Assessment and improvement of the accuracy of tacholess instantaneous speed estimation

Abstract

Accurately estimating instantaneous angular speed (IAS) is crucial for effective fault diagnosis under nonstationary operating conditions. Although numerous methods have been developed to improve the accuracy of IAS estimation, finding a suitable technique (method and vibration component) for a specific application remains challenging. This is because accurately assessing the accuracy of the estimated IAS without a tachometer is not straightforward. Thus, since the accuracy of the estimated IAS is unknown, its accuracy cannot be further improved.To address the challenge of accurately estimating and further improving the accuracy of IAS without a tachometer, this study proposes a method to enhance the accuracy of IAS obtained from different techniques (methods and vibration components). The study first investigates the influence of IAS error on order analysis. Next, it discusses three indicators that can assess the accuracy of the estimated IAS without a tachometer. These indicators are designed to compare the energy of the deterministic and random parts of the signal, assess the sparsity of the order spectrum, and quantify the smearing of the vibration component of interest on the order spectrum. The signal and all IASs are divided into windowed sections to determine the most accurate IAS in each section using the three indicators separately or a resultant indicator obtained by merging them using the entropy weighting method. The most accurate IAS are then merged together to obtain the resultant IAS.The study finds that the indicators can effectively evaluate the accuracy of the IAS without a tachometer. The proposed method also produces a more accurate IAS than the IAS candidates obtained by different techniques.