll coating process (OFC-FRCP). The mathematical formulations for OFC-FRCP are converted into ordinary differential equations using lubrication approximation theory (LAT) and suitable dimensionless parameters. The analytic expressions for velocity distribution, pressure and pressure gradients have been obtained using Adomian’s Decomposition Method (ADM) for several situations. The fluid’s separation point and coating thickness are then calculated using integration. The reference dataset for LMB-SNN has been obtained by ADM for multiple scenarios by adjusting various physical variables. The neural network’s training, testing and validation are carried out in parallel to minimize the mean-squared error function (MSEF). The proposed LMB-SNN’s effectiveness was confirmed by analyzing the MSEF, error histograms and regression plots. The precision of the method is verified by the close agreement between numerical outputs and dataset values. The intelligent numerical results computed using ADM and LMB-SNN are presented in various graphs and tables. The method’s accuracy is demonstrated by the numerical outputs closest to the built and dataset values at levels between [Formula: see text] and [Formula: see text].

A stagnation point flow of cross nanofluid flow: Levenberg Marquardt backpropagation computational approach

AbstractIn the present article, the intelligence based numerical investigation by Levenberg–Marquardt algorithm with Back‐propagated Artificial Neural Networks (LMA–BANN) is exploited to analyze the nonlinear radiative stagnation point flow of cross nanofluid system (NRS‐CNFS) past a stretching surface. The designed NRS‐CNFS, initially represented by system of partial differential equations (PDEs), is converted into system of non‐linear ODEs through the applicability of mathematical conversion analysis. The desired reference solution in the form of a dataset for LMA–BANN is achieved from the Adam method for NRS‐CNFS's different scenarios by varying the magnetic parameter (Ha), Prandlt number (Pr), thermal radiation parameter(R), Eckert number (Ec), thermophoresis diffusion coefficients (Nt), and non‐dimensional activation energy (E). The obtained results are construed for the NRS‐CNFS through the performances of the testing, validation, and training. In addition, the comparison is provided using the LMA ‐BANN analysis, which is validated in terms of regression, fitness estimations, and histograms along with MSE performances. Hence, the article provides an impeccable innovative approach to problem‐solving by implementing a soft computation paradigm. It promotes the use of an effective and dependable alternative framework that is based on soft computing environments and emphasizes descriptive analysis to successfully handle the problems brought on by varied physical features.

Deriving groundwater major ions from electrical conductivity using artificial neural networks supported by analytical hydrochemical solutions

Tracing groundwater chemistry requires broad and continuous sampling campaigns to keep track of groundwater deterioration. The need is pressing in coastal aquifers due to high water demand to meet rapid population growth. This paper uses artificial intelligence (AI) to develop an algorithm capable of predicting seven major chemical ions in water (Ca2+, Mg2+, Na+, K+, HCO3−, Cl− and SO42−) from electrical conductivity (EC), a single input that can be easily acquired manually or via automated monitoring programs. The algorithm mainly relied on artificial neural network (ANN) to tackle complex and nonlinear relationships. It included a set of procedures among others, Levenberg-Marquardt back propagation multilayer perceptron (MLP), conjugate gradient back propagation with Fletcher-Reeves updates MLP, and support vector regression. The low accuracy of ANN models in predicting Ca2+ and HCO3− concentrations justified their calculation via advanced hydrochemical analytical solutions that were embedded in one AI scheme. The full scheme was trained and tested over an EC20 range of 400–3000 μS/cm, and then validated for new measurements elsewhere. It showed <15% discrepancy for the four ions that constitute the majority of groundwater chemical composition (Cl−, Na+, HCO3− and Ca2+). Higher uncertainties were associated with concentrations derived for EC values outside the training limits. The developed algorithm can save time and money in return of labor and laboratory work. It is an open source code publicly available for further improvements by training it on additional data covering a wider EC range.

