Babji Srinivasan

Fouling degrades the overall efficiency of the heat exchanger networks (HENs), which results in a significant economic loss. The mitigation of fouling in an operational HEN is carried out by optimizing the cleaning schedules of the heat exchangers. Although such approach can save costs, it is subjected to the exact implementation of the optimal cleaning schedule. Usually, the small and medium-scale process industries face difficulties in implementing such solutions due to limited resources, which forces them to rely on suboptimal cleaning schedules, such as postponing or avoiding few cleaning tasks. This work addresses this gap by optimizing the cleaning schedule considering the maintenance resource limitation. Our approach considers a mixed-integer linear programming (MILP) based optimization considering groupings of heat exchangers based on their spatial locations for ease of maintenance The proposed formulation is applied on a HEN with linear and asymptotic fouling, with and without cleaning cost. The results show that the approach can prevent a considerable economic loss, which would incur due to suboptimal cleaning schedules due to resource limitations.

Safety in process plants is of paramount importance. With the predominant contributor to accidents in process industries being repeatedly identified as human error, it is necessary to have skilled operators to prevent accidents and minimise the impact of abnormal situations. Such knowledge and skills are imparted to operators using operator training simulator (OTS) which offer a simulated environment of the real process. However, these techniques emphasize on assessing operator's ability to follow standard guidelines – assessment of the operator's process knowledge and imparting an adequate mental model to the operator is not addressed. Further, understanding cognitive behavior of operators, identified as crucial to enhancing their skills and abilities is often neglected. In this work, we develop a systems engineering framework to operator training with emphasis on accounting for the cognitive abilities of the human-in-the-loop. The framework consists of three distinct components: (1) Design of suitable training tasks, (2) Measure and analyse operator's cognitive response while performing the tasks, and (3) Infer operator's mental model through knowledge and capability assessment. Consider the operator as a system whose input is information acquired from the process through the human machine interface (HMI) and output are actions taken on the process (such as manipulating valves). We demonstrate in this paper that the available input (from eye tracking) and output (operator actions) data when suitably analysed with respect to the process state can aid in inferring the operator's mental model at any given time. Based on the model, the operator's current knowledge can be deduced and gaps identified. New training tasks can then be designed to address these gaps. In this article, we describe the proposed framework for operator learning and illustrate it using experimental studies.

In this contribution we propose an active Fault Tolerant Control (FTC) strategy which enables the isolation and identification of valve stiction and valve blocking, in addition to the additive faults like sensor and actuator biases. The developed method is an extension of the original method proposed by Prakash et al. (2002). This method is based on the Kalman filter and is developed under the assumption that the monitored system is Linear Time Invariant (LTI). It has been shown to work well for additive faults such as sensor and actuator biases. Within this method the fault isolation and identification task is based on the Generalized Likelihood Ratio (GLR) test by which the most plausible fault type in a library of faults is selected following estimation of fault parameters. © 2010 Elsevier B.V.

Process industries are highly hazardous, and these hazards often lead to accidents. Over 70% of these accidents are attributed to human errors. With the advancements in technology and changing role of operators to the one involving an emphasis on cognitive aspects, most of these errors occur due to limitations in cognitive performance. One of the major constructs to understand cognitive performance is the cognitive workload. An increase in cognitive workload often leads to degradation in performance. Eye tracking has been used in several domains to assess cognitive workload. In this work, we propose a methodology to assess cognitive workload of control room operators during tasks that involve tackling process abnormalities. The methodology employs the fusion of metrics obtained from pupil and gaze data. Our results reveal that fusion of metrics provides better accuracies of classifying cognitive workload at three levels—low, medium and high workload.

Human operator performance is a key factor for proper operation of safety critical industries like process control, nuclear power plant and aviation. Several incident reports identify human error as one of the causes responsible for accidents. One way to reduce human error is to impart proper operator training which requires understanding how operators learn during training. Traditionally, learning is evaluated on the basis of subjective assessments and outcomes of the task execution. It obscures the cognitive aspect of the learning process like understanding how operator gives attention to various information sources and develops the process of decision making during the training program. In this study, we use pupil size measurement from eye tracking system to study and understand learning as participants undergo repetitive trials of simulated flight operations tasks. A derived measure based on time-frequency analysis of pupil size variation is observed to be sensitive to the process of learning during task repetition. We observe improvement in the task performance accompanied by the decrease in the proposed measure.

Human errors are a root cause of majority of accidents occurring in the process industry. These errors are often a result of excessive workload on operators, especially during abnormal situations. Understanding and measurement of cognitive workload (overload), experienced by human operators while performing key safety critical tasks, is thus important to the understanding of human errors. Subjective measurements of workload are often not reliable and there is a need for physiological based parameters of workload. In this work, we propose a methodology to measure cognitive load of a control room operator in terms of a biomarker, specifically theta/alpha ratio, obtained from a single electrode EEG signal. Real-time detection of the biomarker can enable minimize errors and improve safety.

