Combination Of Input Parameters Research Articles

Machine learning-based prediction for airflow velocity in unpressured water-conveyance tunnels

Spillway and drainage tunnels have an open-channel flow pattern when operating under unpressured condition, above which air flow is driven and resisted by water flow, wall friction, and pressure difference. Unpressured tunnels present many airflow-related safety and environmental issues, including water flow fluctuation, gate vibration, shaft cover blow-off, and odor emission; therefore, it is valuable to study and predict their airflow velocity. Given the difficulty in accurate prediction of airflow velocity in unpressured tunnels and complicated influences of hydraulic, structural, and boundary pressure parameters, this study focuses on establishing high-performance prediction models and understanding the importance and independent and coupled influences of each parameter using machine learning. It is found that the water Froude number, ratio of free-surface width to unwetted perimeter, relative ventilation area, and relative tunnel length are four key parameters. By including these parameters in the input parameter combination, the machine learning models can well predict the airflow velocity in unpressured tunnels, achieving significantly higher performance than the existing empirical and theoretical models. Among these models, the models built by Random Forest and XGBoost demonstrate best performance with R2 ≥ 0.911. The interpretability analysis reveals the highest importance of the water Froude number and the ratio of free-surface width to unwetted perimeter, increases in which generally result in enhancement of the airflow velocity. The water Froude number plays a dominant role when it is ≤11.5, and a continuous increase exhibits a significantly marginal effect. The relative ventilation area and relative length of tunnels have close importances, with an increase in either generally promoting the airflow velocity. To help researchers and engineers unfamiliar with machine learning to easily and accurately predict the airflow velocity in unpressured tunnels, GPlearn algorithm is employed to establish explicit expressions, which is validated to have good performance with R2 close to 0.900.

Mechanical, thermal and morphological analysis of polylactic acid‐woven glass fiber composites fabricated by additive manufacturing

AbstractPolylactic acid (PLA), a biodegradable thermoplastic, is being used significantly in a wide range of applications because of its environmentally friendly nature. 3D printing is an innovative technique that is used to fabricate polymer composites with complex contours and high precision. Conventional 3D printing technologies use fiber reinforcement in the form of filament. However, fibers in the form of sheets, when used in composite materials, improve the distribution of load and enhance the multi‐directional strength. This study investigates the improvement in the tensile strength of PLA‐based composite by integrating woven glass fibers (WGF) during 3D printing processing. The composites were fabricated using an FDM‐based 3D printer where the WGF sheet was used as reinforcement and PLA filament was used as the matrix material. The main input parameters consist of the number of WGF layers and the orientation of the WGF. Taguchi orthogonal array was used to fabricate composite samples with different combinations of input parameters and then these samples were tested to evaluate the tensile properties. The finding shows that PLA composites reinforced with four layers of glass fiber at 0/90° fiber orientation showed the highest tensile strength of 52.85 MPa and a thermal degradation temperature of 330.49 °C.Highlights 3D printing of WGF‐PLA composite and mechanical characterization. The composite specimen showed the highest tensile strength of 52.85 MPa. The thermal degradation temperature reached 330.49 °C in TGA analysis. ANN predicted the optimal combination of 4 fiber layers with 0/90° orientation. Uniform PLA deposition reduced voids and enhanced interlayer adhesion.

Prediction of contraction channel scour depth: based on interpretability analysis and PCA-enhanced SVR