A novel solution for the optimal reactive power dispatch problem using an artificial neural network integrated with the firefly optimization algorithm

This article applies a novel intelligence technique to solve power system issues faced daily. Compensation for reactive power is a significant issue faced by power system operators in research. The solution can be obtained by handling a multi-objective task and multiconstraints by reducing the active power loss and minimizing the voltage deviation at the load end. The novelty of the research focuses on integrating artificial neural network techniques with the firefly algorithm, a novel optimization algorithm for attaining an objective function. The Levenberg–Marquardt back-propagation algorithm is most suited for proper tuning of the control variables. The objective of this research can be attained by appropriately tuning the control variables connected with the IEEE test bus systems, which helps to maximally improve the voltage profile. Existing research studies have focused on reactive power management, which is attained by solving optimal reactive power flow problems employing nature-inspired approach techniques such as the symbiotic organism search algorithm, the cuckoo search algorithm, the black hole algorithm, the krill herd algorithm, and whale optimization. The evolving strategy, the firefly algorithm (FFA), minimizes the multiconstraint functions more competently and effectively than any conventional algorithm. To showcase the strength of the firefly algorithm incorporating AI, it is examined on standard IEEE test bus systems, namely, the 14-, 30-, and 58-bus networks. The obtained results quantify the effectiveness of the proposed methodology, that is, the artificial intelligence technique implementing the firefly algorithm gives better results than conventional methods.

Optimized artificial neural network model for accurate prediction of compressive strength of normal and high strength concrete

This study develops and presents an Artificial Neural Network (ANN) model employing the Levenberg-Marquardt Backpropagation (LMBP) training algorithm to predict the compressive strength of both normal and high strength concrete. The model's robustness was evaluated using an extensive dataset comprising 1637 samples. Eight input variables, including the cement content, blast furnace slag, fly ash, fine aggregate, coarse aggregate, water content, superplasticizer, and testing age, were considered. The optimal number of hidden layers and neurons in the layer were identified through analysis, and the effectiveness of the model was assessed through k-fold cross-validation and statistical measures, including correlation coefficient (R), coefficient of determination (R2), Root Mean Square Error (RMSE), and Mean Absolute Error (MEA). Comparison with other models was carried out, and the perturbation/super-position method was employed for parametric studies to investigate the effect of each input variable on the output variable. The k-fold cross-validation confirmed the generalizability of the model, and statistical measures showed good results, with unit cement content and superplasticizers having the highest impact on compressive strength. The findings demonstrate that the suggested ANN model is an extremely precise, economical, and practical predictive tool for concrete compressive strength.

Artificial intelligent solvers for the HIV-1 system including AIDS based on the cancer cells

The present investigations are related to provide the numerical performances of the HIV-1 dynamical infection model in patients with cancer (HIV-DIMC) by applying the artificial intelligence (AI) scheme based on Levenberg-Marquardt backpropagation neural networks (LBMBP-NNs). The current biological system is presented into three dynamics including cells of cancer population (T), healthy (H), and infected HIV (I). The substantiations, training and testing measures are used as sample statics to solve the HIV-DIMC. These performances with statistical ratios have been chosen as 75% training, substantiations 13% and testing 12% in order to solve the dynamical model. The correctness of achieved performances based on the HIV-DIMC is observed by using the assessment of the obtained and reference results. The absolute error is performed around 10−06 to 10−07 describe the efficiency of the scheme. The achieved measures of the dynamical system are stated to reduce the mean square error in interval 10−11-10−13. To perceive the effectiveness, credibility and aptitude of AI based LBMBP-NNs, the computing performances are proficient to analyze the convergence based on the histogram diagrams and correlation catalogue.

Detailed experimentation and prediction of thermophysical properties in lauric acid-based nanocomposite phase change material using artificial neural network

The thermal conductivity (TC) of a nanocomposite phase change material (NPCM) may be improved by adding nanostructured materials to a Phase Change Material (PCM). To assess the heat transfer rate during the process of phase change, such as melting and freezing, an accurate TC prediction of NPCM is required. A Field Emission Scanning Electron Microscope (FESEM) was utilized to examine the nanoparticle morphological study, and X-Ray Diffraction (XRD) analysis evaluated the crystalline structure. NPCMs were verified using Fourier Transform Infrared Spectroscopy (FTIR). The goal of this research is to create an Artificial Neural Network (ANN) that guesses the TC and viscosity of Lauric Acid (LA) embedded with dispersed copper oxide (CuO) and aluminium oxide (Al2O3). A multi-layered feed-forward ANN (MLFFANN) is trained using the Levenberg-Marquardt (LM) backpropagation algorithm. There are 130 experimental datasets in total, obtained from experiments with nanoparticle mass fractions ranging from 1.25 to 10 wt%. The minimum mean square error (MSE) for TC and viscosity is 4.6815 × 10−5 and 2.4681 × 10−5, respectively. The average absolute deviation (AAD) for TC and viscosity is 0.004249 and 0.003596, respectively, while the mean absolute percentage error (MAPE) is 2.1835 % and 4.8197 % for TC and viscosity, and the correlation coefficients(R) are 0.992 and 0.975, respectively. The largest percentage variation between experimental values and ANN computed values for the liquid and solid phases, respectively, is 3.54 % and 0.832 %. It demonstrates that the constructed ANN prediction model predicts the increased TC and viscosity of NPCM for different nanoparticle loadings, temperatures, and oxide nanoparticles.

