BITS Faculty Publications

Permanent URI for this communityhttp://localhost:4000/handle/123456789/1867

Browse

Has files: NoSubject: search.filters.subject.Machine Learning

Search Results

Now showing 1 - 10 of 64
  • No Thumbnail Available
    Item
    Machine Learning Aided Predictions for Capacity Fade of Li-Ion Batteries
    (IOP, 2022-05) Roy, Tribeni
    Future demands high power and high energy density devices that can be sustainably built and easily maintained. It is seen that among various energy storage devices, the demanding role lithium-ion batteries play in powering electronic gadgets to electric vehicles, is highly significant. Hence, the researchers around the world are trying to solve the riddles of the lithium-ion batteries and make it more efficient. One such problem that researchers are trying to solve is battery degradation and capacity fade. In this work, we made a battery forecasting model that can predict the capacity fade using electrochemical impedance spectroscopy (EIS) data. Two machine learning techniques like, support vector regression (SVR) and multi-linear regression (MLR) were utilized to analyse the data and predict the capacity fade for lithium-ion battery. Principal component analysis was also carried out to determine the most relevant feature from the data. From the analysis it was found that that SVR has a better prediction accuracy than MLR or pre-existing Gaussian process regression (GPR) results and among the two kernels of support vector regression, radial basis function (rbf) kernel has better prediction accuracy with R2 score of 0.9194 than the linear kernel with R2 score of 0.6559.
  • No Thumbnail Available
    Item
    Research Article Investigation of Counterflow Microchannel Heat Exchanger with Hybrid Nanoparticles and PCM Suspension as a Coolant
    (MDPI, 2021-03) Bhattacharyya, Suvanjan
    A circular tube fitted with novel corrugated spring tape inserts has been investigated. Air was used as the working fluid. A thorough literature review has been done and this geometry has not been studied previously, neither experimentally nor theoretically. A novel experimental investigation of this enhanced geometry can, therefore, be treated as a new substantial contribution in the open literature. Three different spring ratio and depth ratio has been used in this study. Increase in thermal energy transport coefficient is noticed with increase in depth ratio. Corrugated spring tape shows promising results towards heat transfer enhancement. This geometry performs significantly better (60% to 75% increase in heat duty at constant pumping power and 20% to 31% reduction in pumping power at constant heat duty) than simple spring tape. This paper also presented a statistical analysis of the heat transfer and fluid flow by developing an artificial neural network (ANN)- based machine learning (ML) model. The model is evaluated to have an accuracy of 98.00% on unknown test data. These models will help the researchers working in heat transfer enhancementbased experiments to understand and predict the output. As a result, the time and cost of the experiments will reduce. The results of this investigation can be used in designing heat exchangers
  • No Thumbnail Available
    Item
    Application of New Artificial Neural Network to Predict Heat Transfer and Thermal Performance of a Solar Air-Heater Tube
    (MDPI, 2021-07) Bhattacharyya, Suvanjan
    In the present study, the heat transfer and thermal performance of a helical corrugation with perforated circular disc solar air-heater tubes are predicted using a machine learning regression technique. This paper describes a statistical analysis of heat transfer by developing an artificial neural network-based machine learning model. The effects of variation in the corrugation angle (θ), perforation ratio (k), corrugation pitch ratio (y), perforated disc pitch ratio (s), and Reynolds number have been analyzed. An artificial neural network model is used for regression analysis to predict the heat transfer in terms of Nusselt number and thermohydraulic efficiency, and the results showed high prediction accuracies. The artificial neural network model is robust and precise, and can be used by thermal system design engineers for predicting output variables. Two different models are trained based on the features of experimental data, which provide an estimation of experimental output based on user-defined input parameters. The models are evaluated to have an accuracy of 97.00% on unknown test data. These models will help the researchers working in heat transfer enhancement-based experiments to understand and predict the output. As a result, the time and cost of the experiments will reduce
  • No Thumbnail Available
    Item
    Heat and Fluid Flow Analysis and ANN-Based Prediction of A Novel Spring Corrugated Tape
    (MDPI, 2021-03) Bhattacharyya, Suvanjan
