A Study of Wear Rate Estimation of Casting Parts by Support Vector Machine

Abrasive wear is one of the prime and mostly costly causes of secondary failure in the design and operation of mechanical machines and equipment. Wear limits must be known and measured to assure quality and durability of products by an experimental process to determine an optimum material composition. The development of wear monitoring systems for industrial processes is well recognised in the industry due to the continued demand for improved product quality and productivity. In this paper, a novel method of wear rate identification, based on a Support Vector Machine (SVM) is proposed. SVM is used to relate the wear rate and technological parameters of the wear resistant drip moulding. The SVM model for determining the wear rate of white iron casting with a low chromium content, was trained and tested by using the existing exploitation data from the Bor Flotation Plant, Serbia. The simulated results of wear prediction show that the accuracy rate of the SVM is 97%.