人工智能/机器学习

石油行业数字化时代仍需人性化

该行业在压榨石油产量的同时,也在平衡大脑和机器人。

ALCE-世代-面板.JPG
从左至右依次为:马拉松公司的主持人 Wilfried Manfoumbi、Camilleri & Associates SAS 的 Lawrence Camilleri、哈里伯顿公司的 Frank Corredor、Oxy 公司的 Courtney Richardson 和雪佛龙公司的 Amine Zejli。
来源:Jennifer Pallanich/JPT

数字技术所能做的事情比以往任何时候都多,但人类的脑力仍然发挥着重要影响。

8 月 20 日,在SPE 人工举升会议和展览会的小组讨论会上,专家们指出,数字技术正在迅速发展,业界正在寻找利用这些进步来生产更多碳氢化合物的方法。

首先,不同的数字工具已经达到了不同的成熟度。

Oxy 公司 QA/QC 的 RCM 人工举升部门负责人考特尼·理查森 (Courtney Richardson) 表示,预测设计软件为设计工作提供了基本基础,尽管仍然需要一些实际操作,但自动验证的井测试已经取得了进一步的进展。

她说:“我们已经有了这些智能流程,可以自动为我们大批量地验证井测试,这为我们节省了一些时间去做其他事情,比如去寻找桶。”

然而,她指出,随着行业实施智能技术和算法,“了解其背后的逻辑非常重要”。

她说,如果不理解底层逻辑,“你很可能会自我验证”导致生产下滑。“当我们利用这些技术时,理解其中的逻辑并围绕所有可能的情况进行构建是非常重要的。”

雪佛龙技术中心高级石油工程师阿明·泽吉利 (Amine Zejli) 表示,雪佛龙的数字解决方案一度受到学科的限制。

“现场操作员、办公室工程师、设施、工程师、资产管理所使用的大量数据和大量工作流程都是不同的。因此,他们对现场发生的事情的看法略有不同,”他说。

这样做的一个问题是,每当雪佛龙工程师创建一种创新工具或工作流程时,它都无法在整个企业范围内轻松推广。另一方面,他说,来自外部供应商的工具的支持和许可成本可能很高。

为了解决这个问题,雪佛龙基于外部供应商的平台创建了自己的系统,该系统已经取代了雪佛龙多个油田和业务部门的旧系统。

“我们还设立了一个中央组织和多学科组织来支持、维持和发展该解决方案,”Zejli 表示,并指出该解决方案支持自动化和实时优化。“当你有大量油井由少数工程师管理时,你必须实施自动化。”

他说,该平台的建模工具堆栈允许工程师构建代表整个价值链的油藏和全油田模型。另一个工具堆栈,即工程协调器,允许工程师构建自己的工作流程,而数据管理器工具堆栈则可汇总整个企业的数据。

他说:“我们已经能够让工程师构建他们自己的模型,构建他们需要的工作流程,或者与中心合作构建它们,构建他们自己的仪表板,或者再次与中心合作构建并满足该特定领域的工程需求,并同时重用来自其他领域的任何功能。”

他说,这些工作流程可以经过审查流程在整个公司范围内扩展,从而在过多控制和无控制之间取得平衡。

“你不会真的想束缚工程师。他们会想办法建造他们想建造的东西,对吧?”他问道。“我们试着找到一个中间立场。”

工作流程需要数据, 哈里伯顿人工举升 Intelevate 经理 Frank Corredor 表示,技术使得获取数据比以往任何时候都更加容易。

他说道:“物联网 (oT) 设备正在帮助我们增强在现场捕获更多数据的方式。”

Camilleri & Associates 首席执行官 Lawrence Camilleri 表示,实时数据可用于三步计划以提高产量。首先,他说,数据可以清楚地表明情况。

“速率和压力是油井的经度和纬度。这告诉我们我们在哪里,”他说,并指出下一步是了解油井的潜力。“一旦我们有了速率和压力,就可以轻而易举地描述油井的流入情况并确定油井的全部潜力。”

一旦了解了潜力,就该规划出油井的去向以及如何到达那里。

他说:“实时数据是关键,我们已经在很多形式的人工举升中看到了它。”

除此之外,卡米莱里还看到了实时数据有助于电力优化的潜力。

他说道:“有了实时数据,你就可以持续地实现自动化。”

