Enhanced Monitoring of Environmental and Chemical Processes
Grant
Overview
Affiliation
View All
Overview
abstract
Fault detection (FD) is essential for safe and reliable operation of various environmental and chemical processes. For example, online monitoring of ozone and other air pollutants is crucial for the safety of humans and the environment. Also, detecting anomalies and malfunctioning units or sensors in chemical processes is vital for their proper and safe operation. Various process monitoring or FD techniques have been developed and applied in practice. When a process model is available, statistical hypothesis testing methods, such as the generalized likelihood ratio (GLR) test, have shown enhanced FD abilities over conventional methods. Unfortunately, process models that can accurately predict the process behavior are not always available. Also, measured process data are usually contaminated with errors that mask the important features in the data and limit their usefulness in practice. The main objective of this project is to develop enhanced FD techniques that can better deal with these practical challenges and help improve monitoring various environmental and chemical processes. Specifically, the following objectives will be sought. The first objective of this project is to exploit the advantages of hypothesis testing in the absence of process models by developing linear and nonlinear latent variable based hypothesis testing FD methods. The developed methods will integrate the advantages of hypothesis testing with those of the conventional methods to achieve enhanced performance. In fact, they will have the ability to simultaneously detect and isolate faults. Another objective of this project is to extend the developed hypothesis testing FD techniques to account for uncertainty (or measurement errors) in the data. Two main approaches will be followed to achieve this objective, one using multiscale representation of data (which is a powerful signal processing and feature extraction tool), and the other using interval PCA, which quantifies uncertainty in the data using intervals within which the data may fall. These approaches are expected to enhance the reliability of FD by decreasing the rate of false alarms. A third objective of the project is to apply the developed FD techniques to enhance monitoring various environmental and chemical processes. For example, they will be utilized to improve air quality control through monitoring the concentration levels of ozone and other air pollutants. Also, they will be used to improve monitoring the operation of chemical processes from the local Qatari industry, such as OryxGTL. The proposed project aligns well with Qatar National Vision 2030 and Qatar National Research Strategy (QNRS) as it addresses (and provides solutions) to a problem of vital importance to Qatarâ s goal of achieving a cleaner environment and a safer workplace. The project will also help accumulate a human capital in Qatar who will be able to provide support to various sectors during and after the completion of the project.
