Graduate Student Research Seminar Day ‑ Jun 30, 2021

You are cordially invited to the Graduate Student Research Seminar of the Department of Industrial Engineering

Date: Wednesday, June 30, 2021
Time: 1:00 - 4:00 PM
Venue: Online Event

 

Schedule:

1:00-1:30 PM Tessa Bulmer
Analysis and modeling of the thrombolysis process for acute ischemic stroke patients at urban and rural hospitals

1:30-2:00 PM Ryan O'Neil
Optimal joint selective maintenance and orienteering: A case study in offshore wind energy

2:00-2:30 PM Lauryne Rodrigues
An analysis technique of ship availability for assessing the capacity of the maritime network after an earthquake: A case of study in Vancouver

2:30-3:00 PM Pardis Pourmohammadi
A robust simulation-optimization approach for designing hybrid renewable

3:00-3:30 PM Reyhaneh Sadeghi
Validation in system safety hazard analysis in safety-critical industries

3:30-4:00 PM Talah Al Sharkawi
A Bayesian Network Sub-model for risk-analysis system on oil spill response effectiveness in the Canadian Arctic

Abstracts:

Analysis and modeling of the thrombolysis process for acute ischemic stroke patients at urban and rural hospitals
Tessa Bulmer, MASc. Student

Background: Thrombolysis is the process of treating acute ischemic stroke (AIS) patients with tissue plasminogen activator (tPA), where effectiveness critically relies on rapid treatment. Fast treatment with tPA has been reported in many urban hospitals, but rural hospitals struggle to reduce treatment times. The study objectives are as follows: (1) to analyze healthcare professionals’ views on various treatment topics in Nova Scotia; (2) to map and compare the thrombolysis treatment process in urban and rural settings in Nova Scotia; (3) to provide a detailed conceptual framework of the thrombolysis process focusing on intra-hospital activities; and (4) to assess the potential impact of process improvements that can result in faster door-to-needle time (DNT) when applied to urban and rural settings.

Methods: Structured interviews were conducted with healthcare professionals involved in stroke treatment across three Nova Scotian hospitals (1 urban and 2 rural). Interview data were used to develop a detailed process map for each site, which provided the foundation to then create a conceptual framework. The interview results and conceptual framework were used to develop an ARENA discrete-event simulation (DES) model to replicate treatment processes at both urban and rural hospitals.

Results: There were 23 health care professionals interviewed at 3 sites. The analysis of the interview data showed a total of 11 urban-rural treatment differences. Additionally, 11 patientrelated and 29 system treatment delays were found. Five scenarios were tested with the DES model, using median DNT as the primary outcome measure. The scenario results include the following maximum DNT reductions: patients arriving via Emergency Medical Services (EMS) remaining on the EMS stretcher to imaging (10.3%), administration of tPA in the imaging area (12.7%), preregistering patients arriving via EMS (2.7%), reducing both the treatment decision time and tPA preparation time by 35% (11.9%), and combining all scenarios (28.3%).

Conclusions: The majority of treatment delays encountered are system delays. There is a general consensus that there is an urban-rural treatment gap, and physicians in rural settings are more hesitant to treat with tPA. The detailed conceptual framework further clarifies intra-hospital logistics of the thrombolysis process. Significant DNT improvements are achievable in urban and rural settings through implementing process changes and reducing activity durations

Optimal joint selective maintenance and orienteering: A case study in offshore wind energy
Ryan O’Neil, MASc. Student

The quest for sustainable energy production is fueling the growth of offshore wind electricity generation. Energy producing offshore wind turbines are typically dispersed across several remote wind farms and must be maintained and operated with high reliability levels for long time periods separated by scheduled maintenance rotations. Due to resource constraints such as travel time, cost, and availability of repair crews, only a subset of turbines and their components can be selected for maintenance operations during maintenance trips. A novel joint selective maintenance and orienteering framework is presented to address the selection of turbines to visit, the components to maintain, the maintenance levels to be performed, and the assignment and routing of repair crews. The goal of the proposed model is to minimize total cost while satisfying a minimum required reliability threshold during the next operating mission until the following maintenance rotation. Several numerical experiments demonstrate the validity of the proposed model and the benefit of jointly optimizing selective maintenance and orienteering decisions.

