Adaptive time series decompositions can be used as a basis for oscillation detection in control loops at chemical plant. In general, the decomposition extracts a process time series into different intrinsic mode functions (IMFs). Later, IMFs are characterized to detect the oscillation. This paper presents a comparative study of three well-known adaptive time series decomposition, i.e. empirical mode decomposition (EMD), empirical wavelet transform (EWT) and variational mode decomposition (VMD). The study compares significancy, regularity and sparseness of each method with different process time series from chemical plant. The time series represents oscillation due to process disturbance, mistuned controller, sticking valve and faulty sensor. The comparison shows that EWT outperforms in consistent detection results over EMD and VMD.