Stream Systems’ simulation software opens new doors to oil and gas business operations modelling

Image: Stream Systems

It was a simple enough announcement in the spring of 2015: two kingpins of the Canadian midstream industry—Keyera, which provides gathering, processing, storage and transportation for producers, and Kinder Morgan, North America’s largest energy infrastructure company with 84,000 miles of pipe and 180 terminals—are joining forces to build the $672-million Baseline Terminal near Edmonton.

The terminal will be developed in phases, with the first tanks expected to be commissioned in the second half of next year. When the full initial stage is complete, it will consist of 12 above-ground tanks with a total capacity of 4.8 million barrels of crude oil and natural gas liquids. There will also be room on the site adjacent to Keyera’s Alberta EnviroFuels complex for another 1.9 million barrels of storage capacity, pending commercial commitments.

It will be connected via pipeline to Kinder Morgan’s Edmonton terminals and will be capable of sourcing all crude streams handled by Kinder Morgan for delivery to multiple destinations, including its Trans Mountain Pipeline and two Edmonton rail terminals.

Dealing with complexity

But it’s the complexity of the project—a greenfield merchant oil terminal nested within two existing terminals that are already bound by a web of connections to feeder pipelines, export pipelines, meter manifolds, an oil products terminal storage facility, railcar transloading and delivery facilities, and several long pipelines connecting all three terminals—that had engineers at both Keyera and Kinder Morgan searching for a way to validate their concept for the terminal. That’s when they called in the simulation experts at Stream Systems, which puts massive computing power to work solving complex integrated business modelling problems.

The initial plan by the joint-venture partners was to save costs by using shared lines instead of dedicated lines to link the various storage tanks, but since Kinder Morgan had no operational experience with this type of arrangement, it wasn’t sure it could meet throughput commitments while maintaining product quality—keeping heavy sour crude apart from light sweet streams.

“They weren’t able to get the answers they wanted,” Stream president and chief executive officer Allan Chegus explains. “The operations people said, ‘We don’t think we can run our customer contracts effectively with the current design but don’t know what to recommend to engineering.’ They came to us and asked if we could help. We said we could build a prototype—a small model—to show what’s possible operationally.”

Kevin MacFarlane, director of shipper services for Kinder Morgan Canada, says the partners in the Baseline Terminal project were interested in finding a software company that could do simulations of the rather complex design they had in mind for the project. Stream, he says, fit the bill perfectly.

“What we tried to do through our engineering process was streamline the costs of this project, and that meant reducing the amount of assets deployed and sharing assets and things like that,” he says. “We understood it would work; it really was a question of how much throughput could we achieve through these reduced assets.”

Within about three weeks, Stream had a working model. In another week or so, engineering and operations people at Keyera and Kinder Morgan were getting the answers they needed. Then there were a few months of testing and validating the simulation outputs and running new scenarios—not an easy process when you’re dealing with a greenfield project with no real-world statistics to back up the simulation results, Chegus says.

“We were able to show, statistically and mathematically, that what the simulation was showing was valid, and once we got beyond that hurdle, it was like the whole world opened up for them,” he says. “The end result is they got the project approved because of the work we did for them and saved a fair amount in capital costs in the process.”

MacFarlane says the validation process took some time largely because the project partners were working with a new suite of software. They needed to take the time to become certain that the software was doing the right calculations and working with appropriate assumptions.

“When you’re spending nearly a billion dollars on a project, you really need to do your due diligence,” he says. “Working with the software is part of that, and becoming comfortable is part of that.”

Stream has been providing simulation services since 2014, and for the most part, Chegus and his team follow a relatively traditional methodology themselves: clients come to them with a problem, they define the problem and design a simulation model to fit, then the client is brought in for various reviews and demonstrations along the way, and eventually the model is completed.

After that, the client is provided the opportunity to play with a stand-alone model on its own internal computer network—try out various scenarios, enter different inputs, take it for a spin. Or, the client can have Stream do all that and generate a report in a few weeks or months with recommended courses of action.

Working in the cloud

But all that is about to change. As one of the companies attached to GE’s Customer Innovation Centre’s Zone Startups Calgary, Stream will be launching its Software as a Service (SaaS) application this fall on Predix, GE’s Industrial Internet of Things (IIoT) platform. Once the simulation application is on Predix, everything that Stream has been able to do for its clients as a contracted service will be available to anyone, anywhere in the world, on their own desktop computers or even on their tablets. The drag-and-drop interface allows users to model, simulate and optimize business problems in a risk-free virtual world, and because it will reside on the IIoT, it will allow teams from around the world to collaborate on a single project or on any number of projects.

“It is democratizing the technology,” Chegus says. “That is a lot of what is going on in the market space today. All of the old revenue models, all of the old distribution models are breaking down or are gone altogether. This is no different—this is getting technology into the hands of people who need it and making it cost-effective, making it easy and intuitive and geographically distributed.”

Headquarters might have a stand-alone modelling program, but it can only be used in-house. Workers at other locations don’t have access. With the SaaS application, the same technology is available to anyone in the company’s network, offering a flexible, interactive tool that is accessible on a subscription basis to anyone, Chegus says.

“You have a Visa card? You need it for the month, that’s great. Everybody around the world can download the same application, do what they want, work together. You can share data back and forth. It’s really the new way of working together and leveraging all of the networking,” he says.

