Discharge performance of plan view of multi-cycle W-form and circular arc labyrinth weir using machine learning

Abstract

Weirs serve the dual purpose of flow measurement and flow regulation in open channels. In the past, weirs of improved plan form have been used to increase the discharge through the weir and to regulate the afflux. This study presents the results of the experiments that were carried out to investigate the discharge capacity of a multi-cycle W-form labyrinth weir and a sharp-crested circular arc weir. The experiments were performed in a rectangular flume under free-flow conditions. For the labyrinth weir, observations were recorded corresponding to five cycles while varying the weir height from 5 cm to 30 cm in steps of 5 cm. For the circular arc crested weir, the arc length was changed by considering five different values of the apex angle, and the crest height was varied from 5 cm to 30 cm in steps of 5 cm. The data obtained from the experiments were used for building predictive models using Multiple Linear regression (MLR), Support Vector Machine (SVM) and Artificial Neural network (ANN). In SVM, a regression model was constructed while in the ANN, feed-forward back propagation neural network with multiple perceptron layers was used to predict the discharge. Metrics like average percentage error (APE), Coefficient of Determination (R2) and Coefficient of efficiency (CE) were used for validating the models. From the present study, it was observed that the SVM regression model performed better than the rest of the models.