ABSTRACT Due to the numerous uncertain factors affecting contraction scour depth, although many traditional empirical formulas have been proposed in past research, their prediction accuracy is generally low. In recent years, with advancements in machine learning (ML) technology, these techniques have been able to accurately capture the nonlinear characteristics of scour-depth data. However, in pursuit of higher prediction accuracy, researchers have explored a wide range of diverse ML models that require various combinations of input parameters. These input parameter combinations often lack reliability, and the models themselves have poor interpretability, increasing the ‘black-box effect.’ Therefore, this study uses a principal component analysis (PCA)-enhanced support vector regression (SVR) model to construct a scour depth prediction model, combined with the interpretability method of SHapley Additive exPlanations (SHAP). The results show that the SVR model's predictions are highly consistent with physical experimental laws, and the model primarily identifies features that are strongly linearly correlated with the dependent variable (scour depth and SHAP values). The application of PCA enhances the correlation, and when using the CC-PCA-4 input parameter combination, the SVR model achieves high accuracy (R2 = 0.971, mean absolute percentage error = 7.54%). Moreover, its comprehensive evaluation in terms of stability, accuracy, and conservativeness surpassed that of other ML models and empirical formulas.

Prediction and Optimization of Surface Roughness and Cutting Forces in Turning Process Using ANN, SHAP Analysis, and Hybrid MCDM Method

As manufacturing technologies advance, the integration of artificial neural networks in machining high-hardness materials and optimization of multi-objective parameters is becoming increasingly prevalent. By employing modeling and optimization strategies during the machining of such materials, manufacturers can improve surface roughness and tool life while minimizing cutting time, tool vibrations, and cutting forces. In this paper, the aim was to analyze the impact of input parameters (cutting speed, feed rate, depth of cut, and insert radius) on surface roughness and cutting forces during the machining of 90MnCrV7 using feed-forward neural network models and SHAP analysis. Afterward, multi-criteria optimization was applied to determine the optimal parameter levels to achieve minimum surface roughness and cutting forces using the modified PSI-TOPSIS method. According to the SHAP analysis, the insert radius has the most significant impact on the surface roughness and passive force, while in the multi-criteria analysis, according to ANOVA results, the insert radius has the most significant impact on all considered outputs. The results show that an insert radius of 0.8 mm, a cutting speed of 260 m/min, a feed rate of 0.08 mm, and a depth of cut of 0.5 mm are the optimal combination of input parameters.

Comparative analysis of ensemble learning algorithms in water quality prediction

ABSTRACT Water is an essential resource necessary for the survival of all life forms, yet it is continually at risk of contamination. Accurate water quality prediction is essential for protecting ecosystem health. This study aims to assess the effectiveness of ensemble learning techniques, namely AdaBoost, gradient boosting, XGBoost, CatBoost, and LightGBM, in predicting water quality parameters in the Bara River Basin, Pakistan. Initially, a random forest model was used to determine the input water quality parameters combination for the selected target water quality variable. Then, the ML models were developed for each combination of input parameters and target water quality variables. The ML model's performance was assessed via statistical performance indicators, namely R2, mean squared error, and mean absolute error. The most suitable model was highlighted using compromise programming. The results reveal that the XGBoost and gradient boosting models outperform other algorithms based on statistical indicators, displaying remarkable predictive ability with near-perfect R2 values for HCO3, CO3, and Mg on the XGBoost model and electrical conductivity, SO4, Temp, and Ca on the gradient boosting model. Whereas CatBoost and LightGBM have a more robust performance on some parameters, such as pH and dissolved solids while its performance in other water quality parameters was weak.

Asteroid material classification based on multi-parameter constraints using artificial intelligence