Exploring the impact of Hall and ion slip effects on mixed convective flow of Casson fluid Model: A stochastic investigation through non-Fourier double diffusion theories using ANNs techniques

The analysis of Non-Newtonian Casson fluid flow using Artificial Neural Networks (ANNs) has enticed the attention of researchers and scientists due to their tremendous role in modern technologies and development. The generalised Ohm law with mass and thermal transport is employed. The mechanism of mass and energy transmission is based on generalised Fick’s and Fourier laws. An incompressible magnetohydrodynamics (MHD) boundary layer flow of Darcy-Forchheimer Casson fluid across a stretching sheet is elaborated in the current study. The flow incorporates the effects of Hall and ion slip phenomena and occurs steadily over a stretchable linear surface through porous medium. The Casson fluid flow has been expressed in form of system of PDEs, which are further simplified to non-linear system of ODEs through similarity replacements. The ANN-LMBOA algorithm is used to crack these equations (ODEs). To validate the ANN-LMBOA results, the dataset is produced by employing the (FDM) finite difference method (Lobatto IIIA) in Matlab package using bvp4c-solver. The dataset is generated for diverse scenarios of flow constraints, testing, and validation of the neural network. The accurateness of the ANN-LMBOA (Artificial Neural Network- Levenberg Marquardt Backpropagation Optimization Algorithm) is evaluated via several statistical neural network tools, i.e., RP, MSE, EH and CF graphs.

Levenberg–Marquardt neural network-based intelligent computation for the non-Newtonian polymer during forward roll coating

Scientists and researchers widely recognize the effectiveness of artificial intelligence (AI)-based machine learning and intelligent computing solvers, demonstrating qualities such as resilience, robustness, stability, and rapid convergence. One particularly significant and rapidly growing field within AI is artificial neural networks. This research uses a supervised neural network model based on Levenberg–Marquardt backpropagation (LMB-SNNs) to examine the Sisko fluid model for the forward roll coating process (SFM-FRCP). A suitable transformation is applied to the partial differential equations based SFM-FRCP mathematical model, resulting in a set of nonlinear ordinary differential equations. The perturbation method has been used to find the analytical solutions for the velocity profile, pressure gradient, and pressure profile. A dataset for varying the pertinent parameters is generated, and the LMB-SNNs technique has been used to estimate the velocity profile, pressure gradient, and pressure profile behavior during FRCP for numerous scenarios. The numerical solution for SFM-FRCP in different scenarios, such as the validation, training, and testing procedures of LMB-SNNs, is carried out. Moreover, the state transition index, fitness outline, mean square error, histogram error, and regression presentation also endorse the strength and reliability of the solver LMB-SNNs for SFM-FRCP. The comparative analyses and performance studies through outputs of regression drawings, absolute error, and error histograms validate the effectiveness of the suggested solver LMB-SNNs. The method's precision is verified by the closest numerical outputs of both built and dataset values with similar levels 10−11–10−14. Furthermore, it has been observed that as the non-Newtonian parameter increases, the fluid velocity decreases. The research work carried out in this paper is original and fills a gap in the existing research by showing the rheological properties of the Sisko fluid model and the implementation of the LMB-SNNs during the FRCP.

