Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Eine Plattform für die Wissenschaft: Bauingenieurwesen, Architektur und Urbanistik
Prediction of EDMed micro-hole quality characteristics using hybrid bio-inspired machine learning-based predictive approaches
https://orcid.org/http://orcid.org/0000-0002-1102-3801
Micro-meter range material removal is an indispensable requirement in today’s industrial scenario for precision components manufacturing in the medical, optical, automobile, and aerospace industries. The dimensional accuracy and material removal rate are highly correlated and important drilled micro-hole characteristics in the micro-electric discharge machining (µEDM) process. In this investigation, predictive models based on bio-inspired intelligent hybrid machine learning algorithms are proposed for machining micro-holes with accurate dimensions on Inconel 718 superalloy by µEDM process. The recast layer thickness and radial overcut are considered as the produced hole dimensional quality features and the material removal rate as the production rate indicator. The proposed predictive models are based on the integration of adaptive neuro-fuzzy inference system (ANFIS) with genetic algorithm (ANFIS-GA) and particle swarm optimization algorithm (ANFIS-PSO). Here, the precision of the ANFIS model was improved by optimizing its algorithmic parameters with the application of GA and PSO individually. Experimentally measured machining response data was used for training, testing, and validation of the models. The performance of the proposed predictive models was assessed based on the statistical parameters. The predicted training and testing values for the responses with regression, ANN, ANFIS, ANFIS-GA, and ANFIS-PSO models were in good agreement with their corresponding experimentally measured values. The estimated values of statistical parameters representing the µEDM responses suggest that the ANFIS-PSO model predicts more accurately as compared with other models. A comparative analysis of the predictability of the proposed models with respect to the experimental results is also presented for confirming the reliability of the ANFIS-PSO model for accurate perdition of the µEDM process.
Prediction of EDMed micro-hole quality characteristics using hybrid bio-inspired machine learning-based predictive approaches
https://orcid.org/http://orcid.org/0000-0002-1102-3801
Micro-meter range material removal is an indispensable requirement in today’s industrial scenario for precision components manufacturing in the medical, optical, automobile, and aerospace industries. The dimensional accuracy and material removal rate are highly correlated and important drilled micro-hole characteristics in the micro-electric discharge machining (µEDM) process. In this investigation, predictive models based on bio-inspired intelligent hybrid machine learning algorithms are proposed for machining micro-holes with accurate dimensions on Inconel 718 superalloy by µEDM process. The recast layer thickness and radial overcut are considered as the produced hole dimensional quality features and the material removal rate as the production rate indicator. The proposed predictive models are based on the integration of adaptive neuro-fuzzy inference system (ANFIS) with genetic algorithm (ANFIS-GA) and particle swarm optimization algorithm (ANFIS-PSO). Here, the precision of the ANFIS model was improved by optimizing its algorithmic parameters with the application of GA and PSO individually. Experimentally measured machining response data was used for training, testing, and validation of the models. The performance of the proposed predictive models was assessed based on the statistical parameters. The predicted training and testing values for the responses with regression, ANN, ANFIS, ANFIS-GA, and ANFIS-PSO models were in good agreement with their corresponding experimentally measured values. The estimated values of statistical parameters representing the µEDM responses suggest that the ANFIS-PSO model predicts more accurately as compared with other models. A comparative analysis of the predictability of the proposed models with respect to the experimental results is also presented for confirming the reliability of the ANFIS-PSO model for accurate perdition of the µEDM process.
Prediction of EDMed micro-hole quality characteristics using hybrid bio-inspired machine learning-based predictive approaches
https://orcid.org/http://orcid.org/0000-0002-1102-3801
Micro-meter range material removal is an indispensable requirement in today’s industrial scenario for precision components manufacturing in the medical, optical, automobile, and aerospace industries. The dimensional accuracy and material removal rate are highly correlated and important drilled micro-hole characteristics in the micro-electric discharge machining (µEDM) process. In this investigation, predictive models based on bio-inspired intelligent hybrid machine learning algorithms are proposed for machining micro-holes with accurate dimensions on Inconel 718 superalloy by µEDM process. The recast layer thickness and radial overcut are considered as the produced hole dimensional quality features and the material removal rate as the production rate indicator. The proposed predictive models are based on the integration of adaptive neuro-fuzzy inference system (ANFIS) with genetic algorithm (ANFIS-GA) and particle swarm optimization algorithm (ANFIS-PSO). Here, the precision of the ANFIS model was improved by optimizing its algorithmic parameters with the application of GA and PSO individually. Experimentally measured machining response data was used for training, testing, and validation of the models. The performance of the proposed predictive models was assessed based on the statistical parameters. The predicted training and testing values for the responses with regression, ANN, ANFIS, ANFIS-GA, and ANFIS-PSO models were in good agreement with their corresponding experimentally measured values. The estimated values of statistical parameters representing the µEDM responses suggest that the ANFIS-PSO model predicts more accurately as compared with other models. A comparative analysis of the predictability of the proposed models with respect to the experimental results is also presented for confirming the reliability of the ANFIS-PSO model for accurate perdition of the µEDM process.
Prediction of EDMed micro-hole quality characteristics using hybrid bio-inspired machine learning-based predictive approaches
Int J Interact Des Manuf
Rao, Thella Babu (Autor:in)
01.04.2023
18 pages
Aufsatz (Zeitschrift)
Elektronische Ressource
Englisch
µ-EDM , Recast layer thickness , Radial overcut , Material removal rate , ANN , ANFIS , ANFIS-GA , ANFIS-PSO Engineering , Engineering, general , Engineering Design , Mechanical Engineering , Computer-Aided Engineering (CAD, CAE) and Design , Electronics and Microelectronics, Instrumentation , Industrial Design
Improvements on surface quality of EDMED surfaces
| British Library Online Contents | 2003
Surface characteristics analysis of dry EDMed AISI D2 steel using modified tool design
| British Library Online Contents | 2015
Microstructure and tool electrode erosion in EDMed of TiN/Si3N4 composites
| British Library Online Contents | 2003