1,720,992 research outputs found
Real time monitoring of train wheels and track conditions based on time series analysis and multivariate data analysis
Get PDFAs the railways are a significant section of transportation infrastructure, it is crucial to use good maintenance procedures for railway networks. Condition-monitoring of railway wheels and tracks is especially important where extremely unfortunate failure happens. So regular examination of railway tracks and wheels health is required to maintain safe and reliable train operations. The traditional method of manually inspecting rail tracks is inefficient and prone to human error and bias. This study aims to improve the aging railway system by using an automated solution to address these difficulties.
A series of experimental tests were carried out before data collection. A LabVIEW application was made for recording data and tested with a function generator and Data Acquisition (DAQ) device. The provided Cemit Data Collection (CDC) time-series data were converted into frequency domain through Fast Fourier Transform (FFT) and Wavelet Transform (WT) in MATLAB. Multivariate data analysis was done through FFT and WT data for fault detection using Principal Component Analysis (PCA) and Partial Least Square Regression (PLS-R).
During PCA, the score plot of FFT data depicted the important samples present in it. However, the score plot of WT data illustrated that important samples were absent. Hence, after further analysis of FFT data using PLSR, the calibration model was built which was validated from test set validation. From this, Y predicted was almost equal to the Y reference, hence the model performance can be significant.
From comparing the score plots of FFT and WT data, it is concluded that the WT data is not useful for discovering faults in the tracks using multivariate data analysis because no relevant samples for further investigation were detected. To get better results, it's also a good idea to analyze multi-channel or more sensor data
Real time monitoring of train wheels and track conditions based on acoustic measurements and multivariate data analysis
Get PDFWith rail-vehicle transport considered as greener and more sustainable means of travel in recent global times, there is a need for study and exploration of a real-time monitoring system for rail and its vehicle structures, that can facilitate identification of track defects
and the vehicle’s structural monitoring. The method proposed for this is acousticchemometrics, is a data-driven approach, where acoustics, or vibrational data iscollected across three orthogonal axes and analyzed for structures or patterns that can suggest and/or indicate any faults or irregularities as can validate some of the hypotheses
proposed by from Cemit.
With an objective to understand and apply the mechanism of acoustic chemometrics for such purposes, pre-existing acoustic data from Cemit’s data collector has been used. The accelerometer data was converted from time-domain to frequency domain, and plotted across a spectrum to have an initial understanding of the dominant frequencies at play.
Based on visual observations of the existing dominant frequencies, unusual or interesting spots along the original track of Brevikbanen, which was the track used for experiment, was suggested for deeper analysis. An initial chemometric analysis on averaged spectra was performed as an exploratory analysis, and modifications were suggested for more productive and specific chemometric analysis. The need for more experiments to produce reference data is also talked of, and a brief review of few studies that has worked with modelling and simulating track defects and rail-wheel interactions were also brought into light
Real time monitoring of train wheels and track conditions based on time series analysis and multivariate data analysis
No full textAs the railways are a significant section of transportation infrastructure, it is crucial to use good maintenance procedures for railway networks. Condition-monitoring of railway wheels and tracks is especially important where extremely unfortunate failure happens. So regular examination of railway tracks and wheels health is required to maintain safe and reliable train operations. The traditional method of manually inspecting rail tracks is inefficient and prone to human error and bias. This study aims to improve the aging railway system by using an automated solution to address these difficulties.
A series of experimental tests were carried out before data collection. A LabVIEW application was made for recording data and tested with a function generator and Data Acquisition (DAQ) device. The provided Cemit Data Collection (CDC) time-series data were converted into frequency domain through Fast Fourier Transform (FFT) and Wavelet Transform (WT) in MATLAB. Multivariate data analysis was done through FFT and WT data for fault detection using Principal Component Analysis (PCA) and Partial Least Square Regression (PLS-R).
During PCA, the score plot of FFT data depicted the important samples present in it. However, the score plot of WT data illustrated that important samples were absent. Hence, after further analysis of FFT data using PLSR, the calibration model was built which was validated from test set validation. From this, Y predicted was almost equal to the Y reference, hence the model performance can be significant.
From comparing the score plots of FFT and WT data, it is concluded that the WT data is not useful for discovering faults in the tracks using multivariate data analysis because no relevant samples for further investigation were detected. To get better results, it's also a good idea to analyze multi-channel or more sensor data
Monitoring and Control of Shale Shaker Performance with Focus on Caving Detection and Volume
No full textDrilling operations often face challenges from unstable wellbore conditions, where cavings, larger rock fragments, can signal early signs of failure. This thesis explores automated techniques for detecting and classifying such cavings, using a controlled test setup and machine learning methods.
A dataset was created using castings of real cavings, recorded under simulated shaker conditions. Segmentation and shape classification tools were developed, and object detection models were trained using YOLOv11. The system performed reliably in segmentation, but classification accuracy was affected by dataset imbalance.
Additionally, the project investigated stereo vision to enhance cavings detection and for estimating the volume of drilled cuttings. Although limited by noisy and inconsistent depth data, the theory has shown potential with more advanced depth sensors.
