Wei Qiu’s research centres around marine hydrodynamics, in the hopes of decreasing noise made by ship propellers, increasing their efficiency as they move through the water, and curbing the negative effects the noise has on marine mammals such as whales. “When ships travel through water, they make noise,” says the department head of ocean and naval architectural engineering in Memorial University’s faculty of engineering and applied science. “As a result of water pressure on the propeller becoming lower than the vapour pressure, propeller cavitation occurs — a phenomenon in which rapid changes of pressure in liquid leads to the formation of small vapour-filled cavities. These vapour-filled cavities, or bubbles, collapse and cause loud popping noises and vibrations, making propeller cavitation the most significant source of noise from a ship. The noise levels jump substantially when propeller cavitation begins. My work is in ship and propeller design, and performance evaluation. I look at how to improve propulsion efficiency – the amount of fuel a vessel consumes — and reduce ship noise by assessing the design and then working to improve it.” The noise issue has traditionally been more critical for navy and research vessels, rather than commercial ones, but in recent years, scientists have discovered the negative effect the noise has on marine mammals, so Qiu’s work also has environmental benefits. “Design is related to the geometry of the ship and how it works with the propeller. If both are designed well, it can help reduce fuel consumption and noise.” Qiu also looks at how ships maneuver in waves, so they don’t capsize, for example. “I want to ensure a ship is designed to operate safely in waves,” he says. In all of his research, the professor often collaborates with industrial partners and government agencies. “I’m working with a propeller manufacturer in Ottawa called Dominis Engineering,” he says. “We’re looking at the effect of manufacturing defects on propeller performance. We started looking into it two years ago and we can see very small defects due to manufacturing can have a pretty big impact on propeller cavitation and therefore noise performance. This partnership helps Dominis manufacture propellers with greater precision.” In his research, Qiu conducts experiments and numerical simulations with computational fluid dynamics (CFD) methods, which he uses to analyze and solve problems involving fluid flows. For his CFD simulations, Qiu relies on Compute Canada and ACENET resources. “We use CFD to simulate the performance of a ship’s propeller,” he says. “This requires pretty extensive computer resources. Without them, I really cannot accomplish my work.” In his own lab, he has a relatively small computer cluster with 700 cores, but through ACENET’s resources, he and his team can access tens of thousands of cores. “For one job, we may need to use hundreds of computer cores for a few days or a few weeks depending on the problem,” he says. “It’s pretty high demand on computer resources. For example, for a ship in waves, we’re dealing with a large area surrounding the ship. That could be 10 times as long as the ship. It implies millions and millions of grids are needed to approximate the domain.”
Showcase Tag: Computational Fluid Dynamics
Predicting the Influence of the Ocean on Machines and Structures
Understanding how machines and structures perform in the ocean has important implications for everything from the way ships manoeuvre to how tidal power turbines respond to massive hydraulic forces. One University of New Brunswick scientist is working hard to increase that understanding and he’s employing some powerful computational tools. Andrew Gerber is a professor of mechanical engineering at UNB. He’s an expert in the field of computational fluid dynamics (CFD) and high performance computing, a discipline that focuses on solving the equations that govern fluid movement using supercomputers. Supercomputers are needed to study in detail how ocean turbulent fluid motion interacts with machines and structures situated within it. Oceans are extremely demanding environments. For man-made systems to survive requires detailed understanding, something that supercomputer simulations can provide. Gerber is involved with a number of important projects including a tidal power study in the Bay of Fundy and a project with Defence Research and Development Canada to study the performance of Canadian naval submarines. “We do a lot of work with submarines,” says Gerber. “We simulate extreme manoeuvres on the computer to see what’s happening to the forces and moments acting on the body of the submarine as it moves through the water. It’s very expensive for the navy to do experiments of this type so our simulations allow them to make predictions without the need for extensive experimentation”. Gerber’s work with tidal power focuses on making detailed fluid flow predictions to support tidal power turbine deployments in the Bay of Fundy. He and his colleagues at UNB are working with the Fundy Ocean Research Centre for Energy (FORCE), and with researchers at Acadia University and Dalhousie University, to measure tidal flow and energy potential in the Minas Passage near Parrsboro and in Grand Passage at Brier Island, Nova Scotia. This work will have a huge impact on where turbine arrays are deployed considering issues such as survivability and maximizing power output. The tidal study involves huge volumes of water requiring complex calculations. Twice each day 10 cubic kilometers of seawater are forced through the Minas Passage – 10 billion tons of water that represent an outflow more than 40 times the amount that flows from the Saint Lawrence River over the same period of time. With technical support from ACENET and funding from the Canada Foundation for Innovation, Gerber set up a “contributed system,” a computer cluster operated by Gerber’s UNB laboratory and managed by ACENET. Under the system, any unused cycles are turned over to ACENET for other researchers to use. The system is powered by the latest Graphical Processing Unit (GPU) hardware, and combined with a CFD simulation software (EXN/Aero) that can efficiently utilize the new hardware, the complex tidal simulations can be completed much more rapidly with high-resolution. In addition to his work at UNB, Gerber is also a partner in a spinoff company called Envenio Inc., a Fredericton-based firm that provides computational fluid mechanics services and engineering software development to engineers and companies. Its flagship product is EXN/Aero, which is specifically designed for next generation hardware. Gerber says that as climate change becomes more prevalent, understanding the movements of fluids in the oceans and atmosphere and their impact on infrastructure will become crucial. “One of the goals of Envenio is to help engineers build better designs for extreme weather events in the ocean or atmosphere and to provide the computational tools to do so.”