他指出,降低功率可以减少运营成本,同时延长设备的使用寿命。

他说道:“如果你优化了效率,那么设备磨损所消耗的电力就会减少。”

寻找新人才

随着获得石油工程学位的毕业生数量持续下降,行业领导者正在寻找鼓励人们进入该行业的方法。

理查森表示,如果她还处于职业生涯的早期,她会被专注于投资碳捕获和低碳技术的公司所吸引。第二个吸引她的是涉及人工智能和机器学习的工作。

她说,年轻一代“在数字时代长大”,很可能会倾向于使用这些技术来工作。

同时,她强调行业从业人员打下扎实的专业基础非常重要。

她说:“我们必须通过这些方式吸引他们进入这个行业,但我们还必须确保增长他们的知识并为他们打下良好的基础。”

泽利说,石油行业可以通过帮助其他学科的工程师学习石油行业来弥补这一差距,但该行业的周期性往往会对其不利。

他说道:“职业稳定性是很多工程师、很多年轻学生最关心的问题。”

他在招聘活动中做的一件事就是强调该行业的规模。

他说道:“我会拿来这样一个趣闻:你知道,你建立了一个工作流工具、一个流程,将其部署到一个日产 12 万桶石油的油田,你将获得 3% 的增幅,你将独自帮助在一年内实现日产 100 万桶石油的产量。”

Zejli 将这份工作与他之前在一家科技公司担任软件工程师的工作进行了比较,“开发一个可能吸引一些用户的小部件,也可能没有,或者以编程方式移动像素。影响并不大。所以,我试着关注影响,关注问题的规模,”他说。“这引起了很大的共鸣。”

科雷多表示,实习可以帮助学生更好地了解石油和天然气行业的机遇。

他说道:“这对我们来说不只是在早期阶段吸引人才的好方法,也是在大学里播下种子的好方法,因为实习结束后,他们会回去向朋友们讲述他们的经历。”

卡米莱里表示,行业内采取更全面的方法可能会鼓励更多的人进入该行业。

“我对未来的愿景是,我们不会只是人工举升工程师。我们不会是生产工程师。我们不会是油藏工程师。这有点像约翰·列侬的愿景,想象世界是一体的,工程师认为数学和物理的应用没有限制,这将吸引更多的年轻人,吸引更广泛领域的人才,”他说。

原文链接/JPT
AI/machine learning

The Human Touch Still Needed in Oil Industry's Digital Age

The industry is balancing brains and bots as it squeezes out barrels of oil production.

ALCE-Generations-Panel.JPG
From left, moderator Wilfried Manfoumbi of Marathon, Lawrence Camilleri of Camilleri & Associates SAS, Frank Corredor of Halliburton, Courtney Richardson of Oxy, and Amine Zejli of Chevron.
Source: Jennifer Pallanich/JPT

Digital technology can do more than ever before, but human brain power still brings something important to the table.

Speaking during a panel session at SPE’s Artificial Lift Conference and Exhibition on 20 August, experts noted that digital tech is rapidly evolving and that the industry is finding ways to harness such advances to produce more hydrocarbons.

For starters, different digital tools have reached different levels of maturity.

Courtney Richardson, Oxy’s lead for QA/QC–ORCM artificial lift, said predictive design software provides a fundamental base for design work although it still requires some hands-on effort, but auto-validated well testing is further along.

“We've got these smart processes in place that can auto-validate well tests for us in large batches, and that frees up some time for us to do other things, like go hunt barrels,” she said.

She noted, however, that as the industry implements smart technologies and algorithms, “it's incredibly important for you to understand the logic behind it.”

Without understanding the underlying logic, she said, “you can very well auto-validate yourself” into production decline. “It's just very important that as we leverage these technologies that we also understand the logic and we build it around all the possible scenarios.”

Amine Zejli, senior petroleum engineer at Chevron Technical Center, said that at one time, digital solutions at Chevron were fairly well siloed by discipline.

“A lot of the data, a lot of the workflows that were used by the operators out in the field, by engineers in the office, the facility, engineers, asset management, were all different. So, they all had slightly different views of what's going on in the field,” he said.

