In this study, a genetic algorithm-based laser beam (LB) path optimization method is presented to improve laser-based additive manufacturing (LBAM). To emulate the LBAM process, LB irradiation of a ...thin metal substrate is applied. The LB path generation is formulated as the search for the optimal sequence of LB irradiation into the cells on the substrate that minimizes the fitness function, which is composed of two components, i.e., thermal fitness and process fitness. The thermal fitness is expressed by the average thermal gradient, and a simple thermal model is developed to simulate the effects of laser-induced heat input on the temperature distribution in the substrate. The process fitness regulates the suitability of the proposed LB path for the implementation of the LBAM process. In addition to standardized tool paths (i.e., raster, spiral, etc.), novel LB path generators are proposed to define the initial population of LB path solutions. To implement a genetic algorithm-based LB path optimization, a framework is proposed, and custom initialization, crossover, and mutation operators are developed for application in LBAM. The effectiveness of the proposed approach is demonstrated through a simulation case study aiming to identify LB paths that minimize the fitness function and thus provide more suitable LB path solutions with respect to the defined fitness function. Compared with the traditional trial-and-error LB path formulations, the proposed approach provides an improved and automated method for an efficient laser beam path selection in LBAM.
This study presents the results of acoustic emission (AE) measurements and characterization in the loading of biocomposites at room and low temperatures that can be observed in the aviation industry. ...The fiber optic sensors (FOS) that can outperform electrical sensors in challenging operational environments were used. Standard features were extracted from AE measurements, and a convolutional autoencoder (CAE) was applied to extract deep features from AE signals. Different machine learning methods including discriminant analysis (DA), neural networks (NN), and extreme learning machines (ELM) were used for the construction of classifiers. The analysis is focused on the classification of extracted AE features to classify the source material, to evaluate the predictive importance of extracted features, and to evaluate the ability of used FOS for the evaluation of material behavior under challenging low-temperature environments. The results show the robustness of different CAE configurations for deep feature extraction. The combination of classic and deep features always significantly improves classification accuracy. The best classification accuracy (80.9%) was achieved with a neural network model and generally, more complex nonlinear models (NN, ELM) outperform simple models (DA). In all the considered models, the selected combined features always contain both classic and deep features.
•Various model structures are compared for natural gas consumption forecasting.•Data sets include local distribution company data and individual house data.•Linear (ARX) and nonlinear (neural ...networks, SVM) forecasting models are applied.•Impact of solar radiation in improving the forecasting accuracy is investigated.•Results confirm that solar radiation clearly improves the forecasting accuracy.
Natural gas is known as a clean energy source used for space heating in residential buildings. Residential sector is a major natural gas consumer that usually demands significant amount of total natural gas supplied in distribution systems. Since demands of all consumers should be satisfied and distribution systems have limited capacity, accurate planning and forecasting in high seasons has become critical and important. In this paper, the influence of solar radiation on forecasting residential natural gas consumption was investigated. Solar radiation impact was tested on two data sets, namely on natural gas consumption data of a model house, and on natural gas consumption data of a local distribution company. Various forecasting models with one day ahead forecasting horizon were compared in this study, including linear models (auto-regressive model with exogenous inputs, stepwise regression) and nonlinear models (neural networks, support vector regression). Results confirmed that solar radiation clearly influences natural gas consumption, and included as input variable in the forecasting model improves the forecasting results. Consequently it is recommended to use solar radiation as input variable in building forecasting models.
Specimens of sensitized type AISI 304 stainless steel were subjected to constant load and exposed to an aqueous sodium thiosulphate solution. Intergranular stress-corrosion cracking was monitored ...simultaneously for electrochemical noise, acoustic emission, and specimen elongation. A section of the gauge length was monitored optically with subsequent analysis by digital image correlation. Correlations between the results were observed and analysed. Electrochemical noise and elongation are associated with crack propagation from the early stages, whereas acoustic emission is associated with the final stages of fracture. Digital image correlation analysis is sensitive to crack development, and is used to measure crack length and crack openings.
Characterization of acoustic emission (AE) signals in loaded materials can reveal structural damage and consequently provide early warnings about product failures. Therefore, extraction of the most ...informative features from AE signals is an important part of the characterization process. This study considers the characterization of AE signals obtained from bending experiments for carbon fiber epoxy (CFE) and glass fiber epoxy (GFE) composites. The research is focused on the recognition of material structure (CFE or GFE) based on the analysis of AE signals. We propose the extraction of deep features using a convolutional autoencoder (CAE). The deep features are compared with extracted standard AE features. Then, the different feature sets are analyzed through decision trees and discriminant analysis, combined with feature selection, to estimate the predictive potential of various feature sets. Results show that the application of deep features increases recognition accuracy. By using only standard AE-based features, a classification accuracy of around 80% is obtained, and adding deep features improves the classification accuracy to above 90%. Consequently, the application of deep feature extraction is encouraged for the characterization of loaded CFE composites.
