10 Sep Pipe Dream – Digital Innovation in the Pipeline Sector
Pipeline companies lag other network asset operators in adopting the digital innovations that lead to greater autonomous operations. That could change.
I was researching for my upcoming presentation to the Canadian Energy Pipeline Association Foundation on artificial intelligence and machine learning, when I came across a press announcement about a novel industrial robot.
With the inaugural voyage of their new driverless rail operation in Western Australia, Rio Tinto issued a statement laying claim (temporarily) to the title of owning and operating the world’s largest robot. I say temporarily, because a rail authority in China has announced that they intend to build the world’s first driverless bullet train. Given the size of the China bullet train network, now closing in at 17,000 kilometers, they will handily and permanently claim the title when they deliver.
Perfectionists might observe that the actual robot part of the Rio system is in fact just the locomotives—they pull cars loaded with iron ore to port. The rails are, of course, stationary, and probably not internet enabled. But let’s go with the definition as laid down by Rio—a large fixed-in-place network asset that operates autonomously.
Pipelines partly fit this definition—like rails, once laid down, pipes don’t move (or shouldn’t). And pipelines can be pretty large network assets. The Interprovincial Pipeline System holds the Guinness World Record for length, at 3787 kilometers. The US has the longest pipeline system in the world, at some 402,300 kilometers, followed up by Russia and Canada. The pipeline equivalent to a locomotive is the compression station that pushes the product (oil or gas) from terminal to terminal.
Rio is suddenly a market leader in autonomous operations. Their autonomous rail complements their autonomous heavy haulers, and I suspect we will soon see autonomous port operations, shipping and processing. Syama is another digital innovator, who aim to launch the first fully autonomous underground mine, in Africa.
With digital innovation arriving at such traditional and heavily unionized industries as mining and rail, I wondered how the pipeline industry is responding to digital transformation. To get at this puzzle, I reviewed investor presentations and websites for 50 different Canadian and US pipeline companies, including a more thorough review of some of the largest North American players. I conducted an extensive search of media and published news stories about digital innovation in the midstream. Finally, I interviewed investment analysts who study the midstream industry and provide market guidance, profit analysis and price recommendations.
My conclusion is that, with few exceptions, there is little published evidence of efforts to digitalize operations in the same manner and to the same degree as mining and rail. Smart pigs (sensor equipped devices that flow through a pipeline in operations looking for cracks, dents and erosion) get a lot of mention, but not much else. There is little overt discussion of drones, augmented reality, artificial intelligence, machine learning, blockchain, cloud computing and agile business methods. I was unable to find a public midstream company that deliberately chooses to differentiate itself from its peers via digital innovation.
Getting the conditions right
Why has the mining industry jumped out in front of the resource sector in the drive to adopt digital innovation? Simply put, they needed to. Rio faced a compelling set of conditions that made investment in autonomous rail and heavy haulers worthy of spending shareholder capital.
Demanding work conditions
Rio’s iron ore mines are located in far off Western Australia. It’s said that Western Australia is not the end of the earth, but it’s probably true you can see the end of the earth from Western Australia. It’s hot, remote, dry and rugged. It doesn’t appeal to everyone. Canada’s oil sands mines are equally if not more demanding.
I’ve spent years working in Australia, and it’s a labour constrained market. Mining is some 20% of Australian GDP, and not that many Australians want to work in the mining sector. It’s tough to attract talent and with the skills the industry needs.
High cost labour
Mining workers have commanded very good wages for their services, and deservedly so. Oil and gas labour rates have also been among the highest of any job class for years.
Cost of human-centric assets
Assets that rely on a lot of people to run are going to be costly. Humans need basics like food, shelter and rest facilities. The assets need human centric controls, like seats and steering wheels, and safety features like harnesses, lighting and signals. Add in the cost of supervision, multiple shifts, transportation to and from work, training and coaching. People-dependent business models are very expensive, and these costs don’t go down.