An analysis technique of ship availability for assessing the capacity of the maritime network after an earthquake: A case of study in Vancouver
Lauryne Rodrigues, MASc. Student

Maritime logistics play a vital role in support emergency relief logistics for communities that are dependent on this transportation mode. By its geographical nature, this concerns specifically island populations. The viability to perform post-disaster operations depends on the availability of infrastructure elements, such as ports, waterways navigation support, and ship availability. During an earthquake event, especially when followed by a tsunami, there is a substantial risk of damage to vessels operating in coastal areas. This research investigates this risk in a Cascadia type earthquake event in British Columbia. In particular, a model is proposed to estimate the probability of ships being available to support the humanitarian supply chain operations in the disaster response phase. The study uses spatial analysis tools with vessel movement data from the Automatic Information System. First, their origin and destination ports are determined, and routes and trajectories patterns are extracted from the data. Then, the model investigates the risk of damage to ferries and tugs on points along a specific path. The developed model considers various spatial and attribute components, such as the distance from collapsing structures, tsunami zone, safe depth areas, tsunami arrival time, and other nautical features. The results indicate that many small ferries and some tugs have a substantial probability of being unavailable to support emergency logistics, whereas larger ferries are less affected. Various results, such as the probability of certain parts of the fleet being unavailable, maps of dangerous navigational areas, and routes with reduced transportation capacity, can be used as a resource to support disaster preparedness and mitigation actions. Despite some uncertainties related to exact ship location, tsunami data, and some model simplifications, the findings can thus be used to inform regional emergency preparedness decision-making.

A robust simulation-optimization approach for designing hybrid renewable energy systems
Pardis Pourmohammadi, MASc. Student

Stand-alone hybrid renewable energy systems (HRES) provide a viable alternative to satisfy energy demand in remote and isolated communities. We consider a PV/Wind/Diesel/Battery HRES and propose a cost-minimization design approach that uses a finite number of supply and demand scenarios with uncertain probabilities, extracted from limited data through kmeans clustering. Using an ambiguity set based on phi-divergence, a novel robust simulationoptimization approach that estimates a surrogate objective function through Response Surface Methodology (RSM) is proposed. Results obtained from implementing the proposed approach on a real case study show that it outperforms classical risk-neutral methods on external data samples.

Validation in system safety hazard analysis in safety-critical industries
Reyhaneh Sadeghi, PhD Candidate

Validation is a critical issue, which has not received much explicit attention in the safety science field, although it has been highlighted as an important focus theme. Whereas validation is studied in some subfields of safety science (e.g. safety culture), the literature in the validation of model-based safety analysis, and more specifically in system safety hazard analysis is lacking. This research intends to provide an understanding of the extent of this problem in both academia and industry, to establish a baseline on which future developments can take place. For this reason, first, a literature analysis spanning a decade (2010-2019) was made to assess the state of the practice in validation of model-based safety analysis in sociotechnical systems. Then, the scope will narrow down to system safety hazard analysis methods, the state of the practice in validation of which will be investigated among practitioners through an exploratory investigation. Semi-structured interviews will be performed with system safety experts in safety-critical industries in Canada. Based on the findings of the literature analysis and interviews, and the gained experience from validation literature in other fields of studies, a logical structure will be developed to address the issue of the validity of the STPA hazard analysis method. This validation framework will be integrated into the STPA method and will be called STPA+. It is important to test the assumptions of the developed validation framework and examine how the developed framework will affect the results of STPA. Thus, STPA and STPA+ will be applied to a case study, and their results will be compared.

A Bayesian Network Sub-model for risk-analysis system on oil spill response effectiveness in the Canadian Arctic
Talah Al Sharkawi, MASc Student

There are various marine-related activities in the Arctic such as shipping activities, tourism, fisheries, research, mining, and offshore oil drilling. When focusing on potential oil spills from shipping activities alone, they can have serious negative consequences to marine ecosystems, lead to important economic costs, and have widespread socio-economic, cultural and health impacts. Therefore, determining the efficiency of oil spill responses will help mitigate some of these negative consequences. To do this, a sub-model needs to be created as an analysis support where different spill types, clean-up technologies, human and environmental conditions are considered. Developing a model for emergency response planning for oil spill incidents has a lot of complexity and uncertainty as there are various variables needed to be considered. Some variables include oil spill location, oil spill incidents, and oil spill size. A Bayesian Network Model will be used to aid in the intricacy of oil spill responses by identifying and developing scenarios for planning and seeking to understand vulnerability to potential spill responses. The model will guide in the analysis of various oil spill response equipment efficiency.

Contact Person:
Prof. Dr. Floris Goerlandt
email: floris.goerlandt@dal.ca