Context. Material types of asteroids provide key clues to their evolutionary history and contained resources. The Gaia mission has released extensive low-resolution spectral observation data of small Solar System bodies. However, methods for classifying asteroids based on low-resolution space-based spectra are still inadequate, and do not fully leverage the complementary features of spectra and multiple intrinsic attributes of asteroids to achieve precise material classification. Aims. Our goal is to propose a method with a higher generalization accuracy for asteroid material classification by integrating multi-source information, identifying optimal feature combinations for model inputs, and deepening the understanding of relationships among asteroid parameters. Methods. The effective asteroid photometric, physical, and orbital parameters were screened using the information gain ratio and Spearman’s rank correlation coefficient. Then, artificial intelligence techniques were employed to combine asteroid spectra with the selected various parameters for six-class material classification. By comparing five machine learning models, we identified network structures with higher validation accuracy and stable generalization performance. Meanwhile, feature ablation experiments were conducted to determine the input parameter combinations suitable for different scenarios. Finally, based on the statistical results and model outputs, the constraint relationships among asteroid parameters were visualized and analyzed. Results. The proposed AsterRF model achieved a validation accuracy of 92.2%, an improvement of approximately 7.8 percentage points compared to existing methods that use only spectra. V-type asteroids exhibited the highest classification accuracy, followed by A-type and D-type. X-type asteroids had the lowest precision and recall, and were easily confused with C-type. The model generally showed higher classification confidence for S-type asteroids. The top five attributes that the model focused on are the phase slope parameter (G), orbital type, albedo, H magnitude, and effective diameter. Additionally, the correlations between asteroid materials and other parameters were generally below 0.4. Conclusions. Incorporating optimal asteroid parameter combinations can significantly enhance classification accuracy based on spectra. A dual-channel network that processes spectra and parameter inputs separately, and employs a self-attention mechanism for feature fusion is effective in combining multi-source asteroid information. Both the statistical correlations and model performance-based importance rankings of parameters contribute to understanding the constraint relationships among asteroid attributes.

A combined support vector regression with a firefly algorithm for prediction of energy consumption in wastewater treatment plants.

Wastewater treatment plants (WWTPs) comprise energy-intensive processes, serving as primary contributors to overall WWTP costs. This research study proposes a novel approach that integrates support vector regression (SVR) with the firefly algorithm (FFA) for the prediction of energy consumption in a WWTP in Chlef City, Algeria. The database comprises a comprehensive set of 1,653 samples, capturing diverse information categories. It includes chemical and physical characteristics, encompassing chemical oxygen demand, 5-day biochemical oxygen demand, potential of hydrogen, water temperature, total suspended sediment in water and basin, influent N-NH3 concentration, number of aerators, and operating time. Additionally, the hydraulic and energy-related parameters are represented by the flow entered at the station and the energy consumed by aerators, respectively. Finally, meteorological data, comprising rainfall, temperature, relative humidity, and the aridity index, are part of the dataset required for analysis. In this regard, 15 different models that correspond to 15 different combinations of input parameters are assessed in this study. The results show that the SVR-FFA-15 can render an improvement in the prediction accuracy of energy consumption in WWTPs. This study provides a useful tool for managing the energy consumption of wastewater treatment and makes insightful recommendations for future energy savings.

Comparative evaluation of machine learning models for assessment of seabed liquefaction using finite element data

Predicting wave-induced liquefaction around submarine pipelines is crucial for marine engineering safety. However, the complex of interactions between ocean dynamics and seabed sediments makes rapid and accurate assessments challenging with traditional numerical methods. Although machine learning approaches are increasingly applied to wave-induced liquefaction problems, the comparative accuracy of different models remains under-explored. We evaluate the predictive accuracy of four classical machine learning models: Gradient Boosting (GB), Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), and Random Forest (RF). The results indicate that the GB model exhibits high stability and accuracy in predicting wave-induced liquefaction, due to its strong ability to handle complex nonlinear geological data. Prediction accuracy varies across output parameters, with higher accuracy for seabed predictions than for pipeline surroundings. The combination of different input parameters significantly influences model predictive accuracy. Compared to traditional finite element numerical methods, employing machine learning models significantly reduces computation time, offering an effective tool for rapid disaster assessment and early warning in marine engineering. This research contributes to the safety of marine pipeline protections and provides new insights into the intersection of marine geological engineering and artificial intelligence.