Testing Flexible Marine Structures; A Vital Study for the Ocean Industry
For those working in the offshore oil and gas industry, the Deepwater Horizon disaster was a major wakeup call. The tragic explosion, fire and subsequent undersea blowout that took place in the Gulf of Mexico off Louisiana on April 20, 2010 will have profound effect on the worldwide oil and gas industry for years to come, and has industry leaders in the Atlantic region taking a close look at the fail-safes in place in their industry. A new study at Memorial University (MUN) may provide knowledge to help deal with future oil disasters. Erkan Cakir and Anup Radhakrishnan are Master’s students in Engineering at Memorial University. Their project is funded by Newfoundland’s R&D Corporation through an Ignite Project directed by NRC’s Dr. Ayhan Akinturk and ACENET’s Dr. Alejandro Allievi. Using a powerful computer system, the two students designed a series of elaborate tests to investigate the suitability of flexible structures in the harsh marine environment. “If we can better understand how submerged flexible structures work, that could help in places like the Gulf of Mexico,” says Cakir. Cakir and Radhkrishnan designed an experimental rig that could carry flexible structures towed along a wave tank, such as the ones available at MUN or at NRC’s OCRE Institute in St. John’s. By towing the rig along the tank, they could simulate the ocean’s hydrodynamic forces under the effect of currents and waves – a highly complicated task with flexible structures. “Fluid forces are based on the shape of the object,” says Cakir. “But with a flexible structure the shape changes considerably under the effect of underwater environmental forces. You need to compute forces depending on the instantaneous shape of the structure, and that is a highly intensive computational process.” With about six million computation points contained in the computer model of the experimental rig structure that Cakir and Radhkrishnan studied, billions of calculations must be carried out. “I’ve got a $5,000 computer on my desk, but that’s not nearly powerful enough,” says Cakir. The research team turned to the distributed processing power of ACENET’s Placentia Cluster at MUN instead, an invaluable resource, according to Cakir. “What would have taken us months took us a few weeks or in some cases just days.” Allievi is ACENET’s Computational Research Consultant at MUN, providing computational, scientific and engineering advice to Placentia users. For the flexible structures project Allievi provided computational and engineering guidance, including in the use of the sophisticated multi-Physics simulation software called ANSYS. “Computationally, it’s an extremely intensive task,” he says. “Even with the computational resources available at Placentia, it can take as much as a couple of months to complete the simulation of the entire experimental rig.” ACENET provides researchers and innovators with advanced computing expertise along with one of Canada’s most powerful computer networks. The only Atlantic Canada-wide facility that provides such services, ACENET is a collaboration of Atlantic Universities and a partner consortium of Compute Canada, the organization responsible for advanced research computing in Canada. The Placentia Cluster headquartered at Memorial University is the largest ACENET cluster, housing about half of the 7,000-8000 CPU cores available. Allievi says that the research carried out by Cakir and Radhkrishnan has considerable potential for oil and gas subsea exploration that goes well beyond the flexible structures studied in this specific project. Placentia Cluster and ANSYS software can be configured to investigate many other ocean-related problems requiring complex fluid flow and structural analyses. “With the burgeoning offshore industry in Newfoundland and Labrador, there are plenty of projects happening in St John’s that could certainly benefit from our computational/engineering expertise and experience. Enhancement of our resources would most definitely improve our ability to provide quicker turnaround of results for researchers, innovators and the local economy.” Cakir and Radhkrishnan presented preliminary results of their works at Compute Canada’s High Performance Computing Symposium 2014 conference in Halifax in June. Cakir says that it could serve as a model for future studies of this type – studies that may help engineers design underwater flexible devices to contain oil spills. It’s the type of research that is vitally important for an ocean technology industry that already generates more than $2 billion to the Atlantic Canadian economy every year.