Rio’s product, iron ore, behaves like any other commodity. The global market sets the price, and companies compete on tonnage produced for each dollar of invested asset, and the cost to produce that volume. When commodity prices fall, as they did in 2012 for mining, and for oil and gas in 2014, there’s strong incentive to remove cost from the business. People costs are incredibly sticky, but Rio believed that the technologies were ready to tackle these structural costs.
Rising operating cost
I’m speculating, but it’s a good bet that Rio’s mines are like a lot of mines around the world. When mines start up, the seams with the least amount of contaminant, and the highest concentration of the best ore, are produced first. This helps pay off the huge up front capital investment. As time goes on, the distance from the face of the mine to the port gets longer, more overburden needs to be removed, the pits deepen, and so on. Operating costs per ton go up.
These same conditions are also motivating Suncor, CNRL and other oil sands mining companies to invest in digital innovation. The technologies that power robotic rail—low cost sensors, enormous amounts of data, advances in artificial intelligence and machine learning, digitally controlled actuators, highly reliable and high bandwidth telecommunications, smart people—are equally available to all.
Pipeline Conditions Are Similar, But Not Quite
Pipelines have many of the same competitive pressures as Rio’s rail. The work to keep the assets running is demanding, labour rates are high, and competitive pressures abound. But there are big differences. Here’s just a few.
Structural disincentives to innovate
Many pipelines are set up as natural monopolies. As monopolies, there’s little incentive to innovate, even when it would result in significant business improvements. The gains from innovation may not accrue to the shareholders, and the costs of innovation may be handed to the customer as a toll increase.
Pipeline customers will often configure their businesses around the vagaries of pipeline design (to take advantage of bottlenecks, for example), and they object to changes that are detrimental. Years ago, I helped a pipeline model out the impact of making a change to its asset mix using game theory. We discovered that the intended asset would make money for shareholders, but would principally benefit all of the marginal customers of the pipeline and penalize the largest and most reliable customer. Needless to say, the investment didn’t proceed.
Infrequent asset turn arounds
Unlike a locomotive asset, which can be taken out of service for repairs and upgrades without incurring a huge network impact, pipelines must generally remain in service 24/7. Pipeline assets are simply going to take longer to upgrade, and when upgrades do take place, emphasis will likely be to improve reliability or operating performance, rather than digitalization. Of late, pipelines have focused on finding and addressing micro fractures and other wear and tear issues as a priority.
Restrictive commercial structures
To secure the capital to build these long life assets, pipelines need the commercial structures in place that lock in volumes and prices for years. These commercial structures are then hard to alter once in place. Rateable schedules are painful to design. Network changes that require coordination across multiple shippers, refineries and utilities (akin to Rio’s autonomous haulers feeding its autonomous rail) will run headlong into commercial formulas that will resist change.
I see three pathways that pipelines could follow to advance their level of digital innovation.
Shift some digital innovation to the corporate center, where it can be more profitably directed for the benefit of the group. I like how Google restructured to create Alphabet, the new corporate business to hold its investments and innovations in home automation, search, urban innovation, moonshots, drone developments, infrastructure, and artificial intelligence. These structures accelerate digital innovation and adoption.
Focus on greenfields
Set the bar on the next greenfield project to achieve totally lights out operations, a 2 times improvement in build cycle time, and zero emissions. Pipelines build new infrastructure on a regular basis, and it’s much easier to design digital into a new pipeline than it is to layer digital innovation as an afterthought or to overhaul a brownfield asset.
Seek new suppliers
Shake things up by inviting an unconventional supplier to challenge the status quo. Traditional technology suppliers strive to capture their clients technology spend, block third party innovation and inhibit integration with others. Exxon Mobil invited a defence contractor to help design its next refinery for this reason.
Large Robots – They’re Coming
What is Rio getting from its large robot? Everything you could possibly want—no bathroom breaks, no shift changes, no errant behaviour, high reliability, optimal energy use, higher overall asset utilization and lower cost.
Pipelines can get the same benefits and with just a few changes, could become digital leaders.