GIS-based solar radiation modelling for photovoltaic potential in cities: A sensitivity analysis for the evaluation of output variability range

Currently, there is increasing attention to matching the buildings’ energy demand with energy produced by renewable sources. This approach serves to both reduce the environmental footprint, the target of several European policies and initiatives and ensure access to energy for everyone worldwide, aligning with one of the aims of the United Nations Agenda 2030. Urban environments are suitable for the installation of photovoltaic panels, which can be easily integrated within roofs to minimise land use and installation costs. Multiple studies focused on solar radiation as the most critical element affecting the estimation of the photovoltaic potential, in particular by resorting to Geographic Information Systems (GIS). Among different tools, which are useful for integrating layers with different spatial resolutions, the r.sun function embedded in GRASS GIS has been selected as the most suitable to capture complex variations resulting from weather and geographical parameters. However, a gap in the literature has emerged as no study is available to define the precision degree of the tool when inputting parameters derived from various data sources based on different model assumptions. The scope of this research is to collect open-access weather inputs and quantify the solar radiation output along with its variability range. Based on 28 test plans, resulting from the combination of input parameters from alternative data sources, this research quantified the output variability range to be ±20 %, thus requiring additional processing to increase the precision. The paper concludes with a critical assessment of the strengths and weaknesses of the current research, emphasizing the importance of validating the results with measured values in forthcoming research, to move from precision to accuracy evaluation.

Assessing the Tribological Impact of 3D Printed Carbon-Reinforced ABS Composite Cylindrical Gears

The tribological performance of carbon-reinforced acrylonitrile butadiene styrene (ABS) composites is very important in determining their suitability for advanced engineering applications. This study employs response surface methodology (RSM) to evaluate the effects of printing temperature and post-processing annealing on the wear resistance and frictional properties of these composites. A central composite design is used to systematically explore the interaction between these two factors, enabling the development of predictive models for key tribological parameters. The results reveal that both the coefficient of friction (COF) and wear are affected by printing and annealing temperatures, although in a non-linear manner. Moderate printing temperatures and lower annealing temperatures were found to reduce friction and wear, with annealing temperature having a more pronounced effect on wear. To further optimize these responses, the desirability approach was applied for predicting the optimal conditions. The optimal combination of input parameters for minimizing both COF and wear was found to be a printing temperature of 256 °C and an annealing temperature of 126 °C. This research provides valuable insights for optimizing additive manufacturing processes of carbon-reinforced ABS composites, contributing to enhanced material durability in practical applications.

Optimizing Cylindrical Grinding Proficiency and Surface Finish for EN24 Steel Alloy with the Use of Taguchi Analysis and Artificial Neural Network

Grinding is a material removal procedure used at the end of manufacturing to obtain high dimensional precision and a desirable surface polish. External cylindrical grinding is a type of grinding procedure that used to grind cylindrical things such as engine shafts, connecting rods, spindles, and axles. This study examines how control settings affect response parameters using EN24 alloy steel on a cylindrical grinding machine. Response parameters are material removal rate (MRR) and surface roughness (Ra), whereas control parameters are workpiece speed (v), feed rate (f), and depth of cut (d). Taguchi DOE is employed with L16 orthogonal array to optimize control parameter levels, followed by cylindrical grinding tests.An artificial neural network (ANN) model was created to validate and predict cylindrical grinding reaction parameters. To train the network, experimental data was used, and ANN predictions were compared to real data. Data collection has also completed.In the purposed work minimum Ra is 0.52 µm and maximum MRR is 0.6588 g/sec found and results indicate that feed rate and workpiece speed are the most dominating parameters of cylindrical grinding. The experimental results can be employed by various manufacturing and industrial firms and they can select suitable combinations of input parameters for desired Ra and MRR

Development on Surrogate Models for Predicting Plume Evolution Features of Groundwater Contamination with Natural Attenuation

Predicting the key plume evolution features of groundwater contamination are crucial for assessing uncertainty in contamination control and remediation, while implementing detailed complex numerical models for a large number of scenario simulations is time-consuming and sometimes even impossible. This work develops surrogate models with an effective and practicable pathway for predicting the key plume evolution features, such as the distance of maximum plume spreading, of groundwater contamination with natural attenuation. The representative various scenarios of the input parameter combinations were effectively generated by the orthogonal experiment method and the corresponding numerical simulations were performed by the reliable Groundwater Modeling System. The PSO-SVM surrogate models were first developed, and the accuracy was gradually enhanced from 0.5 to 0.9 under a multi-objective condition by effectively increasing the sample data size from 54 sets to 78 sets and decreasing the input variables from 25 of all the considered parameters to a smaller number of the key controlling factors. The statistical surrogate models were also constructed with the fitting degree all above 0.85. The achieved findings provide effective generic surrogate models along with a scientific basis and investigation approach reference for the environmental risk management and remediation of groundwater contamination, particularly with limited data.