    A circular tube fitted with novel corrugated spring tape inserts has been investigated. Air was used as the working fluid. A thorough literature review has been done and this geometry has not been studied previously, neither experimentally nor theoretically. A novel experimental investigation of this enhanced geometry can, therefore, be treated as a new substantial contribution in the open literature. Three different spring ratio and depth ratio has been used in this study. Increase in thermal energy transport coefficient is noticed with increase in depth ratio. Corrugated spring tape shows promising results towards heat transfer enhancement. This geometry performs significantly better (60% to 75% increase in heat duty at constant pumping power and 20% to 31% reduction in pumping power at constant heat duty) than simple spring tape. This paper also presented a statistical analysis of the heat transfer and fluid flow by developing an artificial neural network (ANN)-based machine learning (ML) model. The model is evaluated to have an accuracy of 98.00% on unknown test data. These models will help the researchers working in heat transfer enhancement-based experiments to understand and predict the output. As a result, the time and cost of the experiments will reduce. The results of this investigation can be used in designing heat exchangers.
  • No Thumbnail Available
    Item
    Turbulent Flow Heat Transfer through a Circular Tube with Novel Hybrid Grooved Tape Inserts: Thermohydraulic Analysis and Prediction by Applying Machine Learning Model
    (MDPI, 2021-03) Bhattacharyya, Suvanjan
    The present experimental work is performed to investigate the convection heat transfer (HT), pressure drop (PD), irreversibility, exergy efficiency and thermal performance for turbulent flow inside a uniformly heated circular channel fitted with novel geometry of hybrid tape. Air is taken as the working fluid and the Reynolds number is varied from 10,000 to 80,000. Hybrid tape is made up of a combination of grooved spring tape and wavy tape. The results obtained with the novel hybrid tape show significantly better performance over individual tapes. A correlation has been developed for predicting the friction factor (f) and Nusselt number (Nu) with novel hybrid tape. The results of this investigation can be used in designing heat exchangers. This paper also presented a statistical analysis of the heat transfer and fluid flow by developing an artificial neural network (ANN)-based machine learning (ML) model. The model is trained based on the features of experimental data, which provide an estimation of experimental output based on user-defined input parameters. The model is evaluated to have an accuracy of 98.00% on unknown test data. These models will help the researchers working in heat transfer enhancement-based experiments to understand and predict the output. As a result, the time and cost of the experiments will reduce.
  • No Thumbnail Available
    Item
    Analysis of Density of Laser Powder Bed Fusion Fabricated Part Using Decision Tree Algorithm
    (Springer, 2023-05) Mishra, Radha Raman
    Additive manufacturing (AM) enabled manufacturing industries to fabricate metallic components with complex shapes. However, the properties of additively manufactured parts need further improvements to compete with the performance of traditionally manufactured parts. Machine learning (ML) models provide an alternative to study the correlation between the process parameters–properties of the fabricated parts. In the present work, the ML approach has been applied to understand the effect of AM process parameters on the density of additively built parts. The decision tree model was developed for the laser powder bed fusion-processed parts based on the input parameters such as laser power, scan speed, hatching space, energy density, and build rate. The model was trained and tested with experimental data obtained from the relevant literature. The process parameters were optimized to achieve the desired density of the part. A good agreement was indicated between the predicted and experimental data. The study revealed the applicability and potential of the model to determine and predict the density of the additively manufactured parts.
  • No Thumbnail Available
    Item
    A Phase-wise Analysis of Machine Learning based Human Activity Recognition using Inertial Sensors
    (IEEE, 2020) Kala, Prateek
    Advancements in sensor technology, expanded analytical skills, and advancement in the field of Machine Learning (ML) and Deep Learning (DL), have all resulted in a substantial increase in popularity and wearable sensor performance for Human Activity Recognition (HAR). The real challenge is to spot the events in an unsupervised environment. In this study, we have tried to build a user-friendly, as well as an effective HAR system using an inertial sensor for eight everyday activities performed. The data was collected by our research team, using a single inertial sensor in a fully unsupervised setup. The eight tasks include: standing, sitting, sleeping, running, walking, cycling, upstairs and downstairs. This paper aims to present a detailed analysis and comparison for three primary aspects of a general HAR which contributes to the overall system performance. This involves analyzing the effects of pre-processing, comparing several extraction and selection methods for generating features from time-series data, and finally building and validating the performance of various classification methods to obtain the best combination of the three. The classification methods included in this study are Logistic Regression, K-Nearest Neighbors, Support Vector Machines and Artificial Neural Networks. After choosing the best parameters and techniques, we achieve a remarkable performance for recognizing the eight activities with an overall accuracy of 93.6%.
