Monitoring complex industrial plants is a very important task in order to ensure the management, reliability, safety and maintenance of the desired product quality. Early detection of abnormal events allows actions to prevent more serious consequences, improve the system's performance and reduce manufacturing costs. In this work, a new methodology for fault detection is introduced, based on time series models and statistical process control (MSPC). The proposal explicitly accounts for both dynamic and non-linearity properties of the system. A dynamic feature selection is carried out to interpret the dynamic relations by characterizing the auto-and cross-correlations for every variable. After that, a time-series based model framework is used to obtain and validate the best descriptive model of the plant (either linear o non-linear). Fault detection is based on finding anomalies in the temporal residual signals obtained from the models by univariate and multivariate statistical process control charts. Finally, the performance of the method is validated on two benchmarks, a wastewater treatment plant and the Tennessee Eastman Plant. A comparison with other classical methods clearly demonstrates the over performance and feasibility of the proposed monitoring scheme.
Fuzzy rule-based systems (FRBSs) are a common alternative for applying fuzzy logic in different areas and real-world problems. The schemes and algorithms used to generate these types of systems imply that their performance can be analyzed from different points of view, not only model accuracy. Any model, including fuzzy models, needs to be sufficiently accurate, but other perspectives, such as interpretability, are also possible for the FRBSs. Thus, the Accuracy-Interpretability trade-off arises as a challenge for fuzzy systems, as approaches are currently able to generate FRBSs with different trade-offs. Here, rule Relevance is added to Accuracy and Interpretability for a better trade-off in FRBSs. These three factors are involved in this approach to make a rule selection using a multi-objective evolutionary algorithm. The proposal has been tested and compared with nine datasets, two linguistic and two scatter fuzzy algorithms, four measures of interpretability and two rule relevance formulations. The results have been analyzed for different views of Interpretability, Accuracy and Relevance, and the statistical tests have shown that significant improvements have been achieved. On the other hand, the Relevance-based role of fuzzy rules has been checked, and it has been shown that low Relevance rules have a relevant role for trade-off, $ This work has been partially supported by the Spanish Ministry of Economy and Competitiveness through the Project no. DPI2015-67341-C2-2-R
The construction of tunnels has serious geomechanical uncertainties involving matters of both safety and budget. Nowadays, modern machinery gathers very useful information about the drilling process: the so-called Monitor While Drilling (MWD) data. So, one challenge is to provide support for the tunnel construction based on this on-site data .Here, an MWD based methodology to support tunnel construction is introduced: a Rock Mass Rating (RMR) estimation is provided by an MWD rocky based characterization of the excavation front and expert knowledge [1].Well-known machine learning (ML) and computational intelligence (CI) techniques are used. In addition, a collectible and "interpretable" base of knowledge is obtained, linking MWD characterized excavation fronts and RMR.The results from a real tunnel case show a good and serviceable performance: the accuracy of the RMR estimations is high, Errortest ∼ = 3%, using a generated knowledge base of 15 fuzzy rules, 3 linguistic variables and 3 linguistic terms.This proposal is, however, is open to new algorithms to reinforce its performance.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.