Uncertainty and Latin Hypercube Sampling in Geotechnical Earthquake Engineering

A soil–foundation–structure system (SFSS) often exhibits different responses compared to a fixed-base structure when subjected to earthquake ground motion. Both kinematic and inertial soil–foundation–structure interactions can significantly influence the structural performance of buildings. Numerous parameters within an SFSS affect its overall response, introducing inherent uncertainty into the solution. Performing time history analyses, even for a linear elastic coupled SFSS, requires considerable computational effort. To reduce the computational cost without compromising accuracy, the use of the Latin Hypercube Sampling (LHS) technique is proposed herein. Sampling techniques are rarely employed in soil–foundation–structure interaction analyses, yet they are highly beneficial. These methodologies allow analyses determined by sampling to be conducted using commercial codes designed for deterministic analyses, without requiring any modifications. The advantage is that the number of analyses determined by the sampling size is significantly reduced as compared to considering all combinations of input parameters. After identifying the important samples, one can evaluate the seismic demand of selected soil–foundation–bridge pier systems using finite element numerical software. This paper indicates that LHS reduces computational effort by 60%, whereas structural response components (translation, rocking) show distinct trends for different systems.

Control factor optimization for friction stir processing of AA8090/SiC surface composites

Friction Stir Processing is a state-of-the-art technology for microstructure refinement, material property enhancement, and surface composites fabrication. This investigation concentrates on AA8090/SiC surface composites produced via friction stir processing. Experiments were conducted by varying the following friction stir processing parameters: Tool rotational speed (800–1400 rpm), Tool traverse speed (25–75 mm/min), and Groove width (1.0–1.8 mm). Response measures encompassed Ultimate Tensile Strength and surface roughness. Central Composite Design of Response Surface Methodology designed the experiments and mathematical relationships established between input parameters and ultimate tensile strength and surface roughness. Analysis of variance was used to test the model's adequacy. The models examined individual and interaction effects of input factors on ultimate tensile strength and surface roughness of surface composites. A combinations of input parameters was identified that yields the maximum ultimate tensile strength and minimum surface roughness. The current work employs the friction stir processing approach to synthesis near-net-shaped surface composites without additional machining by systematically optimizing process parameters. Results indicate that increasing tool rotational speed produces well-finished AA8090/SiC surface composites with decreased strength. In contrast, increased tool traverse speed and groove width generate surface composites with rougher surfaces and higher strength. Surface and contour plots further explored the influence of parameter interactions on responses.

Analytic Parameterization of Longwave Optical Properties of Bulk Vegetation Layer Permitting Non‐Zero Leaf Reflectivity and Its Implementation in CLM5

AbstractFor modern land surface models (LSMs) representing a singular bulk vegetation layer, the longwave optical properties (i.e., emissivity, reflectivity, and transmittivity) of vegetation layer are derived with a simplified constraint of assuming zero leaf reflectivity. This constraint is necessary, for instance, to the Beer–Lambert (B–L) law to establish a relationship between the optical properties and leaf area index. However, the simplified constraint leads to an overestimation of land surface emissivity in the vegetated regions. In this study, we introduce a new scheme considering realistic leaf reflectivity values rather than assuming zero. This new scheme is based on the relationship derived from the B–L law, but it is statistically augmented to consider the effects of leaf reflections. It is designed to emulate a multi‐vegetation‐layer numerical model known as the Norman model, which is capable of numerical calculations of multi‐reflections among leaves. This new method consists of only a couple of simple equations; despite its simplicity, it very closely mimics the Norman model; The discrepancy of the results between the new method and the Norman model is less than measurement uncertainties for any combination of input parameters. When the new scheme is implemented in the Community Land Model version 5 (CLM5), the land surface emissivity values are simulated much more consistently with global measurements, resulting in significant alterations of land surface energy budget. The enhanced realism through our new scheme is poised to contribute to more accurate numerical weather and climate simulations.

