Farbod Akhavan Niaki

A wavelet-based data-driven modelling for tool wear assessment of difficult to machine materials

In this work, wavelet packet decomposition along with principle component analysis are utilised for feature extraction using two low cost sensing methods: vibration and power sensors, in end-milling of gamma prime-strengthened alloy. The high wear rate of this material induces a rapid transition from a sharp state to a dull state of the tool, and hence limits the number of available data for model establishment. To address this challenge, a neural network with Bayesian regularisation is designed and its performance is compared with regression and time-series models. A maximum of 4% estimation error for Bayesian regularisation neural network, compared to 33% and 17% estimation error of the latter models, shows the good potential of this technique when a limited dataset is available. In addition, the use of low cost measuring sensors in this paper aligned well with the industrial applications to detect and avoid unscheduled downtime in machining situations.

Wavelet based sensor fusion for tool condition monitoring of hard to machine materials

Tool condition monitoring in modern manufacturing systems is gaining more attention due to the fact that excessive tool damage can cause workpiece surface deterioration and increase idle time. Therefore, monitoring tool condition from the initial to final stages of tool life is a task that is critical yet difficult, especially in hard-to-machine materials. In this work, Wavelet Packet Decomposition is used for extracting statistical features in the time-frequency domain of two low cost sensing technologies, i.e. vibration and power, in addition to Principal Component Analysis to reduce the dimensionality of feature vectors. A Recurrent Neural Network is then trained with Bayesian regularization backpropagation method and the estimated tool wear is compared to the actual measured wear. Results show a maximum of 13% relative error in estimating tool wear which proves the effectiveness of implemented sensory data fusion method to be used in automated control of manufacturing processes.

A probabilistic-based study on fused direct and indirect methods for tracking tool flank wear of Rene-108, nickel-based alloy

Nickel-based alloys are a class of hard-to-machine materials that exhibit a unique combination of high strength at high temperature. While they are widely used in industry, the high tool wear rate of these materials makes machining them a challenging task. Furthermore, machining tolerances, residual stresses, and tool runout impose uncertainty in tracking tool wear. In this work, the current view on tracking progressive tool wear is shifted from deterministic domains into the stochastic domain to study the probability distribution of the tool wear during the process. To do so, Bayesian-based estimation was used for accurate model inference and the result was fed into the Kalman filter for tracking tool wear in end-milling Rene-108 Ni-based alloy. To improve the accuracy, a direct laser measuring system for capturing tool length change was fused with the indirect power measurement. The results show a significant improvement in accuracy over the indirect method, with less than 8 µm root mean square error. This measuring strategy improved the accuracy, while preserving the automated platform for monitoring the tool’s health.

Parameter Estimation Using Markov Chain Monte Carlo Method in Mechanistic Modeling of Tool Wear During Milling

Several models have been proposed to describe the relationship between cutting parameters and machining outputs such as cutting forces and tool wear. However, these models usually cannot be generalized, due to the inherent uncertainties that exist in the process. These uncertainties may originate from machining, workpiece material composition, and measurements. A stochastic approach can be utilized to compensate for the lack of certainty in machining, particularly for tool wear evolution. The Markov Chain Monte Carlo (MCMC) method is a powerful tool for addressing uncertainties in machining parameter estimation. The Hybrid Metropolis-Gibbs algorithm has been chosen in this work to estimate the unknown parameters in a mechanistic tool wear model for end milling of difficult-to-machine alloys. The results show a good potential of the Markov Chain Monte Carlo modeling in prediction of parameters in the presence of uncertainties.Copyright

Investigation of the Relationship Between Vibration Data and Tool Wear During End-Milling of Gamma-Prime Strengthened Alloys

Condition Based Maintenance (CBM) systems are crucial for today’s high accuracy machining of exotic materials. For reliable results, CBM systems need early and reliable warning based on prediction models that use multiple types of sensors. In this study, tool flank wear during end milling difficult-to-machine alloys was measured using an optical microscope. Then, vibration data collected with an accelerometer was investigated for its relationship to tool flank wear. The developed relationship between accelerometer output and tool flank wear was validated with further experiments. It was observed from frequency domain responses of these outputs that specific harmonics of the tool pass frequency were dominant, and tool flank wear can be related to the amplitude of these harmonics during machining. This way, it was shown that through accurate online prediction of tool wear, premature interruption of the process as well as machining with a worn tool can both be avoided, improving end-product quality as well as reducing machining costs.© 2015 ASME