Development of an Artificial Neural Network Model to Minimize Power Consumption in the Milling of Heat-Treated and Untreated Wood

Abstract

Aim of study: The power consumption of machining operations is an important part of the total production cost. Therefore, in this study, an artificial neural network (ANN) model was developed to model the effects of treatment, rotation speed, cutting depth, and feed rate on power consumption in the wood milling process. Material and methods: A multilayer feed-forward ANN was employed for the prediction of power consumption. The accuracy of the model was assessed by performance indicators such as MAPE, RMSE, and R². Main results: It has been observed that the ANN model yielded very satisfactory results with acceptable deviations. The MAPE, RMSE, and R 2 values were obtained as 7.533, 0.027, and 0.9737 %, respectively, in the testing phase. Furthermore, it was found that power consumption decreased with decreasing of feed rate and cutting depth. Research highlights: The findings of this study can be used effectively in the forest industry to reduce the experimental time and costs.

Optimization of wood machining parameters using artificial neural network in CNC router

“…Prommul et al [24] stated that ANN was as a satisfactory approach to predict SR. Besides, it was indicated that the ANN approach was an acceptable tool to predict power consumption in the wood milling process [17,22]. In this respect, it can be said that even similar better results were obtained with the literature in terms of the PEC.…”

“…The training of neural network stopped when there was no improvement in accuracy or loss on the validation set over a certain number of epochs. The mean square error (MSE) was used in the evaluation of the performance of ANN models, and it is given in Equation (2) [17]. …”

“…They observed that power requirements were increased with wood density and cutting depth. Ozsahin and Singer [17] developed an ANN model to minimise power consumption in milling of beech wood. They found that power consumption decreased as the FR, SS, and DoC decreased.…”

“…ANNs are inspired by the neural networks of the human brain and are designed to process data [14][15][16]. It is stated that ANNs are suitable for a wide variety of engineering applications such as optimisation prediction, data processing, classification, pattern recognition, and tool condition monitoring [3,[17][18][19][20].…”

“…As a result of model outputs, optimum SR values were obtained in conditions of high SS and low FR using 2 and 4 mm diameters of cutting tools. Ozsahin and Singer [17] used ANN to optimise CP in the milling of beech wood. As a result of the experimental study, they stated that CP consumption increased as the feed speed, SS, and DoC increased.…”