A Benchmarking Method to Rank the Performance of Physics-Based Earthquake Simulations

Abstract Physics-based earthquake simulators are an increasingly popular modeling tool in earthquake forecasting for seismic hazard as well as fault rupture behavior studies. Their popularity comes from their ability to overcome completeness limitations of real catalogs, and also because they allow reproducing complex fault rupture and interaction patterns via modeling the physical processes involved in earthquake nucleation and propagation. One important challenge of these models revolves around selecting the physical input parameters that will yield the better similarity to earthquake relationships observed in nature, for instance, the frictional parameters of the rate-and-state law—a and b—or the initial normal and shear stresses. Because of the scarcity of empirical data, such input parameters are often selected by trial–error exploration and predominantly manual model performance analyses, which can overall be time consuming. We present a new benchmarking approach to analyze and rank the relative performance of simultaneous earthquake simulation catalogs by quantitatively scoring their combined fit to three reference function types: (1) earthquake-scaling relationships, (2) the shape of the magnitude–frequency distributions, and (3) the rates of the surface ruptures from paleoseismology or paleoearthquake occurrences. The approach provides an effective and potentially more efficient approximation to easily identify the models and input parameter combinations that fit more closely to earthquake relations and behavior. The approach also facilitates the exhaustive analysis of many input parameter combinations, identifying systematic correlations between parameters and model outputs that can potentially improve the overall understanding of the physics-based models. Finally, we demonstrate how the method results agree with the published findings in other earthquake simulation evaluations, a fact that reinforces its overall usefulness. The model ranking outputs can be useful for subsequent analyses, particularly in seismic hazard applications, such as the selection of appropriate earthquake occurrence rate models and their weighting for a logic tree.

Modeling flow resistance and geometry of dunes bed form in alluvial channels using hybrid RANN–AHA and GEP models

Dunes formation in sandy rivers significantly impacts flow resistance, subsequently affecting water levels, flow velocity, river navigation, and hydraulic structures performance. Accurate prediction of flow resistance and dune geometry (length and height) is essential for environmental engineering and river management. The current paper introduces two models to evaluate the flow resistance and geometry of dunes formed in sand-bed channels. The first model, RANN–AHA is a hybrid artificial intelligence model using the recurrent artificial neural network (RANN) linked with the artificial hummingbird optimization algorithm (AHA) to optimize the biases and weights of the neural network model. The second model uses gene expression programming (GEP) as a nonlinear approach based on a genetic algorithm (GA) and genetic programming (GP) to explicitly determine dune characteristics. For both models, the input parameters include flow and sediment characteristics, while Manning's roughness coefficient (nM), and relative dune height, h/H or h/L, were used as output parameters where h is the dune height, H is the flow depth above the dune crest, and L is the dune length. Five different published flume data sets were compiled for the analysis. Sensitivity analysis was done using different combinations of input parameters. It was found that the combination of hydraulic radius divided by median diameter (RH/d50), Reynolds number (Re), Particle densimetric Froude number (F∗), and grain Froude number (FG) yielded the best prediction accuracy for estimating Manning nM and relative height, h/H or h/L, with a root mean square error (RMSE) = 0.00027, 0.0504, and 0.0078 and a correlation coefficient (R) = 0.9989, 0.942, and 0.9272, respectively. Model verification proved that the RANN–AHA model outperformed the GEP model and most of the previous studies available in the literature when predicting the roughness coefficient and dune geometry in sand bed channels.