Cloud-sharing of simulation options is the coming “innovation in innovation,” says Birgit Juergensen, who came to Stream a year ago as vice-president, operations, after 10 years working with refinery optimization models throughout Europe and another decade with Shell in its supply chain optimization group. She emigrated to Canada in 2012 and joined Enbridge as a senior petroleum quality specialist, where she introduced a rule-based model for monthly crude oil nominations, before joining Stream 18 months ago.

Enabling collaboration

“You can’t protect your software product anymore—the platforms will come and eat your lunch, so you better be part of the platform, especially if your solution is like ours,” she says. “We want to address the whole company, but we want certain teams to work together, so we [will] have it in the cloud and actually have people collaborate. With our core technology, we let other people build on it and create demand from there.”

Kinder Morgan’s MacFarlane says the cloud-based option should prove valuable in the future as it offers the potential to take much of the back and forth out of the process. In the Baseline Terminal project, much time was spent between Stream and the project engineers working out details, testing assumptions, validating inputs. Working in a virtual environment would certainly be more efficient from that perspective.

“You can imagine having an entire team doing some engineering. There’s time pressure on getting these projects launched, and so you can’t take forever to make some of these decisions,” he notes. “The closer it is to the user [being] able to run a bunch of different simulations without needing help from a programmer is of huge benefit.”

The cloud-based solution will use three layers of simulation, Chegus says. The first layer is mainly discrete simulation—the kind that’s been done for years using Excel spreadsheets. That’s a linear simulation: stable enough, but very static and quite limited in the answers it can provide.

On top of that is added more complex simulations like fluid dynamics, which takes into account how fluids move through pipes and manifolds and tanks. Again, discrete simulation by a hydraulic engineer would answer certain questions, but it would take weeks to do those simulations.

“With our simulation model, we can watch how a batch moves through a pipeline or in and out of a tank, so we can recalculate the composition of the tank or the pipe to optimize how a particular fluid would move in it,” Juergensen explains.

The third layer—and this is what sets Stream’s simulation software apart from the crowd and opens up so many potential applications outside the midstream sector—is agent-based modelling. As Chegus explains it, agent-based modelling offers the ability for a discrete object to be self-aware.

“In other words, you can take a drop of oil, drop it in a pipeline here and have it get to the Gulf Coast,” Chegus explains. “You don’t care how it gets to the Gulf Coast; you tell it you want it to get from here to there, and it will decide whether to go to the left pipeline, the right pipeline, it doesn’t matter. The [software] can now decide what’s the most efficient way to get that drop of oil from Edmonton to the Gulf Coast.”

Agent-based modelling balances a host of factors—power consumption, quality issues, demand for that drop of oil on the Gulf Coast versus, say, the Chicago refining district—to decide what is the best route and, in fact, the best destination if that’s the desired outcome.

Juergensen says one of the advantages of incorporating agent-based modelling into the simulation software is that it removes the need to worry about complexity. As long as the rules are known—what crudes can move in what pipelines, the batch specifications for each stream—then running a simulation and dealing with the outcomes become much easier.

“In the old rigid models, you might end up with hundreds of thousands of potential routes getting from here to the Gulf Coast because there are so many options, but with this technology, as long as I know what is allowed where, then it’s easy to maintain. If you have another commodity, you just add it in and [the software] does the rest,” she says.

External factors can also be incorporated into the simulation, Chegus says, and are limited only by your imagination. Weather impacts, arbitrage opportunities—crude netbacks on the Gulf Coast versus netbacks in Chicago, for example—demand patterns in Asia versus Europe, drilling activity in western Canada and the impact that might have on future production levels can all be factored in and form part of the solution.

Drag and drop

The actual cloud-based application, Chegus says, will be as easy to operate as any of the multiplayer games available on the Internet. It will boast a full library of components—tanks, pumps, pipe in various diameters, meters, manifolds, compressors and the like—that users will be able to incorporate or switch out at their pleasure.

“You’ll just drag and drop, drag and drop,” he explains. “You can assemble a template in minutes, where before with spreadsheets, you would have spent hours putting a project together.”

At the end of the day, the user can either look at the model and admire what it’s done or split everything into a database and really start to dig into what’s been created, Juergensen says.

“We can preconfigure various analyses, compare cases and see how things have changed,” she says. “What’s the impact on quality? What’s the impact on throughput? Compare and contrast, see what’s feasible, can I accept this one over that one, does it hit the throughput I want to have.”

And it can be easily integrated into a computer-aided design (CAD) environment: optimize the project using the modelling software then feed the result into the CAD application to actually design the plant down to every nut, bolt and flange required to make it work.

Juergensen says a couple of challenges still stand in the way of broad adoption of cloud-based modelling of the kind Stream will offer. One is that many companies—especially in oil and gas—are working with relatively dated desktop computers. The other is that they must train staff to use the software, and day work keeps getting in the way of that.

“It’s a matter of making time,” she says. “We have trained [people] to do the analysis, and they are very happy with it, but then they have day-to-day stuff, so they keep coming back to us to do studies. But ideally, we would actually like the companies to do it in-house because that keeps the trust level up.”

While Stream has designed the application initially for its “sweet spot” in the oil and gas midstream, the principals of the simulation software can be applied in virtually any industry, in virtually any business situation. SAGD projects in Alberta’s oilsands, mass transit in any major city in the world, mining, municipal infrastructure—all can benefit.

“What’s stopping us?” Chegus asks. “As Canadians, we’re so humble, why can’t we change the world? Why can’t we transform things? Why not?”

Like this? Check out the latest edition of Oilweek.