Hybridization of harmonic search algorithm in training radial basis function with dynamic decay adjustment for condition monitoring

In recent decades, hybridization of superior attributes few algorithms was proposed to aid in covering more areas complex application as well improves performance. Condition monitoring is a major component predictive maintenance which monitors the condition and identifies significant changes machinery parameter perform early detection prevent equipment defects that could cause unplanned downtime or incur unnecessary expenditures. An effective model helpful reduce frequency unexpected breakdown incidents thus, facilitates maintenance. ANN has shown various fault applications. popular due its capability identifying nonlinear relationships among features large dataset hence, it can with an accurate prediction. However, drawback performance sensitive parameters (i.e., number hidden neurons initial values connection weights) architecture where settings these are subject tuning on trial-and-error basis. Hence, wide range studies have been focused determining optimal weight models neurons. this research work, motivation develop autonomous learning based adaptive metaheuristic algorithm for optimizing so network adaptation flexible way handle classification tasks accurately industries, such power systems. This paper presents intelligent system integrating Radial Basis Function Network Dynamic Decay Adjustment (RBFN-DDA) Harmony Search (HS) industrial processes. RBFN-DDA performs incremental wherein structure expands by adding new units include information. As such, training reach stability shorter time compared gradient-descent methods. To achieve performance, HS optimize center width each unit trained RBFN. By algorithm, neural (RBFN-DDA-HS) improve performances from original 2.2% up 22.5% two benchmarks datasets, numerical records bearing steel plate condition-monitoring plant circulating water (CW) system). The results also show RBFN-DDA-HS compatible, if not better than, other state-of-the-art machine