SOFT COMPUTING ARTICIAL INTELLIGENCE OF SCHRöDINGER TIME INDEPENDENT EQUATION ARISES IN WHEELER-DEWITT MODEL OF QUANTUM COSMOLOGY

Schrödinger time independent equation arises in Wheeler–DeWitt model of quantum cosmology to develop a wave function for the corrections of cosmologies. In this work, k=0 rödinger–Wheeler–DeWitt model is developed by using canonical transformation to transformed Hamiltonian equation. The exact solutions for flat universe () model is presented and then soft computing technique through Levenberg–Marquardt backpropagation neural networks (LMB-NNs) is implemented. The data set of the flat universe model is generated for four different cases of separation constant B with step size 0.1 by Mathematica and imported into LMB-NNs for training, testing, and validation of the our proposed technique. The performance of LMB-NNs is provided by the validation of mean square error, error histogram, and regression analysis. The Statistical analysis; mean, minimum, maximum, and standard deviation, is also presented.

Predicting the strengths of date fiber reinforced concrete subjected to elevated temperature using artificial neural network, and Weibull distribution

Date palm fiber (DPF) is normally used as fiber material in concrete. Though its addition to concrete leads to decline in durability and mechanical strengths performance. Additionally, due to its high ligno-cellulose content and organic nature, when used in concrete for high temperature application, the DPF can easily degrade causing reduction in strength and increase in weight loss. To reduce these effects, the DPF is treated using alkaline solutions. Furthermore, pozzolanic materials are normally added to the DPF composites to reduce the effects of the ligno-cellulose content. Therefore, in this study silica fume was used as supplementary cementitious material in DPF reinforced concrete (DPFRC) to reduce the negative effects of elevated temperature. Hence this study aimed at predicting the residual strengths of DPFRC enhanced/improved with silica fume subjected to elevated temperature using different models such as artificial neural network (ANN), multi-variable regression analysis (MRA) and Weibull distribution. The DPFRC is produced by adding DPF in proportions of 0%, 1%, 2% and 3% by mass. Silica fume was used as partial substitute to cement in dosages of 0%, 5%, 10% and 15% by volume. The DPFRC was then subjected to elevated temperatures between 200 and 800 °C. The weight loss, residual compressive strength and relative strengths were measured. The residual compressive strength and relative strength of the DPFRC declined with addition of DPF at any temperature. Silica fume enhanced the residual and relative strengths of the DPFRC when heated to a temperature up to 400 °C. To forecast residual compressive strength (RCS) and relative strength (RS), we provide two distinct ANN models. The first layer's inputs include DPF (%), silica fume (%), temperature (°C), and weight loss (%). The hidden layer is thought to have ten neurons. M-I is the scenario in which we use RCS as an output, whereas M-II is the scenario in which we use RS as an output. The ANN models were trained using the Levenberg–Marquardt backpropagation algorithm (LMBA). Both neural networking models exhibit a significant correlation between the predicted and actual values, as seen by their respective R = 0.99462 and R = 0.98917. The constructed neural models M-I and M-II are highly accurate at predicting RCS and RS values. MRA and Weibull distribution were used for prediction of the strengths of the DPFRC under high temperature. The developed MRA was found to have a good prediction accuracy. The residual compressive strength and relative strength followed the two-parameter Weibull distribution.

Hybrid Algorithm of Backpropagation and Relevance Vector Machine with Radial Basis Function Kernel for Hydro-Climatological Data Prediction

Hydro-climatological data serves a pivotal role in monitoring climatic alterations and facilitating agricultural planning, inclusive of evapotranspiration estimation, water management, and crop pattern design. The necessity to accurately and expeditiously model and forecast this data underscores the need for effective methodologies. This paper introduces a hybrid algorithm, integrating backpropagation and relevance vector machine (BP-RVM) with a radial basis function (RBF) kernel. A comparative analysis was conducted between RBF and Logsig activation functions in conjunction with resilient backpropagation (trainrp) and Levenberg-Marquardt backpropagation (trainlm). The algorithm was employed to predict and categorize rainfall, temperature, wind speed, humidity, and sunshine duration data. Through extensive trials, the architecture parameters in the training-testing process of the BP-RVM algorithm were meticulously determined. Mean squared error (MSE) and mean absolute percentage error (MAPE) values were classified as indicating high forecast accuracy (&lt;10%). Despite the RBF-trainlm kernel function combination exhibiting a faster epoch completion rate, the BP-RVM algorithm with the RBF-trainrp kernel function combination is recommended for future data prediction stages due to its lower error generation. The BP-RVM-RBF-trainrp algorithm outperformed BP-RVM-RBF-trainlm, with an average error difference of 1.39% in the training process and 2.28% in the testing process. The identified algorithms and architectures present potential for future applications in evapotranspiration calculation and crop pattern planning based on hydro-climatological data.

