New investigation on diagnosing steam production systems from multivariate time series applied to thermal power plants

► We model a new process for multivariate detection of steady states in power plants. ► This modular computing structure ensures flexible adaptation to diagnosis techniques. ► We introduced an automatic procedure for sequencing storage of pre-processed data. ► We grouped several disciplines in engineering for performing operational diagnosis. ► Outcomes can be applied to early warning systems according to specific requirements.