This paper investigates the influence of laser-beam intensity distribution (LBID) on the performance of the annular laser-beam directed energy deposition (DED) process with axial powder delivery. ...Three different LBIDs: Gaussian-like (G-LBID), top-hat-like (TH-LBID) and ring (R-LBID) at two LBID diameters were used. The process performance was characterised qualitatively in terms of the melt-pool shape and the process stability and quantitatively by powder-catchment efficiency, selected geometrical and metallurgical properties of the clad. The observed influence of LBID on process performance, as determined by the relationship between LBID and powder stream density distribution (PSDD), decreased with increasing mean surface-energy density and was more significant at larger LBID diameter. The highest powder-catchment efficiencies (90% and 87%) were achieved with the G-LBID and TH-LBID, whose high-intensity centre is aligned with the peak of the Gaussian-like PSDD. A lower powder-catchment efficiency of 77% was achieved with the R-LBID, whose high-intensity region is located at the edge of the melt pool with minimum powder density. However, this also results in the highest and most uniform dilution, the highest metallurgical bond ratio and the lowest lack of fusion porosity at the clad-substrate interface. In addition, the process was stable at the lower values of mean surface-energy density with R-LBID, while balling instability was observed with G-LBID and TH-LBID. It can be concluded that the use of R-LBID at lower values of mean surface-energy density improves the performance of the DED process with axial powder feed in terms of process stability and metallurgical properties of the clad.
Display omitted
•Influence of laser-beam intensity distribution in metal-powder DED was investigated.•Gaussian-like, top hat, and ring laser-beam intensity distributions were considered.•Effect of intensity distribution was more pronounced at a low surface-energy density.•Gaussian-like distribution provided the highest powder-catchment efficiency.•Ring-intensity distribution improved process stability and clad properties.
Display omitted
•Methodology for 48-hour multi-step heat demand forecasting in a DH system.•Gaussian process regression outperforms considered machine learning methods.•Accurate temperature forecasts ...are important, solar irradiation forecasts are not.•Forecasting errors for 48 h ahead below 3% of the max. heating power.•Proposed forecasting solution can be fitted to different DH systems.
Short-term heat demand forecasting in district heating (DH) systems is essential for a sufficient heat supply and optimal operation of the DH. In this study, a machine learning based multi-step short-term heat demand forecasting approach using the data of the largest Slovenian DH system is considered. The proposed approach involved feature extraction and comparative analysis of different representative machine learning based forecasting models. Nonlinear models performed better than linear models, and the best forecasting results were obtained by Gaussian process regression (GPR), where the mean absolute normalized error was 2.94% of the maximum heating power of the DH system. The analysis confirmed the importance of accurate temperature forecasts but did not confirm the relevance of using future solar irradiation forecasts. The optimal length of training data is shown to be 3 years, and past data of up to 4 days can be used as input to increase the forecasting accuracy. The forecasting model (GPR) proposed in this study can be fitted to different DH systems. In the presented case study, it was selected to implement the online forecasting solution for the DH of Ljubljana and has been generating forecasts with a mean absolute normalized error of 2.70% since November 2019.
The subject of this study is the vertical mass-spring-like oscillation of a pendant droplet and its resonant detachment, which was experimentally observed in the process of laser droplet generation ...from a metal wire. The process was characterized by various time series, which were generated from a sequence of infrared intensity images of the process. Following a visual inspection of pendant droplet images and an analysis of a wavelet based time-frequency map of the droplet’s vertical displacement time series, the pendant droplet’s oscillation is described by a time-variable mass-spring system. Based on the characteristics of the time-frequency map, the resonant nature of the pendant droplet detachment was demonstrated. Additionally, an algebraic expression was formulated, which can be used to predict the detached droplet’s diameter as a function of the laser pulse frequency.
•Oscillating metal pendant droplet detachment was studied using time series analysis.•A t–f map of the pendant droplet vertical displacement time series was calculated.•Based on the t–f map, the pendant droplet was described as a mass-spring system.•The t–f map showed the resonant nature of the pendant droplet detachment.•The detached droplet diameter was predicted by an algebraic formula.
•Semi-supervised approach to condition monitoring of compressors is proposed.•Problem of a large vibration data set without prior class definitions is addressed.•Method combines feature extraction, ...statistical analysis, and classification.•Properly designed nonlinear classifiers are recommended for industrial application.•Extreme learning machines are a useful tool for vibration-based classification.
Semi-supervised vibration-based classification and condition monitoring of the reciprocating compressors installed in refrigeration appliances is proposed in this paper. The method addresses the problem of industrial condition monitoring where prior class definitions are often not available or difficult to obtain from local experts. The proposed method combines feature extraction, principal component analysis, and statistical analysis for the extraction of initial class representatives, and compares the capability of various classification methods, including discriminant analysis (DA), neural networks (NN), support vector machines (SVM), and extreme learning machines (ELM). The use of the method is demonstrated on a case study which was based on industrially acquired vibration measurements of reciprocating compressors during the production of refrigeration appliances. The paper presents a comparative qualitative analysis of the applied classifiers, confirming the good performance of several nonlinear classifiers. If the model parameters are properly selected, then very good classification performance can be obtained from NN trained by Bayesian regularization, SVM and ELM classifiers. The method can be effectively applied for the industrial condition monitoring of compressors.
An annular laser beam based powder cladding head, which enables an axial powder feeding and variation of the laser beam intensity distribution (LBID) on the workpiece surface is presented. The ...influence of typical LBIDs, including Ring, Tophat(−), Tophat(+), and Gaussian-like, on a cladding process has been characterized based on the process and melt pool visualization, powder catchment efficiency, clad layer geometry, and porosity. The results showed that the most stable process without plasma formation but with low dilution and porosity of the clad layer can be achieved within the range from a Ring to a Tophat(−) LBID. Additionally, axial powder feeding results in a high powder catchment efficiency above 80%.