Sundeep Sanghavi, DataRPM
While many might assume a multibillion dollar industry would be technologically savvy, the oil and gas industry isn’t up to par. Despite its precious cargo, the majority of pipelines were built more than two decades ago. They still operate using old technology, which makes them expensive to run and maintain. And as the years go on, the pipelines become more susceptible to undetected leakages.
It doesn’t help that the control system built to operate and monitor these pipelines are outdated, too. So when a fault occurs, a huge workforce is required on site to rectify the situation. And if they don’t? Well, we all remember what happened with the Deepwater Horizon and BP oil spill.
Despite the huge risk to the environment, platform workers and oil and gas companies’ bottom lines, leading oil and gas companies have been slow to modernize their maintenance systems. However, luckily, now emerging sensors and data technology is offering the ability to manage oil production and maintenance—and take a smarter, safer approach to pipeline maintenance.
So sit back and relax, oil and gas industry players. We’re about to cover why the time is now to leverage cognitive outcomes for predictive maintenance that matter.
Cognitive can now learn faster, easing implementation concerns.
Artificial intelligence (AI) is getting smarter and at a faster rate than expected. Notably, we’ve seen some huge advancement in deep reinforcement learning—a technique that enables machines to associate a positive outcome with the steps that led up to it, effectively meaning it can “learn” without instruction.
It’s something that would heavily benefit big tech players in their race to implement more advanced deep learning networks—and yes, soon advance predictive pipeline maintenance capabilities as well. Innovation in the AI industry is going at rapid speed, so it’s best to invest in the technology now to get a head start.
Dan Walker, who leads the emerging technology team in BP’s Group Technology, says that AI is helping to optimize well design and specify procedures to ensure that every well is drilled as efficiently and safely as possible.
“It could help us to improve equipment reliability and predict maintenance requirements of our facilities,” he said.
Technology implementation costs are decreasing.
Innovation democratizes access to technology. So when large corporations expand their AI and machine learning efforts, it’s good news for us all.
We’ve seen this happen in a number of instances. Chinese company Baidu opened a U.S.-based AI lab, as did Didi Chuxing, China’s version of Uber. Didi Chuxing’s also raised $5.5 billion for AI research. Likewise, the Chinese government has pledged to invest $15 billion in research by 2018. Even more, Google has open-sourced its AI library TensorFlow to drive innovation forward.
Now that AI has been popularized by large corporations, the technology is set for larger use—particularly by startups that offer cheaper solutions. Take Ufora, for example. The company helps hedge funds and other players on Wall Street to run complex data models. And while this would have required $1 million in hardware and taken months to complete just five years ago, it can now be done with Amazon Web Services in the time it takes to make a cup of coffee.
Prominent oil and gas industry players are innovating with AI, as well. Just last year, oil producer Gazprom Neft partnered with IBM to develop strategies that use cognitive data analysis, machine learning and high-performance computing. This would optimize processes such as geological prospecting, project management, field development and operation.
With oil and gas prices low, there is a need for savings across the board.
Low oil prices are the new norm. While oil recovered from below $30 a barrel in early 2016 to more than $50 at the end of the year, the price of oil has now fallen more than 20% in 2017 alone. This leads to tons of volatility in the industry.
However although it may seem counterproductive to invest in sophisticated technology and sensors in times like these, writes IBM's Marceli Bouzein, doing so can secure a long-term future. It’s “spending money to save money,” he says.
It’s true that many in the industry have gotten as far as incorporating these sensors—the average offshore production platform has 40,000 data tags. However, not everyone has invested in putting the insights to use. The data comes from expansive well sites, large pipeline projects, high-tech equipment and gas-gathering systems—and flows right into outdated legacy systems. Likewise, oftentimes the data is stored in unstructured forms, making it difficult for computers to understand.
Investing in cognitive computing gives industry players access to this goldmine of data, enabling them to make more data-driven insights and save money across the board. According to The Oil and Gas Technology Centre, data optimization can push plant performance beyond 95%, increase production by 2% to 5%, improve efficiency and thus, reduce costs by 10%–and perhaps most importantly, it can help prevent disasters like oil spills or fires.
With improved technology, more innovation and an increasing need to cut back on spending, there’s never been a better time to invest in cognitive predictive maintenance. So oil and gas industry players, what are you waiting for?
Sundeep Sanghavi is a cognitive computing veteran and Industrial IoT pioneer based in Washington DC. He is CEO and co-founder of DataRPM, a Progress company. It was acquired for $30 million by the NASDAQ listed software giant in March.
Gas supplies have been tight in Southern California for years due to the pipeline limitations and reduced availability of the utility's biggest storage field at Aliso Canyon.
Cognitive predictive maintenance is a way to leverage real-time data to take a smarter, safer approach to service and repairs.
Approach improves priorities without calling for compromises.