Master of Science (MS)


Construction Management

Document Type



Pipelines transport millions of barrels of petroleum products every day. Oil and gas pipelines have become important assets of the economic development of almost any country. Government regulations or internal policies regulate the safety of the assets for the population and environment where these pipelines run. Various strategies and technologies have been introduced for monitoring pipelines, but the most common technology to protect pipelines from occasional hazardous incidents is Computational Pipeline Monitoring (CPM). This technique collects and gathers information from the field related to pressures, flows, and temperatures to estimate the hydraulic behavior of the product being transported. Using the gathered information CPM systems compare its values with standard values and provides a notification if any anomaly or unexpected situation occurs. The result is an alarm to an operator in a supervisory control room. According to Hollifield, it is becoming an increasing problem that there is no standard for plant operators yet, whereas improved design can lead to better performance (Hollifield et. al., 2007). So, the objective of this experiment was to explore the effect of different alarm interfaces on controller response at different alarm rates. A simulated liquid pipeline system was developed and a between subject experimental design was performed to evaluate three different types of alarm window interfaces (Categorical, Chronological, and Revised Categorical), two alarm rates (10 in 10 minutes and 20 in 10 minutes), and three levels of alarms (high, medium, and low). Thirty one participants participated in this research, and the performance of participants was measured in terms of acknowledgement time, response time and the accuracy of response. Results showed that the participants’ performance in terms of response time, acknowledgement time, and accuracy of response was significantly different between chronological, categorical, and revised categorical displays. Data analysis showed that the means were shorter in revised categorical display in terms of response time, acknowledgement time, and accuracy of response. This study will be useful in developing new standards for alarm display.



Document Availability at the Time of Submission

Release the entire work immediately for access worldwide.

Committee Chair

Harvey, M. Craig