Study of biological quality of lake waters based on phycocyanin using tree‐based methodologies

AbstractThe provision of drinking water, agricultural, and industrial applications by reservoirs has made lake exploration and monitoring unavoidable. The features of the ecosystem, particularly physical and chemical elements, influence the evaluation of the quality of water resources. Lakes undergo extensive qualitative changes due to their vast amount of water. In general, these bodies of water represent geological conditions as well as water contamination produced by natural and human activities. In the present research, the prediction of the amount of phycocyanin (fPC) in the water of Lake Michigan has been implemented employing four tree‐based machine learning techniques based on seasonality factors. Phycocyanin has significant effects on quality parameters such as turbidity, chlorophyll concentration, algal bloom, and dissolved oxygen in water by affecting the photosynthesis process of algae. Therefore, in this study, the prediction of the amount of phycocyanin dissolved in the lake water using the mentioned variables, along with the temperature of the water, specific conductance, and pH, has been able to interpret the quality of the water and the occurrence of phenomena such as algal blooms. The results of the models in predicting fPCs equal to 0.44 and 0.55 μg/L were consistent with the natural conditions of the lake, and it seems that ensemble tree–based models, along with the biological index of fPC, formed the right combination of input and output parameters in modeling and obtained the lowest prediction error (root‐mean‐square error [RMSE] boosted trees = 0.0140 and RMSE random forests = 0.0141 μg/L).

Prediction of effective parameters for 3D printing of poly lactic acid-carbon fibre composites using intelligent frameworks based on mechanical response

Purpose This paper aims to find the effective 3D printing process parameters based on mechanical characteristics such as tensile strength and hardness of poly lactic acid (PLA)/carbon fibre composites (CF-PLA) by implementing intelligent frameworks. Design/methodology/approach The experiment trials are conducted based on design of experiments (DoE) using Taguchi L9 orthogonal array with three factors (speed, infill % and pattern type) and three levels. The factors have been optimized by solving the regression equation which is obtained from analysis of variance (ANOVA). The contour plots are generated by response surface methodology (RSM). The influencing parameters are found by using Box–Behnken design. The second order response surface model demonstrated the optimal combination of input parameters for higher tensile strength and hardness. Findings The influencing parameters are found by using Box–Behnken design. The second order response surface model demonstrated the optimal combination of input parameters for higher tensile strength and hardness. The results obtained from RSM are also confirmed by implementing the machine learning classifiers, such as logistic regression, ridge classifier, random forest, K nearest neighbour and support vector classifier (SVC). The results show that the SVC can predict the optimized process parameters with an accuracy of 95.65%. Originality/value 3D printing parameters which are considered in this work such as pattern types for PLA/CF-PLA composites based on intelligent frameworks has not been attempted previously.

An in-silico approach towards multivariate acceptable ranges in biopharmaceutical manufacturing

Multivariate interactions between process parameters can heavily impact product quality and process performance in biopharmaceutical manufacturing processes. Thus, multivariate interactions should be identified and appropriately controlled. This article describes an in-silico approach to establish multivariate acceptable ranges; these ranges help to illustrate the combined impact of multiple input variables on product quality and process performance. Additionally, this article includes a case study for a monoclonal antibody polishing application.Proven acceptable ranges are set by changing only one input parameter at a time while keeping all others constant to understand the impact of process variability on product quality or process performance, but the impact of synergistic variables are not evaluated. Within multivariate acceptable ranges, any combination of input parameters of a unit operation yields the desired product quality and process performance. The layered approach applied in this article is based on risk assessment and statistical models to leverage prior knowledge and existing data. The risk assessment is specific for a manufacturing facility but is applicable to multiple products manufactured in the same facility. No additional wet-lab experiments are required for building the statistical models when development and process characterization are executed using a design of experiments approach, compared to a univariate evaluation of data. The established multivariate acceptable range justifies revised normal operating ranges to ensure process control. Further, the determination of multivariate acceptable ranges adds to overall process knowledge, ultimately supporting the implementation of a more effective control strategy.