  • No Thumbnail Available
    Item
    Development and comparison of machine-learning algorithms for anomaly detection in 3D printing using vibration data
    (Springer, 2023-06) Sangwan, Kuldip Singh
    3D printing is an emerging technology that converts digital models directly into physical objects. However, abnormal vibrations during the 3D printing process significantly affect the product quality, and also lead to possible failures of the printer components. This paper aims at developing machine-learning algorithms for anomaly detection or abnormal behavior of a 3D printer using vibration data. The proposed algorithms utilize vibration data from a sensor mounted on the printer. Data are then trained and validated developing four machine-learning algorithms to detect anomalies due to the structural or mechanical defects of the printer. Performances of the proposed four algorithms were evaluated and compared. It was found that the proposed long short-term memory (LSTM) algorithm has the best accuracy of 97.17% as compared to other algorithms. The novelty of the present work lies in detecting anomalies with high accuracy due to structural or mechanical faults in 3D printers using a low-cost sensor. The significance of the current work lies in its ability to achieve error-free 3D printing, resulting in less material waste, reduced human intervention and costs, and improved product quality by detecting potential anomalies during printing. The proposed algorithm terminates the printing if any anomaly is detected.
  • No Thumbnail Available
    Item
    Development of a cyber physical production system framework for 3D printing analytics
    (Elsevier, 2023-10) Sangwan, Kuldip Singh
    3D printing technology is considered one of the emerging areas to deal with global sustainability challenges and to facilitate the Industry 4.0 adoption. However, 3D printing technology is still immature due to several limitations and negative perceptions about its quality and performance. The goal of this paper is to propose a cyber physical production system (CPPS) framework for a 3D printer to (i) monitor the process, parameters, and carbon footprint, (ii) predict the nozzle’s remaining useful life (RUL), and (iii) prescribe optimum 3D printing parameters for minimizing carbon footprint and printing time, simultaneously at the targeted surface quality. Experiments were designed based on Taguchi L-27 orthogonal array to investigate the relationship between printing parameters and performance characteristics. The usefulness of the proposed framework has been demonstrated for a 3D printer to predict the remaining useful life of the printer nozzle (prognostic model), and to find an optimal combination of printing parameters for the simultaneous optimization of sustainability and productivity at the targeted surface quality (prescriptive model). Layer height was found to have a statically significant impact on the specific carbon footprint followed by scale and bed temperature. Layer height is the only statically significant contributor to the surface roughness of 3D printed parts. The scale and layer height followed by infill have significant effect on the printing time. The significance of the present work lies in enhancing the performance of a conventional 3D printer using low-cost smart sensors, devices, and open-source software. The usefulness of the proposed CPPS framework is demonstrated as a decision support tool for a 3D printer real-time monitoring, visualization, and control. The proposed CPPS framework and its application for prognostic and prescriptive analytics is generic in nature, and is transferable and applicable to other FDM 3D printers, irrespective of brand and size.
  • No Thumbnail Available
    Item
    Development of Machine Learning Algorithm for Characterization and Estimation of Energy Consumption of Various Stages during 3D Printing
    (Elsevier, 2022) Sangwan, Kuldip Singh
    Energy usage in industries is one of the major contributors for climate change, biodiversity loss and resource scarcity. Technological advancements in digitalization led by Industry 4.0 facilitates affordable energy monitoring systems. This allows comprehensive understanding of the primary energy needs and improvement in the areas of inefficiency of a modern manufacturing system. Machine learning has the potential to reveal untapped insights, providing decision support for sustainable manufacturing by improving environmental performances, significant savings, and operational opportunities. The objectives of this research paper are to develop a machine learning algorithm for characterization, and to estimate the energy consumption of various stages in 3D printing. Machine learning model is developed using long short-term memory algorithm, and is trained, validated, and deployed for the classification of various stages during 3D printing process. Furthermore, energy consumption in each stage is estimated based on Simpson’s rule.