Overall, the work demonstrates a promising pipeline for real-time cavings detection, with recommendations for future improvements in data quality, depth sensing, and labelling accuracy
Dynamic system multivariate calibration with low-sampling-rate y data
No full textWhen the data in principal component regression (PCR) or partial least squares regression (PLSR) form time series, it may be possible to improve the prediction/estimation results by utilizing the correlation between neighboring observations. The estimators may then be identified from experimental data using system identification methods. This is possible also in cases where the response variables in the experimental data are sampled at a low and possibly irregular rate, while the regressor variables are sampled at a higher rate. After a discussion of the options available, the paper shows how the autocorrelation of the regressor variables in such multirate sampling cases may be utilized by identification of parsimonious output error (OE) estimators. An example using acoustic power spectrum regressor data is finally presented
CO2 capture by MEA solvent: Chemical Speciation models of CO2 derived species & total solvent alkalinity by multivariate data analysis of FTIR spectra from the CO2 Technology Centre Mongstad
No full textThis study is the next step of ongoing research at University of South-eastern Norway (USN) to enable multivariate analysis for industrial scale capture process.
This study has been done in collaboration with Technology Centre Mongstad (TCM) as one of the largest post-combustion capture test centers in the world. In 2015 and 2017, TCM operated two comprehensive test campaigns using the benchmark aqueous 30 wt% Monoethanolamine (MEA) solvent.
Through collaboration with TCM, USN has been provided with the laboratory test results of the collected samples and analytical data including Fourier Transform Infrared (FTIR) spectra from these two campaigns.
The received FTIR spectra as a multivariate data source contains the plenty of important chemical information of the samples. To extract these information, partial least square regression (PLSR) method has been used in this study.
The PLSR models of Total Inorganic Carbon (TIC) and Total Alkalinity (Tot-Alk) which have been prepared by using FTIR spectra from these campaigns are presented. From this study, it is evident that online monitoring integrated with spectroscopic analysis is an appropriate method for CO2 capture plant online monitoring. Through this, it is possible to reduce the time consuming and expensive conventional laboratory analyses of samples from CO2 capture plants.
Finally, the predictability of PLSR models for preparation of two campaigns was is tested and error of predictions were studied
Chemometric analysis of a new benchmark CO2 capture solvent
Get PDFGreenhouse gases have challenging consequences on climate change and CO2 is one of the most effective elements in this issue. the methods to mitigate these effects such as capturing the carbon dioxide are the aim of this project. CESAR1 (27 wt% 2-amino-2-methyl-1-propanol+13 wt% Piperazine) as an aqueous solution is the main solution to consider in this project. Many experiments have been implemented in this study to evaluate this solution. FTIR, pH as well as density measurements are thoroughly studied. All data is evaluated in Unscrambler with the PLS-R method. However, NMR results could not be prepared because of a delayed response from the SINTEF company as well as titration due to several errors. But all experiments that have been completed during this study, were given practical and considerable information about the CESAR1 solution. pH and density measurements are performed to give results about the effects of loading CO2 in the CESAR1 solution. FTIR test is performed to determine the composition of the CESAR1. A series of tests have been done to analyze the effect of CO2 on CESAR1 solution. According to the PLS-R analysis, it has been reached to the composition of the CESAR1 solution and how the CO2 loaded affected the investigations. From my findings, I can conclude that the FTIR technique is an effective method for the decomposition of this solution to capture carbon dioxide
Monitoring a New Benchmark Solvent for CO2 Capture: Pushing Technical Boundaries
Get PDFGlobal warming of the earth is increasing significantly due to the emission of greenhouse gases where CO2 is the main influential gas for this adverse effect. To mitigate this problem CO2 capture with aqueous amine solvent is a most promising technology for many years.
The study focuses on a new benchmark solvent for CO2 capture, CESAR1, which consists of 2-amino-2-methylpropan-1-ol (AMP) and piperazine (PZ) blend. The objectives of this thesis are to prepare different CO2 loading and unloading sample with various range of CESAR1 concentration and measured the physical properties of density, pH, conductivity of the prepared sample to predict the solvent performance. Using FTIR Spectroscopy spectral regions of interest for estimation of α-CO2 loading, AMP, PZ, density, pH, and conductivity are identified as respectively, 1575-1213 cm-1, 1095 - 877 cm−1, 1213 - 1096 cm−1, 1080 to 502 cm-1, 1835 to 502 cm-1, and 1810 to 502 cm-1.
The methodology involves preparation of aqueous amine samples and estimation of density, pH, and conductivity of the amine solvents. To understand the solvent behaviour and correlation between the multivariable data partial least square regression (PLS-R) modelling has been conducted in this study.
The key finding investigates the effect of CO2 loading in the CESAR1 blend samples. Different species concerned with CESAR1 blend (AMP+PZ) are identified through speciation of pre-processed FTIR spectra including some stretching band. Six PLS-R models are developed to predict α-CO2 loading, AMP, PZ, density, pH, and conductivity of the observed CESAR1 solvent. The model accurately predicts RMSEP values of 0.0234 mol/mol, 0.1338 mol/kg, 0.1611 mol/kg, 0.0048 g/cm3, 0.1842, and 0.6321 mS/cm for α-CO2 loading, AMP, PZ, density, pH, and conductivity respectively with a good fit performance.
From this study, it is quite evident that online monitoring integrated with FTIR spectroscopy analysis is an appropriate method for CO2 capture
- …
