Date of Award


Document Type


Degree Name

Doctor of Philosophy (PhD)


Computer Engineering and Sciences

First Advisor

Marco Carvalho

Second Advisor

Alexandre Lucas Stephane

Third Advisor

Patrick Millot

Fourth Advisor

William Allen


Analyzing potential causes for recent industrial accidents and production losses highlights that the majority of these incidents are due to human factor failure: operator error, poor decision-making and violations (Dźwiarek, 2004). In the meantime, the process equipment design, fabrication, metallurgy and control system has improved significantly to combat these unpredictable failures. This research focus on the errors due to human-machine interaction (HMI) in chemical plant. This paper presents an advanced support system for operators in chemical plants, to perform safe and reliable plant operation and maintenance (O&M). In chemical industry, almost every tasks involving human interaction with the machines are automated using standard operating procedures (SOP) and checklists to reduce failures. Paper-based system are currently used by the chemical industry to assist operators for safe operation of plants. Management of interconnections among paper-based operational documents can be difficult for workers. There are some cases, where operators are unaware of the existing procedures for certain situations exist. Human error is inevitable in this complex environments, but it can be improvised with the advancement in science and technology. An Onboard context-sensitive information system (OCSIS) was developed for pilots to provide appropriate operational content at the right time to reduce human-factor failures (Tan & Boy G.A, 2016). By adopting this method, and customizing it with suitable technology, the system can assist operators with the right information for safe and effective plant Operation and Maintenance. Inappropriate design of these life-critical systems can lead to horrible disasters; therefore, it is very important to design the system using the human-centered approach. CAOSS is a mobile/tablet based application integrated with beacons to provide context aware field readings, procedures, and checklists for effective field operation. The chat feature in the CAOSS facilitates better employee communication. The proposed system doesn’t change any of the existing methods based on process safety management (PSM) in the plant operation. It upgrades the existing paper-based method with context based digital mobile system, which aims to reduce human factor failures. The system also considers the safety critical tasks (SCTs), inspection of elevated structures, such as chimneys, and flaring stacks, cooling towers for any defects and long distance pipeline for any leak detection. Workers have to climb up hundreds-of-meters high towers using scaffolding for the inspections, these tasks are considered high risk. Moreover, some plant operating systems are forced to shut down while these critical tasks are being performed, and increasing the potential for an accident to occur within the plant. CAOSS provides solution through semi-autonomous drones equipped with advanced sensors and payloads. CAOSS was tested in terms of usability with chemical plant experts. The result shows that the CAOSS system is practical and useful for safe chemical plant operation. Specifically, the context-based information is more effective during time-sensitive emergency situations. Furthermore, they believed that the drones can reduce human physical task and operational down-time significantly by replacing human operators in high-risk preliminary inspections. To check the assessment of the CAOSS, a scenario-based test based on the Bhopal disaster was performed with students from FIT. The preliminary results, proved that it reduces operational time and cognitive load through better situational awareness. We need to conduct further tests and discover emerging properties, and their impacts, that the introduction of this system might generate. The CAOSS system is believed to reduce industrial catastrophes due to human errors.