Multilayer neural networks for studying three-dimensional flow of non-Newtonian fluid flow with the impact of magnetic dipole and gyrotactic microorganisms

The current paper explores the three-dimensional flow of an Oldroyd-B liquid with the impact of a magnetic dipole that occurred by stretching a flat surface placed in the plane with a linear velocity variation in two directions containing motile gyrotactic microorganisms. Using proper similarity transformations, the governing equations are reduced into nonlinear coupled ordinary differential equations (ODEs). The ODEs are then solved using Runge–Kutta-Fehlberg (RKF) method. The training, testing, and validation processes are carried out in parallel to adapt neural networks and calculate an approximate solution for the considered model. This helps to reduce the mean square error (MSE) function by Levenberg–Marquardt backpropagation. The efficiency of the suggested backpropagated neural networks methodology has been demonstrated by utilizing outcomes such as MSE, error histograms, correlation and regression. Results reveal that the heat transport augments for increased Biot number values. The mass transport declines for improved chemical reaction rate parameter values. A higher Peclet number will result in a lower motile diffusivity and result in a decline in the micro-organism’s density profile. For the least value of Mu and gradient, better convergence of the findings can be achieved with better network testing and training.

Numerical investigations of ion slip and hall effects on Cattaneo-Christov heat and mass fluxes in darcy-forchheimer flow of Casson fluid within a porous medium, utilizing non-fourier double diffusion theories through artificial neural networks ANNs

The current research focuses on harnessing the computational power of neural networks, notably the Levenberg-Marquardt backpropagation optimization algorithm (ANN-LMBOA), to simulate the behavior of magnetohydrodynamic (MHD) three-dimensional non-Newtonian Darcy-Forchheimer flow of Casson nanofluid over a stretching surface situated within a porous medium. The fluid flow is subject to the influence of Hall and ion slip forces, contributing to its behavior. Additionally, the Cattaneo–Christov heat flux model is incorporated to represent the temperature balance in the boundary layer accurately. Non-Newtonian fluids improve thermal transmission due to their shear-thinning propensity, increased turbulence, and more excellent convective heat transfer. The thermal conductivity of nanofluids is increased by the inclusion of nanoparticles, enabling more effective heat transmission. Buongiorno's model is used for the nanofluid phenomena that concentrate on thermophoresis and Brownian motion. To tackle the flow problem efficiently, the governing equations are transformed into non-linear differential equations using self-similar variables. The ANN-LMBOA algorithm is then employed to solve these equations, providing a practical and accurate approach to the simulation. To train and validate the ANN-LMBOA model, a dataset is generated using the bvp4c numerical method, which can easily handle highly non-linear coupled differential equations. The dataset is created for different scenarios of flow parameters, facilitating comprehensive training, testing, and authentication of the neural network. The accuracy of the ANN-LMBOA model is assessed through various statistical neural network tools, such as MSE, RP, CF graphs, and EH. The study thoroughly investigates flow model parameters related to momentum, energy, and concentration profiles, which are presented using visual representations to understand the system's behavior and characteristics comprehensively. The effects of various parameters of interest like Casson, ion slip, porosity, thermal relaxation, and magnetic parameters on velocity, temperature, and concentration distribution have been studied numerically via ANN-LMBOA networks. The absolute error of reference and target data is obtained in the range of 10−4 − 10−5which proves the best accuracy performance of ANN-LMBOA networks. By raising the Prandtl and Schmidt numbers, solutal, thermal, and diffusional variables, one can reduce the diffusion of mass and heat.

Prediction modeling using artificial neural network (ANN) for the performance and emission characteristics of catalytic co-pyrolytic fuel blended with diesel in a CI engine.