A combined cycle power plant (CCPP) is a complex system with a Gas Turbine, Steam Turbine and a Heat Recovery Steam Generator (HRSG) working together. These three units work together and make the process highly interdependent. The onset of any fault in one of the above units would results in a significant reduction in overall efficiency and potentially lead to catastrophic accidents. Such failures can occur due to process faults because of large abrupt variations of operating conditions and structural faults due to corrosion, uneven stresses due to frequent cyclic operations. Conventionally, the identification of such leakage locations is made via visual inspection which is a time consuming and tedious. In the present work, we discuss a fault diagnosis strategy for an actual industrial HRSG present in a CCPP. Various steady state models at different loads of CCPP as well as a dynamic model are developed. Various structural faults in the form of leakages are incorporated in the heat exchangers. An Artificial Neural Network (ANN) model is developed based on data from the above simulations to detect the leaking heat exchangers.

The significance of operator training has dramatically increased due to complex automation strategies in modern process plants. It is reported that human errors account for 70% of the accidents in process industries, with inadequate training cited as one of the most common reasons for these incidents. Our previous work has shown the potential of eye-tracking to infer the mental state of control room operators. In this work, we propose a methodology that combines multi-scale data from the process simulator, control actions performed, and eye gaze data of the operators to evaluate their training outcomes. Specifically, we use fixation transition entropy, an eye-tracking metric, which can help infer the mental models of the process abnormalities developed by the operators during repeated control room tasks. Results indicate that the fixation transition entropy decreases on account of development of correct mental models of process while it remain at higher values when operator fails to update their mental models during plant abnormalities. Thus, the proposed metric can be used to gauge the development of operator's mental models during training to understand the transition from novice to becoming experts.

A combined cycle gas turbine (CCGT) is a complex interconnected system that uses gas turbine's hot exhaust to power the steam power plant for achieving higher thermal efficiency. Since performance of CCGT power plants is governed by huge number of parameter linked to various components (such as compressor, combustion unit and turbine), alarms are configured to inform system operators of the abnormal operating conditions (Wong et al., 2013). Alarm systems also play a vital role in ensuring safe operation of power plants. However poor management of alarm systems result in alarm flooding, nuisance alarms, chattering alarms and false alarms that can distract the operator from determining the true state of the process. Studies indicate poor alarm system to be one of the major causes for incidents (such as BP Texas refinery and Texaco's Oil Refinery, Milford Haven) in process industries (Shu et al., 2016). Therefore, alarm management has received much attention over the past with guidelines available for alarm rationalization, prioritization and elimination of nuisance alarms. In this work, we studied the alarm data collected from Dhuvaran CCGT power plant in Gujarat, India. Event log obtained for 6 days from the plant indicates that 12186 alarms are flagged per day (compared to 144 alarms as per EEMUA-191 guidelines) with approximately 8 alarms every minute (EEMUA-191,2013). We implemented univariate (such as chattering index) and multivariate data analytic tools (such as correspondence analysis) on this alarm data obtained from combined cycle gas turbine power plant to perform alarm rationalization. We removed chattering alarms (by varying threshold based on guidelines in the literature), reduced the replicated alarms and grouped multiple alarms based on correspondence analysis to introduce unit level alarms. After performing these tasks, the average alarm count reduced to 698 per day, a reduction of 94.3% alarms. We will discuss the details of the proposed alarm management system and demonstrate the results obtained from its applicability to CCGT power plant.

Operators’ knowledge during abnormal situations that are faced in chemical process industries is critical to ensure safety. Operators expand their knowledge base through training programmes that assess their comprehension and skills using simple success and failure criteria, process-based measures, and operator actions. However, these assessment techniques often overlook factors relevant to the evaluation of their cognitive capabilities such as information acquisition pattern, cognitive workload and decision-making strategy. In this work, we present a methodology for evaluating operators’ performance during training that blends process-based measurements with eye-tracking-derived cognitive behaviour. Our methodology is based on Self- Organizing Map (SOM), an unsupervised neural network that allows optimum visualization of complex data. Accordingly, we trained two different SOM networks, one using the process data and the other using eye-tracking data to obtain information about operators’ performance during training. Results indicate that when operators learn the process dynamics, the number of neuronal clusters hit by the process as well as operators’ eye gaze trajectory decrease. The decrease in the number of clusters on SOM trained using process data indicates improved operator performance in terms of successful completion of the task and correct control action with appropriate magnitude. On the other hand, the decrease in the number of clusters hit on SOM trained using eye gaze data signifies that the operator attends to only a few regions on HMI that are critical to the current disturbance/abnormality in the process. Thus, the proposed methodology can be used to gauge the operators’ learning progress during training to understand the transition from novice to expert.