One problem with that was that whenever a Chevron engineer created an innovative tool or workflow, it couldn’t easily be scaled across the enterprise. On the other hand, support and licensing costs for tools from external vendors could be significant, he said.

To solve this issue, Chevron created its own system based on a platform from an external vendor, and the system has replaced legacy systems in a number of Chevron fields and business units.

“We also set up a central organization and multidisciplinary organization to support, sustain, and grow the solution,” Zejli said, noting it supports automation and real-time optimization. “When you have a large number of wells being managed by a small number of engineers, you have to have automation in place.”

The platform’s modeling tool stack allows engineers to build reservoir and full-field models that represent the entire value chain, he said. Another tool stack, the engineering orchestrator, allows engineers to build their own workflows, while the data manager tool stack aggregates data across the enterprise.

“We have been able to enable engineers to build their own models, build their own workflows that they need to, or collaborate with the center to build them, build their own dashboard, or, again, collaborate with the center to build up and meet the engineering needs for that specific field, and at the same time reuse any capabilities” from other fields, he said.

Those workflows can be scaled across the company following a review process, providing a balance between too many controls and no controls, he said.

“You don't really want to handcuff engineers. They will figure out a way to build what they want to build, right?” he asked. “We try to find a middle ground.”

Workflows need data, and Frank Corredor, artificial lift Intelevate manager at Halliburton, said technology has made acquiring data easier than ever.

“IoT (Internet of Things) devices are helping us to enhance how we're capturing more data in the field,” he said.

Lawrence Camilleri, CEO of Camilleri & Associates, said real-time data could be used in a three-step plan to improve production. First, he said, data makes it clear where things stand.

“Rates and pressure are the longitude and latitude of the well. It tells us where we are,” he said, noting the next step is understanding the well’s potential. “And once we have rates and pressure, it's a walk in the park to characterize the inflow of the well and identify the full potential of the well.”

Once the potential is understood, it’s time to map out where the well should go and how it can get there.

“Real-time data is key, and we already see it in a lot of forms of artificial lift,” he said.

Beyond that, Camilleri sees potential for real-time data to help with power optimization.

“With the availability of real-time data, you can automate it on a continuous basis,” he said.

Reducing the power reduces operating costs while increasing the run life of equipment, he noted.

“If you optimize the efficiency, there's less power spent wearing out equipment,” he said.

Finding New Talent

As the number of people graduating with petroleum engineering degrees continues to decline, industry leaders are looking for ways to encourage people to enter the industry.

Richardson said if she were early in her career, she would be drawn by companies focused on investing in carbon capture and low-carbon technologies. A second draw would be in the form of work that involves artificial intelligence and machine learning.

The younger generation “grew up in the digital age” and would likely gravitate to work using those technologies, she said.

At the same time, she stressed the importance for the industry’s workforce to have a solid foundation in the fundamentals of their disciplines.

“We've got to draw them into the industry on those things, but we also have to make sure that we are growing their knowledge and giving them a good foundational base,” she said.

Zejli said the industry can bridge the gap by helping engineers from other disciplines learn the petroleum industry, but that the industry’s cyclicality often works against it.

“Stability of the career is a major concern for a lot of engineers, a lot of young students,” he said.

One thing he does during recruiting events is highlight the scale of what the industry does.

“I would anecdotally say that, you know, you build a workflow tool, a process, deploy it to a field that produces 120,000 barrels, and you get 3% uplift, and you will single- handedly have helped bring a million barrels online over a year,” he said.

Zejli compared that to one of his previous jobs as a software engineer at a tech company, “working on a widget that may get some users, or it may not, or programmatically moving pixels around. The impact is not there. So, I try to focus on the impact, the size of the problem,” he said. “That resonates a lot.”

Corredor said internships can help students gain a better understanding of the opportunities that exist in the oil and gas industry.

“It's a good way for us not just to capture talent in an early stage, but also to start planting the seed at universities, because after the internship, they're going back and they're going to start telling their friends about their experience,” he said.

Camilleri said a more holistic approach within the industry might encourage more to enter it.

“My vision of the future is we won't just be artificial lift engineers. We won't be production engineers. We won't be reservoir engineers. A bit like John Lennon's vision, imagining the world altogether, acting as one, and the engineer sees no limits to his application of maths and physics and that would attract more young people and would attract people from the broader spectrum,” he said.