In the present research work, artificial neural network (ANN) is used to model the performance and emission parameters in a four-stroke, single-cylinder diesel engine combusting a blended fuel of diesel and catalytic co-pyrolysis oil produced from seeds of Pongamia pinnata, waste LDPE, and calcium oxide catalyst. The optimum yield of oil obtained was 92.5% at 500 °C temperature. Physical properties of the obtained oil, such as calorific value (44.85 MJ/kg) and density (797 kg/m3), level it by that of diesel while the flash point and fire point were found to be lower than that of pure diesel. An ANN model was then generated for the prediction of performance characteristics (BTE and BSFC) and emission characteristics (NOx and smoke) under varying loads, braking power, brake mean effective pressure, and torque as inputs using the Levenberg-Marquardt back-propagation training technique. The regression coefficients (R2) for BTE, BSFC, smoke, and NOx predictions were determined to be close to unity at 0.99859, 0.99814, 0.96129, and 0.92505, respectively (all values being close to unity). It has been discovered that ANN makes an effective simulation and prediction tool for blended fuels in CI engines. It is also suggested to predict the mechanical efficiency, volumetric efficiency, and CO, CO2, HC emissions using ANN in its future work.

Response surface methodology-artificial neural network application for bisphenol A treatment using carbon nanotubes manganese oxide composite

Bisphenol A (BPA) is an emerging pollutant that easily escapes conventional treatment techniques. It requires application of novel composite materials along with mathematical modelling for optimisation and evaluation of the treatment process. In the present study, manganese (IV) oxide (MnO2) nanoparticles were doped onto the surface of multi-walled carbon nanotubes to develop an adsorptive–oxidative composite. The composite was characterised using transmission electron microscopy, X-ray diffraction, Raman spectroscopy, X-ray photoelectron spectroscopy, Fourier transform infrared spectroscopy and surface area analysis to confirm composite formation and study its properties. Conventional optimisation of pH (4–10), initial BPA concentration (10–50 mg/l) and contact time (0–60 min) was carried out and found to fit well with the Freundlich isotherm model (R 2 &gt; 0.99) and followed a pseudo-second-order kinetic reaction. A central composite design model was applied using response surface methodology (RSM) to study individual parameters and their interaction effects to enhance process efficiency. Further, the experimental data sets and their responses from RSM were analysed using an artificial neural network. From random experimental sets (80%) of which (10%) each to train, validate and test were selected to analyse the variance of models for higher efficiency using Levenberg–Marquardt back-propagation (LM-BP) algorithm. Additionally, BPA-spiked simulated pharmaceutical waste water was treated with the composite to explore its treatment potential. This systematic experimental and computational approach aided in optimising treatment efficiency for real-time application.

Machine learning prediction of biochar yield based on biomass characteristics

Slow pyrolysis is a widely used thermochemical pathway that can convert organic waste into biochar. We employed six machine learning models to predictively model 13 selected variables using pearson feature selection. Additionally, partial dependence analysis is used to reveal the deep relationship between feature variables. Both the gradient boosting decision tree and the Levenberg-Marquardt backpropagation neural network achieved training set R2 > 0.9 and testing set R2 > 0.8. But the other models displayed lower performance on the testing set, with R2 < 0.8. The partial dependence plot demonstrates that pyrolysis conditions have greater impact on biochar yield than biomass composition. Furthermore, the highest treatment temperature, being the sole consistently changing feature, can serve as a guiding factor for regulating biochar yield. This study highlights the immense potential of machine learning in experimental prediction, providing a scientific reference for reducing time and economic costs in pyrolysis experiments and process development.

Multiple faults diagnosis for ocean-going marine diesel engines based on different neural network algorithms.

Fault diagnosis technologies for ocean-going marine diesel engines play an important role in the safety and reliability of ship navigation. Although many fault diagnosis technologies have achieved acceptable results for single fault of diesel engines, the diagnosis of multiple faults is rarely involved. Due to the strong correlation, non-linearity and randomness of multiple faults, it is extremely difficult to make an accurate diagnosis. In this study, diagnosis methods based on thermal parametric analysis combined with different neural network algorithms were established and used for the diagnosis of multiple faults in the ocean-going marine diesel engine. The results show that the Levenberg Marquardt back propagation neural network has the highest diagnostic accuracy rate of 88.89% and 100% for multiple faults and single faults, respectively, and its diagnostic time is also relatively short, 0.78 s. The Bayesian regularization back propagation neural network can give a diagnostic accuracy rate of 100% for single faults, but for multiple faults, the diagnostic accuracy rate is only 55.56%, and the diagnosis time for the entire sample is the longest. As for the probabilistic neural network, although it has the fastest diagnosis speed, it has the lowest diagnostic accuracy rate for both single faults and multiple faults. The results may provide references for the online diagnosis of single faults and multiple faults in ocean-going marine diesel engines.